1 2013-11-18 Jan Hubicka <jh@suse.cz>
3 * profile.c (compute_branch_probabilities): Do not sanity check run_max.
5 2013-11-18 Kenneth Zadeck <zadeck@naturalbridge.com>
7 * tree.c (int_fits_type_p): Change GET_MODE_BITSIZE to
9 * fold-const.c (fold_single_bit_test_into_sign_test)
10 (fold_binary_loc): Change GET_MODE_BITSIZE to
13 2013-11-18 Teresa Johnson <tejohnson@google.com>
15 * gcc/cfgrtl.c (cfg_layout_initialize): Assert if we
16 try to go into cfglayout after bb reordering.
17 * gcc/passes.def: Move compgotos before bb reordering
18 since it goes into cfglayout.
20 2013-11-18 Bernd Schmidt <bernds@codesourcery.com>
22 * cgraphunit.c (ipa_passes): Don't execute all_lto_gen_passes.
23 * lto-streamer-out.c (lto_output, produce_asm_for_decls): No longer
25 (pass_data_ipa_lto_gimple_out, pass_ipa_lto_gimple_out,
26 make_pass_ipa_lto_gimple_out, pass_data_ipa_lto_finish_out,
27 pass_ipa_lto_finish_out, make_pass_ipa_lto_finish_out): Remove.
28 * lto-streamer.h (lto_output, produce_asm_for_decls): Declare.
29 * pass-manager.h (GCC_PASS_LISTS, class pass_manager):
30 Remove all_lto_gen_passes.
31 * passes.c (pass_manager::dump_passes): Remove its use.
32 (pass_manager::register_pass): Likewise.
33 (ipa_read_summaries, ipa_read_optimization_summaries): Likewise.
34 (pass_manager::pass_manager): Don't initialize or use it.
35 (write_lto): New static function.
36 (ipa_write_summaries_1, ipa_write_optimization_summaries): Use it
37 instead of using all_lto_gen_passes.
38 * passes.def (all_lto_gen_passes, pass_ipa_lto_gimple_out,
39 pass_ipa_lto_finish_out): Delete.
40 * tree-pass.h (make_pass_ipa_lto_gimple_out,
41 make_pass_ipa_lto_finish_out): Don't declare.
43 2013-11-18 Jeff Law <law@redhat.com>
45 * tree-ssa-threadupdate.c (redirection_data): Record two
46 duplicated blocks instead of just one.
47 (local_info): Explain why we don't create a template for the
48 second duplicated block in a thread path.
49 (create_block_for_threading): Accept argument indicating array
50 index into redirection_data to store its result.
51 (lookup_redirection_data): Initialize both duplicate blocks.
52 (ssa_create_duplicates): If a jump threading path needs multiple
53 blocks duplicated, then duplicate them.
54 (ssa_fix_duplicate_block_edges): Corresponding changes.
55 (ssa_fixup_template_block, thread_single_edge): Likewise.
57 2013-11-18 Marek Polacek <polacek@redhat.com>
59 * doc/invoke.texi: Extend -fsanitize=undefined documentation.
61 2013-11-18 Andrew Pinski <apinski@cavium.com>
62 Steve Ellcey <sellcey@mips.com>
65 * config/mips/mips.md (*mov<GPR:mode>_on_<MOVECC:mode>): Remove
66 type restriction from equality_operator on conditonal move.
67 (*mov<SCALARF:mode>_on_<MOVECC:mode>): Ditto.
68 (*mov<GPR:mode>_on_<GPR2:mode>_ne): New.
70 2013-11-18 Jeff Law <law@redhat.com>
72 * tree-ssa-threadupdate.c: Fix file block comment.
73 Fix minor indention issue.
75 2013-11-18 Uros Bizjak <ubizjak@gmail.com>
77 * config/i386/i386.c (ix86_decompose_address): Use REG_P instead of
78 ix86_address_subreg_operand. Move subreg checks to
79 ix86_validate_address_register. Move address override check to
80 ix86_legitimate_address_p.
81 (ix86_validate_address_register): New function.
82 (ix86_legitimate_address_p): Call ix86_validate_address_register
83 to validate base and index registers. Add address override check
84 from ix86_decompose_address.
85 (ix86_decompose_address): Remove.
87 2013-11-18 Richard Biener <rguenther@suse.de>
89 PR tree-optimization/59125
90 PR tree-optimization/54570
91 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): When inlining
92 is not complete do not treat component-references with offset zero
93 but different fields as equal.
94 * tree-object-size.c: Include tree-phinodes.h and ssa-iterators.h.
95 (compute_object_sizes): Apply TLC. Propagate the constant
96 results into all uses and fold their stmts.
97 * passes.def (pass_all_optimizations): Move pass_object_sizes
98 after the first pass_forwprop and before pass_fre.
100 2013-11-18 Richard Sandiford <rdsandiford@googlemail.com>
102 * tree.h (tree_to_uhwi): Return an unsigned HOST_WIDE_INT.
103 * tree.c (tree_to_uhwi): Return an unsigned HOST_WIDE_INT.
104 (tree_ctz): Remove cast to unsigned type.
105 * builtins.c (fold_builtin_memory_op): Likewise.
106 * dwarf2out.c (descr_info_loc): Likewise.
107 * godump.c (go_output_typedef): Likewise.
108 * omp-low.c (expand_omp_simd): Likewise.
109 * stor-layout.c (excess_unit_span): Likewise.
110 * tree-object-size.c (addr_object_size): Likewise.
111 * tree-sra.c (analyze_all_variable_accesses): Likewise.
112 * tree-ssa-forwprop.c (simplify_builtin_call): Likewise.
113 (simplify_rotate): Likewise.
114 * tree-ssa-strlen.c (adjust_last_stmt, handle_builtin_memcpy)
115 (handle_pointer_plus): Likewise.
116 * tree-switch-conversion.c (check_range): Likewise.
117 * tree-vect-patterns.c (vect_recog_rotate_pattern): Likewise.
118 * tsan.c (instrument_builtin_call): Likewise.
119 * cfgexpand.c (defer_stack_allocation): Add cast to HOST_WIDE_INT.
120 * trans-mem.c (tm_log_add): Likewise.
121 * config/aarch64/aarch64.c (aapcs_vfp_sub_candidate): Likewise.
122 * config/arm/arm.c (aapcs_vfp_sub_candidate): Likewise.
123 * config/rs6000/rs6000.c (rs6000_aggregate_candidate): Likewise.
124 * config/mips/mips.c (r10k_safe_mem_expr_p): Make offset unsigned.
126 2013-11-18 Richard Sandiford <rdsandiford@googlemail.com>
128 * tree.h (host_integerp, tree_low_cst): Delete.
129 * tree.c (host_integerp, tree_low_cst): Delete.
131 2013-11-18 Richard Sandiford <rdsandiford@googlemail.com>
133 * expr.h: Update comments to refer to tree_to_[su]hwi rather
135 * fold-const.c (fold_binary_loc): Likewise.
136 * expr.c (store_constructor): Use tree_to_uhwi rather than
138 * ipa-utils.h (possible_polymorphic_call_target_p): Likewise.
139 * stmt.c (emit_case_dispatch_table): Likewise.
140 * tree-switch-conversion.c (emit_case_bit_tests): Likewise.
142 2013-11-18 Richard Sandiford <rdsandiford@googlemail.com>
144 * alias.c, asan.c, builtins.c, cfgexpand.c, cgraph.c,
145 config/aarch64/aarch64.c, config/alpha/predicates.md,
146 config/arm/arm.c, config/darwin.c, config/epiphany/epiphany.c,
147 config/i386/i386.c, config/iq2000/iq2000.c, config/m32c/m32c-pragma.c,
148 config/mep/mep-pragma.c, config/mips/mips.c,
149 config/picochip/picochip.c, config/rs6000/rs6000.c, cppbuiltin.c,
150 dbxout.c, dwarf2out.c, emit-rtl.c, except.c, expr.c, fold-const.c,
151 function.c, gimple-fold.c, godump.c, ipa-cp.c, ipa-prop.c, omp-low.c,
152 predict.c, sdbout.c, stor-layout.c, trans-mem.c, tree-object-size.c,
153 tree-sra.c, tree-ssa-ccp.c, tree-ssa-forwprop.c,
154 tree-ssa-loop-ivcanon.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-niter.c,
155 tree-ssa-loop-prefetch.c, tree-ssa-strlen.c, tree-stdarg.c,
156 tree-switch-conversion.c, tree-vect-generic.c, tree-vect-loop.c,
157 tree-vect-patterns.c, tree-vrp.c, tree.c, tsan.c, ubsan.c, varasm.c:
158 Replace tree_low_cst (..., 1) with tree_to_uhwi throughout.
160 2013-11-18 Richard Sandiford <rdsandiford@googlemail.com>
162 * builtins.c, cilk-common.c, config/aarch64/aarch64.c,
163 config/alpha/alpha.c, config/arm/arm.c, config/c6x/predicates.md,
164 config/i386/i386.c, config/ia64/predicates.md, config/s390/s390.c,
165 coverage.c, dbxout.c, dwarf2out.c, except.c, explow.c, expr.c, expr.h,
166 fold-const.c, gimple-fold.c, godump.c, ipa-prop.c, omp-low.c,
167 predict.c, rtlanal.c, sdbout.c, stmt.c, stor-layout.c, targhooks.c,
168 tree-cfg.c, tree-data-ref.c, tree-inline.c, tree-ssa-forwprop.c,
169 tree-ssa-loop-prefetch.c, tree-ssa-phiopt.c, tree-ssa-sccvn.c,
170 tree-ssa-strlen.c, tree-stdarg.c, tree-vect-data-refs.c,
171 tree-vect-patterns.c, tree.c, tree.h, var-tracking.c, varasm.c:
172 Replace tree_low_cst (..., 0) with tree_to_shwi throughout.
174 2013-11-18 Richard Sandiford <rdsandiford@googlemail.com>
176 * tree.h (tree_to_shwi, tree_to_uhwi): Declare, with inline expansions.
177 * tree.c (tree_to_shwi, tree_to_uhwi): New functions.
179 2013-11-18 Richard Sandiford <rdsandiford@googlemail.com>
181 * expr.h: Update comments to refer to tree_fits_[su]hwi_p rather
184 2013-11-18 Richard Sandiford <rdsandiford@googlemail.com>
186 * builtins.c, config/alpha/alpha.c, config/iq2000/iq2000.c,
187 config/mips/mips.c, dbxout.c, dwarf2out.c, expr.c, fold-const.c,
188 gimple-fold.c, godump.c, omp-low.c, predict.c, sdbout.c, stor-layout.c,
189 tree-dfa.c, tree-sra.c, tree-ssa-forwprop.c, tree-ssa-loop-prefetch.c,
190 tree-ssa-phiopt.c, tree-ssa-sccvn.c, tree-ssa-strlen.c,
191 tree-ssa-structalias.c, tree-vect-data-refs.c, tree-vect-patterns.c,
192 tree.c, varasm.c, alias.c, cfgexpand.c, config/aarch64/aarch64.c,
193 config/arm/arm.c, config/epiphany/epiphany.c, config/i386/i386.c,
194 config/m32c/m32c-pragma.c, config/mep/mep-pragma.c,
195 config/rs6000/rs6000.c, config/sparc/sparc.c, emit-rtl.c, function.c,
196 gimplify.c, ipa-prop.c, stmt.c, trans-mem.c, tree-cfg.c,
197 tree-object-size.c, tree-ssa-ccp.c, tree-ssa-loop-ivcanon.c,
198 tree-stdarg.c, tree-switch-conversion.c, tree-vect-generic.c,
199 tree-vrp.c, tsan.c, ubsan.c: Replace host_integerp (..., 1) with
200 tree_fits_uhwi_p throughout.
202 2013-11-18 Richard Sandiford <rdsandiford@googlemail.com>
204 * builtins.c, config/alpha/alpha.c, config/c6x/predicates.md,
205 config/ia64/predicates.md, config/iq2000/iq2000.c, config/mips/mips.c,
206 config/s390/s390.c, dbxout.c, dwarf2out.c, except.c, explow.c, expr.c,
207 expr.h, fold-const.c, gimple-fold.c, gimple-ssa-strength-reduction.c,
208 gimple.c, godump.c, graphite-scop-detection.c, graphite-sese-to-poly.c,
209 omp-low.c, predict.c, rtlanal.c, sdbout.c, simplify-rtx.c,
210 stor-layout.c, tree-data-ref.c, tree-dfa.c, tree-pretty-print.c,
211 tree-sra.c, tree-ssa-alias.c, tree-ssa-forwprop.c,
212 tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c, tree-ssa-math-opts.c,
213 tree-ssa-phiopt.c, tree-ssa-reassoc.c, tree-ssa-sccvn.c,
214 tree-ssa-strlen.c, tree-ssa-structalias.c, tree-vect-data-refs.c,
215 tree-vect-patterns.c, tree-vectorizer.h, tree.c, var-tracking.c,
216 varasm.c: Replace host_integerp (..., 0) with tree_fits_shwi_p
219 2013-11-18 Richard Sandiford <rdsandiford@googlemail.com>
221 * tree.h (tree_fits_shwi_p, tree_fits_uhwi_p): Declare.
222 * tree.c (tree_fits_shwi_p, tree_fits_uhwi_p): Define.
224 2013-11-18 Kirill Yukhin <kirill.yukhin@intel.com>
226 * gcc/config/ia64/ia64.c (ia64_split_tmode_move): Mark load with
227 `dead' flag if it kills address, not its post-increment.
229 2013-11-18 Ilya Enkovich <ilya.enkovich@intel.com>
231 * builtin-types.def (BT_FN_PTR_CONST_PTR_VAR): New.
232 * chkp-builtins.def (BUILT_IN_CHKP_BIND_BOUNDS): New.
233 * cfgexpand.c (expand_call_stmt): Expand BUILT_IN_CHKP_BIND_BOUNDS.
234 * gimple.c (gimple_call_get_nobnd_arg_index): Remove.
235 * gimple.h (gf_mask): Add GF_CALL_WITH_BOUNDS.
236 (gimple_call_with_bounds_p): New.
237 (gimple_call_set_with_bounds): New.
238 (gimple_call_num_nobnd_args): Remove.
239 (gimple_call_nobnd_arg): Remove.
240 * tree.h (CALL_WITH_BOUNDS_P): New.
241 * rtl.h (CALL_EXPR_WITH_BOUNDS_P): New.
243 2013-11-18 Trevor Saunders <tsaunders@mozilla.com>
245 * cgraph.h (symtab_node_asm_name): Rename to symtab_node::asm_name.
246 (symtab_node_name): Rename to symtab_node::name.
247 (cgraph_node_asm_name): Remove.
248 (varpool_node_asm_name): Remove.
249 * cgraph.c cgraphclones.c cgraphunit.c ipa-cp.c ipa-devirt.c
250 ipa-inline-analysis.c ipa-inline-transform.c ipa-inline.c
251 ipa-profile.c ipa-prop.c ipa-pure-const.c ipa-ref.c ipa-reference.c
252 ipa-utils.c ipa.c symtab.c tree-inline.c tree-sra.c
253 tree-ssa-structalias.c value-prof.c varpool.c Adjust.
255 2013-11-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
257 * config/arm/aarch-cost-tables.h (cortexa53_extra_costs): New table.
258 * config/arm/arm.c (arm_cortex_a53_tune): New.
259 * config/arm/arm-cores.def (cortex-a53): Use cortex_a53 tuning struct.
261 2013-11-12 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
263 * config.gcc (i[34567]86-*-linux* | ...): Add bdver4.
264 (case ${target}): Add bdver4.
265 * config/i386/bdver3.md: Add bdver4.
266 * config/i386/driver-i386.c: (host_detect_local_cpu): Let
267 -march=native recognize bdver4 processors.
268 * config/i386/i386-c.c (ix86_target_macros_internal): Add
270 * config/i386/i386.c (struct processor_costs bdver4_cost): New.
271 (m_BDVER4): New definition.
272 (m_AMD_MULTIPLE): Includes m_BDVER4.
273 (processor_target_table): Add bdver4 entry.
274 (static const char *const cpu_names): Add bdver4 entry.
275 (software_prefetching_beneficial_p): Add bdver3.
276 (ix86_option_override_internal): Add bdver4 instruction sets.
277 (ix86_issue_rate): Add bdver4.
278 (ix86_adjust_cost): Add bdver4.
279 (ia32_multipass_dfa_lookahead): Add bdver4.
280 (enum processor_model): Add M_AMDFAM15H_BDVER4.
281 (struct _arch_names_table): Add M_AMDFAM15H_BDVER4.
282 (has_dispatch): Add bdver4.
283 * config/i386/i386.h (TARGET_BDVER4): New definition.
284 (enum target_cpu_default): Add TARGET_CPU_DEFAULT_bdver4.
285 (enum processor_type): Add PROCESSOR_BDVER4.
286 * config/i386/i386.md (define_attr "cpu"): Add bdver4.
287 * config/i386/i386.opt (flag_dispatch_scheduler): Add bdver4.
288 * gcc/doc/extend.texi: Add details about bdver4.
289 * gcc/doc/invoke.texi: Add details about bdver4. Add
290 fma4 and fsgsbase for bdver3. Add fma4 for bdver2.
292 2013-11-17 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
294 * config/rs6000/rs6000.c (rs6000_emit_move): Use low word of
295 sdmode_stack_slot also in little-endian mode.
297 2013-11-17 Jan Hubicka <jh@suse.cz>
299 * doc/md.texi (setmem, movstr): Update documentation.
300 * builtins.c (determine_block_size): New function.
301 (expand_builtin_memcpy): Use it and pass it to emit_block_move_hints.
302 (expand_builtin_memset_args): Use it and pass it to
303 set_storage_via_setmem.
304 * expr.c (emit_block_move_via_movmem): Add min_size/max_size
305 parameters; update call to expander.
306 (emit_block_move_hints): Add min_size/max_size parameters.
307 (clear_storage_hints): Likewise.
308 (set_storage_via_setmem): Likewise.
309 (clear_storage): Update.
310 * expr.h (emit_block_move_hints, clear_storage_hints,
311 set_storage_via_setmem): Update prototypes.
312 * i386.c (ix86_expand_set_or_movmem): Add bounds; export.
313 (ix86_expand_movmem, ix86_expand_setmem): Remove.
314 (ix86_expand_movmem, ix86_expand_setmem): Remove.
315 * i386.md (movmem, setmem): Pass parameters.
317 2013-11-17 Uros Bizjak <ubizjak@gmail.com>
320 * config/i386/i386.c (ix86_address_subreg_operand): Do not
321 reject non-integer subregs.
322 (ix86_decompose_address): Do not reject invalid CONST_INT RTXes.
323 Move check for invalid x32 constant addresses ...
324 (ix86_legitimate_address_p): ... here.
326 2011-11-17 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
328 * config/rs6000/rs6000.c (rs6000_frame_related): Add split_reg
329 parameter and use it in REG_FRAME_RELATED_EXPR note.
330 (emit_frame_save): Call rs6000_frame_related with extra NULL_RTX
332 (rs6000_emit_prologue): Likewise, but for little endian VSX
333 stores, pass the source register of the store instead.
335 2013-11-17 Andrew MacLeod <amacleod@redhat.com>
337 * gimple.h: Reorder prototypes to match .c declaration order, and
338 remove protyotypes for functions not in gimple.c.
339 (LABEL): Move to tree-into-ssa.c.
340 * gimple.c: Remove unused prototypes.
341 (get_base_address): Move to tree.c.
342 * tree.c (get_base_address): Relocate from gimple.c.
343 * builtins.h (validate_gimple_arglist): Add prototype.
344 * trans-mem.h (compute_transaction_bits, is_tm_ending): Add prototype.
345 * cfgexpand.h: New File.
346 (gimple_assign_rhs_to_tree, estimated_stack_frame_size): Add protoype.
347 * tree.h (build_addr): Move to tree-nested.h.
348 * tree-nested.h: New File.
349 (build_addr, lower_nested_functions, insert_field_into_struct): Add
351 * tree-inline.h (estimated_stack_frame_size): Remove prototype.
352 * ipa-inline-analysis.c: Include cfgexpand.h.
353 * cgraphunit.c: Include tree-nested.h.
354 * omp-low.c: Likewise.
355 * tree-parloops.c: Likewise.
356 * gimple-low.h: Likewise.
357 * tree-profile.h: Likewise.
358 * expr.c: Include cfgexpand.h.
359 * tree-affine.c: Likewise.
360 * tree-ssa.c: Likewise.
361 * tree-ssa-loop-im.c: Include trans-mem.h.
362 * tree-ssa-tail-merge.c: Likewise.
363 * value-prof.c: Include builtins.h and tree-nested.h.
364 * tree-into-ssa.c (LABEL): Define here.
366 2013-11-16 Joern Rennecke <joern.rennecke@embecosm.com>
368 * config/arc/arc.c (arc_predicate_delay_insns): New function.
369 (pass_data_arc_predicate_delay_insns): New pass_data instance.
370 (pass_arc_predicate_delay_insns): New subclass of rtl_opt_class.
371 (make_pass_arc_predicate_delay_insns): New function.
372 (arc_init): Register pass_arc_predicate_delay_insns if
373 flag_delayed_branch is active.
375 2013-11-16 Joern Rennecke <joern.rennecke@embecosm.com>
377 * config/arc/constraints.md (Rcq): Simplify register number test.
379 2013-11-15 Aldy Hernandez <aldyh@redhat.com>
381 * gimple.h (enum gf_mask): Change the ordering of GF_OMP_* bits.
383 2013-11-15 Kaz Kojima <kkojima@gcc.gnu.org>
385 * config/sh/sh.c (barrier_align): Return 0 when barrier_or_label
388 2013-11-15 Aldy Hernandez <aldyh@redhat.com>
390 * Makefile.in (C_COMMON_OBJS): Depend on c-cilkplus.o.
391 * gimple-pretty-print.c (dump_omp_for): Add case for
392 GF_OMP_FOR_KIND_CILKSIMD.
393 * gimple.h (enum gf_mask): Restructure entries to add
394 GF_OMP_FOR_KIND_CILKSIMD.
395 * gimplify.c (is_gimple_stmt): Add case for CILK_SIMD.
396 (gimplify_omp_for): Handle CILK_SIMD.
397 (gimplify_expr): Add ccase for CILK_SIMD.
398 * omp-low.c (extract_omp_for_data): Handle CILK_SIMD.
399 (build_outer_var_ref): Same.
400 (check_omp_nesting_restrictions): Same.
401 (lower_rec_input_clauses): Same.
402 (lower_lastprivate_clauses): Same.
403 (expand_omp_for): Same.
404 (execute_expand_omp): Check flag_enable_cilkplus.
405 (execute_lower_omp): Same.
406 (diagnose_sb_0): Handle CILK_SIMD.
407 (diagnose_omp_structured_block_errors): Check flag_enable_cilkplus.
408 (setjmp_or_longjmp_p): New.
409 (scan_omp_1_stmt): Error on setjmp/longjmp in a simd construct.
410 * tree-pretty-print.c (dump_generic_node): Add case for CILK_SIMD.
411 * tree.def: Add tree code for CILK_SIMD.
413 2013-11-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
415 * config/rs6000/altivec.md (UNSPEC_VPERM_X, UNSPEC_VPERM_UNS_X):
417 (altivec_vperm_<mode>): Revert earlier little endian change.
418 (*altivec_vperm_<mode>_internal): Remove.
419 (altivec_vperm_<mode>_uns): Revert earlier little endian change.
420 (*altivec_vperm_<mode>_uns_internal): Remove.
421 * config/rs6000/vector.md (vec_realign_load_<mode>): Revise commentary.
423 2013-11-15 Jeff Law <law@redhat.com>
425 * basic-block.h (has_abnormal_or_eh_outgoing_edge): Renamed from
426 has_abnormal_or_outgoing_edge. Check for EH edges as well.
427 * gimple-ssa-isolate-paths.c
428 (find_implicit_erroneous_behaviour): Corresponding changes.
429 Do not check stmt_ends_bb_p or GIMPLE_RETURN anymore.
430 (find_explicit_erroneous_behaviour): Likewise.
432 2013-11-15 Jeff Law <law@redhat.com>
434 * ifcvt.c (find_cond_trap): Properly handle case where
437 2013-11-15 Andreas Schwab <schwab@linux-m68k.org>
439 * configure: Regenerate.
441 2013-11-15 James Greenhalgh <james.greenhalgh@arm.com>
443 * config/aarch64/aarch64-simd.md: Remove simd_type from all patterns.
444 * config/aarch64/aarch64.md: Likewise, correct "type" attribute
445 where it is incorrect or missing.
447 2013-11-15 Richard Sandiford <rdsandiford@googlemail.com>
449 * dwarf2out.c (gen_enumeration_type_die): Remove unnecessary
451 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
452 Use TREE_INT_CST_LOW rather than tree_low_cst when reading the
454 * fold-const.c (fold_binary_loc): Replace a host_integerp/tree_low_cst
455 pair with a TREE_CODE test and TREE_INT_CST_LOW.
456 * tree-vect-generic.c (expand_vector_divmod): Likewise.
458 2013-11-15 Richard Biener <rguenther@suse.de>
460 PR tree-optimization/50262
461 * tree-ssa-alias.h (struct pt_solution): Split
462 vars_contains_global into vars_contains_nonlocal,
463 vars_contains_escaped and vars_contains_escaped_heap.
464 * tree-ssa-structalias.c (label_visit): Expand comment.
465 (handle_lhs_call): Adjust comment.
466 (set_uids_in_ptset): Set the new flags appropriately.
467 (pt_solution_set): Adjust.
468 (pt_solution_set_var): Likewise.
469 (pt_solution_ior_into): Likewise.
470 (pt_solution_includes_global): Likewise.
471 (pt_solutions_intersect_1): Optimize escaped handling.
472 (compute_points_to_sets): Remove heap variable globalization.
473 (ipa_escaped_pt): Adjust initializer.
474 (pass_data_ipa_pta): Do not run TODO_update_ssa.
475 * gimple-pretty-print.c (pp_points_to_solution): Print split flags.
476 * tree-ssa-alias.c (dump_points_to_solution): Likewise.
478 2013-11-15 Richard Biener <rguenther@suse.de>
480 * tree-loop-distribution.c (tree_loop_distribution): Make sure
481 to distribute all stores.
483 2013-11-15 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
485 * doc/invoke.texi (-mabi=elfv1, -mabi=elfv2): Document.
487 2013-11-15 Joseph Myers <joseph@codesourcery.com>
489 * acinclude.m4 (GCC_GLIBC_VERSION_GTE_IFELSE): New configure macro.
490 * configure.ac: Determine target_header_dir earlier.
491 (--with-glibc-version): New configure option.
492 Use GCC_GLIBC_VERSION_GTE_IFELSE in enable_gnu_unique_object,
493 gcc_cv_libc_provides_ssp and gcc_cv_target_ldbl128 tests.
494 * configure: Regenerate.
495 * doc/install.texi (--enable-gnu-unique-object): Don't refer to
496 native toolchains for default.
497 (--with-glibc-version): Document.
499 2013-11-15 Eric Botcazou <ebotcazou@adacore.com>
501 * fold-const.c (fold_binary_loc) <comparisons>: Reuse local variable.
503 2013-11-15 Uros Bizjak <ubizjak@gmail.com>
505 * lto-streamer-in.c (input function): Call cgraph_create_node if
506 cgraph_get_node failed.
508 2013-11-14 Olivier Hainque <hainque@adacore.com>
510 * cfgexpand.c (defer_stack_allocation): When optimization is enabled,
511 defer allocation of DECL_IGNORED_P variables at toplevel unless really
512 small. Factorize size threshold computation from the existing one.
513 (expand_used_vars): Refine comment.
515 2013-11-14 Cong Hou <congh@google.com>
517 * tree-vectorizer.h (struct dr_with_seg_len): Remove the base
518 address field as it can be obtained from dr. Rename the struct.
519 * tree-vect-data-refs.c (comp_dr_with_seg_len_pair): Consider
520 steps of data references during sort.
521 (vect_prune_runtime_alias_test_list): Adjust with the change to
522 struct dr_with_seg_len.
523 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
524 Adjust with the change to struct dr_with_seg_len.
526 2013-11-14 Jeff Law <law@redhat.com>
529 * basic-block.h (has_abnormal_outgoing_edge_p): Moved here from...
530 * tree-inline.c (has_abnormal_outgoing_edge_p): Remove.
531 * gimple-ssa-isolate-paths.c: Include tree-cfg.h.
532 (find_implicit_erroneous_behaviour): If a block has abnormal outgoing
533 edges, then ignore it. If the statement exhibiting erroneous
534 behaviour ends basic blocks, with the exception of GIMPLE_RETURNs,
535 then we can not optimize.
536 (find_explicit_erroneous_behaviour): Likewise.
538 2013-11-14 Andrew MacLeod <amacleod@redhat.com>
540 * gimplify-me.h: New file. Add prototypes.
541 * gimplify.h: Don't include gimple.h.
542 (struct gimplify_hasher, struct gimplify_ctx, is_gimple_sizepos):
543 Relocate from gimple.h.
544 * gimple.h (struct gimplify_hasher, struct gimplify_ctx,
545 is_gimple_sizepos): Move to gimplify.h.
546 (gimplify_hasher::hash, gimplify_hasher::equal): Move to gimplify.c.
547 (enum gsi_iterator_update): Move to gimple-iterator.h.
548 * gimple-iterator.h (enum gsi_iterator_update): Relocate from gimple.h.
549 * gimplify-me.c: New File.
550 (force_gimple_operand_1, force_gimple_operand,
551 force_gimple_operand_gsi_1, force_gimple_operand_gsi,
552 gimple_regimplify_operands): Relocate from gimplify.c.
553 * gimplify.c (force_gimple_operand_1, force_gimple_operand,
554 force_gimple_operand_gsi_1, force_gimple_operand_gsi,
555 gimple_regimplify_operands): Move to gimplify-me.c.
556 (gimplify_hasher::hash, gimplify_hasher::equal): Relocate
558 * Makefile.in (OBJS): Add gimplify-me.o
559 * asan.c: Include only gimplify.h, gimplify-me.h, and/or gimple.h as
561 * cfgloopmanip.c: Likewise.
562 * cgraphunit.c: Likewise.
563 * cilk-common.c: Likewise.
564 * fold-const.c: Likewise.
565 * function.c: Likewise.
566 * gimple-expr.c: Likewise.
567 * gimple-fold.c: Likewise.
568 * gimple-ssa-strength-reduction.c: Likewise.
569 * gimple.c: Likewise.
570 * graphite-clast-to-gimple.c: Likewise.
571 * graphite-sese-to-poly.c: Likewise.
572 * ipa-prop.c: Likewise.
573 * ipa-split.c: Likewise.
575 * langhooks.c: Likewise.
576 * omp-low.c: Likewise.
578 * stor-layout.c: Likewise.
579 * targhooks.c: Likewise.
580 * trans-mem.c: Likewise.
581 * tree-affine.c: Likewise.
582 * tree-cfg.c: Likewise.
583 * tree-cfgcleanup.c: Likewise.
584 * tree-complex.c: Likewise.
585 * tree-if-conv.c: Likewise.
586 * tree-inline.c: Likewise.
587 * tree-loop-distribution.c: Likewise.
588 * tree-nested.c: Likewise.
589 * tree-parloops.c: Likewise.
590 * tree-predcom.c: Likewise.
591 * tree-profile.c: Likewise.
592 * tree-scalar-evolution.c: Likewise.
593 * tree-sra.c: Likewise.
594 * tree-ssa-address.c: Likewise.
595 * tree-ssa-ccp.c: Likewise.
596 * tree-ssa-dce.c: Likewise.
597 * tree-ssa-forwprop.c: Likewise.
598 * tree-ssa-ifcombine.c: Likewise.
599 * tree-ssa-loop-im.c: Likewise.
600 * tree-ssa-loop-ivopts.c: Likewise.
601 * tree-ssa-loop-manip.c: Likewise.
602 * tree-ssa-loop-niter.c: Likewise.
603 * tree-ssa-loop-prefetch.c: Likewise.
604 * tree-ssa-loop-unswitch.c: Likewise.
605 * tree-ssa-math-opts.c: Likewise.
606 * tree-ssa-phiopt.c: Likewise.
607 * tree-ssa-phiprop.c: Likewise.
608 * tree-ssa-pre.c: Likewise.
609 * tree-ssa-propagate.c: Likewise.
610 * tree-ssa-reassoc.c: Likewise.
611 * tree-ssa-sccvn.c: Likewise.
612 * tree-ssa-strlen.c: Likewise.
613 * tree-ssa.c: Likewise.
614 * tree-switch-conversion.c: Likewise.
615 * tree-tailcall.c: Likewise.
616 * tree-vect-data-refs.c: Likewise.
617 * tree-vect-generic.c: Likewise.
618 * tree-vect-loop-manip.c: Likewise.
619 * tree-vect-loop.c: Likewise.
620 * tree-vect-patterns.c: Likewise.
621 * tree-vect-stmts.c: Likewise.
624 * value-prof.c: Likewise.
625 * config/aarch64/aarch64.c: Likewise.
626 * config/alpha/alpha.c: Likewise.
627 * config/darwin.c: Likewise.
628 * config/i386/i386.c: Likewise.
629 * config/ia64/ia64.c: Likewise.
630 * config/mep/mep.c: Likewise.
631 * config/mips/mips.c: Likewise.
632 * config/rs6000/rs6000.c: Likewise.
633 * config/s390/s390.c: Likewise.
634 * config/sh/sh.c: Likewise.
635 * config/sparc/sparc.c: Likewise.
636 * config/spu/spu.c: Likewise.
637 * config/stormy16/stormy16.c: Likewise.
638 * config/tilegx/tilegx.c: Likewise.
639 * config/tilepro/tilepro.c: Likewise.
640 * config/xtensa/xtensa.c: Likewise.
642 2013-11-14 Joern Rennecke <joern.rennecke@embecosm.com>
644 * config/arc/arc.md (doloop_begin_i): Remove extra alignment;
647 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
649 * config/rs6000/sysv4le.h (LINUX64_DEFAULT_ABI_ELFv2): Define.
651 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
652 Alan Modra <amodra@gmail.com>
654 * config/rs6000/rs6000.h (RS6000_SAVE_AREA): Handle ABI_ELFv2.
655 (RS6000_SAVE_TOC): Remove.
656 (RS6000_TOC_SAVE_SLOT): New macro.
657 * config/rs6000/rs6000.c (rs6000_parm_offset): New function.
658 (rs6000_parm_start): Use it.
659 (rs6000_function_arg_advance_1): Likewise.
660 (rs6000_emit_prologue): Use RS6000_TOC_SAVE_SLOT.
661 (rs6000_emit_epilogue): Likewise.
662 (rs6000_call_aix): Likewise.
663 (rs6000_output_function_prologue): Do not save/restore r11
664 around calling _mcount for ABI_ELFv2.
666 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
667 Alan Modra <amodra@gmail.com>
669 * config/rs6000/rs6000-protos.h (rs6000_reg_parm_stack_space):
671 * config/rs6000/rs6000.h (RS6000_REG_SAVE): Remove.
672 (REG_PARM_STACK_SPACE): Call rs6000_reg_parm_stack_space.
673 * config/rs6000/rs6000.c (rs6000_parm_needs_stack): New function.
674 (rs6000_function_parms_need_stack): Likewise.
675 (rs6000_reg_parm_stack_space): Likewise.
676 (rs6000_function_arg): Do not replace BLKmode by Pmode when
677 returning a register argument.
679 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
680 Michael Gschwind <mkg@us.ibm.com>
682 * config/rs6000/rs6000.h (FP_ARG_MAX_RETURN): New macro.
683 (ALTIVEC_ARG_MAX_RETURN): Likewise.
684 (FUNCTION_VALUE_REGNO_P): Use them.
685 * config/rs6000/rs6000.c (TARGET_RETURN_IN_MSB): Define.
686 (rs6000_return_in_msb): New function.
687 (rs6000_return_in_memory): Handle ELFv2 homogeneous aggregates.
688 Handle aggregates of up to 16 bytes for ELFv2.
689 (rs6000_function_value): Handle ELFv2 homogeneous aggregates.
691 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
692 Michael Gschwind <mkg@us.ibm.com>
694 * config/rs6000/rs6000.h (AGGR_ARG_NUM_REG): Define.
695 * config/rs6000/rs6000.c (rs6000_aggregate_candidate): New function.
696 (rs6000_discover_homogeneous_aggregate): Likewise.
697 (rs6000_function_arg_boundary): Handle homogeneous aggregates.
698 (rs6000_function_arg_advance_1): Likewise.
699 (rs6000_function_arg): Likewise.
700 (rs6000_arg_partial_bytes): Likewise.
701 (rs6000_psave_function_arg): Handle BLKmode arguments.
703 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
704 Michael Gschwind <mkg@us.ibm.com>
706 * config/rs6000/rs6000.h (AGGR_ARG_NUM_REG): Define.
707 * config/rs6000/rs6000.c (rs6000_aggregate_candidate): New function.
708 (rs6000_discover_homogeneous_aggregate): Likewise.
709 (rs6000_function_arg_boundary): Handle homogeneous aggregates.
710 (rs6000_function_arg_advance_1): Likewise.
711 (rs6000_function_arg): Likewise.
712 (rs6000_arg_partial_bytes): Likewise.
713 (rs6000_psave_function_arg): Handle BLKmode arguments.
715 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
717 * config/rs6000/rs6000.c (machine_function): New member
719 (rs6000_emit_prologue): Set r2_setup_needed if necessary.
720 (rs6000_output_mi_thunk): Set r2_setup_needed.
721 (rs6000_output_function_prologue): Output global entry point
722 prologue and local entry point marker if needed for ABI_ELFv2.
723 Output -mprofile-kernel code here.
724 (output_function_profiler): Do not output -mprofile-kernel
725 code here; moved to rs6000_output_function_prologue.
726 (rs6000_file_start): Output ".abiversion 2" for ABI_ELFv2.
728 (rs6000_emit_move): Do not handle dot symbols for ABI_ELFv2.
729 (rs6000_output_function_entry): Likewise.
730 (rs6000_assemble_integer): Likewise.
731 (rs6000_elf_encode_section_info): Likewise.
732 (rs6000_elf_declare_function_name): Do not create dot symbols
733 or .opd section for ABI_ELFv2.
735 (rs6000_trampoline_size): Update for ABI_ELFv2 trampolines.
736 (rs6000_trampoline_init): Likewise.
737 (rs6000_elf_file_end): Call file_end_indicate_exec_stack for ABI_ELFv2.
739 (rs6000_call_aix): Handle ELFv2 indirect calls. Do not check
740 for function descriptors in ABI_ELFv2.
742 * config/rs6000/rs6000.md ("*call_indirect_aix<mode>"): Support
743 on ABI_AIX only, not ABI_ELFv2.
744 ("*call_value_indirect_aix<mode>"): Likewise.
745 ("*call_indirect_elfv2<mode>"): New pattern.
746 ("*call_value_indirect_elfv2<mode>"): Likewise.
748 * config/rs6000/predicates.md ("symbol_ref_operand"): Do not
749 check for function descriptors in ABI_ELFv2.
750 ("current_file_function_operand"): Likewise.
752 * config/rs6000/ppc-asm.h [__powerpc64__ && _CALL_ELF == 2]:
754 (FUNC_NAME): Define ELFv2 variant.
755 (JUMP_TARGET): Likewise.
756 (FUNC_START): Likewise.
757 (HIDDEN_FUNC): Likewise.
758 (FUNC_END): Likeiwse.
760 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
762 * config.gcc [powerpc*-*-* | rs6000-*-*]: Support --with-abi=elfv1
763 and --with-abi=elfv2.
764 * config/rs6000/option-defaults.h (OPTION_DEFAULT_SPECS): Add "abi".
765 * config/rs6000/rs6000.opt (mabi=elfv1): New option.
766 (mabi=elfv2): Likewise.
767 * config/rs6000/rs6000-opts.h (enum rs6000_abi): Add ABI_ELFv2.
768 * config/rs6000/linux64.h (DEFAULT_ABI): Do not hard-code to AIX_ABI
770 (ELFv2_ABI_CHECK): New macro.
771 (SUBSUBTARGET_OVERRIDE_OPTIONS): Use it to decide whether to set
772 rs6000_current_abi to ABI_AIX or ABI_ELFv2.
773 (GLIBC_DYNAMIC_LINKER64): Support ELFv2 ld.so version.
774 * config/rs6000/rs6000-c.c (rs6000_cpu_cpp_builtins): Predefine
775 _CALL_ELF and __STRUCT_PARM_ALIGN__ if appropriate.
777 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Handle ABI_ELFv2.
778 (debug_stack_info): Likewise.
779 (rs6000_file_start): Treat ABI_ELFv2 the same as ABI_AIX.
780 (rs6000_legitimize_tls_address): Likewise.
781 (rs6000_conditional_register_usage): Likewise.
782 (rs6000_emit_move): Likewise.
783 (init_cumulative_args): Likewise.
784 (rs6000_function_arg_advance_1): Likewise.
785 (rs6000_function_arg): Likewise.
786 (rs6000_arg_partial_bytes): Likewise.
787 (rs6000_output_function_entry): Likewise.
788 (rs6000_assemble_integer): Likewise.
789 (rs6000_savres_strategy): Likewise.
790 (rs6000_stack_info): Likewise.
791 (rs6000_function_ok_for_sibcall): Likewise.
792 (rs6000_emit_load_toc_table): Likewise.
793 (rs6000_savres_routine_name): Likewise.
794 (ptr_regno_for_savres): Likewise.
795 (rs6000_emit_prologue): Likewise.
796 (rs6000_emit_epilogue): Likewise.
797 (rs6000_output_function_epilogue): Likewise.
798 (output_profile_hook): Likewise.
799 (output_function_profiler): Likewise.
800 (rs6000_trampoline_size): Likewise.
801 (rs6000_trampoline_init): Likewise.
802 (rs6000_elf_output_toc_section_asm_op): Likewise.
803 (rs6000_elf_encode_section_info): Likewise.
804 (rs6000_elf_reloc_rw_mask): Likewise.
805 (rs6000_elf_declare_function_name): Likewise.
806 (rs6000_function_arg_boundary): Treat ABI_ELFv2 the same as ABI_AIX,
807 except that rs6000_compat_align_parm is always assumed false.
808 (rs6000_gimplify_va_arg): Likewise.
809 (rs6000_call_aix): Update comment.
810 (rs6000_sibcall_aix): Likewise.
811 * config/rs6000/rs6000.md ("tls_gd_aix<TLSmode:tls_abi_suffix>"):
812 Treat ABI_ELFv2 the same as ABI_AIX.
813 ("*tls_gd_call_aix<TLSmode:tls_abi_suffix>"): Likewise.
814 ("tls_ld_aix<TLSmode:tls_abi_suffix>"): Likewise.
815 ("*tls_ld_call_aix<TLSmode:tls_abi_suffix>"): Likewise.
816 ("load_toc_aix_si"): Likewise.
817 ("load_toc_aix_di"): Likewise.
819 ("call_value"): Likewise.
820 ("*call_local_aix<mode>"): Likewise.
821 ("*call_value_local_aix<mode>"): Likewise.
822 ("*call_nonlocal_aix<mode>"): Likewise.
823 ("*call_value_nonlocal_aix<mode>"): Likewise.
824 ("*call_indirect_aix<mode>"): Likewise.
825 ("*call_value_indirect_aix<mode>"): Likewise.
826 ("sibcall"): Likewise.
827 ("sibcall_value"): Likewise.
828 ("*sibcall_aix<mode>"): Likewise.
829 ("*sibcall_value_aix<mode>"): Likewise.
830 * config/rs6000/predicates.md ("symbol_ref_operand"): Likewise.
831 ("current_file_function_operand"): Likewise.
833 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
835 * config/rs6000/rs6000.c (rs6000_arg_partial_bytes): Simplify logic
836 by making use of the fact that for vector / floating point arguments
837 passed both in VRs/FPRs and in the fixed parameter area, the partial
838 bytes mechanism is in fact not used.
840 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
842 * config/rs6000/rs6000.c (rs6000_psave_function_arg): New function.
843 (rs6000_finish_function_arg): Likewise.
844 (rs6000_function_arg): Use rs6000_psave_function_arg and
845 rs6000_finish_function_arg to handle both vector and floating
846 point arguments that are also passed in GPRs / the stack.
848 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
850 * config/rs6000/rs6000.c (USE_FP_FOR_ARG_P): Remove TYPE argument.
851 (USE_ALTIVEC_FOR_ARG_P): Likewise.
852 (rs6000_darwin64_record_arg_advance_recurse): Update uses.
853 (rs6000_function_arg_advance_1):Likewise.
854 (rs6000_darwin64_record_arg_recurse): Likewise.
855 (rs6000_function_arg): Likewise.
856 (rs6000_arg_partial_bytes): Likewise.
858 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
860 * config/rs6000/rs6000.c (rs6000_option_override_internal): Replace
861 "DEFAULT_ABI != ABI_AIX" test by testing for ABI_V4 or ABI_DARWIN.
862 (rs6000_savres_strategy): Likewise.
863 (rs6000_return_addr): Likewise.
864 (rs6000_emit_load_toc_table): Replace "DEFAULT_ABI != ABI_AIX" by
865 testing for ABI_V4 (since ABI_DARWIN is impossible here).
866 (rs6000_emit_prologue): Likewise.
867 (legitimate_lo_sum_address_p): Simplify DEFAULT_ABI test.
868 (rs6000_elf_declare_function_name): Remove duplicated test.
869 * config/rs6000/rs6000.md ("load_toc_v4_PIC_1"): Explicitly test
870 for ABI_V4 (instead of "DEFAULT_ABI != ABI_AIX" test).
871 ("load_toc_v4_PIC_1_normal"): Likewise.
872 ("load_toc_v4_PIC_1_476"): Likewise.
873 ("load_toc_v4_PIC_1b"): Likewise.
874 ("load_toc_v4_PIC_1b_normal"): Likewise.
875 ("load_toc_v4_PIC_1b_476"): Likewise.
876 ("load_toc_v4_PIC_2"): Likewise.
877 ("load_toc_v4_PIC_3b"): Likewise.
878 ("load_toc_v4_PIC_3c"): Likewise.
879 * config/rs6000/rs6000.h (RS6000_REG_SAVE): Simplify DEFAULT_ABI test.
880 (RS6000_SAVE_AREA): Likewise.
881 (FP_ARG_MAX_REG): Likewise.
882 (RETURN_ADDRESS_OFFSET): Likewise.
883 * config/rs6000/sysv.h (TARGET_TOC): Test for ABI_V4 instead
885 (SUBTARGET_OVERRIDE_OPTIONS): Likewise.
886 (MINIMAL_TOC_SECTION_ASM_OP): Likewise.
888 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
890 * config/rs6000/rs6000.c (rs6000_call_indirect_aix): Rename to ...
891 (rs6000_call_aix): ... this. Handle both direct and indirect calls.
892 Create call insn directly instead of via various gen_... routines.
893 Mention special registers used by the call in CALL_INSN_FUNCTION_USAGE.
894 (rs6000_sibcall_aix): New function.
895 * config/rs6000/rs6000.md (TOC_SAVE_OFFSET_32BIT): Remove.
896 (TOC_SAVE_OFFSET_64BIT): Likewise.
897 (AIX_FUNC_DESC_TOC_32BIT): Likewise.
898 (AIX_FUNC_DESC_TOC_64BIT): Likewise.
899 (AIX_FUNC_DESC_SC_32BIT): Likewise.
900 (AIX_FUNC_DESC_SC_64BIT): Likewise.
901 ("call" expander): Call rs6000_call_aix.
902 ("call_value" expander): Likewise.
903 ("call_indirect_aix<ptrsize>"): Replace this pattern ...
904 ("call_indirect_aix<ptrsize>_nor11"): ... and this pattern ...
905 ("*call_indirect_aix<mode>"): ... by this insn pattern.
906 ("call_value_indirect_aix<ptrsize>"): Replace this pattern ...
907 ("call_value_indirect_aix<ptrsize>_nor11"): ... and this pattern ...
908 ("*call_value_indirect_aix<mode>"): ... by this insn pattern.
909 ("*call_nonlocal_aix32", "*call_nonlocal_aix64"): Replace by ...
910 ("*call_nonlocal_aix<mode>"): ... this pattern.
911 ("*call_value_nonlocal_aix32", "*call_value_nonlocal_aix64"): Replace
912 ("*call_value_nonlocal_aix<mode>"): ... by this pattern.
913 ("*call_local_aix<mode>"): New insn pattern.
914 ("*call_value_local_aix<mode>"): Likewise.
915 ("sibcall" expander): Call rs6000_sibcall_aix.
916 ("sibcall_value" expander): Likewise. Move earlier in file.
917 ("*sibcall_nonlocal_aix<mode>"): Replace by ...
918 ("*sibcall_aix<mode>"): ... this pattern.
919 ("*sibcall_value_nonlocal_aix<mode>"): Replace by ...
920 ("*sibcall_value_aix<mode>"): ... this pattern.
921 * config/rs6000/rs6000-protos.h (rs6000_call_indirect_aix): Remove.
922 (rs6000_call_aix): Add prototype.
923 (rs6000_sibcall_aix): Likewise.
925 2013-11-14 Jakub Jelinek <jakub@redhat.com>
928 * asan.c (asan_emit_stack_protection): Ensure -fsection-anchors
929 isn't confused by the artificial decl.
931 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
933 * config/rs6000/rs6000.c (rs6000_emit_prologue): Do not place a
934 RTX_FRAME_RELATED_P marker on the UNSPEC_MOVESI_FROM_CR insn.
935 Instead, add USEs of all modified call-saved CR fields to the
936 insn storing the result to the stack slot, and provide an
937 appropriate REG_FRAME_RELATED_EXPR for that insn.
938 * config/rs6000/rs6000.md ("*crsave"): New insn pattern.
939 * config/rs6000/predicates.md ("crsave_operation"): New predicate.
941 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
942 Alan Modra <amodra@gmail.com>
944 * function.c (assign_parms): Use all.reg_parm_stack_space instead
945 of re-evaluating REG_PARM_STACK_SPACE target macro.
946 (locate_and_pad_parm): New parameter REG_PARM_STACK_SPACE. Use it
947 instead of evaluating target macro REG_PARM_STACK_SPACE every time.
948 (assign_parm_find_entry_rtl): Update call.
949 * calls.c (initialize_argument_information): Update call.
950 (emit_library_call_value_1): Likewise.
951 * expr.h (locate_and_pad_parm): Update prototype.
953 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
955 * calls.c (store_unaligned_arguments_into_pseudos): Skip PARALLEL
958 2013-11-14 DJ Delorie <dj@redhat.com>
960 * config/rx/rx.c (rx_mode_dependent_address_p): Allow offsets up
963 2013-11-14 Jeff Law <law@redhat.com>
965 * tree-ssa-threadedge.c (thread_through_normal_block): Only push the
966 EDGE_START_JUMP_THREAD marker if the jump threading path is empty.
968 2013-11-14 James Greenhalgh <james.greenhalgh@arm.com>
970 * doc/invoke.texi: Update documentation for AArch64's -mcpu
973 2013-11-14 James Greenhalgh <james.greenhalgh@arm.com>
975 * config/aarch64/aarch64-cores.def (example-1): Remove.
976 (example-2): Likewise.
977 * config/aarch64/aarch64-tune.md: Regenerate.
978 * config/aarch64/aarch64.md: Do not include "large.md" or "small.md".
979 (generic_sched): Remove "large", "small".
980 * config/aarch64/large.md: Delete.
981 * config/aarch64/small.md: Delete.
983 2013-11-14 James Greenhalgh <james.greenhalgh@arm.com>
985 * config/aarch64/aarch64-cores.def (cortex-a57): Tune for cortexa15.
986 * config/aarch64/aarch64-tune.md: Regenerate.
987 * config/aarch64/aarch64.md: Include cortex-a15 pipeline model.
988 (generic_sched): "no" if we are tuning for cortexa15.
989 * config/arm/cortex-a15.md: Include cortex-a15-neon.md by
992 2013-11-14 James Greenhalgh <james.greenhalgh@arm.com>
994 * config/aarch64/aarch64-arches.def (armv8-a): Tune for cortex-a53.
995 * config/aarch64/aarch64.md: Do not include aarch64-generic.md.
996 * config/aarch64/aarch64.c (aarch64_tune): Initialize to cortexa53.
997 (all_cores): Use cortexa53 when tuning for "generic".
998 (aarch64_override_options): Fix comment.
999 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Set to cortexa53.
1000 * config/aarch64/aarch64-generic.md: Delete.
1002 2013-11-14 James Greenhalgh <james.greenhalgh@arm.com>
1004 * config/aarch64/aarch64.c (all_architectures): Remove "generic".
1006 2013-11-14 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1008 * config/aarch64/aarch64.c: Include aarch-cost-tables.h.
1009 (generic_rtx_cost_table): Remove.
1010 (aarch64_rtx_costs): Use fields from cpu_cost_table.
1011 * config/aarch64/aarch64-protos.h (tune_params): Use cpu_cost_table for
1013 (cpu_rtx_cost_table): Remove.
1015 2013-11-14 Julian Brown <julian@codesourcery.com>
1016 Joey Ye <joey.ye@arm.com>
1018 * config/arm/arm.c (arm_cortex_m_branch_cost): New.
1019 (arm_v7m_tune): New.
1020 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune, arm_9e_tune,
1021 arm_v6t2_tune, arm_cortex_tune, arm_cortex_a15_tune,
1022 arm_cortex_a5_tune, arm_v6m_tune): Add comments for Sched adj cost.
1023 * config/arm/arm-cores.def (cortex-m4, cortex-m3): Use arm_v7m_tune.
1025 2013-11-14 Kirill Yukhin <kirill.yukhin@intel.com>
1028 * config/ia64/ia64.c (ia64_split_tmode_move): Relax `dead'
1031 2013-11-14 Jakub Jelinek <jakub@redhat.com>
1032 Uros Bizjak <ubizjak@gmail.com>
1035 * config/i386/i386.md (*anddi_2): Only allow CCZmode if
1036 operands[2] satisfies_constraint_Z that might have bit 31 set.
1038 2013-11-13 Jeff Law <law@redhat.com>
1040 PR tree-optimization/59102
1041 * gimple-ssa-isolate-paths.c
1042 (insert_trap_and_remove_trailing_statments): Ensure STMT is a
1043 gimple assignment before looking at gimple_assign_lhs.
1045 2013-11-13 Vladimir Makarov <vmakarov@redhat.com>
1047 * ira.c: Add comment about threads at the top of file.
1049 2013-11-13 Vladimir Makarov <vmakarov@redhat.com>
1051 * ira-color.c (coalesce_allocnos): Don't allocate and free
1054 2013-11-14 Tom de Vries <tom@codesourcery.com>
1056 * tree-ssa-tail-merge.c (gimple_equal_p): Add test for structural
1057 equality for GIMPLE_ASSIGN.
1059 2013-11-14 Tom de Vries <tom@codesourcery.com>
1061 * tree-ssa-tail-merge.c (gimple_operand_equal_value_p): Factor new
1063 (gimple_equal_p): ... here.
1065 2013-11-14 Tom de Vries <tom@codesourcery.com>
1067 * trans-mem.c (is_tm_ending): New function.
1068 * gimple.h (is_tm_ending): Declare.
1069 * tree-ssa-tail-merge.c (gimple_equal_p): Remove test on
1071 (find_duplicate): Use is_tm_ending instead of is_tm_ending_fndecl.
1073 2013-11-14 Tom de Vries <tom@codesourcery.com>
1075 * tree-ssa-tail-merge.c (gimple_equal_p): Remove equal variable.
1077 2013-11-13 Andrew MacLeod <amacleod@redhat.com>
1079 * gimple-walk.h: New File. Relocate prototypes from gimple.h.
1080 (struct walk_stmt_info): Relocate here from gimple.h.
1081 * gimple-iterator.h: New File. Relocate prototypes from gimple.h.
1082 (struct gimple_stmt_iterator_d): Relocate here from gimple.h.
1083 (gsi_start_1, gsi_none, gsi_start_bb, gsi_last_1, gsi_last_bb,
1084 gsi_end_p, gsi_one_before_end_p, gsi_next, gsi_prev, gsi_stmt,
1085 gsi_after_labels, gsi_next_nondebug, gsi_prev_nondebug,
1086 gsi_start_nondebug_bb, gsi_start_nondebug_after_labels_bb,
1087 gsi_last_nondebug_bb, gsi_bb, gsi_seq): Relocate here from gimple.h.
1088 * gimple.h (struct gimple_stmt_iterator_d): Move to gimple-iterator.h.
1089 (gsi_start_1, gsi_none, gsi_start_bb, gsi_last_1, gsi_last_bb,
1090 gsi_end_p, gsi_one_before_end_p, gsi_next, gsi_prev, gsi_stmt,
1091 gsi_after_labels, gsi_next_nondebug, gsi_prev_nondebug,
1092 gsi_start_nondebug_bb, gsi_start_nondebug_after_labels_bb,
1093 gsi_last_nondebug_bb, gsi_bb, gsi_seq): Move to gimple-iterator.h.
1094 (struct walk_stmt_info): Move to gimple-walk.h.
1095 (gimple_seq_set_location): Move to gimple.c
1096 * gimple-walk.c: New File.
1097 (walk_gimple_seq_mod, walk_gimple_seq, walk_gimple_asm, walk_gimple_op,
1098 walk_gimple_stmt, get_base_loadstore, walk_stmt_load_store_addr_ops,
1099 walk_stmt_load_store_ops): Relocate here from gimple.c.
1100 * gimple-iterator.c: Include gimple-iterator.h.
1101 * gimple.c (walk_gimple_seq_mod, walk_gimple_seq, walk_gimple_asm,
1102 walk_gimple_op, walk_gimple_stmt, get_base_loadstore,
1103 walk_stmt_load_store_addr_ops, walk_stmt_load_store_ops): Move to
1105 (gimple_seq_set_location): Relocate from gimple.h.
1106 * tree-phinodes.h (set_phi_nodes): Move to tree-phinodes.c.
1107 * tree-phinodes.c (set_phi_nodes): Relocate from tree-phinodes.h.
1108 * gengtype.c (open_base_files): Add gimple-iterator.h to include list.
1109 * Makefile.in (OBJS): Add gimple-walk.o
1110 * asan.c: Update Include list as required for gimple-iterator.h and
1112 * cfgexpand.c: Likewise.
1113 * cfgloop.c: Likewise.
1114 * cfgloopmanip.c: Likewise.
1115 * cgraph.c: Likewise.
1116 * cgraphbuild.c: Likewise.
1117 * cgraphunit.c: Likewise.
1118 * gimple-fold.c: Likewise.
1119 * gimple-low.c: Likewise.
1120 * gimple-pretty-print.c: Likewise.
1121 * gimple-ssa-isolate-paths.c: Likewise.
1122 * gimple-ssa-strength-reduction.c: Likewise.
1123 * gimple-streamer-in.c: Likewise.
1124 * gimple-streamer-out.c: Likewise.
1125 * gimplify.c: Likewise.
1126 * graphite-blocking.c: Likewise.
1127 * graphite-clast-to-gimple.c: Likewise.
1128 * graphite-dependences.c: Likewise.
1129 * graphite-interchange.c: Likewise.
1130 * graphite-optimize-isl.c: Likewise.
1131 * graphite-poly.c: Likewise.
1132 * graphite-scop-detection.c: Likewise.
1133 * graphite-sese-to-poly.c: Likewise.
1134 * graphite.c: Likewise.
1135 * ipa-inline-analysis.c: Likewise.
1136 * ipa-profile.c: Likewise.
1137 * ipa-prop.c: Likewise.
1138 * ipa-pure-const.c: Likewise.
1139 * ipa-split.c: Likewise.
1140 * lto-streamer-in.c: Likewise.
1141 * lto-streamer-out.c: Likewise.
1142 * omp-low.c: Likewise.
1143 * predict.c: Likewise.
1144 * profile.c: Likewise.
1146 * tracer.c: Likewise.
1147 * trans-mem.c: Likewise.
1148 * tree-call-cdce.c: Likewise.
1149 * tree-cfg.c: Likewise.
1150 * tree-cfgcleanup.c: Likewise.
1151 * tree-complex.c: Likewise.
1152 * tree-data-ref.c: Likewise.
1153 * tree-dfa.c: Likewise.
1154 * tree-eh.c: Likewise.
1155 * tree-emutls.c: Likewise.
1156 * tree-if-conv.c: Likewise.
1157 * tree-inline.c: Likewise.
1158 * tree-into-ssa.c: Likewise.
1159 * tree-loop-distribution.c: Likewise.
1160 * tree-nested.c: Likewise.
1161 * tree-nrv.c: Likewise.
1162 * tree-object-size.c: Likewise.
1163 * tree-outof-ssa.c: Likewise.
1164 * tree-parloops.c: Likewise.
1165 * tree-predcom.c: Likewise.
1166 * tree-profile.c: Likewise.
1167 * tree-scalar-evolution.c: Likewise.
1168 * tree-sra.c: Likewise.
1169 * tree-ssa-ccp.c: Likewise.
1170 * tree-ssa-coalesce.c: Likewise.
1171 * tree-ssa-copy.c: Likewise.
1172 * tree-ssa-copyrename.c: Likewise.
1173 * tree-ssa-dce.c: Likewise.
1174 * tree-ssa-dom.c: Likewise.
1175 * tree-ssa-dse.c: Likewise.
1176 * tree-ssa-forwprop.c: Likewise.
1177 * tree-ssa-ifcombine.c: Likewise.
1178 * tree-ssa-live.c: Likewise.
1179 * tree-ssa-loop-ch.c: Likewise.
1180 * tree-ssa-loop-im.c: Likewise.
1181 * tree-ssa-loop-ivcanon.c: Likewise.
1182 * tree-ssa-loop-ivopts.c: Likewise.
1183 * tree-ssa-loop-manip.c: Likewise.
1184 * tree-ssa-loop-niter.c: Likewise.
1185 * tree-ssa-loop-prefetch.c: Likewise.
1186 * tree-ssa-loop.c: Likewise.
1187 * tree-ssa-math-opts.c: Likewise.
1188 * tree-ssa-phiopt.c: Likewise.
1189 * tree-ssa-phiprop.c: Likewise.
1190 * tree-ssa-pre.c: Likewise.
1191 * tree-ssa-propagate.c: Likewise.
1192 * tree-ssa-reassoc.c: Likewise.
1193 * tree-ssa-sink.c: Likewise.
1194 * tree-ssa-strlen.c: Likewise.
1195 * tree-ssa-structalias.c: Likewise.
1196 * tree-ssa-tail-merge.c: Likewise.
1197 * tree-ssa-ter.c: Likewise.
1198 * tree-ssa-threadedge.c: Likewise.
1199 * tree-ssa-threadupdate.c: Likewise.
1200 * tree-ssa-uncprop.c: Likewise.
1201 * tree-ssa-uninit.c: Likewise.
1202 * tree-ssa.c: Likewise.
1203 * tree-stdarg.c: Likewise.
1204 * tree-switch-conversion.c: Likewise.
1205 * tree-tailcall.c: Likewise.
1206 * tree-vect-data-refs.c: Likewise.
1207 * tree-vect-generic.c: Likewise.
1208 * tree-vect-loop-manip.c: Likewise.
1209 * tree-vect-loop.c: Likewise.
1210 * tree-vect-patterns.c: Likewise.
1211 * tree-vect-slp.c: Likewise.
1212 * tree-vect-stmts.c: Likewise.
1213 * tree-vectorizer.c: Likewise.
1214 * tree-vrp.c: Likewise.
1217 * value-prof.c: Likewise.
1218 * vtable-verify.c: Likewise.
1220 2013-11-13 Steven Bosscher <steven@gcc.gnu.org>
1222 * gimple-ssa-isolate-paths.c (pass_isolate_erroneous_paths): Comment
1225 2013-11-13 Jeff Law <law@redhat.com>
1227 * PR middle-end/59119
1228 * gimple-ssa-isolate-paths.c (find_implicit_erroneous_behaviour): New
1229 function, extracted from gimple_ssa_isolate_erroneous_paths.
1230 (find_explicit_erroneous_behaviour): Similarly.
1231 (insert_trap_and_remove_trailing_statements): Remove statements
1234 2013-11-13 Steven Bosscher <steven@gcc.gnu.org>
1236 * cfgrtl.c (can_fallthru): Reorder code to move tablejump check up.
1237 Make that check explicit. BB_HEAD cannot be NULL, remove check for it.
1238 * haifa-sched.c (ready_remove_first_dispatch): Check INSN_P before
1239 looking at INSN_CODE.
1240 * reload1.c (delete_dead_insn) Do not expect JUMP_TABLE_DATA to be an
1241 active_insn_p object, respect basic block boundaries.
1242 * reorg.c (follow_jumps): Use invariant that JUMP_TABLE_DATA always
1243 follows immediately after the jump table data label.
1244 * config/nds32/nds32.c (nds32_output_casesi_pc_relative): Likewise.
1245 * config/sh/sh.c (barrier_align): Likewise. Rearrange code such
1246 that JUMP_TABLE_DATA is not expected to be an active_insn_p object.
1248 2013-11-13 Teresa Johnson <tejohnson@google.com>
1251 * predict.c (drop_profile): Error is currently too strict.
1252 (handle_missing_profiles): Pass call_count to drop_profile.
1254 2013-11-13 Teresa Johnson <tejohnson@google.com>
1257 * ipa-inline.c (edge_badness): Fix overflow.
1259 2013-11-13 Vladimir Makarov <vmakarov@redhat.com>
1261 PR rtl-optimization/59036
1262 * ira-color.c (struct allocno_color_data): Add new members
1263 first_thread_allocno, next_thread_allocno, thread_freq.
1264 (sorted_copies): New static var.
1265 (allocnos_conflict_by_live_ranges_p, copy_freq_compare_func): Move up.
1266 (allocno_thread_conflict_p, merge_threads)
1267 (form_threads_from_copies, form_threads_from_bucket)
1268 (form_threads_from_colorable_allocno, init_allocno_threads): New
1270 (bucket_allocno_compare_func): Add comparison by thread frequency
1272 (add_allocno_to_ordered_bucket): Rename to
1273 add_allocno_to_ordered_colorable_bucket. Remove parameter.
1274 (push_only_colorable): Call form_threads_from_bucket.
1275 (color_pass): Call init_allocno_threads. Use
1276 consideration_allocno_bitmap instead of coloring_allocno_bitmap
1277 for nuillify allocno color data.
1278 (ira_initiate_assign, ira_finish_assign): Allocate/free sorted_copies.
1279 (coalesce_allocnos): Use static sorted copies.
1281 2013-11-13 Jakub Jelinek <jakub@redhat.com>
1283 * passes.c (execute_todo): Don't call do_per_function if
1285 (execute_one_ipa_transform_pass, execute_one_pass): Don't call
1286 execute_function_dump if dump_file is NULL.
1288 2013-11-13 Martin Jambor <mjambor@suse.cz>
1290 * cgraph.c (cgraph_get_create_node): Do what
1291 cgraph_get_create_real_symbol_node used to do.
1292 (cgraph_get_create_real_symbol_node): Removed. Changed all users to
1293 call cgraph_get_create_node.
1294 * cgraph.h (cgraph_get_create_real_symbol_node): Removed.
1295 * lto-streamer-in.c (input_function): Call cgraph_get_node instead of
1296 cgraph_get_create_node. Assert we get a node.
1298 2013-11-13 Tejas Belagod <tejas.belagod@arm.com>
1300 * config/aarch64/aarch64-simd.md (vec_extract): New.
1302 2013-11-13 Tejas Belagod <tejas.belagod@arm.com>
1304 * config/aarch64/aarch64-simd.md (vec_set<mode>): Add w -> w option to
1307 2013-11-13 Eric Botcazou <ebotcazou@adacore.com>
1309 * cfgexpand.c (expand_used_vars): Allocate space for partitions based
1310 on PARM_DECLs or RESULT_DECLs only if they are ignored for debug info
1311 or if optimization is enabled.
1312 * tree-ssa-coalesce.c (coalesce_ssa_name): If optimization is disabled,
1313 require that all the names based on a PARM_DECL or a RESULT_DECL that
1314 isn't ignored for debug info be coalesced.
1316 2013-11-13 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1318 * config/c6x/c6x.c: Include "gimple-expr.h".
1320 2013-11-13 Richard Biener <rguenther@suse.de>
1322 * gimple-streamer-out.c (output_gimple_stmt): Also wrap
1323 decls in ADDR_EXPR operands inside a MEM_REF and optimize that.
1324 * gimple-streamer-in.c (input_gimple_stmt): Remove now dead code
1325 dealing with type mismatches inside component reference chains.
1327 2013-11-13 Marc Glisse <marc.glisse@inria.fr>
1329 PR tree-optimization/59077
1330 * ipa-pure-const.c (better_state): Update *state.
1332 2013-11-13 Christophe Lyon <christophe.lyon@linaro.org>
1334 * config/aarch64/aarch64.h (FRAME_GROWS_DOWNWARD): Define to 1.
1335 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
1336 Update offset calculations.
1338 2013-11-13 Eric Botcazou <ebotcazou@adacore.com>
1341 * dwarf2out.c (add_byte_size_attribute): Also use int_size_in_bytes
1342 for fields. Do not add the attribute if the size is negative.
1344 2013-11-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1346 * config/arm/arm.c: Include aarch-cost-tables.h.
1347 (generic_extra_costs): Move from here...
1348 * config/arm/aarch-cost-tables.h: ... To here. New file.
1350 2013-11-13 Alexander Ivchenko <alexander.ivchenko@intel.com>
1351 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1352 Sergey Lega <sergey.s.lega@intel.com>
1353 Anna Tikhonova <anna.tikhonova@intel.com>
1354 Ilya Tocar <ilya.tocar@intel.com>
1355 Andrey Turetskiy <andrey.turetskiy@intel.com>
1356 Ilya Verbin <ilya.verbin@intel.com>
1357 Kirill Yukhin <kirill.yukhin@intel.com>
1358 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1360 * config/i386/i386.c (ix86_print_operand): Support z-masking.
1361 * config/i386/predicate.md (const_0_to_4_operand): New.
1362 (const_0_to_5_operand): Ditto.
1363 * config/i386/sse.md (UNSPEC_COMPRESS): New.
1364 (UNSPEC_COMPRESS_STORE): Ditto.
1365 (UNSPEC_EXPAND): Ditto.
1366 (UNSPEC_EMBEDDED_ROUNDING): Ditto.
1367 (define_mode_attr ssescalarsize): Ditto.
1368 (avx512f_load<mode>_mask): Ditto.
1369 (avx512f_store<mode>_mask): Ditto.
1370 (avx512f_storedqu<mode>_mask): Ditto.
1371 (avx512f_vmcmp<mode>3_mask): Ditto.
1372 (avx512f_fmadd_<mode>_mask): Ditto.
1373 (avx512f_fmadd_<mode>_mask3): Ditto.
1374 (avx512f_fmsub_<mode>_mask): Ditto.
1375 (avx512f_fmsub_<mode>_mask3): Ditto.
1376 (avx512f_fnmadd_<mode>_mask): Ditto.
1377 (avx512f_fnmadd_<mode>_mask3): Ditto.
1378 (avx512f_fnmsub_<mode>_mask): Ditto.
1379 (avx512f_fnmsub_<mode>_mask3): Ditto.
1380 (avx512f_fmaddsub_<mode>_mask): Ditto.
1381 (avx512f_fmaddsub_<mode>_mask3): Ditto.
1382 (avx512f_fmsubadd_<mode>_mask): Ditto.
1383 (avx512f_fmsubadd_<mode>_mask3): Ditto.
1384 (vec_unpacku_float_lo_v16si): Ditto.
1385 (avx512f_vextract<shuffletype>32x4_mask): Ditto.
1386 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
1387 (avx512f_vextract<shuffletype>64x4_mask): Ditto.
1388 (vec_extract_lo_<mode>_maskm): Ditto.
1389 (vec_extract_hi_<mode>_maskm): Ditto.
1390 (avx512f_vternlog<mode>_mask): Ditto.
1391 (avx512f_shufps512_mask): Ditto.
1392 (avx512f_fixupimm<mode>_mask): Ditto.
1393 (avx512f_shufpd512_mask): Ditto.
1394 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
1395 (avx512f_<code>v8div16qi2_mask/trunc): Ditto.
1396 (*avx512f_<code>v8div16qi2_store_mask): Ditto.
1397 (ashr<mode>3<mask_name>): Ditto.
1398 (avx512f_vinsert<shuffletype>32x4_mask): Ditto.
1399 (avx512f_vinsert<shuffletype>64x4_mask): Ditto.
1400 (avx512f_shuf_<shuffletype>64x2_mask): Ditto.
1401 (avx512f_shuf_<shuffletype>32x4_mask): Ditto.
1402 (avx512f_pshufdv3_mask): Ditto.
1403 (avx512f_perm<mode>_mask): Ditto.
1404 (avx512f_vpermi2var<mode>3_mask): Ditto.
1405 (avx512f_vpermt2var<mode>3_mask): Ditto.
1406 (avx512f_compress<mode>_mask): Ditto.
1407 (avx512f_compressstore<mode>_mask): Ditto.
1408 (avx512f_expand<mode>_mask): Ditto.
1409 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>): Extend
1411 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
1412 (<plusminus_insn><mode>3<mask_name>): Ditto.
1413 (*<plusminus_insn><mode>3<mask_name>): Ditto.
1414 (mul<mode>3<mask_name>): Ditto.
1415 (*mul<mode>3<mask_name>): Ditto.
1416 (<sse>_div<mode>3<mask_name>): Ditto.
1417 (<mask_codefor>rcp14<mode><mask_name>): Ditto.
1418 (<sse>_sqrt<mode>2<mask_name>): Ditto.
1419 (<mask_codefor>rsqrt14<mode><mask_name>): Ditto.
1420 (<code><mode>3<mask_name>/smaxmin): Ditto.
1421 (*<code><mode>3_finite<mask_name>/smaxmin): Ditto.
1422 (*<code><mode>3<mask_name>/smaxmin): Ditto.
1423 (float<sseintvecmodelower><mode>2<mask_name>): Ditto.
1424 (ufloatv16siv16sf2<mask_name>): Ditto.
1425 (<mask_codefor>avx512f_fix_notruncv16sfv16si<mask_name>): Ditto.
1426 (<mask_codefor>avx512f_ufix_notruncv16sfv16si<mask_name>): Ditto.
1427 (<fixsuffix>fix_truncv16sfv16si2<mask_name>): Ditto.
1428 (float<si2dfmodelower><mode>2<mask_name>): Ditto.
1429 (ufloatv8siv8df<mask_name>): Ditto.
1430 (<mask_codefor>avx512f_cvtpd2dq512<mask_name>): Ditto.
1431 (avx512f_ufix_notruncv8dfv8si<mask_name>): Ditto.
1432 (<fixsuffix>fix_truncv8dfv8si2<mask_name>): Ditto.
1433 (<mask_codefor>avx512f_cvtpd2ps512<mask_name>): Ditto.
1434 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix><mask_name>): Ditto.
1435 (<mask_codefor>avx512f_unpckhps512<mask_name>): Ditto.
1436 (<mask_codefor>avx512f_unpcklps512<mask_name>): Ditto.
1437 (<mask_codefor>avx512f_movshdup512<mask_name>): Ditto.
1438 (<mask_codefor>avx512f_movsldup512<mask_name>): Ditto.
1439 (<mask_codefor>avx512f_vextract<shuffletype>32x4_1<mask_name>): Ditto.
1440 (vec_extract_lo_<mode><mask_name>): Ditto.
1441 (vec_extract_hi_<mode><mask_name>): Ditto.
1442 (<mask_codefor>avx512f_unpckhpd512<mask_name>): Ditto.
1443 (avx512f_movddup512<mask_name>): Ditto.
1444 (avx512f_unpcklpd512<mask_name>): Ditto.
1445 (*avx512f_unpcklpd512<mask_name>): Ditto.
1446 (*avx512f_vmscalef<mode>): Ditto.
1447 (avx512f_scalef<mode><mask_name>): Ditto.
1448 (avx512f_getexp<mode><mask_name>): Ditto.
1449 (<mask_codefor>avx512f_align<mode><mask_name>): Ditto.
1450 (avx512f_rndscale<mode><mask_name>): Ditto.
1451 (avx512f_shufps512_1<mask_name>): Ditto.
1452 (avx512f_shufpd512_1<mask_name>): Ditto.
1453 (<plusminus_insn><mode>3<mask_name>): Ditto.
1454 (*<plusminus_insn><mode>3<mask_name>): Ditto.
1455 (vec_widen_umult_even_v16si<mask_name>): Ditto.
1456 (*vec_widen_umult_even_v16si<mask_name>): Ditto.
1457 (vec_widen_smult_even_v16si<mask_name>): Ditto.
1458 (*vec_widen_smult_even_v16si<mask_name>): Ditto.
1459 (mul<mode>3<mask_name>): Ditto.
1460 (*<sse4_1_avx2>_mul<mode>3<mask_name>): Ditto.
1461 (<shift_insn><mode>3<mask_name>): Ditto.
1462 (avx512f_<rotate>v<mode><mask_name>/rotate): Ditto.
1463 (avx512f_<rotate><mode><mask_name>): Ditto.
1464 (<code><mode>3<mask_name>/maxmin): Ditto.
1465 (*avx2_<code><mode>3<mask_name>/maxmin): Ditto.
1466 (<sse2_avx2>_andnot<mode>3<mask_name>): Ditto.
1467 (*andnot<mode>3<mask_name>): Ditto.
1468 (<mask_codefor><code><mode>3<mask_name>/any_logic): Ditto.
1469 (<mask_codefor>avx512f_interleave_highv16si<mask_name>): Ditto.
1470 (<mask_codefor>avx512f_interleave_lowv16si<mask_name>): Ditto.
1471 (<mask_codefor>avx512f_vinsert<shuffletype>32x4_1<mask_name>): Ditto.
1472 (vec_set_lo_<mode><mask_name>): Ditto.
1473 (vec_set_hi_<mode><mask_name>): Ditto.
1474 (avx512f_shuf_<shuffletype>64x2_1<mask_name>): Ditto.
1475 (avx512f_shuf_<shuffletype>32x4_1<mask_name>): Ditto.
1476 (avx512f_pshufd_1<mask_name>): Ditto.
1477 (<mask_codefor>abs<mode>2<mask_name>): Ditto.
1478 (<mask_codefor>avx512f_<code>v16qiv16si2<mask_name>): Ditto.
1479 (avx512f_<code>v16hiv16si2<mask_name>/any_extend): Ditto.
1480 (avx512f_<code>v8qiv8di2<mask_name>/any_extend): Ditto.
1481 (avx512f_<code>v8hiv8di2<mask_name>/any_extend): Ditto.
1482 (avx512f_<code>v8siv8di2<mask_name>/any_extend): Ditto.
1483 (avx512er_exp2<mode><mask_name>): Ditto.
1484 (<mask_codefor>avx512er_rcp28<mode><mask_name>): Ditto.
1485 (<mask_codefor>avx512er_rsqrt28<mode><mask_name>): Ditto.
1486 (<avx2_avx512f>_permvar<mode><mask_name>): Ditto.
1487 (<avx2_avx512f>_perm<mode>_1<mask_name>): Ditto.
1488 (<mask_codefor>avx512f_vec_dup<mode><mask_name>): Ditto.
1489 (<mask_codefor>avx512f_broadcast<mode><mask_name>/V16FI): Ditto.
1490 (<mask_codefor>avx512f_broadcast<mode><mask_name>/V8FI): Ditto.
1491 (<mask_codefor>avx512f_vec_dup_gpr<mode><mask_name>): Ditto.
1492 (<mask_codefor>avx512f_vec_dup_mem<mode><mask_name>): Ditto.
1493 (<sse2_avx_avx512f>_vpermil<mode><mask_name>/VF2): Ditto.
1494 (<sse2_avx_avx512f>_vpermil<mode><mask_name>/VF1): Ditto.
1495 (*<sse2_avx_avx512f>_vpermilp<mode><mask_name>): Ditto.
1496 (<sse2_avx_avx512f>_vpermilvar<mode>3<mask_name>): Ditto.
1497 (<avx2_avx512f>_ashrv<mode><mask_name>): Ditto.
1498 (<avx2_avx512f>_<shift_insn>v<mode><mask_name>): Ditto.
1499 (<mask_codefor>avx512f_vcvtph2ps512<mask_name>): Ditto.
1500 (<mask_codefor>avx512f_vcvtps2ph512<mask_name>): Ditto.
1501 (avx512f_getmant<mode><mask_name>): Ditto.
1502 (clz<mode>2<mask_name>): Ditto.
1503 (<mask_codefor>conflict<mode><mask_name>): Ditto.
1504 (*srcp14<mode>): Remove visibility.
1505 (*rsqrt14<mode>): Ditto.
1506 (*fma_fmsub_<mode>): Ditto.
1507 (*fma_fnmadd_<mode>): Ditto.
1508 (*avx512f_rndscale<mode>): Ditto.
1509 * config/i386/subst.md: New file.
1511 2013-11-13 Joseph Myers <joseph@codesourcery.com>
1513 * doc/extend.texi (Statement Exprs, Typeof): Discuss __auto_type.
1514 * ginclude/stdatomic.h (kill_dependency, atomic_store_explicit)
1515 (atomic_load_explicit, atomic_exchange_explicit)
1516 (atomic_compare_exchange_strong_explicit)
1517 (atomic_compare_exchange_weak_explicit): Use __auto_type to
1518 declare variable initialized with PTR argument.
1520 2013-11-12 Jeff Law <law@redhat.com>
1522 * tree-ssa-threadedge.c (thread_around_empty_blocks): New argument
1523 backedge_seen_p. Set, use and pass it to children appropriately.
1524 (thread_through_normal_block): Similarly.
1525 (thread_across_edge): Similarly.
1527 * gimple-ssa-isolate-paths.c (check_loadstore): Mark discovered
1528 memory references as volatile.
1529 (insert_trap_and_remove_trailing_statements): Fix comment.
1531 2013-11-12 Vladimir Makarov <vmakarov@redhat.com>
1534 * ira-costs.c (record_operand_costs): Check operands number for
1537 2013-11-12 Michael Meissner <meissner@linux.vnet.ibm.com>
1540 * config/rs6000/rs6000.md (movdi_internal32): Eliminate
1541 constraints that would allow DImode into the traditional Altivec
1542 registers, but cause undesirable code generation when loading 0 as
1544 (movdi_internal64): Likewise.
1545 (cmp<mode>_fpr): Do not use %x for CR register output.
1546 (extendsfdf2_fpr): Fix constraints when -mallow-upper-df and
1547 -mallow-upper-sf debug switches are used.
1549 2013-11-12 Andrew MacLeod <amacleod@redhat.com>
1551 * gimple-expr.h (create_tmp_var_name, create_tmp_var_raw,
1552 create_tmp_var, create_tmp_reg, mark_addressable, is_gimple_reg_rhs):
1553 Relocate prototypes from gimple.h.
1554 * gimplify.h: New File. Relocate some prototypes from gimple.h here.
1555 (gimple_predicate, enum fallback, enum gimplify_status): Relocate
1557 * gimple.h: Move some prototypes to gimplify.h.
1558 (gimple_predicate, enum fallback, enum gimplify_status): Move to
1560 (gimple_do_not_emit_location_p, gimple_set_do_not_emit_location):
1561 Relocate from gimpify.c.
1562 * gimple-expr.c (remove_suffix, tmp_var_id_num, create_tmp_var_name,
1563 create_tmp_var_raw, create_tmp_var, create_tmp_reg, mark_addressable,
1564 is_gimple_reg_rhs) Relocate from gimplify.c.
1565 * gimplify.c (mark_addressable): Move to gimple-expr.c.
1566 (gimple_seq_add_stmt_without_update): Move to gimple.c.
1567 (remove_suffix, tmp_var_id_num, create_tmp_var_name,
1568 create_tmp_var_raw, create_tmp_var, create_tmp_reg,
1569 is_gimple_reg_rhs): Move to gimple-expr.c.
1570 (should_carry_location_p): Move to gimple.c.
1571 (gimple_do_not_emit_location_p, gimple_set_do_not_emit_location): Move
1573 (annotate_one_with_location, annotate_all_with_location_after,
1574 annotate_all_with_location): Move to gimple.c.
1575 (compare_case_labels, sort_case_labels,
1576 preprocess_case_label_vec_for_gimple): Move to gimple.c.
1577 (rhs_predicate_for): Make static.
1578 (gimplify_assign): Relocate from gimple.c.
1579 * gimple.c (gimplify_assign): Move to gimplify.c.
1580 (gimple_seq_add_stmt_without_update, should_carry_location_p,
1581 annotate_one_with_location, annotate_all_with_location_after,
1582 annotate_all_with_location, compare_case_labels, sort_case_labels,
1583 preprocess_case_label_vec_for_gimple): Relocate from gimplify.c.
1584 * tree.h (unshare_expr, unshare_expr_without_location,
1585 mark_addressable): Move prototypes to gimplify.h.
1586 * Makefile.in (GTFILES): gimple-expr.c now has the GTY tag for
1588 * asan.c: Include gimplify.h rather than gimple.h.
1589 * cfgloopmanip.c: Likewise.
1590 * cgraphunit.c: Likewise.
1591 * cilk-common.c: Likewise.
1592 * dwarf2out.c: Dont include gimple.h.
1593 * fold-const.c: Include gimplify.h rather than gimple.h.
1594 * function.c: Likewise.
1595 * gimple-fold.c: Likewise.
1596 * gimple-ssa-strength-reduction.c: Likewise.
1597 * graphite-clast-to-gimple.c: Likewise.
1598 * graphite-sese-to-poly.c: Likewise.
1599 * ipa-prop.c: Likewise.
1600 * ipa-split.c: Likewise.
1602 * langhooks.c: Dont include gimple.h.
1603 * loop-init.c: Include gimplify.h rather than gimple.h.
1604 * omp-low.c: Likewise.
1606 * stor-layout.c: Likewise.
1607 * targhooks.c: Likewise.
1608 * trans-mem.c: Likewise.
1609 * tree-affine.c: Likewise.
1610 * tree-cfg.c: Likewise.
1611 * tree-cfgcleanup.c: Likewise.
1612 * tree-complex.c: Likewise.
1613 * tree-if-conv.c: Likewise.
1614 * tree-inline.c: Likewise.
1615 * tree-iterator.c: Likewise.
1616 * tree-loop-distribution.c: Likewise.
1617 * tree-nested.c: Likewise.
1618 * tree-parloops.c: Likewise.
1619 * tree-predcom.c: Likewise.
1620 * tree-profile.c: Likewise.
1621 * tree-scalar-evolution.c: Likewise.
1622 * tree-sra.c: Likewise.
1623 * tree-ssa-address.c: Likewise.
1624 * tree-ssa-ccp.c: Likewise.
1625 * tree-ssa-dce.c: Likewise.
1626 * tree-ssa-forwprop.c: Likewise.
1627 * tree-ssa-ifcombine.c: Likewise.
1628 * tree-ssa-loop-im.c: Likewise.
1629 * tree-ssa-loop-ivopts.c: Likewise.
1630 * tree-ssa-loop-manip.c: Likewise.
1631 * tree-ssa-loop-niter.c: Likewise.
1632 * tree-ssa-loop-prefetch.c: Likewise.
1633 * tree-ssa-loop-unswitch.c: Likewise.
1634 * tree-ssa-math-opts.c: Likewise.
1635 * tree-ssa-phiopt.c: Likewise.
1636 * tree-ssa-phiprop.c: Likewise.
1637 * tree-ssa-pre.c: Likewise.
1638 * tree-ssa-propagate.c: Likewise.
1639 * tree-ssa-reassoc.c: Likewise.
1640 * tree-ssa-sccvn.c: Likewise.
1641 * tree-ssa-strlen.c: Likewise.
1642 * tree-ssa.c: Likewise.
1643 * tree-switch-conversio: Likewise.n.c
1644 * tree-tailcall.c: Likewise.
1645 * tree-vect-data-refs.c: Likewise.
1646 * tree-vect-generic.c: Likewise.
1647 * tree-vect-loop-manip.c: Likewise.
1648 * tree-vect-loop.c: Likewise.
1649 * tree-vect-patterns.c: Likewise.
1650 * tree-vect-stmts.c: Likewise.
1652 * value-prof.c: Likewise.
1653 * config/aarch64/aarch64.c: Include gimplify.h instead of gimple.h.
1654 * config/alpha/alpha.c: Likewise.
1655 * config/darwin.c: Likewise.
1656 * config/i386/i386.c: Likewise.
1657 * config/ia64/ia64.c: Likewise.
1658 * config/mep/mep.c: Likewise.
1659 * config/mips/mips.c: Likewise.
1660 * config/rs6000/rs6000.c: Likewise.
1661 * config/s390/s390.c: Likewise.
1662 * config/sh/sh.c: Likewise.
1663 * config/sparc/sparc.c: Likewise.
1664 * config/spu/spu.c: Likewise.
1665 * config/stormy16/stormy16.c: Likewise.
1666 * config/tilegx/tilegx.c: Likewise.
1667 * config/tilepro/tilepro.c: Likewise.
1668 * config/xtensa/xtensa.c: Likewise.
1670 2013-11-12 Adam Butcher <adam@jessamine.co.uk>
1672 * tree.c (grow_tree_vec_stat): New function ...
1673 * tree.h (grow_tree_vec_stat) (grow_tree_vec): ... and its declaration
1674 and macro front-end.
1676 2013-11-12 Marek Polacek <polacek@redhat.com>
1678 * final.c (update_alignments): Initialize label to NULL_RTX.
1680 2013-11-12 Jeff Law <law@redhat.com>
1682 * gimple-ssa-isolate-paths.c (check_loadstore): New function.
1683 (insert_trap_and_remove_trailing_statements): New argument OP which
1684 is the NULL pointer. Emit the trap after the load/store through
1685 the NULL pointer. Simplify the RHS of a store through a NULL pointer
1686 when trivial to do so.
1687 (isolate_path): Corresponding changes.
1688 (gimple_ssa_isolate_erroneous_path): Likewise.
1690 2013-11-12 Teresa Johnson <tejohnson@google.com>
1691 Jan Hubicka <jh@suse.cz>
1693 * predict.c (drop_profile): New function.
1694 (handle_missing_profiles): Ditto.
1695 (counts_to_freqs): Don't overwrite estimated frequencies
1696 when function has no profile counts.
1697 * predict.h (handle_missing_profiles): Declare.
1698 * tree-inline.c (freqs_to_counts): New function.
1699 (copy_cfg_body): Invoke freqs_to_counts as needed.
1700 * tree-profile.c (tree_profiling): Invoke handle_missing_profiles.
1702 2013-11-12 H.J. Lu <hongjiu.lu@intel.com>
1705 * config/i386/x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL):
1707 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Set for m_HASWELL.
1709 2013-11-12 H.J. Lu <hongjiu.lu@intel.com>
1712 * config/i386/i386.c (ix86_option_override_internal): Check
1713 X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL and
1714 X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL for
1715 MASK_AVX256_SPLIT_UNALIGNED_LOAD and
1716 MASK_AVX256_SPLIT_UNALIGNED_STORE.
1718 * config/i386/x86-tune.def (X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL):
1719 Clear m_COREI7_AVX and update comments.
1720 (X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL): Likewise.
1722 2013-11-12 Martin Jambor <mjambor@suse.cz>
1724 PR rtl-optimization/10474
1725 * ira.c (interesting_dest_for_shprep): New function.
1726 (split_live_ranges_for_shrink_wrap): Likewise.
1727 (find_moveable_pseudos): Move calculation of dominance info,
1728 df_analysios and the final anlyses to...
1729 (ira): ...here, call split_live_ranges_for_shrink_wrap.
1731 2013-11-12 Bin Cheng <bin.cheng@arm.com>
1733 * tree-ssa-loop-ivopts.c (force_expr_to_var_cost): Refactor the code.
1734 Handle type conversion.
1736 2013-11-11 Martin Liska <marxin.liska@gmail.com>
1737 Jan Hubicka <jh@suse.cz>
1739 * cgraph.c (dump_cgraph_node): Profile dump added.
1740 * cgraph.h (struct cgraph_node): New time profile variable added.
1741 * cgraphclones.c (cgraph_clone_node): Time profile is cloned.
1742 * gcov-io.h (gcov_type): New profiler type introduced.
1743 * ipa-profile.c (lto_output_node): Streaming for time profile added.
1744 (input_node): Time profiler is read from LTO stream.
1745 * predict.c (maybe_hot_count_p): Hot prediction changed.
1746 * profile.c (instrument_values): New case for time profiler added.
1747 (compute_value_histograms): Read of time profile.
1748 * tree-pretty-print.c (dump_function_header): Time profiler is dumped.
1749 * tree-profile.c (init_ic_make_global_vars): Time profiler
1751 (gimple_init_edge_profiler): TP function instrumentation.
1752 (gimple_gen_time_profiler): New.
1753 * value-prof.c (gimple_add_histogram_value): Support for time profiler
1755 (dump_histogram_value): TP type added to dumps.
1756 (visit_hist): More sensitive check that takes TP into account.
1757 (gimple_find_values_to_profile): TP instrumentation.
1758 * value-prof.h (hist_type): New histogram type added.
1759 (struct histogram_value_t): Pointer to struct function added.
1760 * libgcc/Makefile.in: New GCOV merge function for TP added.
1761 * libgcov.c: function_counter variable introduced.
1762 (_gcov_merge_time_profile): New.
1763 (_gcov_time_profiler): New.
1765 2013-11-11 Marc Glisse <marc.glisse@inria.fr>
1766 Jeff Law <law@redhat.com>
1768 * tree-ssa-alias.c (stmt_kills_ref_p_1): Use
1769 ao_ref_init_from_ptr_and_size for builtins.
1771 2013-11-11 Uros Bizjak <ubizjak@gmail.com>
1772 H.J. Lu <hongjiu.lu@intel.com>
1775 * config/i386/x86-tune.def
1776 (X86_TUNE_MISALIGNED_MOVE_STRING_PRO_EPILOGUES): Rename from
1777 TARGET_MISALIGNED_MOVE_STRING_PROLOGUES.
1778 * config/i386/i386.h
1779 (TARGET_MISALIGNED_MOVE_STRING_PRO_EPILOGUES): Rename from
1780 TARGET_MISALIGNED_MOVE_STRING_PROLOGUES_EPILOGUES. Update for renamed
1781 X86_TUNE_MISALIGNED_MOVE_STRING_PRO_EPILOGUES.
1782 * config/i386/i386.c (ix86_expand_set_or_movmem): Use
1783 TARGET_MISALIGNED_MOVE_STRING_PRO_EPILOGUES to calculate
1784 misaligned_prologue_used. Check that
1785 desired_aling <= epilogue_size_needed.
1787 2013-11-11 Cong Hou <congh@google.com>
1789 PR tree-optimization/59050
1790 * tree-vect-data-refs.c (comp_dr_addr_with_seg_len_pair): Bug fix.
1792 2013-11-11 Joern Rennecke <joern.rennecke@embecosm.com>
1795 * expmed.c (emit_store_flag): Fail for const-const comparison.
1797 2013-11-11 Tristan Gingold <gingold@adacore.com>
1798 Eric Botcazou <ebotcazou@adacore.com>
1800 * tree.h (CONSTRUCTOR_NO_CLEARING): Define.
1801 * tree-core.h (CONSTRUCTOR_NO_CLEARING): Document it.
1802 * tree.def (CONSTRUCTOR): Likewise.
1803 * doc/generic.texi (CONSTRUCTOR): Likewise. Update description.
1804 * gimplify.c (gimplify_init_constructor): Do not clear the object when
1805 the constructor is incomplete and CONSTRUCTOR_NO_CLEARING is set.
1807 2013-11-11 Basile Starynkevitch <basile@starynkevitch.net>
1809 * toplev.c (toplev_main): Move PLUGIN_FINISH invocation before
1812 2013-11-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1814 * config/arm/arm.c (arm_new_rtx_costs): Return after handling
1817 2013-11-11 Joern Rennecke <joern.rennecke@embecosm.com>
1819 * config/arc/arc.h (LOGICAL_OP_NON_SHORT_CIRCUIT): Define.
1821 2013-11-08 Jeff Law <law@redhat.com>
1823 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
1824 threading paths first, then perform PHI node checks if applicable.
1826 2013-11-10 Karlson2k <k2k@narod.ru>
1827 Kai Tietz <ktietz@redhat.com>
1830 * configure.ac: Adding for exported symbols check
1831 and for rdynamic-check executable-extension.
1832 * configure: Regenerated.
1834 2013-11-10 Uros Bizjak <ubizjak@gmail.com>
1836 * mode-switching.c (optimize_mode_switching): Mark block as
1837 nontransparent, if last_mode at block exit is different from no_mode.
1839 2013-11-09 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1841 * function.c (NAME__MAIN): Move to...
1842 * cfgexpand.c (NAME__MAIN): ...here.
1844 2013-11-09 Richard Sandiford <rdsandiford@googlemail.com>
1846 * target.def (can_use_doloop_p): New hook.
1847 * doc/tm.texi.in (TARGET_CAN_USE_DOLOOP_P): Add.
1848 * doc/tm.texi: Regenerate.
1849 * doc/md.texi (doloop_begin, doloop_end): Update documentation.
1850 * hooks.h (hook_bool_dint_dint_uint_true): Declare.
1851 * hooks.c (hook_bool_dint_dint_uint_true): New function.
1852 * targhooks.h (can_use_doloop_if_innermost): Declare.
1853 * targhooks.c (can_use_doloop_if_innermost): New function.
1854 * target.h: Include double-int.h.
1855 * loop-doloop.c (doloop_optimize): Call targetm.can_use_doloop_p.
1856 Remove iteration count, maximum iteration count, loop depth and
1857 enter-at-top inputs from doloop_begin and doloop_end.
1858 * config/arc/arc.md (doloop_begin, doloop_end): Update for new
1860 * config/arc/arc.c (arc_can_use_doloop_p): New function.
1861 (TARGET_CAN_USE_DOLOOP_P): Define.
1862 * config/arm/thumb2.md (doloop_end): Update for new interface.
1863 * config/arm/arm.c (TARGET_CAN_USE_DOLOOP_P): Define.
1864 * config/bfin/bfin.md (doloop_end): Update for new interface.
1865 * config/bfin/bfin.c (bfin_can_use_doloop_p): New function.
1866 (TARGET_CAN_USE_DOLOOP_P): Define.
1867 * config/c6x/c6x.md (doloop_end): Update for new interface.
1868 * config/ia64/ia64.md (doloop_end): Update for new interface.
1869 * config/ia64/ia64.c (TARGET_CAN_USE_DOLOOP_P): Define.
1870 * config/mep/mep.md (doloop_begin, doloop_end): Update for new
1872 * config/mep/mep.c (mep_emit_doloop): Likewise.
1873 (TARGET_CAN_USE_DOLOOP_P): Define.
1874 * config/rs6000/rs6000.md (doloop_end): Update for new interface.
1875 * config/rs6000/rs6000.c (TARGET_CAN_USE_DOLOOP_P): Define.
1876 * config/s390/s390.md (doloop_end): Update for new interface.
1877 * config/sh/sh.md (doloop_end): Likewise.
1878 * config/spu/spu.md (doloop_end): Likewise.
1879 * config/spu/spu.c (TARGET_CAN_USE_DOLOOP_P): Define.
1880 * config/tilegx/tilegx.md (doloop_end): Update for new interface.
1881 * config/tilegx/tilegx.c (TARGET_CAN_USE_DOLOOP_P): Define.
1882 * config/tilepro/tilepro.md (doloop_end): Update for new interface.
1883 * config/tilepro/tilepro.c (TARGET_CAN_USE_DOLOOP_P): Define.
1884 * config/v850/v850.md (doloop_begin, doloop_end): Update for new
1886 * config/v850/v850.c (TARGET_CAN_USE_DOLOOP_P): Define.
1888 2013-11-08 H.J. Lu <hongjiu.lu@intel.com>
1891 * doc/extend.texi: Move Cilk Plus Builtins node before Other
1894 2013-11-08 Andrew MacLeod <amacleod@redhat.com>
1895 Joseph Myers <joseph@codesourcery.com>
1897 * ginclude/stdatomic.h: New file.
1898 * Makefile.in (USER_H): Add stdatomic.h.
1900 2013-11-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1902 * config/arm/arm.c (arm_new_rtx_costs): Break after handling
1905 2013-11-08 Jeff Law <law@redhat.com>
1907 * tree-ssa-threadupdate.h (delete_thread_path): Declare.
1908 * tree-ssa-threadupdate.c (delete_thread_path): New function.
1909 (ssa_redirect_edges, thread_block_1): Use it.
1910 (thread_through_loop_header, mark_threaded_blocks): Likewise.
1911 (thread_through_all_blocks, register_jump_thread): Likewise.
1912 * tree-ssa-threadedge.c (thread_across_edge): Likewise.
1914 2013-11-08 James Greenhalgh <james.greenhalgh@arm.com>
1916 * config/arm/aarch-common.c
1917 (search_term): New typedef.
1918 (shift_rtx_costs): New array.
1919 (arm_rtx_shift_left_p): New.
1920 (arm_find_sub_rtx_with_search_term): Likewise.
1921 (arm_find_sub_rtx_with_code): Likewise.
1922 (arm_early_load_addr_dep): Add sanity checking.
1923 (arm_no_early_alu_shift_dep): Likewise.
1924 (arm_no_early_alu_shift_value_dep): Likewise.
1925 (arm_no_early_mul_dep): Likewise.
1926 (arm_no_early_store_addr_dep): Likewise.
1928 2013-11-08 Richard Biener <rguenther@suse.de>
1930 PR tree-optimization/59047
1931 * tree-predcom.c (ref_at_iteration): Handle bitfield accesses properly.
1933 2013-11-08 Ilya Enkovich <ilya.enkovich@intel.com>
1935 * common.opt (fcheck-pointer-bounds): Move to ...
1936 * c-family/c.opt: ... here.
1937 * langhooks-def.h (LANG_HOOKS_CHKP_SUPPORTED): Remove.
1938 (LANG_HOOKS_INITIALIZER): Remove LANG_HOOKS_CHKP_SUPPORTED.
1939 * langhooks.h (lang_hooks): Remove chkp_supported field.
1940 * toplev.c (process_options): Remove chkp_supported check.
1942 2013-11-08 Richard Biener <rguenther@suse.de>
1944 PR tree-optimization/59038
1945 PR tree-optimization/58955
1946 * tree-loop-distribution.c (pg_add_dependence_edges): Revert
1947 previous change. Handle known dependences correctly.
1949 2013-11-08 Tom de Vries <tom@codesourcery.com>
1951 * config/rs6000/t-xilinx: Remove duplicate contents.
1953 2013-11-07 Andrew MacLeod <amacleod@redhat.com>
1954 Joseph Myers <joseph@codesourcery.com>
1956 * tree-core.h (enum cv_qualifier): Add TYPE_QUAL_ATOMIC.
1957 (enum tree_index): Add TI_ATOMICQI_TYPE, TI_ATOMICHI_TYPE,
1958 TI_ATOMICSI_TYPE, TI_ATOMICDI_TYPE and TI_ATOMICTI_TYPE.
1959 (struct tree_base): Add atomic_flag field.
1960 * tree.h (TYPE_ATOMIC): New accessor macro.
1961 (TYPE_QUALS, TYPE_QUALS_NO_ADDR_SPACE): Add TYPE_QUAL_ATOMIC.
1962 (TYPE_QUALS_NO_ADDR_SPACE_NO_ATOMIC): New macro.
1963 (atomicQI_type_node, atomicHI_type_node, atomicSI_type_node)
1964 (atomicDI_type_node, atomicTI_type_node): New macros for type nodes.
1965 * tree.c (set_type_quals): Set TYPE_ATOMIC.
1966 (find_atomic_core_type): New function.
1967 (build_qualified_type): Adjust alignment for qualified types.
1968 (build_atomic_base): New function
1969 (build_common_tree_nodes): Build atomicQI_type_node,
1970 atomicHI_type_node, atomicSI_type_node, atomicDI_type_node and
1972 * print-tree.c (print_node): Print atomic qualifier.
1973 * tree-pretty-print.c (dump_generic_node): Print atomic type attribute.
1974 * target.def (atomic_assign_expand_fenv): New hook.
1975 * doc/tm.texi.in (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): New @hook.
1976 * doc/tm.texi: Regenerate.
1977 * targhooks.c (default_atomic_assign_expand_fenv): New function.
1978 * targhooks.h (default_atomic_assign_expand_fenv): Declare.
1979 * sync-builtins.def (__atomic_feraiseexcept): New built-in function.
1980 * config/i386/i386-builtin-types.def (VOID_FTYPE_PUSHORT): New
1982 * config/i386/i386.c (enum ix86_builtins): Add
1983 IX86_BUILTIN_FNSTENV, IX86_BUILTIN_FLDENV, IX86_BUILTIN_FNSTSW and
1984 IX86_BUILTIN_FNCLEX.
1985 (bdesc_special_args): Add __builtin_ia32_fnstenv,
1986 __builtin_ia32_fldenv, __builtin_ia32_fnstsw and __builtin_ia32_fnclex.
1987 (ix86_expand_builtin): Handle the new built-in functions.
1988 (ix86_atomic_assign_expand_fenv): New function.
1989 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): New macro.
1990 * config/i386/i386.md (UNSPECV_FNSTENV, UNSPECV_FLDENV)
1991 (UNSPECV_FNSTSW, UNSPECV_FNCLEX): New unspecs.
1992 (fnstenv, fldenv, fnstsw, fnclex): New insns.
1994 2013-11-07 Steve Ellcey <sellcey@mips.com>
1996 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add fp64 directory.
1997 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add -mfp64 flag.
1998 (MULTILIB_DIRNAMES): Add fp64 directory.
1999 (MULTILIB_EXCEPTIONS): Add new exclusions.
2001 2013-11-07 Aldy Hernandez <aldyh@redhat.com>
2003 * gimplify.c (gimple_regimplify_operands): Do not set
2005 * graphite-sese-to-poly.c (remove_simple_copy_phi): Same.
2006 (rewrite_close_phi_out_of_ssa): Same.
2007 (rewrite_phi_out_of_ssa): Same.
2008 (rewrite_degenerate_phi): Same.
2009 (handle_scalar_deps_crossing_scop_limits): Same.
2010 * tree-if-conv.c (predicate_scalar_phi): Same.
2011 * tree-parloops.c (create_loads_for_reductions): Same.
2012 (create_final_loads_for_reduction): Same.
2013 (create_loads_and_stores_for_name): Same.
2014 (transform_to_exit_first_loop): Same.
2015 (create_parallel_loop): Same.
2016 * tree-ssa-loop-im.c
2017 (move_computations_dom_walker::before_dom_children): Same.
2018 * tree-ssa-loop-manip.c (rewrite_phi_with_iv): Same.
2019 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children): Same.
2020 * tree-ssa-propagate.c (substitute_and_fold): Same.
2021 * tree-vect-loop.c (vect_finalize_reduction): Same.
2022 * tree-vect-stmts.c (vectorizable_call): Same.
2024 2013-11-07 Mike Stump <mikestump@comcast.net>
2026 * config/pdp11/pdp11.c: Include dbxout.h.
2027 * config/picochip/picochip.c: Likewise.
2029 2013-11-07 Cong Hou <congh@google.com>
2031 PR tree-optimization/56764
2032 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
2033 Combine alias checks if it is possible to amortize the runtime
2034 overhead. Return the number of alias checks after merging.
2035 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
2036 Use the function vect_create_cond_for_alias_checks () to check
2037 the number of alias checks.
2039 2013-11-07 Jeff Law <law@redhat.com>
2041 * varpool.c (ctor_for_folding): Fix typo in comment.
2043 2013-11-07 Joern Rennecke <joern.rennecke@embecosm.com>
2045 * config/arc/arc.c (arc_ifcvt): Use commutativity, e.g.:
2046 reg_a := reg_b + reg_a ==> reg_a := reg_a + reg_b
2048 2013-11-07 Jeff Law <law@redhat.com>
2050 * doc/invoke.texi (-fisolate-erroneous-paths): Document.
2052 * gimple-ssa-isolate-paths.c (gate_isolate_erroneous_paths):
2053 No longer check if we have __builtin_trap, assume it's available.
2055 2013-11-07 Diego Novillo <dnovillo@google.com>
2057 * attribs.c (lookup_scoped_attribute_spec): Make static.
2058 (get_attribute_namespace): Likewise.
2059 * builtins.c (more_const_call_expr_args_p): Move from tree.h.
2060 (validate_arglist): Move earlier in the file. Make static.
2061 (expand_stack_restore): Move from stmt.c
2062 (expand_stack_save): Move from stmt.c
2063 (rewrite_call_expr_array): Move earlier in the file.
2064 (rewrite_call_expr_valist): Likewise.
2065 * cfgexpand.c: Include hard-reg-set.h before tree.h
2068 (expand_asm_loc): Move from stmt.c.
2069 (n_occurrences): Move from stmt.c.
2070 (check_operand_nalternatives): Move from stmt.c.
2071 (tree_conflicts_with_clobbers_p): Move from stmt.c.
2072 (expand_asm_operands): Move from stmt.c
2073 (expand_asm_stmt): Move from stmt.c
2074 (expand_computed_goto): Move from stmt.c
2075 (expand_goto): Move from stmt.c
2076 (expand_null_return_1): Move from stmt.c
2077 (expand_null_return): Move from stmt.c
2078 (expand_value_return): Move from stmt.c
2079 (expand_return): Move from stmt.c
2080 (expand_main_function): Move from function.c
2081 (stack_protect_prologue): Move from function.c
2082 * cgraphclones.c (build_function_type_skip_args): Move from tree.c.
2083 (build_function_decl_skip_args): Move from tree.c.
2084 * explow.c (tree_expr_size): Move from tree.c.
2085 * expr.c (addr_expr_of_non_mem_decl_p): Remove.
2086 (fields_length): Move from tree.c.
2087 * fold-const.c (size_low_cst): Move from tree.c.
2088 (tree_expr_nonzero_warnv_p): Make static. Move earlier in the file.
2089 (tree_expr_nonzero_p): Make static. Move earlier in the file.
2090 (fold_build3_initializer_loc): Remove.
2091 (tree_invalid_nonnegative_warnv_p): Make static.
2092 * function.c (expand_main_function): Move to cfgexpand.c.
2093 (stack_protect_prologue): Move to cfgexpand.c.
2094 (set_insn_locations): Move earlier in the file.
2095 * gimple-fold.c: Include langhooks.h.
2096 (truth_type_for): Move from tree.c.
2097 * print-tree.c (print_vec_tree): Remove.
2098 * stmt.c (expand_computed_goto): Move to cfgexpand.c.
2099 (expand_goto): Move to cfgexpand.c.
2100 (n_occurrences): Move to cfgexpand.c.
2101 (expand_asm_loc): Move to cfgexpand.c
2102 (tree_conflicts_with_clobbers_p): Move to cfgexpand.c.
2103 (expand_asm_operands): Move to cfgexpand.c.
2104 (expand_asm_stmt): Move to cfgexpand.c.
2105 (check_operand_nalternatives): Move to cfgexpand.c
2106 (expand_null_return): Move to cfgexpand.c.
2107 (expand_value_return): Move to cfgexpand.c.
2108 (expand_null_return_1): Move to cfgexpand.c.
2109 (expand_return): Move to cfgexpand.c.
2110 (expand_stack_save): Move to builtins.c.
2111 (expand_stack_restore): Move to builtins.c
2112 * symtab.c: Include output.h.
2113 (decl_assembler_name_hash): Move from tree.c.
2114 (decl_assembler_name_equal): Move from tree.c.
2115 * trans-mem.c (is_tm_safe_or_pure): Move from tree.h.
2116 * tree-eh.c (in_array_bounds_p): Move from tree.c.
2117 (range_in_array_bounds_p): Move from tree.c.
2118 * tree-object-size.c (fini_object_sizes): Make static.
2119 * tree-ssa-dom.c (iterative_hash_exprs_commutative): Move from tree.h.
2120 * tree-vrp.c (ssa_name_nonnegative_p): Remove.
2121 * tree.c (decl_assembler_name_equal): Move to symtab.c.
2122 (tree_expr_size): Move to explow.c.
2123 (decl_assembler_name_hash): Move to symtab.c.
2124 (real_twop): Remove.
2125 (tree_expr_size): Move to explow.c.
2126 (stabilize_reference_1): Move earlier in the file. Make static.
2127 (omp_remove_redundant_declare_simd_attrs): Remove.
2128 (simple_cst_list_equal): Move earlier in the file. Make static.
2129 (size_low_cst): Move to fold-const.c.
2130 (build_type_no_quals): Remove.
2131 (build_function_type_skip_args): Move to cgraphclones.c.
2132 (build_function_decl_skip_args): Move to cgraphclones.c.
2133 (in_array_bounds_p): Move to tree-eh.c.
2134 (range_in_array_bounds_p): Move to tree-eh.c.
2135 (truth_type_for): Move to gimple-fold.c.
2136 (list_equal_p): Remove.
2137 * tree.h (decl_assembler_name_equal): Remove.
2138 (decl_assembler_name_hash): Remove.
2139 (truth_type_for): Remove.
2140 (build_type_no_quals): Remove.
2141 (build_function_decl_skip_args): Remove.
2142 (in_array_bounds_p): Remove.
2143 (range_in_array_bounds_p): Remove.
2144 (size_low_cst): Remove.
2145 (omp_remove_redundant_declare_simd_attrs): Remove.
2146 (tree_expr_size): Remove.
2147 (fields_length): Remove.
2148 (stabilize_reference_1): Remove.
2149 (expand_goto): Remove.
2150 (expand_stack_save): Remove.
2151 (expand_stack_restore): Remove.
2152 (expand_return): Remove.
2153 (fold_build3_initializer_loc): Remove.
2154 (tree_expr_nonzero_p): Remove.
2155 (tree_invalid_nonnegative_warnv_p): Remove.
2156 (tree_expr_nonzero_warnv_p): Remove.
2157 (fold_builtin_snprintf_chk): Remove.
2158 (validate_arglist): Remove.
2159 (iterative_hash_exprs_commutative): Move to tree-ssa-dom.c.
2160 (simple_cst_list_equal): Remove.
2161 (real_twop): Remove.
2162 (expand_main_function): Remove.
2163 (stack_protect_prologue): Remove.
2164 (print_vec_tree): Remove.
2165 (lookup_scoped_attribute_spec): Remove.
2166 (get_attribute_namespace): Remove.
2167 (expand_computed_goto): Remove.
2168 (expand_asm_stmt): Remove.
2169 (list_equal_p): Remove.
2170 (ssa_name_nonnegative_p): Remove.
2171 (fini_object_sizes): Remove.
2172 (addr_expr_of_non_mem_decl_p): Remove.
2173 (is_tm_safe_or_pure): Move to trans-mem.c.
2174 (more_const_call_expr_args_p): Remove.
2175 (save_vtable_map_decl): Remove.
2177 2013-11-07 Thomas Schwinge <thomas@codesourcery.com>
2179 * doc/sourcebuild.texi (Top Level) <lto-plugin>: GNU ld can use
2180 linker plugins, too.
2182 * config/arc/arc.h (LINK_COMMAND_SPEC): For -ftree-parallelize-loops=*,
2183 link to libgomp and its dependencies.
2184 * config/ia64/hpux.h (LIB_SPEC): Likewise.
2185 * config/pa/pa-hpux11.h (LIB_SPEC): Likewise.
2186 * config/pa/pa64-hpux.h (LIB_SPEC): Likewise.
2187 * gcc.c (GOMP_SELF_SPECS): Update comment about libgomp's dependencies.
2189 2013-11-07 Jakub Jelinek <jakub@redhat.com>
2191 * tree-ssa-loop-niter.c: Include tree-ssanames.h.
2192 (determine_value_range): Add loop argument. Use get_range_info to
2194 (bound_difference): Adjust caller.
2196 2013-11-07 Richard Biener <rguenther@suse.de>
2197 Jakub Jelinek <jakub@redhat.com>
2199 * tree-vrp.c (find_assert_locations): Pre-seed live bitmaps for loop
2200 latches from header PHI arguments from the latch edge.
2202 2013-11-07 Paolo Carlini <paolo.carlini@oracle.com>
2205 * varasm.c (output_constant): Handle NULLPTR_TYPE.
2207 2013-11-07 H.J. Lu <hongjiu.lu@intel.com>
2209 * config/i386/i386.c (ix86_expand_set_or_movmem): Don't set
2210 misaligned_prologue_used when it has been set.
2212 2013-11-07 Yury Gribov <y.gribov@samsung.com>
2213 Jakub Jelinek <jakub@redhat.com>
2216 * asan.c (get_mem_refs_of_builtin_call): Allow
2217 integer literals as addresses in instrumented builtins.
2219 2013-11-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2221 * config/aarch64/aarch64.c (aarch64_legitimize_reload_address):
2222 Explain why plus_constant is not used.
2224 2013-11-07 Richard Biener <rguenther@suse.de>
2226 * tree-ssa-ccp.c (canonicalize_float_value): Rename to ...
2227 (canonicalize_value): ... this. Also handle stripping of
2229 (get_value, set_lattice_value, get_value_for_expr): Adjust.
2230 * gimple-fold.c (canonicalize_constructor_val): Strip TREE_OVERFLOW.
2231 * tree-ssa-threadedge.c (set_ssa_name_value): Likewise.
2233 2013-11-07 Richard Biener <rguenther@suse.de>
2235 * tree-dfa.c (get_ref_base_and_extent): Fix casting.
2237 2013-11-07 H.J. Lu <hongjiu.lu@intel.com>
2240 * config/i386/i386.md (push peepholer/splitter): Use Pmode
2241 with stack_pointer_rtx.
2243 2013-11-07 Bin Cheng <bin.cheng@arm.com>
2245 * tree-ssa-loop-ivopts.c (get_shiftadd_cost): Check equality
2246 using operand_equal_p.
2248 2013-11-07 Bin Cheng <bin.cheng@arm.com>
2250 * tree-ssa-loop-ivopts.c (alloc_iv): Lower address expressions.
2251 * tree-affine.c (get_inner_reference_aff): Return base.
2252 * tree-affine.h (get_inner_reference_aff): Change prototype.
2254 2013-11-06 Tobias Burnus <burnus@net-b.de>
2256 * doc/invoke.texi (Wdate-time): Fix typo.
2258 2013-11-06 Oleg Endo <olegendo@gcc.gnu.org>
2260 * config/sh/sh.md (addsf3, divsf3, divsf3_i, rsqrtsf2, cmpgtdf_t,
2261 cmpeqdf_t, *ieee_ccmpeqdf_t, negdf2, sqrtdf2, absdf2): Use
2262 fp_arith_reg_operand instead of arith_reg_operand.
2264 2013-11-06 Oleg Endo <olegendo@gcc.gnu.org>
2266 * config/sh/sh.md (adddi3): Remove empty constraints.
2267 Remove can_create_pseudo_p and arith_reg_operand check.
2268 (adddi3_compact, subdi3_compact, *negdi2): Remove constraints.
2269 Split before reload.
2271 2013-11-06 Jeff Law <law@redhat.com>
2272 Tom Tromey <tromey@redhat.com>
2274 * gdbinit.in: Disable strict type checking.
2276 2013-11-06 Vladimir Makarov <vmakarov@redhat.com>
2278 * tree-pass.h (make_pass_live_range_shrinkage): New external.
2279 * timevar.def (TV_LIVE_RANGE_SHRINKAGE): New.
2280 * sched-rgn.c (gate_handle_live_range_shrinkage): New.
2281 (rest_of_handle_live_range_shrinkage): Ditto
2282 (class pass_live_range_shrinkage): Ditto.
2283 (pass_data_live_range_shrinkage): Ditto.
2284 (make_pass_live_range_shrinkage): Ditto.
2285 * sched-int.h (initialize_live_range_shrinkage): New prototype.
2286 (finish_live_range_shrinkage): Ditto.
2287 * sched-deps.c (create_insn_reg_set): Make void return value.
2288 * passes.def: Add pass_live_range_shrinkage.
2289 * ira.c (update_equiv_regs): Don't move if flag_live_range_shrinkage.
2290 * haifa-sched.c (live_range_shrinkage_p): New.
2291 (initialize_live_range_shrinkage, finish_live_range_shrinkage):
2293 (rank_for_schedule): Add code for pressure relief through live
2295 (schedule_insn): Print more debug info.
2296 (sched_init): Setup SCHED_PRESSURE_WEIGHTED for pressure relief
2297 through live range shrinkage.
2298 * doc/invoke.texi (-flive-range-shrinkage): New.
2299 * common.opt (flive-range-shrinkage): New.
2301 2013-11-06 Uros Bizjak <ubizjak@gmail.com>
2304 * config/i386/i386.c (ix86_avx_u128_mode_needed): Require
2305 AVX_U128_DIRTY mode for call_insn RTXes that use AVX256 registers.
2306 (ix86_avx_u128_mode_needed): Return AVX_U128_DIRTY mode for call_insn
2307 RTXes that return in AVX256 register.
2309 2013-11-06 Richard Biener <rguenther@suse.de>
2311 PR tree-optimization/58653
2312 * tree-predcom.c (ref_at_iteration): Rewrite to generate a MEM_REF.
2313 (prepare_initializers_chain): Adjust.
2315 2013-11-06 Andrew MacLeod <amacleod@redhat.com>
2317 * gimple.h (block_in_transaction): Move to basic-block.h and rename.
2318 (gimple_in_transaction): Use bb_in_transaction.
2319 * basic-block.h (bb_in_transaction): Relocate here and rename.
2320 * tree-ssa-loop-im.c (execute_sm): Use bb_in_transaction.
2322 2013-11-06 Richard Biener <rguenther@suse.de>
2324 * tree.c (drop_tree_overflow): New function.
2325 * tree.h (drop_tree_overflow): Declare.
2326 * gimplify.c (gimplify_expr): Drop TREE_OVERFLOW.
2327 * tree-vrp.c (range_int_cst_singleton_p): Use
2328 is_overflow_infinity instead of testing TREE_OVERFLOW.
2329 (extract_range_from_assert): Likewise.
2330 (zero_nonzero_bits_from_vr): Likewise.
2331 (extract_range_basic): Likewise.
2332 (register_new_assert_for): Use drop_tree_overflow.
2333 (vrp_visit_phi_node): Likewise.
2335 2013-11-06 Eric Botcazou <ebotcazou@adacore.com>
2337 * config/i386/i386.c (ix86_expand_prologue): Optimize stack
2338 checking for leaf functions without dynamic stack allocation.
2339 * config/ia64/ia64.c (ia64_emit_probe_stack_range): Adjust.
2340 (ia64_expand_prologue): Likewise.
2341 * config/mips/mips.c (mips_expand_prologue): Likewise.
2342 * config/rs6000/rs6000.c (rs6000_emit_prologue): Likewise.
2343 * config/sparc/sparc.c (sparc_expand_prologue): Likewise.
2344 (sparc_flat_expand_prologue): Likewise.
2346 2013-11-06 James Greenhalgh <james.greenhalgh@arm.com>
2348 * config/aarch64/arm_neon.h (__ST2_LANE_FUNC): Better model data size.
2349 (__ST3_LANE_FUNC): Likewise.
2350 (__ST4_LANE_FUNC): Likewise.
2352 2013-11-06 Nick Clifton <nickc@redhat.com>
2354 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define the
2355 name returned by msp430_mcu_name.
2356 (LIB_SPEC): If a -T option has not been specified then set a
2357 default, mcu-specific, linker script.
2358 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more mcu names.
2359 * config/msp430/msp430.c (msp430x_names): Likewise.
2360 Alpha sort the names for ease of comparison.
2361 (msp430_mcu_name): New function: Returns a string suitable for
2362 use as a C preprocessor symbol based upon the name of the MCU
2364 (msp430_option_override): Accept msp430x and msp430xv2 as generic
2366 * config/msp430/msp430-protos.h (msp430_mcu_name): Prototype.
2368 * gcc.c (do_spec_1): Do not insert a space after a %* substitution
2369 unless it is the last part of a spec substring.
2370 * doc/invoke.texi (Spec Files): Document space insertion
2373 2013-11-06 Christian Bruel <christian.bruel@st.com>
2375 * config/sh/sh-mem.cc (sh_expand_cmpnstr, sh_expand_cmpstr):
2376 Factorize probabilities, Use adjust_address instead of
2377 adjust_automodify_address when possible. Enable for optimize.
2378 (sh_expand_strlen): New function.
2379 * config/sh/sh-protos.h (sh_expand_strlen): Declare.
2380 * config/sh/sh.md (strlensi): New pattern.
2381 (UNSPEC_BUILTIN_STRLEN): Define.
2383 2013-11-06 Jakub Jelinek <jakub@redhat.com>
2386 * expr.c (get_bit_range): Handle *offset == NULL_TREE.
2387 (expand_assignment): If *bitpos is negative, set *offset
2388 and adjust *bitpos, so that it is not negative.
2390 2013-11-06 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
2392 * config/i386/bdver3.md : Added two additional decoder units
2393 to support issue rate of 4 and remodeled vector unit.
2394 * config/i386/i386.c (ix86_issue_rate): Issue rate for BD
2395 architectures is set to 4.
2396 * config/i386/i386.c (ia32_multipass_dfa_lookahead): DFA
2397 lookahead is set to 4 for BD architectures.
2399 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2401 * config/rs6000/rs6000.c (rs6000_option_override_internal):
2402 Remove restriction against use of VSX instructions when generating
2403 code for little endian mode.
2405 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2407 * config/rs6000/altivec.md (mulv4si3): Ensure we generate vmulouh
2408 for both big and little endian.
2409 (mulv8hi3): Swap input operands for merge high and merge low
2410 instructions for little endian.
2412 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2414 * config/rs6000/altivec.md (vec_widen_umult_even_v16qi): Change
2415 define_insn to define_expand that uses even patterns for big
2416 endian and odd patterns for little endian.
2417 (vec_widen_smult_even_v16qi): Likewise.
2418 (vec_widen_umult_even_v8hi): Likewise.
2419 (vec_widen_smult_even_v8hi): Likewise.
2420 (vec_widen_umult_odd_v16qi): Likewise.
2421 (vec_widen_smult_odd_v16qi): Likewise.
2422 (vec_widen_umult_odd_v8hi): Likewise.
2423 (vec_widen_smult_odd_v8hi): Likewise.
2424 (altivec_vmuleub): New define_insn.
2425 (altivec_vmuloub): Likewise.
2426 (altivec_vmulesb): Likewise.
2427 (altivec_vmulosb): Likewise.
2428 (altivec_vmuleuh): Likewise.
2429 (altivec_vmulouh): Likewise.
2430 (altivec_vmulesh): Likewise.
2431 (altivec_vmulosh): Likewise.
2433 2013-11-05 Mike Stump <mikestump@comcast.net>
2435 * Makefile.in (mostlyclean): Remove c-family objects.
2437 2013-11-05 Ian Lance Taylor <iant@google.com>
2439 * config/i386/sync.md (atomic_compare_and_swap<dwi>_doubleword):
2440 If possible, add .cfi directives to record change to bx.
2441 * config/i386/i386.c (ix86_emit_cfi): New function.
2442 * config/i386/i386-protos.h (ix86_emit_cfi): Declare.
2444 2013-11-05 Steven Bosscher <steven@gcc.gnu.org>
2446 * rtlanal.c (tablejump_p): Expect a JUMP_TABLE_DATA to always follow
2447 immediately after a label for a tablejump pattern.
2449 * config/arm/arm.c (is_jump_table): Remove.
2450 (create_fix_barrier): Use tablejump_p instead.
2451 (arm_reorg): Likewise.
2452 (thumb1_output_casesi): Expect JUMP_TABLE_DATA to always be NEXT_INSN.
2453 (thumb2_output_casesi): Likewise.
2454 * config/aarch64/aarch64.c (aarch64_output_casesi): Likewise.
2455 * config/sh/sh.md (casesi_worker_1, casesi_worker_2,
2456 casesi_shift_media, casesi_load_media): Likewise.
2457 * config/iq2000/iq2000.md: Likewise (in anonymous define_insn).
2458 * config/microblaze/microblaze.md: Likewise.
2460 2013-11-05 Tobias Burnus <burnus@net-b.de>
2462 * doc/invoke.texi (-Wdate-time): Document.
2464 2013-11-05 Richard Sandiford <rdsandiford@googlemail.com>
2466 * double-int.c (lshift_double, rshift_double): Remove
2467 SHIFT_COUNT_TRUNCATED handling.
2469 2013-11-05 Jeff Law <law@redhat.com>
2471 * Makefile.in (OBJS): Add gimple-ssa-isolate-paths.o
2472 * common.opt (-fisolate-erroneous-paths): Add option and documentation.
2473 * gimple-ssa-isolate-paths.c: New file.
2474 * gimple.c (check_loadstore): New function.
2475 (infer_nonnull_range): Moved into gimple.c from tree-vrp.c
2476 Verify OP is in the argument list and the argument corresponding
2477 to OP is a pointer type. Use operand_equal_p rather than
2478 pointer equality when testing if OP is on the nonnull list.
2479 Use check_loadstore rather than count_ptr_derefs. Handle
2480 GIMPLE_RETURN statements.
2481 * tree-vrp.c (infer_nonnull_range): Remove.
2482 * gimple.h (infer_nonnull_range): Declare.
2483 * opts.c (default_options_table): Add OPT_fisolate_erroneous_paths.
2484 * passes.def: Add pass_isolate_erroneous_paths.
2485 * timevar.def (TV_ISOLATE_ERRONEOUS_PATHS): New timevar.
2486 * tree-pass.h (make_pass_isolate_erroneous_paths): Declare.
2487 * tree-ssa.c (struct count_ptr_d): Remove.
2488 (count_ptr_derefs, count_uses_and_derefs): Remove.
2489 * tree-ssa.h (count_uses_and_derefs): Remove.
2491 2013-11-05 Jakub Jelinek <jakub@redhat.com>
2493 PR rtl-optimization/58997
2494 * loop-iv.c (iv_subreg): For IV_UNKNOWN_EXTEND, expect
2495 get_iv_value to be in iv->mode rather than iv->extend_mode.
2496 (iv_extend): Likewise. Otherwise, if iv->extend != extend,
2497 use lowpart_subreg on get_iv_value before calling simplify_gen_unary.
2498 * loop-unswitch.c (may_unswitch_on): Make sure op[i] is in the right
2501 2013-11-05 Andrew MacLeod <amacleod@redhat.com>
2503 * gimple.h: Move some prototypes to gimple-expr.h and add to include
2505 (extract_ops_from_tree, gimple_call_addr_fndecl, is_gimple_reg_type):
2506 Move to gimple-expr.h.
2507 * gimple-expr.h: New file. Relocate some prototypes from gimple.h.
2508 (types_compatible_p, is_gimple_reg_type, is_gimple_variable,
2509 is_gimple_id, virtual_operand_p, is_gimple_addressable,
2510 is_gimple_constant, extract_ops_from_tree, gimple_call_addr_fndecl):
2512 * gimple.c (extract_ops_from_tree_1, gimple_cond_get_ops_from_tree,
2513 gimple_set_body, gimple_body, gimple_has_body_p, is_gimple_lvalue,
2514 is_gimple_condexpr, is_gimple_addressable, is_gimple_constant,
2515 is_gimple_address, is_gimple_invariant_address,
2516 is_gimple_ip_invariant_address, is_gimple_min_invariant,
2517 is_gimple_ip_invariant, is_gimple_variable, is_gimple_id,
2518 virtual_operand_p, is_gimple_reg, is_gimple_val, is_gimple_asm_val,
2519 is_gimple_min_lval, is_gimple_call_addr, is_gimple_mem_ref_addr,
2520 gimple_decl_printable_name, useless_type_conversion_p,
2521 types_compatible_p, gimple_can_coalesce_p, copy_var_decl): Move to
2523 * gimple-expr.c: New File.
2524 (useless_type_conversion_p, gimple_set_body, gimple_body,
2525 gimple_has_body_p, gimple_decl_printable_name, copy_var_decl,
2526 gimple_can_coalesce_p, extract_ops_from_tree_1,
2527 gimple_cond_get_ops_from_tree, is_gimple_lvalue, is_gimple_condexpr,
2528 is_gimple_address, is_gimple_invariant_address,
2529 is_gimple_ip_invariant_address, is_gimple_min_invariant,
2530 is_gimple_ip_invariant, is_gimple_reg, is_gimple_val,
2531 is_gimple_asm_val, is_gimple_min_lval, is_gimple_call_addr,
2532 is_gimple_mem_ref_addr): Relocate here.
2533 * Makefile.in (OBJS): Add gimple-expr.o.
2535 2013-11-05 David Malcolm <dmalcolm@redhat.com>
2537 * gengtype-parse.c (struct_field_seq): Support empty structs.
2539 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
2541 * config/i386/t-rtems (MULTILIB_MATCHES): Fix option typos.
2543 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
2545 * config/i386/i386-c.c (ix86_target_macros): Define _SOFT_FLOAT
2547 * config/i386/rtemself.h (TARGET_OS_CPP_BUILTINS): Do not define
2549 (LONG_DOUBLE_TYPE_SIZE): New define.
2550 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Ditto.
2552 2013-11-05 Paolo Carlini <paolo.carlini@oracle.com>
2555 * doc/extend.texi [visibility ("visibility_type")]: Add example
2556 about visibility attribute on namespace declaration.
2558 2013-11-05 Richard Biener <rguenther@suse.de>
2561 * passes.def (all_passes): Start with pass_fixup_cfg again.
2563 2013-11-05 Richard Biener <rguenther@suse.de>
2565 PR tree-optimization/58955
2566 * tree-loop-distribution.c (pg_add_dependence_edges): Fix
2569 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2571 * config/rs6000/vector.md (vec_pack_sfix_trunc_v2df): Adjust for
2573 (vec_pack_ufix_trunc_v2df): Likewise.
2575 2013-11-05 H.J. Lu <hongjiu.lu@intel.com>
2578 * doc/md.texi (@code{movmem@var{m}}): Specify Pmode as mode of
2579 pattern, instead of word_mode.
2581 * expr.c (emit_block_move_via_movmem): Don't use mode wider than
2583 (set_storage_via_setmem): Likewise.
2585 2013-11-05 Andrew MacLeod <amacleod@redhat.com>
2587 * tree-outof-ssa.c (queue_phi_copy_p): Combine phi_ssa_name_p from
2588 gimple.h and the rest of the condition in eliminate_build.
2589 (eliminate_build): Call new routine.
2590 * gimple.h (phi_ssa_name_p): Delete.
2592 2013-11-05 Trevor Saunders <tsaunders@mozilla.com>
2594 * vec.c (vec_prefix::calculate_allocation): Don't try to handle the
2595 case of no prefix and reserving zero slots, because when that's the
2596 case we'll never get here.
2597 * vec.h (va_heap::reserve): Don't try and handle
2598 vec_prefix::calculate_allocation returning zero because that should
2601 2013-11-05 Richard Biener <rguenther@suse.de>
2604 * tree-dfa.c (get_ref_base_and_extent): Merge common code
2605 in MEM_REF and TARGET_MEM_REF handling. Make sure to
2606 process trailing array detection before diving into the
2607 view-converted object (and possibly apply some extra offset).
2609 2013-11-05 Joseph Myers <joseph@codesourcery.com>
2611 * config/i386/i386.c (ix86_float_exceptions_rounding_supported_p):
2613 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): Define.
2615 2013-11-05 Marc Glisse <marc.glisse@inria.fr>
2617 PR tree-optimization/58958
2618 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Use
2619 get_addr_base_and_unit_offset instead of get_ref_base_and_extent.
2621 2013-11-05 Marc Glisse <marc.glisse@inria.fr>
2623 * tree-ssa-alias.h (ranges_overlap_p): Handle negative offsets.
2624 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Likewise.
2626 2013-11-05 Jakub Jelinek <jakub@redhat.com>
2628 PR tree-optimization/58984
2629 * ipa-prop.c (ipa_load_from_parm_agg_1): Add SIZE_P argument,
2630 set *SIZE_P if non-NULL on success.
2631 (ipa_load_from_parm_agg, ipa_analyze_indirect_call_uses): Adjust
2633 (ipcp_transform_function): Likewise. Punt if size of access
2634 is different from TYPE_SIZE on v->value's type.
2636 2013-11-05 Tobias Burnus <burnus@net-b.de>
2638 * doc/invoke.texi (-fopenmp-simd): Document new option.
2639 * gimplify.c (gimplify_body): Accept -fopenmp-simd.
2640 * omp-low.c (execute_expand_omp, execute_lower_omp): Ditto.
2641 * tree.c (attribute_value_equal): Ditto.
2643 2013-11-04 Wei Mi <wmi@google.com>
2645 * sched-rgn.c (add_branch_dependences): Keep insns in
2646 a SCHED_GROUP at the end of BB to remain their location.
2648 2013-11-04 Wei Mi <wmi@google.com>
2650 * config/i386/i386.c (memory_address_length): Extract a part
2651 of code to rip_relative_addr_p.
2652 (rip_relative_addr_p): New Function.
2653 (ix86_macro_fusion_p): Ditto.
2654 (ix86_macro_fusion_pair_p): Ditto.
2655 * config/i386/i386.h: Add new tune features about macro-fusion.
2656 * config/i386/x86-tune.def (DEF_TUNE): Ditto.
2657 * doc/tm.texi: Generated.
2658 * doc/tm.texi.in: Ditto.
2659 * haifa-sched.c (try_group_insn): New Function.
2660 (group_insns_for_macro_fusion): Ditto.
2661 (sched_init): Call group_insns_for_macro_fusion.
2662 * target.def: Add two hooks: macro_fusion_p and
2663 macro_fusion_pair_p.
2665 2013-11-04 Kostya Serebryany <kcc@google.com>
2667 Update to match the changed asan API.
2668 * asan.c (asan_function_start): New function.
2669 (asan_emit_stack_protection): Update the string stored in the
2670 stack red zone to match new API. Store the PC of the current
2671 function in the red zone.
2672 (asan_global_struct): Update the __asan_global definition to match
2674 (asan_add_global): Ditto.
2675 * asan.h (asan_function_start): New prototype.
2676 * final.c (final_start_function): Call asan_function_start.
2677 * sanitizer.def (BUILT_IN_ASAN_INIT): Rename __asan_init_v1
2680 2013-11-04 Wei Mi <wmi@google.com>
2682 * config/i386/i386-c.c (ix86_target_macros_internal): Separate
2683 PROCESSOR_COREI7_AVX out from PROCESSOR_COREI7.
2684 * config/i386/i386.c (ix86_option_override_internal): Ditto.
2685 (ix86_issue_rate): Ditto.
2686 (ix86_adjust_cost): Ditto.
2687 (ia32_multipass_dfa_lookahead): Ditto.
2688 (ix86_sched_init_global): Ditto.
2689 (get_builtin_code_for_version): Ditto.
2690 * config/i386/i386.h (enum target_cpu_default): Ditto.
2691 (enum processor_type): Ditto.
2692 * config/i386/x86-tune.def (DEF_TUNE): Ditto.
2694 2013-11-04 Vladimir Makarov <vmakarov@redhat.com>
2696 PR rtl-optimization/58967
2697 * config/rs6000/rs6000.c (legitimate_lo_sum_address_p): Remove
2698 !lra_in_progress for mode sizes bigger word.
2700 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2702 * config/rs6000/altivec.md (vec_widen_umult_hi_v16qi): Swap
2703 arguments to merge instruction for little endian.
2704 (vec_widen_umult_lo_v16qi): Likewise.
2705 (vec_widen_smult_hi_v16qi): Likewise.
2706 (vec_widen_smult_lo_v16qi): Likewise.
2707 (vec_widen_umult_hi_v8hi): Likewise.
2708 (vec_widen_umult_lo_v8hi): Likewise.
2709 (vec_widen_smult_hi_v8hi): Likewise.
2710 (vec_widen_smult_lo_v8hi): Likewise.
2712 2013-11-04 Ian Lance Taylor <iant@google.com>
2714 * builtins.def (ATTR_NOTHROWCALL_LEAF_LIST): Define.
2715 * sync-builtins.def: Use ATTR_NOTHROWCALL_LEAF_LIST for all sync
2716 builtins that take pointers.
2717 * lto-opts.c (lto_write_options): Write -fnon-call-exceptions if set.
2718 * lto-wrapper.c (merge_and_complain): Collect OPT_fnon_call_exceptions.
2719 (run_gcc): Pass -fnon-call-exceptions.
2721 2013-11-04 Jakub Jelinek <jakub@redhat.com>
2723 * optabs.c (expand_vec_perm): Revert one incorrect line from
2726 PR tree-optimization/58978
2727 * tree-vrp.c (all_imm_uses_in_stmt_or_feed_cond): Don't modify
2728 use_stmt by single_imm_use directly. Only call single_imm_use
2731 2013-11-04 Vladimir Makarov <vmakarov@redhat.com>
2733 PR rtl-optimization/58968
2734 * lra-spills.c (return_regno_p): New function.
2735 (lra_final_code_change): Use it.
2737 2013-11-04 Joseph Myers <joseph@codesourcery.com>
2739 * doc/cpp.texi (__GCC_IEC_559, __GCC_IEC_559_COMPLEX): Document macros.
2740 * target.def (float_exceptions_rounding_supported_p): New hook.
2741 * targhooks.c (default_float_exceptions_rounding_supported_p): New
2743 * targhooks.h (default_float_exceptions_rounding_supported_p): Declare.
2744 * doc/tm.texi.in (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P):
2746 * doc/tm.texi: Regenerate.
2747 * config.gcc (powerpc*-*-linux*): Set extra_objs.
2748 * config/rs6000/rs6000-linux.c: New file.
2749 * config/rs6000/rs6000-protos.h
2750 (rs6000_linux_float_exceptions_rounding_supported_p): Declare.
2751 * config/rs6000/linux.h
2752 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): New macro.
2753 * config/rs6000/linux64.h
2754 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): Likewise.
2755 * config/rs6000/t-linux (rs6000-linux.o): New rule.
2756 * config/rs6000/t-linux64 (rs6000-linux.o): Likewise.
2758 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2760 * config/rs6000/vsx.md (*vsx_le_perm_store_<mode> for VSX_D):
2761 Replace the define_insn_and_split with a define_insn and two
2762 define_splits, with the split after reload re-permuting the source
2763 register to its original value.
2764 (*vsx_le_perm_store_<mode> for VSX_W): Likewise.
2765 (*vsx_le_perm_store_v8hi): Likewise.
2766 (*vsx_le_perm_store_v16qi): Likewise.
2768 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2770 * config/rs6000/vector.md (vec_pack_trunc_v2df): Adjust for
2773 2013-11-04 Jakub Jelinek <jakub@redhat.com>
2775 PR tree-optimization/58946
2776 * tree-ssa-reassoc.c (maybe_optimize_range_tests): Update all
2777 bbs with bbinfo[idx].op != NULL before all blocks with
2778 bbinfo[idx].op == NULL.
2780 2013-11-04 Richard Sandiford <rdsandiford@googlemail.com>
2782 * config/avr/avr-log.c (avr_double_int_pop_digit): Delete.
2783 (avr_dump_double_int_hex): Likewise.
2784 (avr_log_vadump): Remove %D and %X handling.
2785 * config/avr/avr.c (avr_double_int_push_digit): Delete.
2786 (avr_map_op_t): Change map from double_int to unsigned int.
2787 (avr_map_op): Update accordingly.
2788 (avr_map, avr_map_metric, avr_has_nibble_0xf, avr_map_decompose)
2789 (avr_move_bits, avr_out_insert_bits, avr_fold_builtin): Operate on
2790 unsigned ints rather than double_ints.
2792 2013-11-03 Marek Polacek <polacek@redhat.com>
2794 Implement -fsanitize=vla-bound.
2795 * opts.c (common_handle_option): Handle vla-bound.
2796 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE): Define.
2797 * flag-types.h (enum sanitize_code): Add SANITIZE_VLA.
2798 * asan.c (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR.
2800 2013-11-02 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
2802 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Adjust for
2805 2013-11-02 Uros Bizjak <ubizjak@gmail.com>
2807 * config/i386/constraints.md (Ts, Tv): New address constrains.
2808 * config/i386/i386.md (*lea<mode>, *<mode>_<bndcheck>): Use Ts
2809 constraint for address_no_seg_operand.
2810 * config/i386/sse.md (*avx512pf_gatherpf<mode>_mask)
2811 (*avx512pf_gatherpf<mode>, *avx512pf_scatterpf<mode>_mask)
2812 (*avx512pf_scatterpf<mode>, *avx2_gathersi<mode>)
2813 (*avx2_gathersi<mode>_2, *avx2_gatherdi<mode>, *avx2_gatherdi<mode>_2)
2814 (*avx2_gatherdi<mode>_3, *avx2_gatherdi<mode>_4)
2815 (*avx512f_gathersi<mode>, *avx512f_gathersi<mode>_2)
2816 (*avx512f_gatherdi<mode>, *avx512f_gatherdi<mode>_2)
2817 (*avx512f_scattersi<mode> *avx512f_scatterdi<mode>): Use Tv
2818 constraint for vsib_address_operand.
2820 2013-11-02 Steven Bosscher <steven@gcc.gnu.org>
2822 * gcse.c (pre_delete): Remove references to regmove from comments.
2823 * recog.c: (validate_replace_rtx_1): Likewise.
2824 * config/rl78/rl78.c: Likewise.
2825 * config/v850/v850.h: Likewise, and remove unused ENABLE_REGMOVE_PASS.
2826 * common/config/m32r/m32r-common.c: Don't manipulate OPT_fregmove.
2827 * common/config/mmix/mmix-common.c: Likewise.
2829 2013-11-01 Trevor Saunders <tsaunders@mozilla.com>
2831 * function.c (reorder_blocks): Convert block_stack to a stack_vec.
2832 * gimplify.c (gimplify_compound_lval): Likewise.
2833 * graphite-clast-to-gimple.c (gloog): Likewise.
2834 * graphite-dependences.c (loop_is_parallel_p): Likewise.
2835 * graphite-scop-detection.c (scopdet_basic_block_info): Likewise.
2836 (limit_scop); Likewise.
2837 (build_scops): Likewise.
2838 (dot_scop): Likewise.
2839 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
2840 (build_scop_drs): Likewise.
2841 (insert_stmts): Likewise.
2842 (insert_out_of_ssa_copy): Likewise.
2843 (remove_phi): Likewise.
2844 (rewrite_commutative_reductions_out_of_ssa_close_phi): Likewise.
2845 * hw-doloop.c (discover_loop): Likewise.
2846 * tree-call-cdce.c (shrink_wrap_one_built_in_call): Likewise.
2847 * tree-dfa.c (dump_enumerated_decls): Likewise.
2848 * tree-if-conv.c (if_convertable_loop_p): Likewise.
2849 * tree-inline.c (tree_function_versioning): Likewise.
2850 * tree-loop-distribution.c (build_rdg): Likewise.
2851 (rdg_flag_vertex_and_dependent): Likewise.
2852 (distribute_loop): Likewise.
2853 * tree-parloops.c (loop_parallel_p): Likewise.
2854 (eliminate_local_variables): Likewise.
2855 (separate_decls_in_region): Likewise.
2856 * tree-predcom.c (tree_predictive_commoning_loop): Likewise.
2857 * tree-ssa-phiopt.c (cond_if_else_store_replacement): Likewise.
2858 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
2859 * tree-vect-loop.c (vect_analyze_scaler_cycles_1): Likewise.
2860 * tree-vect-patterns.c (vect_pattern_recog): Likewise.
2861 * tree-vect-stmts.c (vect_mark_stmts_to_be_vectorized): Likewise.
2862 (vectorizable_condition): Likewise.
2864 2013-11-01 Uros Bizjak <ubizjak@gmail.com>
2866 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): Always define as 0/1.
2867 * configure: Regenerate.
2868 * config/i386/i386.md (*movdi_internal): Change
2869 HAVE_AS_IX86_INTERUNIT_MOVQ to runtime check.
2870 (*movdf_internal): Ditto.
2871 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
2872 * config/i386/sse.md (vec_concatv2di): Output interunit movq
2873 for HAVE_AS_IX86_INTERUNIT_MOVQ targets.
2875 2013-10-31 Robert Suchanek <Robert.Suchanek@imgtec.com>
2877 * lra-spills.c (assign_spill_hard_regs): Remove statement terminator
2880 2013-10-31 David Malcolm <dmalcolm@redhat.com>
2882 Automated part of renaming of symtab_node_base to symtab_node.
2884 Patch autogenerated by rename_symtab.py from
2885 https://github.com/davidmalcolm/gcc-refactoring-scripts
2886 revision 58bb219cc090b2f4516a9297d868c245495ee622
2887 with ChangeLog entry fixed up by hand.
2889 * cgraph.c (x_cgraph_nodes_queue): Rename symtab_node_base to
2891 (cgraph_node_for_asm): Likewise.
2892 * cgraph.h (symtab_node_base): Likewise.
2893 (cgraph_node): Likewise.
2894 (varpool_node): Likewise.
2895 (is_a_helper <cgraph_node>::test): Likewise.
2896 (is_a_helper <varpool_node>::test): Likewise.
2897 (symtab_nodes): Likewise.
2898 (symtab_register_node): Likewise.
2899 (symtab_unregister_node): Likewise.
2900 (symtab_remove_node): Likewise.
2901 (symtab_get_node): Likewise.
2902 (symtab_node_for_asm): Likewise.
2903 (symtab_node_asm_name): Likewise.
2904 (symtab_node_name): Likewise.
2905 (symtab_insert_node_to_hashtable): Likewise.
2906 (symtab_add_to_same_comdat_group): Likewise.
2907 (symtab_dissolve_same_comdat_group_list): Likewise.
2908 (dump_symtab_node): Likewise.
2909 (debug_symtab_node): Likewise.
2910 (dump_symtab_base): Likewise.
2911 (verify_symtab_node): Likewise.
2912 (verify_symtab_base): Likewise.
2913 (symtab_used_from_object_file_p): Likewise.
2914 (symtab_alias_ultimate_target): Likewise.
2915 (symtab_resolve_alias): Likewise.
2916 (fixup_same_cpp_alias_visibility): Likewise.
2917 (symtab_for_node_and_aliases): Likewise.
2918 (symtab_nonoverwritable_alias): Likewise.
2919 (availability symtab_node_availability): Likewise.
2920 (symtab_semantically_equivalent_p): Likewise.
2921 (fixup_same_cpp_alias_visibility): Likewise.
2922 (symtab_prevail_in_asm_name_hash): Likewise.
2924 (varpool): Likewise.
2925 (varpool_first_variable): Likewise.
2926 (varpool_next_variable): Likewise.
2927 (varpool_first_static_initializer): Likewise.
2928 (varpool_next_static_initializer): Likewise.
2929 (varpool_first_defined_variable): Likewise.
2930 (varpool_next_defined_variable): Likewise.
2931 (cgraph_first_defined_function): Likewise.
2932 (cgraph_next_defined_function): Likewise.
2933 (cgraph_first_function): Likewise.
2934 (cgraph_next_function): Likewise.
2935 (cgraph_first_function_with_gimple_body): Likewise.
2936 (cgraph_next_function_with_gimple_body): Likewise.
2937 (symtab_alias_target): Likewise.
2938 (symtab_real_symbol_p): Likewise.
2939 (symtab_can_be_discarded): Likewise.
2940 * cgraphbuild.c (mark_address): Likewise.
2941 (mark_load): Likewise.
2942 (mark_store): Likewise.
2943 * cgraphunit.c (decide_is_symbol_needed): Likewise.
2945 (enqueue_node): Likewise.
2946 (referred_to_p): Likewise.
2947 (cgraph_process_same_body_aliases): Likewise.
2948 (analyze_functions): Likewise.
2949 (handle_alias_pairs): Likewise.
2950 (output_weakrefs): Likewise.
2951 (compile): Likewise.
2952 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
2953 * ipa-inline-analysis.c (inline_write_summary): Likewise.
2954 * ipa-prop.c (remove_described_reference): Likewise.
2955 (try_decrement_rdesc_refcount): Likewise.
2956 (ipa_edge_duplication_hook): Likewise.
2957 * ipa-ref.c (ipa_record_reference): Likewise.
2958 (ipa_maybe_record_reference): Likewise.
2959 (ipa_clone_ref): Likewise.
2960 (ipa_clone_references): Likewise.
2961 (ipa_clone_referring): Likewise.
2962 (ipa_find_reference): Likewise.
2963 (ipa_remove_stmt_references): Likewise.
2964 (ipa_clear_stmts_in_references): Likewise.
2965 * ipa-ref.h (symtab_node_base): Likewise.
2966 (ipa_ref): Likewise.
2967 (ipa_record_reference): Likewise.
2968 (ipa_maybe_record_reference): Likewise.
2969 (ipa_clone_references): Likewise.
2970 (ipa_clone_referring): Likewise.
2971 (ipa_clone_ref): Likewise.
2972 (ipa_find_reference): Likewise.
2973 (ipa_remove_stmt_references): Likewise.
2974 (ipa_clear_stmts_in_references): Likewise.
2975 * ipa-reference.c (ipa_reference_write_optimization_summary):
2977 * ipa.c (enqueue_node): Likewise.
2978 (process_references): Likewise.
2979 (walk_polymorphic_call_targets): Likewise.
2980 (symtab_remove_unreachable_nodes): Likewise.
2981 (address_taken_from_non_vtable_p): Likewise.
2982 (comdat_can_be_unshared_p_1): Likewise.
2983 (comdat_can_be_unshared_p): Likewise.
2984 (can_replace_by_local_alias): Likewise.
2985 (function_and_variable_visibility): Likewise.
2986 * is-a.h: Likewise (within example in comment).
2987 * lto-cgraph.c (input_cgraph_opt_summary): Likewise.
2988 (lto_symtab_encoder_encode): Likewise.
2989 (lto_symtab_encoder_delete_node): Likewise.
2990 (lto_symtab_encoder_in_partition_p): Likewise.
2991 (lto_set_symtab_encoder_in_partition): Likewise.
2992 (output_refs): Likewise.
2993 (compute_ltrans_boundary): Likewise.
2994 (output_symtab): Likewise.
2995 (input_node): Likewise.
2996 (input_ref): Likewise.
2997 (input_edge): Likewise.
2998 (input_cgraph_1): Likewise.
2999 (input_refs): Likewise.
3000 (output_cgraph_opt_summary): Likewise.
3001 (input_node_opt_summary): Likewise.
3002 (input_cgraph_opt_section): Likewise.
3003 * lto-section-in.c (lto_free_function_in_decl_state_for_node):
3005 * lto-streamer-out.c (lto_output): Likewise.
3006 (output_symbol_p): Likewise.
3007 (produce_symtab): Likewise.
3008 * lto-streamer.h (lto_encoder_entry): Likewise.
3009 (lto_free_function_in_decl_state_for_node): Likewise.
3010 (lto_symtab_encoder_encode): Likewise.
3011 (lto_symtab_encoder_delete_node): Likewise.
3012 (lto_symtab_encoder_in_partition_p): Likewise.
3013 (lto_set_symtab_encoder_in_partition): Likewise.
3014 (lto_symtab_encoder_lookup): Likewise.
3015 (lsei_node): Likewise.
3016 (lto_symtab_encoder_deref): Likewise.
3017 * symtab.c (symtab_hash): Likewise.
3018 (assembler_name_hash): Likewise.
3019 (symtab_nodes): Likewise.
3020 (hash_node): Likewise.
3021 (eq_node): Likewise.
3022 (hash_node_by_assembler_name): Likewise.
3023 (eq_assembler_name): Likewise.
3024 (insert_to_assembler_name_hash): Likewise.
3025 (unlink_from_assembler_name_hash): Likewise.
3026 (symtab_prevail_in_asm_name_hash): Likewise.
3027 (symtab_register_node): Likewise.
3028 (symtab_insert_node_to_hashtable): Likewise.
3029 (symtab_unregister_node): Likewise.
3030 (symtab_get_node): Likewise.
3031 (symtab_remove_node): Likewise.
3032 (symtab_initialize_asm_name_hash): Likewise.
3033 (symtab_node_for_asm): Likewise.
3034 (symtab_add_to_same_comdat_group): Likewise.
3035 (symtab_dissolve_same_comdat_group_list): Likewise.
3036 (symtab_node_asm_name): Likewise.
3037 (symtab_node_name): Likewise.
3038 (dump_symtab_base): Likewise.
3039 (dump_symtab_node): Likewise.
3040 (dump_symtab): Likewise.
3041 (debug_symtab_node): Likewise.
3042 (verify_symtab_base): Likewise.
3043 (verify_symtab_node): Likewise.
3044 (verify_symtab): Likewise.
3045 (symtab_used_from_object_file_p): Likewise.
3046 (symtab_node_availability): Likewise.
3047 (symtab_alias_ultimate_target): Likewise.
3048 (fixup_same_cpp_alias_visibility): Likewise.
3049 (symtab_resolve_alias): Likewise.
3050 (symtab_for_node_and_aliases): Likewise.
3051 (symtab_for_node_and_aliases): Likewise.
3052 (symtab_nonoverwritable_alias_1): Likewise.
3053 (symtab_nonoverwritable_alias): Likewise.
3054 (symtab_semantically_equivalent_p): Likewise.
3055 * value-prof.c (init_node_map): Likewise.
3056 * varasm.c (find_decl): Likewise.
3057 * varpool.c (varpool_node_for_asm): Likewise.
3058 (varpool_remove_unreferenced_decls): Likewise.
3060 2013-10-31 David Malcolm <dmalcolm@redhat.com>
3062 Manual part of renaming of symtab_node_base to symtab_node.
3064 * ipa-ref.h (symtab_node): Remove typedef to pointer type, as it
3065 clashes with the preferred name for the base class.
3066 (const_symtab_node): Remove redundant typedef.
3068 2013-10-31 Jakub Jelinek <jakub@redhat.com>
3070 * optabs.c (expand_vec_perm): Avoid vector mode punning
3071 SUBREGs in SET_DEST.
3072 * expmed.c (store_bit_field_1): Likewise.
3073 * config/i386/sse.md (movdi_to_sse, vec_pack_sfix_trunc_v2df,
3074 vec_pack_sfix_v2df, vec_shl_<mode>, vec_shr_<mode>,
3075 vec_interleave_high<mode>, vec_interleave_low<mode>): Likewise.
3076 * config/i386/i386.c (ix86_expand_vector_move_misalign,
3077 ix86_expand_sse_movcc, ix86_expand_int_vcond, ix86_expand_vec_perm,
3078 ix86_expand_sse_unpack, ix86_expand_args_builtin,
3079 ix86_expand_vector_init_duplicate, ix86_expand_vector_set,
3080 emit_reduc_half, expand_vec_perm_blend, expand_vec_perm_pshufb,
3081 expand_vec_perm_interleave2, expand_vec_perm_pshufb2,
3082 expand_vec_perm_vpshufb2_vpermq,
3083 expand_vec_perm_vpshufb2_vpermq_even_odd, expand_vec_perm_even_odd_1,
3084 expand_vec_perm_broadcast_1, expand_vec_perm_vpshufb4_vpermq2,
3085 ix86_expand_sse2_mulv4si3, ix86_expand_pinsr): Likewise.
3086 (expand_vec_perm_palignr): Likewise. Modify a copy of *d rather
3089 2013-10-31 Uros Bizjak <ubizjak@gmail.com>
3091 * config/i386/i386.c (ix86_expand_sse2_abs): Rename function arguments.
3092 Use gcc_unreachable for unhandled modes. Do not check results of
3093 expand_simple_binop. If not expanded to target, move the result.
3095 2013-10-31 Chung-Ju Wu <jasonwucj@gmail.com>
3096 Shiva Chen <shiva0217@gmail.com>
3098 * config.gcc (nds32*-*-*): Add nds32 target.
3099 * config/nds32/nds32.c: New file.
3100 * config/nds32/nds32.h: New file.
3101 * config/nds32/nds32.md: New file.
3102 * config/nds32/constants.md: New file.
3103 * config/nds32/constraints.md: New file.
3104 * config/nds32/iterators.md: New file.
3105 * config/nds32/nds32-doubleword.md: New file.
3106 * config/nds32/nds32-intrinsic.md: New file.
3107 * config/nds32/nds32_intrinsic.h: New file.
3108 * config/nds32/nds32-modes.def: New file.
3109 * config/nds32/nds32-multiple.md: New file.
3110 * config/nds32/nds32.opt: New file.
3111 * config/nds32/nds32-opts.h: New file.
3112 * config/nds32/nds32-protos.h: New file.
3113 * config/nds32/nds32-peephole2.md: New file.
3114 * config/nds32/pipelines.md: New file.
3115 * config/nds32/predicates.md: New file.
3116 * config/nds32/t-mlibs: New file.
3117 * common/config/nds32: New directory and files.
3119 * doc/invoke.texi (NDS32 options): Document nds32 specific options.
3120 * doc/md.texi (NDS32 family): Document nds32 specific constraints.
3121 * doc/install.texi (Cross-Compiler-Specific Options): Document
3122 --with-nds32-lib for nds32 target.
3123 * doc/extend.texi (Function Attributes, Target Builtins): Document
3124 nds32 specific attributes.
3126 2013-10-31 Vladimir Makarov <vmakarov@redhat.com>
3128 * lra-constraints (process_alt_operands): Use the result
3129 elimination register for operand when matching constraints.
3131 2013-10-31 Jakub Jelinek <jakub@redhat.com>
3133 * tree-vrp.c (maybe_set_nonzero_bits): New function.
3134 (remove_range_assertions): Call it.
3136 * tree.c (tree_ctz): New function.
3137 * tree.h (tree_ctz): New prototype.
3138 * tree-ssanames.h (get_range_info, get_nonzero_bits): Change
3139 first argument from tree to const_tree.
3140 * tree-ssanames.c (get_range_info, get_nonzero_bits): Likewise.
3141 * tree-vectorizer.h (vect_generate_tmps_on_preheader): New prototype.
3142 * tree-vect-loop-manip.c (vect_generate_tmps_on_preheader): No longer
3144 * expr.c (highest_pow2_factor): Reimplemented using tree_ctz.
3145 * tree-vect-loop.c (vect_analyze_loop_operations,
3146 vect_transform_loop): Don't force scalar loop for bound just because
3147 number of iterations is unknown, only do it if it is not known to be
3148 a multiple of vectorization_factor.
3149 * builtins.c (get_object_alignment_2): Use tree_ctz on offset.
3151 * gimple-pretty-print.c (dump_ssaname_info): Print newline also
3152 in case of VR_VARYING. Print get_nonzero_bits if not all ones.
3153 * tree-ssanames.h (struct range_info_def): Add nonzero_bits field.
3154 (set_nonzero_bits, get_nonzero_bits): New prototypes.
3155 * tree-ssa-ccp.c (get_default_value): Use get_range_info to see if
3156 a default def isn't partially constant.
3157 (ccp_finalize): If after IPA, set_range_info if integral SSA_NAME
3158 is known to be partially zero.
3159 (evaluate_stmt): If we'd return otherwise VARYING, use get_range_info
3160 to see if a default def isn't partially constant.
3161 * tree-ssanames.c (set_range_info): Initialize nonzero_bits upon
3162 creation of a range, if VR_RANGE, try to improve nonzero_bits from
3164 (set_nonzero_bits, get_nonzero_bits): New functions.
3166 * tree-cfg.c (assert_unreachable_fallthru_edge_p): New function.
3167 * tree-cfg.h (assert_unreachable_fallthru_edge_p): New prototype.
3168 * tree-vrp.c (all_imm_uses_in_stmt_or_feed_cond): New function.
3169 (remove_range_assertions): If ASSERT_EXPR_VAR has no other immediate
3170 uses but in the condition and ASSERT_EXPR and the other successor of
3171 the predecessor bb is __builtin_unreachable (), set_range_info of the
3172 ASSERT_EXPR_VAR to the range info of the ASSERT_EXPR's lhs.
3174 2013-10-31 Martin Jambor <mjambor@suse.cz>
3176 PR rtl-optimization/58934
3178 2013-10-30 Martin Jambor <mjambor@suse.cz>
3179 PR rtl-optimization/10474
3180 * ira.c (find_moveable_pseudos): Do not calculate dominance info
3182 (interesting_dest_for_shprep): New function.
3183 (split_live_ranges_for_shrink_wrap): Likewise.
3184 (ira): Calculate dominance info and df analysis. Call
3185 split_live_ranges_for_shrink_wrap.
3187 2013-10-31 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
3188 Yury Gribov <y.gribov@samsung.com>
3191 * asan.c (asan_clear_shadow): Allocate a new vreg for temporary
3192 shadow pointer to avoid clobbering the main one.
3194 2013-10-31 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3196 * lower-subreg.c (resolve_simple_move): Copy REG_INC note.
3198 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
3201 * ira-color.c (update_costs_from_copies): Add new parameter. Use
3202 it for calling update_costs_from_allocno.
3203 (assign_hard_reg): Call restore_costs_from_copies only for
3204 !retry_p. Pass new argument to update_costs_from_copies.
3205 (color_pass): Pass new argument to update_costs_from_copies.
3206 (ira_mark_allocation_change): Ditto.
3208 2013-10-30 Sharad Singhai <singhai@google.com>
3211 * opts.c (common_handle_option): Remove deprecated option
3212 -ftree-vectorizer-verbose.
3213 * doc/invoke.texi (Debugging Options): Ditto.
3214 * opts-global.c (handle_common_deferred_options): Ditto.
3215 (dump_remap_tree_vectorizer_verbose): Delete.
3216 * common.opt: Set -ftree-vectorizer-verbose as an ignored option.
3218 2013-10-30 DJ Delorie <dj@redhat.com>
3220 * config/rx/rx.c (ADD_RX_BUILTIN0): New macro, used for builtins
3221 that take no arguments.
3223 2013-10-30 Joern Rennecke <joern.rennecke@embecosm.com>
3226 * reload1.c (update_eliminables_and_spill): New function, broken
3228 (reload): Use it. Check for frame size change after frame size
3229 alignment, and call update_eliminables_and_spill first if continue-ing.
3231 2013-10-30 Cong Hou <congh@google.com>
3234 * config/i386/i386-protos.h (ix86_expand_sse2_abs): New function.
3235 * config/i386/i386.c (ix86_expand_sse2_abs): New function.
3236 * config/i386/sse.md: Add SSE2 support to abs (8/16/32-bit-int).
3238 2013-10-18 Mikael Pettersson <mikpelinux@gmail.com>
3240 PR rtl-optimization/58369
3241 * reload1.c (compute_reload_subreg_offset): New function.
3242 (choose_reload_regs): Use it to pass endian-correct
3243 offset to subreg_regno_offset.
3245 2013-10-30 Tobias Burnus <burnus@net-b.de>
3248 * tree-cfg.c (replace_loop_annotate): Replace warning by
3251 2013-10-30 Jason Merrill <jason@redhat.com>
3253 * configure.ac (loose_warn): Add -Wno-format if
3254 --disable-build-format-warnings.
3256 2013-10-30 David Malcolm <dmalcolm@redhat.com>
3258 * cgraphunit.c (analyze_functions): Split symtab_node declarations
3259 onto multiple lines to make things easier for rename_symtab.py.
3261 * symtab.c (symtab_dissolve_same_comdat_group_list): Likewise.
3262 (symtab_semantically_equivalent_p): Likewise.
3264 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
3267 * lra.c (check_rtl): Remove address check before LRA work.
3269 2013-10-30 Marc Glisse <marc.glisse@inria.fr>
3271 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Look for a
3272 POINTER_PLUS_EXPR in the defining statement.
3274 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
3276 * regmove.c: Remove.
3277 * tree-pass.h (make_pass_regmove): Remove.
3278 * timevar.def (TV_REGMOVE): Remove.
3279 * passes.def (pass_regmove): Remove.
3280 * opts.c (default_options_table): Remove entry for regmove.
3281 * doc/passes.texi: Remove regmove pass description.
3282 * doc/invoke.texi (-foptimize-register-move, -fregmove): Remove
3284 (-fdump-rtl-regmove): Ditto.
3285 * common.opt (foptimize-register-move, fregmove): Ignore.
3286 * Makefile.in (OBJS): Remove regmove.o.
3287 * regmove.c: Remove.
3288 * ira-int.h (struct ira_allocno_pref, ira_pref_t): New structure
3290 (struct ira_allocno) New member allocno_prefs.
3291 (ALLOCNO_PREFS): New macro.
3292 (ira_prefs, ira_prefs_num): New external vars.
3293 (ira_setup_alts, ira_get_dup_out_num, ira_debug_pref): New prototypes.
3294 (ira_debug_prefs, ira_debug_allocno_prefs, ira_create_pref): Ditto.
3295 (ira_add_allocno_pref, ira_remove_pref, ira_remove_allocno_prefs):
3297 (ira_add_allocno_copy_to_list): Remove prototype.
3298 (ira_swap_allocno_copy_ends_if_necessary): Ditto.
3299 (ira_pref_iterator): New type.
3300 (ira_pref_iter_init, ira_pref_iter_cond): New functions.
3301 (FOR_EACH_PREF): New macro.
3302 * ira.c (commutative_constraint_p): Move from ira-conflicts.c.
3303 (ira_get_dup_out_num): Ditto. Rename from get_dup_num. Modify the
3305 (ira_setup_alts): New function.
3306 (decrease_live_ranges_number): New function.
3307 (ira): Call the above function.
3308 * ira-build.c (ira_prefs, ira_prefs_num): New global vars.
3309 (ira_create_allocno): Initialize allocno prefs.
3310 (pref_pool, pref_vec): New static vars.
3311 (initiate_prefs, find_allocno_pref, ira_create_pref): New
3313 (add_allocno_pref_to_list, ira_add_allocno_pref, print_pref): Ditto.
3314 (ira_debug_pref, print_prefs, ira_debug_prefs): Ditto.
3315 (print_allocno_prefs, ira_debug_allocno_prefs, finish_pref): Ditto.
3316 (ira_remove_pref, ira_remove_allocno_prefs, finish_prefs): Ditto.
3317 (ira_add_allocno_copy_to_list): Make static. Rename to
3318 add_allocno_copy_to_list.
3319 (ira_swap_allocno_copy_ends_if_necessary): Make static. Rename to
3320 swap_allocno_copy_ends_if_necessary.
3321 (remove_unnecessary_allocnos, remove_low_level_allocnos): Call
3322 ira_remove_allocno_prefs.
3323 (ira_flattening): Ditto.
3324 (ira_build): Call initiate_prefs, print_prefs.
3325 (ira_destroy): Call finish_prefs.
3326 * ira-color.c (struct update_cost_record): New.
3327 (struct allocno_color_data): Add new member update_cost_records.
3328 (update_cost_record_pool): New static var.
3329 (init_update_cost_records, get_update_cost_record): New functions.
3330 (free_update_cost_record_list, finish_update_cost_records): Ditto.
3331 (struct update_cost_queue_elem): Add member from.
3332 (initiate_cost_update): Call init_update_cost_records.
3333 (finish_cost_update): Call finish_update_cost_records.
3334 (queue_update_cost, get_next_update_cost): Add new param from.
3335 (Update_allocno_cost, update_costs_from_allocno): New functions.
3336 (update_costs_from_prefs): Ditto.
3337 (update_copy_costs): Rename to update_costs_from_copies.
3338 (restore_costs_from_copies): New function.
3339 (update_conflict_hard_regno_costs): Don't go back.
3340 (assign_hard_reg): Call restore_costs_from_copies. Add printing
3342 (pop_allocnos): Add priniting more debug info.
3343 (color_allocnos): Remove prefs for conflicting hard regs.
3344 Call update_costs_from_prefs.
3345 * ira-conflicts.c (commutative_constraint_p): Move to ira.c
3346 (get_dup_num): Rename, modify, and move to ira.c
3347 (process_regs_for_copy): Add prefs.
3348 (add_insn_allocno_copies): Put src as first arg of
3349 process_regs_for_copy. Remove dead code. Call ira_setup_alts.
3350 * ira-costs.c (record_reg_classes): Modify and move code into
3351 record_operands_costs.
3352 (find_costs_and_classes): Create prefs for the hard reg of small
3354 (process_bb_node_for_hard_reg_moves): Add prefs.
3356 2013-10-30 Richard Biener <rguenther@suse.de>
3359 * basic-block.h (pre_and_rev_post_order_compute_fn): New function.
3360 * cfganal.c (pre_and_rev_post_order_compute_fn): New function
3362 (pre_and_rev_post_order_compute): ... which now wraps it.
3363 * graph.c (draw_cfg_nodes_no_loops): Use
3364 pre_and_rev_post_order_compute_fn to avoid ICEing and dependence
3367 2013-10-30 Christian Bruel <christian.bruel@st.com>
3369 * config/sh/sh-mem.cc (sh_expand_cmpnstr): New function.
3370 (sh_expand_cmpstr): Handle known align and schedule improvements.
3371 * config/sh/sh-protos.h (sh_expand_cmpstrn): Declare.
3372 * config/sh/sh.md (cmpstrnsi): New pattern.
3374 2013-10-30 Martin Jambor <mjambor@suse.cz>
3376 PR rtl-optimization/10474
3377 * ira.c (find_moveable_pseudos): Do not calculate dominance info
3379 (interesting_dest_for_shprep): New function.
3380 (split_live_ranges_for_shrink_wrap): Likewise.
3381 (ira): Calculate dominance info and df analysis. Call
3382 split_live_ranges_for_shrink_wrap.
3384 2013-10-30 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3387 * config/arm/arm.c (arm_expand_epilogue_apcs_frame): Emit blockage.
3389 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
3391 * tree-core.h (tree_index): Add TI_POINTER_BOUNDS_TYPE.
3392 * tree.h (POINTER_BOUNDS_P): New.
3393 (BOUNDED_TYPE_P): New.
3395 (pointer_bounds_type_node): New.
3396 * tree.c (build_common_tree_nodes): Initialize
3397 pointer_bounds_type_node.
3398 * gimple.h (gimple_call_get_nobnd_arg_index): New.
3399 (gimple_call_num_nobnd_args): New.
3400 (gimple_call_nobnd_arg): New.
3401 (gimple_return_retbnd): New.
3402 (gimple_return_set_retbnd): New
3403 * gimple.c (gimple_build_return): Increase number of ops
3404 for return statement.
3405 (gimple_call_get_nobnd_arg_index): New.
3406 * gimple-pretty-print.c (dump_gimple_return): Print second op.
3408 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
3410 * ipa.c (cgraph_build_static_cdtor_1): Support contructors
3411 with "chkp ctor" and "bnd_legacy" attributes.
3412 * gimplify.c (gimplify_init_constructor): Avoid infinite
3413 loop during gimplification of bounds initializer.
3415 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
3417 * c-family/c-common.c (handle_bnd_variable_size_attribute): New.
3418 (handle_bnd_legacy): New.
3419 (c_common_attribute_table): Add bnd_variable_size and bnd_legacy.
3420 * doc/extend.texi: Document bnd_variable_size and bnd_legacy
3423 2013-10-29 Ilya Enkovich <ilya.enkovich@intel.com>
3425 * builtin-types.def (BT_FN_VOID_CONST_PTR): New.
3426 (BT_FN_PTR_CONST_PTR): New.
3427 (BT_FN_CONST_PTR_CONST_PTR): New.
3428 (BT_FN_PTR_CONST_PTR_SIZE): New.
3429 (BT_FN_PTR_CONST_PTR_CONST_PTR): New.
3430 (BT_FN_VOID_PTRPTR_CONST_PTR): New.
3431 (BT_FN_VOID_CONST_PTR_SIZE): New.
3432 (BT_FN_PTR_CONST_PTR_CONST_PTR_SIZE): New.
3433 * chkp-builtins.def: New.
3434 * builtins.def: include chkp-builtins.def.
3435 (DEF_CHKP_BUILTIN): New.
3436 * builtins.c (expand_builtin): Support BUILT_IN_CHKP_INIT_PTR_BOUNDS,
3437 BUILT_IN_CHKP_NULL_PTR_BOUNDS, BUILT_IN_CHKP_COPY_PTR_BOUNDS,
3438 BUILT_IN_CHKP_CHECK_PTR_LBOUNDS, BUILT_IN_CHKP_CHECK_PTR_UBOUNDS,
3439 BUILT_IN_CHKP_CHECK_PTR_BOUNDS, BUILT_IN_CHKP_SET_PTR_BOUNDS,
3440 BUILT_IN_CHKP_NARROW_PTR_BOUNDS, BUILT_IN_CHKP_STORE_PTR_BOUNDS,
3441 BUILT_IN_CHKP_GET_PTR_LBOUND, BUILT_IN_CHKP_GET_PTR_UBOUND,
3442 BUILT_IN_CHKP_BNDMK, BUILT_IN_CHKP_BNDSTX, BUILT_IN_CHKP_BNDCL,
3443 BUILT_IN_CHKP_BNDCU, BUILT_IN_CHKP_BNDLDX, BUILT_IN_CHKP_BNDRET,
3444 BUILT_IN_CHKP_INTERSECT, BUILT_IN_CHKP_ARG_BND, BUILT_IN_CHKP_NARROW,
3445 BUILT_IN_CHKP_EXTRACT_LOWER, BUILT_IN_CHKP_EXTRACT_UPPER.
3446 * common.opt (fcheck-pointer-bounds): New.
3447 * toplev.c (process_options): Check Pointer Bounds Checker is
3449 * doc/extend.texi: Document Pointer Bounds Checker built-in functions.
3451 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
3453 * target.def (builtin_chkp_function): New.
3454 (chkp_bound_type): New.
3455 (chkp_bound_mode): New.
3456 (fn_abi_va_list_bounds_size): New.
3457 (load_bounds_for_arg): New.
3458 (store_bounds_for_arg): New.
3459 * targhooks.h (default_load_bounds_for_arg): New.
3460 (default_store_bounds_for_arg): New.
3461 (default_fn_abi_va_list_bounds_size): New.
3462 (default_chkp_bound_type): New.
3463 (default_chkp_bound_mode): New.
3464 (default_builtin_chkp_function): New.
3465 * targhooks.c (default_load_bounds_for_arg): New.
3466 (default_store_bounds_for_arg): New.
3467 (default_fn_abi_va_list_bounds_size): New.
3468 (default_chkp_bound_type): New.
3469 (default_chkp_bound_mode); New.
3470 (default_builtin_chkp_function): New.
3471 * doc/tm.texi.in (TARGET_FN_ABI_VA_LIST_BOUNDS_SIZE): New.
3472 (TARGET_LOAD_BOUNDS_FOR_ARG): New.
3473 (TARGET_STORE_BOUNDS_FOR_ARG): New.
3474 (TARGET_BUILTIN_CHKP_FUNCTION): New.
3475 (TARGET_CHKP_BOUND_TYPE): New.
3476 (TARGET_CHKP_BOUND_MODE): New.
3477 * doc/tm.texi: Regenerated.
3478 * langhooks.h (lang_hooks): Add chkp_supported field.
3479 * langhooks-def.h (LANG_HOOKS_CHKP_SUPPORTED): New.
3480 (LANG_HOOKS_INITIALIZER); Add LANG_HOOKS_CHKP_SUPPORTED.
3482 2013-10-29 Andrew Pinski <apinski@cavium.com>
3484 * tree-ssa-ifcombine.c: Include rtl.h and tm_p.h.
3485 (ifcombine_ifandif): Handle cases where maybe_fold_and_comparisons
3486 fails, combining the branches anyways.
3487 (tree_ssa_ifcombine): Inverse the order of the basic block walk,
3488 increases the number of combinings.
3489 * gimple.h (gsi_start_nondebug_after_labels_bb): New function.
3491 2013-10-29 Mike Stump <mikestump@comcast.net>
3493 * machmode.def (PARTIAL_INT_MODE): Add precision and name.
3494 * genmodes.c (PARTIAL_INT_MODE): Add precision and name.
3495 (make_vector_mode): Increase namebuf to 16.
3496 (emit_insn_modes_h): When processing BImode, don't
3497 also match partial int modes.
3498 (emit_class_narrowest_mode): Likewise.
3500 * config/bfin/bfin-modes.def: Add precision to PDI.
3501 * config/m32c/m32c-modes.def: Add precision to PSI.
3502 * config/msp430/msp430-modes.def: Add precision to PSI.
3503 * config/rs6000/rs6000-modes.def: Add precision to PTI.
3504 * config/sh/sh-modes.def: Add precision to PSI and PDI.
3506 2013-10-29 Oleg Endo <olegendo@gcc.gnu.org>
3509 * config/sh/sh.md (*addc): Rename existing variations to ...
3510 (*addc_r_r_1, *addc_2r_1, *addc_r_1): ... these.
3511 (*addc_r_lsb, *addc_r_r_lsb, *addc_r_lsb_r, *addc_2r_lsb, *addc_r_msb,
3512 *addc_r_r_msb, *addc_2r_msb): New insn_and_split patterns.
3513 * config/sh/sh.c (addsubcosts): Handle some addc special cases.
3515 2013-10-29 Teresa Johnson <tejohnson@google.com>
3518 * tree-ssa-tail-merge.c (replace_block_by): Tolerate profile
3519 insanities when updating probabilities.
3521 2013-10-29 David Malcolm <dmalcolm@redhat.com>
3523 * gdbhooks.py (CGraphNodePrinter.to_string): Update gdb
3524 prettyprinter for cgraph_node to reflect the conversion of the
3525 symtable types to a C++ class hierarchy: it now *is* a
3526 symtab_node_base, rather than having one (named "symbol").
3528 2013-10-29 Balaji V. Iyer <balaji.v.iyer@intel.com>
3530 * builtins.c (is_builtin_name): Added a check for __cilkrts_detach and
3531 __cilkrts_pop_frame. If matched, then return true for built-in
3533 (expand_builtin): Added BUILT_IN_CILK_DETACH and
3534 BUILT_IN_CILK_POP_FRAME case.
3535 * langhooks-def.h (lhd_install_body_with_frame_cleanup): New prototype.
3536 (lhs_cilk_detect_spawn): Likewise.
3537 (LANG_HOOKS_DECLS): Added LANG_HOOKS_CILKPLUS.
3538 (LANG_HOOKS_CILKPLUS_DETECT_SPAWN_AND_UNWRAP): New #define.
3539 (LANG_HOOKS_CILKPLUS_FRAME_CLEANUP): Likewise.
3540 (LANG_HOOKS_CILKPLUS_GIMPLIFY_SPAWN): Likewise.
3541 (LANG_HOOKS_CILKPLUS): Likewise.
3542 * tree.h (CILK_SPAWN_FN): Likewise.
3543 * builtin.def (DEF_CILK_BUILTIN_STUB): Likewise.
3544 * Makefile.in (C_COMMON_OBJS): Added c-family/cilk.o.
3545 (OBJS): Added cilk-common.o.
3546 (BUILTINS_DEF): Added cilk-builtins.def.
3547 * langhooks.c (lhd_install_body_with_frame_cleanup): New function.
3548 (lhd_cilk_detect_spawn): Likewise.
3549 * langhooks.h (lang_hooks_for_cilkplus): New struct.
3550 (struct lang_hooks): Added new field called "cilkplus."
3551 * cilk-common.c: New file.
3553 * cilk-builtins.def: Likewise.
3554 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Added
3555 "__cilk" macro and set it to 200.
3556 * function.h (struct function::cilk_frame_decl): New field.
3557 (struct function::is_cilk_function): Likewise.
3558 (struct function::calls_cilk_spawn): Likewise.
3559 * gimplify.c (gimplify_call_expr): Added a check if the function call
3560 being gimplified is a spawn detach point. If so, then add pop_frame
3561 and detach function calls.
3562 (gimplify_expr): Added a CILK_SPAWN_STMT and CILK_SYNC_STMT case
3563 for gimplifying _Cilk_spawn and _Cilk_sync statements.
3564 (gimplify_return_expr): Added a check for _Cilk_spawn usage in
3565 function. If so, added a _Cilk_sync and gimplified it.
3566 (gimplify_modify_expr): Added a check for _Cilk_spawn in MODIFY and
3567 INIT_EXPRs. If so, then call gimplify_cilk_spawn.
3568 * ipa-inline-analysis (initialize_inline_failed): Prevent inlining of
3570 (can_inline_edge_p): Prevent inling of spawnee function.
3571 * ira.c (ira_setup_eliminable_regset): Force usage of frame pointer
3572 for functions that use Cilk keywords.
3573 * tree-inline.h (struct copy_body_data::remap_var_for_cilk): New field.
3574 * tree-pretty-print.c (dump_generic_node): Added CILK_SPAWN_STMT and
3575 CILK_SYNC_STMT cases.
3576 * tree.def (DEFTREECODE): Added CILK_SPAWN_STMT and CILK_SYNC_STMT
3578 * generic.texi (CILK_SPAWN_STMT): Added documentation for _Cilk_spawn.
3579 (CILK_SYNC_STMT): Added documentation for _Cilk_sync.
3580 * passes.texi (Cilk Keywords): New section that describes the compiler
3581 code changes for handling Cilk Keywords.
3583 2013-10-29 David Malcolm <dmalcolm@redhat.com>
3585 Patch autogenerated by refactor_symtab.py from
3586 https://github.com/davidmalcolm/gcc-refactoring-scripts
3587 revision 58bb219cc090b2f4516a9297d868c245495ee622
3589 * asan.c (asan_finish_file): Update for conversion of symtab types to
3590 a true class hierarchy.
3591 * cfgexpand.c (estimated_stack_frame_size): Likewise.
3592 * cgraph.c (cgraph_get_body): Likewise.
3593 (cgraph_get_create_real_symbol_node): Likewise.
3594 (verify_cgraph_node): Likewise.
3595 (verify_edge_corresponds_to_fndecl): Likewise.
3596 (verify_edge_count_and_frequency): Likewise.
3597 (cgraph_will_be_removed_from_program_if_no_direct_calls): Likewise.
3598 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
3599 (cgraph_can_remove_if_no_direct_calls_and_refs_p): Likewise.
3600 (cgraph_node_cannot_return): Likewise.
3601 (cgraph_set_pure_flag_1): Likewise.
3602 (cgraph_set_const_flag_1): Likewise.
3603 (cgraph_set_nothrow_flag_1): Likewise.
3604 (cgraph_make_node_local_1): Likewise.
3605 (cgraph_for_node_and_aliases): Likewise.
3606 (cgraph_for_node_thunks_and_aliases): Likewise.
3607 (cgraph_node_can_be_local_p): Likewise.
3608 (cgraph_node_cannot_be_local_p_1): Likewise.
3609 (cgraph_function_body_availability): Likewise.
3610 (dump_cgraph_node): Likewise.
3611 (cgraph_rtl_info): Likewise.
3612 (cgraph_mark_address_taken_node): Likewise.
3613 (cgraph_remove_node): Likewise.
3614 (cgraph_release_function_body): Likewise.
3615 (cgraph_update_edges_for_call_stmt_node): Likewise.
3616 (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
3617 (cgraph_make_edge_direct): Likewise.
3618 (cgraph_resolve_speculation): Likewise.
3619 (cgraph_speculative_call_info): Likewise.
3620 (cgraph_turn_edge_to_speculative): Likewise.
3621 (cgraph_create_edge_1): Likewise.
3622 (cgraph_set_call_stmt): Likewise.
3623 (cgraph_node_for_asm): Likewise.
3624 (cgraph_add_thunk): Likewise.
3625 (cgraph_same_body_alias): Likewise.
3626 (cgraph_create_function_alias): Likewise.
3627 (cgraph_create_node): Likewise.
3628 (cgraph_create_empty_node): Likewise.
3629 (record_function_versions): Likewise.
3630 (used_from_object_file_p): Likewise.
3631 * cgraph.h (symtab_can_be_discarded): Likewise.
3632 (symtab_real_symbol_p): Likewise.
3633 (cgraph_mark_force_output_node): Likewise.
3634 (cgraph_edge_recursive_p): Likewise.
3635 (symtab_alias_target): Likewise.
3636 (varpool_all_refs_explicit_p): Likewise.
3637 (varpool_can_remove_if_no_refs): Likewise.
3638 (cgraph_only_called_directly_or_aliased_p): Likewise.
3639 (cgraph_next_function_with_gimple_body): Likewise.
3640 (cgraph_first_function_with_gimple_body): Likewise.
3641 (cgraph_function_with_gimple_body_p): Likewise.
3642 (cgraph_next_function): Likewise.
3643 (cgraph_first_function): Likewise.
3644 (cgraph_next_defined_function): Likewise.
3645 (cgraph_first_defined_function): Likewise.
3646 (varpool_next_defined_variable): Likewise.
3647 (varpool_first_defined_variable): Likewise.
3648 (varpool_next_static_initializer): Likewise.
3649 (varpool_first_static_initializer): Likewise.
3650 (varpool_next_variable): Likewise.
3651 (varpool_first_variable): Likewise.
3652 (varpool_node_name): Likewise.
3653 (varpool): Likewise.
3655 (is_a_helper <varpool_node>::test): Likewise.
3656 (is_a_helper <cgraph_node>::test): Likewise.
3657 (varpool_variable_node): Likewise.
3658 (cgraph_function_or_thunk_node): Likewise.
3659 (varpool_alias_target): Likewise.
3660 (cgraph_alias_target): Likewise.
3661 (cgraph_node_name): Likewise.
3662 (varpool_node_asm_name): Likewise.
3663 (cgraph_node_asm_name): Likewise.
3664 * cgraphbuild.c (remove_cgraph_callee_edges): Likewise.
3665 (cgraph_rebuild_references): Likewise.
3666 (rebuild_cgraph_edges): Likewise.
3667 (record_eh_tables): Likewise.
3668 (build_cgraph_edges): Likewise.
3669 (mark_store): Likewise.
3670 (mark_load): Likewise.
3671 (mark_address): Likewise.
3672 (record_type_list): Likewise.
3673 (record_reference): Likewise.
3674 * cgraphclones.c (cgraph_materialize_all_clones): Likewise.
3675 (cgraph_materialize_clone): Likewise.
3676 (cgraph_function_versioning): Likewise.
3677 (cgraph_copy_node_for_versioning): Likewise.
3678 (update_call_expr): Likewise.
3679 (cgraph_find_replacement_node): Likewise.
3680 (cgraph_create_virtual_clone): Likewise.
3681 (cgraph_clone_node): Likewise.
3682 * cgraphunit.c (compile): Likewise.
3683 (output_weakrefs): Likewise.
3684 (output_in_order): Likewise.
3685 (expand_function): Likewise.
3686 (assemble_thunks_and_aliases): Likewise.
3687 (expand_thunk): Likewise.
3688 (mark_functions_to_output): Likewise.
3689 (handle_alias_pairs): Likewise.
3690 (analyze_functions): Likewise.
3691 (walk_polymorphic_call_targets): Likewise.
3692 (varpool_finalize_decl): Likewise.
3693 (process_function_and_variable_attributes): Likewise.
3694 (cgraph_process_same_body_aliases): Likewise.
3695 (analyze_function): Likewise.
3696 (cgraph_add_new_function): Likewise.
3697 (cgraph_finalize_function): Likewise.
3698 (referred_to_p): Likewise.
3699 (cgraph_reset_node): Likewise.
3700 (cgraph_process_new_functions): Likewise.
3701 (enqueue_node): Likewise.
3702 (decide_is_symbol_needed): Likewise.
3703 * coverage.c (coverage_compute_profile_id): Likewise.
3704 * dbxout.c (dbxout_expand_expr): Likewise.
3705 * dwarf2out.c (premark_types_used_by_global_vars_helper): Likewise.
3706 (reference_to_unused): Likewise.
3707 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
3708 * gimplify.c (unvisit_body): Likewise.
3709 (unshare_body): Likewise.
3710 * ipa-cp.c (ipcp_generate_summary): Likewise.
3711 (ipcp_decision_stage): Likewise.
3712 (identify_dead_nodes): Likewise.
3713 (decide_whether_version_node): Likewise.
3714 (decide_about_value): Likewise.
3715 (perhaps_add_new_callers): Likewise.
3716 (create_specialized_node): Likewise.
3717 (update_profiling_info): Likewise.
3718 (ipcp_propagate_stage): Likewise.
3719 (estimate_local_effects): Likewise.
3720 (good_cloning_opportunity_p): Likewise.
3721 (devirtualization_time_bonus): Likewise.
3722 (propagate_constants_accross_call): Likewise.
3723 (initialize_node_lattices): Likewise.
3724 (ipcp_cloning_candidate_p): Likewise.
3725 (determine_versionability): Likewise.
3726 (print_all_lattices): Likewise.
3727 (print_lattice): Likewise.
3728 (ipcp_discover_new_direct_edges): Likewise.
3729 * ipa-devirt.c (ipa_devirt): Likewise.
3730 (likely_target_p): Likewise.
3731 (update_type_inheritance_graph): Likewise.
3732 (possible_polymorphic_call_target_p): Likewise.
3733 (dump_possible_polymorphic_call_targets): Likewise.
3734 (devirt_variable_node_removal_hook): Likewise.
3735 (record_binfo): Likewise.
3736 (maybe_record_node): Likewise.
3737 (build_type_inheritance_graph): Likewise.
3738 * ipa-inline-analysis.c (inline_write_summary): Likewise.
3739 (inline_generate_summary): Likewise.
3740 (inline_analyze_function): Likewise.
3741 (do_estimate_growth): Likewise.
3742 (simple_edge_hints): Likewise.
3743 (estimate_node_size_and_time): Likewise.
3744 (estimate_edge_devirt_benefit): Likewise.
3745 (compute_inline_parameters): Likewise.
3746 (estimate_function_body_sizes): Likewise.
3747 (compute_bb_predicates): Likewise.
3748 (initialize_inline_failed): Likewise.
3749 (dump_inline_summary): Likewise.
3750 (dump_inline_edge_summary): Likewise.
3751 * ipa-inline-transform.c (inline_transform): Likewise.
3752 (preserve_function_body_p): Likewise.
3753 (save_inline_function_body): Likewise.
3754 (inline_call): Likewise.
3755 (clone_inlined_nodes): Likewise.
3756 (can_remove_node_now_p): Likewise.
3757 (can_remove_node_now_p_1): Likewise.
3758 * ipa-inline.c (early_inliner): Likewise.
3759 (early_inline_small_functions): Likewise.
3760 (inline_always_inline_functions): Likewise.
3761 (ipa_inline): Likewise.
3762 (flatten_function): Likewise.
3763 (inline_small_functions): Likewise.
3764 (speculation_useful_p): Likewise.
3765 (recursive_inlining): Likewise.
3766 (update_caller_keys): Likewise.
3767 (reset_edge_caches): Likewise.
3768 (update_edge_key): Likewise.
3769 (edge_badness): Likewise.
3770 (relative_time_benefit): Likewise.
3771 (want_inline_self_recursive_call_p): Likewise.
3772 (want_inline_small_function_p): Likewise.
3773 (want_early_inline_function_p): Likewise.
3774 (num_calls): Likewise.
3775 (can_early_inline_edge_p): Likewise.
3776 (can_inline_edge_p): Likewise.
3777 (report_inline_failed_reason): Likewise.
3778 * ipa-profile.c (ipa_profile): Likewise.
3779 (ipa_propagate_frequency): Likewise.
3780 (ipa_propagate_frequency_1): Likewise.
3781 (ipa_profile_generate_summary): Likewise.
3782 * ipa-prop.c (ipcp_transform_function): Likewise.
3783 (read_replacements_section): Likewise.
3784 (ipa_prop_read_section): Likewise.
3785 (ipa_modify_call_arguments): Likewise.
3786 (ipa_print_node_params): Likewise.
3787 (propagate_controlled_uses): Likewise.
3788 (update_indirect_edges_after_inlining): Likewise.
3789 (remove_described_reference): Likewise.
3790 (ipa_make_edge_direct_to_target): Likewise.
3791 (ipa_analyze_node): Likewise.
3792 (ipa_analyze_params_uses): Likewise.
3793 (ipa_compute_jump_functions): Likewise.
3794 (ipa_get_callee_param_type): Likewise.
3795 (ipa_print_node_jump_functions): Likewise.
3796 (ipa_initialize_node_params): Likewise.
3797 (ipa_populate_param_decls): Likewise.
3798 (ipa_func_spec_opts_forbid_analysis_p): Likewise.
3799 (write_agg_replacement_chain): Likewise.
3800 (ipa_write_node_info): Likewise.
3801 (ipa_edge_duplication_hook): Likewise.
3802 (try_decrement_rdesc_refcount): Likewise.
3803 * ipa-pure-const.c (propagate_nothrow): Likewise.
3804 (propagate_pure_const): Likewise.
3805 (pure_const_read_summary): Likewise.
3806 (pure_const_write_summary): Likewise.
3807 (analyze_function): Likewise.
3808 * ipa-ref-inline.h (ipa_ref_referred_ref_list): Likewise.
3809 (ipa_ref_referring_ref_list): Likewise.
3810 * ipa-ref.c (ipa_clear_stmts_in_references): Likewise.
3811 (ipa_remove_stmt_references): Likewise.
3812 (ipa_find_reference): Likewise.
3813 (ipa_dump_referring): Likewise.
3814 (ipa_dump_references): Likewise.
3815 (ipa_record_reference): Likewise.
3816 * ipa-reference.c (ipa_reference_read_optimization_summary): Likewise.
3817 (ipa_reference_write_optimization_summary): Likewise.
3818 (write_node_summary_p): Likewise.
3819 (propagate): Likewise.
3820 (read_write_all_from_decl): Likewise.
3821 (generate_summary): Likewise.
3822 (analyze_function): Likewise.
3823 (propagate_bits): Likewise.
3824 (ipa_reference_get_not_written_global): Likewise.
3825 (ipa_reference_get_not_read_global): Likewise.
3826 * ipa-split.c (execute_split_functions): Likewise.
3827 (split_function): Likewise.
3828 * ipa-utils.c (ipa_merge_profiles): Likewise.
3829 (dump_cgraph_node_set): Likewise.
3830 (ipa_reverse_postorder): Likewise.
3831 (ipa_edge_within_scc): Likewise.
3832 (ipa_get_nodes_in_cycle): Likewise.
3833 (ipa_free_postorder_info): Likewise.
3834 (ipa_reduced_postorder): Likewise.
3835 (searchc): Likewise.
3836 (recursive_call_p): Likewise.
3837 * ipa.c (ipa_cdtor_merge): Likewise.
3838 (record_cdtor_fn): Likewise.
3839 (function_and_variable_visibility): Likewise.
3840 (varpool_externally_visible_p): Likewise.
3841 (cgraph_externally_visible_p): Likewise.
3842 (comdat_can_be_unshared_p): Likewise.
3843 (comdat_can_be_unshared_p_1): Likewise.
3844 (address_taken_from_non_vtable_p): Likewise.
3845 (ipa_discover_readonly_nonaddressable_vars): Likewise.
3846 (symtab_remove_unreachable_nodes): Likewise.
3847 (walk_polymorphic_call_targets): Likewise.
3848 (process_references): Likewise.
3849 (enqueue_node): Likewise.
3850 (has_addr_references_p): Likewise.
3851 (cgraph_non_local_node_p_1): Likewise.
3852 * is-a.h (varpool_analyze_node): Likewise.
3853 * lto-cgraph.c (input_symtab): Likewise.
3854 (merge_profile_summaries): Likewise.
3855 (input_cgraph_1): Likewise.
3856 (input_edge): Likewise.
3857 (input_varpool_node): Likewise.
3858 (input_node): Likewise.
3859 (input_overwrite_node): Likewise.
3860 (compute_ltrans_boundary): Likewise.
3861 (output_refs): Likewise.
3862 (lto_output_varpool_node): Likewise.
3863 (lto_output_node): Likewise.
3864 (reachable_from_other_partition_p): Likewise.
3865 (referenced_from_other_partition_p): Likewise.
3866 (lto_output_edge): Likewise.
3867 (output_node_opt_summary): Likewise.
3868 (add_node_to): Likewise.
3869 (reachable_from_this_partition_p): Likewise.
3870 (lto_set_symtab_encoder_in_partition): Likewise.
3871 (lto_symtab_encoder_in_partition_p): Likewise.
3872 (lto_set_symtab_encoder_encode_initializer): Likewise.
3873 (lto_symtab_encoder_encode_initializer_p): Likewise.
3874 (lto_set_symtab_encoder_encode_body): Likewise.
3875 (lto_symtab_encoder_encode_body_p): Likewise.
3876 * lto-section-in.c (lto_free_function_in_decl_state_for_node):
3878 * lto-streamer-in.c (lto_read_body): Likewise.
3879 (fixup_call_stmt_edges): Likewise.
3880 (fixup_call_stmt_edges_1): Likewise.
3881 * lto-streamer-out.c (produce_symtab): Likewise.
3882 (output_symbol_p): Likewise.
3883 (write_symbol): Likewise.
3884 (lto_output): Likewise.
3885 (copy_function): Likewise.
3886 (output_function): Likewise.
3887 * passes.c (function_called_by_processed_nodes_p): Likewise.
3888 (ipa_write_optimization_summaries): Likewise.
3889 (ipa_write_summaries): Likewise.
3890 (do_per_function_toporder): Likewise.
3891 (do_per_function): Likewise.
3892 (dump_passes): Likewise.
3893 * symtab.c (symtab_semantically_equivalent_p): Likewise.
3894 (symtab_nonoverwritable_alias): Likewise.
3895 (symtab_nonoverwritable_alias_1): Likewise.
3896 (symtab_for_node_and_aliases): Likewise.
3897 (symtab_resolve_alias): Likewise.
3898 (fixup_same_cpp_alias_visibility): Likewise.
3899 (symtab_alias_ultimate_target): Likewise.
3900 (symtab_used_from_object_file_p): Likewise.
3901 (verify_symtab_base): Likewise.
3902 (dump_symtab_base): Likewise.
3903 (symtab_node_name): Likewise.
3904 (symtab_node_asm_name): Likewise.
3905 (symtab_dissolve_same_comdat_group_list): Likewise.
3906 (symtab_add_to_same_comdat_group): Likewise.
3907 (symtab_unregister_node): Likewise.
3908 (symtab_insert_node_to_hashtable): Likewise.
3909 (symtab_register_node): Likewise.
3910 (unlink_from_assembler_name_hash): Likewise.
3911 (insert_to_assembler_name_hash): Likewise.
3912 (eq_assembler_name): Likewise.
3913 (hash_node_by_assembler_name): Likewise.
3914 (eq_node): Likewise.
3915 (hash_node): Likewise.
3916 * toplev.c (wrapup_global_declaration_2): Likewise.
3917 * trans-mem.c (ipa_tm_execute): Likewise.
3918 (ipa_tm_transform_clone): Likewise.
3919 (ipa_tm_transform_transaction): Likewise.
3920 (ipa_tm_transform_calls_redirect): Likewise.
3921 (ipa_tm_insert_gettmclone_call): Likewise.
3922 (ipa_tm_insert_irr_call): Likewise.
3923 (ipa_tm_create_version): Likewise.
3924 (ipa_tm_create_version_alias): Likewise.
3925 (ipa_tm_mark_forced_by_abi_node): Likewise.
3926 (ipa_tm_mark_force_output_node): Likewise.
3927 (ipa_tm_diagnose_tm_safe): Likewise.
3928 (ipa_tm_mayenterirr_function): Likewise.
3929 (ipa_tm_scan_irr_function): Likewise.
3930 (ipa_tm_note_irrevocable): Likewise.
3931 (ipa_tm_scan_calls_clone): Likewise.
3932 (get_cg_data): Likewise.
3933 * tree-eh.c (tree_could_trap_p): Likewise.
3934 * tree-emutls.c (ipa_lower_emutls): Likewise.
3935 (create_emultls_var): Likewise.
3936 (lower_emutls_function_body): Likewise.
3937 (gen_emutls_addr): Likewise.
3938 (emutls_decl): Likewise.
3939 (new_emutls_decl): Likewise.
3940 * tree-inline.c (tree_function_versioning): Likewise.
3941 (optimize_inline_calls): Likewise.
3942 (expand_call_inline): Likewise.
3943 (estimate_num_insns): Likewise.
3944 (copy_bb): Likewise.
3945 (delete_unreachable_blocks_update_callgraph): Likewise.
3946 * tree-nested.c (gimplify_all_functions): Likewise.
3947 (create_nesting_tree): Likewise.
3948 (check_for_nested_with_variably_modified): Likewise.
3949 * tree-pretty-print.c (dump_function_header): Likewise.
3950 * tree-profile.c (tree_profiling): Likewise.
3951 * tree-sra.c (ipa_sra_preliminary_function_checks): Likewise.
3952 (modify_function): Likewise.
3953 (convert_callers): Likewise.
3954 (convert_callers_for_node): Likewise.
3955 * tree-ssa-structalias.c (ipa_pta_execute): Likewise.
3956 (associate_varinfo_to_alias): Likewise.
3957 (create_variable_info_for): Likewise.
3958 (get_constraint_for_ssa_var): Likewise.
3959 * tree-vectorizer.c (increase_alignment): Likewise.
3960 * tree.c (find_decls_types_in_var): Likewise.
3961 (find_decls_types_in_node): Likewise.
3962 (free_lang_data_in_decl): Likewise.
3963 * value-prof.c (gimple_ic_transform): Likewise.
3964 (gimple_ic): Likewise.
3965 (check_ic_target): Likewise.
3966 (init_node_map): Likewise.
3967 * varasm.c (decl_binds_to_current_def_p): Likewise.
3968 (default_binds_local_p_1): Likewise.
3969 (dump_tm_clone_pairs): Likewise.
3970 (assemble_alias): Likewise.
3971 (find_decl): Likewise.
3972 (mark_decl_referenced): Likewise.
3973 * varpool.c (varpool_for_node_and_aliases): Likewise.
3974 (varpool_extra_name_alias): Likewise.
3975 (varpool_create_variable_alias): Likewise.
3976 (add_new_static_var): Likewise.
3977 (varpool_finalize_named_section_flags): Likewise.
3978 (varpool_remove_unreferenced_decls): Likewise.
3979 (enqueue_node): Likewise.
3980 (varpool_assemble_decl): Likewise.
3981 (assemble_aliases): Likewise.
3982 (varpool_analyze_node): Likewise.
3983 (cgraph_variable_initializer_availability): Likewise.
3984 (varpool_add_new_variable): Likewise.
3985 (ctor_for_folding): Likewise.
3986 (dump_varpool_node): Likewise.
3987 (varpool_remove_initializer): Likewise.
3988 (varpool_remove_node): Likewise.
3989 (varpool_node_for_decl): Likewise.
3990 (varpool_create_empty_node): Likewise.
3991 * config/i386/i386.c (ix86_generate_version_dispatcher_body): Likewise.
3992 (ix86_get_function_versions_dispatcher): Likewise.
3994 2013-10-29 David Malcolm <dmalcolm@redhat.com>
3996 * cgraph.h (symtab_node_base): Convert to a class;
3997 add GTY((desc ("%h.type"), tag ("SYMTAB_SYMBOL"))), and take
3998 chain_next/prev from symtab_node_def.
3999 (cgraph_node): Inherit from symtab_node; add GTY option
4000 tag ("SYMTAB_FUNCTION").
4001 (varpool_node): Inherit from symtab_node; add GTY option
4002 tag ("SYMTAB_VARIABLE").
4003 (symtab_node_def): Remove.
4004 (is_a_helper <cgraph_node>::test (symtab_node_def *)): Convert to...
4005 (is_a_helper <cgraph_node>::test (symtab_node_base *)): ...this.
4006 (is_a_helper <varpool_node>::test (symtab_node_def *)): Convert to...
4007 (is_a_helper <varpool_node>::test (symtab_node_base *)): ...this.
4009 * ipa-ref.h (symtab_node_def): Drop.
4010 (symtab_node): Change underlying type from symtab_node_def to
4012 (const_symtab_node): Likwise.
4014 * is-a.h: Update examples in comment.
4016 * symtab.c (symtab_hash): Change symtab_node_def to symtab_node_base.
4017 (assembler_name_hash): Likewise.
4019 2013-10-29 Martin Liska <marxin.liska@gmail.com>
4021 * doc/tree-ssa.texi (gimple_phi_result): Document.
4022 (gimple_phi_num_args, gimple_phi_arg): Likewise.
4023 (gimple_phi_arg_edge, gimple_phi_arg_def): Likewise.
4024 (PHI_RESULT, PHI_NUM_ARGS): Remove.
4025 (PHI_ARG_ELT, PHI_ARG_EDGE, PHI_ARG_DEF): Likewise.
4027 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
4029 * expr.h: Revert change and include tree-core.h.
4030 * rtl.h: Revert change and don't include tree-core.h.
4032 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
4034 * config/darwin.c: Include gimple.h.
4035 * config/i386/winnt.c: Likewise.
4037 2013-10-29 Marc Glisse <marc.glisse@inria.fr>
4039 PR tree-optimization/19831
4040 * tree-ssa-alias.c (stmt_kills_ref_p_1): Handle BUILT_IN_FREE.
4042 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
4044 * tree-outof-ssa.h: Remove include files.
4045 * tree-outof-ssa.c: Add required include files from tree-outof-ssa.h.
4047 * tree-ssa-coalesce.c: Likewise.
4048 * cfgexpand.c: Likewise.
4049 * tree-ssa-ter.c: Likewise.
4050 * ipa-prop.h: Remove gimple.h and tree-core.h from include list.
4051 * lto-streamer.h: Likewise.
4052 * cgraphbuild.c: Add gimple.h to include list.
4053 * data-streamer-in.c: Likewise.
4054 * ipa-cp.c: Likewise.
4055 * tree-streamer.c: Likewise.
4056 * lto-compress.c: Likewise.
4057 * ipa-reference.c: Likewise.
4058 * data-streamer-out.c: Likewise.
4059 * lto-cgraph.c: Likewise.
4060 * cgraphclones.c: Likewise.
4061 * ipa-utils.c: Likewise.
4062 * data-streamer.c: Likewise.
4063 * ipa-split.c: Likewise.
4064 * lto-section-in.c: Likewise.
4065 * tree-streamer-out.c: Likewise.
4066 * ipa-prop.c: Likewise.
4067 * tree-streamer-in.c: Likewise.
4068 * symtab.c: Likewise.
4069 * opts-global.c: Likewise.
4070 * lto-opts.c: Likewise.
4071 * lto-section-out.c: Likewise.
4072 * lto-streamer.c: Likewise.
4073 * rtl.h: Add tree-core.h to include list.
4074 * expr.h: Remove tree-core.h from include list.
4075 * gimple.h: Likewise.
4076 * ipa-utils.h: Likewise.
4077 * streamer-hooks.h: Likewise.
4078 * streamer-hooks.c: Include input.h.
4080 2013-10-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4082 * config/arm/arm.c (cortexa7_extra_costs): New table.
4083 (arm_cortex_a7_tune): New.
4084 * config/arm/arm-cores.def: Use cortex_a7 tuning for cortex-a7.
4086 2013-10-29 Eric Botcazou <ebotcazou@adacore.com>
4088 * expr.c (expand_expr_real_1) <MEM_EXPR>: Eliminate small redundancy.
4090 2013-10-29 David Malcolm <dmalcolm@redhat.com>
4092 * doc/gty.texi ("Inheritance and GTY"): Make it clear that
4093 to use autogenerated markers for a class-hierarchy, every class
4094 must have a GTY marker.
4095 * gengtype.h (struct type): Add linked list of subclasses to
4096 the "s" member of the union.
4097 (add_subclass): New decl.
4098 * gengtype-state.c (read_state_struct_type): Set up subclass
4100 * gengtype.c (get_ultimate_base_class): New.
4101 (add_subclass): New.
4102 (new_structure): Set up subclass linked list.
4103 (set_gc_used_type): Propagate usage information to subclasses.
4104 (output_mangled_typename): Use get_ultimate_base_class.
4105 (walk_subclasses): Use the subclass linked list, avoiding an
4106 O(N^2) when writing out all types.
4107 (walk_type): Issue an error if the base class is missing a tag,
4108 rather than generating bogus C code. Add a gcc_unreachable
4109 default case, in case people omit tags from concrete subclasses,
4110 or get the values wrong.
4111 (write_func_for_structure): Issue an error for subclasses for
4112 which the base doesn't have a "desc", since otherwise the
4113 autogenerated routines for the base would silently fail to visit
4114 any subclass fields.
4115 (write_root): Use get_ultimate_base_class, tweaking constness of
4116 tp to match that function's signature.
4118 2013-10-29 David Malcolm <dmalcolm@redhat.com>
4120 * doc/gty.texi (GTY Options): Add note about inheritance to
4121 description of desc and tag.
4122 (Inheritance and GTY): New.
4124 2013-10-29 David Malcolm <dmalcolm@redhat.com>
4126 * gengtype-parse.c (opts_have): Drop "static" so that
4127 we can use this from gengtype.c.
4128 * gengtype.c (set_gc_used_type): Mark any base class as used;
4129 update field traversal to visit inherited fields.
4130 (output_mangled_typename): Convert references to classes within
4131 an inheritance hierarchy to reference the ultimate base class,
4132 since only it will have gt_ functions.
4133 (get_string_option): New.
4134 (walk_subclasses): New.
4135 (walk_type): Treat GTY structs that have a "desc" as being the
4136 root of an inheritance hierarchy. Generate a switch on it
4137 within the marking function which walks all subclasses, adding
4138 cases for them via walk_subclasses. For subclasses, visit all
4139 fields of the type (including inherited ones).
4140 (write_func_for_structure): Don't write fns for subclasses, only
4141 for the ultimate base class within an inheritance hierarchy.
4142 Subclasses-marking will be handled by the base class marking functions.
4143 (write_types): Likewise.
4144 (write_local_func_for_structure): Likewise.
4145 (USED_BY_TYPED_GC_P): Emit allocators for subclasses that have
4146 a "tag" option (and are thus concrete subclasses).
4147 (write_root): Use the marker function for the ultimate base class.
4148 * gengtype.h (FOR_ALL_INHERITED_FIELDS): New.
4149 (opts_have): Add declaration.
4151 2013-10-28 Vladimir Makarov <vmakarov@redhat.com>
4153 * lra-spills.c (lra_final_code_change): Remove useless move insns
4154 originated from moves of pseudos.
4156 2013-10-28 Jeff Law <law@redhat.com>
4158 * lower-subreg.c (resolve_simple_move): Fix comment typo.
4160 2013-10-28 Trevor Saunders <tsaunders@mozilla.com>
4162 * df-scan.c (df_collection_rec): Adjust.
4163 (copy_defs): New constant.
4164 (copy_uses): Likewise.
4165 (copy_eq_uses): Likewise.
4166 (copy_mw): Likewise.
4167 (copy_all): Likewise.
4168 (df_insn_rescan): Adjust.
4169 (df_notes_rescan): Likewise.
4170 (df_swap_refs): Likewise.
4171 (df_sort_and_compress_refs): Likewise.
4172 (df_sort_and_compress_mws): Likewise.
4173 (df_install_refs): Likewise.
4174 (df_install_mws): Likewise.
4175 (df_refs_add_to_chains): Add flags parameter controlling which vectors
4177 (df_bb_refs_record): Adjust.
4178 (df_record_entry_block_defs): Likewise.
4179 (df_record_exit_block_defs): Likewise.
4180 (df_refs_verify): Likewise.
4181 (df_mws_verify): Likewise.
4182 (df_insn_refs_verify): Likewise.
4183 (df_bb_verify): Likewise.
4184 * ipa-pure-const.c (finish_state): Remove.
4185 (propagate): Adjust.
4186 * tree-data-ref.c tree-ssa-alias.c tree-ssa-loop-ivcanon.c
4187 tree-ssa-threadedge.c tree-vect-loop-manip.c tree-vect-slp.c
4188 var-tracking.c: Adjust.
4189 * vec.c (stack_vecs): Remove.
4190 (register_stack_vec): Likewise.
4191 (stack_vec_register_index): Likewise.
4192 (unregister_stack_vec): Likewise.
4193 * vec.h (struct va_stack): Remove.
4194 (struct vec<T, A, vl_ptr>): Specialize as
4195 struct vec<T, va_heap, vl_ptr> instead since va_heap is the only
4196 allocation strategy compatable with the vl_ptr layout.
4197 (struct vec<T, va_gc, vl_ptr>): Remove because it now gets an empty
4198 specialization anyway.
4199 (class stack_vec): New class.
4200 (vec_stack_alloc): Remove.
4201 (vec<T, va_heap, vl_ptr>::using_auto_storage): New method.
4203 2013-10-28 Alexander Ivchenko <alexander.ivchenko@intel.com>
4204 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4205 Sergey Lega <sergey.s.lega@intel.com>
4206 Anna Tikhonova <anna.tikhonova@intel.com>
4207 Ilya Tocar <ilya.tocar@intel.com>
4208 Andrey Turetskiy <andrey.turetskiy@intel.com>
4209 Ilya Verbin <ilya.verbin@intel.com>
4210 Kirill Yukhin <kirill.yukhin@intel.com>
4211 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4213 * config/i386/i386.md (prefetch): Allow TARGET_AVX512PF.
4214 (*prefetch_avx512pf_<mode>): New.
4215 * config/i386/sse.md (avx512f_vmcmp<mode>3): Ditto.
4216 (avx512f_maskcmp<mode>3): Ditto.
4217 (vashrv16si3): Ditto.
4219 2013-10-28 Alexander Ivchenko <alexander.ivchenko@intel.com>
4220 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4221 Sergey Lega <sergey.s.lega@intel.com>
4222 Anna Tikhonova <anna.tikhonova@intel.com>
4223 Ilya Tocar <ilya.tocar@intel.com>
4224 Andrey Turetskiy <andrey.turetskiy@intel.com>
4225 Ilya Verbin <ilya.verbin@intel.com>
4226 Kirill Yukhin <kirill.yukhin@intel.com>
4227 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4229 * config/i386/i386.md (any_truncate): New.
4230 (trunsuffix): Ditto.
4231 * config/i386/predicates.md (const_8_to_9_operand): New.
4232 (const_10_to_11_operand): Ditto.
4233 (const_12_to_13_operand): Ditto.
4234 (const_14_to_15_operand): Ditto.
4235 (const_16_to_19_operand): Ditto.
4236 (const_20_to_23_operand): Ditto.
4237 (const_24_to_27_operand): Ditto.
4238 (const_28_to_31_operand): Ditto.
4239 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_FIX_NOTRUNC.
4240 (cvtusi2<ssescalarmodesuffix>32): New.
4241 (cvtusi2<ssescalarmodesuffix>64): Ditto.
4242 (ufloatv16siv16sf2): Ditto.
4243 (avx512f_fix_notruncv16sfv16si): Ditto.
4244 (avx512f_ufix_notruncv16sfv16si): Ditto.
4245 (avx512f_vcvtss2usi): Ditto.
4246 (avx512f_vcvtss2usiq): Ditto.
4247 (avx512f_vcvttss2usi): Ditto.
4248 (avx512f_vcvttss2usiq): Ditto.
4249 (avx512f_vcvtsd2usi): Ditto.
4250 (avx512f_vcvtsd2usiq): Ditto.
4251 (avx512f_vcvttsd2usi): Ditto.
4252 (avx512f_vcvttsd2usiq): Ditto.
4253 (ufloatv8siv8df): Ditto.
4254 (avx512f_cvtdq2pd512_2): Ditto.
4255 (avx512f_cvtpd2dq512): Ditto.
4256 (avx512f_ufix_notruncv8dfv8si): Ditto.
4257 (avx512f_cvtpd2ps512): Ditto.
4258 (vec_unpacks_lo_v16sf): Ditto.
4259 (vec_unpacks_hi_v16sf): Ditto.
4260 (vec_unpacks_float_hi_v16si): Ditto.
4261 (vec_unpacks_float_lo_v16si): Ditto.
4262 (avx512f_unpckhps512): Ditto.
4263 (avx512f_unpcklps512): Ditto.
4264 (avx512f_movshdup512): Ditto.
4265 (avx512f_movsldup512): Ditto.
4266 (vec_extract_lo_v32hi): Ditto.
4267 (vec_extract_hi_v32hi): Ditto.
4268 (vec_extract_lo_v64qi): Ditto.
4269 (vec_extract_hi_v64qi): Ditto.
4270 (avx512f_unpckhpd512): Ditto.
4271 (avx512f_movddup512): Ditto.
4272 (avx512f_unpcklpd512): Ditto.
4273 (*avx512f_unpcklpd512): Ditto.
4274 (avx512f_shufps512_1): Ditto.
4275 (avx512f_shufpd512_1): Ditto.
4276 (avx512f_interleave_highv8di): Ditto.
4277 (avx512f_interleave_lowv8di): Ditto.
4278 (PMOV_DST_MODE): Ditto.
4279 (pmov_src_mode): Ditto.
4280 (pmov_src_lower): Ditto.
4282 (*avx512f_<code><pmov_src_lower><mode>2): Ditto.
4283 (*avx512f_<code>v8div16qi2): Ditto.
4284 (*avx512f_<code>v8div16qi2_store): Ditto.
4285 (vec_widen_umult_even_v16si): Ditto.
4286 (*vec_widen_umult_even_v16si): Ditto.
4287 (vec_widen_smult_even_v16si): Ditto.
4288 (*vec_widen_smult_even_v16si): Ditto.
4289 (avx512f_interleave_highv16si): Ditto.
4290 (avx512f_interleave_lowv16si): Ditto.
4291 (avx512f_<code>v16qiv16si2): Ditto.
4292 (avx512f_<code>v16hiv16si2): Ditto.
4293 (avx512f_<code>v8qiv8di2): Ditto.
4294 (avx512f_<code>v8hiv8di2): Ditto.
4295 (avx512f_<code>v8siv8di2): Ditto.
4296 (avx512cd_maskb_vec_dupv8di): Ditto.
4297 (avx512cd_maskw_vec_dupv16si): Ditto.
4298 (avx512f_vcvtph2ps512): Ditto.
4299 (avx512f_vcvtps2ph512): Ditto.
4300 (VEC_EXTRACT_MODE): Extened with wider modes.
4301 (VEC_PERM_AVX2): Ditto.
4302 (VEC_PERM_CONST): Ditto.
4304 2013-10-28 Joern Rennecke <joern.rennecke@embecosm.com>
4306 * config/arc/arc.c (arc_ccfsm_post_advance):
4307 Add comment about TYPE_RETURN.
4309 2013-10-28 Bin Cheng <bin.cheng@arm.com>
4311 * tree-ssa-loop-ivopts.c (strip_offset_1): Change parameter type.
4312 Count DECL_FIELD_BIT_OFFSET for COMPONENT_REF.
4313 (strip_offset): Convert offset to unsigned number.
4315 2013-10-27 Tom de Vries <tom@codesourcery.com>
4317 * cfgexpand.c (gimple_expand_cfg): Remove test for parm_birth_insn.
4318 Don't commit insertions after NOTE_INSN_FUNCTION_BEG.
4320 2013-10-27 Oleg Endo <olegendo@gcc.gnu.org>
4322 * config/sh/sh.c (MSW, LSW): Move and rename macros to...
4323 * config/sh/sh.h (SH_REG_MSW_OFFSET, SH_REG_LSW_OFFSET): ... here.
4324 (TARGET_BIG_ENDIAN): New macro.
4325 * config/sh/sh.md: Use it instead of !TARGET_LITTLE_ENDIAN.
4326 Use SH_REG_MSW_OFFSET and SH_REG_LSW_OFFSET.
4327 * config/sh/sh.c: Likewise.
4328 * config/sh/sh.h: Likewise.
4330 2013-10-27 Hans-Peter Nilsson <hp@axis.com>
4332 * config/cris/cris.c (cris_emit_trap_for_misalignment): Replace the
4333 removed PRED_MUDFLAP with PRED_NORETURN. Correct file-path in comment.
4335 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
4337 * config/sh/sh.md (movmemsi): Remove empty constraints and predicates.
4339 (cmpstr_t, cmpstrsi): Fix formatting.
4341 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
4344 * config/sh/predicates.md (general_movdst_operand): Allow reg+reg
4345 addressing, do not use general_operand for memory operands.
4347 2013-10-26 Vladimir Makarov <vmakarov@redhat.com>
4350 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
4351 * lra-spills.c (lra_final_code_change): Remove useless move insns.
4353 2013-10-26 Jeff Law <law@redhat.com>
4355 * predict.c (PRED_MUDFLAP): Remove.
4356 * targhooks.c (build_va_arg_indirect_ref): Remove mudflap support.
4358 * Makefile.in (C_COMMON_OBJS): Remove tree-mudflap.
4359 (OBJS): Remove tree-nomudflap.o
4360 (GTFILES): Remove tree-mudflap.c
4361 * builtins.c (expand_builtin_alloc): Remove mudflap support.
4362 * gcc.c (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
4363 (mfwrap_spec, mflib_spec): Likewise.
4364 (cpp_unique_options, cc1_options, static_specs): Likewise.
4365 * gimplify (gimplify_vla_decl, build_va_arg_indirect_ref): Likewise.
4366 * passes.def: Likewise.
4367 * toplev.c (compile_file, process_options): Likewise.
4368 * tree-inline.c (copy_tree_r): Likewise.
4369 * tree-pass.,h (make_pass_mudflap_1, make_pass_mudflap_2): Likewise.
4370 * varasm.c (make_decl_rtl, make_decl_rtl_for_debug): Likewise.
4371 (build_constant_desc, output_constant_def_contents): Likewise.
4372 (categorize_decl_for_section): Likewise.
4373 * tree-mudflap.c: Removed.
4374 * tree-mudflap.h: Removed.
4375 * tree-nomudflap.c: Removed.
4376 * bfin/uclinux.h (MFWRAP_SPEC): Remove.
4377 * moxie/uclinux.h (MFWRAP_SPEC): Likewise.
4378 * rs6000/aix.h (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
4379 * config/sol2.h (MFLIB_SPEC): Likewise.
4380 * doc/install.texi: Remove mudflap references.
4381 * doc/passes.texi: Similarly.
4382 * doc/sourcebuild.texi: Similarly.
4383 * doc/invoke.texi: Remove mudlfap related options.
4385 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
4387 PR rtl-optimization/58759
4388 * lra-constraints.c (lra_constraints): Remove wrong condition to
4389 remove insn setting up an equivalent pseudo.
4391 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
4393 * config/rs6000/rs6000-protos.h
4394 (rs6000_secondary_memory_needed_mode): New prototype.
4395 * config/rs6000/rs6000.c: Include ira.h.
4396 (TARGET_LRA_P): Redefine.
4397 (rs6000_legitimate_offset_address_p): Call
4398 legitimate_constant_pool_address_p in strict mode for LRA.
4399 (rs6000_legitimate_address_p): Ditto.
4400 (legitimate_lo_sum_address_p): Add code for LRA. Use lra_in_progress.
4401 (rs6000_emit_move): Add LRA version of code to generate load/store
4403 (rs6000_secondary_memory_needed_mode): New.
4404 (rs6000_alloc_sdmode_stack_slot): Do nothing for LRA.
4405 (rs6000_secondary_reload_class): Return NO_REGS for LRA for
4406 constants, memory, and FP registers.
4407 (rs6000_lra_p): New.
4408 * config/rs6000/rs6000.h (SECONDARY_MEMORY_NEEDED_MODE): New macro.
4409 * config/rs6000/rs6000.opt (mlra): New option.
4410 * lra-spills.c (lra_final_code_change): Remove useless move insns.
4412 2013-10-25 Yufeng Zhang <yufeng.zhang@arm.com>
4414 * tree-ssa-math-opts.c (convert_plusminus_to_widen): Call
4415 has_single_use () and not do the conversion if has_single_use ()
4416 returns false for the multiplication result.
4418 2013-10-25 David Malcolm <dmalcolm@redhat.com>
4420 * tree.h (EXCEPTIONAL_CLASS_P): Rename parameter from "CODE"
4421 to "NODE", since this works on a "tree", not an
4423 (CONSTANT_CLASS_P): Likewise.
4426 (INDIRECT_REF_P): Likewise.
4427 (REFERENCE_CLASS_P): Likewise.
4428 (COMPARISON_CLASS_P): Likewise.
4429 (UNARY_CLASS_P): Likewise.
4430 (BINARY_CLASS_P): Likewise.
4431 (STATEMENT_CLASS_P): Likewise.
4432 (VL_EXP_CLASS_P): Likewise.
4433 (EXPRESSION_CLASS_P): Likewise.
4434 (IS_TYPE_OR_DECL_P): Likewise.
4436 2013-10-25 Marc Glisse <marc.glisse@inria.fr>
4438 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Look for an
4439 ADDR_EXPR in the defining statement.
4441 2013-10-25 Richard Biener <rguenther@suse.de>
4443 PR tree-optimization/58626
4444 * tree-loop-distribution.c (enum rdg_dep_type): Remove
4445 anti_dd, output_dd and input_dd.
4446 (struct rdg_edge): Remove level and relation members.
4447 (RDGE_LEVEL, RDGE_RELATION): Remove.
4448 (dot_rdg_1): Adjust.
4449 (create_rdg_edge_for_ddr): Remove.
4450 (create_rdg_edges_for_scalar): Adjust.
4451 (create_edge_for_control_dependence): Likewise.
4452 (create_rdg_edges): Split into ...
4453 (create_rdg_flow_edges): ... this
4454 (create_rdg_cd_edges): ... and this.
4456 (build_rdg): Likewise, do not compute data dependences or
4458 (pg_add_dependence_edges): New function.
4460 (distribute_loop): First apply all non-dependence based
4461 partition mergings. Then compute dependences between partitions
4462 and merge and order partitions according to them.
4464 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
4466 PR rtl-optimization/58831
4467 * alias.c (init_alias_analysis): At the beginning of each iteration,
4468 set the reg_seen[N] bit if static_reg_base_value[N] is non-null.
4470 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
4472 * recog.c (search_ofs): New static variable moved from...
4473 (peep2_find_free_register): ...here.
4474 (peephole2_optimize): Initialize it.
4476 2013-10-25 Tobias Burnus <burnus@net-b.de>
4478 * doc/invoke.texi (fopenmp): Change supported OpenMP version to 4.0.
4480 2013-10-25 Uros Bizjak <ubizjak@gmail.com>
4482 * config/i386/i386.h (TARGET_MPX): New define.
4483 (TARGET_MPX_P): Ditto.
4485 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
4487 * config/i386/constraints.md (B): New.
4490 * config/i386/i386-c.c (ix86_target_macros_internal): Add __MPX__.
4491 * config/i386/i386-modes.def (BND32): New.
4493 * config/i386/i386-protos.h (ix86_bnd_prefixed_insn_p): New.
4494 * config/i386/i386.c (isa_opts): Add mmpx.
4495 (regclass_map): Add bound registers.
4496 (dbx_register_map): Likewise.
4497 (dbx64_register_map): Likewise.
4498 (svr4_dbx_register_map): Likewise.
4500 (ix86_option_override_internal): Support MPX ISA.
4501 (ix86_conditional_register_usage): Support bound registers.
4502 (print_reg): Likewise.
4503 (ix86_code_end): Add MPX bnd prefix.
4504 (output_set_got): Likewise.
4505 (ix86_output_call_insn): Likewise.
4506 (ix86_print_operand): Add '!' (MPX bnd) print prefix support.
4507 (ix86_print_operand_punct_valid_p): Likewise.
4508 (ix86_print_operand_address): Support UNSPEC_BNDMK_ADDR and
4510 (ix86_class_likely_spilled_p): Add bound regs support.
4511 (ix86_hard_regno_mode_ok): Likewise.
4512 (x86_order_regs_for_local_alloc): Likewise.
4513 (ix86_bnd_prefixed_insn_p): New.
4514 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Fix to new value.
4515 (FIXED_REGISTERS): Add bound registers.
4516 (CALL_USED_REGISTERS): Likewise.
4517 (REG_ALLOC_ORDER): Likewise.
4518 (HARD_REGNO_NREGS): Likewise.
4520 (VALID_BND_REG_MODE): New.
4521 (FIRST_BND_REG): New.
4522 (LAST_BND_REG): New.
4523 (reg_class): Add BND_REGS.
4524 (REG_CLASS_NAMES): Likewise.
4525 (REG_CLASS_CONTENTS): Likewise.
4527 (ANY_BND_REG_P): New.
4529 (HI_REGISTER_NAMES): Add bound registers.
4530 * config/i386/i386.md (UNSPEC_BNDMK): New.
4531 (UNSPEC_BNDMK_ADDR): New.
4532 (UNSPEC_BNDSTX): New.
4533 (UNSPEC_BNDLDX): New.
4534 (UNSPEC_BNDLDX_ADDR): New.
4535 (UNSPEC_BNDCL): New.
4536 (UNSPEC_BNDCU): New.
4537 (UNSPEC_BNDCN): New.
4538 (UNSPEC_MPX_FENCE): New.
4541 (type): Add mpxmov, mpxmk, mpxchk, mpxld, mpxst.
4542 (length_immediate): Likewise.
4543 (prefix_0f): Likewise.
4545 (prefix_rep): Check for bnd prefix.
4546 (length_nobnd): New.
4547 (length): Use length_nobnd if specified.
4552 (*jcc_1): Add bnd prefix and rename length attr to length_nobnd.
4555 (simple_return_internal): Likewise.
4556 (simple_return_pop_internal): Likewise.
4557 (*indirect_jump): Add MPX bnd prefix.
4558 (*tablejump_1): Likewise.
4559 (simple_return_internal_long): Likewise.
4560 (simple_return_indirect_internal): Likewise.
4564 (*mov<mode>_internal_mpx): New.
4565 (<mode>_<bndcheck>): New.
4566 (*<mode>_<bndcheck>): New.
4571 * config/i386/predicates.md (lea_address_operand): Rename to...
4572 (address_no_seg_operand): ... this.
4573 (address_mpx_no_base_operand): New.
4574 (address_mpx_no_index_operand): New.
4575 (bnd_mem_operator): New.
4576 * config/i386/i386.opt (mmpx): New.
4577 * doc/invoke.texi: Add documentation for the flags -mmpx, -mno-mpx.
4578 * doc/rtl.texi Add documentation for BND32mode and BND64mode.
4580 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
4582 * mode-classes.def (MODE_POINTER_BOUNDS): New.
4583 * tree.def (POINTER_BOUNDS_TYPE): New.
4584 * genmodes.c (complete_mode): Support MODE_POINTER_BOUNDS.
4585 (POINTER_BOUNDS_MODE): New.
4586 (make_pointer_bounds_mode): New.
4587 * machmode.h (POINTER_BOUNDS_MODE_P): New.
4588 * stor-layout.c (int_mode_for_mode): Support MODE_POINTER_BOUNDS.
4589 (layout_type): Support POINTER_BOUNDS_TYPE.
4590 * tree-pretty-print.c (dump_generic_node): Support POINTER_BOUNDS_TYPE.
4591 * tree.c (build_int_cst_wide): Support POINTER_BOUNDS_TYPE.
4592 (type_contains_placeholder_1): Likewise.
4593 * tree.h (POINTER_BOUNDS_TYPE_P): New.
4594 * varasm.c (output_constant): Support POINTER_BOUNDS_TYPE.
4595 * doc/rtl.texi (MODE_POINTER_BOUNDS): New.
4597 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
4599 * expr.c (expand_expr_real_1): Use mode of memory reference rather than
4600 mode of address computation when calling memory_address_addr_space.
4602 2013-08-24 Richard Henderson <rth@twiddle.net>
4605 * optabs.c (maybe_emit_atomic_exchange): Use create_input_operand
4606 instead of create_convert_operand_to.
4607 (maybe_emit_sync_lock_test_and_set): Likewise.
4608 (expand_atomic_compare_and_swap): Likewise.
4609 (maybe_emit_compare_and_swap_exchange_loop): Don't convert_modes.
4611 2013-08-24 Sriraman Tallam <tmsriram@google.com>
4613 * cgraph.c (cgraph_fnver_htab): Move GTY((...)) to be before htab_t.
4614 Change param_is to use the struct and not the pointer to the struct.
4616 2013-10-24 Andrew MacLeod <amacleod@redhat.com>
4618 * builtins.c (dummy_object, gimplify_va_arg_expr): Move to gimplify.c.
4619 * gimplify.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
4620 Move to targhooks.c.
4621 (dummy_object, gimplify_va_arg_expr): Relocate from builtins.c.
4622 * targhooks.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
4623 Relocate from gimplify.c.
4624 * targhooks.h: Add 2 prototypes.
4625 * tree.h: Delete 2 prototypes.
4627 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
4629 * tree-ssa-loop-ivopts.c (multiplier_allowed_in_address_p ): Check both
4630 [reg+mult*reg] and [mult*reg] to determine if multiplier is allowed.
4632 2013-10-24 Cong Hou <congh@google.com>
4634 * convert.c (convert_to_real): Guard those unsafe math function
4635 convertions with flag_unsafe_math_optimizations. Handle sqrt()
4638 2013-10-24 Markus Trippelsdorf <markus@trippelsdorf.de>
4641 * cgraph.c (cgraph_create_edge_1): Add indirect_unknown_callee
4643 (cgraph_create_edge): Use the new argument.
4644 (cgraph_create_indirect_edge): Likewise.
4646 2013-10-24 Joern Rennecke <joern.rennecke@embecosm.com>
4648 * config/arc/arc.c (arc_ccfsm_post_advance): Also handle
4650 (arc_ifcvt) <case 1 and 2>: Check that arc_ccfsm_post_advance
4651 changes statep->state.
4653 2013-10-24 Tobias Burnus <burnus@net-b.de>
4656 * tree-cfg.c (replace_loop_annotate): New function.
4657 (execute_build_cfg): Call it.
4658 * gimplify.c (gimple_boolify, gimplify_expr): Handle ANNOTATE_EXPR.
4659 * internal-fn.c (expand_ANNOTATE): New function.
4660 * internal-fn.def (ANNOTATE): Define as new internal function.
4661 * tree-core.h (tree_node_kind): Add annot_expr_ivdep_kind.
4662 * tree-pretty-print.c (dump_generic_node): Handle ANNOTATE_EXPR.
4663 * tree.def (ANNOTATE_EXPR): New DEFTREECODE.
4664 * doc/extend.texi (Pragmas): Document #pragma ivdep.
4665 * doc/generic.texi (Expressions): Document ANNOTATE_EXPR.
4667 2013-10-17 Ian Bolton <ian.bolton@arm.com>
4668 Marcus Shawcroft <marcus.shawcroft@arm.com>
4670 * config/aarch64/aarch64.c (aarch64_preferred_reload_class):
4671 Special case reload SP+C into none GENERAL_REGS.
4673 2013-10-24 Michael Matz <matz@suse.de>
4675 * gengtype.c (is_file_equal): Check that files will be same length.
4677 2013-10-25 Christian Bruel <christian.bruel@st.com>
4679 * config.gcc (sh-*): Add sh-mem.o to extra_obj.
4680 * config/sh/t-sh (sh-mem.o): New rule.
4681 * config/sh/sh-mem.cc (expand_block_move): Moved here.
4682 (sh_expand_cmpstr): New function.
4683 * config/sh/sh.c (force_into, expand_block_move): Move to sh-mem.c.
4684 * config/sh/sh-protos.h (sh_expand_cmpstr): Declare.
4685 * config/sh/sh.md (cmpstrsi, cmpstr_t): New patterns.
4686 (rotlhi3_8): Rename.
4688 2013-10-24 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4690 * configure.ac (ZW_PROG_COMPILER_DEPENDENCIES): Use CXX instead of CC.
4691 * Makefile.in (CXXDEPMODE): Assign and change users.
4692 (CCDEPMODE): Delete.
4693 * configure: Regenerate.
4695 2013-10-23 David Malcolm <dmalcolm@redhat.com>
4697 * gengtype-parse.c (require_without_advance): New.
4698 (type): For GTY-marked types that are not GTY((user)), parse any
4699 base classes, requiring them to be single-inheritance, and not
4700 be templates. For non-GTY-marked types and GTY((user)),
4701 continue to skip over any C++ inheritance specification.
4702 * gengtype-state.c (state_writer::write_state_struct_type):
4703 Write base class of type (if any).
4704 (read_state_struct_type): Read base class of type (if any).
4705 * gengtype.c (new_structure): Add a "base_class" parameter.
4706 (create_optional_field_): Update for new parameter to new_structure.
4707 (adjust_field_rtx_def): Likewise.
4708 (adjust_field_tree_exp): Likewise.
4709 * gengtype.h (struct type): Add "base_class" field to the s
4711 (new_structure): Add "base" parameter.
4713 2013-10-23 Sriraman Tallam <tmsriram@google.com>
4716 * config/i386/i386.c (ix86_option_override_internal):
4717 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...)
4718 (ix86_valid_target_attribute_tree):
4719 Change TARGET_64BIT to TARGET_64BIT_P (opts->...)
4720 Change TARGET_SSE to TARGET_SSE_P (opts->...)
4722 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
4724 * tree-ssa-loop.h: Remove include files.
4725 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
4726 * cfgloopmanip.c: Move required includes from tree-ssa-loop.h.
4727 * graphite-clast-to-gimple.c: Likewise.
4728 * graphite-scop-detection.c: Likewise.
4729 * graphite-sese-to-poly.c: Likewise.
4730 * ipa-inline-analysis.c: Likewise.
4731 * ipa-pure-const.c: Likewise.
4732 * loop-init.c: Likewise.
4733 * passes.c: Likewise.
4734 * predict.c: Likewise.
4735 * tree-cfg.c: Likewise.
4736 * tree-cfgcleanup.c: Likewise.
4737 * tree-chrec.c: Likewise.
4738 * tree-data-ref.c: Likewise.
4739 * tree-loop-distribution.c: Likewise.
4740 * tree-parloops.c: Likewise.
4741 * tree-predcom.c: Likewise.
4742 * tree-scalar-evolution.c: Likewise.
4743 * tree-ssa-address.c: Likewise.
4744 * tree-ssa.c: Likewise.
4745 * tree-ssa-dce.c: Likewise.
4746 * tree-ssa-loop.c: Likewise.
4747 * tree-ssa-loop-im.c: Likewise.
4748 * tree-ssa-loop-ivcanon.c: Likewise.
4749 * tree-ssa-loop-ivopts.c: Likewise.
4750 * tree-ssa-loop-manip.c: Likewise.
4751 * tree-ssa-loop-niter.c: Likewise.
4752 * tree-ssa-loop-prefetch.c: Likewise.
4753 * tree-ssa-loop-unswitch.c: Likewise.
4754 * tree-ssa-reassoc.c: Likewise.
4755 * tree-vect-data-refs.c: Likewise.
4756 * tree-vect-loop.c: Likewise.
4757 * tree-vect-loop-manip.c: Likewise.
4758 * tree-vectorizer.c: Likewise.
4759 * tree-vect-stmts.c: Likewise.
4760 * tree-vrp.c: Likewise.
4762 2013-10-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4764 * config/rs6000/altivec.md (mulv8hi3): Adjust for little endian.
4766 2013-10-23 Jakub Jelinek <jakub@redhat.com>
4768 PR tree-optimization/58775
4769 PR tree-optimization/58791
4770 * tree-ssa-reassoc.c (reassoc_stmt_dominates_stmt_p): New function.
4771 (insert_stmt_after): Rewritten, don't move the stmt, but really
4773 (get_stmt_uid_with_default): Remove.
4774 (build_and_add_sum): Use insert_stmt_after and
4775 reassoc_stmt_dominates_stmt_p. Fix up uid if bb contains only labels.
4776 (update_range_test): Set uid on stmts added by
4777 force_gimple_operand_gsi. Don't immediately modify statements
4778 in inter-bb optimization, just update oe->op values.
4779 (optimize_range_tests): Return bool whether any changed have been made.
4780 (update_ops): New function.
4781 (struct inter_bb_range_test_entry): New type.
4782 (maybe_optimize_range_tests): Perform statement changes here.
4783 (not_dominated_by, appears_later_in_bb, get_def_stmt,
4784 ensure_ops_are_available): Remove.
4785 (find_insert_point): Rewritten.
4786 (rewrite_expr_tree): Remove MOVED argument, add CHANGED argument,
4787 return LHS of the (new resp. old) stmt. Don't call
4788 ensure_ops_are_available, don't reuse SSA_NAMEs, recurse first
4789 instead of last, move new stmt at the right place.
4790 (linearize_expr, repropagate_negates): Don't reuse SSA_NAMEs.
4791 (negate_value): Likewise. Set uids.
4792 (break_up_subtract_bb): Initialize uids.
4793 (reassociate_bb): Adjust rewrite_expr_tree caller.
4794 (do_reassoc): Don't call renumber_gimple_stmt_uids.
4796 2013-10-23 David Edelsohn <dje.gcc@gmail.com>
4799 * config/rs6000/rs6000.md (mulsi3_internal1 and splitter): Add
4800 TARGET_32BIT final condition.
4801 (mulsi3_internal2 and splitter): Same.
4803 2013-10-23 Jeff Law <law@redhat.com>
4805 * tree-ssa-threadedge.c (thread_across_edge): Do not allow threading
4806 through joiner blocks with abnormal outgoing edges.
4808 * tree-ssa-threadupdate.c (thread_block_1): Renamed from thread_block.
4809 Add parameter JOINERS, to allow/disallow threading through joiner
4811 (thread_block): New. Call thread_block_1.
4812 (mark_threaded_blocks): Remove code to filter out certain cases
4813 of threading through joiner blocks.
4814 (thread_through_all_blocks): Document how we can have a dangling
4815 edge AUX field and clear it.
4817 2013-10-23 Ian Lance Taylor <iant@google.com>
4819 * doc/invoke.texi (Option Summary): Remove -fno-default-inline.
4820 (C++ Dialect Options): Likewise.
4821 (Optimize Options): Likewise.
4823 2013-10-23 Tom de Vries <tom@codesourcery.com>
4825 PR tree-optimization/58805
4826 * tree-ssa-tail-merge.c (stmt_local_def): Add gimple_vdef check.
4828 2013-10-23 Jakub Jelinek <jakub@redhat.com>
4830 * tree-vect-patterns.c (vect_recog_divmod_pattern): Optimize
4831 sequence based on get_range_info returned range.
4833 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
4835 * tree-ssa.h: Remove all #include's
4836 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
4837 * alias.c: Move required includes from tree-ssa.h.
4839 * builtins.c: Likewise.
4840 * calls.c: Likewise.
4841 * cfgexpand.c: Likewise.
4842 * cfghooks.c: Likewise.
4843 * cfgloop.c: Likewise.
4844 * cfgloopmanip.c: Likewise.
4845 * cgraph.c: Likewise.
4846 * cgraphbuild.c: Likewise.
4847 * cgraphclones.c: Likewise.
4848 * cgraphunit.c: Likewise.
4850 * except.c: Likewise.
4852 * final.c: Likewise.
4853 * fold-const.c: Likewise.
4854 * ggc-page.c: Likewise.
4855 * gimple-builder.c: Likewise.
4856 * gimple-fold.c: Likewise.
4857 * gimple-iterator.c: Likewise.
4858 * gimple-low.c: Likewise.
4859 * gimple-pretty-print.c: Likewise.
4860 * gimple-ssa-strength-reduction.c: Likewise.
4861 * gimple-streamer-in.c: Likewise.
4862 * gimple-streamer-out.c: Likewise.
4863 * gimplify.c: Likewise.
4864 * graphite-blocking.c: Likewise.
4865 * graphite-clast-to-gimple.c: Likewise.
4866 * graphite-dependences.c: Likewise.
4867 * graphite-interchange.c: Likewise.
4868 * graphite-optimize-isl.c: Likewise.
4869 * graphite-poly.c: Likewise.
4870 * graphite-scop-detection.c: Likewise.
4871 * graphite-sese-to-poly.c: Likewise.
4872 * graphite.c: Likewise.
4873 * ipa-cp.c: Likewise.
4874 * ipa-inline-analysis.c: Likewise.
4875 * ipa-inline-transform.c: Likewise.
4876 * ipa-inline.c: Likewise.
4877 * ipa-prop.c: Likewise.
4878 * ipa-pure-const.c: Likewise.
4879 * ipa-reference.c: Likewise.
4880 * ipa-split.c: Likewise.
4881 * ipa-utils.c: Likewise.
4882 * loop-init.c: Likewise.
4883 * lto-cgraph.c: Likewise.
4884 * lto-section-in.c: Likewise.
4885 * lto-section-out.c: Likewise.
4886 * lto-streamer-in.c: Likewise.
4887 * lto-streamer-out.c: Likewise.
4888 * lto-streamer.c: Likewise.
4889 * omp-low.c: Likewise.
4890 * passes.c: Likewise.
4891 * predict.c: Likewise.
4892 * print-tree.c: Likewise.
4893 * profile.c: Likewise.
4895 * targhooks.c: Likewise.
4896 * tracer.c: Likewise.
4897 * trans-mem.c: Likewise.
4898 * tree-call-cdce.c: Likewise.
4899 * tree-cfg.c: Likewise.
4900 * tree-cfgcleanup.c: Likewise.
4901 * tree-chrec.c: Likewise.
4902 * tree-complex.c: Likewise.
4903 * tree-data-ref.c: Likewise.
4904 * tree-dfa.c: Likewise.
4905 * tree-eh.c: Likewise.
4906 * tree-emutls.c: Likewise.
4907 * tree-if-conv.c: Likewise.
4908 * tree-inline.c: Likewise.
4909 * tree-into-ssa.c: Likewise.
4910 * tree-loop-distribution.c: Likewise.
4911 * tree-mudflap.c: Likewise.
4912 * tree-nested.c: Likewise.
4913 * tree-nrv.c: Likewise.
4914 * tree-object-size.c: Likewise.
4915 * tree-outof-ssa.c: Likewise.
4916 * tree-parloops.c: Likewise.
4917 * tree-phinodes.c: Likewise.
4918 * tree-predcom.c: Likewise.
4919 * tree-pretty-print.c: Likewise.
4920 * tree-profile.c: Likewise.
4921 * tree-scalar-evolution.c: Likewise.
4922 * tree-sra.c: Likewise.
4923 * tree-ssa-address.c: Likewise.
4924 * tree-ssa-alias.c: Likewise.
4925 * tree-ssa-ccp.c: Likewise.
4926 * tree-ssa-coalesce.c: Likewise.
4927 * tree-ssa-copy.c: Likewise.
4928 * tree-ssa-copyrename.c: Likewise.
4929 * tree-ssa-dce.c: Likewise.
4930 * tree-ssa-dom.c: Likewise.
4931 * tree-ssa-dse.c: Likewise.
4932 * tree-ssa-forwprop.c: Likewise.
4933 * tree-ssa-ifcombine.c: Likewise.
4934 * tree-ssa-live.c: Likewise.
4935 * tree-ssa-loop-ch.c: Likewise.
4936 * tree-ssa-loop-im.c: Likewise.
4937 * tree-ssa-loop-ivcanon.c: Likewise.
4938 * tree-ssa-loop-ivopts.c: Likewise.
4939 * tree-ssa-loop-manip.c: Likewise.
4940 * tree-ssa-loop-niter.c: Likewise.
4941 * tree-ssa-loop-prefetch.c: Likewise.
4942 * tree-ssa-loop-unswitch.c: Likewise.
4943 * tree-ssa-loop.c: Likewise.
4944 * tree-ssa-math-opts.c: Likewise.
4945 * tree-ssa-operands.c: Likewise.
4946 * tree-ssa-phiopt.c: Likewise.
4947 * tree-ssa-phiprop.c: Likewise.
4948 * tree-ssa-pre.c: Likewise.
4949 * tree-ssa-propagate.c: Likewise.
4950 * tree-ssa-reassoc.c: Likewise.
4951 * tree-ssa-sccvn.c: Likewise.
4952 * tree-ssa-sink.c: Likewise.
4953 * tree-ssa-strlen.c: Likewise.
4954 * tree-ssa-structalias.c: Likewise.
4955 * tree-ssa-tail-merge.c: Likewise.
4956 * tree-ssa-ter.c: Likewise.
4957 * tree-ssa-threadedge.c: Likewise.
4958 * tree-ssa-threadupdate.c: Likewise.
4959 * tree-ssa-uncprop.c: Likewise.
4960 * tree-ssa-uninit.c: Likewise.
4961 * tree-ssa.c: Likewise.
4962 * tree-ssanames.c: Likewise.
4963 * tree-stdarg.c: Likewise.
4964 * tree-streamer-in.c: Likewise.
4965 * tree-switch-conversion.c: Likewise.
4966 * tree-tailcall.c: Likewise.
4967 * tree-vect-data-refs.c: Likewise.
4968 * tree-vect-generic.c: Likewise.
4969 * tree-vect-loop-manip.c: Likewise.
4970 * tree-vect-loop.c: Likewise.
4971 * tree-vect-patterns.c: Likewise.
4972 * tree-vect-slp.c: Likewise.
4973 * tree-vect-stmts.c: Likewise.
4974 * tree-vectorizer.c: Likewise.
4975 * tree-vrp.c: Likewise.
4978 * value-prof.c: Likewise.
4979 * var-tracking.c: Likewise.
4980 * varpool.c: Likewise.
4981 * vtable-verify.c: Likewise.
4983 2013-10-23 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4985 * config/tilegx/tilegx.c: Include "tree.h".
4987 2013-10-23 Jakub Jelinek <jakub@redhat.com>
4989 * gimple-pretty-print.c (dump_ssaname_info): Always print "# " before
4990 the info, not after it.
4991 (gump_gimple_phi): Add COMMENT argument, if true, print "# " after
4992 dump_ssaname_info call.
4993 (pp_gimple_stmt_1): Adjust caller.
4994 (dump_phi_nodes): Likewise. Don't print "# " here.
4996 2013-10-22 Jan Hubicka <jh@suse.cz>
4998 * i386.h (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES_EPILOGUES): New
5000 * x86-tune.def (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES): Define it.
5001 * i386.c (expand_small_movmem_or_setmem): New function.
5002 (expand_set_or_movmem_prologue_epilogue_by_misaligned_moves): New
5004 (alg_usable_p): Add support for value ranges; cleanup.
5005 (ix86_expand_set_or_movmem): Add support for misaligned moves.
5007 2013-10-22 Sterling Augustine <saugustine@google.com>
5009 * doc/invoke.texi: Document -ggnu-pubnames.
5010 * common.opt: Add new option -ggnu-pubnames and modify -gpubnames
5012 * dwarf2out.c: Include gdb/gdb-index.h.
5013 (DEBUG_PUBNAMES_SECTION, DEBUG_PUBTYPES_SECTION): Handle
5014 debug_generate_pub_sections.
5015 (is_java, output_pubtables, output_pubname): New functions.
5016 (include_pubname_in_output): Handle debug_generate_pub_sections at
5018 (size_of_pubnames): Use new local space_for_flags based on
5019 debug_generate_pub_sections.
5020 (output_pubnames): Unify pubnames and pubtypes output logic.
5021 Genericize comments. Call output_pubname.
5022 (dwarf2out_finish): Move logic to output_pubnames and call it.
5024 2013-10-22 Uros Bizjak <ubizjak@gmail.com>
5027 * config/i386/i386.c (put_condition_code) <case GTU, case LEU>:
5028 Remove CCCmode handling.
5029 <case LTU>: Return 'c' suffix for CCCmode.
5030 <case GEU>: Return 'nc' suffix for CCCmode.
5031 (ix86_cc_mode) <case GTU, case LEU>: Do not generate overflow checks.
5032 * config/i386/i386.md (*sub<mode>3_cconly_overflow): Remove.
5033 (*sub<mode>3_cc_overflow): Ditto.
5034 (*subsi3_zext_cc_overflow): Ditto.
5036 2013-10-22 Steve Ellcey <sellcey@mips.com>
5038 * config/mips/mips.c (mips_rtx_costs): Fix cost estimate for nor
5039 (AND (NOT OP1) (NOT OP2)).
5041 2013-10-22 Bill Schmidt <wschmidt@vnet.ibm.com>
5043 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
5044 meaning of merge-high and merge-low masks for little endian; avoid
5045 use of vector-pack masks for little endian for mismatched modes.
5047 2013-10-22 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5049 * config/tilepro/tilepro.c: Include "tree.h".
5051 2013-10-22 Andreas Schwab <schwab@suse.de>
5053 * config/m68k/m68k.c (notice_update_cc): Handle register conflict
5056 2013-10-22 Paolo Carlini <paolo.carlini@oracle.com>
5058 * doc/contrib.texi ([Fran@,{c}ois Dumont], [Tim Shen],
5059 [Ed Smith-Rowland]): New entries.
5060 ([Stephen M. Webb]): Update.
5062 2013-10-22 Andrew MacLeod <amacleod@redhat.com>
5064 * tree-ssa-ter.h: Remove duplicate copy of file contents.
5066 2013-10-21 Marek Polacek <polacek@redhat.com>
5069 * fold-const.c (fold_range_test): Return 0 if the type is not
5072 2013-10-21 Richard Sandiford <rdsandiford@googlemail.com>
5074 * system.h: Move hwint.h include further down.
5075 * hwint.h (sext_hwi, zext_hwi): Define unconditionally. Add
5076 gcc_checking_asserts.
5077 * hwint.c (sext_hwi, zext_hwi): Delete ENABLE_CHECKING versions.
5079 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
5081 Fix volatile issues in optimize_bit_field_compare.
5082 * fold-const.c (optimize_bit_field_compare): Bail out if
5083 lvolatilep or rvolatilep.
5085 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
5087 Fix DECL_BIT_FIELD depencency on flag_strict_volatile_bitfields
5088 and get_inner_reference returning different pmode for non-volatile
5089 bit-field members dependent on flag_strict_volatile_bitfields.
5090 * stor-layout.c (layout_decl): Remove special handling of
5091 flag_strict_volatile_bitfields.
5092 * expr.c (get_inner_reference): Don't use DECL_BIT_FIELD
5093 if flag_strict_volatile_bitfields > 0 and TREE_THIS_VOLATILE.
5095 2013-10-21 Paulo Matos <pmatos@broadcom.com>
5097 * ipa-inline.c (edge_badness): Cap edge->count at max_count for badness
5100 2013-10-21 Jeff Law <law@redhat.com>
5102 * tree-ssa-threadedge.c (thread_through_normal_block): New
5103 argument VISITED. Remove VISISTED as a local variable. When we
5104 have a threadable jump, verify the destination of the jump has not
5106 (thread_across_edge): Allocate VISITED bitmap once at function
5107 scope and use it throughout. Make sure to set appropriate bits in
5108 VISITED for E (start of jump thread path).
5109 * tree-ssa-threadupdate.c (mark_threaded_blocks): Reject threading
5110 through a joiner if any edge on the path has a recorded jump thread.
5112 2013-10-21 Ian Lance Taylor <iant@google.com>
5114 * doc/invoke.texi (Optimize Options): For -fno-toplevel-reorder,
5115 don't imply that attributes can solve all problems.
5116 (Directory Options): Fix typo.
5118 2013-10-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5120 * config/arm/arm.c (cortexa9_extra_costs): Update mult costs for
5121 extend and extend_add.
5123 2013-10-21 Richard Biener <rguenther@suse.de>
5125 PR tree-optimization/58794
5126 * fold-const.c (operand_equal_p): Compare FIELD_DECL operand
5127 of COMPONENT_REFs with OEP_CONSTANT_ADDRESS_OF left in place.
5129 2013-10-21 Richard Biener <rguenther@suse.de>
5132 * fold-const.c (fold_binary_loc): Fold ((T) (X /[ex] C)) * C
5133 to (T) X for sign-changing conversions (or no conversion).
5135 2013-10-20 Uros Bizjak <ubizjak@gmail.com>
5137 * config/i386/i386.md (kxnor<mode>): Add FLAGS_REG clobber.
5139 2013-10-20 Jan Hubicka <jh@suse.cz>
5141 * config/i386/i386-tune.def: Add comment; organize into categories
5143 2013-10-21 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
5145 * config/i386/i386.c (expand_set_or_movmem_via_loop): Add issetmem
5146 argument. Update function comment.
5147 (expand_set_or_movmem_via_rep): New function combining
5148 expand_movmem_via_rep_mov and expand_setmem_via_rep_stos.
5149 (expand_movmem_via_rep_mov): Remove.
5150 expand_setmem_via_rep_stos): Remove.
5151 (expand_movmem_epilogue): Update calls correspondingly.
5152 (expand_setmem_epilogue_via_loop): Likewise.
5154 (expand_setmem_epilogue): Add VEC_VALUE argument, refactor.
5155 (expand_set_or_movmem_prologue): New function combining
5156 expand_movmem_prologue and expand_setmem_prologue.
5157 (expand_movmem_prologue): Remove.
5158 (expand_setmem_prologue): Remove.
5159 (expand_set_or_movmem_constant_prologue): New function combining
5160 expand_constant_movmem_prologue and expand_constant_setmem_prologue.
5161 (expand_constant_movmem_prologue): Remove.
5162 (expand_constant_setmem_prologue): Remove.
5163 (promote_duplicated_reg): Allow vector-const0 value.
5164 (ix86_expand_set_or_movmem): New function combining ix86_expand_movmem
5165 and ix86_expand_setmem.
5166 (ix86_expand_movmem): Call ix86_expand_set_or_movmem.
5167 (ix86_expand_setmem): Call ix86_expand_set_or_movmem.
5169 2013-10-21 Diego Novillo <dnovillo@google.com>
5171 * asan.c: Include tree.h
5172 * bb-reorder.c: Likewise.
5173 * cfgcleanup.c: Likewise.
5174 * cfgloopmanip.c: Likewise.
5175 * data-streamer-in.c: Likewise.
5176 * data-streamer-out.c: Likewise.
5177 * data-streamer.c: Likewise.
5178 * dwarf2cfi.c: Likewise.
5179 * graphite-blocking.c: Likewise.
5180 * graphite-clast-to-gimple.c: Likewise.
5181 * graphite-dependences.c: Likewise.
5182 * graphite-interchange.c: Likewise.
5183 * graphite-optimize-isl.c: Likewise.
5184 * graphite-poly.c: Likewise.
5185 * graphite-scop-detection.c: Likewise.
5186 * graphite-sese-to-poly.c: Likewise.
5187 * graphite.c: Likewise.
5188 * ipa-devirt.c: Likewise.
5189 * ipa-profile.c: Likewise.
5192 * loop-init.c: Likewise.
5193 * loop-unroll.c: Likewise.
5194 * lower-subreg.c: Likewise.
5195 * lto/lto-object.c: Likewise.
5196 * recog.c: Likewise.
5197 * reginfo.c: Likewise.
5198 * tree-loop-distribution.c: Likewise.
5199 * tree-parloops.c: Likewise.
5200 * tree-ssa-strlen.c: Likewise.
5201 * tree-streamer.c: Likewise.
5202 * value-prof.c: Likewise.
5203 * target-globals.c: Likewise.
5204 * expr.h: Include tree-core.h instead of tree.h.
5205 * gimple.h: Likewise.
5206 * ipa-prop.h: Likewise.
5207 * ipa-utils.h: Likewise.
5208 * lto-streamer.h: Likewise.
5209 * streamer-hooks.h: Likewise.
5210 * ipa-reference.h: Include cgraph.h instead of tree.h.
5211 * cgraph.h: Include basic-block.h instead of tree.h.
5212 * tree-streamer.h: Do not include tree.h.
5213 * genattrtab.c (write_header): Generate inclusion of tree.h.
5214 * genautomata.c (main): Likewise.
5215 * genemit.c: Likewise.
5216 * genopinit.c: Likewise.
5217 * genoutput.c (output_prologue): Likewise.
5218 * genpeep.c: Likewise.
5220 2013-10-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5222 * config/rs6000/altivec.md (vec_unpacku_hi_v16qi): Adjust for
5224 (vec_unpacku_hi_v8hi): Likewise.
5225 (vec_unpacku_lo_v16qi): Likewise.
5226 (vec_unpacku_lo_v8hi): Likewise.
5228 2013-10-20 Jan Hubicka <jh@suse.cz>
5230 * config/i386/x86-tune.def (X86_TUNE_SLOW_IMUL_IMM32_MEM,
5231 X86_TUNE_SLOW_IMUL_IMM8): Keep enabled only for K8 and AMDFAM10.
5232 (X86_TUNE_USE_VECTOR_FP_CONVERTS): Disable for generic.
5234 2013-10-20 Richard Sandiford <rdsandiford@googlemail.com>
5236 * config/mips/mips.h (ISA_HAS_WSBH): Define.
5237 * config/mips/mips.md (UNSPEC_WSBH, UNSPEC_DSBH, UNSPEC_DSHD): New
5239 (bswaphi2, bswapsi2, bswapdi2, wsbh, dsbh, dshd): New patterns.
5241 2013-10-19 John David Anglin <danglin@gcc.gnu.org>
5244 * system.h: Undef m_slot.
5246 2013-10-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5248 * config/rs6000/rs6000.c (vspltis_constant): Make sure we check
5249 all elements for both endian flavors.
5251 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
5254 * config/i386/i386.c (ix86_function_value_regno): Add DX_REG,
5255 ST1_REG and XMM1_REG for 32bit and 64bit targets. Also add DI_REG
5256 and SI_REG for 64bit SYSV ABI targets.
5258 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
5260 * mode-switching.c (create_pre_exit): Rename maybe_builtin_apply
5261 to multi_reg_return. Clarify that we are skipping USEs of multiple
5262 return registers. Use bool type where appropriate.
5264 2013-10-18 Jan Hubicka <jh@suse.cz>
5266 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Disable accumulation
5268 * x86-tune.def (X86_TUNE_USE_LEAVE): Update comment.
5269 (X86_TUNE_PUSH_MEMORY): Likewise.
5270 (X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL,
5271 X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL): New.
5272 (X86_TUNE_ACCUMULATE_OUTGOING_ARGS, X86_TUNE_ALWAYS_FANCY_MATH_387):
5274 * i386.c (x86_accumulate_outgoing_args, x86_arch_always_fancy_math_387,
5275 x86_avx256_split_unaligned_load, x86_avx256_split_unaligned_store):
5277 (ix86_option_override_internal): Update to use tune features instead
5280 2013-10-18 Cong Hou <congh@google.com>
5282 PR tree-optimization/58508
5283 * tree-vect-loop-manip.c (vect_loop_versioning): Hoist loop invariant
5284 statement that contains data refs with zero-step.
5286 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
5288 * tree-ssa.h: Don't include gimple-low.h, tree-ssa-address.h,
5289 sbitmap.h, tree-ssa-threadedge.h, tree-ssa-dom.h and tree-cfgcleanup.h.
5290 * gimple-low.c (gimple_check_call_arg,
5291 gimple_check_call_matching_types): Move to cgraph.c.
5292 * gimple-low.h: Remove prototype.
5293 * cgraph.c: (gimple_check_call_arg, gimple_check_call_matching_types):
5294 Relocate from gimple-low.c.
5295 * cgraph.h: Add prototype. Don't include basic-block.h.
5296 * gimplify.c: Add gimple-low to include list.
5297 * omp-low.c: Add gimple-low and tree-cfgcleanup.h to include list.
5298 * tree-eh.c: Add gimple-low to include list.
5299 * tree-nested.c: Likewise.
5300 * cfgexpand.c: Add tree-ssa-address.h to include list.
5302 * gimple-fold.c: Likewise.
5303 * gimple-ssa-strength-reduction.c: Likewise.
5304 * trans-mem.c: Likewise.
5305 * tree-mudflap.c: Likewise.
5306 * tree-ssa-loop-ivopts.c: Likewise.
5307 * tree-ssa-dom.c: Include tree-ssa-threadedge.h and tree-ssa-dom.h.
5308 (degenerate_phi_result): Move to tree-phinodes.c.
5309 * tree-ssa-loop-ch.c: Include tree-ssa-threadedge.h.
5310 * tree-ssa-threadedge.c: Likewise.
5311 * tree-vrp.c: Likewise.
5312 * tree-phinodes.c (degenerate_phi_result): Relocate here.
5313 * tree-ssa-dom.h (degenerate_phi_result): Remove Prototype.
5314 * tree-phinodes.h (degenerate_phi_result): Add prototype.
5315 * tree-ssa-copy.c: Include tree-ssa-dom.h.
5316 * tree-ssa-forwprop.c: Likewise.
5317 * tree-cfgcleanup.c (execute_cleanup_cfg_post_optimizing,
5318 pass_data_cleanup_cfg_post_optimizing,
5319 make_pass_cleanup_cfg_post_optimizing): Relocate from tree-optimize.c.
5320 * tree-optimize.c: Delete File.
5321 * graphite.c: Include tree-cfgcleanup.h.
5322 * passes.c: Likewise.
5323 * tree-cfg.c: Likewise.
5324 * tree-profile.c: Likewise.
5325 * tree-ssa-dse.c: Likewise.
5326 * tree-ssa-loop-ivcanon.c: Likewise.
5327 * tree-switch-conversion.c: Don't include tree-ssa-operands.h.
5328 * tree-outof-ssa.c: Include sbitmap.h.
5329 * tree-ssa-live.c: Likewise.
5330 * tree-ssa-propagate.c: Likewise.
5331 * tree-ssa-structalias.c: Likewise.
5332 * tree-stdarg.c: Likewise.
5333 * Makefile.in (OBJS): Delete tree-optimize.o.
5334 * basic-block.h (gcov_type, gcov_type_unsigned): Move to coretypes.h.
5335 * coretypes.h (gcov_type, gcov_type_unsigned): Relocate here.
5336 * varasm.c: Include basic-block.h.
5337 * cfgloop.h: Include function.h instead of basic-block.h
5338 (bb_loop_depth): Move to cfgloop.c.
5339 * cfgloop.c (bb_loop_depth): Relocate from cfgloop.h.
5341 2013-10-18 Teresa Johnson <tejohnson@google.com>
5343 * predict.c (probably_never_executed): Compare frequency-based
5344 count to number of training runs.
5345 * params.def (UNLIKELY_BB_COUNT_FRACTION): New parameter.
5347 2013-10-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5349 * config/arm/arm.c (cortexa9_extra_costs): New table.
5350 (arm_cortex_a9_tune): Use cortexa9_extra_costs.
5352 2013-10-18 Jeff Law <law@redhat.com>
5354 * tree-ssa-threadupdate.c: Do not include "tm.h" or "tm_p.h".
5356 * tree-ssa-threadupdate.c: Include "dbgcnt.h".
5357 (register_jump_thread): Add "registered_jump_thread" debug
5359 * dbgcnt.def (registered_jump_thread): New debug counter.
5361 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
5363 * config/rs6000/rs6000.c: Include cgraph.h.
5365 2013-10-18 Teresa Johnson <tejohnson@google.com>
5367 * tree-ssa-tail-merge.c (replace_block_by): Update edge
5368 weights during merging.
5370 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
5372 * tree-cfg.h: Rename from tree-flow.h. Remove #includes.
5373 * tree-ssa.h: Relocate required #includes from tree-cfg.h.
5374 * tree-ssa-operands.h: Remove prototype.
5375 * tree-ssa-operands.c (virtual_operand_p): Move to gimple.c.
5376 * gimple.c (virtual_operand_p): Relocate from gimple.c.
5377 * gimple.h: Add prototype.
5378 * gimple-ssa.h: Include tree-ssa-operands.h.
5379 * tree-dump.c: Add tree-cfg.h to include list.
5380 * tree-ssa-alias.c: Add ipa-reference.h to include list.
5381 * config/alpha/alpha.c: Include gimple-ssa.h instead of tree-flow.h.
5382 * config/i386/i386.c: Don't include tree-flow.h.
5383 * config/rs6000/rs6000.c: Likewise.
5385 2013-10-18 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5387 * config/frv/frv.c (frv_init_cumulative_args): Fix wrong cast.
5389 2013-10-18 Richard Biener <rguenther@suse.de>
5391 * stor-layout.c (layout_type): Do not change TYPE_PRECISION
5392 or TYPE_UNSIGNED of integral types.
5393 (set_min_and_max_values_for_integral_type): Leave TYPE_MIN/MAX_VALUE
5394 NULL_TREE for zero-precision integral types.
5396 2013-10-18 James Greenhalgh <james.greenhalgh@arm.com>
5398 * config/aarch64/arm_neon.h
5399 (vcvt<ds>_n_<fsu><32,64>_<fsu><32,64>): Correct argument types.
5401 2013-10-17 Sriraman Tallam <tmsriram@google.com>
5404 * opth-gen.awk: Define target_flags_explicit.
5406 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
5408 * config/rs6000/rs6000.c (enum rs6000_reload_reg_type): Add new
5409 fields to the reg_addr array that describes the valid addressing
5410 mode for any register, general purpose registers, floating point
5411 registers, and Altivec registers.
5412 (FIRST_RELOAD_REG_CLASS): Likewise.
5413 (LAST_RELOAD_REG_CLASS): Likewise.
5414 (struct reload_reg_map_type): Likewise.
5415 (reload_reg_map_type): Likewise.
5416 (RELOAD_REG_VALID): Likewise.
5417 (RELOAD_REG_MULTIPLE): Likewise.
5418 (RELOAD_REG_INDEXED): Likewise.
5419 (RELOAD_REG_OFFSET): Likewise.
5420 (RELOAD_REG_PRE_INCDEC): Likewise.
5421 (RELOAD_REG_PRE_MODIFY): Likewise.
5422 (reg_addr): Likewise.
5423 (mode_supports_pre_incdec_p): New helper functions to say whether
5424 a given mode supports PRE_INC, PRE_DEC, and PRE_MODIFY.
5425 (mode_supports_pre_modify_p): Likewise.
5426 (rs6000_debug_vector_unit): Rearrange the -mdebug=reg output to
5427 print the valid address mode bits for each mode.
5428 (rs6000_debug_print_mode): Likewise.
5429 (rs6000_debug_reg_global): Likewise.
5430 (rs6000_setup_reg_addr_masks): New function to set up the address
5431 mask bits for each type.
5432 (rs6000_init_hard_regno_mode_ok): Use memset to clear arrays.
5433 Call rs6000_setup_reg_addr_masks to set up the address mask bits.
5434 (rs6000_legitimate_address_p): Use mode_supports_pre_incdec_p and
5435 mode_supports_pre_modify_p to determine if PRE_INC, PRE_DEC, and
5436 PRE_MODIFY are supported.
5437 (rs6000_output_move_128bit): Change to use {src,dest}_vmx_p for altivec
5438 registers, instead of {src,dest}_av_p.
5439 (rs6000_print_options_internal): Tweak the debug output slightly.
5441 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
5443 * config/i386/sse.md (*vec_widen_smult_even_v8si): Remove
5446 2013-10-17 Andrew MacLeod <amacleod@redhat.com>
5448 * tree-flow.h (struct omp_region): Move to omp-low.c.
5449 Remove omp_ prototypes and variables.
5450 * gimple.h (omp_reduction_init): Move prototype to omp-low.h.
5451 (copy_var_decl): Relocate prototype from tree-flow.h.
5452 * gimple.c (copy_var_decl): Relocate from omp-low.c.
5453 * tree.h: Move prototype to omp-low.h.
5454 * omp-low.h: New File. Relocate prototypes here.
5455 * omp-low.c (struct omp_region): Make local here.
5456 (root_omp_region): Make static.
5457 (copy_var_decl) Move to gimple.c.
5458 (new_omp_region): Make static.
5459 (make_gimple_omp_edges): New. Refactored from tree-cfg.c make_edges.
5460 * tree-cfg.c: Include omp-low.h.
5461 (make_edges): Factor out OMP specific bits to make_gimple_omp_edges.
5462 * gimplify.c: Include omp-low.h.
5463 * tree-parloops.c: Likewise.
5465 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
5467 * config/i386/i386.c (ix86_fixup_binary_operands): When both source
5468 operands are in memory, prefer to force non-matched operand 1 to
5471 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
5474 * config/rs6000/rs6000.c (rs6000_legitimate_address_p): Only
5475 restrict TImode addresses to single indirect registers if both
5476 -mquad-memory and -mvsx-timode are used.
5477 (rs6000_output_move_128bit): Use quad_load_store_p to determine if
5478 we should emit load/store quad. Remove using %y for quad memory
5481 * config/rs6000/rs6000.md (mov<mode>_ppc64, TI/PTImode): Add
5482 constraints to allow load/store quad on machines where TImode is
5483 not allowed in VSX registers. Use 'n' instead of 'F' constraint
5484 for TImode to load integer constants.
5486 2013-10-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5488 * config/aarch64/aarch64.c (aarch64_print_operand): Handle 'c'.
5490 2013-10-17 Marcus Shawcroft <marcus.shawcroft@arm.com>
5492 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Adjust
5493 handling of STACK_REG.
5495 2013-10-17 Richard Biener <rguenther@suse.de>
5497 PR tree-optimization/58143
5498 * tree-ssa-loop-im.c (arith_code_with_undefined_signed_overflow):
5500 (rewrite_to_defined_overflow): Likewise.
5501 (move_computations_dom_walker::before_dom): Rewrite stmts
5502 with undefined signed overflow that are not always executed
5503 into unsigned arithmetic.
5505 2013-10-16 Michael Meissner <meissner@linux.vnet.ibm.com>
5508 * config/rs6000/rs6000.opt (rs6000_isa_flags_explicit): Move the
5509 explicit isa flag to be an options variable, instead of using
5510 global_options_set. Remove define from rs6000.h.
5511 * config/rs6000/rs6000.h (rs6000_isa_flags_explicit): Likewise.
5513 * config/rs6000/rs6000.c (rs6000_option_override_internal):
5514 Initialize rs6000_isa_flags_explicit.
5515 (rs6000_function_specific_save): Add gcc_options* parameter, so
5516 that the powerpc builds after the 2013-10-15 changes.
5517 (rs6000_function_specific_restore): Likewise.
5519 2013-10-16 DJ Delorie <dj@redhat.com>
5521 * config/rl78/rl78.c (rl78_alloc_address_registers_macax): Verify
5522 op is a REG before checking REGNO.
5523 (rl78_alloc_physical_registers): Verify pattern is a SET before
5526 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5528 * config/rs6000/vector.md (vec_unpacks_hi_v4sf): Correct for
5530 (vec_unpacks_lo_v4sf): Likewise.
5531 (vec_unpacks_float_hi_v4si): Likewise.
5532 (vec_unpacks_float_lo_v4si): Likewise.
5533 (vec_unpacku_float_hi_v4si): Likewise.
5534 (vec_unpacku_float_lo_v4si): Likewise.
5536 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5538 * config/rs6000/vsx.md (vsx_concat_<mode>): Adjust output for LE.
5539 (vsx_concat_v2sf): Likewise.
5541 2013-10-16 James Greenhalgh <james.greenhalgh@arm.com>
5543 * config/aarch64/aarch64.md
5544 (*mov<mode>_aarch64): Fix output template for DUP (element) Scalar.
5546 2013-10-16 Andrew MacLeod <amacleod@redhat.com>
5548 PR tree-optimization/58697
5549 * cfgloop.c (get_estimated_loop_iterations_int): Rename from
5550 estimated_loop_iterations_int.
5551 (max_stmt_executions_int): Call get_max_loop_iterations_int.
5552 (get_max_loop_iterations_int): New. HWINT version of
5553 get_max_loop_iterations.
5554 * cfgloop.h: Add prototypes.
5555 * loop-iv.c (find_simple_exit): call get_estimated_loop_iterations_int.
5556 * loop-unroll.c (decide_peel_once_rolling): Call
5557 get_estimated_loop_iterations_int.
5558 * tree-ssa-loop-niter.c (estimated_loop_iterations_int): Add back.
5559 * tree-ssa-loop-niter.h: Tweak prototypes.
5561 2013-10-16 David Malcolm <dmalcolm@redhat.com>
5563 * gengtype-parse.c (struct_field_seq): Ignore access-control
5564 keywords ("public:" etc).
5566 2013-10-16 Marcus Shawcroft <marcus.shawcroft@arm.com>
5568 * config/aarch64/aarch64.c (aarch64_regno_regclass): Classify
5569 FRAME_POINTER_REGNUM and ARG_POINTER_REGNUM as POINTER_REGS.
5571 2013-10-16 Yvan Roux <yvan.roux@linaro.org>
5573 * config/arm/arm.opt (mlra): New option.
5574 * config/arm/arm.c (arm_lra_p): New function.
5575 (TARGET_LRA_P): Define.
5577 2013-10-16 Paulo Matos <pmatos@broadcom.com>
5579 * tree-core.h (tree_code_name): Remove.
5580 * tree.h (get_tree_code_name): New prototype.
5581 * tree.c (tree_code_name): Make static.
5582 (get_tree_code_name): New function.
5583 (dump_tree_statistics, tree_check_failed, tree_not_check_failed,
5584 tree_class_check_failed, tree_range_check_failed,
5585 tree_not_class_check_failed, omp_clause_check_failed,
5586 tree_contains_struct_check_failed, tree_operand_check_failed): Use new
5587 wrapper get_tree_code_name instead of calling tree_code_name directly.
5588 * tree-vrp.c (dump_asserts_for): Likewise.
5589 * tree-dump.c (dequeue_and_dump): Likewise.
5590 * tree-pretty-print.c (do_niy, dump_generic_node): Likewise.
5591 * tree-pretty-print.h (pp_unsupported_tree): Likewise.
5592 * lto-streamer-out.c (lto_write_tree, DFS_write_tree): Likewise.
5593 * tree-ssa-dom.c (print_expr_hash_elt): Likewise.
5594 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs,
5595 dump_ternary_rhs, dump_gimple_assign, dump_gimple_cond,
5596 dump_gimple_omp_for): Likewise.
5597 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Likewise.
5598 * tree-ssa-pre.c (print_pre_expr): Likewise.
5599 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Likewise.
5600 * print-tree.c (print_node_brief, print_node): Likewise.
5601 * gimple.c (gimple_check_failed): Likewise.
5602 * lto-streamer.c (lto_tag_name, print_lto_report): Likewise.
5603 * config/frv/frv.c (frv_init_cumulative_args): Likewise.
5604 * config/mep/mep.c (mep_validate_vliw): Likewise.
5605 * config/iq2000/iq2000.c (init_cumulative_args): Likewise.
5606 * config/rs6000/rs6000.c (init_cumulative_args): Likewise.
5608 2013-10-16 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
5610 * config/i386/i386.c (ix86_option_override_internal): Enable FMA4
5613 2013-10-16 Hans-Peter Nilsson <hp@axis.com>
5615 * config/cris/t-elfmulti (MULTILIB_OPTIONS, MULTILIB_DIRNAMES)
5616 (MULTILIB_MATCHES): Add multilib for -march=v8.
5618 2013-10-15 Sriraman Tallam <tmsriram@google.com>
5621 * optc-save-gen.awk: Add extra parameter to the save and restore
5623 * opth-gen.awk: Generate new TARGET_* macros to accept a parameter.
5624 * tree.c (build_optimization_node): New parameter. Add extra parameter
5625 to call to cl_optimization_save.
5626 (build_target_option_node): New parameter. Add extra parameter
5627 to call to cl_target_option_save.
5628 * tree.h (build_optimization_node): New parameter.
5629 (build_target_option_node): New parameter.
5630 * c-family/c-common.c (handle_optimize_attribute): Fix calls to
5631 build_optimization_node and build_target_option_node.
5632 * c-family/c-pragma.c (handle_pragma_optimize): Ditto.
5633 (handle_pragma_push_options): Ditto.
5634 * toplev.c (process_options): Ditto.
5635 * opts.c (init_options_struct): Check for opts_set non-null.
5636 * target.def (target_option.save): New parameter.
5637 (target_option.restore): New parameter.
5638 * tm.texi: Generate.
5639 * config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
5640 (ix86_pragma_target_parse): Ditto.
5641 * config/i386/i386-protos.h (ix86_valid_target_attribute_tree): New
5643 * config/rs6000/rs6000.c (rs6000_option_override_internal): Fix calls
5644 to build_optimization_node and build_target_option_node.
5645 (rs6000_valid_attribute_p): Ditto.
5646 (rs6000_pragma_target_parse): Ditto.
5647 * config/i386/i386.opt (x_ix86_target_flags_explicit): New TargetSave
5649 * config/i386/i386.h:
5650 TARGET_64BIT_P: New Macro
5651 TARGET_MMX_P: New Macro.
5652 TARGET_3DNOW_P: New Macro.
5653 TARGET_3DNOW_A_P: New Macro.
5654 TARGET_SSE_P: New Macro.
5655 TARGET_SSE2_P: New Macro.
5656 TARGET_SSE3_P: New Macro.
5657 TARGET_SSSE3_P: New Macro.
5658 TARGET_SSE4_1_P: New Macro.
5659 TARGET_SSE4_2_P: New Macro.
5660 TARGET_AVX_P: New Macro.
5661 TARGET_AVX2_P: New Macro.
5662 TARGET_AVX512F_P: New Macro.
5663 TARGET_AVX512PF_P: New Macro.
5664 TARGET_AVX512ER_P: New Macro.
5665 TARGET_AVX512CD_P: New Macro.
5666 TARGET_FMA_P: New Macro.
5667 TARGET_SSE4A_P: New Macro.
5668 TARGET_FMA4_P: New Macro.
5669 TARGET_XOP_P: New Macro.
5670 TARGET_LWP_P: New Macro.
5671 TARGET_ABM_P: New Macro.
5672 TARGET_BMI_P: New Macro.
5673 TARGET_BMI2_P: New Macro.
5674 TARGET_LZCNT_P: New Macro.
5675 TARGET_TBM_P: New Macro.
5676 TARGET_POPCNT_P: New Macro.
5677 TARGET_SAHF_P: New Macro.
5678 TARGET_MOVBE_P: New Macro.
5679 TARGET_CRC32_P: New Macro.
5680 TARGET_AES_P: New Macro.
5681 TARGET_PCLMUL_P: New Macro.
5682 TARGET_CMPXCHG16B_P: New Macro.
5683 TARGET_FSGSBASE_P: New Macro.
5684 TARGET_RDRND_P: New Macro.
5685 TARGET_F16C_P: New Macro.
5686 TARGET_RTM_P: New Macro.
5687 TARGET_HLE_P: New Macro.
5688 TARGET_RDSEED_P: New Macro.
5689 TARGET_PRFCHW_P: New Macro.
5690 TARGET_ADX_P: New Macro.
5691 TARGET_FXSR_P: New Macro.
5692 TARGET_XSAVE_P: New Macro.
5693 TARGET_XSAVEOPT_P: New Macro.
5694 TARGET_LP64_P: New Macro.
5695 TARGET_X32_P: New Macro.
5696 TARGET_FPMATH_DEFAULT_P: New Macro.
5697 TARGET_FLOAT_RETURNS_IN_80387_P: New Macro.
5698 * config/i386/i386.c (ix86_option_override_internal): New parameters.
5700 Change ix86_tune_string to access opts->x_ix86_tune_string.
5701 Change ix86_isa_flags to access opts->x_ix86_isa_flags.
5702 Change ix86_arch_string to access opts->x_ix86_arch_string.
5703 Change ix86_stringop_alg to access opts->x_ix86_stringop_alg.
5704 Change ix86_pmode to access opts->x_ix86_pmode.
5705 Change ix86_abi to access opts->x_ix86_abi.
5706 Change ix86_cmodel to access opts->x_ix86_cmodel.
5707 Change ix86_asm_dialect to access opts->x_ix86_asm_dialect.
5708 Change ix86_isa_flags_explicit to access
5709 opts->x_ix86_isa_flags_explicit.
5710 Change ix86_dump_tunes to access opts->x_ix86_dump_tunes.
5711 Change ix86_regparm to access opts->x_ix86_regparm.
5712 Change ix86_branch_cost to access opts->x_ix86_branch_cost.
5713 Change ix86_preferred_stack_boundary_arg to access
5714 opts->x_ix86_preferred_stack_boundary_arg.
5715 Change ix86_force_align_arg_pointer to access
5716 opts->x_ix86_force_align_arg_pointer.
5717 Change ix86_incoming_stack_boundar_arg to access
5718 opts->x_ix86_incoming_stack_boundar_arg.
5719 Change ix86_fpmath to access opts->x_ix86_fpmath.
5720 Change ix86_veclibabi_type to access opts->x_ix86_veclibabi_type.
5721 Change ix86_recip_name to access opts->x_ix86_recip_name.
5722 Change ix86_stack_protector_guard to access
5723 opts->x_ix86_stack_protector_guard.
5724 Change ix86_tune_memcpy_strategy to access
5725 opts->x_ix86_tune_memcpy_strategy.
5726 Change ix86_tune_memset_strategy to access
5727 opts->x_ix86_tune_memset_strategy.
5728 Change global_options to access opts.
5729 Change global_options_set to access opts_set.
5730 Change TARGET_64BIT to TARGET_64BIT_P (opts->...).
5731 Change TARGET_MMX to TARGET_MMX_P (opts->...).
5732 Change TARGET_3DNOW to TARGET_3DNOW_P (opts->...).
5733 Change TARGET_3DNOW_A to TARGET_3DNOW_A_P (opts->...).
5734 Change TARGET_SSE to TARGET_SSE_P (opts->...).
5735 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...).
5736 Change TARGET_SSE3 to TARGET_SSE3_P (opts->...).
5737 Change TARGET_SSSE3 to TARGET_SSSE3_P (opts->...).
5738 Change TARGET_SSE4_1 to TARGET_SSE4_1_P (opts->...).
5739 Change TARGET_SSE4_2 to TARGET_SSE4_2_P (opts->...).
5740 Change TARGET_AVX to TARGET_AVX_P (opts->...).
5741 Change TARGET_AVX2 to TARGET_AVX2_P (opts->...).
5742 Change TARGET_AVX512F to TARGET_AVX512F_P (opts->...).
5743 Change TARGET_AVX512PF to TARGET_AVX512PF_P (opts->...).
5744 Change TARGET_AVX512ER to TARGET_AVX512ER_P (opts->...).
5745 Change TARGET_AVX512CD to TARGET_AVX512CD_P (opts->...).
5746 Change TARGET_FMA to TARGET_FMA_P (opts->...).
5747 Change TARGET_SSE4A to TARGET_SSE4A_P (opts->...).
5748 Change TARGET_FMA4 to TARGET_FMA4_P (opts->...).
5749 Change TARGET_XOP to TARGET_XOP_P (opts->...).
5750 Change TARGET_LWP to TARGET_LWP_P (opts->...).
5751 Change TARGET_ABM to TARGET_ABM_P (opts->...).
5752 Change TARGET_BMI to TARGET_BMI_P (opts->...).
5753 Change TARGET_BMI2 to TARGET_BMI2_P (opts->...).
5754 Change TARGET_LZCNT to TARGET_LZCNT_P (opts->...).
5755 Change TARGET_TBM to TARGET_TBM_P (opts->...).
5756 Change TARGET_POPCNT to TARGET_POPCNT_P (opts->...).
5757 Change TARGET_SAHF to TARGET_SAHF_P (opts->...).
5758 Change TARGET_MOVBE to TARGET_MOVBE_P (opts->...).
5759 Change TARGET_CRC32 to TARGET_CRC32_P (opts->...).
5760 Change TARGET_AES to TARGET_AES_P (opts->...).
5761 Change TARGET_PCLMUL to TARGET_PCLMUL_P (opts->...).
5762 Change TARGET_CMPXCHG16B to TARGET_CMPXCHG16B_P (opts->...).
5763 Change TARGET_FSGSBASE to TARGET_FSGSBASE_P (opts->...).
5764 Change TARGET_RDRND to TARGET_RDRND_P (opts->...).
5765 Change TARGET_F16C to TARGET_F16C_P (opts->...).
5766 Change TARGET_RTM to TARGET_RTM_P (opts->...).
5767 Change TARGET_HLE to TARGET_HLE_P (opts->...).
5768 Change TARGET_RDSEED to TARGET_RDSEED_P (opts->...).
5769 Change TARGET_PRFCHW to TARGET_PRFCHW_P (opts->...).
5770 Change TARGET_ADX to TARGET_ADX_P (opts->...).
5771 Change TARGET_FXSR to TARGET_FXSR_P (opts->...).
5772 Change TARGET_XSAVE to TARGET_XSAVE_P (opts->...).
5773 Change TARGET_XSAVEOPT to TARGET_XSAVEOPT_P (opts->...).
5774 Change TARGET_LP64 to TARGET_LP64_P (opts->...).
5775 Change TARGET_X32 to TARGET_X32_P (opts->...).
5776 Change TARGET_FPMATH_DEFAULT to TARGET_FPMATH_DEFAULT_P (opts->...).
5777 Change TARGET_FLOAT_RETURNS_IN_80387 to
5778 TARGET_FLOAT_RETURNS_IN_80387_P (opts->...).
5779 (ix86_function_specific_save): New parameter. Use opts-> fields
5780 to replace global fields.
5781 (ix86_function_specific_restore): Ditto.
5782 (ix86_valid_target_attribute_inner_p): New parameters.
5784 Fix call to ix86_handle_option and set_option.
5785 (ix86_valid_target_attribute_tree): New parameters.
5786 Change global_options to access opts.
5787 Change global_options_set to access opts_set.
5788 Fix call to ix86_valid_target_attribute_inner_p.
5789 Change ix86_tune_string to access opts->x_ix86_tune_string.
5790 Change ix86_arch_string to access opts->x_ix86_arch_string.
5791 Change ix86_fpmath to access opts->x_ix86_fpmath
5792 Fix call to ix86_option_override_internal.
5793 Fix call to ix86_add_new_builtins.
5794 Fix calls to build_optimization_node and build_target_option_node.
5795 (ix86_valid_target_attribute_p): Remove access to global_options.
5796 Use new gcc_options structure func_options.
5797 Fix call to ix86_valid_target_attribute_tree.
5798 Fix call to build_optimization_node.
5799 (get_builtin_code_for_version): Fix call to
5800 ix86_valid_target_attribute_tree.
5802 2013-10-15 David Malcolm <dmalcolm@redhat.com>
5804 * Makefile.in (PICFLAG): New.
5805 (enable_host_shared): New.
5806 (INTERNAL_CFLAGS): Use PICFLAG.
5807 (LIBIBERTY): Use pic build of libiberty.a if configured with
5808 --enable-host-shared.
5809 * configure.ac: Add --enable-host-shared, setting up new
5811 * configure: Regenerate.
5812 * doc/install.texi (--enable-shared): Add note contrasting it with ...
5813 (--enable-host-shared): New option.
5815 2013-10-15 Richard Biener <rguenther@suse.de>
5817 * tree-tailcall.c (find_tail_calls): Don't use tail-call recursion
5818 for built-in functions.
5820 2013-10-15 Zhenqiang Chen <zhenqiang.chen@arm.com>
5822 * tree-ssa-reassoc.c: Include rtl.h and tm_p.h.
5823 (optimize_range_tests_1): New function,
5824 extracted from optimize_range_tests.
5825 (optimize_range_tests_xor): Similarly.
5826 (optimize_range_tests_diff): New function.
5827 (optimize_range_tests): Use optimize_range_tests_1.
5829 2013-10-15 Cong Hou <congh@google.com>
5831 * tree-vect-loop.c (vect_is_simple_reduction_1): Relax the
5832 requirement of the reduction pattern so that one operand of the
5833 reduction operation can come from outside of the loop.
5835 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
5837 * config/arm/neon-schedgen.ml: Remove.
5838 * config/arm/cortex-a9-neon.md: Remove comment regarding
5841 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
5843 * config/arm/types: Remove old neon types.
5845 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
5847 * config/arm/cortex-a7.md
5848 (cortex_a7_neon_type): New.
5849 (cortex_a7_neon_mul): Update for new types.
5850 (cortex_a7_neon_mla): Likewise.
5851 (cortex_a7_neon): Likewise.
5853 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
5855 * config/arm/cortex-a15-neon.md
5856 (cortex_a15_neon_type): New,
5858 (cortex_a15_neon_int_1): Remove.
5859 (cortex_a15_neon_int_2): Likewise.
5860 (cortex_a15_neon_int_3): Likewise.
5861 (cortex_a15_neon_int_4): Likewise.
5862 (cortex_a15_neon_int_5): Likewise.
5863 (cortex_a15_neon_vqneg_vqabs): Likewise.
5864 (cortex_a15_neon_vmov): Likewise.
5865 (cortex_a15_neon_vaba): Likewise.
5866 (cortex_a15_neon_vaba_qqq): Likewise.
5867 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
5868 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
5869 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
5871 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
5872 (cortex_a15_neon_mla_qqq_8_16): Likewise.
5873 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
5874 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
5875 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
5876 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
5877 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
5878 (cortex_a15_neon_shift_1): Likewise.
5879 (cortex_a15_neon_shift_2): Likewise.
5880 (cortex_a15_neon_shift_3): Likewise.
5881 (cortex_a15_neon_vshl_ddd): Likewise.
5882 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
5883 (cortex_a15_neon_vsra_vrsra): Likewise.
5884 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
5885 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
5886 (cortex_a15_neon_bp_3cycle): Likewise.
5887 (cortex_a15_neon_ldm_2): Likewise.
5888 (cortex_a15_neon_stm_2): Likewise.
5889 (cortex_a15_neon_mcr): Likewise.
5890 (cortex_a15_neon_mrc): Likewise.
5891 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
5892 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
5893 (cortex_a15_neon_fp_vmul_ddd): Likewise.
5894 (cortex_a15_neon_fp_vmul_qqd): Likewise.
5895 (cortex_a15_neon_fp_vmla_ddd): Likewise.
5896 (cortex_a15_neon_fp_vmla_qqq): Likewise.
5897 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
5898 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
5899 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
5900 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
5901 (cortex_a15_neon_bp_simple): Likewise.
5902 (cortex_a15_neon_bp_2cycle): Likewise.
5903 (cortex_a15_neon_bp_3cycle): Likewise.
5904 (cortex_a15_neon_vld1_1_2_regs): Likewise.
5905 (cortex_a15_neon_vld1_3_4_regs): Likewise.
5906 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
5907 (cortex_a15_neon_vld2_4_regs): Likewise.
5908 (cortex_a15_neon_vld3_vld4): Likewise.
5909 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
5910 (cortex_a15_neon_vst1_3_4_regs): Likewise.
5911 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
5912 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
5913 (cortex_a15_neon_vst3_vst4): Rename to...
5914 (cortex_a15_neon_vst4): This, update for new attributes.
5915 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
5916 (cortex_a15_neon_vld3_vld4_lane): Likewise.
5917 (cortex_a15_neon_vst1_vst2_lane): Likewise.
5918 (cortex_a15_neon_vst3_vst4_lane): Likewise.
5919 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
5920 (cortex_a15_neon_ldm_2): Likewise.
5921 (cortex_a15_neon_stm_2): Likewise.
5922 (cortex_a15_neon_mcr): Likewise.
5923 (cortex_a15_neon_mcr_2_mcrr): Likewise.
5924 (cortex_a15_neon_mrc): Likewise.
5925 (cortex_a15_neon_mrrc): Likewise.
5927 (cortex_a15_neon_abd): New.
5928 (cortex_a15_neon_abd_q): Likewise.
5929 (cortex_a15_neon_aba): Likewise.
5930 (cortex_a15_neon_aba_q): Likewise.
5931 (cortex_a15_neon_acc): Likewise.
5932 (cortex_a15_neon_acc_q): Likewise.
5933 (cortex_a15_neon_arith_basic): Likewise.
5934 (cortex_a15_neon_arith_complex): Likewise.
5935 (cortex_a15_neon_multiply): Likewise.
5936 (cortex_a15_neon_multiply_q): Likewise.
5937 (cortex_a15_neon_mla): Likewise.
5938 (cortex_a15_neon_mla_q): Likewise.
5939 (cortex_a15_neon_sat_mla_long): Likewise.
5940 (cortex_a15_neon_shift_acc): Likewise.
5941 (cortex_a15_neon_shift_imm_basic): Likewise.
5942 (cortex_a15_neon_shift_imm_complex): Likewise.
5943 (cortex_a15_neon_shift_reg_basic): Likewise.
5944 (cortex_a15_neon_shift_reg_basic_q): Likewise.
5945 (cortex_a15_neon_shift_reg_complex): Likewise.
5946 (cortex_a15_neon_shift_reg_complex_q): Likewise.
5947 (cortex_a15_neon_fp_negabs): Likewise
5948 (cortex_a15_neon_fp_arith): Likewise
5949 (cortex_a15_neon_fp_arith_q): Likewise
5950 (cortex_a15_neon_fp_cvt_int): Likewise
5951 (cortex_a15_neon_fp_cvt_int_q): Likewise
5952 (cortex_a15_neon_fp_cvt_16): Likewise
5953 (cortex_a15_neon_fp_mul): Likewise
5954 (cortex_a15_neon_fp_mul_q): Likewise
5955 (cortex_a15_neon_fp_mla): Likewise
5956 (cortex_a15_neon_fp_mla_q): Likewise
5957 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
5958 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
5959 (cortex_a15_neon_bitops): Likewise.
5960 (cortex_a15_neon_bitops_q): Likewise.
5961 (cortex_a15_neon_from_gp): Likewise.
5962 (cortex_a15_neon_from_gp_q): Likewise.
5963 (cortex_a15_neon_tbl3_tbl4): Likewise.
5964 (cortex_a15_neon_zip_q): Likewise.
5965 (cortex_a15_neon_to_gp): Likewise.
5966 (cortex_a15_neon_load_a): Likewise.
5967 (cortex_a15_neon_load_b): Likewise.
5968 (cortex_a15_neon_load_c): Likewise.
5969 (cortex_a15_neon_load_d): Likewise.
5970 (cortex_a15_neon_load_e): Likewise.
5971 (cortex_a15_neon_load_f): Likewise.
5972 (cortex_a15_neon_store_a): Likewise.
5973 (cortex_a15_neon_store_b): Likewise.
5974 (cortex_a15_neon_store_c): Likewise.
5975 (cortex_a15_neon_store_d): Likewise.
5976 (cortex_a15_neon_store_e): Likewise.
5977 (cortex_a15_neon_store_f): Likewise.
5978 (cortex_a15_neon_store_g): Likewise.
5979 (cortex_a15_neon_store_h): Likewise.
5980 (cortex_a15_vfp_to_from_gp): Likewise.
5982 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
5984 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
5986 (cortex_a9_neon_vshl_ddd): Remove.
5987 (cortex_a9_neon_vst3_vst4): Likewise.
5988 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
5990 (cortex_a9_neon_bit_ops_q): New.
5992 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
5993 (cortex_a9_neon_int_2): Likewise.
5994 (cortex_a9_neon_int_3): Likewise.
5995 (cortex_a9_neon_int_4): Likewise.
5996 (cortex_a9_neon_int_5): Likewise.
5997 (cortex_a9_neon_vqneg_vqabs): Likewise.
5998 (cortex_a9_neon_vmov): Likewise.
5999 (cortex_a9_neon_vaba): Likewise.
6000 (cortex_a9_neon_vaba_qqq): Likewise.
6001 (cortex_a9_neon_shift_1): Likewise.
6002 (cortex_a9_neon_shift_2): Likewise.
6003 (cortex_a9_neon_shift_3): Likewise.
6004 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
6005 (cortex_a9_neon_vsra_vrsra): Likewise.
6006 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6007 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
6008 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
6010 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6011 (cortex_a9_neon_mla_qqq_8_16): Likewise.
6012 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
6014 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
6015 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
6016 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
6017 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
6018 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
6019 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
6020 (cortex_a9_neon_fp_vsum): Likewise.
6021 (cortex_a9_neon_fp_vmul_ddd): Likewise.
6022 (cortex_a9_neon_fp_vmul_qqd): Likewise.
6023 (cortex_a9_neon_fp_vmla_ddd): Likewise.
6024 (cortex_a9_neon_fp_vmla_qqq): Likewise.
6025 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
6026 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
6027 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
6028 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
6029 (cortex_a9_neon_bp_simple): Likewise.
6030 (cortex_a9_neon_bp_2cycle): Likewise.
6031 (cortex_a9_neon_bp_3cycle): Likewise.
6032 (cortex_a9_neon_ldr): Likewise.
6033 (cortex_a9_neon_str): Likewise.
6034 (cortex_a9_neon_vld1_1_2_regs): Likewise.
6035 (cortex_a9_neon_vld1_3_4_regs): Likewise.
6036 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
6037 (cortex_a9_neon_vld2_4_regs): Likewise.
6038 (cortex_a9_neon_vld3_vld4): Likewise.
6039 (cortex_a9_neon_vld1_vld2_lane): Likewise.
6040 (cortex_a9_neon_vld3_vld4_lane): Likewise.
6041 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
6042 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
6043 (cortex_a9_neon_vst1_3_4_regs): Likewise.
6044 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
6045 (cortex_a9_neon_vst1_vst2_lane): Likewise.
6046 (cortex_a9_neon_vst3_vst4_lane): Likewise.
6047 (cortex_a9_neon_mcr): Likewise.
6048 (cortex_a9_neon_mcr_2_mcrr): Likewise.
6049 (cortex_a9_neon_mrc): Likewise.
6050 (cortex_a9_neon_mrrc): Likewise.
6052 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
6054 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
6056 (cortex_a8_neon_vshl_ddd): Remove.
6057 (cortex_a8_neon_vst3_vst4): Likewise.
6058 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
6060 (cortex_a8_neon_bit_ops_q): New.
6062 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
6063 (cortex_a8_neon_int_2): Likewise..
6064 (cortex_a8_neon_int_3): Likewise.
6065 (cortex_a8_neon_int_5): Likewise.
6066 (cortex_a8_neon_vqneg_vqabs): Likewise.
6067 (cortex_a8_neon_int_4): Likewise.
6068 (cortex_a8_neon_vaba): Likewise.
6069 (cortex_a8_neon_vaba_qqq): Likewise.
6070 (cortex_a8_neon_shift_1): Likewise.
6071 (cortex_a8_neon_shift_2): Likewise.
6072 (cortex_a8_neon_shift_3): Likewise.
6073 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
6074 (cortex_a8_neon_vsra_vrsra): Likewise.
6075 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6076 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
6077 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
6079 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6080 (cortex_a8_neon_mla_qqq_8_16): Likewise.
6081 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
6083 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
6084 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
6085 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
6086 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
6087 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
6088 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
6089 (cortex_a8_neon_fp_vsum): Likewise.
6090 (cortex_a8_neon_fp_vmul_ddd): Likewise.
6091 (cortex_a8_neon_fp_vmul_qqd): Likewise.
6092 (cortex_a8_neon_fp_vmla_ddd): Likewise.
6093 (cortex_a8_neon_fp_vmla_qqq): Likewise.
6094 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
6095 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
6096 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
6097 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
6098 (cortex_a8_neon_bp_simple): Likewise.
6099 (cortex_a8_neon_bp_2cycle): Likewise.
6100 (cortex_a8_neon_bp_3cycle): Likewise.
6101 (cortex_a8_neon_ldr): Likewise.
6102 (cortex_a8_neon_str): Likewise.
6103 (cortex_a8_neon_vld1_1_2_regs): Likewise.
6104 (cortex_a8_neon_vld1_3_4_regs): Likewise.
6105 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
6106 (cortex_a8_neon_vld2_4_regs): Likewise.
6107 (cortex_a8_neon_vld3_vld4): Likewise.
6108 (cortex_a8_neon_vld1_vld2_lane): Likewise.
6109 (cortex_a8_neon_vld3_vld4_lane): Likewise.
6110 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
6111 (cortex_a8_neon_vst1_3_4_regs): Likewise.
6112 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
6113 (cortex_a8_neon_vst1_vst2_lane): Likewise.
6114 (cortex_a8_neon_vst3_vst4_lane): Likewise.
6115 (cortex_a8_neon_mcr): Likewise.
6116 (cortex_a8_neon_mcr_2_mcrr): Likewise.
6117 (cortex_a8_neon_mrc): Likewise.
6118 (cortex_a8_neon_mrrc): Likewise.
6120 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
6122 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
6124 * config/aarch64/aarch64-simd.md (neon_type): Remove.
6125 (aarch64_simd_dup<mode>): Add "type" attribute.
6126 (aarch64_dup_lane<mode>): Likewise.
6127 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
6128 (*aarch64_simd_mov<mode>): Likewise.
6129 (aarch64_simd_mov_from_<mode>low): Likewise.
6130 (aarch64_simd_mov_from_<mode>high): Likewise.
6131 (orn<mode>3): Likewise.
6132 (bic<mode>3): Likewise.
6133 (add<mode>3): Likewise.
6134 (sub<mode>3): Likewise.
6135 (mul<mode>3): Likewise.
6136 (*aarch64_mul3_elt<mode>): Likewise.
6137 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
6138 (*aarch64_mul3_elt_to_128df): Likewise.
6139 (*aarch64_mul3_elt_to_64v2df): Likewise.
6140 (neg<mode>2): Likewise.
6141 (abs<mode>2): Likewise.
6142 (abd<mode>_3): Likewise.
6143 (aba<mode>_3): Likewise.
6144 (fabd<mode>_3): Likewise.
6145 (*fabd_scalar<mode>3): Likewise.
6146 (and<mode>3): Likewise.
6147 (ior<mode>3): Likewise.
6148 (xor<mode>3): Likewise.
6149 (one_cmpl<mode>2): Likewise.
6150 (aarch64_simd_vec_set<mode>): Likewise.
6151 (aarch64_simd_lshr<mode>): Likewise.
6152 (aarch64_simd_ashr<mode>): Likewise.
6153 (aarch64_simd_imm_shl<mode>): Likewise.
6154 (aarch64_simd_reg_sshl<mode): Likewise.
6155 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
6156 (aarch64_simd_reg_shl<mode>_signed): Likewise.
6157 (aarch64_simd_vec_setv2di): Likewise.
6158 (aarch64_simd_vec_set<mode>): Likewise.
6159 (aarch64_mla<mode>): Likewise.
6160 (*aarch64_mla_elt<mode>): Likewise.
6161 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
6162 (aarch64_mls<mode>): Likewise.
6163 (*aarch64_mls_elt<mode>): Likewise.
6164 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
6165 (<su><maxmin><mode>3): Likewise.
6166 (move_lo_quad_<mode>): Likewise.
6167 (aarch64_simd_move_hi_quad_<mode>): Likewise.
6168 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
6169 (vec_pack_trunc_<mode>): Likewise.
6170 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
6171 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
6172 (*aarch64_<su>mlal_lo<mode>): Likewise.
6173 (*aarch64_<su>mlal_hi<mode>): Likewise.
6174 (*aarch64_<su>mlsl_lo<mode>): Likewise.
6175 (*aarch64_<su>mlsl_hi<mode>): Likewise.
6176 (*aarch64_<su>mlal<mode>): Likewise.
6177 (*aarch64_<su>mlsl<mode>): Likewise.
6178 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
6179 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
6180 (add<mode>3): Likewise.
6181 (sub<mode>3): Likewise.
6182 (mul<mode>3): Likewise.
6183 (div<mode>3): Likewise.
6184 (neg<mode>2): Likewise.
6185 (abs<mode>2): Likewise.
6186 (fma<mode>4): Likewise.
6187 (*aarch64_fma4_elt<mode>): Likewise.
6188 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
6189 (*aarch64_fma4_elt_to_128df): Likewise.
6190 (*aarch64_fma4_elt_to_64v2df): Likewise.
6191 (fnma<mode>4): Likewise.
6192 (*aarch64_fnma4_elt<mode>): Likewise.
6193 (*aarch64_fnma4_elt_<vswap_width_name><mode>
6194 (*aarch64_fnma4_elt_to_128df): Likewise.
6195 (*aarch64_fnma4_elt_to_64v2df): Likewise.
6196 (<frint_pattern><mode>2): Likewise.
6197 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
6198 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
6199 (vec_unpacks_lo_v4sf): Likewise.
6200 (aarch64_float_extend_lo_v2df): Likewise.
6201 (vec_unpacks_hi_v4sf): Likewise.
6202 (aarch64_float_truncate_lo_v2sf): Likewise.
6203 (aarch64_float_truncate_hi_v4sf): Likewise.
6204 (aarch64_vmls<mode>): Likewise.
6205 (<su><maxmin><mode>3): Likewise.
6206 (<maxmin_uns><mode>3): Likewise.
6207 (reduc_<sur>plus_<mode>): Likewise.
6208 (reduc_<sur>plus_v2di): Likewise.
6209 (reduc_<sur>plus_v2si): Likewise.
6210 (reduc_<sur>plus_<mode>): Likewise.
6211 (aarch64_addpv4sf): Likewise.
6212 (clz<mode>2): Likewise.
6213 (reduc_<maxmin_uns>_<mode>): Likewise.
6214 (reduc_<maxmin_uns>_v2di): Likewise.
6215 (reduc_<maxmin_uns>_v2si): Likewise.
6216 (reduc_<maxmin_uns>_<mode>): Likewise.
6217 (reduc_<maxmin_uns>_v4sf): Likewise.
6218 (aarch64_simd_bsl<mode>_internal): Likewise.
6219 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
6220 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
6221 (aarch64_get_lane<mode>): Likewise.
6222 (*aarch64_combinez<mode>): Likewise.
6223 (aarch64_combine<mode>): Likewise.
6224 (aarch64_simd_combine<mode>): Likewise.
6225 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
6226 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
6227 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
6228 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
6229 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
6230 (aarch64_<sur>h<addsub><mode>): Likewise.
6231 (aarch64_<sur><addsub>hn<mode>): Likewise.
6232 (aarch64_<sur><addsub>hn2<mode>): Likewise.
6233 (aarch64_pmul<mode>): Likewise.
6234 (aarch64_<su_optab><optab><mode>): Likewise.
6235 (aarch64_<sur>qadd<mode>): Likewise.
6236 (aarch64_sqmovun<mode>): Likewise.
6237 (aarch64_<sur>qmovn<mode>): Likewise.
6238 (aarch64_s<optab><mode>): Likewise.
6239 (aarch64_sq<r>dmulh<mode>): Likewise.
6240 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
6241 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
6242 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
6243 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
6244 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
6245 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
6246 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
6247 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
6248 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
6249 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
6250 (aarch64_sqdmull<mode>): Likewise.
6251 (aarch64_sqdmull_lane<mode>_internal): Likewise.
6252 (aarch64_sqdmull_n<mode>): Likewise.
6253 (aarch64_sqdmull2<mode>_internal): Likewise.
6254 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
6255 (aarch64_sqdmull2_n<mode>_internal): Likewise.
6256 (aarch64_<sur>shl<mode>): Likewise.
6257 (aarch64_<sur>q<r>shl<mode>
6258 (aarch64_<sur>shll_n<mode>): Likewise.
6259 (aarch64_<sur>shll2_n<mode>): Likewise.
6260 (aarch64_<sur>shr_n<mode>): Likewise.
6261 (aarch64_<sur>sra_n<mode>): Likewise.
6262 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
6263 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
6264 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
6265 (aarch64_cm<optab><mode>): Likewise.
6266 (aarch64_cm<optab>di): Likewise.
6267 (aarch64_cm<optab><mode>): Likewise.
6268 (aarch64_cm<optab>di): Likewise.
6269 (aarch64_cmtst<mode>): Likewise.
6270 (aarch64_cmtstdi): Likewise.
6271 (aarch64_cm<optab><mode>): Likewise.
6272 (*aarch64_fac<optab><mode>): Likewise.
6273 (aarch64_addp<mode>): Likewise.
6274 (aarch64_addpdi): Likewise.
6275 (sqrt<mode>2): Likewise.
6276 (vec_load_lanesoi<mode>): Likewise.
6277 (vec_store_lanesoi<mode>): Likewise.
6278 (vec_load_lanesci<mode>): Likewise.
6279 (vec_store_lanesci<mode>): Likewise.
6280 (vec_load_lanesxi<mode>): Likewise.
6281 (vec_store_lanesxi<mode>): Likewise.
6282 (*aarch64_mov<mode>): Likewise.
6283 (aarch64_ld2<mode>_dreg): Likewise.
6284 (aarch64_ld2<mode>_dreg): Likewise.
6285 (aarch64_ld3<mode>_dreg): Likewise.
6286 (aarch64_ld3<mode>_dreg): Likewise.
6287 (aarch64_ld4<mode>_dreg): Likewise.
6288 (aarch64_ld4<mode>_dreg): Likewise.
6289 (aarch64_tbl1<mode>): Likewise.
6290 (aarch64_tbl2v16qi): Likewise.
6291 (aarch64_combinev16qi): Likewise.
6292 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
6293 (aarch64_st2<mode>_dreg): Likewise.
6294 (aarch64_st2<mode>_dreg): Likewise.
6295 (aarch64_st3<mode>_dreg): Likewise.
6296 (aarch64_st3<mode>_dreg): Likewise.
6297 (aarch64_st4<mode>_dreg): Likewise.
6298 (aarch64_st4<mode>_dreg): Likewise.
6299 (*aarch64_simd_ld1r<mode>): Likewise.
6300 (aarch64_frecpe<mode>): Likewise.
6301 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
6302 (aarch64_frecps<mode>): Likewise.
6304 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
6306 * config/arm/iterators.md (V_elem_ch): New.
6308 (VQH_type): Likewise.
6309 * config/arm/arm.md (is_neon_type): New.
6310 (conds): Use is_neon_type.
6311 (anddi3_insn): Update type attribute.
6312 (xordi3_insn): Likewise.
6313 (one_cmpldi2): Likewise.
6314 * config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
6315 * config/arm/neon.md (neon_mov): Update type attribute.
6316 (*movmisalign<mode>_neon_store): Likewise.
6317 (*movmisalign<mode>_neon_load): Likewise.
6318 (vec_set<mode>_internal): Likewise.
6319 (vec_set<mode>_internal): Likewise.
6320 (vec_setv2di_internal): Likewise.
6321 (vec_extract<mode>): Likewise.
6322 (vec_extract<mode>): Likewise.
6323 (vec_extractv2di): Likewise.
6324 (*add<mode>3_neon): Likewise.
6325 (adddi3_neon): Likewise.
6326 (*sub<mode>3_neon): Likewise.
6327 (subdi3_neon): Likewise.
6328 (fma<VCVTF:mode>4): Likewise.
6329 (fma<VCVTF:mode>4_intrinsic): Likewise.
6330 (*fmsub<VCVTF:mode>4): Likewise.
6331 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
6332 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
6333 (ior<mode>3): Likewise.
6334 (and<mode>3): Likewise.
6335 (orn<mode>3_neon): Likewise.
6336 (orndi3_neon): Likewise.
6337 (bic<mode>3_neon): Likewise.
6338 (bicdi3_neon): Likewise.
6339 (xor<mode>3): Likewise.
6340 (one_cmpl<mode>2): Likewise.
6341 (abs<mode>2): Likewise.
6342 (neg<mode>2): Likewise.
6343 (negdi2_neon): Likewise.
6344 (*umin<mode>3_neon): Likewise.
6345 (*umax<mode>3_neon): Likewise.
6346 (*smin<mode>3_neon): Likewise.
6347 (*smax<mode>3_neon): Likewise.
6348 (vashl<mode>3): Likewise.
6349 (vashr<mode>3_imm): Likewise.
6350 (vlshr<mode>3_imm): Likewise.
6351 (ashl<mode>3_signed): Likewise.
6352 (ashl<mode>3_unsigned): Likewise.
6353 (neon_load_count): Likewise.
6354 (ashldi3_neon_noclobber): Likewise.
6355 (ashldi3_neon): Likewise.
6356 (signed_shift_di3_neon): Likewise.
6357 (unsigned_shift_di3_neon): Likewise.
6358 (ashrdi3_neon_imm_noclobber): Likewise.
6359 (lshrdi3_neon_imm_noclobber): Likewise.
6360 (<shift>di3_neon): Likewise.
6361 (widen_ssum<mode>3): Likewise.
6362 (widen_usum<mode>3): Likewise.
6363 (quad_halves_<code>v4si): Likewise.
6364 (quad_halves_<code>v4sf): Likewise.
6365 (quad_halves_<code>v8hi): Likewise.
6366 (quad_halves_<code>v16qi): Likewise.
6367 (reduc_splus_v2di): Likewise.
6368 (neon_vpadd_internal<mode>): Likewise.
6369 (neon_vpsmin<mode>): Likewise.
6370 (neon_vpsmax<mode>): Likewise.
6371 (neon_vpumin<mode>): Likewise.
6372 (neon_vpumax<mode>): Likewise.
6373 (*ss_add<mode>_neon): Likewise.
6374 (*us_add<mode>_neon): Likewise.
6375 (*ss_sub<mode>_neon): Likewise.
6376 (*us_sub<mode>_neon): Likewise.
6377 (neon_vadd<mode>_unspec): Likewise.
6378 (neon_vaddl<mode>): Likewise.
6379 (neon_vaddw<mode>): Likewise.
6380 (neon_vhadd<mode>): Likewise.
6381 (neon_vqadd<mode>): Likewise.
6382 (neon_vaddhn<mode>): Likewise.
6383 (neon_vmul<mode>): Likewise.
6384 (neon_vfms<VCVTF:mode>): Likewise.
6385 (neon_vmlal<mode>): Likewise.
6386 (neon_vmls<mode>): Likewise.
6387 (neon_vmlsl<mode>): Likewise.
6388 (neon_vqdmulh<mode>): Likewise.
6389 (neon_vqdmlal<mode>): Likewise.
6390 (neon_vqdmlsl<mode>): Likewise.
6391 (neon_vmull<mode>): Likewise.
6392 (neon_vqdmull<mode>): Likewise.
6393 (neon_vsub<mode>_unspec): Likewise.
6394 (neon_vsubl<mode>): Likewise.
6395 (neon_vsubw<mode>): Likewise.
6396 (neon_vqsub<mode>): Likewise.
6397 (neon_vhsub<mode>): Likewise.
6398 (neon_vsubhn<mode>): Likewise.
6399 (neon_vceq<mode>): Likewise.
6400 (neon_vcge<mode>): Likewise.
6401 (neon_vcgeu<mode>): Likewise.
6402 (neon_vcgt<mode>): Likewise.
6403 (neon_vcgtu<mode>): Likewise.
6404 (neon_vcle<mode>): Likewise.
6405 (neon_vclt<mode>): Likewise.
6406 (neon_vcage<mode>): Likewise.
6407 (neon_vcagt<mode>): Likewise.
6408 (neon_vtst<mode>): Likewise.
6409 (neon_vabd<mode>): Likewise.
6410 (neon_vabdl<mode>): Likewise.
6411 (neon_vaba<mode>): Likewise.
6412 (neon_vabal<mode>): Likewise.
6413 (neon_vmax<mode>): Likewise.
6414 (neon_vmin<mode>): Likewise.
6415 (neon_vpaddl<mode>): Likewise.
6416 (neon_vpadal<mode>): Likewise.
6417 (neon_vpmax<mode>): Likewise.
6418 (neon_vpmin<mode>): Likewise.
6419 (neon_vrecps<mode>): Likewise.
6420 (neon_vrsqrts<mode>): Likewise.
6421 (neon_vqabs<mode>): Likewise.
6422 (neon_vqneg<mode>): Likewise.
6423 (neon_vcls<mode>): Likewise.
6424 (clz<mode>2): Likewise.
6425 (popcount<mode>2): Likewise.
6426 (neon_vrecpe<mode>): Likewise.
6427 (neon_vrsqrte<mode>): Likewise.
6428 (neon_vget_lane<mode>_sext_internal): Likewise.
6429 (neon_vget_lane<mode>_zext_internal): Likewise.
6430 (neon_vdup_n<mode>): Likewise.
6431 (neon_vdup_n<mode>): Likewise.
6432 (neon_vdup_nv2di): Likewise.
6433 (neon_vdup_lane<mode>_interal): Likewise.
6434 (*neon_vswp<mode>): Likewise.
6435 (neon_vcombine<mode>): Likewise.
6436 (float<mode><V_cvtto>2): Likewise.
6437 (floatuns<mode><V_cvtto>2): Likewise.
6438 (fix_trunc<mode><V_cvtto>2): Likewise.
6439 (fixuns_trunc<mode><V_cvtto>2
6440 (neon_vcvt<mode>): Likewise.
6441 (neon_vcvt<mode>): Likewise.
6442 (neon_vcvtv4sfv4hf): Likewise.
6443 (neon_vcvtv4hfv4sf): Likewise.
6444 (neon_vcvt_n<mode>): Likewise.
6445 (neon_vcvt_n<mode>): Likewise.
6446 (neon_vmovn<mode>): Likewise.
6447 (neon_vqmovn<mode>): Likewise.
6448 (neon_vqmovun<mode>): Likewise.
6449 (neon_vmovl<mode>): Likewise.
6450 (neon_vmul_lane<mode>): Likewise.
6451 (neon_vmul_lane<mode>): Likewise.
6452 (neon_vmull_lane<mode>): Likewise.
6453 (neon_vqdmull_lane<mode>): Likewise.
6454 (neon_vqdmulh_lane<mode>): Likewise.
6455 (neon_vqdmulh_lane<mode>): Likewise.
6456 (neon_vmla_lane<mode>): Likewise.
6457 (neon_vmla_lane<mode>): Likewise.
6458 (neon_vmlal_lane<mode>): Likewise.
6459 (neon_vqdmlal_lane<mode>): Likewise.
6460 (neon_vmls_lane<mode>): Likewise.
6461 (neon_vmls_lane<mode>): Likewise.
6462 (neon_vmlsl_lane<mode>): Likewise.
6463 (neon_vqdmlsl_lane<mode>): Likewise.
6464 (neon_vext<mode>): Likewise.
6465 (neon_vrev64<mode>): Likewise.
6466 (neon_vrev32<mode>): Likewise.
6467 (neon_vrev16<mode>): Likewise.
6468 (neon_vbsl<mode>_internal): Likewise.
6469 (neon_vshl<mode>): Likewise.
6470 (neon_vqshl<mode>): Likewise.
6471 (neon_vshr_n<mode>): Likewise.
6472 (neon_vshrn_n<mode>): Likewise.
6473 (neon_vqshrn_n<mode>): Likewise.
6474 (neon_vqshrun_n<mode>): Likewise.
6475 (neon_vshl_n<mode>): Likewise.
6476 (neon_vqshl_n<mode>): Likewise.
6477 (neon_vqshlu_n<mode>): Likewise.
6478 (neon_vshll_n<mode>): Likewise.
6479 (neon_vsra_n<mode>): Likewise.
6480 (neon_vsri_n<mode>): Likewise.
6481 (neon_vsli_n<mode>): Likewise.
6482 (neon_vtbl1v8qi): Likewise.
6483 (neon_vtbl2v8qi): Likewise.
6484 (neon_vtbl3v8qi): Likewise.
6485 (neon_vtbl4v8qi): Likewise.
6486 (neon_vtbl1v16qi): Likewise.
6487 (neon_vtbl2v16qi): Likewise.
6488 (neon_vcombinev16qi): Likewise.
6489 (neon_vtbx1v8qi): Likewise.
6490 (neon_vtbx2v8qi): Likewise.
6491 (neon_vtbx3v8qi): Likewise.
6492 (neon_vtbx4v8qi): Likewise.
6493 (*neon_vtrn<mode>_insn): Likewise.
6494 (*neon_vzip<mode>_insn): Likewise.
6495 (*neon_vuzp<mode>_insn): Likewise.
6496 (neon_vld1<mode>): Likewise.
6497 (neon_vld1_lane<mode>): Likewise.
6498 (neon_vld1_lane<mode>): Likewise.
6499 (neon_vld1_dup<mode>): Likewise.
6500 (neon_vld1_dup<mode>): Likewise.
6501 (neon_vld1_dupv2di): Likewise.
6502 (neon_vst1<mode>): Likewise.
6503 (neon_vst1_lane<mode>): Likewise.
6504 (neon_vst1_lane<mode>): Likewise.
6505 (neon_vld2<mode>): Likewise.
6506 (neon_vld2<mode>): Likewise.
6507 (neon_vld2_lane<mode>): Likewise.
6508 (neon_vld2_lane<mode>): Likewise.
6509 (neon_vld2_dup<mode>): Likewise.
6510 (neon_vst2<mode>): Likewise.
6511 (neon_vst2<mode>): Likewise.
6512 (neon_vst2_lane<mode>): Likewise.
6513 (neon_vst2_lane<mode>): Likewise.
6514 (neon_vld3<mode>): Likewise.
6515 (neon_vld3qa<mode>): Likewise.
6516 (neon_vld3qb<mode>): Likewise.
6517 (neon_vld3_lane<mode>): Likewise.
6518 (neon_vld3_lane<mode>): Likewise.
6519 (neon_vld3_dup<mode>): Likewise.
6520 (neon_vst3<mode>): Likewise.
6521 (neon_vst3qa<mode>): Likewise.
6522 (neon_vst3qb<mode>): Likewise.
6523 (neon_vst3_lane<mode>): Likewise.
6524 (neon_vst3_lane<mode>): Likewise.
6525 (neon_vld4<mode>): Likewise.
6526 (neon_vld4qa<mode>): Likewise.
6527 (neon_vld4qb<mode>): Likewise.
6528 (neon_vld4_lane<mode>): Likewise.
6529 (neon_vld4_lane<mode>): Likewise.
6530 (neon_vld4_dup<mode>): Likewise.
6531 (neon_vst4<mode>): Likewise.
6532 (neon_vst4qa<mode>): Likewise.
6533 (neon_vst4qb<mode>): Likewise.
6534 (neon_vst4_lane<mode>): Likewise.
6535 (neon_vst4_lane<mode>): Likewise.
6536 (neon_vec_unpack<US>_lo_<mode>): Likewise.
6537 (neon_vec_unpack<US>_hi_<mode>): Likewise.
6538 (neon_vec_<US>mult_lo_<mode>): Likewise.
6539 (neon_vec_<US>mult_hi_<mode>): Likewise.
6540 (neon_vec_<US>shiftl_<mode>): Likewise.
6541 (neon_unpack<US>_<mode>): Likewise.
6542 (neon_vec_<US>mult_<mode>): Likewise.
6543 (vec_pack_trunc_<mode>): Likewise.
6544 (neon_vec_pack_trunc_<mode>): Likewise.
6545 (neon_vabd<mode>_2): Likewise.
6546 (neon_vabd<mode>_3): Likewise.
6548 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
6550 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
6551 (load_pair): Update type attribute.
6552 (store_pair): Update type attribute.
6553 * config/aarch64/iterators.md (q): New.
6555 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
6557 * config/arm/types.md: Add new types for Neon insns.
6559 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
6560 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6561 Sergey Lega <sergey.s.lega@intel.com>
6562 Anna Tikhonova <anna.tikhonova@intel.com>
6563 Ilya Tocar <ilya.tocar@intel.com>
6564 Andrey Turetskiy <andrey.turetskiy@intel.com>
6565 Ilya Verbin <ilya.verbin@intel.com>
6566 Kirill Yukhin <kirill.yukhin@intel.com>
6567 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6569 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
6570 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
6571 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
6573 (srcp14<mode>): Ditto.
6574 (rsqrt14<mode>): Ditto.
6575 (rsqrt14<mode>): Ditto.
6576 (avx512f_vmscalef<mode>): Ditto.
6577 (avx512f_scalef<mode>): Ditto.
6578 (avx512f_getexp<mode>): Ditto.
6579 (avx512f_sgetexp<mode>): Ditto.
6580 (avx512f_fixupimm<mode>): Ditto.
6581 (avx512f_sfixupimm<mode>): Ditto.
6582 (avx512f_rndscale<mode>): Ditto.
6583 (*avx512er_exp2<mode>): Ditto.
6584 (*avx512er_rcp28<mode>): Ditto.
6585 (avx512er_rsqrt28<mode>): Ditto.
6586 (avx512f_getmant<mode>): Ditto.
6587 (avx512f_getmant<mode>): Ditto.
6588 (avx512f_rndscale<mode>): Fix formatting.
6590 2013-10-15 Martin Jambor <mjambor@suse.cz>
6592 * ipa-utils.h (ipa_edge_within_scc): Declare.
6593 * ipa-cp.c (edge_within_scc): Moved...
6594 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
6596 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
6597 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6598 Sergey Lega <sergey.s.lega@intel.com>
6599 Anna Tikhonova <anna.tikhonova@intel.com>
6600 Ilya Tocar <ilya.tocar@intel.com>
6601 Andrey Turetskiy <andrey.turetskiy@intel.com>
6602 Ilya Verbin <ilya.verbin@intel.com>
6603 Kirill Yukhin <kirill.yukhin@intel.com>
6604 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6606 * config/i386/predicates.md (const_8_to_15_operand): New.
6607 (const_16_to_31_operand): Ditto.
6608 * config/i386/sse.md (V8FI): New.
6610 (reduc_splus_v8df): Ditto.
6611 (reduc_splus_v16sf): Ditto.
6612 (avx512f_vextract<shuffletype>32x4_1): Ditto.
6613 (vec_extract_hi_<mode>): Ditto.
6614 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
6615 (vec_set_lo_<mode>): Ditto.
6616 (vec_set_hi_<mode>): Ditto.
6617 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
6618 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
6619 (avx512f_pshufd_1): Ditto.
6620 (avx512f_broadcast<mode>): Ditto.
6621 (avx512f_broadcast<mode>): Ditto.
6622 (define_split): Split vec_extract_lo into move.
6623 (ssequartermode): Ditto.
6624 (ssedoublemode): Extened with wider modes.
6625 (vec_extract_lo_<mode>): Ditto.
6627 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
6628 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6629 Sergey Lega <sergey.s.lega@intel.com>
6630 Anna Tikhonova <anna.tikhonova@intel.com>
6631 Ilya Tocar <ilya.tocar@intel.com>
6632 Andrey Turetskiy <andrey.turetskiy@intel.com>
6633 Ilya Verbin <ilya.verbin@intel.com>
6634 Kirill Yukhin <kirill.yukhin@intel.com>
6635 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6637 * config/i386/predicates.md (register_or_constm1_operand): New.
6638 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
6639 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
6640 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
6641 UNSPEC_SCATTER_PREFETCH
6643 (avx512f_ucmp<mode>3): Ditto.
6644 (avx512f_vternlog<mode>): Ditto.
6645 (avx512f_align<mode>): Ditto.
6646 (<shift_insn><mode>3): Ditto.
6647 (avx512f_<rotate>v<mode>): Ditto.
6648 (avx512f_<rotate><mode>): Ditto.
6649 (avx512f_eq<mode>3): Ditto.
6650 (avx512f_eq<mode>3_1): Ditto.
6651 (avx512f_gt<mode>3): Ditto.
6652 (avx512f_testm<mode>3): Ditto.
6653 (avx512f_testnm<mode>3): Ditto.
6654 (avx512pf_gatherpf<mode>): Ditto.
6655 (*avx512pf_gatherpf<mode>_mask): Ditto.
6656 (*avx512pf_gatherpf<mode>): Ditto.
6657 (avx512pf_scatterpf<mode>): Ditto.
6658 (*avx512pf_scatterpf<mode>_mask): Ditto.
6659 (*avx512pf_scatterpf<mode>): Ditto.
6660 (avx512f_vec_dup_gpr<mode>): Ditto.
6661 (clz<mode>2): Ditto.
6662 (conflict<mode>): Ditto.
6663 (REDUC_SMINMAX_MODE): Extened with wider modes.
6664 (reduc_<code>_<mode>): Ditto.
6665 (vlshr<mode>3): Ditto.
6666 (vashl<mode>3): Ditto.
6668 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
6669 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6670 Sergey Lega <sergey.s.lega@intel.com>
6671 Anna Tikhonova <anna.tikhonova@intel.com>
6672 Ilya Tocar <ilya.tocar@intel.com>
6673 Andrey Turetskiy <andrey.turetskiy@intel.com>
6674 Ilya Verbin <ilya.verbin@intel.com>
6675 Kirill Yukhin <kirill.yukhin@intel.com>
6676 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6678 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
6681 (avx512fmaskmode): Ditto.
6682 (bcstscalarsuff): Ditto.
6683 (avx512f_blendm<mode>): Ditto.
6684 (cmp_imm_predicate): Ditto.
6685 (avx512f_cmp<mode>3): Ditto.
6686 (avx512f_vec_dup<mode>): Ditto.
6687 (avx512f_vec_dup_mem<mode>): Ditto.
6688 (avx512f_vpermi2var<mode>3): Ditto.
6689 (avx512f_vpermt2var<mode>3): Ditto.
6690 (vec_init<mode>): Ditto.
6691 (avx512f_gathersi<mode>): Ditto.
6692 (*avx512f_gathersi<mode>): Ditto.
6693 (*avx512f_gathersi<mode>_2): Ditto.
6694 (avx512f_gatherdi<mode>): Ditto.
6695 (*avx512f_gatherdi<mode>): Ditto.
6696 (*avx512f_gatherdi<mode>_2): Ditto.
6697 (avx512f_scattersi<mode>): Ditto.
6698 (*avx512f_scattersi<mode>): Ditto.
6699 (avx512f_scatterdi<mode>): Ditto.
6700 (*avx512f_scatterdi<mode>): Ditto.
6701 (sseintprefix): Extened with wider modes.
6702 (VEC_GATHER_IDXSI): Ditto.
6703 (VEC_GATHER_IDXDI): Ditto.
6704 (VEC_GATHER_SRCDI): Ditto.
6706 2013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
6707 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6709 * config/arm/t-aprofile: New file.
6710 * config.gcc: Handle --with-multilib-list option.
6712 2013-10-15 Bernd Schmidt <bernds@codesourcery.com>
6714 * reload1.c (reloads_unique_chain_p): Ensure that r1 is
6717 2013-10-15 Richard Biener <rguenther@suse.de>
6719 * tree-loop-distribution.c (build_empty_rdg): Inline into
6721 (rdg_flag_vertex): Inline into single user.
6722 (rdg_flag_vertex_and_dependent): Likewise.
6723 (build_rdg_partition_for_vertex): Remove processed bitmap.
6724 (rdg_build_partitions): Simplify.
6726 2013-10-15 Richard Biener <rguenther@suse.de>
6728 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
6729 Restructure forwarding through conversions and copies to
6730 avoid performing copy-propagation the wrong way. Adjust
6731 recursion invocations.
6732 (forward_propagate_addr_expr): Add argument stating if we
6733 are recursing from a single-use.
6734 (ssa_forward_propagate_and_combine): Adjust.
6736 2013-10-14 David Malcolm <dmalcolm@redhat.com>
6738 * dumpfile.h (gcc::dump_manager): New class, to hold state
6739 relating to dumpfile management.
6740 (get_dump_file_name): Remove in favor of method of dump_manager.
6741 (dump_initialized_p): Likewise.
6742 (dump_start): Likewise.
6743 (dump_finish): Likewise.
6744 (dump_switch_p): Likewise.
6745 (dump_register): Likewise.
6746 (get_dump_file_info): Likewise.
6747 * context.c (gcc::context::context): Construct the dump_manager
6749 * context.h (gcc::context::get_dumps): New.
6750 (gcc::context::m_dumps): New.
6751 * coverage.c (coverage_init): Port to dump_manager API.
6752 * dumpfile.c (extra_dump_files): Convert to field of gcc::dump_manager.
6753 (extra_dump_files_in_use): Likewise.
6754 (extra_dump_files_alloced): Likewise.
6755 (gcc::dump_manager::dump_manager): New.
6756 (dump_register): Convert to...
6757 (gcc::dump_manager::dump_register): ...method, replacing
6758 function-static next_dump with m_next_dump field.
6759 (get_dump_file_info): Convert to...
6760 (gcc::dump_manager::get_dump_file_info): ...method.
6761 (get_dump_file_name): Convert to...
6762 (gcc::dump_manager::get_dump_file_name): ...method.
6763 (dump_start): Convert to...
6764 (gcc::dump_manager::dump_start): ...method.
6765 (dump_finish): Convert to...
6766 (gcc::dump_manager::dump_finish): ...method.
6767 (dump_begin): Replace body with...
6768 (gcc::dump_manager::dump_begin): ...new method.
6769 (dump_phase_enabled_p): Convert to...
6770 (gcc::dump_manager::dump_phase_enabled_p): ...method.
6771 (dump_phase_enabled_p): Convert to...
6772 (gcc::dump_manager::dump_phase_enabled_p): ...method.
6773 (dump_initialized_p): Convert to...
6774 (gcc::dump_manager::dump_initialized_p): ...method.
6775 (dump_flag_name): Replace body with...
6776 (gcc::dump_manager::dump_flag_name): ...new method.
6777 (dump_enable_all): Convert to...
6778 (gcc::dump_manager::dump_enable_all): ...new method.
6779 (opt_info_enable_passes): Convert to...
6780 (gcc::dump_manager::opt_info_enable_passes): ...new method.
6781 (dump_switch_p_1): Convert to...
6782 (gcc::dump_manager::dump_switch_p_1): ...new method.
6783 (dump_switch_p): Convert to...
6784 (gcc::dump_manager::dump_switch_p): ...new method.
6785 (opt_info_switch_p): Port to dump_manager API.
6786 (enable_rtl_dump_file): Likewise.
6787 * opts-global.c (handle_common_deferred_options): Port to new
6789 * passes.c (pass_manager::finish_optimization_passes): Likewise.
6790 (pass_manager::register_one_dump_file): Likewise.
6791 (pass_manager::register_pass): Likewise.
6792 (pass_init_dump_file): Likewise.
6793 (pass_fini_dump_file): Likewise.
6794 * statistics.c (statistics_early_init): Likewise.
6796 2013-10-14 Richard Biener <rguenther@suse.de>
6798 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
6799 iterative_hash_canonical_type, gimple_canonical_type_hash,
6800 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
6801 gimple_register_canonical_type, print_gimple_types_stats,
6802 free_gimple_type_tables): Move to lto/lto.c
6803 (gt-gimple.h): Do not include.
6804 * gimple.h (gimple_register_canonical_type,
6805 print_gimple_types_stats, free_gimple_type_tables): Remove.
6806 * Makefile.in (GTFILES): Remove gimple.c.
6808 2013-10-14 Travis Snoozy <quandary@remstate.com>
6811 * config/msp430/msp430.c (msp430_option_override): Correct thinko
6812 scanning for msp430x targets.
6814 2013-10-14 Eric Botcazou <ebotcazou@adacore.com>
6817 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
6818 (registers_ok_for_ldd_peep): Move around.
6819 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
6820 * config/sparc/sparc.md (widening peepholes): Use it.
6822 2013-10-14 Richard Biener <rguenther@suse.de>
6826 * gimple.c (iterative_hash_canonical_type): Make sure to
6827 record the hash into the correct hashtable slot.
6829 2013-10-13 Eric Botcazou <ebotcazou@adacore.com>
6831 PR rtl-optimization/58662
6832 * combine.c (try_combine): Take into account death nodes on I2 when
6833 splitting a PARALLEL of two independent SETs. Fix dump message.
6835 2013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
6838 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
6839 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
6840 * config/sh/t-sh (sh_treg_combine.o): New entry.
6841 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
6842 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
6843 (register_sh_passes): New function. Register sh_treg_combine pass.
6844 (sh_option_override): Invoke it.
6845 (sh_canonicalize_comparison): Handle op0_preserve_value.
6846 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
6847 opportunities. Canonicalize branch condition.
6848 (nott): Allow only if pseudos can be created for non-SH2A.
6850 2013-10-12 H.J. Lu <hongjiu.lu@intel.com>
6853 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
6854 (ix86_expand_movmem): Replace copy_addr_to_reg with
6855 ix86_copy_addr_to_reg.
6856 (ix86_expand_setmem): Likewise.
6858 2013-10-12 Alexander Monakov <amonakov@ispras.ru>
6860 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
6861 provided by ix86_fp_compare_mode instead of CCFPUmode.
6863 2013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
6865 * config/aarch64/arm_neon.h
6866 (vtbx<1,3>_<psu>8): Fix register constriants.
6868 2013-10-11 Jeff Law <law@redhat.com>
6870 PR tree-optimization/58640
6871 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
6872 threading paths that cross over two loop entry points.
6874 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6876 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
6877 handle vector float as well.
6878 (*vsx_le_perm_load_v4si): Likewise.
6879 (*vsx_le_perm_store_v2di): Likewise.
6880 (*vsx_le_perm_store_v4si): Likewise.
6882 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6884 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
6885 directly to circumvent subtract from splat{31} workaround.
6886 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
6888 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
6889 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
6890 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
6891 (altivec_vperm_<mode>): Convert to define_insn_and_split to
6892 separate big and little endian logic.
6893 (*altivec_vperm_<mode>_internal): New define_insn.
6894 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
6895 separate big and little endian logic.
6896 (*altivec_vperm_<mode>_uns_internal): New define_insn.
6897 (vec_permv16qi): Add little endian logic.
6899 2013-10-11 Marc Glisse <marc.glisse@inria.fr>
6901 * doc/extend.texi (returns_nonnull): Remove arguments.
6903 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
6904 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6905 Sergey Lega <sergey.s.lega@intel.com>
6906 Anna Tikhonova <anna.tikhonova@intel.com>
6907 Ilya Tocar <ilya.tocar@intel.com>
6908 Andrey Turetskiy <andrey.turetskiy@intel.com>
6909 Ilya Verbin <ilya.verbin@intel.com>
6910 Kirill Yukhin <kirill.yukhin@intel.com>
6911 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6913 * config/i386/sse.md (VI48F_256_512): New.
6914 (avx2_permvar<mode>): Change to ...
6915 (<avx2_avx512f>_permvar<mode>): This.
6917 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
6918 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6919 Sergey Lega <sergey.s.lega@intel.com>
6920 Anna Tikhonova <anna.tikhonova@intel.com>
6921 Ilya Tocar <ilya.tocar@intel.com>
6922 Andrey Turetskiy <andrey.turetskiy@intel.com>
6923 Ilya Verbin <ilya.verbin@intel.com>
6924 Kirill Yukhin <kirill.yukhin@intel.com>
6925 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6927 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
6928 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
6929 * config/i386/sse.md (VI4_AVX): New.
6930 (sf2simodelower): Ditto.
6931 (sse2_cvtps2dq): Change to ...
6932 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
6934 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
6935 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6936 Sergey Lega <sergey.s.lega@intel.com>
6937 Anna Tikhonova <anna.tikhonova@intel.com>
6938 Ilya Tocar <ilya.tocar@intel.com>
6939 Andrey Turetskiy <andrey.turetskiy@intel.com>
6940 Ilya Verbin <ilya.verbin@intel.com>
6941 Kirill Yukhin <kirill.yukhin@intel.com>
6942 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6944 * config/i386/sse.md (V_512): New.
6946 (vcond<V_512:mode><VF_512:mode>): Ditto.
6947 (vcond<V_512:mode><VI_512:mode>): Ditto.
6948 (vcondu<V_512:mode><VI_512:mode>): Ditto.
6950 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
6951 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6952 Sergey Lega <sergey.s.lega@intel.com>
6953 Anna Tikhonova <anna.tikhonova@intel.com>
6954 Ilya Tocar <ilya.tocar@intel.com>
6955 Andrey Turetskiy <andrey.turetskiy@intel.com>
6956 Ilya Verbin <ilya.verbin@intel.com>
6957 Kirill Yukhin <kirill.yukhin@intel.com>
6958 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6960 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
6961 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
6963 (fma<mode>4): Removed condition.
6964 (fms<mode>4): Ditto.
6965 (fnma<mode>4): Ditto.
6966 (fnms<mode>4): Ditto.
6967 (fma4i_fmadd_<mode>): Ditto.
6968 (*fma_fmadd_<mode>): Ditto.
6969 (*fma_fmsub_<mode>): Ditto.
6970 (*fma_fnmadd_<mode>): Ditto.
6971 (*fma_fnmsub_<mode>): Ditto.
6972 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
6973 (*fma_fmaddsub_<mode>): Ditto.
6974 (*fma_fmsubadd_<mode>): Ditto.
6975 (*fmai_fmadd_<mode>): Ditto.
6976 (*fmai_fmsub_<mode>): Ditto.
6977 (*fmai_fnmadd_<mode>): Ditto.
6978 (*fmai_fnmsub_<mode>): Ditto.
6980 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
6981 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6982 Sergey Lega <sergey.s.lega@intel.com>
6983 Anna Tikhonova <anna.tikhonova@intel.com>
6984 Ilya Tocar <ilya.tocar@intel.com>
6985 Andrey Turetskiy <andrey.turetskiy@intel.com>
6986 Ilya Verbin <ilya.verbin@intel.com>
6987 Kirill Yukhin <kirill.yukhin@intel.com>
6988 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6990 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
6991 (VI124_256): Changed to ...
6992 (VI124_256_48_512): This.
6993 (ssepackmode): Extended with wider modes.
6994 (<code><mode>3): Changed iterator.
6995 (*avx2_<code><mode>3): Ditto.
6996 (vec_pack_trunc_<mode>): Ditto.
6998 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
6999 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
7000 Sergey Lega <sergey.s.lega@intel.com>
7001 Anna Tikhonova <anna.tikhonova@intel.com>
7002 Ilya Tocar <ilya.tocar@intel.com>
7003 Andrey Turetskiy <andrey.turetskiy@intel.com>
7004 Ilya Verbin <ilya.verbin@intel.com>
7005 Kirill Yukhin <kirill.yukhin@intel.com>
7006 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
7008 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
7009 (VI8F_256_512): Ditto.
7010 (abs<mode>2): Changed iterator.
7011 (avx2_perm<mode>): Changed to ...
7012 (<avx2_avx512f>_perm<mode>): This.
7013 (avx2_perm<mode>_1): Changed to ...
7014 (<avx2_avx512f>_perm<mode>_1): This.
7016 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
7017 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
7018 Sergey Lega <sergey.s.lega@intel.com>
7019 Anna Tikhonova <anna.tikhonova@intel.com>
7020 Ilya Tocar <ilya.tocar@intel.com>
7021 Andrey Turetskiy <andrey.turetskiy@intel.com>
7022 Ilya Verbin <ilya.verbin@intel.com>
7023 Kirill Yukhin <kirill.yukhin@intel.com>
7024 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
7026 * config/i386/sse.md (VI48_AVX512F): New.
7027 (VI48_AVX2): Changed to ...
7028 (VI48_AVX2_48_AVX512F): This.
7029 (avx2_ashrv<mode>): Changed to ...
7030 (<avx2_avx512f>_ashrv<mode>): This.
7031 (avx2_<shift_insn>v<mode>): Changed to ...
7032 (<avx2_avx512f>_<shift_insn>v<mode>): This.
7034 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
7035 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
7036 Sergey Lega <sergey.s.lega@intel.com>
7037 Anna Tikhonova <anna.tikhonova@intel.com>
7038 Ilya Tocar <ilya.tocar@intel.com>
7039 Andrey Turetskiy <andrey.turetskiy@intel.com>
7040 Ilya Verbin <ilya.verbin@intel.com>
7041 Kirill Yukhin <kirill.yukhin@intel.com>
7042 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
7044 * config/i386/sse.md (VI4_AVX512F): New.
7045 (VI8_AVX2_AVX512F): Ditto.
7046 (mul<mode>3): Extended with wider modes.
7047 (*<sse4_1_avx2>_mul<mode>3): Ditto.
7048 (mul<mode>3): Ditto.
7049 (vec_widen_<s>mult_odd_<mode>): Ditto.
7051 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
7052 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
7053 Sergey Lega <sergey.s.lega@intel.com>
7054 Anna Tikhonova <anna.tikhonova@intel.com>
7055 Ilya Tocar <ilya.tocar@intel.com>
7056 Andrey Turetskiy <andrey.turetskiy@intel.com>
7057 Ilya Verbin <ilya.verbin@intel.com>
7058 Kirill Yukhin <kirill.yukhin@intel.com>
7059 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
7061 * config/i386/sse.md (VI2_AVX512F): New.
7062 (VI124_AVX512F): Ditto.
7063 (sseunpackmode): Extended with wider modes.
7064 (sseunpackfltmode): Ditto.
7065 (vec_unpacks_float_hi_<mode>): Ditto.
7066 (vec_unpacks_float_lo_<mode>): Ditto.
7067 (vec_unpacku_float_hi_<mode>): Ditto.
7068 (vec_unpacku_float_lo_<mode>): Ditto.
7069 (vec_unpacks_lo_<mode>): Ditto.
7070 (vec_unpacks_hi_<mode>): Ditto.
7071 (vec_unpacku_lo_<mode>): Ditto.
7072 (vec_unpacku_hi_<mode>): Ditto.
7074 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
7075 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
7076 Sergey Lega <sergey.s.lega@intel.com>
7077 Anna Tikhonova <anna.tikhonova@intel.com>
7078 Ilya Tocar <ilya.tocar@intel.com>
7079 Andrey Turetskiy <andrey.turetskiy@intel.com>
7080 Ilya Verbin <ilya.verbin@intel.com>
7081 Kirill Yukhin <kirill.yukhin@intel.com>
7082 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
7084 * config/i386/i386.md (multdiv): New.
7085 (multdiv_mnemonic): Ditto.
7086 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
7087 (<sse>_vm<multdiv_mnemonic><mode>3): This.
7088 (<sse>_vmdiv<mode>3): Removed.
7090 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
7091 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
7092 Sergey Lega <sergey.s.lega@intel.com>
7093 Anna Tikhonova <anna.tikhonova@intel.com>
7094 Ilya Tocar <ilya.tocar@intel.com>
7095 Andrey Turetskiy <andrey.turetskiy@intel.com>
7096 Ilya Verbin <ilya.verbin@intel.com>
7097 Kirill Yukhin <kirill.yukhin@intel.com>
7098 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
7100 * config/i386/sse.md (V): Extended with wider modes.
7102 (ssehalfvecmode): Ditto.
7104 (ssepackfltmode): Ditto.
7105 (avx_vec_concat<mode>): Ditto.
7106 (V_256_512): New iterator.
7107 (VF2_512_256): Ditto.
7108 (si2dfmode): New attribute.
7109 (si2dfmodelower): Ditto.
7111 (concat_tg_mode): Ditto.
7112 (floatv4siv4df2): Changed to ...
7113 (float<si2dfmodelower><mode>2): This.
7114 (avx_cvtps2pd256): Changed to ...
7115 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
7116 (vec_pack_trunc_v4df): Changed to ...
7117 (vec_pack_trunc_<mode>): This.
7118 (avx_vpermil<mode>): Changed to ...
7119 (<sse2_avx_avx512f>_vpermil<mode>): This.
7120 (<fixsuffix>fix_truncv8dfv8si2): New.
7121 (vec_pack_sfix_trunc_v8df): Ditto.
7122 (avx512f_rndscale<mode>): Ditto.
7123 (avx512f_roundpd512): Ditto.
7124 (vec_pack_ufix_trunc_<mode>): Updated iterator.
7126 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
7127 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
7128 Sergey Lega <sergey.s.lega@intel.com>
7129 Anna Tikhonova <anna.tikhonova@intel.com>
7130 Ilya Tocar <ilya.tocar@intel.com>
7131 Andrey Turetskiy <andrey.turetskiy@intel.com>
7132 Ilya Verbin <ilya.verbin@intel.com>
7133 Kirill Yukhin <kirill.yukhin@intel.com>
7134 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
7136 * config/i386/i386.md (any_fix): New iterator.
7137 (fixsuffix): New attribute.
7138 * config/i386/sse.md (VF1): Extened with wider modes.
7142 (sseintvecmodelower): Ditto.
7143 (ssescalarmode): Ditto.
7144 (ssescalarnum): Ditto.
7146 (ssexmmmode): Ditto.
7147 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
7148 (<sse>_rcp<mode>2): Change iterator.
7149 (rsqrt<mode>2): Ditto.
7150 (<sse>_rsqrt<mode>2): Ditto.
7151 (avx2_vec_dup<mode>): Ditto.
7152 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
7153 (round<mode>2_sfix): Ditto.
7154 (avx2_pbroadcast<mode>): Ditto.
7155 (*andnot<mode>3): Handle XI mode.
7156 (*<code><mode>3): Ditto.
7157 (AVXTOSSEMODE): Removed.
7158 (avx_vpermil<mode>): Changed to ...
7159 (<sse2_avx_avx512f>_vpermil<mode>): This.
7161 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
7162 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
7163 Sergey Lega <sergey.s.lega@intel.com>
7164 Anna Tikhonova <anna.tikhonova@intel.com>
7165 Ilya Tocar <ilya.tocar@intel.com>
7166 Andrey Turetskiy <andrey.turetskiy@intel.com>
7167 Ilya Verbin <ilya.verbin@intel.com>
7168 Kirill Yukhin <kirill.yukhin@intel.com>
7169 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
7171 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
7172 (<sse>_comi): Ditto.
7173 (<sse>_ucomi): Ditto.
7174 (sse_cvtss2siq_2): Ditto.
7175 (sse2_cvtsd2si): Ditto.
7176 (sse2_cvtsd2siq): Ditto.
7177 (sse2_cvttsd2si): Ditto.
7178 (sse2_cvttsd2siq): Ditto.
7179 (<shift_insn><mode>3): Ditto.
7180 (sse2_cvtsi2sdq): Update constraint and prefix.
7181 (sse_cvtsi2ss): Update prefix.
7182 (sse_cvtsi2ssq): Ditto.
7184 2013-10-11 Jakub Jelinek <jakub@redhat.com>
7186 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
7187 ignore internal calls.
7189 2013-10-11 Richard Biener <rguenther@suse.de>
7191 * tree-pretty-print.c (dump_generic_node): Allow to dump both (D)
7192 and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with (OVF) if
7193 TREE_OVERFLOW is set.
7195 2013-10-11 Thomas Schwinge <thomas@codesourcery.com>
7197 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
7199 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
7200 * gimple.h: Likewise.
7202 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
7204 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
7207 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
7209 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
7210 * configure: Regenerate.
7212 2013-10-11 Jakub Jelinek <jakub@redhat.com>
7214 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
7215 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
7216 formatting fixes, use pp_colon instead of pp_character (..., ':'),
7217 similarly pp_right_paren.
7218 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
7219 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
7220 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
7221 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
7223 (omp_declare_simd_clauses_equal,
7224 omp_remove_redundant_declare_simd_attrs): New functions.
7225 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
7226 (walk_tree_1): Handle new OpenMP 4.0 clauses.
7227 * tree.h (OMP_LOOP_CHECK): Define.
7228 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
7229 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
7230 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
7231 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
7232 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
7233 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
7234 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
7235 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
7236 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
7237 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
7238 OMP_CLAUSE_SIMDLEN_EXPR): Define.
7239 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
7240 (omp_remove_redundant_declare_simd_attrs): New prototype.
7241 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
7242 GIMPLE_OMP_TEAMS): New codes.
7243 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
7244 * omp-low.c (struct omp_context): Add cancel_label and cancellable
7246 (target_nesting_level): New variable.
7247 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
7248 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
7249 for collapse > 1 static schedule unless ordered.
7250 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
7251 (determine_parallel_type): Adjust get_ws_args_for caller.
7252 (install_var_field): Handle mask & 4 for double indirection.
7253 (scan_sharing_clauses): Ignore shared clause on teams construct.
7254 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
7255 (create_omp_child_function): If inside target or declare target
7256 constructs, set "omp declare target" attribute on the child function.
7257 (find_combined_for): New function.
7258 (scan_omp_parallel): Handle combined loops.
7259 (scan_omp_target, scan_omp_teams): New functions.
7260 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
7261 restrictions and set ctx->cancellable for cancellable constructs.
7262 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
7263 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
7264 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
7265 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
7266 (omp_clause_aligned_alignment): New function.
7267 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR on decls.
7268 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
7269 on teams constructs. Handle user defined reductions and new
7271 (lower_reduction_clauses): Don't set placeholder to address of ref
7272 if it has already the right type.
7273 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
7274 (expand_parallel_call): Use the new non-_start suffixed builtins,
7275 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
7276 and GOMP_parallel_end after the call.
7277 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
7278 (expand_omp_for_init_counts): Handle combined loops.
7279 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
7281 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
7282 end of cancellable loops.
7283 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
7284 Likewise. Handle collapse > 1 loops.
7285 (expand_omp_simd): Handle combined loops.
7286 (expand_omp_for): Add inner_stmt argument, adjust callers of
7287 expand_omp_for* functions, use expand_omp_for_static*chunk even
7288 for collapse > 1 unless ordered.
7289 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
7290 of cancellable sections.
7291 (expand_omp_single): Remove need_barrier variable, just rely on
7292 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
7293 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
7294 (expand_omp_atomic_load, expand_omp_atomic_store,
7295 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
7296 (expand_omp_target): New function.
7297 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
7298 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
7299 (build_omp_regions_1): Immediately close region for
7300 GF_OMP_TARGET_KIND_UPDATE.
7301 (maybe_add_implicit_barrier_cancel): New function.
7302 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
7304 (lower_omp_single): Likewise. Add clobber after the barrier.
7305 (lower_omp_taskgroup): New function.
7306 (lower_omp_for): Handle combined loops. Adjust
7307 lower_rec_input_clauses caller. Handle cancellation.
7308 (lower_depend_clauses): New function.
7309 (lower_omp_taskreg): Lower depend clauses. Adjust
7310 lower_rec_input_clauses caller. Add clobber after the call. Handle
7312 (lower_omp_target, lower_omp_teams): New functions.
7313 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
7314 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
7315 and GOMP_cancellation_point calls.
7316 (lower_omp): Fold stmts inside of target region.
7317 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
7318 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
7319 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
7320 (BT_FN_VOID_OMPFN_PTR_UINT,
7321 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
7322 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
7323 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
7324 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
7325 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
7326 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
7327 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
7328 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
7329 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
7330 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
7331 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
7332 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
7333 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
7334 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
7335 BUILT_IN_GOMP_PARALLEL_END.
7336 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
7337 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
7338 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
7339 GF_OMP_FOR_KIND_DISTRIBUTE.
7340 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
7341 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
7342 (dump_gimple_omp_return): Print lhs if it has any.
7343 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
7344 gimple_omp_atomic_seq_cst_p.
7345 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
7346 and GIMPLE_OMP_TEAMS.
7347 * langhooks.c (lhd_omp_mappable_type): New function.
7348 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
7349 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
7351 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
7352 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
7353 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and ORT_TARGET.
7354 (struct gimplify_omp_ctx): Add combined_loop field.
7355 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
7356 on stmts inside of target region.
7357 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
7358 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
7359 ORT_TARGET and ORT_TARGET_DATA.
7360 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
7362 (omp_notice_threadprivate_variable): Complain about threadprivate
7363 variables in target region.
7364 (omp_notice_variable): Complain about vars with non-mappable type
7365 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
7366 (omp_check_private): Ignore ORT_TARGET* regions.
7367 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
7368 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
7369 (find_combined_omp_for): New function.
7370 (gimplify_omp_for): Handle gimplification of combined loops.
7371 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
7373 (gimplify_omp_target_update): New function.
7374 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
7375 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
7376 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
7377 (gimplify_body): If fndecl has "omp declare target" attribute, add
7378 implicit ORT_TARGET context around it.
7379 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
7380 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
7381 * tree-nested.c (convert_nonlocal_reference_stmt,
7382 convert_local_reference_stmt, convert_gimple_call): Handle
7383 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
7384 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
7385 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
7386 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
7387 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
7388 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
7389 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
7390 BUILT_IN_GOMP_LOOP_END_CANCEL,
7391 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
7392 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
7393 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
7394 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
7395 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
7396 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
7397 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
7398 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
7399 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
7400 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
7401 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
7402 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
7403 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
7404 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
7405 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
7406 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
7407 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
7408 gimple_build_omp_teams): New functions.
7409 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
7410 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
7411 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
7412 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
7413 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
7414 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
7415 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
7416 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
7417 GF_OMP_ATOMIC_SEQ_CST): New.
7418 (gimple_build_omp_taskgroup, gimple_build_omp_target,
7419 gimple_build_omp_teams): New prototypes.
7420 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
7421 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
7422 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
7423 GIMPLE_OMP_SINGLE as end of range.
7424 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
7425 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
7426 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
7427 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
7428 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
7429 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
7430 gimple_omp_target_kind, gimple_omp_target_set_kind,
7431 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
7432 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
7433 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
7434 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
7435 gimple_omp_teams_set_clauses): New inlines.
7436 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
7437 and GIMPLE_OMP_TASKGROUP.
7438 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
7439 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
7440 enum omp_clause_proc_bind_kind): New.
7441 (union omp_clause_subcode): Add depend_kind, map_kind and
7442 proc_bind_kind fields.
7443 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
7444 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
7445 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
7446 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
7447 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
7449 2013-10-10 Teresa Johnson <tejohnson@google.com>
7451 * predict.c (tree_estimate_probability): Add new parameter
7452 for estimate_bb_frequencies.
7453 (estimate_bb_frequencies): Add new parameter to force estimation.
7454 (rebuild_frequencies): When max frequency in function is small,
7455 recompute counts from frequencies.
7456 * predict.h (estimate_bb_frequencies): New parameter.
7458 2013-10-10 David Malcolm <dmalcolm@redhat.com>
7460 * ipa-inline.c (ipa_inline): Fix leak of "order" when
7461 optimizations are disabled.
7463 2013-10-10 David Malcolm <dmalcolm@redhat.com>
7465 * coverage.c (coverage_finish): Fix leak of da_file_name.
7467 2013-10-10 Jan Hubicka <jh@suse.cz>
7469 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
7470 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
7473 2013-10-10 Jakub Jelinek <jakub@redhat.com>
7476 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
7477 if any labels are in FALLTHRU_BB, use a special label emitted
7478 immediately after the asm goto insn rather than label_rtx
7480 (expand_asm_stmt): Adjust caller.
7481 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
7482 edge if the last insn in predecessor is a jump with single successor,
7483 but it isn't simplejump_p.
7485 2013-10-10 Richard Biener <rguenther@suse.de>
7487 PR tree-optimization/58656
7488 * tree-ssa-pre.c (phi_translate): Do not cache failed translations.
7490 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
7492 * gimplify.c: Include expr.h and tm_p.h for targets with special
7493 va-arg padding requirements.
7495 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
7497 * tree-flow.h: Move some prototypes to gimple.h.
7498 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
7499 * gimple.h: Relocate some prototypes from tree-flow.h
7500 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
7502 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
7503 (build_va_arg_indirect_ref): Relocate and make static.
7504 (std_gimplify_va_arg_expr): Relocate here.
7505 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
7506 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
7508 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7510 * doc/md.texi: Document the mnemonic attribute.
7512 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7515 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
7516 (set_attr_alternative x ...) when searching for user defined
7519 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
7521 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
7523 2013-10-09 Easwaran Raman <eraman@google.com>
7525 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
7526 * cfgexpand.c (defer_stack_allocation): ...use here
7527 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
7529 2013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
7531 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
7532 (operand_equal_for_value_replacement): New function, extracted from
7533 value_replacement and enhanced to catch more cases.
7534 (value_replacement): Use operand_equal_for_value_replacement.
7536 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
7538 * loop-doloop.c (doloop_modify, doloop_optimize): Use
7539 get_max_loop_iterations.
7541 2013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7543 * config/arm/aarch-common.c (arm_early_load_addr_dep):
7544 Place comment above function.
7546 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
7548 * tree-flow.h: Remove all remaining prototypes, enums and structs that
7549 are not related to tree-cfg.c.
7550 * tree-ssa-address.h: New file. Relocate prototypes.
7551 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
7552 (addr_for_mem_ref): New. Combine call to get_address_description and
7553 return addr_for_mem_ref.
7554 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
7555 * tree-ssa-live.h: Adjust prototypes.
7556 * passes.c: Include tree-ssa-live.h.
7557 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
7558 * graphite.c (graphite_transform_loops): Make static.
7559 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
7560 make_pass_graphite, pass_data_graphite_transforms,
7561 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
7562 * ipa-pure-const.c (warn_function_noreturn): Make static.
7563 (execute_warn_function_noreturn, gate_warn_function_noreturn,
7564 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
7565 Relocate from tree-cfg.c
7566 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
7568 (execute_warn_function_noreturn, gate_warn_function_noreturn,
7569 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
7570 Move to ipa-pure-const.c.
7571 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
7572 Relocate from tree-optimize.c.
7573 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
7574 make_pass_fixup_cfg): Move to tree-cfg.c.
7575 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
7576 Relocate some prototypes.
7577 * tree-data-ref.h (tree_check_data_deps) Add prototype.
7578 * tree-dump.c (dump_function_to_file): Remove prototype.
7579 Add tree-flow.h to the include file.
7580 * tree-dump.h: Remove prototype.
7581 * tree-parloops.h: New File. Add prototypes.
7582 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
7583 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
7584 from tree-ssa-loop.c.
7585 * tree-predcom.c (run_tree_predictive_commoning,
7586 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
7587 Relocate here from tree-ssa-loop.c.
7588 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
7589 ssa_name_values.release ().
7590 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
7591 (ssa_name_values): Relocate from tree-flow.h.
7592 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
7593 * tree-ssa-loop.c (run_tree_predictive_commoning,
7594 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
7595 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
7596 make_pass_graphite, pass_data_graphite_transforms,
7597 make_pass_graphite_transforms, gate_tree_parallelize_loops,
7598 tree_parallelize_loops, pass_data_parallelize_loops,
7599 make_pass_parallelize_loops): Move to other files.
7600 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
7602 * tree.h: Remove prototypes from tree-address.c.
7604 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
7606 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
7607 (struct int_tree_map): Move to tree-hasher.h
7608 (SCALE, LABEL, PERCENT): Move to gimple.h
7609 * tree-flow-inline.h: Delete. Move functions to other files.
7610 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
7611 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
7612 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
7613 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
7614 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
7615 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
7616 * tree-hasher.h: Don't include tree-flow.h.
7617 (struct int_tree_map): Relocate from tree-flow.h.
7618 * tree-sra.c (contains_view_convert_expr_p): Relocate from
7619 tree-flow-inline.h and make static.
7620 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
7622 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
7623 tree-flow-inline.h and make static.
7624 * tree.h (is_global_var, may_be_aliased): Relocate from
7626 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
7627 * value-prof.c: No longer include tree-flow-inline.h.
7628 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
7630 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
7632 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
7633 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
7634 (enum move_pos): Move to tree-ssa-loop-im.h
7635 * cfgloop.h: Move some prototypes.
7636 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
7637 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
7638 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
7639 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
7640 (loop_containing_stmt): Relocate from tree-flow-inline.h.
7641 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
7642 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
7643 (enum move_pos): Relocate here.
7644 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
7646 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
7647 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
7648 make_pass_lim): Relocate here from tree-ssa-loop.c.
7649 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
7651 (loop_edge_to_cancel, unloop_loops): Make static.
7652 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
7653 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
7654 (tree_complete_unroll, gate_tree_complete_unroll,
7655 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
7656 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
7657 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
7658 * tree-ssa-loop-ivopts.c: Remove local prototypes.
7659 (stmt_invariant_in_loop_p): Remove unused function.
7660 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
7661 * tree-ssa-loop-manip.h: New file. Add prototypes.
7662 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
7663 (gcov_type_to_double_int): Move to cfgloop.h.
7664 (double_int_cmp, bound_index,
7665 estimate_numbers_of_iterations_loop): Make static.
7666 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
7667 (max_loop_iterations): Factor out get_max_loop_iterations.
7668 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
7670 * tree-ssa-loop-niter.h: New file. Add prototypes.
7671 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
7672 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
7673 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
7674 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
7675 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
7676 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
7677 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
7678 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
7679 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
7680 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
7681 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
7682 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
7683 pass_data_complete_unroll, make_pass_complete_unroll,
7684 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
7685 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
7686 tree-ssa-loop-ivcanon.c.
7687 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
7688 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
7689 tree-ssa-loop-prefetch.c.
7690 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
7692 (get_lsm_tmp_name): Relocate and add suffix parameter.
7693 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
7694 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
7695 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
7696 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
7697 (get_estimated_loop_iterations): Factor out accessor from
7698 estimated_loop_iterations in tree-ssa-loop-niter.c.
7699 (get_max_loop_iterations): Factor out accessor from
7700 _max_loop_iterations in tree-ssa-niter.c.
7701 * loop-unroll.c (decide_unroll_constant_iterations,
7702 decide_unroll_runtime_iterations, decide_peel_simple,
7703 decide_unroll_stupid): Use new get_* accessors.
7705 2013-10-09 Marc Glisse <marc.glisse@inria.fr>
7707 PR tree-optimization/20318
7708 * doc/extend.texi (returns_nonnull): New function attribute.
7709 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
7711 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
7712 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
7714 2013-10-09 Eric Botcazou <ebotcazou@adacore.com>
7717 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
7718 false if both components are bitfields.
7720 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
7722 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
7723 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
7724 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
7725 * config/aarch64/aarch64.h
7726 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
7727 * config/aarch64/aarch64-simd-builtins.def
7728 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
7730 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
7732 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
7734 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
7736 * config/aarch64/arm_neon.h (vdiv_f64): Added.
7738 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
7740 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
7741 (vneg_f64): New intrinsic.
7742 (vneg_s8): Asm replaced with C.
7743 (vneg_s16): Likewise.
7744 (vneg_s32): Likewise.
7745 (vneg_s64): New intrinsic.
7746 (vnegq_f32): Asm replaced with C.
7747 (vnegq_f64): Likewise.
7748 (vnegq_s8): Likewise.
7749 (vnegq_s16): Likewise.
7750 (vnegq_s32): Likewise.
7751 (vnegq_s64): Likewise.
7753 2013-10-09 Renlin Li <Renlin.Li@arm.com>
7755 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
7757 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7759 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
7760 packed stack special handling.
7761 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
7762 back to fixed stack slots for FPRs saved due to stdarg.
7764 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7766 * config/s390/s390.c (s390_frame_info): Restructure function.
7768 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7770 * config/s390/s390.c (struct s390_frame_layout): New field
7772 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
7773 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
7774 regs_ever_clobbered to char*.
7775 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
7776 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
7777 (s390_register_info): Enable FPR save slots. Move/Copy some
7778 functionality into ...
7779 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
7780 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
7782 (s390_frame_info): Do gpr slot allocation here now. stdarg does
7783 not imply a stack frame.
7784 (s390_init_frame_layout): Remove variable clobbered_regs.
7785 (s390_update_register_info): Remove function.
7786 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
7787 cannot be used for register renaming.
7788 (s390_hard_regno_scratch_ok): New function.
7789 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
7790 (s390_initial_elimination_offset): Change offset calculation of
7791 the return address pointer.
7792 (save_gprs): Deal with only r6 being saved from the call-saved regs.
7793 (restore_gprs): Set frame related flag.
7794 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
7795 (s390_emit_prologue): Call s390_register_info instead of
7796 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
7797 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
7798 (s390_optimize_prologue): Call s390_optimize_register_info.
7799 Try to remove also FPR slot save/restore INSNs. Remove frame
7800 related flags from restore INSNs.
7802 2013-10-08 DJ Delorie <dj@redhat.com>
7804 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
7807 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
7808 avoid conflict with the MI use of %c.
7809 * config/rl78/rl78-real.md: change %c to %C throughout.
7810 * config/rl78/rl78-virt.md: Likewise.
7812 2013-10-08 Jan Hubicka <jh@suse.cz>
7814 * config/i386/i386.c (ix86_option_override_internal): Switch
7815 to SSE math for -ffast-math when target ISA supports SSE2.
7817 2013-10-08 Andrew MacLeod <amacleod@redhat.com>
7819 * tree-flow.h: Remove some prototypes.
7820 * tree.h: Remove some protypes, add a couple.
7821 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
7822 using_eh_for_cleanups_p): Add interface routines for front ends.
7823 * tree-eh.h: New file. Add protoptyes.
7824 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
7825 (add_stmt_to_eh_lp_fn): Make static.
7826 (lower_try_finally): Use new using_eh_for_cleanups_p.
7827 * emit-rtl.c: Include tree-eh.h.
7828 * gimple.h: Include tree-eh.h.
7830 2013-10-08 Marc Glisse <marc.glisse@inria.fr>
7832 PR tree-optimization/58480
7833 * tree-vrp.c (infer_nonnull_range): New function.
7834 (infer_value_range): Call infer_nonnull_range.
7836 2013-10-08 Dehao Chen <dehao@google.com>
7838 PR tree-optimization/58619
7839 * tree-inline.c (copy_phis_for_bb): Combine location data
7842 2013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
7845 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
7846 RTX_FRAME_RELATED_P on INSN.
7848 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7850 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
7851 (altivec_expand_vec_perm_const): Call it.
7853 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7855 * config/rs6000/vector.md (mov<mode>): Emit permuted move
7856 sequences for LE VSX loads and stores at expand time.
7857 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
7859 * config/rs6000/rs6000.c (rs6000_const_vec): New.
7860 (rs6000_gen_le_vsx_permute): New.
7861 (rs6000_gen_le_vsx_load): New.
7862 (rs6000_gen_le_vsx_store): New.
7863 (rs6000_gen_le_vsx_move): New.
7864 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
7865 (*vsx_le_perm_load_v4si): New.
7866 (*vsx_le_perm_load_v8hi): New.
7867 (*vsx_le_perm_load_v16qi): New.
7868 (*vsx_le_perm_store_v2di): New.
7869 (*vsx_le_perm_store_v4si): New.
7870 (*vsx_le_perm_store_v8hi): New.
7871 (*vsx_le_perm_store_v16qi): New.
7872 (*vsx_xxpermdi2_le_<mode>): New.
7873 (*vsx_xxpermdi4_le_<mode>): New.
7874 (*vsx_xxpermdi8_le_V8HI): New.
7875 (*vsx_xxpermdi16_le_V16QI): New.
7876 (*vsx_lxvd2x2_le_<mode>): New.
7877 (*vsx_lxvd2x4_le_<mode>): New.
7878 (*vsx_lxvd2x8_le_V8HI): New.
7879 (*vsx_lxvd2x16_le_V16QI): New.
7880 (*vsx_stxvd2x2_le_<mode>): New.
7881 (*vsx_stxvd2x4_le_<mode>): New.
7882 (*vsx_stxvd2x8_le_V8HI): New.
7883 (*vsx_stxvd2x16_le_V16QI): New.
7885 2013-10-07 Renlin Li <Renlin.Li@arm.com>
7887 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
7889 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7891 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
7892 loop to work also for 31bit ABI.
7893 Save the stack pointer for frame_size > 0.
7895 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7897 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
7898 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
7899 constraint letters from expanders.
7900 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
7901 retry count to general_operand.
7902 ("tabort"): Give operand 0 a mode.
7903 ("tabort_1"): Add mode and constraint letter for operand 0.
7904 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
7906 2013-10-04 Jeff Law <law@redhat.com>
7908 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
7910 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
7912 (thread_across_edge): Here. Call it.
7914 2013-10-04 Cary Coutant <ccoutant@google.com>
7916 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
7917 discarding a location list expression (or a piece of one).
7919 2013-10-03 Jan Hubicka <jh@suse.cz>
7921 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
7922 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
7923 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
7925 2013-10-03 Jan Hubicka <jh@suse.cz>
7927 * config/i386/i386.c (ix86_option_override_internal): Do not enable
7928 accumulate-outgoing-args when producing unwind info.
7930 2013-10-03 Wei Mi <wmi@google.com>
7932 * lra-constraints.c (insert_move_for_subreg): New function
7933 extracted from simplify_operand_subreg.
7934 (simplify_operand_subreg): Add reload for paradoxical subreg.
7936 2013-10-03 Rong Xu <xur@google.com>
7938 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
7939 the candidate of builtin_expect such that we should fix the
7940 size/time estimation.
7941 (estimate_function_body_sizes): Do the acutally size/time fix-up
7944 2013-10-03 Rong Xu <xur@google.com>
7946 * predict.c (tree_predict_by_opcode): Get the probability
7947 for builtin_expect from param builtin_expect_probability.
7948 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
7949 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
7950 * doc/invoke.texi: Add documentation for builtin-expect-probability.
7952 2013-10-03 Marc Glisse <marc.glisse@inria.fr>
7955 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
7956 * calls.c (alloca_call_p): Use get_callee_fndecl.
7957 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
7958 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
7960 (vrp_visit_stmt): Remove duplicated code.
7962 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
7964 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
7965 ceildf2, btruncdf2, instead of vsx_* name.
7967 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
7968 iterators to only do V2DF and V4SF here. Move the DF code to
7969 rs6000.md where it is combined with SF mode. Replace <VSv> with
7970 just 'v' since only vector operations are handled with these insns
7971 after moving the DF support to rs6000.md.
7972 (vsx_sub<mode>3): Likewise.
7973 (vsx_mul<mode>3): Likewise.
7974 (vsx_div<mode>3): Likewise.
7975 (vsx_fre<mode>2): Likewise.
7976 (vsx_neg<mode>2): Likewise.
7977 (vsx_abs<mode>2): Likewise.
7978 (vsx_nabs<mode>2): Likewise.
7979 (vsx_smax<mode>3): Likewise.
7980 (vsx_smin<mode>3): Likewise.
7981 (vsx_sqrt<mode>2): Likewise.
7982 (vsx_rsqrte<mode>2): Likewise.
7983 (vsx_fms<mode>4): Likewise.
7984 (vsx_nfma<mode>4): Likewise.
7985 (vsx_copysign<mode>3): Likewise.
7986 (vsx_btrunc<mode>2): Likewise.
7987 (vsx_floor<mode>2): Likewise.
7988 (vsx_ceil<mode>2): Likewise.
7989 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
7990 (vsx_sminsf3): Likewise.
7991 (vsx_fmadf4): Likewise.
7992 (vsx_fmsdf4): Likewise.
7993 (vsx_nfmadf4): Likewise.
7994 (vsx_nfmsdf4): Likewise.
7995 (vsx_cmpdf_internal1): Likewise.
7997 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
7998 simpler to select whether a target has SPE or traditional floating
7999 point support in iterators.
8000 (TARGET_DF_SPE): Likewise.
8001 (TARGET_SF_FPR): Likewise.
8002 (TARGET_DF_FPR): Likewise.
8003 (TARGET_SF_INSN): Macros to say whether floating point support
8004 exists for a given operation for expanders.
8005 (TARGET_DF_INSN): Likewise.
8007 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
8008 combining of SF/DF mode operations, using both traditional and VSX
8016 (abs<mode>2): Combine SF/DF modes using traditional floating point
8017 instructions. Add support for using the upper DF registers with
8018 VSX support, and SF registers with power8-vector support. Update
8019 expanders for operations supported by both the SPE and traditional
8020 floating point units.
8021 (abs<mode>2_fpr): Likewise.
8022 (nabs<mode>2): Likewise.
8023 (nabs<mode>2_fpr): Likewise.
8024 (neg<mode>2): Likewise.
8025 (neg<mode>2_fpr): Likewise.
8026 (add<mode>3): Likewise.
8027 (add<mode>3_fpr): Likewise.
8028 (sub<mode>3): Likewise.
8029 (sub<mode>3_fpr): Likewise.
8030 (mul<mode>3): Likewise.
8031 (mul<mode>3_fpr): Likewise.
8032 (div<mode>3): Likewise.
8033 (div<mode>3_fpr): Likewise.
8034 (sqrt<mode>3): Likewise.
8035 (sqrt<mode>3_fpr): Likewise.
8036 (fre<Fs>): Likewise.
8037 (rsqrt<mode>2): Likewise.
8038 (cmp<mode>_fpr): Likewise.
8039 (smax<mode>3): Likewise.
8040 (smin<mode>3): Likewise.
8041 (smax<mode>3_vsx): Likewise.
8042 (smin<mode>3_vsx): Likewise.
8043 (negsf2): Delete SF operations that are merged with DF.
8050 (fmasf4_fpr): Likewise.
8051 (fmssf4_fpr): Likewise.
8052 (nfmasf4_fpr): Likewise.
8053 (nfmssf4_fpr): Likewise.
8054 (sqrtsf2): Likewise.
8055 (rsqrtsf_internal1): Likewise.
8056 (smaxsf3): Likewise.
8057 (sminsf3): Likewise.
8058 (cmpsf_internal1): Likewise.
8059 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
8060 (negdf2): Delete DF operations that are merged with SF.
8062 (nabsdf2): Likewise.
8068 (rsqrtdf_internal1): Likewise.
8069 (fmadf4_fpr): Likewise.
8070 (fmsdf4_fpr): Likewise.
8071 (nfmadf4_fpr): Likewise.
8072 (nfmsdf4_fpr): Likewise.
8073 (sqrtdf2): Likewise.
8074 (smaxdf3): Likewise.
8075 (smindf3): Likewise.
8076 (cmpdf_internal1): Likewise.
8077 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
8078 (btrunc<mode>2): Delete separate expander, and combine with the
8079 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
8080 (btrunc<mode>2_fpr): Likewise.
8081 (ceil<mode>2): Likewise.
8082 (ceil<mode>2_fpr): Likewise.
8083 (floor<mode>2): Likewise.
8084 (floor<mode>2_fpr): Likewise.
8085 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
8086 Add support for using the upper registers with VSX and
8087 power8-vector. Move insns to be closer to the define_expands. On
8088 VSX systems, prefer the traditional form of FMA over the VSX
8089 version, since the traditional form allows the target not to
8090 overlap with the inputs.
8091 (fms<mode>4_fpr): Likewise.
8092 (nfma<mode>4_fpr): Likewise.
8093 (nfms<mode>4_fpr): Likewise.
8095 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8096 Richard Earnshaw <richard.earnshaw@arm.com>
8098 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
8099 (struct mult_cost_table): Likewise.
8100 (struct mem_cost_table): Likewise.
8101 (struct fp_cost_table): Likewise.
8102 (struct vector_cost_table): Likewise.
8103 (cpu_cost_table): Likewise.
8104 * config/arm/arm.opt (mold-rts-costs): New option.
8105 (mnew-generic-costs): Likewise.
8106 * config/arm/arm.c (generic_extra_costs): New table.
8107 (cortexa15_extra_costs): Likewise.
8108 (arm_slowmul_tune): Use NULL as new costs.
8109 (arm_fastmul_tune): Likewise.
8110 (arm_strongarm_tune): Likewise.
8111 (arm_xscale_tune): Likewise.
8112 (arm_9e_tune): Likewise.
8113 (arm_v6t2_tune): Likewise.
8114 (arm_cortex_a5_tune): Likewise.
8115 (arm_cortex_a9_tune): Likewise.
8116 (arm_v6m_tune): Likewise.
8117 (arm_fa726te_tune): Likewise.
8118 (arm_cortex_a15_tune): Use cortex15_extra_costs.
8119 (arm_cortex_tune): Use generict_extra_costs.
8120 (shifter_op_p): New function.
8121 (arm_unspec_cost): Likewise.
8122 (LIBCALL_COST): Define.
8123 (arm_new_rtx_costs): New function.
8124 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
8125 table is available. Use old costs otherwise unless mnew-generic-costs
8127 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
8128 (cpu_cost_table): Declare.
8130 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
8133 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
8134 (*subs_mul_imm_<mode>)
8135 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
8136 (*sub_<shift>_<mode>)
8137 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
8138 Remove k constraint.
8140 2013-10-03 Ian Bolton <ian.bolton@arm.com>
8142 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
8144 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
8146 2013-10-02 Teresa Johnson <tejohnson@google.com>
8148 * predict.c (probably_never_executed): New function.
8149 (probably_never_executed_bb_p): Invoke probably_never_executed.
8150 (probably_never_executed_edge_p): Ditto.
8151 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
8152 Treat profile insanities conservatively.
8154 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
8156 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
8158 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
8160 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
8161 use it. Use smaller increase for scratch. Don't increase reject
8162 for early clobber scratch.
8163 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
8164 setting eliminated regs except setting fp from hfp.
8165 (lra_eliminate): Check lra_insn_recog_data on NULL.
8167 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
8170 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
8171 setting -mvsx-timode by default until the underlying problem is fixed.
8172 (RS6000_CPU, power7 defaults): Likewise.
8174 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
8176 * config/x-linux (host-linux.o): Remove header dependencies.
8177 Use $(COMPILE) and $(POSTCOMPILE).
8178 * config/t-linux-android (linux-android.o): Ditto.
8180 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
8182 * Makefile.in (expmed.o-warn): Remove.
8184 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
8186 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
8187 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
8189 2013-10-02 Teresa Johnson <tejohnson@google.com>
8191 * dojump.c (do_jump_1): Divide probability between
8192 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
8194 2013-10-02 Tom Tromey <tromey@redhat.com>
8196 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
8198 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
8200 * tree-flow.h: Remove some prototypes.
8201 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
8202 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
8203 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
8204 mark_virtual_phi_result_for_renaming): Relocate here.
8205 * tree-into-ssa.h: Add prototypes.
8206 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
8207 single_pred_before_succ_order.
8208 (blocks_in_phiopt_order): Rename and move to cfganal.c.
8209 (nonfreeing_call_p) Move to gimple.c.
8210 * cfganal.c (single_pred_before_succ_order): Move and renamed from
8212 * basic-block.h (single_pred_before_succ_order): Add prototype.
8213 * gimple.c (nonfreeing_call_p): Relocate here.
8214 * gimple.h: Add prototype.
8215 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
8216 * tree-ssa-dom.h: New file. Relocate prototypes here.
8217 * tree-ssa.h: Include tree-ssa-dom.h.
8219 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
8221 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
8222 Use $(COMPILE) and $(POSTCOMPILE).
8224 * config/alpha/x-alpha (driver-alpha.o): Ditto.
8226 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
8228 * tree-flow.h: Remove some prototypes.
8229 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
8230 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
8231 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
8232 propagate_tree_value*): Move from here to...
8233 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
8234 propagate_tree_value*): Relocate here.
8235 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
8236 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
8237 * gimple-fold.c: Remove gimple-fold.h from include list.
8238 * tree-vrp.c: Remove gimple-fold.h from include list.
8239 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
8240 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
8241 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
8242 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
8243 * sese.c: Add tree-ssa-propagate.h to include list.
8245 2013-10-02 Richard Biener <rguenther@suse.de>
8247 * tree-loop-distribution.c: Include tree-vectorizer.h for
8249 (enum partition_kind): Remove PKIND_REDUCTION.
8250 (struct partition_s): Remove has_writes member, add reduction_p member.
8251 (partition_alloc): Adjust.
8252 (partition_builtin_p): Likewise.
8253 (partition_has_writes): Remove.
8254 (partition_reduction_p): New function.
8255 (partition_merge_into): Likewise.
8256 (generate_code_for_partition): Commonize builtin partition
8258 (rdg_cannot_recompute_vertex_p): Remove.
8259 (already_processed_vertex_p): Likewise.
8260 (rdg_flag_vertex): Do not set has_writes.
8261 (classify_partition): Adjust.
8262 (rdg_build_partitions): Do not set has_writes, treat all
8263 partitions as useful.
8264 (distribute_loop): Record number of library calls generated. Adjust.
8265 (tree_loop_distribution): Report number of loops and library
8266 calls generated as opt-info.
8268 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
8270 * tree-flow.h: Include new .h files. Move prototypes.
8271 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
8272 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
8273 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
8274 * tree-pretty-print.h: Add prototypes from tree-flow.h.
8275 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
8276 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
8277 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
8278 * tree.h (get_ref_base_and_extent): Move prototype out.
8279 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
8281 * gimple-low.h: New File. Add prototypes from tree-flow.h.
8282 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
8283 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
8284 * tree-scalar-evolution.c: Include tree.h.
8285 * sese.c: Include tree.h.
8286 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
8287 * dwarf2out.c: Include tree-dfa.h.
8288 * tree-chrec.c: Include tree.h.
8289 * tree-data-ref.c: Include tree.h.
8291 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
8293 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
8296 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8298 * config/t-sol2 (sol2-c.o): Remove header dependencies.
8299 Use $(COMPILE) and $(POSTCOMPILE).
8300 (sol2-cxx.o): Likewise.
8301 (sol2-stubs.o): Likewise.
8303 * config/x-solaris (host-solaris.o): Likewise.
8305 * config/sparc/t-sparc (sparc.o): Remove.
8306 (sparc-c.o): Remove header dependencies.
8307 Use $(COMPILE) and $(POSTCOMPILE).
8308 * config/sparc/x-sparc: Likewise.
8310 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
8312 * config/arc/arc-opts.h: Add 2013 to Copyright years.
8313 * config/arc/arc700.md: Likewise.
8314 * config/arc/arc-modes.def: Likewise.
8315 * config/arc/arc-simd.h: Likewise.
8316 * config/arc/t-arc-uClibc: Likewise.
8317 * config/arc/t-arc-newlib: Likewise.
8319 2013-10-02 Renlin Li <renlin.li@arm.com>
8321 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
8323 (aarch64_expand_epilogue): Likewise.
8325 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8326 Yufeng Zhang <yufeng.zhang@arm.com>
8328 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
8330 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
8331 'base_in' represent a conversion and legal_cast_p_1 holds; set
8332 'base_in' with the returned value from get_unwidened.
8334 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8336 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
8337 plus_constant is not used.
8339 2013-10-01 Wei Mi <wmi@google.com>
8341 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
8342 * config/i386/i386.md: Add define_peephole2 to
8343 break partial reg stall for cvtss2sd/cvtsd2ss.
8345 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
8347 * config/arc/arc.c (pass_arc_ifcvt::clone):
8348 Update for ctxt_ -> m_ctxt change.
8350 2013-10-01 Jeff Law <law@redhat.com>
8352 * tree-ssa-threadupdate.c (struct redirection_data): Delete
8353 outgoing_edge and intermediate_edge fields. Instead store the path.
8354 (redirection_data::hash): Hash on the last edge's destination index.
8355 (redirection_data::equal): Check the entire thread path.
8356 (lookup_redirectio_data): Corresponding changes.
8357 (create_edge_and_update_destination_phis): Likewise.
8358 (thread_single_edge): Likewise.
8360 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
8361 Diego Novillo <dnovillo@google.com>
8363 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
8364 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
8365 (vld32wh_insn, vld32wl_insn): Delete commented-out old
8366 versions of these patterns.
8368 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
8369 (__builtin_arc_aligned): Likewise.
8371 * config/arc/arc.md: Expand adc_0 comment stating the intended
8372 purpose and why it isn't ready.
8373 Replace commented out call_value_via_label_mixed with a
8374 plain comment about bl_s.
8376 * config/arc/arc.c (stdio.h): Don't include directly.
8377 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
8378 Remove if (1) condition.
8379 (arc_encode_section_info): Fix comment.
8381 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
8383 * config/arc/arc.c (arc_conditional_register_usage):
8384 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
8385 Also set reg_alloc_order for DMA config regs.
8387 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
8388 Jeremy Bennett <jeremy.bennett@embecosm.com>
8390 * doc/install.texi (--with-cpu): Mention ARC.
8391 (arc-*-elf32): New paragraph.
8392 (arc-linux-uclibc): Likewise.
8393 * doc/md.texi (Machine Constraints): Add ARC part.
8394 * doc/invoke.texi: (menu): Add ARC Options.
8395 (Machine Dependent Options) <ARC Options>: Add synopsis.
8396 (node ARC Options): Add.
8397 * doc/extend.texi (long_call / short_call attribute): Add ARC.
8398 (ARC Built-in Functions): New section defining
8399 generic ARC built-in functions.
8400 (ARC SIMD Built-in Functions): New section defining SIMD specific
8402 (Declaring Attributes of Functions): Extended
8403 description of short_call and long_call attributes for ARC and
8404 added index entries.
8406 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
8407 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
8408 Joern Rennecke <joern.rennecke@embecosm.com>
8409 Muhammad Khurram Riaz <khurram.riaz@arc.com>
8410 Brendan Kehoe <brendan@zen.org>
8411 Michael Eager <eager@eagercon.com>
8412 Simon Cook <simon.cook@embecosm.com>
8413 Jeremy Bennett <jeremy.bennett@embecosm.com>
8415 * config/arc, common/config/arc: New directories.
8417 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
8418 Brendan Kehoe <brendan@zen.org>
8419 Simon Cook <simon.cook@embecosm.com>
8421 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
8423 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
8425 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
8426 * tree-ssa-coalesce.h: New. Move prototype to here.
8427 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
8428 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
8429 (gimple_can_coalesce_p): Move to...
8430 * gimple.c (gimple_can_coalesce_p): Here.
8432 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
8434 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
8435 (dump_decl_set): Move to gimple.c.
8436 * gimple.h: Don't include tree-ssa-operands.h.
8437 (dump_decl_set): Add prototype.
8438 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
8439 Move to gimple-ssa.h.
8440 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
8441 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
8442 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
8443 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
8444 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
8446 (dump_decl_set): Relocate here.
8447 * gimple-ssa.h: New file.
8448 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
8449 Relocate from gimple.h.
8450 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
8451 * tree-ssa-operands.c (swap_ssa_operands): Rename from
8452 swap_tree_operands and remove non-ssa path.
8453 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
8454 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
8456 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
8457 vect_is_simple_reduction_1): Use swap_ssa_operands.
8458 * tree-flow.h: Move various prototypes to tree-phinodes.h.
8459 (enum need_phi_state): Move to tree-into-ssa.c.
8460 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
8461 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
8462 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
8463 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
8464 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
8465 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
8466 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
8467 num_imm_uses): Move to ssa-iterators.h.
8468 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
8469 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
8471 (op_iter_done, op_iter_next_def, op_iter_next_tree,
8472 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
8473 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
8474 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
8475 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
8476 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
8477 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
8478 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
8479 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
8480 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
8481 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
8482 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
8483 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
8484 (set_phi_nodes): Move to tree-phinodes.h.
8485 * tree-ssa-operands.h (enum ssa_op_iter_type,
8486 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
8487 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
8488 (dump_decl_set): Remove prototype.
8489 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
8490 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
8491 (set_phi_nodes): Relocate from tree-flow-inline.h.
8492 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
8494 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
8495 include list. Temporarily add gimple.h to include list.
8496 * ssa-iterators.h: New file.
8497 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
8498 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
8499 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
8500 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
8501 Relocate from tree-ssa-operands.h.
8502 (delink_imm_use, link_imm_use_to_list, link_imm_use,
8503 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
8504 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
8505 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
8506 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
8507 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
8508 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
8509 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
8510 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
8511 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
8512 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
8513 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
8514 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
8515 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
8516 Relocate from tree-flow-inline.h.
8517 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
8519 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
8522 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
8523 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
8524 Insert '\t' to output template.
8525 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
8526 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
8527 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
8528 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
8530 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
8532 * doc/install.texi (Host/target specific installation notes for GCC):
8533 Put @anchor before @heading.
8534 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
8535 Use @email for email addresses.
8537 2013-10-01 Jeff Law <law@redhat.com>
8539 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
8540 a vec. Only delete the path if we create one without successfully
8541 registering a jump thread.
8542 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
8544 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
8545 (paths): New vector of jump threading paths.
8546 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
8547 (THREAD_PATH): New accessor macro for the entire thread path.
8548 (lookup_redirection_data): Get intermediate and final outgoing edge
8549 from the thread path.
8550 (create_edge_and_update_destination_phis): Copy the threading path.
8551 (ssa_fix_duplicate_block_edges): Get edges and block types from the
8552 jump threading path.
8553 (ssa_redirect_edges): Get edges and block types from the jump threading
8554 path. Free the path vector.
8555 (thread_block): Get edges from the jump threading path. Look at the
8556 entire path to see if we thread to a loop exit. If we cancel a jump
8557 thread request, then free the path vector.
8558 (thread_single_edge): Get edges and block types from the jump threading
8559 path. Free the path vector.
8560 (thread_through_loop_header): Get edges and block types from the jump
8561 threading path. Free the path vector.
8562 (mark_threaded_blocks): Iterate over the vector of paths and store
8563 the path on the appropriate edge. Get edges and block types from the
8564 jump threading path.
8565 (mark_threaded_blocks): Get edges and block types from the jump
8566 threading path. Free the path vector.
8567 (thread_through_all_blocks): Use the vector of paths rather than
8568 a vector of 3-edge sets.
8569 (register_jump_thread): Accept pointer to a path vector rather
8570 than the path vector itself. Store the path vector for later use.
8573 2013-10-01 Jakub Jelinek <jakub@redhat.com>
8574 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8577 * config/s390/s390.c (s390_split_branches): Modify check for table
8579 (s390_chunkify_start): Rearrange table jump insn check in order to
8580 deal with compare and branch insns correctly.
8582 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
8586 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
8587 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
8588 define_insn_and_split.
8589 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
8590 (shiftsi3_compare): New pattern.
8592 * config/arm/unspecs.md (UNSPEC_RRX): New.
8594 2013-10-01 Alan Modra <amodra@gmail.com>
8596 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
8597 casing inout operands.
8599 2013-10-01 Richard Biener <rguenther@suse.de>
8601 PR tree-optimization/58553
8602 * tree-loop-distribution.c (struct partition_s): Add niter member.
8603 (classify_partition): Populate niter member for the partition
8604 and properly identify whether the relevant store happens before
8605 or after the loop exit.
8606 (generate_memset_builtin): Use niter member from the partition.
8607 (generate_memcpy_builtin): Likewise.
8609 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
8611 * vec.h (vec_prefix, vec): Prefix member names with "m_".
8612 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
8614 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
8616 * basic-block.h (edge_list): Prefix member names with "m_".
8617 * context.h (context): Likewise.
8618 * domwalk.h (dom_walker): Likewise.
8619 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
8620 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
8621 * hash-table.h (hash_table): Likewise.
8622 * machmode.h (bit_field_mode_iterator): Likewise.
8623 * pass_manager.h (pass_list): Likewise.
8624 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
8625 * tree-pass.h (pass_data): Likewise.
8626 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
8627 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
8628 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
8629 * asan.c (pass_data_asan): Update accordingly.
8630 * cfganal.c (control_dependences::find_control_dependence): Likewise.
8631 (control_dependences::control_dependences): Likewise.
8632 (control_dependences::~control_dependences): Likewise.
8633 (control_dependences::~control_dependences): Likewise.
8634 (control_dependences::get_edges_dependent_on): Likewise.
8635 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
8636 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
8637 * context.c (gcc::context::context): Likewise.
8638 * cprop.c (pass_rtl_cprop::clone): Likewise.
8639 * domwalk.c (dom_walker::walk): Likewise.
8640 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
8641 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
8642 * mode-switching.c (pass_mode_switching::clone): Likewise.
8643 * passes.c (opt_pass::opt_pass): Likewise.
8644 (pass_manager::pass_manager): Likewise.
8645 * predict.c (pass_strip_predict_hints::clone): Likewise.
8646 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
8647 (pass_split_all_insns::clone): Likewise.
8648 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
8650 (bit_field_mode_iterator::next_mode): Likewise.
8651 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
8652 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
8653 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
8654 * tree-complex.c (pass_lower_complex::clone): Likewise.
8655 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
8656 * tree-object-size.c (pass_object_sizes::clone): Likewise.
8657 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
8658 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
8659 (pass_fold_builtins::clone): Likewise.
8660 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
8661 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
8662 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
8663 (pass_cd_dce::clone): Likewise.
8664 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
8665 (pass_phi_only_cprop::clone): Likewise.
8666 * tree-ssa-dse.c (pass_dse::clone): Likewise.
8667 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
8668 * tree-ssa-loop.c (pass_lim::clone): Likewise.
8669 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
8670 * tree-ssa-pre.c (pass_fre::clone): Likewise.
8671 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
8672 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
8673 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
8674 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
8675 * tree-vrp.c (pass_vrp::clone): Likewise.
8676 * tsan.c (pass_tsan::clone): Likewise.
8678 2013-09-30 Jakub Jelinek <jakub@redhat.com>
8681 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
8682 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
8685 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
8686 optimization, punt if sign_bit_p looked through any zero extension.
8688 2013-09-30 Teresa Johnson <tejohnson@google.com>
8690 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
8691 Update redirected out edge count in joiner case.
8692 (ssa_redirect_edges): Common the joiner and non-joiner cases
8693 so that joiner case gets profile updates.
8695 2013-09-30 Richard Biener <rguenther@suse.de>
8697 PR tree-optimization/58554
8698 * tree-loop-distribution.c (classify_partition): Require
8699 unconditionally executed stores for memcpy and memset recognition.
8700 (tree_loop_distribution): Calculate dominance info.
8702 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
8704 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
8705 (NO_PROFILE_COUNTERS): Likewise.
8706 (PROFILE_HOOK): Likewise.
8707 (FUNCTION_PROFILER): Likewise.
8708 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
8710 2013-09-30 Iain Sandoe <iain@codesourcery.com>
8712 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
8713 calls and defines in TARGET_MACHO conditional.
8714 (load_macho_picbase_di): Likewise.
8715 (reload_macho_picbase): Likewise.
8716 (reload_macho_picbase_si): Likewise.
8717 (reload_macho_picbase_di): Likewise.
8718 (nonlocal_goto_receiver): Likewise.
8720 2013-09-30 Nick Clifton <nickc@redhat.com>
8722 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
8723 that use the MSP430X ISA.
8724 (msp430_option_override): Scan -mmcu command line option for any
8725 MCU name that supports the MSP430X ISA.
8726 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
8727 -mmcu options which enable the MSP430X ISA.
8729 2013-09-30 Richard Biener <rguenther@suse.de>
8732 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
8733 before looking for setjmp-like calls.
8735 2013-09-29 Iain Sandoe <iain@codesourcery.com>
8738 * config/darwin-protos.h (machopic_get_function_picbase): New.
8739 * config/darwin.c (machopic_get_function_picbase): New.
8740 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
8741 label for a new func. (load_macho_picbase_di): Likewise.
8742 (reload_macho_picbase): New expand.
8743 (reload_macho_picbase_si): New insn.
8744 (reload_macho_picbase_di): New insn.
8745 (nonlocal_goto_receiver): New define and split.
8746 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
8747 (unspecv enum): Add UNSPECV_NLGR.
8749 2013-09-29 Iain Sandoe <iain@codesourcery.com>
8751 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
8752 that altivec registers are correctly sized on Darwin.
8754 2013-09-29 Iain Sandoe <iain@codesourcery.com>
8756 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
8757 darwin-driver.o): Use COMPILE and POSTCOMPILE.
8758 * config/x-darwin (host-darwin.o): Likewise.
8759 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
8760 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
8761 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
8763 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
8765 * doc/invoke.texi: Fix usage of @tie{} command.
8767 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
8769 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
8771 2013-09-28 Jan Hubicka <jh@suse.cz>
8773 * config/i386/x86-tune.def: Add documentation for each of the options;
8776 2013-09-28 Jan Hubicka <jh@suse.cz>
8778 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
8780 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
8781 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
8782 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
8783 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
8784 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
8786 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
8788 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
8789 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
8790 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
8791 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
8792 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
8793 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
8794 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
8795 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
8796 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
8797 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
8798 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
8799 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
8800 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
8801 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
8802 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
8803 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
8804 gimple.h, godump.c, graphite-clast-to-gimple.c,
8805 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
8806 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
8807 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
8808 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
8809 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
8810 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
8811 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
8812 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
8813 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
8814 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
8815 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
8816 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
8817 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
8818 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
8819 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
8820 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
8821 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
8822 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
8823 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
8824 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
8825 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
8826 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
8827 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
8828 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
8829 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
8830 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
8831 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
8832 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
8833 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
8834 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
8835 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
8836 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
8837 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
8838 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
8839 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
8840 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
8841 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
8842 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
8843 whitespace before "(".
8845 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
8847 * expr.h (extract_bit_field): Remove packedp parameter.
8848 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
8849 from forward declaration.
8850 (store_split_bit_field): Remove packedp arg from calls to
8851 extract_fixed_bit_field.
8852 (extract_bit_field_1): Remove packedp parameter and packedp
8853 argument from recursive calls and calls to extract_fixed_bit_field.
8854 (extract_bit_field): Remove packedp parameter and corresponding
8855 arg to extract_bit_field_1.
8856 (extract_fixed_bit_field): Remove packedp parameter. Remove code
8858 (extract_split_bit_field): Remove packedp arg from call to
8859 extract_fixed_bit_field.
8860 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
8861 (copy_blkmode_from_reg): Likewise.
8862 (copy_blkmode_to_reg): Likewise.
8863 (read_complex_part): Likewise.
8864 (store_field): Likewise.
8865 (expand_expr_real_1): Likewise.
8866 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
8867 to extract_bit_field.
8868 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
8869 call to extract_bit_field.
8870 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
8871 call to extract_bit_field.
8872 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
8873 and special packedp behavior from -fstrict-volatile-bitfields
8876 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8878 * lra-eliminations.c (init_elim_table): Guard value_p.
8880 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
8882 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
8883 DFmode, DImode, and SFmode in the upper VSX registers based on the
8884 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
8885 if -mpower8-vector. Combine -mvsx-timode handling with the rest
8886 of the VSX register handling.
8888 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
8890 (zero_extendsidi2_lfiwzx): Add support for loading into the
8891 Altivec registers with -mpower8-vector. Use wu/wv constraints to
8892 only do VSX memory options on Altivec registers.
8893 (extendsidi2_lfiwax): Likewise.
8894 (extendsfdf2_fpr): Likewise.
8895 (mov<mode>_hardfloat, SF/SD modes): Likewise.
8896 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
8897 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
8898 (movdi_internal64): Likewise.
8900 2013-09-27 Xinliang David Li <davidxl@google.com>
8902 * opts.c (finish_options): Adjust parameters
8903 according to vect cost model.
8904 (common_handle_option): Set dynamic vect cost
8906 targhooks.c (default_add_stmt_cost): Compute stmt cost
8908 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
8909 Use helper function.
8910 * tree-vectorizer.h (unlimited_cost_model): New function.
8911 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
8912 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
8914 (vect_enhance_data_refs_alignment): Ditto.
8915 * flag-types.h: New enum.
8916 * common/config/i386/i386-common.c (ix86_option_init_struct):
8917 No need to initialize vect_cost_model flag.
8918 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
8921 2013-09-27 Diego Novillo <dnovillo@google.com>
8923 * gimple.h (enum ssa_mode): Remove.
8925 2013-09-27 Paulo Matos <pmatos@broadcom.com>
8927 * cfgloop.h (number_of_loops): Fix typo in check for null.
8929 2013-09-27 Jakub Jelinek <jakub@redhat.com>
8932 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
8933 are children of outermost saved_cfun's loop, and set it up to
8934 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
8935 if loop != loop0 and SESE region contains bbs that belong to loop0.
8937 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
8939 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
8940 (binary_scale_code_p, get_base_term, get_index_term): New functions.
8941 (set_address_segment, set_address_base, set_address_index)
8942 (set_address_disp): Accept the argument unconditionally.
8943 (baseness): Remove must_be_base_p and must_be_index_p checks.
8944 (decompose_normal_address): Classify as much as possible in the
8947 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
8949 * cse.c (count_reg_usage): Handle INT_LIST.
8950 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
8951 * reginfo.c (reg_scan_mark_refs): Likewise.
8952 * reload1.c (eliminate_regs_1): Likewise.
8954 2013-09-27 Iain Sandoe <iain@codesourcery.com>
8957 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
8960 2013-09-27 Richard Biener <rguenther@suse.de>
8962 PR tree-optimization/58459
8963 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
8964 restriction not propagating into loops.
8966 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
8968 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
8969 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
8970 * doc/tree-ssa.texi (Walking use-def chains): Delete.
8972 2013-09-26 Richard Biener <rguenther@suse.de>
8974 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
8976 2013-09-26 Richard Biener <rguenther@suse.de>
8978 * alias.h (component_uses_parent_alias_set): Rename to ...
8979 (component_uses_parent_alias_set_from): ... this.
8980 * alias.c (component_uses_parent_alias_set): Rename to ...
8981 (component_uses_parent_alias_set_from): ... this and return
8983 (reference_alias_ptr_type_1): Use the result from
8984 component_uses_parent_alias_set_from instead of stripping
8985 components one at a time.
8986 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
8988 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
8990 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
8991 Move prototypes to...
8992 * tree-ssa-ter.h: New File. Move prototypes here.
8993 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
8994 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
8995 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
8996 from is_replaceable_p.
8997 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
8998 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
8999 (process_replaceable): Use ter_is_replaceable_p.
9000 (find_replaceable_in_bb): Use ter_is_replaceable_p.
9001 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
9002 newly refactored ssa_is_replaceable_p.
9003 * cfgexpand.c: Include tree-outof-ssa.h.
9004 * ssaexpand.h: Delete.
9006 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
9008 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
9009 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
9011 (create_gimple_tmp): Delete.
9012 (get_expr_type, build_assign, build_type_cast): Move to...
9013 * gimple-builder.c: New File.
9014 (get_expr_type): Relocate from gimple.c.
9015 (build_assign, build_type_cast): Change to only create ssanames.
9016 * gimple.h: Move prototypes to...
9017 * gimple-builder.h: New File. Here.
9018 * tree-ssa.h: And here.
9019 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
9020 count_uses_and_derefs): Relocate from gimple.c.
9021 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
9022 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
9023 * tree-ssa-math-opts (execute_cse_reciprocals): Use
9024 gimple_replace_ssa_lhs.
9025 * asan.c: Include gimple-builder.h.
9026 * Makefile.in: Add gimple-builder.o.
9028 2013-09-26 Richard Biener <rguenther@suse.de>
9030 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
9031 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
9032 (loe_visit_block): Use gcc_checking_assert.
9033 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
9034 gimple_assign_ssa_name_copy_p.
9035 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
9038 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
9040 * config/rs6000/t-rs6000 (rs6000.o): Remove.
9041 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
9043 2013-09-26 Richard Biener <rguenther@suse.de>
9045 PR tree-optimization/58539
9046 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
9047 the fact that debug statements are not taking part in loop-closed
9050 2013-09-26 Nick Clifton <nickc@redhat.com>
9052 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
9053 time warning message.
9054 (msp430_print_operand_raw): Delete unused letter parameter.
9055 (TARGET_PRINT_OPERAND_ADDRESS): Define.
9056 (msp430_print_operand_address): New function.
9057 (msp430_print_operand): Move address printing code from here to
9059 * config/msp430/msp430.md (movsipsi2): Add comment in generated
9061 (zero_extendpsisi2): Likewise.
9062 (extendpsisi2): New pattern.
9063 (andneghi3): New pattern.
9065 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
9067 * config/aarch64/aarch64.opt (mlra): New option.
9068 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
9069 (TARGET_LRA_P): Define.
9071 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
9073 * expr.c (expand_assignment): Remove obsolete comment.
9075 2013-09-25 Jeff Law <law@redhat.com>
9077 * tree-flow.h (thread_through_all_blocks): Prototype moved into
9078 tree-ssa-threadupdate.h.
9079 (register_jump_thread): Similarly.
9080 * tree-ssa-threadupdate.h: New header file.
9081 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
9082 * tree-vrp.c: Likewise.
9083 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
9084 (thread_around_empty_blocks): Change type of path vector argument to
9085 an edge,type pair from just an edge. Initialize both elements when
9086 appending to a jump threading path. Tweak references to elements
9088 (thread_across_edge): Similarly. Release memory for the elements
9090 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
9091 (dump_jump_thread_path): New function broken out from
9092 register_jump_thread.
9093 (register_jump_thread): Use dump_jump_thread_path. Change type of
9094 path vector entries. Search the path for NULL edges and dump
9095 the path if one is found. Tweak the conversion of path to 3-edge
9096 form to use the block copy type information embedded in the path.
9098 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
9100 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
9102 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
9103 Vladimir Makarov <vmakarov@redhat.com>
9105 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
9106 from the least significant bit.
9107 (strip_address_mutations): Add bitfield operations handling.
9108 (must_be_index_p): Add shifting and rotate operations handling.
9109 (set_address_base): Use must_be_base_p predicate.
9110 (set_address_index): Use must_be_index_p predicate.
9112 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
9113 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
9114 Sergey Lega <sergey.s.lega@intel.com>
9115 Anna Tikhonova <anna.tikhonova@intel.com>
9116 Ilya Tocar <ilya.tocar@intel.com>
9117 Andrey Turetskiy <andrey.turetskiy@intel.com>
9118 Ilya Verbin <ilya.verbin@intel.com>
9119 Kirill Yukhin <kirill.yukhin@intel.com>
9120 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
9122 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
9124 (ix86_expand_vector_move_misalign): Support new unaligned load and
9125 stores and use new names.
9126 (CODE_FOR_sse2_storedqu): Rename to ...
9127 (CODE_FOR_sse2_storedquv16qi): ... this.
9128 (CODE_FOR_sse2_loaddqu): Rename to ...
9129 (CODE_FOR_sse2_loaddquv16qi): ... this.
9130 (CODE_FOR_avx_loaddqu256): Rename to ...
9131 (CODE_FOR_avx_loaddquv32qi): ... this.
9132 (CODE_FOR_avx_storedqu256): Rename to ...
9133 (CODE_FOR_avx_storedquv32qi): ... this.
9134 * config/i386/i386.md (fpint_logic): New.
9135 * config/i386/sse.md (VMOVE): Extend for AVX512.
9139 (VI_UNALIGNED_LOADSTORE): Ditto.
9140 (sse2_avx_avx512f): Ditto.
9141 (sse2_avx2): Extend for AVX512.
9142 (sse4_1_avx2): Ditto.
9143 (avx2_avx512f): New.
9144 (sse): Extend for AVX512.
9147 (avxsizesuffix): Ditto.
9148 (sseintvecmode): Ditto.
9150 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
9151 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
9152 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
9153 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
9154 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
9155 (<sse2_avx_avx512f>_storedqu<mode): ... this.
9156 (<sse>_movnt<mode>): Replace constraint "x" with "v".
9157 (STORENT_MODE): Extend for AVX512.
9158 (*absneg<mode>2): Replace constraint "x" with "v".
9159 (*mul<mode>3): Ditto.
9160 (*ieee_smin<mode>3): Ditto.
9161 (*ieee_smax<mode>3): Ditto.
9162 (avx_cmp<mode>3): Replace VF with VF_128_256.
9163 (*<sse>_maskcmp<mode>3_comm): Ditto.
9164 (<sse>_maskcmp<mode>3): Ditto.
9165 (<sse>_andnot<mode>3): Extend for AVX512.
9166 (<code><mode>3, anylogic): Replace VF with VF_128_256.
9167 (<code><mode>3, fpint_logic): New.
9168 (*<code><mode>3): Extend for AVX512.
9169 (avx512flogicsuff): New.
9170 (avx512f_<logic><mode>): Ditto.
9171 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
9173 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
9174 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
9175 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
9176 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
9177 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
9178 (xop_vpermil2<mode>3): Ditto.
9179 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
9180 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
9181 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
9182 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
9184 2013-09-25 Tom Tromey <tromey@redhat.com>
9186 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
9187 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
9188 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
9189 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
9190 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
9191 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
9192 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
9193 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
9194 (GRAPHITE_HTAB_H): Remove.
9196 2013-09-25 Tom Tromey <tromey@redhat.com>
9198 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
9200 2013-09-25 Tom Tromey <tromey@redhat.com>
9202 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
9204 2013-09-25 Tom Tromey <tromey@redhat.com>
9206 * config/i386/t-i386 (i386.o): Remove.
9207 (i386-c.o): Use COMPILE and POSTCOMPILE.
9209 2013-09-25 Tom Tromey <tromey@redhat.com>
9211 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
9213 2013-09-25 Tom Tromey <tromey@redhat.com>
9215 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
9216 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
9217 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
9218 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
9219 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
9220 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
9221 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
9222 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
9223 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
9224 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
9225 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
9226 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
9227 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
9228 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
9229 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
9230 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
9231 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
9232 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
9233 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
9234 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
9235 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
9236 (tree-ssa-pre.o, tree-ssa-sccvn.o)
9237 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
9238 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
9239 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
9240 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
9241 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
9242 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
9243 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
9244 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
9245 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
9246 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
9247 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
9248 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
9249 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
9250 (graphite-clast-to-gimple.o, graphite-dependences.o)
9251 (graphite-interchange.o, graphite-poly.o)
9252 (graphite-scop-detection.o, graphite-sese-to-poly.o)
9253 (graphite-optimize-isl.o, tree-vect-loop.o)
9254 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
9255 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
9256 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
9257 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
9258 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
9259 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
9260 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
9261 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
9262 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
9263 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
9264 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
9265 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
9266 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
9267 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
9268 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
9269 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
9270 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
9271 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
9272 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
9273 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
9274 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
9275 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
9276 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
9277 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
9278 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
9279 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
9280 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
9281 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
9282 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
9283 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
9284 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
9285 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
9286 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
9287 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
9288 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
9289 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
9290 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
9291 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
9292 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
9293 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
9294 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
9295 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
9296 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
9297 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
9298 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
9299 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
9300 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
9301 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
9302 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
9303 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
9304 (gcov-dump.o): Remove.
9305 (default-c.o): Use COMPILE and POSTCOMPILE.
9306 (CFLAGS-gcc.o): New variable.
9307 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
9309 2013-09-25 Tom Tromey <tromey@redhat.com>
9311 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
9312 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
9313 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
9314 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
9315 (c-family/c-pragma.o, c-family/c-pretty-print.o)
9316 (c-family/c-semantics.o, c-family/c-ada-spec.o)
9317 (c-family/array-notation-common.o, c-family/stub-objc.o)
9318 (c-family/c-ubsan.o): Remove.
9320 2013-09-25 Tom Tromey <tromey@redhat.com>
9322 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
9324 2013-09-25 Tom Tromey <tromey@redhat.com>
9326 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
9327 to add -DENABLE_SHARED_LIBGCC.
9328 (gcc.o): Don't use subshell.
9330 2013-09-25 Tom Tromey <tromey@redhat.com>
9332 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
9333 * configure.ac: Don't invoke AM_PROG_CC_C_O.
9334 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
9335 * configure, config.in: Rebuild.
9337 2013-09-25 Tom Tromey <tromey@redhat.com>
9339 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
9340 (COMPILE, POSTCOMPILE): New variables.
9341 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
9342 (DEPFILES): New variable.
9343 Include ".Po" files.
9344 * configure.ac: Add checks for dependency checking.
9345 * configure, aclocal.m4: Regenerate.
9347 2013-09-25 Tom Tromey <tromey@redhat.com>
9349 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
9350 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
9352 2013-09-25 Tom Tromey <tromey@redhat.com>
9354 * Makefile.in (generated_files): Add options.h,
9355 target-hooks-def.h, insn-opinit.h,
9356 common/common-target-hooks-def.h, pass-instances.def,
9357 c-family/c-target-hooks-def.h.
9359 2013-09-25 Jeff Law <law@redhat.com>
9361 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
9362 than foo[foo.length () - 1] to access last member in a vec.
9363 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
9365 2013-09-25 Richard Biener <rguenther@suse.de>
9368 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
9370 2013-09-25 Jan Hubicka <jh@suse.cz>
9372 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
9375 2013-09-25 Marek Polacek <polacek@redhat.com>
9378 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
9379 when determining the type name.
9381 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
9383 * config/sh/sh.md: Fix formatting.
9385 2013-09-24 Xinliang David Li <davidxl@google.com>
9387 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
9388 max peel iterations parameter.
9389 * param.def: New parameter.
9390 * doc/invoke.texi: Document New parameter.
9392 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
9394 * gimple-pretty-print.c: Various whitespace tweaks.
9395 * tree-core.h: Likewise.
9396 * tree-pretty-print.c: Likewise.
9397 * tree-ssa-alias.c: Likewise.
9398 * tree-ssa-copy.c: Likewise.
9399 * tree-ssanames.c: Likewise.
9400 * tree-ssanames.h: Likewise.
9401 * tree-vrp.c: Likewise.
9403 2013-09-24 Alan Modra <amodra@gmail.com>
9407 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
9408 for output operands that disallow regs. Don't use EXPAND_WRITE on
9411 2013-09-24 Richard Biener <rguenther@suse.de>
9414 * tree.c (reference_alias_ptr_type): Move ...
9415 * alias.c (reference_alias_ptr_type): ... here and implement
9416 in terms of the new reference_alias_ptr_type_1.
9417 (ref_all_alias_ptr_type_p): New helper.
9418 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
9419 use ref_all_alias_ptr_type_p.
9420 (get_deref_alias_set): Add flag_strict_aliasing check here.
9421 (reference_alias_ptr_type_1): New function, split out from ...
9422 (get_alias_set): ... here.
9423 (alias_ptr_types_compatible_p): New function.
9424 * alias.h (reference_alias_ptr_type): Declare.
9425 (alias_ptr_types_compatible_p): Likewise.
9426 * tree.h (reference_alias_ptr_type): Remove.
9427 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
9428 to compare MEM_REF alias types.
9430 2013-09-24 Richard Biener <rguenther@suse.de>
9432 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
9434 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
9436 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
9437 reload helper function arrays into a single array reg_addr.
9438 (reload_fpr_gpr): Likewise.
9439 (reload_gpr_vsx): Likewise.
9440 (reload_vsx_gpr): Likewise.
9441 (struct rs6000_reg_addr): Likewise.
9442 (reg_addr): Likewise.
9443 (rs6000_debug_reg_global): Change rs6000_vector_reload,
9444 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
9445 (rs6000_init_hard_regno_mode_ok): Likewise.
9446 (rs6000_secondary_reload_direct_move): Likewise.
9447 (rs6000_secondary_reload): Likewise.
9449 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
9450 constraints: wu, ww, and wy. Repurpose wv constraint added during
9451 power8 changes. Put wg constraint in alphabetical order.
9453 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
9454 for future work to add ISA 2.07 VSX single precision support.
9455 (-mvsx-scalar-double): Change default from -1 to 1, update
9456 documentation comment.
9457 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
9458 (-mupper-regs-df): New debug switch to control whether DF values
9459 can go in the traditional Altivec registers.
9460 (-mupper-regs-sf): New debug switch to control whether SF values
9461 can go in the traditional Altivec registers.
9463 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
9465 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
9466 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
9467 Add new constraints, wu/ww/wy. Repurpose wv constraint.
9468 (rs6000_debug_legitimate_address_p): Print if we are running
9469 before, during, or after reload.
9470 (rs6000_secondary_reload): Add a comment.
9471 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
9473 * config/rs6000/constraints.md (wa constraint): Sort w<x>
9474 constraints. Update documentation string.
9475 (wd constraint): Likewise.
9476 (wf constraint): Likewise.
9477 (wg constraint): Likewise.
9478 (wn constraint): Likewise.
9479 (ws constraint): Likewise.
9480 (wt constraint): Likewise.
9481 (wx constraint): Likewise.
9482 (wz constraint): Likewise.
9483 (wu constraint): New constraint for ISA 2.07 SFmode scalar
9485 (ww constraint): Likewise.
9486 (wy constraint): Likewise.
9487 (wv constraint): Repurpose ISA 2.07 constraint that we did not use
9488 in the previous submissions.
9489 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
9491 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
9493 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
9494 (REG_BR_PROB): Say that the probability is stored in an int_list.
9495 * reg-notes.def: Update commentary to mention INT_LIST.
9496 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
9497 (INT_LIST): New rtx.
9498 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
9499 * rtlanal.c (int_reg_note_p): New function.
9500 (alloc_reg_note): Assert that the note does not have an int argument.
9501 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
9502 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
9503 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
9504 rather than an INSN_LIST. Handle INT_LIST.
9505 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
9506 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
9507 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
9508 Manipulate them as ints rather than rtxes.
9509 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
9510 * regmove.c (copy_src_to_dest): Likewise.
9511 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
9513 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
9514 into the cases that need it.
9515 * config/arm/arm.c (arm_unwind_emit): Likewise.
9517 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
9518 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
9519 * loop-doloop.c (add_test, doloop_modify): Likewise.
9520 * loop-unswitch.c (compare_and_jump_seq): Likewise.
9521 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
9522 * predict.c (combine_predictions_for_insn): Likewise.
9523 * print-rtl.c (print_rtx): Handle INT_LIST.
9524 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
9525 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
9526 * config/arm/arm.c (emit_unlikely_jump): Likewise.
9527 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
9528 (ix86_split_fp_branch, predict_jump): Likewise.
9529 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
9530 * config/sh/sh.c (expand_cbranchsi4): Likewise.
9531 * config/spu/spu.c (ea_load_store_inline): Likewise.
9533 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
9534 value of a REG_BR_PROB note.
9535 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
9536 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
9537 * emit-rtl.c (try_split): Likewise.
9538 * predict.c (br_prob_note_reliable_p): Likewise.
9539 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
9540 * reorg.c (mostly_true_jump): Likewise.
9541 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
9542 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
9543 * config/i386/i386.c (ix86_print_operand): Likewise.
9544 * config/ia64/ia64.c (ia64_print_operand): Likewise.
9545 * config/mmix/mmix.c (mmix_print_operand): Likewise.
9546 * config/rs6000/rs6000.c (output_cbranch): Likewise.
9547 * config/s390/s390.c (s390_expand_tbegin): Likewise.
9548 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
9549 * config/sparc/sparc.c (output_cbranch): Likewise.
9550 * config/spu/spu.c (get_branch_target): Likewise.
9551 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
9552 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
9554 2013-09-23 Jan Hubicka <jh@suse.cz>
9556 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
9558 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
9559 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
9561 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
9562 (gimple_fold_call): Dump inconsistent devirtualizations; add
9563 sanity check for type based devirtualizations.
9564 * ipa-prop.c: Include ipa-utils.h
9565 (ipa_intraprocedural_devirtualization): Add sanity check.
9566 (try_make_edge_direct_virtual_call): Likewise.
9568 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
9570 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
9571 assignment statements.
9573 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
9575 * gimple-pretty-print.c (dump_ssaname_info): New function.
9576 (dump_gimple_phi): Call it.
9577 (pp_gimple_stmt_1): Likewise.
9578 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
9579 (range_info_def): Declare.
9580 * tree-pretty-print.c (pp_double_int): New function.
9581 (dump_generic_node): Call it.
9582 * tree-pretty-print.h (pp_double_int): Declare.
9583 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
9584 * tree-ssanames.h (range_info_def): New structure.
9585 (value_range_type): Move definition here.
9586 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
9588 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
9590 (set_range_info): New function.
9591 (get_range_info): Likewise.
9592 (duplicate_ssa_name_range_info): Likewise.
9593 (duplicate_ssa_name_fn): Check pointer type and call
9594 duplicate_ssa_name_range_info.
9595 * tree-ssa-copy.c (fini_copy_prop): Likewise.
9596 * tree-vrp.c (value_range_type): Remove definition, now in
9598 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
9599 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
9600 (SSA_NAME_RANGE_INFO): New macro.
9602 2013-09-23 Richard Biener <rguenther@suse.de>
9604 PR tree-optimization/58464
9605 * tree-ssa-pre.c (phi_trans_lookup): Remove.
9606 (phi_trans_add): Change to add conditionally on being not present.
9607 (phi_translate_1): Remove recursion detection here.
9608 (phi_translate): Pre-seed the cache with NULL to catch
9609 recursion here in a more generic way.
9610 (bitmap_find_leader): Adjust comment.
9611 (get_representative_for): Dump value-numbers.
9612 (create_expression_by_pieces): Likewise.
9613 (insert_into_preds_of_block): Likewise.
9615 2013-09-23 Christian Bruel <christian.bruel@st.com>
9618 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
9619 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
9621 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
9624 2013-09-20 Renlin Li <renlin.li@arm.com>
9626 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
9628 (aarch64_expand_epilogue): Likewise.
9629 (aarch64_legitimize_reload_address): Likewise.
9631 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
9633 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
9634 NULL_TREE before pushing them onto the vector. Likewise for labels.
9636 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
9638 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
9639 comparison operators when -fno-trapping-math is in effect.
9640 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
9641 comparison operators in TFmode and assert that unsupported operators
9643 (ia64_print_operand): Likewise.
9645 2013-09-21 Jan Hubicka <jh@suse.cz>
9647 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
9648 (sse_typeless_stores): Enable for core
9649 (sse_load0_by_pxor): Likewise.
9650 (four_jump_limit): Disable for core.
9651 (pad_returns): Likewise.
9652 (avoid_vector_decode): Likewise.
9653 (fuse_cmp_and_branch): Enable for cores.
9654 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
9656 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
9659 * config/pa/pa.c (pa_option_override): Disable auto increment and
9660 decrement instructions until reload is completed.
9662 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
9663 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
9664 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
9666 2013-09-20 DJ Delorie <dj@redhat.com>
9667 Nick Clifton <nickc@redhat.com>
9669 * config/rl78/rl78.c: Various whitespace and comment tweaks.
9670 (need_to_save): Save bank 0 on interrupts.
9671 (characterize_address): Strip far address wrappers.
9672 (rl78_as_legitimate_address): Likewise.
9673 (transcode_memory_rtx): Likewise.
9674 (rl78_peep_movhi_p): Disable this peephole after devirt.
9675 (rl78_propogate_register_origins): Forget all origins when a
9677 * config/rl78/rl78-virt.md: Various whitespace tweaks.
9678 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
9680 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
9681 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
9682 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
9684 * config/rl78/constraints.md: Various whitespace and paren tweaks.
9686 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
9688 * config/pa/pa.md: In "scc" insn patterns, change output template to
9689 handle const0_rtx in reg_or_0_operand operands.
9691 2013-09-20 Martin Husemann <martin@NetBSD.org>
9694 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
9695 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
9697 2013-09-20 Richard Biener <rguenther@suse.de>
9700 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
9701 remembering SSA name version and block index.
9702 (new_scev_info_str): Adjust.
9703 (hash_scev_info): Likewise. Also hash the block index.
9704 (eq_scev_info): Adjust.
9705 (find_var_scev_info): Likewise.
9706 (struct instantiate_cache_entry): Remove.
9707 (struct instantiate_cache_type): Use a htab to map name, block
9709 (instantiate_cache_type::~instantiate_cache_type): Adjust.
9710 (get_instantiated_value_entry): Likewise.
9711 (hash_idx_scev_info, eq_idx_scev_info): New functions.
9712 (instantiate_scev_name): Adjust.
9714 2013-09-20 Jeff Law <law@redhat.com>
9716 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
9718 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
9720 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
9721 Call aarch64_simd_expand_args to update op[argc].
9723 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
9725 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
9728 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
9730 * gengtype.c (file_rules): Added rule for *.cc files.
9731 (get_output_file_with_visibility): Give fatal message when no
9734 2013-09-20 Renlin Li <renlin.li@arm.com>
9736 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
9737 (aarch64_expand_epilogue): Likewise.
9738 (aarch64_legitimize_reload_address): Likewise.
9740 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
9743 * expr.c (expand_assignment): Remove misalignp code path.
9745 2013-09-20 Marek Polacek <polacek@redhat.com>
9748 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
9749 TYPE_PRECISION. Add asserts.
9751 2013-09-20 Richard Biener <rguenther@suse.de>
9753 PR tree-optimization/58453
9754 * tree-loop-distribution.c (distribute_loop): Apply the cost
9755 model for -ftree-loop-distribute-patterns, too.
9757 2013-09-20 Richard Biener <rguenther@suse.de>
9760 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
9761 make type comparison less strict.
9763 2013-09-20 Alan Modra <amodra@gmail.com>
9765 * configure: Regenerate.
9766 * aclocal.m4: Regenerate.
9768 2013-09-20 Marek Polacek <polacek@redhat.com>
9771 * doc/extend.texi: Document that attribute used is meant to be used
9772 on variables with static storage duration.
9774 2013-09-19 Jakub Jelinek <jakub@redhat.com>
9776 PR tree-optimization/58472
9777 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
9778 simd_lane_access set inv_p = false.
9779 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
9780 the simduid magic VAR_DECL.
9782 2013-09-19 Jan Hubicka <jh@suse.cz>
9784 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
9786 2013-09-17 Jeff Law <law@redhat.com>
9788 * tree-ssa-dom.c (record_temporary_equivalences): New function
9789 split out of dom_opt_dom_walker::after_dom_children.
9790 (dom_opt_dom_walker::thread_across_edge): Move common code
9791 in here from dom_opt_dom_walker::after_dom_children.
9792 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
9794 2013-09-19 Jan Hubicka <jh@suse.cz>
9796 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
9797 (TARGET_GENERIC): Use PROCESOR_GENERIC
9798 (enum processor_type): Unify generic32 and 64.
9799 * i386.md (cpu): Likewise.
9800 * x86-tune.def (use_leave): Enable for generic32.
9801 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
9802 * athlon.md: Change generic64 to generic in all occurences.
9803 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
9804 (ix86_target_macros_internal): Likewise.
9805 * driver-i386.c (host_detect_local_cpu): Likewise.
9806 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
9808 (generic_memcpy, generic_memset, generic_cost): This one.
9809 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
9810 (m_GENERIC32, m_GENERIC64): Remove.
9811 (m_GENERIC): Turn into one flag.
9812 (processor_target): Unify generic tunnings.
9813 (ix86_option_override_internal): Replace generic32/64 by generic.
9814 (ix86_issue_rate): Likewise.
9815 (ix86_adjust_cost): Likewise.
9817 2013-09-19 Jan Hubicka <jh@suse.cz>
9819 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
9820 of speculative flag.
9822 2013-09-19 Jakub Jelinek <jakub@redhat.com>
9824 * omp-low.c (expand_omp_sections): Always pass len - 1 to
9825 GOMP_sections_start, even if !exit_reachable.
9827 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
9829 * lra-constraints.c (need_for_all_save_p): Use macro
9830 HARD_REGNO_CALL_PART_CLOBBERED.
9831 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
9832 set up pseudo conflict hard regs.
9834 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
9837 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
9840 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
9842 PR rtl-optimization/58438
9843 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
9844 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
9845 from previous subpasses.
9847 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
9849 * arm.c (arm_get_frame_offsets): Validate architecture supports
9850 LDRD/STRD before accepting the tuning preference.
9851 (arm_expand_prologue): Likewise.
9852 (arm_expand_epilogue): Likewise.
9854 2013-09-18 Richard Biener <rguenther@suse.de>
9856 PR tree-optimization/58417
9857 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
9858 have chrecs with symbols defined in the loop as operands.
9859 (chrec_fold_multiply): Likewise.
9860 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
9861 parameters before folding binary operations.
9862 (struct instantiate_cache_entry_hasher): Remove.
9863 (struct instantiate_cache_type): Use a pointer-map.
9864 (instantiate_cache_type::instantiate_cache_type): New function.
9865 (instantiate_cache_type::get): Likewise.
9866 (instantiate_cache_type::set): Likewise.
9867 (instantiate_cache_type::~instantiate_cache_type): Adjust.
9868 (get_instantiated_value_entry): Likewise.
9869 (global_cache): New global.
9870 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
9871 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
9872 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
9873 (instantiate_scev_name): Adjust.
9874 (instantiate_scev): Construct global instead of local cache.
9875 (resolve_mixers): Likewise.
9877 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
9878 Paolo Carlini <paolo.carlini@oracle.com>
9881 * doc/implement-cxx.texi: Fix references to the C++ standards.
9883 2013-09-18 Jakub Jelinek <jakub@redhat.com>
9885 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
9886 * tree-vect-data-refs.c (vect_analyze_data_refs): For
9887 simd_lane_access drs, update also DR_ALIGNED_TO.
9889 2013-09-18 Marek Polacek <polacek@redhat.com>
9892 * doc/extend.texi: Document no_sanitize_undefined attribute.
9893 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
9894 no_sanitize_undefined attribute.
9896 2013-09-18 Nick Clifton <nickc@redhat.com>
9898 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
9899 (ASM_DECLARE_FUNCTION_NAME): Define.
9901 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
9903 * compare-elim.c (find_comparison_dom_walker): New class
9904 (find_comparisons_in_bb): Rename to
9905 find_comparison_dom_walker::before_dom_children
9906 (find_comparisons): Adjust
9907 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
9909 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
9910 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
9911 (init_walk_dominator_tree): Remove declaration.
9912 (fini_walk_dominator_tree): Remove declaration.
9913 * fwprop.c (single_def_use_dom_walker): New class
9914 (single_def_use_enter_block): Convert to
9915 single_def_use_dom_walker::before_dom_children.
9916 (single_def_use_leave_block): Convert to
9917 single_def_use_dom_walker::after_dom_children.
9918 (build_single_def_use_links): Adjust.
9919 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
9921 (find_candidates_in_block): Convert to
9922 find_candidates_dom_walker::before_dom_children.
9923 (execute_strength_reduction): Adjust.
9924 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
9925 (sese_dom_walker): New class.
9926 (sese_dom_walker::sese_dom_walker): New constructor.
9927 (sese_dom_walker::~sese_dom_walker): New destructor.
9928 (build_sese_conditions_before): Convert to
9929 sese_dom_walker::before_dom_children.
9930 (build_sese_conditions_after): Convert to
9931 sese_dom_walker::after_dom_children.
9932 (build_poly_scop): Adjust
9933 * tree-into-ssa.c (rewrite_dom_walker): New class
9934 (rewrite_enter_block): Convert to
9935 rewrite_dom_walker::before_dom_children.
9936 (rewrite_leave_block): Convert to
9937 rewrite_dom_walker::after_dom_children.
9938 (rewrite_update_dom_walker): New class.
9939 (rewrite_update_enter_block): Convert to
9940 rewrite_update_dom_walker::before_dom_children.
9941 (rewrite_update_leave_block): Convert to
9942 rewrite_update_dom_walker::after_dom_children.
9943 (rewrite_blocks, rewrite_into_ssa): Adjust.
9944 (mark_def_dom_walker): New class.
9945 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
9946 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
9947 (mark_def_sites_blocks): Convert to
9948 mark_def_dom_walker::before_dom_children.
9949 (mark_def_site_blocks): Remove.
9950 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
9951 (tree_ssa_dominator_optimize): Adjust.
9952 (dom_thread_across_edge): Convert to method
9953 dom_opt_dom_walker::thread_across_edge.
9954 (dom_opt_enter_block): Convert to member function
9955 dom_opt_dom_walker::before_dom_children.
9956 (dom_opt_leave_block): Convert to member function
9957 dom_opt_dom_walker::after_dom_children.
9958 * tree-ssa-dse.c (dse_dom_walker): New class.
9959 (dse_enter_block): Convert to member function
9960 dse_dom_walker::before_dom_children.
9961 (tree_ssa_dse): Adjust.
9962 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
9963 (determine_invariantness_stmt): Convert to method
9964 invariantness_dom_walker::before_dom_children.
9965 (determine_invariantness): Remove
9966 (move_computations_dom_walker): New class.
9967 (move_computations_stmt): Convert to method
9968 move_computations_dom_walker::before_dom_children.
9969 (move_computations, tree_ssa_lim): Adjust.
9970 * tree-ssa-phiopt.c (nontrapping_dom_walker): New class.
9971 (nt_init_block): Convert to method
9972 notrappping_dom_walker::before_dom_children.
9973 (nt_fini_block): Convert to method
9974 method nontrapping_dom_walker::after_dom_children.
9975 (get_non_trapping): Adjust.
9976 * tree-ssa-pre.c (eliminate_dom_walker): New class.
9977 (eliminate_bb): Convert to method
9978 eliminate_dom_walker::before_dom_children.
9979 (eliminate_leave_block): Convert to method
9980 eliminate_dom_walker::after_dom_children.
9981 (eliminate): Adjust.
9982 * tree-ssa-strlen.c (strlen_dom_walker): New class.
9983 (strlen_enter_block): Convert to method
9984 strlen_dom_walker::before_dom_children.
9985 (strlen_leave_block): Convert to method
9986 method strlen_dom_walker::after_dom_children.
9987 (tree_ssa_strlen): Adjust.
9988 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
9989 (tree_ssa_uncprop): Adjust.
9990 (uncprop_leave_block): Convert to method
9991 uncprop_dom_walker::after_dom_children.
9992 (uncprop_leave_block): Convert to method
9993 uncprop_dom_walker::before_dom_children.
9995 2013-09-18 Bin Cheng <bin.cheng@arm.com>
9997 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
9998 branch in same basic block. Check both src and dest of the move insn.
10000 2013-09-17 Nick Clifton <nickc@redhat.com>
10002 * config/rl78/rl78-real.md (bf): New pattern.
10004 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
10005 (rl78_print_operand): Do not put a # before a %B.
10006 * config/rl78/rl78.opt: Tweak doc strings.
10008 2013-09-17 DJ Delorie <dj@redhat.com>
10010 * config/rl78/constraints.md (Wcv): Allow up to $r31.
10011 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
10012 (rl78_option_override): Likewise, if -mallregs.
10013 (is_virtual_register): Likewise.
10014 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
10015 (REGNO_OK_FOR_BASE_P): Likewise.
10016 * config/rl78/rl78.opt (-mallregs): New.
10018 2013-09-17 Nick Clifton <nickc@redhat.com>
10020 * config/rl78/rl78.c (need_to_save): Change return type to bool.
10021 For interrupt functions: save all call clobbered registers if the
10022 interrupt handler is not a leaf function.
10023 (rl78_expand_prologue): Always recompute the frame information.
10024 For interrupt functions: only select bank 0 if one of the bank 0
10025 registers is going to be psuhed.
10027 2013-09-17 DJ Delorie <dj@redhat.com>
10029 * config/rl78/constraints.md: For each W* constraint, rename to C*
10030 and create a W* constraint that checks for an optional ES: prefix
10032 * config/rl78/rl78.md (UNS_ES_ADDR): New.
10033 (es_addr): New. Used to wrap far addresses.
10034 * config/rl78/rl78-protos.h (rl78_es_addr): New.
10035 (rl78_es_base): New.
10036 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
10037 wrapped far addresses.
10038 (rl78_print_operand_1): Unwrap far addresses before processing.
10039 (rl78_lo16): Wrap far addresses in unspecs.
10040 (rl78_es_addr): New.
10041 (rl78_es_base): New.
10042 (insn_ok_now): Check for not-yet-wrapped far addresses.
10043 (transcode_memory_rtx): Properly re-wrap far addresses.
10045 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
10047 * config/sparc/t-rtems: Add leon3 multilibs.
10049 2013-09-17 Cong Hou <congh@google.com>
10051 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
10052 when checking the dot production pattern. The type of rhs operand
10053 of multiply is now checked correctly.
10055 2013-09-17 Jeff Law <law@redhat.com>
10057 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
10058 edge implied equivalences into successor phis.
10059 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
10060 here from tree-ssa-threadedge.c.
10061 (mark_threaded_blocks): When threading through a joiner, if both
10062 successors of the joiner's clone reach the same block, verify the
10063 PHI arguments are equal. If not, cancel the jump threading request.
10064 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
10065 tree-ssa-threadupdate.c
10066 (thread_across_edge): Don't check PHI argument equality when
10067 threading through joiner block here.
10069 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
10071 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
10072 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
10073 (warn_uninit, warn_uninitialized_vars,
10074 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
10075 Move to tree-ssa-uninit.c.
10076 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
10077 (has_undefined_value_p): New. Pass dependant parts of
10078 ssa_undefined_value_p.
10079 (uninit_undefined_value_p): Use has_undefined_value_p.
10080 (warn_uninit, warn_uninitialized_vars,
10081 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
10082 Move from tree-ssa.c.
10083 * tree-ssa.h: Adjust prototypes.
10085 2013-09-17 Jan Hubicka <jh@suse.cz>
10087 PR middle-end/58332
10088 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
10089 * ipa-inline.c (can_inline_edge_p): Do not downgrade
10090 FUNCTION_NOT_OPTIMIZED.
10091 * ipa-inline-analysis.c (compute_inline_parameters): Function
10092 not optimized is not inlinable unless it is alwaysinline.
10093 (inline_analyze_function): Force calls in not optimized
10094 function not inlinable.
10096 2013-09-17 Jan Hubicka <jh@suse.cz>
10098 PR middle-end/58329
10099 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
10101 * ipa.c (function_and_variable_visibility): Likewise.
10102 * ipa-profile.c (ipa_profile): Likewise.
10104 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
10107 * cgraph.c (cgraph_function_body_availability): Check for ifunc
10108 attribute, and don't inline the resolver in this case.
10110 2013-09-17 Teresa Johnson <tejohnson@google.com>
10112 * coverage.c (get_coverage_counts): Add missing newline.
10114 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10116 PR tree-optimization/58088
10117 * fold-const.c (mask_with_trailing_zeros): New function.
10118 (fold_binary_loc): Make sure we don't recurse infinitely
10119 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
10120 Use mask_with_trailing_zeros where appropriate.
10122 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
10124 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
10125 of 'prev' var to get better distance estimation.
10127 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
10129 * tree-inline.h (struct copy_body_data): Add transform_parameter.
10130 * tree-inline.c (is_parameter_of): New predicate.
10131 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
10132 a parameter has been remapped.
10133 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
10134 (optimize_inline_calls): Initialize transform_parameter.
10135 (copy_gimple_seq_and_replace_locals): Likewise.
10136 (tree_function_versioning): Likewise.
10137 (maybe_inline_call_in_expr): Likewise.
10139 2013-09-17 Nick Clifton <nickc@redhat.com>
10141 * config/msp430/msp430-protos.h: Add prototypes for new functions.
10142 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
10143 interrupt handlers.
10144 (is_attr_func): New function.
10145 (msp430_is_interrupt_func): New function.
10146 (is_naked_func): New function.
10147 (is_reentrant_func): New function.
10148 (is_critical_func): New function.
10149 (msp430_start_function): Add annotations for function attributes.
10150 (msp430_attr): New function.
10151 (msp430_attribute_table): New.
10152 (msp430_function_section): New function.
10153 (TARGET_ASM_FUNCTION_SECTION): Define.
10154 (msp430_builtin): New enum.
10155 (msp430_init_builtins): New function.
10156 (msp430_builtin_devl): New function.
10157 (msp430_expand_builtin): New function.
10158 (TARGET_INIT_BUILTINS): Define.
10159 (TARGET_EXPAND_BUILTINS): Define.
10160 (TARGET_BUILTIN_DECL): Define.
10161 (msp430_expand_prologue): Add support for naked, interrupt,
10162 critical and reentrant functions.
10163 (msp430_expand_epilogue): Likewise.
10164 (msp430_print_operand): Handle 'O' character.
10165 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
10167 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
10168 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
10169 (pushm): Use a 'n' rather than an 'i' constraint.
10170 (msp_return): Add generation of the interrupt return instruction.
10171 (disable_interrupts): New pattern.
10172 (enable_interrupts): New pattern.
10173 (push_intr_state): New pattern.
10174 (pop_intr_state): New pattern.
10175 (bic_SR): New pattern.
10176 (bis_SR): New pattern.
10177 * doc/extend.texi: Document MSP430 function attributes and builtin
10180 2013-09-17 Richard Biener <rguenther@suse.de>
10182 PR tree-optimization/58432
10183 * tree-loop-distribution.c (tree_loop_distribution): Also
10184 scan PHIs for outside loop uses and seed a partition from them.
10186 2013-09-17 Bin Cheng <bin.cheng@arm.com>
10188 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
10189 (restructure_reference): Call backtrace_base_for_ref.
10191 2013-09-17 Alan Modra <amodra@gmail.com>
10194 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
10197 2013-09-16 DJ Delorie <dj@redhat.com>
10199 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
10200 vregs location for RL78/G10.
10201 (rl78_expand_prologue): Avoid SEL on G10.
10202 (rl78_expand_epilogue): Likewise.
10203 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
10204 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
10205 __RL78_G10__ when appropriate.
10206 (ASM_SPEC): Pass -mg10 along to the assembler.
10207 * config/rl78/rl78.md (sel_rb): Disable for G10.
10208 * config/rl78/rl78.opt: Add -mg10 option.
10209 * config/rl78/t-rl78: Add -mg10 multilib.
10211 2013-09-16 Xinliang David Li <davidxl@google.com>
10213 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
10214 * omp-low.c (omp_max_vf): Ditto.
10215 (expand_omp_simd): Ditto.
10216 * tree-vectorizer.c (vectorize_loops): Ditto.
10217 (gate_vect_slp): Ditto.
10218 (gate_increase_alignment): Ditto.
10219 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
10220 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
10221 (gate_tree_vectorize): Name change.
10222 (tree_vectorize): Ditto.
10223 (pass_vectorize::gate): Call new function.
10224 (pass_vectorize::execute): Ditto.
10225 * opts.c: O3 default setting change.
10226 (finish_options): Check new flag.
10227 * doc/invoke.texi: Document new flags.
10228 * common.opt: New flags.
10230 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
10232 * doc/tm.texi.in (Cond Exec Macros): Remove node.
10233 (Condition Code): Don't reference it.
10234 * doc/tm.texi: Regenerate.
10236 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
10238 PR middle-end/58418
10239 * lra-constraints.c (undo_optional_reloads): Consider all optional
10240 reload even if it did not get a hard reg.
10242 2013-09-16 Teresa Johnson <tejohnson@google.com>
10244 * dumpfile.c (dump_loc): Remove newline emission.
10245 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
10246 emission to dump_printf_loc calls where missing.
10247 (vect_mark_for_runtime_alias_test): Ditto.
10248 (vect_analyze_data_ref_dependence): Ditto.
10249 (vect_analyze_data_ref_dependences): Ditto.
10250 (vect_slp_analyze_data_ref_dependence): Ditto.
10251 (vect_slp_analyze_data_ref_dependences): Ditto.
10252 (vect_compute_data_ref_alignment): Ditto.
10253 (vect_update_misalignment_for_peel): Ditto.
10254 (vect_verify_datarefs_alignment): Ditto.
10255 (vector_alignment_reachable_p): Ditto.
10256 (vect_get_data_access_cost): Ditto.
10257 (vect_enhance_data_refs_alignment): Ditto.
10258 (vect_find_same_alignment_drs): Ditto.
10259 (vect_analyze_data_refs_alignment): Ditto.
10260 (vect_analyze_group_access): Ditto.
10261 (vect_analyze_data_ref_access): Ditto.
10262 (vect_analyze_data_ref_accesses): Ditto.
10263 (vect_prune_runtime_alias_test_list): Ditto.
10264 (vect_analyze_data_refs): Ditto.
10265 (vect_create_addr_base_for_vector_ref): Ditto.
10266 (vect_create_data_ref_ptr): Ditto.
10267 (vect_grouped_store_supported): Ditto.
10268 (vect_grouped_load_supported): Ditto.
10269 * value-prof.c (check_counter): Ditto.
10270 (check_ic_target): Ditto.
10271 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
10272 (vect_recog_widen_mult_pattern): Ditto.
10273 (vect_recog_widen_sum_pattern): Ditto.
10274 (vect_recog_over_widening_pattern): Ditto.
10275 (vect_recog_widen_shift_pattern): Ditto.
10276 (vect_recog_rotate_pattern): Ditto.
10277 (vect_recog_vector_vector_shift_pattern): Ditto.
10278 (vect_recog_divmod_pattern): Ditto.
10279 (vect_recog_mixed_size_cond_pattern): Ditto.
10280 (vect_recog_bool_pattern): Ditto.
10281 (vect_pattern_recog_1): Ditto.
10282 (vect_pattern_recog): Ditto.
10283 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
10284 (vect_is_simple_iv_evolution): Ditto.
10285 (vect_analyze_scalar_cycles_1): Ditto.
10286 (vect_get_loop_niters): Ditto.
10287 (vect_analyze_loop_1): Ditto.
10288 (vect_analyze_loop_form): Ditto.
10289 (vect_analyze_loop_operations): Ditto.
10290 (vect_analyze_loop_2): Ditto.
10291 (vect_analyze_loop): Ditto.
10292 (report_vect_op): Ditto.
10293 (vect_is_slp_reduction): Ditto.
10294 (vect_is_simple_reduction_1): Ditto.
10295 (vect_get_known_peeling_cost): Ditto.
10296 (vect_estimate_min_profitable_iters): Ditto.
10297 (vect_model_reduction_cost): Ditto.
10298 (vect_model_induction_cost): Ditto.
10299 (get_initial_def_for_induction): Ditto.
10300 (vect_create_epilog_for_reduction): Ditto.
10301 (vectorizable_reduction): Ditto.
10302 (vectorizable_induction): Ditto.
10303 (vectorizable_live_operation): Ditto.
10304 (vect_loop_kill_debug_uses): Ditto.
10305 (vect_transform_loop): Ditto.
10306 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
10307 (vect_stmt_relevant_p): Ditto.
10308 (process_use): Ditto.
10309 (vect_mark_stmts_to_be_vectorized): Ditto.
10310 (vect_model_simple_cost): Ditto.
10311 (vect_model_promotion_demotion_cost): Ditto.
10312 (vect_model_store_cost): Ditto.
10313 (vect_get_store_cost): Ditto.
10314 (vect_model_load_cost): Ditto.
10315 (vect_get_load_cost): Ditto.
10316 (vect_init_vector_1): Ditto.
10317 (vect_get_vec_def_for_operand): Ditto.
10318 (vect_finish_stmt_generation): Ditto.
10319 (vectorizable_call): Ditto.
10320 (vectorizable_conversion): Ditto.
10321 (vectorizable_assignment): Ditto.
10322 (vectorizable_shift): Ditto.
10323 (vectorizable_operation): Ditto.
10324 (vectorizable_store): Ditto.
10325 (vectorizable_load): Ditto.
10326 (vectorizable_condition): Ditto.
10327 (vect_analyze_stmt): Ditto.
10328 (vect_transform_stmt): Ditto.
10329 (vect_is_simple_use): Ditto.
10330 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
10331 (vect_can_advance_ivs_p): Ditto.
10332 (vect_update_ivs_after_vectorizer): Ditto.
10333 (vect_do_peeling_for_loop_bound): Ditto.
10334 (vect_gen_niters_for_prolog_loop): Ditto.
10335 (vect_update_inits_of_drs): Ditto.
10336 (vect_create_cond_for_alias_checks): Ditto.
10337 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
10338 (vect_build_slp_tree_1): Ditto.
10339 (vect_supported_load_permutation_p): Ditto.
10340 (vect_analyze_slp_instance): Ditto.
10341 (vect_analyze_slp): Ditto.
10342 (vect_make_slp_decision): Ditto.
10343 (vect_detect_hybrid_slp): Ditto.
10344 (vect_bb_vectorization_profitable_p): Ditto.
10345 (vect_slp_analyze_bb_1): Ditto.
10346 (vect_update_slp_costs_according_to_vf): Ditto.
10347 (vect_get_mask_element): Ditto.
10348 (vect_transform_slp_perm_load): Ditto.
10349 (vect_schedule_slp_instance): Ditto.
10350 (vect_schedule_slp): Ditto.
10351 (vect_slp_transform_bb): Ditto.
10352 * profile.c (read_profile_edge_counts): Ditto.
10353 (compute_branch_probabilities): Ditto.
10354 * coverage.c (get_coverage_counts): Ditto.
10356 2013-09-16 Diego Novillo <dnovillo@google.com>
10358 * tree-core.h: Add missing comment lines from refactoring of tree.h.
10360 2013-09-16 Jan Hubicka <jh@suse.cz>
10362 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
10363 abstract functions; for static functions check the presence of body.
10365 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
10367 * config/aarch64/aarch64-simd-builtins.def (fma): New.
10368 * config/aarch64/aarch64-simd.md
10369 (aarch64_mla_elt<mode>): New.
10370 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
10371 (aarch64_mls_elt<mode>): Likewise.
10372 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
10373 (aarch64_fma4_elt<mode>): Likewise.
10374 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
10375 (aarch64_fma4_elt_to_128v2df): Likewise.
10376 (aarch64_fma4_elt_to_64df): Likewise.
10377 (fnma<mode>4): Likewise.
10378 (aarch64_fnma4_elt<mode>): Likewise.
10379 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
10380 (aarch64_fnma4_elt_to_128v2df): Likewise.
10381 (aarch64_fnma4_elt_to_64df): Likewise.
10382 * config/aarch64/iterators.md (VDQSF): New.
10383 * config/aarch64/arm_neon.h
10384 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
10385 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
10387 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
10389 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
10390 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
10391 (aarch64_mul3_elt_to_128df): Likewise.
10392 (aarch64_mul3_elt_to_64v2df): Likewise.
10393 * config/aarch64/iterators.md (VEL): Also handle DFmode.
10395 (VMUL_CHANGE_NLANES) Likewise.
10398 * config/aarch64/arm_neon.h
10399 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
10401 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
10403 * config/aarch64/arm_neon.h
10404 (vcvtx_high_f32_f64): Fix parameters.
10406 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
10407 Uros Bizjak <ubizjak@gmail.com>
10409 * config/alpha.c: Include tree-ssanames.h.
10411 2013-09-16 Richard Biener <rguenther@suse.de>
10413 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
10414 (dot_rdg_1): Handle control_dd.
10415 (create_edge_for_control_dependence): New function.
10416 (create_rdg_edges): Add control dependences if asked for.
10417 (build_rdg): Likewise.
10418 (generate_loops_for_partition): If there are not necessary
10419 control stmts remove all their dependencies.
10420 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
10421 (distribute_loop): Pass on control dependences.
10422 (tree_loop_distribution): Compute control dependences and remove
10423 restriction on number of loop nodes.
10425 2013-09-16 Jakub Jelinek <jakub@redhat.com>
10427 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
10428 for internal calls.
10430 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
10432 * cse.c (try_const_anchors): Punt on CC modes.
10434 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
10436 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
10438 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
10441 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
10442 calls to word_mode.
10444 2013-09-14 Iain Sandoe <iain@codesourcery.com>
10447 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
10449 (darwin_objc1_section): Likewise.
10450 (darwin_file_end): Emit Image Info section when required.
10452 2013-09-14 Jan Hubicka <jh@suse.cz>
10454 * tree-into-ssa.c (gate_into_ssa): New.
10455 (pass_data_build_ssa): Use it.
10456 * cgraph.h (expand_thunk): Update prototype.
10457 * cgraphunit.c (analyze_function): Expand thunks early.
10458 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
10459 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
10460 set lowered flag; do not add new function.
10461 (assemble_thunks_and_aliases): Update.
10462 * tree-ssa.c (gate_init_datastructures): New gate.
10463 (pass_data_init_datastructures): Use it.
10465 2013-09-14 Iain Sandoe <iain@codesourcery.com>
10468 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
10469 xmm register set described in the psABI.
10471 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
10473 * dwarf2out.c (should_emit_struct_debug): Add check
10474 for type_decl variable is not NULL.
10476 2013-09-13 Jacek Caban <jacek@codeweavers.com>
10478 * config.gcc: Use new winnt-c.c target hooks
10479 * config/t-winnt: New file
10480 * config/winnt-c.c: New file
10481 * doc/tm.texi.in: Document new hook
10482 * doc/tm.texi: Regenerated
10484 2013-09-13 Jan Hubicka <jh@suse.cz>
10486 PR middle-end/58094
10487 * ipa-inline.c (check_callers): New function.
10488 (check_caller_edge): Remove.
10489 (want_inline_function_to_all_callers_p): Also permit alises that are
10491 (inline_to_all_callers): Terminate the walk when devirtualization
10492 introduce new calls.
10494 2013-09-13 Jan Hubicka <jh@suse.cz>
10496 * ipa-inline-analysis.c (struct growth_data): Add node.
10497 (do_estimate_growth_1): Fix detection of recursion.
10499 2013-09-13 Jakub Jelinek <jakub@redhat.com>
10501 PR tree-optimization/58392
10502 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
10503 to avoid shadowing of outer loop variable. If
10504 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
10505 replace_by_duplicate_decl simduid of loops that have it set and
10506 set dest_cfun->has_simduid_loops and/or
10507 dest_cfun->has_force_vect_loops.
10508 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
10509 instead of maybe_lookup_decl.
10510 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
10511 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
10512 call. Copy over force_vect and copy and remap simduid. Set
10513 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
10514 (copy_cfg_body): Remove blocks_to_copy argument. Use
10515 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
10516 caller. Don't set cfun->has_simduid_loops and/or
10517 cfun->has_force_vect_loops here.
10518 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
10520 (expand_call_inline, tree_function_versioning): Adjust copy_body
10523 2013-09-13 Martin Jambor <mjambor@suse.cz>
10526 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
10527 the assert if the edge was a speculative one.
10529 2013-09-13 Richard Biener <rguenther@suse.de>
10531 * tree-data-ref.h (known_dependences_p): Move here ...
10532 * tree-loop-distribution.c (known_dependences_p): ... from here.
10533 (dump_rdg_component, debug_rdg_component): Remove.
10534 (dump_rdg): Adjust.
10535 (generate_loops_for_partition): Use gimple_uid instead of
10536 relying on matching stmt visit order.
10537 (rdg_build_partitions): Take starting stmt vector.
10538 (ldist_gen): Merge into ...
10539 (distribute_loop): ... this function. Do not compute starting
10541 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
10543 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10545 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
10546 Set type attribute correctly. Set predicable_short_it attribute.
10547 (cmpsi_shiftsi): Remove %? from output template.
10549 2013-09-13 Richard Biener <rguenther@suse.de>
10551 * tree-loop-distribution.c (struct rdg_component,
10552 rdg_defs_used_in_other_loops_p, free_rdg_components,
10553 rdg_build_components): Remove.
10554 (stmts_from_loop): Do not record virtual PHIs.
10555 (generate_loops_for_partition): Skip virtual PHIs.
10556 (build_rdg_partition_for_component): Rename to ...
10557 (build_rdg_partition_for_vertex): ... this and adjust.
10558 (rdg_build_partitions): Take a vector of starting vertices
10559 instead of components. Remove unnecessary leftover handling.
10560 (ldist_gen): Do not build components or record other stores.
10561 (distribute_loop): Do not distribute loops containing stmts
10564 2013-09-13 Christian Bruel <christian.bruel@st.com>
10567 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
10569 2013-09-13 Kai Tietz <ktietz@redhat.com>
10571 * config.gcc: Separate cases for mingw and cygwin targets,
10572 and add 64-bit cygwin target case.
10574 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
10575 dll-export inline-functions.
10576 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
10578 2013-09-13 Jeff Law <law@redhat.com>
10580 PR middle-end/58387
10582 2013-09-06 Jeff Law <law@redhat.com>
10584 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
10585 edge implied equivalences into successor phis.
10587 2013-09-12 DJ Delorie <dj@redhat.com>
10589 * config/rl78/rl78-virt.md: Change from | to \; for asm line
10592 2013-09-12 Brooks Moses <bmoses@google.com>
10595 * Makefile.in: Do not install driver binaries in $(target)/bin.
10597 2013-09-12 DJ Delorie <dj@redhat.com>
10599 * config/rl78/rl78.opt (mrelax): New.
10600 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
10601 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
10603 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
10605 (rl78_expand_epilogue): Likewise.
10607 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
10609 PR middle-end/58335
10610 * lra-eliminations.c (remove_reg_equal_offset_note): New.
10611 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
10612 pointer elimination with using remove_reg_equal_offset_note.
10614 2013-09-12 DJ Delorie <dj@redhat.com>
10616 * config/msp430/: New port.
10617 * config.gcc (msp430): Added.
10618 * doc/invoke.texi: Document MSP430 options.
10619 * doc/install.texi: Document msp430-elf
10620 * doc/md.texi: Document msp430-elf
10621 * doc/contrib.texi: Document msp430-elf
10623 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
10626 2013-09-12 Martin Jambor <mjambor@suse.cz>
10629 * ipa-prop.c (remove_described_reference): Give up if the edge in the
10630 reference descriptor is NULL.
10631 (ipa_edge_removal_hook): If owning a reference descriptor, set its
10634 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
10636 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
10637 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
10638 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
10639 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
10640 make_temp_ssa_name): move to tree-ssanames.h
10641 * tree-ssa-alias.h: Move prototype.
10642 * tree-ssa.h: Include tree-ssanames.h.
10643 * tree-ssanames.c (FREE_SSANAMES): Move to here.
10644 * tree-ssanames.h: New. Move items from tree-flow*.h
10645 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
10647 2013-09-12 Richard Biener <rguenther@suse.de>
10649 PR tree-optimization/58404
10650 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
10651 propagate non-invariant addresses into dereferences wrapped
10652 in component references.
10654 2013-09-12 Richard Biener <rguenther@suse.de>
10656 PR tree-optimization/58402
10657 * passes.def: Move pass_late_warn_uninitialized later.
10659 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
10661 * tree-ssa.h: New. Move content from tree-flow.h and
10662 tree-flow-inline.h.
10663 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
10664 Move prototypes belonging to tree-ssa.c.
10665 * tree-flow-inline.h (redirect_edge_var_map_def,
10666 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
10668 * gimple.h: Adjust prototypes.
10669 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
10671 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
10672 * tree.h: Move prototype to tree-ssa.h.
10673 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
10674 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
10675 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
10676 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
10677 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
10678 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
10679 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
10680 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
10681 graphite-blocking.c, graphite-clast-to-gimple.c,
10682 graphite-dependences.c, graphite-interchange.c,
10683 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
10684 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
10685 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
10686 ipa-reference.c, ipa-split.c, ipa-utils.c,
10687 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
10688 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
10689 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
10690 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
10691 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
10692 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
10693 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
10694 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
10695 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
10696 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
10697 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
10698 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
10699 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
10700 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
10701 value-prof.c, var-tracking.c,
10702 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
10704 2013-09-12 Richard Biener <rguenther@suse.de>
10706 PR tree-optimization/58396
10707 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
10708 (build_rdg): Take a loop-nest parameter, fix memleaks.
10709 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
10711 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
10713 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
10716 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
10718 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
10719 The FMA instruction names should have a 'v' prefix.
10721 2013-09-12 Richard Biener <rguenther@suse.de>
10723 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
10724 (dot_rdg): Use popen instead of system in optional code.
10725 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
10726 (already_processed_vertex_p): Adjust.
10727 (has_anti_or_output_dependence, predecessor_has_mem_write,
10728 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
10729 rdg_flag_uses): Remove.
10730 (rdg_flag_vertex): Simplify.
10731 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
10733 (build_rdg_partition_for_component): Process the first vertex
10734 of a component only.
10735 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
10737 2013-09-12 Alan Modra <amodra@gmail.com>
10739 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
10741 2013-09-11 DJ Delorie <dj@redhat.com>
10742 Nick Clifton <nickc@redhat.com>
10744 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
10745 (rl78_stack_based_mem): New.
10746 * config/rl78/constraints.md (Iv08): New.
10755 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
10757 (movsi): Change from expand to insn-and-split.
10758 (ashrsi3): Clobber AX.
10762 * config/rl78/rl78.md (CC_REG): Fix.
10763 (addsi3): Allow memory and immediate operands.
10764 (addsi3_internal): Split into...
10765 (addsi3_internal_virt): ...new, and ...
10766 (addsi3_internal_real): ...new.
10768 (subsi3_internal_virt): New.
10769 (subsi3_internal_real): New.
10770 (mulsi3): Add memory operand.
10771 (mulsi3_rl78): Likewise.
10772 (mulsi3_g13): Likewise.
10773 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
10774 (cbranchqi4_real): Add more constraint options.
10775 (cbranchhi4_real): Expand pattern.
10776 (cbranchhi4_real_signed): New.
10777 (cbranchhi4_real_inverted): New.
10778 (cbranchsi4_real_lt): New.
10779 (cbranchsi4_real_ge): New.
10780 (cbranchsi4_real_signed): New.
10781 (cbranchsi4_real): New.
10783 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
10785 (lshrsi3_virt): Likewise.
10786 (ashlsi3_virt): Likewise.
10787 (cbranchqi4_virt_signed): New.
10788 (cbranchhi4_virt_signed): New.
10789 (cbranchsi4_virt): New.
10790 * config/rl78/rl78.c: Whitespace fixes throughout.
10791 (move_elim_pass): New.
10792 (pass_data_rl78_move_elim): New.
10793 (pass_rl78_move_elim): New.
10794 (make_pass_rl78_move_elim): New.
10795 (rl78_devirt_info): Run devirt earlier.
10796 (rl78_move_elim_info): New.
10797 (rl78_asm_file_start): Register it.
10798 (rl78_split_movsi): New.
10799 (rl78_as_legitimate_address): Allow virtual base registers when
10801 (rl78_addr_space_convert): Remove spurious debug stuff.
10802 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
10803 (rl78_print_operand): More cases for not printing '#'.
10804 (rl78_expand_compare): Remove most of the logic.
10805 (content_memory): New.
10806 (clear_content_memory): New.
10807 (get_content_index): New.
10808 (get_content_name): New.
10809 (display_content_memory): New.
10810 (update_content): New.
10811 (record_content): New.
10812 (already_contains): New.
10813 (insn_ok_now): Re-recog insns with virtual registers.
10814 (add_postponed_content_update): New.
10815 (process_postponed_content_update): New.
10816 (gen_and_emit_move): New.
10817 (transcode_memory_rtx): Record new location content.
10818 Use gen_and_emit_move.
10819 (force_into_acc): New.
10820 (move_to_acc): Use gen_and_emit_move.
10821 (move_from_acc): Likewise.
10822 (move_acc_to_reg): Likewise.
10823 (move_to_x): Likewise.
10824 (move_to_hl): Likewise.
10825 (move_to_de): Likewise.
10826 (rl78_alloc_physical_registers_op1): Record location content.
10827 (has_constraint): New.
10828 (rl78_alloc_physical_registers_op2): Record location content.
10829 Optimize use of HL.
10830 (rl78_alloc_physical_registers_ro1): Likewise.
10831 (rl78_alloc_physical_registers_cmp): Likewise.
10832 (rl78_alloc_physical_registers_umul): Likewise.
10833 (rl78_alloc_address_registers_macax): New.
10834 (rl78_alloc_physical_registers): Initialize and set location
10835 content memory as needed.
10836 (rl78_reorg): Make sure split2 is called.
10837 (rl78_rtx_costs): New.
10839 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
10841 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
10842 for (not (neg ...)) and (neg (not ...)) cases.
10844 2013-09-11 Richard Biener <rguenther@suse.de>
10846 PR middle-end/58377
10847 * passes.def: Split critical edges before late uninit warning passes.
10848 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
10850 2013-09-11 Jakub Jelinek <jakub@redhat.com>
10852 PR tree-optimization/58385
10853 * fold-const.c (build_range_check): If both low and high are NULL,
10854 use omit_one_operand_loc to preserve exp side-effects.
10856 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10858 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
10860 2013-09-11 Richard Biener <rguenther@suse.de>
10862 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
10863 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
10864 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
10865 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
10866 stmts_from_loop, known_dependences_p, build_empty_rdg,
10867 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
10868 * tree-loop-distribution.c: ... here.
10869 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
10870 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
10871 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
10872 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
10873 * tree-loop-distribution.c: ... here.
10874 * tree-loop-distribution.c: Include gimple-pretty-print.h.
10875 (struct partition_s): Add loops member.
10876 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
10877 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
10878 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
10880 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
10881 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
10882 Sergey Lega <sergey.s.lega@intel.com>
10883 Anna Tikhonova <anna.tikhonova@intel.com>
10884 Ilya Tocar <ilya.tocar@intel.com>
10885 Andrey Turetskiy <andrey.turetskiy@intel.com>
10886 Ilya Verbin <ilya.verbin@intel.com>
10887 Kirill Yukhin <kirill.yukhin@intel.com>
10888 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
10890 * config/i386/constraints.md (k): New.
10892 * config/i386/i386.c (const regclass_map): Add new mask registers.
10893 (dbx_register_map): Ditto.
10894 (dbx64_register_map): Ditto.
10895 (svr4_dbx_register_map): Ditto.
10896 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
10898 (ix86_preferred_reload_class): Disable constants for mask registers.
10899 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
10900 (ix86_hard_regno_mode_ok): Support new mask registers.
10901 (x86_order_regs_for_local_alloc): Ditto.
10902 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
10903 (FIXED_REGISTERS): Add new mask registers.
10904 (CALL_USED_REGISTERS): Ditto.
10905 (REG_ALLOC_ORDER): Ditto.
10906 (VALID_MASK_REG_MODE): New.
10907 (FIRST_MASK_REG): Ditto.
10908 (LAST_MASK_REG): Ditto.
10909 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
10910 (MAYBE_MASK_CLASS_P): New.
10911 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
10912 (REG_CLASS_CONTENTS): Ditto.
10913 (MASK_REGNO_P): New.
10914 (ANY_MASK_REG_P): Ditto.
10915 (HI_REGISTER_NAMES): Add new mask registers.
10916 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
10917 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
10919 (attribute "type"): Add mskmov, msklog.
10920 (attribute "length_immediate"): Support them.
10921 (attribute "memory"): Ditto.
10922 (attribute "prefix_0f"): Ditto.
10923 (*movhi_internal): Support new mask registers.
10924 (*movqi_internal): Ditto.
10925 (define_split): Split out clobber pattern is a logic
10926 insn on mask registers.
10927 (*k<logic><mode>): New.
10928 (*andhi_1): Extend to support mask regs.
10929 (*andqi_1): Extend to support mask regs.
10930 (kandn<mode>): New.
10931 (define_split): Split and-not to and and not if operands
10933 (*<code><mode>_1): Separate HI mode to new pattern...
10934 (*<code>hi_1): This.
10935 (*<code>qi_1): Extend to support mask regs.
10936 (kxnor<mode>): New.
10937 (kortestzhi): Ditto.
10938 (kortestchi): Ditto.
10940 (*one_cmpl<mode>2_1): Remove HImode and handle it...
10941 (*one_cmplhi2_1): ...Here, now with mask registers support.
10942 (*one_cmplqi2_1): Support new mask registers.
10943 (HI/QImode arithmetics splitter): Don't split if mask registers
10945 (HI/QImode not splitter): Ditto.
10946 * config/i386/predicated.md (mask_reg_operand): New.
10947 (general_reg_operand): Ditto.
10949 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
10951 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
10952 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
10954 2013-09-10 Jeff Law <law@redhat.com>
10956 PR tree-optimization/58380
10957 * tree-ssa-threadupdate.c (thread_block): Recognize another case
10958 of threading through a buried loop header.
10960 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
10961 return value for single successor case.
10963 2013-09-10 Jan Hubicka <jh@suse.cz>
10965 * ipa-devirt.c (ipa_devirt): Enable with LTO.
10967 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
10970 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
10971 support conditional execution.
10972 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
10974 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
10976 * lra.c (lra): Clear lra_optional_reload_pseudos before every
10978 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
10979 Check destination too to check move insn.
10980 (undo_optional_reloads): Add check that the original peudo did not
10981 changed its allocation and the optional reload was inherited on last
10982 inheritance pass. Break loop after deciding to keep optional reload.
10983 (lra_undo_inheritance): Add check that inherited pseudo still in
10986 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
10988 * config/aarch64/aarch64.md (generic_sched): New.
10989 * config/aarch64/aarch64-generic.md (load): Make conditional
10990 on generic_sched attribute.
10991 (nonload): Likewise.
10993 2013-09-10 Jan Hubicka <jh@suse.cz>
10995 * lto-cgraph.c: Include ipa-utils.h.
10996 (compute_ltrans_boundary): Also add possible targets into the boundary.
10998 2013-09-10 Jan Hubicka <jh@suse.cz>
11000 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
11001 VAR_DECL of vtable rather than full expression.
11003 2013-09-10 Jan Hubicka <jh@suse.cz>
11004 Paolo Carlini <paolo.carlini@oracle.com>
11006 * cgraphunit.c (analyze_functions): Save input_location, set it
11007 to UNKNOWN_LOCATION and restore it at the end.
11009 2013-09-10 Martin Jambor <mjambor@suse.cz>
11011 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
11012 represented by a thunk.
11014 2013-09-10 Jeff Law <law@redhat.com>
11016 PR tree-optimization/58343
11017 * tree-ssa-threadupdate.c (thread_block): Identify and disable
11018 jump threading requests through loop headers buried in the middle
11019 of a jump threading path.
11021 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
11022 in return value/type.
11024 2013-09-10 Jakub Jelinek <jakub@redhat.com>
11026 PR rtl-optimization/58365
11027 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
11028 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
11031 2013-09-10 Richard Biener <rguenther@suse.de>
11033 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
11034 * tree-data-ref.c (create_rdg_vertices): Collect all data
11035 references, signal failure to the caller, use data-ref API.
11036 (build_rdg): Compute data references only once. Maintain lifetime
11037 of data references and data dependences from within RDG.
11038 (free_rdg): Free dependence relations.
11039 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
11040 inventing extra dependences.
11041 (distribute_loop): Update for RDG API changes.
11043 2013-09-10 Kai Tietz <ktietz@redhat.com>
11045 * doc/invoke.texi (fms-extensions): Document changed
11046 behavior for ms-abi targets.
11047 * config/i386/i386.c (ix86_option_override_internal):
11048 Set default value of option -fms-extension for ms-abi targets.
11050 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
11052 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
11054 2013-09-10 Alan Modra <amodra@gmail.com>
11057 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
11059 2013-09-10 Alan Modra <amodra@gmail.com>
11061 * config/rs6000/predicates.md (add_cint_operand): New.
11062 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
11063 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
11064 using add_cint_operand.
11065 (largetoc_high_plus_aix): Likewise.
11067 2013-09-09 Jakub Jelinek <jakub@redhat.com>
11069 PR tree-optimization/58364
11070 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
11071 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
11072 the current range can't be an unconditional true or false.
11074 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
11076 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
11078 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
11080 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
11082 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
11085 * doc/invoke.texi (-Wdelete-incomplete): Document it.
11087 2013-09-09 Ian Bolton <ian.bolton@arm.com>
11089 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
11090 NO_REGS for immediate that can't be moved directly into FP_REGS.
11092 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11094 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
11095 comparison with negated operand.
11096 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
11099 2013-09-09 Richard Biener <rguenther@suse.de>
11101 PR middle-end/58326
11102 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
11103 of a subloop record all its block as affecting loop-closed SSA form.
11105 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
11107 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
11108 of an rtx/bitpos pair.
11109 (store_fixed_bit_field): Update accordingly.
11111 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
11113 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
11115 * builtins.c (expand_errno_check): Likewise.
11116 * dwarf2cfi.c (init_return_column_size): Likewise.
11117 * except.c (sjlj_mark_call_sites): Likewise.
11118 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
11119 * lra-constraints.c (emit_inc): Likewise.
11120 * ree.c (combine_set_extension): Likewise.
11121 * regmove.c (fixup_match_2): Likewise.
11122 * reload1.c (inc_for_reload): Likewise.
11124 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
11126 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
11127 (force_to_mode, simplify_shift_const_1, simplify_comparison):
11128 Use gen_int_mode with the mode of the associated simplify_* call.
11129 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
11130 * expmed.c (expand_shift_1): Likewise.
11131 * function.c (instantiate_virtual_regs_in_insn): Likewise.
11132 * loop-iv.c (iv_number_of_iterations): Likewise.
11133 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
11134 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
11136 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
11138 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
11139 of the associated expand_* call.
11140 (asan_emit_stack_protection): Likewise.
11141 * builtins.c (round_trampoline_addr): Likewise.
11142 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
11143 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
11144 (emit_store_flag): Likewise.
11145 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
11146 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
11148 * function.c (instantiate_virtual_regs_in_insn): Likewise.
11149 * ifcvt.c (noce_try_store_flag_constants): Likewise.
11150 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
11151 * modulo-sched.c (generate_prolog_epilog): Likewise.
11152 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
11153 (expand_ctz, expand_ffs, expand_unop): Likewise.
11155 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
11157 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
11158 of the associated gen_rtx_* call.
11159 * caller-save.c (init_caller_save): Likewise.
11160 * combine.c (find_split_point, make_extraction): Likewise.
11161 (make_compound_operation): Likewise.
11162 * dwarf2out.c (mem_loc_descriptor): Likewise.
11163 * explow.c (plus_constant, probe_stack_range): Likewise.
11164 * expmed.c (expand_mult_const): Likewise.
11165 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
11166 * reload1.c (init_reload): Likewise.
11167 * valtrack.c (cleanup_auto_inc_dec): Likewise.
11168 * var-tracking.c (adjust_mems): Likewise.
11169 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
11170 rather than gen_rtx_fmt_ee.
11172 2013-09-09 Jan Hubicka <jh@suse.cz>
11174 PR middle-end/58294
11175 * value-prof.c (gimple_ic): Copy also abnormal edges.
11177 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
11179 * asan.c (asan_shadow_cst): Use gen_int_mode.
11181 2013-09-08 Jan Hubicka <jh@suse.cz>
11183 * ipa-profile.c: Add toplevel comment.
11184 (ipa_propagate_frequency_1): Be more conservative when profile is read.
11185 (contains_hot_call_p): New function.
11186 (ipa_propagate_frequency): Set frequencies based on counts when
11188 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
11189 profile; do not tamper with profile after inlining if it is read.
11191 2013-09-08 Jan Hubicka <jh@suse.cz>
11193 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
11196 2013-09-08 Jan Hubicka <jh@suse.cz>
11198 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
11199 summary generation.
11201 2013-09-08 Jeff Law <law@redhat.com>
11204 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
11207 2013-09-08 Andi Kleen <ak@linux.intel.com>
11209 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
11211 2013-09-08 Jan Hubicka <jh@suse.cz>
11213 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
11215 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
11217 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
11218 for non-debug insns.
11219 * lra.c (new_insn_reg): Take the containing insn as a parameter.
11220 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
11221 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
11224 2013-09-08 Jan Hubicka <jh@suse.cz>
11226 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
11227 targets and devirtualize to BUILT_IN_UNREACHABLE.
11228 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
11229 * ipa.c (walk_polymorphic_call_targets): New function.
11230 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
11231 functions; use the new timevar.
11232 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
11233 was removed from the program.
11234 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
11235 not consider it in the walk when its vtable is dead.
11236 (possible_polymorphic_call_targets_1): Pass anonymous flag to
11238 (devirt_variable_node_removal_hook): New function.
11239 (possible_polymorphic_call_targets): Also register
11240 devirt_variable_node_removal_hook.
11241 (ipa_devirt): Do not do non-speculative devirtualization.
11242 (gate_ipa_devirt): One execute if devirtualizing speculatively.
11244 2013-09-08 Jan Hubicka <jh@suse.cz>
11246 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
11247 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
11248 varpool_remove_variable_insertion_hook): Declare.
11249 * varpool.c (varpool_node_hook_list): New structure.
11250 (first_varpool_node_removal_hook,
11251 first_varpool_variable_insertion_hook): New variables.
11252 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
11253 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
11254 varpool_remove_variable_insertion_hook,
11255 varpool_call_variable_insertion_hooks): New functions.
11256 (varpool_remove_node): Use it.
11258 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
11261 * diagnostic.c (diagnostic_build_prefix): When s.file is
11262 "<built-in>" don't output line and column numbers.
11264 2013-09-06 Jan Hubicka <jh@suse.cz>
11266 * cgraphunit.c (expand_thunk): Get body before touching arguments.
11267 * lto-streamer-out.c: Stream thunks, too.
11268 * lto-streamer-in.c (input_function): Pop cfun here
11269 (lto_read_body): Instead of here.
11271 2013-09-06 Caroline Tice <cmtice@google.com>
11273 * doc/install.texi: Add documentation for the --enable-vtable-verify
11274 and the --disable-libvtv configure options.
11276 2013-09-06 Jeff Law <law@redhat.com>
11278 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
11279 edge implied equivalences into successor phis.
11281 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
11283 * resource.c (mark_referenced_resources): Handle COND_EXEC.
11285 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
11287 * resource.c (mark_target_live_regs): Compute resources taking
11288 into account if a call is predicated or not.
11290 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
11292 * toplev.c (output_stack_usage): Be prepared for suffixes created by
11293 the compiler in the function names.
11295 2013-09-06 Jan Hubicka <jh@suse.cz>
11297 PR middle-end/58094
11298 * ipa-inline.c (has_caller_p): New function.
11299 (want_inline_function_to_all_callers_p): Use it.
11300 (sum_callers, inline_to_all_callers): Break out from ...
11301 (ipa_inline): ... here.
11303 2013-09-06 Jan Hubicka <jh@suse.cz>
11305 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
11306 only when AVX is enabled.
11308 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
11310 * config/aarch64/aarch64.md
11311 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
11312 is fpsimd_<load/store>2.
11313 (load_pair<mode>): Likewise.
11314 (store_pair<mode>): Likewise.
11316 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
11318 * config/arm/types.md (type): Add "mrs" type.
11319 * config/aarch64/aarch64.md
11320 (aarch64_load_tp_hard): Make type "mrs".
11321 * config/arm/arm.md
11322 (load_tp_hard): Make type "mrs".
11323 * config/arm/cortex-a15.md: Update with new attributes.
11324 * config/arm/cortex-a5.md: Update with new attributes.
11325 * config/arm/cortex-a53.md: Update with new attributes.
11326 * config/arm/cortex-a7.md: Update with new attributes.
11327 * config/arm/cortex-a8.md: Update with new attributes.
11328 * config/arm/cortex-a9.md: Update with new attributes.
11329 * config/arm/cortex-m4.md: Update with new attributes.
11330 * config/arm/cortex-r4.md: Update with new attributes.
11331 * config/arm/fa526.md: Update with new attributes.
11332 * config/arm/fa606te.md: Update with new attributes.
11333 * config/arm/fa626te.md: Update with new attributes.
11334 * config/arm/fa726te.md: Update with new attributes.
11336 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
11338 * config/aarch64/aarch64.md
11339 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
11340 (*movtf_aarch64): Likewise.
11341 * config/arm/arm.md
11342 (thumb1_movdi_insn): Use "multiple" for type where more than one
11343 instruction is used for a move.
11344 (*arm32_movhf): Likewise.
11345 (*thumb_movdf_insn): Likewise.
11347 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
11349 * config/arm/types.md (type): Rename fcpys to fmov.
11350 * config/arm/vfp.md
11351 (*arm_movsi_vfp): Rename type fcpys as fmov.
11352 (*thumb2_movsi_vfp): Likewise
11353 (*movhf_vfp_neon): Likewise
11354 (*movhf_vfp): Likewise
11355 (*movsf_vfp): Likewise
11356 (*thumb2_movsf_vfp): Likewise
11357 (*movsfcc_vfp): Likewise
11358 (*thumb2_movsfcc_vfp): Likewise
11359 * config/aarch64/aarch64-simd.md
11360 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
11361 * config/aarch64/aarch64.md
11362 (*movsi_aarch64): Replace type mov_reg with fmovs.
11363 (*movdi_aarch64): Likewise
11364 (*movsf_aarch64): Likewise
11365 (*movdf_aarch64): Likewise
11367 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
11368 * config/arm/iwmmxt.md
11369 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
11370 * config/arm/arm1020e.md: Update with new attributes.
11371 * config/arm/cortex-a15-neon.md: Update with new attributes.
11372 * config/arm/cortex-a5.md: Update with new attributes.
11373 * config/arm/cortex-a53.md: Update with new attributes.
11374 * config/arm/cortex-a7.md: Update with new attributes.
11375 * config/arm/cortex-a8-neon.md: Update with new attributes.
11376 * config/arm/cortex-a9.md: Update with new attributes.
11377 * config/arm/cortex-m4-fpu.md: Update with new attributes.
11378 * config/arm/cortex-r4f.md: Update with new attributes.
11379 * config/arm/marvell-pj4.md: Update with new attributes.
11380 * config/arm/vfp11.md: Update with new attributes.
11382 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
11384 * config/aarch64/aarch64.md
11385 (*madd<mode>): Fix type attribute.
11386 (*maddsi_uxtw): Likewise.
11387 (*msub<mode>): Likewise.
11388 (*msubsi_uxtw): Likewise.
11389 (<su_optab>maddsidi4): Likewise.
11390 (<su_optab>msubsidi4): Likewise.
11392 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
11394 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
11395 * config/arm/arm.md (core_cycles): Remove fdiv.
11396 * config/arm/vfp.md:
11397 (*sqrtsf2_vfp): Update for attribute changes.
11398 (*sqrtdf2_vfp): Likewise.
11399 * config/aarch64/aarch64.md:
11400 (sqrt<mode>2): Update for attribute changes.
11401 * config/arm/arm1020e.md: Update with new attributes.
11402 * config/arm/cortex-a15-neon.md: Update with new attributes.
11403 * config/arm/cortex-a5.md: Update with new attributes.
11404 * config/arm/cortex-a53.md: Update with new attributes.
11405 * config/arm/cortex-a7.md: Update with new attributes.
11406 * config/arm/cortex-a8-neon.md: Update with new attributes.
11407 * config/arm/cortex-a9.md: Update with new attributes.
11408 * config/arm/cortex-m4-fpu.md: Update with new attributes.
11409 * config/arm/cortex-r4f.md: Update with new attributes.
11410 * config/arm/marvell-pj4.md: Update with new attributes.
11411 * config/arm/vfp11.md: Update with new attributes.
11413 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
11415 * config/arm/types.md
11416 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
11417 * config/aarch64/aarch64.md
11418 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
11420 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
11421 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
11422 (float<GPI:mode><GPF:mode>2): Likewise.
11423 * config/arm/vfp.md
11424 (*truncsisf2_vfp): Update with new attributes.
11425 (*truncsidf2_vfp): Likewise.
11426 (fixuns_truncsfsi2): Likewise.
11427 (fixuns_truncdfsi2): Likewise.
11428 (*floatsisf2_vfp): Likewise.
11429 (*floatsidf2_vfp): Likewise.
11430 (floatunssisf2): Likewise.
11431 (floatunssidf2): Likewise.
11432 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
11433 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
11434 * config/arm/arm1020e.md: Update with new attributes.
11435 * config/arm/cortex-a15-neon.md: Update with new attributes.
11436 * config/arm/cortex-a5.md: Update with new attributes.
11437 * config/arm/cortex-a53.md: Update with new attributes.
11438 * config/arm/cortex-a7.md: Update with new attributes.
11439 * config/arm/cortex-a8-neon.md: Update with new attributes.
11440 * config/arm/cortex-a9.md: Update with new attributes.
11441 * config/arm/cortex-m4-fpu.md: Update with new attributes.
11442 * config/arm/cortex-r4f.md: Update with new attributes.
11443 * config/arm/marvell-pj4.md: Update with new attributes.
11444 * config/arm/vfp11.md: Update with new attributes.
11446 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
11448 * config/aarch64/arm_neon.h
11449 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
11450 (vqtbx<1,2,3,4><q>_s8): Likewise.
11452 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
11454 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
11455 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
11456 (add<mode>3): Add type attribute.
11457 (add<mode>3): Likewise.
11458 (usadd<mode>3): Likewise.
11459 (ssadd<mode>3): Likewise.
11460 (sub<mode>3): Likewise.
11461 (sub<mode>3): Likewise.
11462 (ussub<mode>3): Likewise.
11463 (sssub<mode>3): Likewise.
11464 (ssmulsa3): Likewise.
11465 (usmulusa3): Likewise.
11466 (arm_usatsihi): Likewise.
11467 * config/arm/vfp.md
11468 (*movdi_vfp): Add types for all instructions.
11469 (*movdi_vfp_cortexa8): Likewise.
11470 (*movhf_vfp_neon): Likewise.
11471 (*movhf_vfp): Likewise.
11472 (*movdf_vfp): Likewise.
11473 (*thumb2_movdf_vfp): Likewise.
11474 (*thumb2_movdfcc_vfp): Likewise.
11475 * config/arm/arm.md: Add type attribute to all insn patterns.
11476 (*thumb1_adddi3): Add type attribute.
11477 (*arm_adddi3): Likewise.
11478 (*adddi_sesidi_di): Likewise.
11479 (*adddi_zesidi_di): Likewise.
11480 (*thumb1_addsi3): Likewise.
11481 (addsi3_compare0): Likewise.
11482 (*addsi3_compare0_scratch): Likewise.
11483 (*compare_negsi_si): Likewise.
11484 (cmpsi2_addneg): Likewise.
11485 (*addsi3_carryin_<optab>): Likewise.
11486 (*addsi3_carryin_alt2_<optab>): Likewise.
11487 (*addsi3_carryin_clobercc_<optab>): Likewise.
11488 (*subsi3_carryin): Likewise.
11489 (*subsi3_carryin_const): Likewise.
11490 (*subsi3_carryin_compare): Likewise.
11491 (*subsi3_carryin_compare_const): Likewise.
11492 (*arm_subdi3): Likewise.
11493 (*thumb_subdi3): Likewise.
11494 (*subdi_di_zesidi): Likewise.
11495 (*subdi_di_sesidi): Likewise.
11496 (*subdi_zesidi_di): Likewise.
11497 (*subdi_sesidi_di): Likewise.
11498 (*subdi_zesidi_ze): Likewise.
11499 (thumb1_subsi3_insn): Likewise.
11500 (*arm_subsi3_insn): Likewise.
11501 (*anddi3_insn): Likewise.
11502 (*anddi_zesidi_di): Likewise.
11503 (*anddi_sesdi_di): Likewise.
11504 (*ne_zeroextracts): Likewise.
11505 (*ne_zeroextracts): Likewise.
11506 (*ite_ne_zeroextr): Likewise.
11507 (*ite_ne_zeroextr): Likewise.
11508 (*anddi_notdi_di): Likewise.
11509 (*anddi_notzesidi): Likewise.
11510 (*anddi_notsesidi): Likewise.
11511 (andsi_notsi_si): Likewise.
11512 (thumb1_bicsi3): Likewise.
11513 (*iordi3_insn): Likewise.
11514 (*iordi_zesidi_di): Likewise.
11515 (*iordi_sesidi_di): Likewise.
11516 (*thumb1_iorsi3_insn): Likewise.
11517 (*xordi3_insn): Likewise.
11518 (*xordi_zesidi_di): Likewise.
11519 (*xordi_sesidi_di): Likewise.
11520 (*arm_xorsi3): Likewise.
11521 (*andsi_iorsi3_no): Likewise.
11522 (*smax_0): Likewise.
11523 (*smax_m1): Likewise.
11524 (*arm_smax_insn): Likewise.
11525 (*smin_0): Likewise.
11526 (*arm_smin_insn): Likewise.
11527 (*arm_umaxsi3): Likewise.
11528 (*arm_uminsi3): Likewise.
11529 (*minmax_arithsi): Likewise.
11530 (*minmax_arithsi_): Likewise.
11531 (*satsi_<SAT:code>): Likewise.
11532 (arm_ashldi3_1bit): Likewise.
11533 (arm_ashrdi3_1bit): Likewise.
11534 (arm_lshrdi3_1bit): Likewise.
11535 (*arm_negdi2): Likewise.
11536 (*thumb1_negdi2): Likewise.
11537 (*arm_negsi2): Likewise.
11538 (*thumb1_negsi2): Likewise.
11539 (*negdi_extendsid): Likewise.
11540 (*negdi_zero_extend): Likewise.
11541 (*arm_abssi2): Likewise.
11542 (*thumb1_abssi2): Likewise.
11543 (*arm_neg_abssi2): Likewise.
11544 (*thumb1_neg_abss): Likewise.
11545 (one_cmpldi2): Likewise.
11546 (extend<mode>di2): Likewise.
11547 (*compareqi_eq0): Likewise.
11548 (*arm_extendhisi2addsi): Likewise.
11549 (*arm_movdi): Likewise.
11550 (*thumb1_movdi_insn): Likewise.
11551 (*arm_movt): Likewise.
11552 (*thumb1_movsi_insn): Likewise.
11553 (pic_add_dot_plus_four): Likewise.
11554 (pic_add_dot_plus_eight): Likewise.
11555 (tls_load_dot_plus_eight): Likewise.
11556 (*thumb1_movhi_insn): Likewise.
11557 (*thumb1_movsf_insn): Likewise.
11558 (*movdf_soft_insn): Likewise.
11559 (*thumb_movdf_insn): Likewise.
11560 (cbranchsi4_insn): Likewise.
11561 (cbranchsi4_scratch): Likewise.
11562 (*negated_cbranchsi4): Likewise.
11563 (*tbit_cbranch): Likewise.
11564 (*tlobits_cbranch): Likewise.
11565 (*tstsi3_cbranch): Likewise.
11566 (*cbranchne_decr1): Likewise.
11567 (*addsi3_cbranch): Likewise.
11568 (*addsi3_cbranch_scratch): Likewise.
11569 (*arm_cmpdi_insn): Likewise.
11570 (*arm_cmpdi_unsig): Likewise.
11571 (*arm_cmpdi_zero): Likewise.
11572 (*thumb_cmpdi_zero): Likewise.
11573 (*deleted_compare): Likewise.
11574 (*mov_scc): Likewise.
11575 (*mov_negscc): Likewise.
11576 (*mov_notscc): Likewise.
11577 (*cstoresi_eq0_thumb1_insn): Likewise.
11578 (cstoresi_nltu_thumb1): Likewise.
11579 (cstoresi_ltu_thu): Likewise.
11580 (thumb1_addsi3_addgeu): Likewise.
11581 (*arm_jump): Likewise.
11582 (*thumb_jump): Likewise.
11583 (*check_arch2): Likewise.
11584 (arm_casesi_internal): Likewise.
11585 (thumb1_casesi_dispatch): Likewise.
11586 (*arm_indirect_jump): Likewise.
11587 (*thumb1_indirect_jump): Likewise.
11589 (*and_scc): Likewise.
11590 (*ior_scc): Likewise.
11591 (*compare_scc): Likewise.
11592 (*cond_move): Likewise.
11593 (*cond_arith): Likewise.
11594 (*cond_sub): Likewise.
11595 (*cmp_ite0): Likewise.
11596 (*cmp_ite1): Likewise.
11597 (*cmp_and): Likewise.
11598 (*cmp_ior): Likewise.
11599 (*ior_scc_scc): Likewise.
11600 (*ior_scc_scc_cmp): Likewise.
11601 (*and_scc_scc): Likewise.
11602 (*and_scc_scc_cmp): Likewise.
11603 (*and_scc_scc_nod): Likewise.
11604 (*negscc): Likewise.
11605 (movcond_addsi): Likewise.
11606 (movcond): Likewise.
11607 (*ifcompare_plus_move): Likewise.
11608 (*if_plus_move): Likewise.
11609 (*ifcompare_move_plus): Likewise.
11610 (*if_move_plus): Likewise.
11611 (*ifcompare_arith_arith): Likewise.
11612 (*if_arith_arith): Likewise.
11613 (*ifcompare_arith_move): Likewise.
11614 (*if_arith_move): Likewise.
11615 (*ifcompare_move_arith): Likewise.
11616 (*if_move_arith): Likewise.
11617 (*ifcompare_move_not): Likewise.
11618 (*if_move_not): Likewise.
11619 (*ifcompare_not_move): Likewise.
11620 (*if_not_move): Likewise.
11621 (*ifcompare_shift_move): Likewise.
11622 (*if_shift_move): Likewise.
11623 (*ifcompare_move_shift): Likewise.
11624 (*if_move_shift): Likewise.
11625 (*ifcompare_shift_shift): Likewise.
11626 (*ifcompare_not_arith): Likewise.
11627 (*ifcompare_arith_not): Likewise.
11628 (*if_arith_not): Likewise.
11629 (*ifcompare_neg_move): Likewise.
11630 (*if_neg_move): Likewise.
11631 (*ifcompare_move_neg): Likewise.
11632 (*if_move_neg): Likewise.
11633 (prologue_thumb1_interwork): Likewise.
11634 (*cond_move_not): Likewise.
11635 (*sign_extract_onebit): Likewise.
11636 (*not_signextract_onebit): Likewise.
11637 (stack_tie): Likewise.
11638 (align_4): Likewise.
11639 (align_8): Likewise.
11640 (consttable_end): Likewise.
11641 (consttable_1): Likewise.
11642 (consttable_2): Likewise.
11643 (consttable_4): Likewise.
11644 (consttable_8): Likewise.
11645 (consttable_16): Likewise.
11646 (*thumb1_tablejump): Likewise.
11647 (prefetch): Likewise.
11648 (force_register_use): Likewise.
11649 (thumb_eh_return): Likewise.
11650 (load_tp_hard): Likewise.
11651 (load_tp_soft): Likewise.
11652 (tlscall): Likewise.
11653 (*arm_movtas_ze): Likewise.
11654 (*arm_rev): Likewise.
11655 (*arm_revsh): Likewise.
11656 (*arm_rev16): Likewise.
11657 * config/arm/thumb2.md
11658 (*thumb2_smaxsi3): Likewise.
11659 (*thumb2_sminsi3): Likewise.
11660 (*thumb32_umaxsi3): Likewise.
11661 (*thumb2_uminsi3): Likewise.
11662 (*thumb2_negdi2): Likewise.
11663 (*thumb2_abssi2): Likewise.
11664 (*thumb2_neg_abss): Likewise.
11665 (*thumb2_movsi_insn): Likewise.
11666 (tls_load_dot_plus_four): Likewise.
11667 (*thumb2_movhi_insn): Likewise.
11668 (*thumb2_mov_scc): Likewise.
11669 (*thumb2_mov_negs): Likewise.
11670 (*thumb2_mov_negs): Likewise.
11671 (*thumb2_mov_nots): Likewise.
11672 (*thumb2_mov_nots): Likewise.
11673 (*thumb2_movsicc_): Likewise.
11674 (*thumb2_movsfcc_soft_insn): Likewise.
11675 (*thumb2_indirect_jump): Likewise.
11676 (*thumb2_and_scc): Likewise.
11677 (*thumb2_ior_scc): Likewise.
11678 (*thumb2_ior_scc_strict_it): Likewise.
11679 (*thumb2_cond_move): Likewise.
11680 (*thumb2_cond_arith): Likewise.
11681 (*thumb2_cond_ari): Likewise.
11682 (*thumb2_cond_sub): Likewise.
11683 (*thumb2_negscc): Likewise.
11684 (*thumb2_movcond): Likewise.
11685 (thumb2_casesi_internal): Likewise.
11686 (thumb2_casesi_internal_pic): Likewise.
11687 (*thumb2_alusi3_short): Likewise.
11688 (*thumb2_mov<mode>_shortim): Likewise.
11689 (*thumb2_addsi_short): Likewise.
11690 (*thumb2_subsi_short): Likewise.
11691 (thumb2_addsi3_compare0): Likewise.
11692 (*thumb2_cbz): Likewise.
11693 (*thumb2_cbnz): Likewise.
11694 (*thumb2_one_cmplsi2_short): Likewise.
11695 (*thumb2_negsi2_short): Likewise.
11696 (*orsi_notsi_si): Likewise.
11697 * config/arm/arm1020e.md: Update with new attributes.
11698 * config/arm/arm1026ejs.md: Update with new attributes.
11699 * config/arm/arm1136jfs.md: Update with new attributes.
11700 * config/arm/arm926ejs.md: Update with new attributes.
11701 * config/arm/cortex-a15.md: Update with new attributes.
11702 * config/arm/cortex-a5.md: Update with new attributes.
11703 * config/arm/cortex-a53.md: Update with new attributes.
11704 * config/arm/cortex-a7.md: Update with new attributes.
11705 * config/arm/cortex-a8.md: Update with new attributes.
11706 * config/arm/cortex-a9.md: Update with new attributes.
11707 * config/arm/cortex-m4.md: Update with new attributes.
11708 * config/arm/cortex-r4.md: Update with new attributes.
11709 * config/arm/fa526.md: Update with new attributes.
11710 * config/arm/fa606te.md: Update with new attributes.
11711 * config/arm/fa626te.md: Update with new attributes.
11712 * config/arm/fa726te.md: Update with new attributes.
11714 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
11716 * config/aarch64/aarch64-simd.md
11717 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
11718 <vwx> iterator to ensure correct register choice.
11719 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
11720 (aarch64_sqdmull_n<mode>): Likewise.
11721 (aarch64_sqdmull2_n<mode>_internal): Likewise.
11722 * config/aarch64/arm_neon.h
11723 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
11724 (vml<as><q>_n_<su>16): Likewise.
11725 (vml<as>l_high_lane<q>_<su>16): Likewise.
11726 (vml<as>l_high_n_<su>16): Likewise.
11727 (vml<as>l_lane<q>_<su>16): Likewise.
11728 (vml<as>l_n_<su>16): Likewise.
11729 (vmul<q>_lane<q>_<su>16): Likewise.
11730 (vmul<q>_n_<su>16): Likewise.
11731 (vmull_lane<q>_<su>16): Likewise.
11732 (vmull_n_<su>16): Likewise.
11733 (vmull_high_lane<q>_<su>16): Likewise.
11734 (vmull_high_n_<su>16): Likewise.
11735 (vqrdmulh<q>_n_s16): Likewise.
11737 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
11739 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
11740 have the correct lane parameter.
11742 2013-09-06 Richard Biener <rguenther@suse.de>
11744 * cfganal.c (control_dependences::~control_dependences):
11745 Properly free all of the vector.
11747 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
11750 * config/i386/i386.c (ix86_conditional_register_usage):
11751 Proper initialize extended SSE registers.
11753 2013-09-06 Jan Hubicka <jh@suse.cz>
11755 PR tree-optimization/58311
11756 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
11758 2013-09-06 Jan Hubicka <jh@suse.cz>
11760 * Makefile.in (tree-sra.o): Update dependencies.
11761 * tree-sra.c: Include ipa-utils.h
11762 (scan_function): Use recursive_call_p.
11763 (has_caller_p): New function.
11764 (cgraph_for_node_and_aliases): Count also callers of aliases.
11766 2013-09-06 Jan Hubicka <jh@suse.cz>
11768 PR middle-end/58094
11769 * cgraph.h (symtab_semantically_equivalent_p): Declare.
11770 * tree-tailcall.c: Include ipa-utils.h.
11771 (find_tail_calls): Use it.
11772 * ipa-pure-const.c (check_call): Likewise.
11773 * ipa-utils.c (recursive_call_p): New function.
11774 * ipa-utils.h (recursive_call_p): Dclare.
11775 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
11776 (symtab_semantically_equivalent_p): New function.
11777 * Makefile.in (tree-tailcall.o): Update dependencies.
11779 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
11781 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
11782 non-inlinable part.
11784 2013-09-06 Richard Biener <rguenther@suse.de>
11786 * lto-streamer.h (lto_global_var_decls): Remove.
11787 * Makefile.in (OBJS): Remove lto-symtab.o.
11788 (lto-symtab.o): Remove.
11789 (GTFILES): Remove lto-symtab.c
11790 * lto-symtab.c: Move to lto/
11792 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11794 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
11795 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
11796 (UNSPEC_FPINT_RINT): New constant definitions.
11797 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
11798 definition with 2 attributes.
11799 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
11800 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
11803 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11805 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
11806 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
11807 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
11808 attribute to "z196_alone".
11809 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
11812 2013-09-06 Richard Biener <rguenther@suse.de>
11814 * basic-block.h (class control_dependences): New.
11815 * tree-ssa-dce.c (control_dependence_map): Remove.
11817 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
11818 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
11819 find_pdom, find_control_dependence, find_all_control_dependences):
11821 (mark_control_dependent_edges_necessary,
11822 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
11823 tree_dce_done, perform_tree_ssa_dce): Adjust.
11824 * cfganal.c (set_control_dependence_map_bit,
11825 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
11826 find_all_control_dependences): Move from tree-ssa-dce.c and
11827 implement as methods of control_dependences class.
11828 (control_dependences::control_dependences): New.
11829 (control_dependences::~control_dependences): Likewise.
11830 (control_dependences::get_edges_dependent_on): Likewise.
11831 (control_dependences::get_edge): Likewise.
11833 2013-09-04 Jan Hubicka <jh@suse.cz>
11835 * tree.c (types_same_for_odr): Drop overactive check.
11836 * ipa-devirt.c (hash_type_name): Likewise.
11838 2013-09-04 Jan Hubicka <jh@suse.cz>
11840 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
11841 (analyze_functions): ... here.
11843 2013-09-04 Jan Hubicka <jh@suse.cz>
11845 PR middle-end/58201
11846 * cgraphunit.c (analyze_functions): Clear AUX fields
11847 after processing; initialize assembler name has.
11849 2013-09-05 Jeff Law <law@redhat.com>
11851 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
11852 from thread_around_empty_block. Record threading path into PATH.
11853 Recurse if threading through the initial block is successful.
11854 (thread_across_edge): Corresponding changes to slightly simplify.
11856 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
11858 * config/aarch64/aarch64.md
11859 (type): Remove frecpe, frecps, frecpx.
11860 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
11861 fix to be a TARGET_SIMD instruction.
11862 (aarch64_frecps): Remove.
11863 * config/aarch64/aarch64-simd.md
11864 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
11865 (aarch64_frecps<mode>): Handle all float/vector of float modes.
11867 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
11868 Sofiane Naci <sofiane.naci@arm.com>
11870 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
11871 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
11872 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
11873 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
11874 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
11875 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
11876 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
11877 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
11878 into "clz, "rbit". Rename "shift" to "shift_imm".
11879 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
11880 changes. Update for attribute changes all occurrences of arlo_* and
11882 * config/arm/arm-fixed.md: Update for attribute changes
11883 all occurrences of arlo_* types.
11884 * config/arm/thumb2.md: Update for attribute changes all occurrences
11886 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
11887 (cortexa7_older_only): Likewise.
11888 (cortexa7_younger): Likewise.
11889 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
11890 (1020alu_shift_op): Likewise.
11891 (1020alu_shift_reg_op): Likewise.
11892 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
11893 (alu_shift_op): Likewise.
11894 (alu_shift_reg_op): Likewise.
11895 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
11896 (11_alu_shift_op): Likewise.
11897 (11_alu_shift_reg_op): Likewise.
11898 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
11899 (9_alu_shift_reg_op): Likewise.
11900 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
11902 (cortex_a15_alu_shift): Likewise.
11903 (cortex_a15_alu_shift_reg): Likewise.
11904 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
11906 (cortex_a5_alu_shift): Likewise.
11907 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
11909 (cortex_a53_alu_shift): Likewise.
11910 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
11912 (cortex_a7_alu_reg): Likewise.
11913 (cortex_a7_alu_shift): Likewise.
11914 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
11916 (cortex_a8_alu_shift): Likewise.
11917 (cortex_a8_alu_shift_reg): Likewise.
11918 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
11919 (cortex_a9_dp_shift): Likewise.
11920 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
11922 * config/arm/cortex-r4.md
11923 (cortex_r4_alu): Update for attribute changes.
11924 (cortex_r4_mov): Likewise.
11925 (cortex_r4_alu_shift_reg): Likewise.
11926 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
11927 (526_alu_shift_op): Likewise.
11928 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
11929 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
11930 (626te_alu_shift_op): Likewise.
11931 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
11932 (726te_alu_shift_op): Likewise.
11933 (726te_alu_shift_reg_op): Likewise.
11934 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
11935 (mp626_alu_shift_op): Likewise.
11936 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
11937 (pj4_alu_conds): Likewise.
11938 (pj4_shift): Likewise.
11939 (pj4_shift_conds): Likewise.
11940 (pj4_alu_shift): Likewise.
11941 (pj4_alu_shift_conds): Likewise.
11942 * config/aarch64/aarch64.md: Update for attribute change
11943 all occurrences of arlo_* and shift* types.
11945 2013-09-05 Mike Stump <mikestump@comcast.net>
11947 * tree.h: Move documentation for tree_function_decl to tree-core.h
11948 with the declaration.
11950 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
11953 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
11954 looking for widest mode.
11956 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
11958 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
11960 2013-09-05 Richard Biener <rguenther@suse.de>
11962 PR tree-optimization/58137
11963 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
11964 Do not create vectors of pointers.
11965 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
11966 types for the components of the vector initializer.
11967 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
11968 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
11970 2013-09-05 Martin Jambor <mjambor@suse.cz>
11972 * ipa-prop.c (remove_described_reference): Accept missing references,
11973 return false if that hppens, otherwise return true.
11974 (cgraph_node_for_jfunc): New function.
11975 (try_decrement_rdesc_refcount): Likewise.
11976 (try_make_edge_direct_simple_call): Use them.
11977 (ipa_edge_removal_hook): Remove references from rdescs.
11978 (ipa_edge_duplication_hook): Clone rdescs and their references
11979 when the new edge has the same caller as the old one.
11980 * cgraph.c (cgraph_resolve_speculation): Remove speculative
11981 reference before removing any edges.
11983 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
11985 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
11987 * thumb2.md (thumb2_storewb_parisi): New pattern.
11989 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
11991 * config/aarch64/aarch64-option-extensions.def: Add
11992 AARCH64_OPT_EXTENSION of 'crc'.
11993 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
11994 (AARCH64_ISA_CRC): Ditto.
11995 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
11996 description of the CRC extension.
11998 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
12000 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
12001 * config/rs6000/linux.h: Ditto.
12002 * alpha/linux.h: Ditto.
12003 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
12004 no_c99_libc_has_function.
12005 * config/c6x/uclinux-elf.h: Ditto.
12006 * config/lm32/uclinux-elf.h: Ditto.
12007 * config/m68k/uclinux.h: Ditto.
12008 * config/moxie/uclinux.h: Ditto.
12009 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
12010 (crisv32-*-linux*, cris-*-linux*): Ditto.
12011 * config/bfin/bfin.c: Include "tm_p.h".
12013 2013-09-05 Richard Biener <rguenther@suse.de>
12015 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
12016 check for a definition without a basic-block.
12018 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
12019 Sofiane Naci <sofiane.naci@arm.com>
12021 * config/aarch64/aarch64.md
12022 (*movti_aarch64): Rename r_2_f and f_2_r.
12023 (*movsf_aarch64): Likewise.
12024 (*movdf_aarch64): Likewise.
12025 (*movtf_aarch64): Likewise.
12026 (aarch64_movdi_<mode>low): Likewise.
12027 (aarch64_movdi_<mode>high): Likewise.
12028 (aarch64_mov<mode>high_di): Likewise.
12029 (aarch64_mov<mode>low_di): Likewise.
12030 (aarch64_movtilow_tilow): Likewise.
12031 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
12032 values to config/arm/types.md
12033 (attribute "conds"): Update for attribute change.
12034 (anddi3_insn): Likewise.
12035 (iordi3_insn): Likewise.
12036 (xordi3_insn): Likewise.
12037 (one_cmpldi2): Likewise.
12038 * config/arm/types.md (type): Add Neon types.
12039 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
12040 use "type" attribute.
12041 (movmisalign<mode>_neon_store): Likewise.
12042 (movmisalign<mode>_neon_load): Likewise.
12043 (vec_set<mode>_internal): Likewise.
12044 (vec_setv2di_internal): Likewise.
12045 (vec_extract<mode>): Likewise.
12046 (vec_extractv2di): Likewise.
12047 (add<mode>3_neon): Likewise.
12048 (adddi3_neon): Likewise.
12049 (sub<mode>3_neon): Likewise.
12050 (subdi3_neon): Likewise.
12051 (mul<mode>3_neon): Likewise.
12052 (mul<mode>3add<mode>_neon): Likewise.
12053 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
12054 (fma<VCVTF:mode>4)): Likewise.
12055 (fma<VCVTF:mode>4_intrinsic): Likewise.
12056 (fmsub<VCVTF:mode>4)): Likewise.
12057 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
12058 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
12059 (ior<mode>3): Likewise.
12060 (and<mode>3): Likewise.
12061 (anddi3_neon): Likewise.
12062 (orn<mode>3_neon): Likewise.
12063 (orndi3_neon): Likewise.
12064 (bic<mode>3_neon): Likewise.
12065 (bicdi3_neon): Likewise.
12066 (xor<mode>3): Likewise.
12067 (one_cmpl<mode>2): Likewise.
12068 (abs<mode>2): Likewise.
12069 (neg<mode>2): Likewise.
12070 (umin<mode>3_neon): Likewise.
12071 (umax<mode>3_neon): Likewise.
12072 (smin<mode>3_neon): Likewise.
12073 (smax<mode>3_neon): Likewise.
12074 (vashl<mode>3): Likewise.
12075 (vashr<mode>3_imm): Likewise.
12076 (vlshr<mode>3_imm): Likewise.
12077 (ashl<mode>3_signed): Likewise.
12078 (ashl<mode>3_unsigned): Likewise.
12079 (neon_load_count): Likewise.
12080 (ashldi3_neon_noclobber): Likewise.
12081 (signed_shift_di3_neon): Likewise.
12082 (unsigned_shift_di3_neon): Likewise.
12083 (ashrdi3_neon_imm_noclobber): Likewise.
12084 (lshrdi3_neon_imm_noclobber): Likewise.
12085 (widen_ssum<mode>3): Likewise.
12086 (widen_usum<mode>3): Likewise.
12087 (quad_halves_<code>v4si): Likewise.
12088 (quad_halves_<code>v4sf): Likewise.
12089 (quad_halves_<code>v8hi): Likewise.
12090 (quad_halves_<code>v16qi): Likewise.
12091 (reduc_splus_v2di): Likewise.
12092 (neon_vpadd_internal<mode>): Likewise.
12093 (neon_vpsmin<mode>): Likewise.
12094 (neon_vpsmax<mode>): Likewise.
12095 (neon_vpumin<mode>): Likewise.
12096 (neon_vpumax<mode>): Likewise.
12097 (ss_add<mode>_neon): Likewise.
12098 (us_add<mode>_neon): Likewise.
12099 (ss_sub<mode>_neon): Likewise.
12100 (us_sub<mode>_neon): Likewise.
12101 (neon_vadd<mode>_unspec): Likewise.
12102 (neon_vaddl<mode>): Likewise.
12103 (neon_vaddw<mode>): Likewise.
12104 (neon_vhadd<mode>): Likewise.
12105 (neon_vqadd<mode>): Likewise.
12106 (neon_vaddhn<mode>): Likewise.
12107 (neon_vmul<mode>): Likewise.
12108 (neon_vmla<mode>): Likewise.
12109 (neon_vmlal<mode>): Likewise.
12110 (neon_vmls<mode>): Likewise.
12111 (neon_vmlsl<mode>): Likewise.
12112 (neon_vqdmulh<mode>): Likewise.
12113 (neon_vqdmlal<mode>): Likewise.
12114 (neon_vqdmlsl<mode>): Likewise.
12115 (neon_vmull<mode>): Likewise.
12116 (neon_vqdmull<mode>): Likewise.
12117 (neon_vsub<mode>_unspec): Likewise.
12118 (neon_vsubl<mode>): Likewise.
12119 (neon_vsubw<mode>): Likewise.
12120 (neon_vqsub<mode>): Likewise.
12121 (neon_vhsub<mode>): Likewise.
12122 (neon_vsubhn<mode>): Likewise.
12123 (neon_vceq<mode>): Likewise.
12124 (neon_vcge<mode>): Likewise.
12125 (neon_vcgeu<mode>): Likewise.
12126 (neon_vcgt<mode>): Likewise.
12127 (neon_vcgtu<mode>): Likewise.
12128 (neon_vcle<mode>): Likewise.
12129 (neon_vclt<mode>): Likewise.
12130 (neon_vcage<mode>): Likewise.
12131 (neon_vcagt<mode>): Likewise.
12132 (neon_vtst<mode>): Likewise.
12133 (neon_vabd<mode>): Likewise.
12134 (neon_vabdl<mode>): Likewise.
12135 (neon_vaba<mode>): Likewise.
12136 (neon_vabal<mode>): Likewise.
12137 (neon_vmax<mode>): Likewise.
12138 (neon_vmin<mode>): Likewise.
12139 (neon_vpaddl<mode>): Likewise.
12140 (neon_vpadal<mode>): Likewise.
12141 (neon_vpmax<mode>): Likewise.
12142 (neon_vpmin<mode>): Likewise.
12143 (neon_vrecps<mode>): Likewise.
12144 (neon_vrsqrts<mode>): Likewise.
12145 (neon_vqabs<mode>): Likewise.
12146 (neon_vqneg<mode>): Likewise.
12147 (neon_vcls<mode>): Likewise.
12148 (clz<mode>2): Likewise.
12149 (popcount<mode>2): Likewise.
12150 (neon_vrecpe): Likewise.
12151 (neon_vrsqrte): Likewise.
12152 (neon_vget_lane<mode>_sext_internal): Likewise.
12153 (neon_vget_lane<mode>_zext_internal): Likewise.
12154 (neon_vdup_n<mode>): Likewise.
12155 (neon_vdup_nv2di): Likewise.
12156 (neon_vdpu_lane<mode>_internal): Likewise.
12157 (neon_vswp<mode>): Likewise.
12158 (float<mode><V_cvtto>2): Likewise.
12159 (floatuns<mode><V_cvtto>2): Likewise.
12160 (fix_trunc<mode><V_cvtto>)2): Likewise
12161 (fixuns_trunc<mode><V_cvtto)2): Likewise.
12162 (neon_vcvt<mode>): Likewise.
12163 (neon_vcvtv4sfv4hf): Likewise.
12164 (neon_vcvtv4hfv4sf): Likewise.
12165 (neon_vcvt_n<mode>): Likewise.
12166 (neon_vmovn<mode>): Likewise.
12167 (neon_vqmovn<mode>): Likewise.
12168 (neon_vqmovun<mode>): Likewise.
12169 (neon_vmovl<mode>): Likewise.
12170 (neon_vmul_lane<mode>): Likewise.
12171 (neon_vmull_lane<mode>): Likewise.
12172 (neon_vqdmull_lane<mode>): Likewise.
12173 (neon_vqdmulh_lane<mode>): Likewise.
12174 (neon_vmla_lane<mode>): Likewise.
12175 (neon_vmlal_lane<mode>): Likewise.
12176 (neon_vqdmlal_lane<mode>): Likewise.
12177 (neon_vmls_lane<mode>): Likewise.
12178 (neon_vmlsl_lane<mode>): Likewise.
12179 (neon_vqdmlsl_lane<mode>): Likewise.
12180 (neon_vext<mode>): Likewise.
12181 (neon_vrev64<mode>): Likewise.
12182 (neon_vrev32<mode>): Likewise.
12183 (neon_vrev16<mode>): Likewise.
12184 (neon_vbsl<mode>_internal): Likewise.
12185 (neon_vshl<mode>): Likewise.
12186 (neon_vqshl<mode>): Likewise.
12187 (neon_vshr_n<mode>): Likewise.
12188 (neon_vshrn_n<mode>): Likewise.
12189 (neon_vqshrn_n<mode>): Likewise.
12190 (neon_vqshrun_n<mode>): Likewise.
12191 (neon_vshl_n<mode>): Likewise.
12192 (neon_vqshl_n<mode>): Likewise.
12193 (neon_vqshlu_n<mode>): Likewise.
12194 (neon_vshll_n<mode>): Likewise.
12195 (neon_vsra_n<mode>): Likewise.
12196 (neon_vsri_n<mode>): Likewise.
12197 (neon_vsli_n<mode>): Likewise.
12198 (neon_vtbl1v8qi): Likewise.
12199 (neon_vtbl2v8qi): Likewise.
12200 (neon_vtbl3v8qi): Likewise.
12201 (neon_vtbl4v8qi): Likewise.
12202 (neon_vtbx1v8qi): Likewise.
12203 (neon_vtbx2v8qi): Likewise.
12204 (neon_vtbx3v8qi): Likewise.
12205 (neon_vtbx4v8qi): Likewise.
12206 (neon_vtrn<mode>_internal): Likewise.
12207 (neon_vzip<mode>_internal): Likewise.
12208 (neon_vuzp<mode>_internal): Likewise.
12209 (neon_vld1<mode>): Likewise.
12210 (neon_vld1_lane<mode>): Likewise.
12211 (neon_vld1_dup<mode>): Likewise.
12212 (neon_vld1_dupv2di): Likewise.
12213 (neon_vst1<mode>): Likewise.
12214 (neon_vst1_lane<mode>): Likewise.
12215 (neon_vld2<mode>): Likewise.
12216 (neon_vld2_lane<mode>): Likewise.
12217 (neon_vld2_dup<mode>): Likewise.
12218 (neon_vst2<mode>): Likewise.
12219 (neon_vst2_lane<mode>): Likewise.
12220 (neon_vld3<mode>): Likewise.
12221 (neon_vld3qa<mode>): Likewise.
12222 (neon_vld3qb<mode>): Likewise.
12223 (neon_vld3_lane<mode>): Likewise.
12224 (neon_vld3_dup<mode>): Likewise.
12225 (neon_vst3<mode>): Likewise.
12226 (neon_vst3qa<mode>): Likewise.
12227 (neon_vst3qb<mode>): Likewise.
12228 (neon_vst3_lane<mode>): Likewise.
12229 (neon_vld4<mode>): Likewise.
12230 (neon_vld4qa<mode>): Likewise.
12231 (neon_vld4qb<mode>): Likewise.
12232 (neon_vld4_lane<mode>): Likewise.
12233 (neon_vld4_dup<mode>): Likewise.
12234 (neon_vst4<mode>): Likewise.
12235 (neon_vst4qa<mode>): Likewise.
12236 (neon_vst4qb<mode>): Likewise.
12237 (neon_vst4_lane<mode>): Likewise.
12238 (neon_vec_unpack<US>_lo_<mode>): Likewise.
12239 (neon_vec_unpack<US>_hi_<mode>): Likewise.
12240 (neon_vec_<US>mult_lo_<mode>): Likewise.
12241 (neon_vec_<US>mult_hi_<mode>): Likewise.
12242 (neon_vec_<US>shiftl_<mode>): Likewise.
12243 (neon_unpack<US>_<mode>): Likewise.
12244 (neon_vec_<US>mult_<mode>): Likewise.
12245 (vec_pack_trunc_<mode>): Likewise.
12246 (neon_vec_pack_trunk_<mode>): Likewise.
12247 (neon_vabd<mode>_2): Likewise.
12248 (neon_vabd<mode>_3): Likewise.
12249 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
12250 (thumb2_movsi_vfp): Likewise.
12251 (movdi_vfp): Likewise.
12252 (movdi_vfp_cortexa8): Likewise.
12253 (movhf_vfp_neon): Likewise.
12254 (movhf_vfp): Likewiwse.
12255 (movsf_vfp): Likewiwse.
12256 (thumb2_movsf_vfp): Likewiwse.
12257 (movdf_vfp): Likewise.
12258 (thumb2_movdf_vfp): Likewise.
12259 (movsfcc_vfp): Likewise.
12260 (thumb2_movsfcc_vfp): Likewise.
12261 (movdfcc_vfp): Likewise.
12262 (thumb2_movdfcc_vfp): Likewise.
12263 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
12264 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
12265 (v10_v2c): Likewise.
12266 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
12268 (cortex_a15_neon_int_2): Likewise.
12269 (cortex_a15_neon_int_3): Likewise.
12270 (cortex_a15_neon_int_4): Likewise.
12271 (cortex_a15_neon_int_5): Likewise.
12272 (cortex_a15_neon_vqneg_vqabs): Likewise.
12273 (cortex_a15_neon_vmov): Likewise.
12274 (cortex_a15_neon_vaba): Likewise.
12275 (cortex_a15_neon_vaba_qqq): Likewise.
12276 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
12277 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
12278 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
12279 scalar_64_32_long_scalar): Likewise.
12280 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
12281 (cortex_a15_neon_mla_qqq_8_16): Likewise.
12282 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
12283 lotype_qdd_64_32_long): Likewise.
12284 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
12285 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
12286 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
12287 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
12288 (cortex_a15_neon_shift_1): Likewise.
12289 (cortex_a15_neon_shift_2): Likewise.
12290 (cortex_a15_neon_shift_3): Likewise.
12291 (cortex_a15_neon_vshl_ddd): Likewise.
12292 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
12293 (cortex_a15_neon_vsra_vrsra): Likewise.
12294 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
12295 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
12296 (cortex_a15_neon_fp_vmul_ddd): Likewise.
12297 (cortex_a15_neon_fp_vmul_qqd): Likewise.
12298 (cortex_a15_neon_fp_vmla_ddd): Likewise.
12299 (cortex_a15_neon_fp_vmla_qqq): Likewise.
12300 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
12301 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
12302 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
12303 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
12304 (cortex_a15_neon_bp_simple): Likewise.
12305 (cortex_a15_neon_bp_2cycle): Likewise.
12306 (cortex_a15_neon_bp_3cycle): Likewise.
12307 (cortex_a15_neon_vld1_1_2_regs): Likewise.
12308 (cortex_a15_neon_vld1_3_4_regs): Likewise.
12309 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
12310 (cortex_a15_neon_vld2_4_regs): Likewise.
12311 (cortex_a15_neon_vld3_vld4): Likewise.
12312 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
12313 (cortex_a15_neon_vst1_3_4_regs): Likewise.
12314 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
12315 (cortex_a15_neon_vst3_vst4): Likewise.
12316 (cortex_a15_neon_vld1_vld2_lane): Likewise.
12317 (cortex_a15_neon_vld3_vld4_lane" 10
12318 (cortex_a15_neon_vst1_vst2_lane): Likewise.
12319 (cortex_a15_neon_vst3_vst4_lane): Likewise.
12320 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
12321 (cortex_a15_neon_ldm_2): Likewise.0
12322 (cortex_a15_neon_stm_2): Likewise.
12323 (cortex_a15_neon_mcr): Likewise.
12324 (cortex_a15_neon_mcr_2_mcrr): Likewise.
12325 (cortex_a15_neon_mrc): Likewise.
12326 (cortex_a15_neon_mrrc): Likewise.
12327 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
12329 (cortex_a15_alu_shift): Likewise.
12330 (cortex_a15_alu_shift_reg): Likewise.
12331 (cortex_a15_mult32): Likewise.
12332 (cortex_a15_mult64): Likewise.
12333 (cortex_a15_block): Likewise.
12334 (cortex_a15_branch): Likewise.
12335 (cortex_a15_load1): Likewise.
12336 (cortex_a15_load3): Likewise.
12337 (cortex_a15_store1): Likewise.
12338 (cortex_a15_store3): Likewise.
12339 (cortex_a15_call): Likewise.
12340 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
12341 (cortex_a5_f2r): Likewise.
12342 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
12344 (cortex_a53_f2r): Likewise.
12345 * config/arm/cortex-a7.md
12346 (cortex_a7_branch): Update for attribute change.
12347 (cortex_a7_call): Likewise.
12348 (cortex_a7_alu_imm): Likewise.
12349 (cortex_a7_alu_reg): Likewise.
12350 (cortex_a7_alu_shift): Likewise.
12351 (cortex_a7_mul): Likewise.
12352 (cortex_a7_load1): Likewise.
12353 (cortex_a7_store1): Likewise.
12354 (cortex_a7_load2): Likewise.
12355 (cortex_a7_store2): Likewise.
12356 (cortex_a7_load3): Likewise.
12357 (cortex_a7_store3): Likewise.
12358 (cortex_a7_load4): Likewise.
12359 (cortex_a7_store4): Likewise.
12360 (cortex_a7_fpalu): Likewise.
12361 (cortex_a7_fconst): Likewise.
12362 (cortex_a7_fpmuls): Likewise.
12363 (cortex_a7_neon_mul): Likewise.
12364 (cortex_a7_fpmacs): Likewise.
12365 (cortex_a7_neon_mla: Likewise.
12366 (cortex_a7_fpmuld: Likewise.
12367 (cortex_a7_fpmacd: Likewise.
12368 (cortex_a7_fpfmad: Likewise.
12369 (cortex_a7_fdivs: Likewise.
12370 (cortex_a7_fdivd: Likewise.
12371 (cortex_a7_r2f: Likewise.
12372 (cortex_a7_f2r: Likewise.
12373 (cortex_a7_f_flags: Likewise.
12374 (cortex_a7_f_loads: Likewise.
12375 (cortex_a7_f_loadd: Likewise.
12376 (cortex_a7_f_stores: Likewise.
12377 (cortex_a7_f_stored: Likewise.
12378 (cortex_a7_neon): Likewise.
12379 * config/arm/cortex-a8-neon.md
12380 (cortex_a8_neon_mrc): Update for attribute change.
12381 (cortex_a8_neon_mrrc): Likewise.
12382 (cortex_a8_neon_int_1): Likewise.
12383 (cortex_a8_neon_int_2): Likewise.
12384 (cortex_a8_neon_int_3): Likewise.
12385 (cortex_a8_neon_int_4): Likewise.
12386 (cortex_a8_neon_int_5): Likewise.
12387 (cortex_a8_neon_vqneg_vqabs): Likewise.
12388 (cortex_a8_neon_vmov): Likewise.
12389 (cortex_a8_neon_vaba): Likewise.
12390 (cortex_a8_neon_vaba_qqq): Likewise.
12391 (cortex_a8_neon_vsma): Likewise.
12392 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
12393 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
12394 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
12395 long_scalar): Likewise.
12396 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
12397 (cortex_a8_neon_mla_qqq_8_16): Likewise.
12398 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
12399 long_scalar_qdd_64_32_long): Likewise.
12400 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
12401 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
12402 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
12403 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
12404 (cortex_a8_neon_shift_1): Likewise.
12405 (cortex_a8_neon_shift_2): Likewise.
12406 (cortex_a8_neon_shift_3): Likewise.
12407 (cortex_a8_neon_vshl_ddd): Likewise.
12408 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
12409 (cortex_a8_neon_vsra_vrsra): Likewise.
12410 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
12411 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
12412 (cortex_a8_neon_fp_vsum): Likewise.
12413 (cortex_a8_neon_fp_vmul_ddd): Likewise.
12414 (cortex_a8_neon_fp_vmul_qqd): Likewise.
12415 (cortex_a8_neon_fp_vmla_ddd): Likewise.
12416 (cortex_a8_neon_fp_vmla_qqq): Likewise.
12417 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
12418 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
12419 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
12420 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
12421 (cortex_a8_neon_bp_simple): Likewise.
12422 (cortex_a8_neon_bp_2cycle): Likewise.
12423 (cortex_a8_neon_bp_3cycle): Likewise.
12424 (cortex_a8_neon_ldr): Likewise.
12425 (cortex_a8_neon_str): Likewise.
12426 (cortex_a8_neon_vld1_1_2_regs): Likewise.
12427 (cortex_a8_neon_vld1_3_4_regs): Likewise.
12428 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
12429 (cortex_a8_neon_vld2_4_regs): Likewise.
12430 (cortex_a8_neon_vld3_vld4): Likewise.
12431 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
12432 (cortex_a8_neon_vst1_3_4_regs): Likewise.
12433 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
12434 (cortex_a8_neon_vst3_vst4): Likewise.
12435 (cortex_a8_neon_vld1_vld2_lane): Likewise.
12436 (cortex_a8_neon_vld3_vld4_lane): Likewise.
12437 (cortex_a8_neon_vst1_vst2_lane): Likewise.
12438 (cortex_a8_neon_vst3_vst4_lane): Likewise.
12439 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
12440 (cortex_a8_neon_mcr): Likewise.
12441 (cortex_a8_neon_mcr_2_mcrr): Likewise.
12442 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
12443 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
12445 (ca9_neon_mrrc): Likewise.
12446 (cortex_a9_neon_int_1): Likewise.
12447 (cortex_a9_neon_int_2): Likewise.
12448 (cortex_a9_neon_int_3): Likewise.
12449 (cortex_a9_neon_int_4): Likewise.
12450 (cortex_a9_neon_int_5): Likewise.
12451 (cortex_a9_neon_vqneg_vqabs): Likewise.
12452 (cortex_a9_neon_vmov): Likewise.
12453 (cortex_a9_neon_vaba): Likewise.
12454 (cortex_a9_neon_vaba_qqq): Likewise.
12455 (cortex_a9_neon_vsma): Likewise.
12456 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
12457 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
12458 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
12459 long_scalar): Likewise.
12460 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
12461 (cortex_a9_neon_mla_qqq_8_16): Likewise.
12462 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
12463 long_scalar_qdd_64_32_long): Likewise.
12464 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
12465 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
12466 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
12467 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
12468 (cortex_a9_neon_shift_1): Likewise.
12469 (cortex_a9_neon_shift_2): Likewise.
12470 (cortex_a9_neon_shift_3): Likewise.
12471 (cortex_a9_neon_vshl_ddd): Likewise.
12472 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
12473 (cortex_a9_neon_vsra_vrsra): Likewise.
12474 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
12475 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
12476 (cortex_a9_neon_fp_vsum): Likewise.
12477 (cortex_a9_neon_fp_vmul_ddd): Likewise.
12478 (cortex_a9_neon_fp_vmul_qqd): Likewise.
12479 (cortex_a9_neon_fp_vmla_ddd): Likewise.
12480 (cortex_a9_neon_fp_vmla_qqq): Likewise.
12481 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
12482 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
12483 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
12484 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
12485 (cortex_a9_neon_bp_simple): Likewise.
12486 (cortex_a9_neon_bp_2cycle): Likewise.
12487 (cortex_a9_neon_bp_3cycle): Likewise.
12488 (cortex_a9_neon_ldr): Likewise.
12489 (cortex_a9_neon_str): Likewise.
12490 (cortex_a9_neon_vld1_1_2_regs): Likewise.
12491 (cortex_a9_neon_vld1_3_4_regs): Likewise.
12492 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
12493 (cortex_a9_neon_vld2_4_regs): Likewise.
12494 (cortex_a9_neon_vld3_vld4): Likewise.
12495 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
12496 (cortex_a9_neon_vst1_3_4_regs): Likewise.
12497 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
12498 (cortex_a9_neon_vst3_vst4): Likewise.
12499 (cortex_a9_neon_vld1_vld2_lane): Likewise.
12500 (cortex_a9_neon_vld3_vld4_lane): Likewise.
12501 (cortex_a9_neon_vst1_vst2_lane): Likewise.
12502 (cortex_a9_neon_vst3_vst4_lane): Likewise.
12503 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
12504 (cortex_a9_neon_mcr): Likewise.
12505 (cortex_a9_neon_mcr_2_mcrr): Likewise.
12506 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
12507 (cortex_a9_fps): Likewise.
12508 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
12510 (cortex_m4_fmuls): Likewise.
12511 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
12513 (cortex_r4_mrc): Likewise.
12514 * config/arm/iterators.md: Update comment referring to neon_type.
12515 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
12516 (iwmmxt_movsi_insn): Likewise.
12517 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
12519 (pj4_core_to_vfp): Likewise.
12520 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
12522 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
12523 (vfp_fstore): Likewise.
12524 * doc/md.texi: Change references to neon_type to refer to type.
12526 2013-09-04 Dodji Seketeli <dodji@redhat.com>
12528 * tree.h (DECL_BUILT_IN): Fix typo in comment.
12530 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
12532 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
12533 lglobl if not weak.
12535 2013-09-04 Easwaran Raman <eraman@google.com>
12537 PR middle-end/57370
12538 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
12539 (build_and_add_sum): Use it.
12540 (appears_later_in_bb): Simplify code.
12542 2013-09-04 Teresa Johnson <tejohnson@google.com>
12544 * dumpfile.c (dump_finish): Don't close stderr/stdout.
12546 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
12548 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
12550 2013-09-04 Jan Hubicka <jh@suse.cz>
12552 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
12553 * ipa-devirt.c: Include diganostic.h
12554 (odr_type_d): Add types and types_set.
12555 (hash_type_name): Work for types with vtables during LTO.
12556 (odr_hasher::remove): Fix comment; destroy types_set.
12557 (add_type_duplicate): New function,
12558 (get_odr_type): Use it.
12559 (dump_type_inheritance_graph): Dump type duplicates.
12560 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
12562 * tree.c (types_same_for_odr): Give exact answers on types with
12565 2013-09-04 Dodji Seketeli <dodji@redhat.com>
12567 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
12568 explaining their differences.
12570 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
12572 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
12574 2013-09-03 Jeff Law <law@redhat.com>
12576 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
12577 when not threading through a joiner block. Pass joiner/no joiner
12578 state to register_jump_thread.
12579 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
12580 state from argument rather than implying on path length.
12581 Dump the entire jump thread path into debugging dump.
12582 * tree-flow.h (register_jump_thread): Update prototype.
12584 2013-08-29 Xinliang David Li <davidxl@google.com>
12586 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
12587 Remove a trivial gcc_assert.
12589 2013-08-29 Xinliang David Li <davidxl@google.com>
12591 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
12592 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
12593 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
12594 Delay base decl alignment adjustment.
12595 * tree-vectorizer.c (vect_destroy_datarefs): New function.
12596 * tree-vectorizer.h: New data structure.
12597 (set_dr_misalignment): New function.
12598 (dr_misalignment): Ditto.
12599 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
12600 (vectorizable_load): Ditto.
12601 (ensure_base_align): New function.
12602 (vectorize_loops): Add dbg_cnt support.
12603 (execute_vect_slp): Ditto.
12604 * dbgcnt.def: New debug counter.
12605 * Makefile: New dependency.
12607 2013-09-03 Meador Inge <meadori@codesourcery.com>
12611 2013-08-30 Meador Inge <meadori@codesourcery.com>
12613 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
12615 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
12617 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
12618 function descriptor.
12620 2013-09-03 Richard Biener <rguenther@suse.de>
12622 * tree-affine.c (add_elt_to_tree): Fix association issue,
12623 avoid useless converts and make sure to always return a
12624 properly typed result.
12626 2013-09-03 Richard Biener <rguenther@suse.de>
12628 PR middle-end/57656
12629 * fold-const.c (negate_expr_p): Fix division case.
12630 (negate_expr): Likewise.
12632 2013-09-03 Richard Biener <rguenther@suse.de>
12635 * tree-streamer-out.c: Include tm.h.
12636 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
12638 2013-09-03 Jan Hubicka <jh@suse.cz>
12640 * tree-profile.c (tree_profiling): Cleanup CFG when done.
12642 2013-09-03 Alan Modra <amodra@gmail.com>
12644 * config.gcc (powerpc*-*-linux*): Add support for little-endian
12645 multilibs to big-endian target and vice versa.
12646 * config/rs6000/t-linux64: Use := assignment on all vars.
12647 (MULTILIB_EXTRA_OPTS): Remove fPIC.
12648 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
12649 * config/rs6000/t-linux64le: New file.
12650 * config/rs6000/t-linux64bele: New file.
12651 * config/rs6000/t-linux64lebe: New file.
12653 2013-09-02 Jan Hubicka <jh@suse.cz>
12655 * ipa-inline-transform.c (inline_transform): Do not
12656 optimize_inline_calls when not optimizing.
12658 2013-09-02 Jan Hubicka <jh@suse.cz>
12660 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
12661 duplicated nodes for assembler names.
12662 * symtab.c (symtab_unregister_node): Do not attempt to unlink
12663 hard registers from assembler name hash.
12665 2013-09-02 Jan Hubicka <jh@suse.cz>
12667 * ipa-split.c (execute_split_functions): Split externally visible
12668 functions called once.
12670 2013-09-02 Martin Jambor <mjambor@suse.cz>
12673 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
12674 linked list. When finding the correct duplicate, also consider also
12675 the caller in additon to its inlined_to node.
12677 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
12679 * config/aarch64/aarch64-simd-builtins.def
12680 (dup_lane_scalar): Remove.
12681 * config/aarch64/aarch64-simd.md
12682 (aarch64_simd_dup): Add 'w->w' alternative.
12683 (aarch64_dup_lane<mode>): Allow for VALL.
12684 (aarch64_dup_lane_scalar<mode>): Remove.
12685 (aarch64_dup_lane_<vswap_width_name><mode>): New.
12686 (aarch64_get_lane_signed<mode>): Add w->w altenative.
12687 (aarch64_get_lane_unsigned<mode>): Likewise.
12688 (aarch64_get_lane<mode>): Likewise.
12689 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
12690 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
12691 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
12692 (VCON): Change container of V2SF.
12693 (vswap_width_name): Likewise.
12694 * config/aarch64/arm_neon.h
12695 (__aarch64_vdup_lane_any): New.
12696 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
12697 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
12698 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
12700 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
12702 PR middle-end/56382
12703 * expr.c (emit_move_complex): Do not move complex FP values as parts if
12704 the source or the destination is a single hard register.
12706 2013-09-02 Richard Biener <rguenther@suse.de>
12708 PR middle-end/57511
12709 * tree-scalar-evolution.c (instantiate_scev_name): Allow
12712 2013-09-02 Richard Biener <rguenther@suse.de>
12714 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
12715 arithmetic to sizetype.
12717 2013-09-02 Bin Cheng <bin.cheng@arm.com>
12719 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
12720 Find auto-increment use both before and after candidate.
12722 2013-09-02 Marek Polacek <polacek@redhat.com>
12724 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
12726 2013-09-01 Jan Hubicka <jh@suse.cz>
12728 * Makefile.in: Add ipa-profile.o
12729 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
12730 * cgraph.c (struct cgraph_propagate_frequency_data,
12731 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
12732 ipa-profile.c; replace cgraph_ by ipa_ prefix.
12733 * cgraph.h (cgraph_propagate_frequency): Remove.
12734 * ipa-inline-analysis.c: Include ipa-utils.h;
12735 drop duplicated cfgloop.h.
12736 (inline_update_callee_summaries): Update.
12737 * ipa-profile.c: New file.
12738 * ipa-utils.h (ipa_propagate_frequency): Declare.
12739 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
12740 data-streamer.h, value-prof.h.
12741 (symtab_remove_unreachable_nodes): Update profile.
12742 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
12743 account_time_size, cmp_counts, dump_histogram,
12744 ipa_profile_generate_summary, ipa_profile_write_summary,
12745 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
12746 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
12747 Move to ipa-profile.c.
12749 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
12751 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
12753 2013-09-01 Jan Hubicka <jh@suse.cz>
12755 * common.opt (fdevirtualize-speculatively): New function.
12756 * invoke.texi (fdevirtualize-speculatively): Document.
12757 * ipa-devirt.c: Include ipa-inline.h
12758 (likely_target_p): New function.
12759 (ipa_devirt): New function.
12760 (gate_ipa_devirt): New function.
12761 (pass_data_ipa_devirt): New static var.
12762 (pass_ipa_devirt): Likewise.
12763 (make_pass_ipa_devirt): New function.
12764 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
12765 (common_handle_option): Disable devirtualization when
12766 value range profiling is available.
12767 * passes.def (pass_ipa_devirt): Add.
12768 * timever.def (TV_IPA_DEVIRT): New timevar.
12769 * tree-pass.h (make_pass_ipa_devirt):
12771 2013-09-01 Iain Sandoe <iain@codesourcery.com>
12773 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
12774 include sanitize(undefined).
12776 2013-08-31 Diego Novillo <dnovillo@google.com>
12778 * Makefile.in (TREE_CORE_H): Define.
12780 (GTFILES): Add tree-core.h.
12781 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
12783 * tree-core.h: New file.
12784 Move all data structures, enum, typedefs, global
12785 declarations and constants from ...
12786 * tree.h: ... here.
12788 2013-08-31 Jan Hubicka <jh@suse.cz>
12790 * bulitins.c (expand_builtin): Do not early exit for gcov
12791 instrumented functions.
12793 2013-08-31 Marek Polacek <polacek@redhat.com>
12795 * ubsan.c: Include tm_p.h.
12797 2013-08-31 Jan Hubicka <jh@suse.cz>
12799 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
12802 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
12803 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
12804 * tree-cfg.c (verify_gimple_label): ... here.
12805 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
12806 (ipa_merge_profiles): New function.
12807 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
12808 (lto_input_function_body): Likewise.
12809 * ipa-utils.h (ipa_merge_profiles): Declare.
12810 * lto-streamer.h (lto_input_function_body): Update prototype.
12811 (emit_label_in_global_context_p): Remove.
12812 * lto-symtab.c: Include ipa-utils.h
12813 (lto_cgraph_replace_node): Use ipa_merge_profiles.
12815 2013-08-31 Jan Hubicka <jh@suse.cz>
12817 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
12819 2013-08-31 Jan Hubicka <jh@suse.cz>
12821 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
12823 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
12825 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
12826 "cmp" RTX before signed_comparison_operator check to account
12827 for "code" changes.
12829 2013-08-30 Jan Hubicka <jh@suse.cz>
12831 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
12832 (detect_type_change_1): Rename to ...
12833 (detect_type_change): ... this one; early return on non-polymorphic
12835 (detect_type_change_ssa): Add comp_type parameter; update
12836 use of detect_type_change.
12837 (compute_complex_assign_jump_func): Add param_type parameter;
12838 update use of detect_type_change_ssa.
12839 (compute_complex_ancestor_jump_func): Likewise.
12840 (ipa_get_callee_param_type): New function.
12841 (ipa_compute_jump_functions_for_edge): Compute parameter type;
12842 update calls to the jump function computation functions.
12844 2013-08-30 Teresa Johnson <tejohnson@google.com>
12845 Steven Bosscher <steven@gcc.gnu.org>
12847 * cfgrtl.c (fixup_new_cold_bb): New routine.
12848 (commit_edge_insertions): Invoke fixup_partitions.
12849 (find_partition_fixes): New routine.
12850 (fixup_partitions): Ditto.
12851 (verify_hot_cold_block_grouping): Update comments.
12852 (rtl_verify_edges): Invoke find_partition_fixes.
12853 (rtl_verify_bb_pointers): Update comments.
12854 (rtl_verify_bb_layout): Ditto.
12855 * basic-block.h (probably_never_executed_edge_p): Declare.
12856 (fixup_partitions): Ditto.
12857 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
12858 * bb-reorder.c (sanitize_hot_paths): New function.
12859 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
12860 sanitize_hot_paths.
12861 * predict.c (probably_never_executed_edge_p): New routine.
12862 * cfg.c (check_bb_profile): Add partition insanity warnings.
12864 2013-08-30 Meador Inge <meadori@codesourcery.com>
12866 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
12868 2013-08-30 Marek Polacek <polacek@redhat.com>
12870 * Makefile.in (ubsan.o): Add.
12871 (c-family/c-ubsan.o): Add.
12872 (builtins.o): Add ubsan.h dependency.
12873 * ubsan.h: New file.
12874 * ubsan.c: New file.
12875 * common.opt: Add -fsanitize=undefined option.
12876 (flag_sanitize): Add variable.
12877 (fsanitize=): Add option. Add Driver.
12878 (fsanitize=thread): Remove option.
12879 (fsanitize=address): Likewise.
12880 (static-libubsan): New option.
12881 * doc/invoke.texi: Document the new flag and -static-libubsan.
12882 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
12883 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
12884 * builtin-attrs.def (ATTR_COLD): Define.
12885 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
12886 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
12887 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
12888 * flag-types.h (sanitize_code): New enum.
12889 * opts.c (common_handle_option): Parse command line arguments
12890 of -fsanitize=. Add -fsanitize=unreachable option.
12891 * varasm.c (get_variable_section): Adjust.
12892 (assemble_noswitch_variable): Likewise.
12893 (assemble_variable): Likewise.
12894 (output_constant_def_contents): Likewise.
12895 (categorize_decl_for_section): Likewise.
12896 (place_block_symbol): Likewise.
12897 (output_object_block): Likewise.
12898 * builtins.def: Likewise.
12899 * toplev.c (compile_file): Likewise.
12900 (process_options): Likewise.
12901 * cppbuiltin.c: Likewise.
12902 * tsan.c (tsan_pass): Likewise.
12903 (tsan_gate): Likewise.
12904 (tsan_gate_O0): Likewise.
12905 * cfgexpand.c (partition_stack_vars): Likewise.
12906 (expand_stack_vars): Likewise.
12907 (defer_stack_allocation): Likewise.
12908 (expand_used_vars): Likewise.
12909 * cfgcleanup.c (old_insns_match_p): Likewise.
12910 * asan.c (asan_finish_file): Likewise.
12911 (asan_instrument): Likewise.
12912 (gate_asan): Likewise.
12913 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
12914 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
12915 (asan_global_struct): Use pointer_sized_int_node instead
12916 calling build_nonstandard_integer_type.
12917 (initialize_sanitizer_builtins): Likewise.
12918 (asan_finish_file): Likewise.
12919 * gcc.c: Document %{%:function(args):X}.
12920 (static_spec_functions): Add sanitize.
12921 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
12922 store funcval != NULL there.
12923 (do_spec_1): Adjust handle_spec_function caller.
12924 (handle_braces): Allow %:function(args) as condition.
12925 (sanitize_spec_function): New function.
12926 (ADD_STATIC_LIBUBSAN_LIBS): Define.
12927 (LIBUBSAN_SPEC): Likewise.
12928 (LIBUBSAN_EARLY_SPEC): Likewise.
12929 (SANITIZER_SPEC): Handle libubsan.
12930 (SANITIZER_EARLY_SPEC): Likewise.
12931 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
12932 instead of fsanitize=address.
12933 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
12934 instead of fsanitize=address*.
12935 * builtins.c: Include ubsan.h.
12936 (fold_builtin_0): Instrument __builtin_unreachable.
12937 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
12938 instead of flag_asan.
12939 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
12940 (pointer_sized_int_node): Define.
12941 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
12943 2013-08-30 Mike Stump <mikestump@comcast.net>
12945 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
12948 2013-08-29 Jan Hubicka <jh@suse.cz>
12950 * cgraph.c (cgraph_function_body_availability): Handle weakref
12952 * passes.def: Remove pass_fixup_cfg.
12953 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
12954 track when we need to remove functions.
12955 (gate_ipa_inline): Execute inlining always; add comment why.
12956 (pass_data_ipa_inline): Remove TODO_remove_functions.
12957 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
12958 do not produce summaries.
12959 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
12960 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
12961 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
12962 constant pool and vtable.
12964 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
12966 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
12967 New arm_neon.h's internal macros to specify 64-bit constants.
12968 Avoid using stdint.h's macros.
12970 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
12972 * recog.c (verify_changes): Verify that changes[i].old is non-zero
12973 before applying REG_P.
12975 2013-08-30 Jakub Jelinek <jakub@redhat.com>
12977 PR tree-optimization/58277
12978 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
12979 after seeing too many stmts with vdef in between dombb and current
12980 bb, invalidate everything.
12982 2013-08-30 Richard Biener <rguenther@suse.de>
12984 * fold-const.c (fold_single_bit_test): Fix overflow test.
12986 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
12988 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
12989 and which can live in a register, always retrieve the value on entry.
12990 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
12991 passed by invisible reference specially.
12992 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
12993 (vt_add_function_parameter): Correctly deal with a parameter passed by
12994 invisible reference.
12996 2013-08-30 Jan Hubicka <jh@suse.cz>
12998 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
13000 2013-08-30 Richard Biener <rguenther@suse.de>
13002 PR tree-optimization/58228
13003 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
13004 allow invariant loads in nested loop vectorization.
13006 2013-08-30 Richard Biener <rguenther@suse.de>
13008 PR tree-optimization/58223
13009 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
13010 (has_anti_or_output_dependence): ... this and adjust to also
13011 look for output dependences.
13012 (mark_nodes_having_upstream_mem_writes): Adjust.
13013 (rdg_flag_uses): Likewise.
13015 2013-08-30 Richard Biener <rguenther@suse.de>
13017 PR tree-optimization/58010
13018 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
13019 assert that we have a loop-closed PHI.
13021 2013-08-29 Jan Hubicka <jh@suse.cz>
13023 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
13024 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
13025 * lto-section-in.c (lto_free_function_in_decl_state): New function.
13026 (lto_free_function_in_decl_state_for_node): New function.
13028 2013-08-29 Xinliang David Li <davidxl@google.com>
13030 * loop-unroll.c (report_unroll_peel): Minor message change.
13031 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
13032 Emit alignment peeling message with default -fopt-info.
13033 (vect_loop_versioning): Emit loop version info message.
13034 * tree-vectorizer.c (vectorize_loops): Minor message change.
13035 (execute_vect_slp): Ditto.
13037 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
13039 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
13040 of the clone from the DECL_NAME of the original function.
13042 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
13044 * passes.c (register_pass): Add overload.
13045 * tree-pass.h (register_pass): Forward declare it. Add comment.
13047 2013-08-29 Jan Hubicka <jh@suse.cz>
13049 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
13050 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
13051 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
13052 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
13053 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
13054 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
13055 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
13056 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
13058 2013-08-29 Teresa Johnson <tejohnson@google.com>
13060 * dumpfile.c (dump_loc): Output column number.
13061 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
13062 * doc/invoke.texi: Document optall -fopt-info flag.
13063 * profile.c (read_profile_edge_counts): Use new dump framework.
13064 (compute_branch_probabilities): Ditto.
13065 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
13066 when pass not in any opt group.
13067 * pass_manager.h (pass_manager::get_pass_profile): New method.
13068 * value-prof.c (check_counter): Use new dump framework.
13069 (check_ic_target): Ditto.
13070 * coverage.c (get_coverage_counts): Ditto.
13071 (coverage_init): Setup new dump framework.
13073 2013-08-29 Richard Biener <rguenther@suse.de>
13075 PR tree-optimization/58246
13076 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
13077 handle the dominance check inside a basic-block.
13079 2013-08-29 Richard Biener <rguenther@suse.de>
13081 PR middle-end/57287
13082 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
13083 of default defs that appear in abnormal PHI nodes.
13085 2013-08-29 Richard Biener <rguenther@suse.de>
13087 PR tree-optimization/57685
13088 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
13089 single-use operands to avoid exponential complexity.
13091 2013-08-28 Dehao Chen <dehao@google.com>
13093 * ipa-inline.c (edge_badness): Fix integer underflow.
13095 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
13097 * gtm-builtins.def (_ITM_free): Declare leaf.
13099 2013-08-28 Jakub Jelinek <jakub@redhat.com>
13102 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
13103 (*tls_local_dynamic_base_64_largepic): Likewise.
13104 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
13105 Remove predicate from call operand.
13106 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
13107 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
13109 2013-08-28 Jeff Law <law@redhat.com>
13111 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
13112 checks for the number of predecessors and successors allowed.
13113 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
13114 which require copying a joiner block if there is a request which
13115 is a subpath that requires no joiner block copying.
13117 2013-08-28 Jan Hubicka <jh@suse.cz>
13119 * lto-streamer-out.c (DFS_write_tree_body): Drop
13120 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
13121 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
13122 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
13123 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
13124 Do not read DECL_ERROR_ISSUED.
13125 (unpack_ts_decl_with_vis_value_fields): Do not read
13127 (lto_input_ts_binfo_tree_pointers): Do not read
13128 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
13129 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
13130 write DECL_ERROR_ISSUED..
13131 (pack_ts_decl_with_vis_value_fields): Do not write
13133 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
13134 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
13135 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
13136 * tree.h (tree_decl_common): Update comment.
13137 (DECL_ERROR_ISSUED): Remove.
13139 2013-08-28 Jakub Jelinek <jakub@redhat.com>
13141 PR middle-end/58257
13142 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
13144 2013-08-28 Jan Hubicka <jh@suse.cz>
13146 * builtins.def (free): Declare leaf.
13148 2013-08-27 David Malcolm <dmalcolm@redhat.com>
13150 * gdbhooks.py: New.
13151 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
13152 * configure: Regenerate.
13154 2013-08-27 Martin Jambor <mjambor@suse.cz>
13156 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
13157 (ipa_ancestor_jf_data): Likewise.
13158 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
13159 (ipa_get_jf_pass_through_type_preserved): New function.
13160 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
13161 (ipa_get_jf_ancestor_type_preserved): New function.
13162 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
13163 (ipa_get_jf_ancestor_result): Likewise.
13164 (propagate_vals_accross_pass_through): Use
13165 ipa_get_jf_pass_through_result to do all the value mappings.
13166 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
13167 type_preserved flag.
13168 (ipa_set_jf_cst_copy): New function.
13169 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
13170 (ipa_set_jf_arith_pass_through): Likewise.
13171 (ipa_set_ancestor_jf): Likewise.
13172 (compute_complex_assign_jump_func): Set type_preserved instead of
13174 (ipa_compute_jump_functions_for_edge): Likewise.
13175 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
13176 (update_jump_functions_after_inlining): Update type_preserved.
13177 Explicitely create jump functions when combining one with pass_through.
13178 (ipa_write_jump_function): Stream the type_preserved flags.
13179 (ipa_read_jump_function): Likewise.
13181 2013-08-27 Jakub Jelinek <jakub@redhat.com>
13182 Aldy Hernandez <aldyh@redhat.com>
13184 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
13185 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
13186 * function.h (struct function): Add has_force_vect_loops and
13188 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
13189 * gimple.c (gimple_build_omp_critical): Add KIND argument and
13191 * gimple.def: Update CLAUSES comments.
13192 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
13193 (gimple_build_omp_for): Add argument to prototype.
13194 (gimple_omp_for_kind): New.
13195 (gimple_omp_for_set_kind): New.
13196 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
13197 GOVD_DATA_SHARE_CLASS.
13198 (enum omp_region_type): Add ORT_SIMD.
13199 (gimple_add_tmp_var): Handle ORT_SIMD.
13200 (gimplify_var_or_parm_decl): Same.
13201 (is_gimple_stmt): Same.
13202 (omp_firstprivatize_variable): Same.
13203 (omp_add_variable): Only use splay_tree_insert if lookup failed.
13204 (omp_notice_variable): Handle ORT_SIMD.
13205 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
13206 (omp_check_private): Handle ORT_SIMD.
13207 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
13208 OMP_CLAUSE_SAFELEN.
13209 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
13210 Handle OMP_CLAUSE_LASTPRIVATE.
13211 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
13212 OMP_CLAUSE_SAFELEN.
13213 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
13214 (gimplify_expr): Handle OMP_SIMD.
13215 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
13216 (expand_GOMP_SIMD_VF): New.
13217 (expand_GOMP_SIMD_LAST_LANE): New.
13218 * internal-fn.def (GOMP_SIMD_LANE): New.
13219 (GOMP_SIMD_VF): New.
13220 (GOMP_SIMD_LAST_LANE): New.
13221 * omp-low.c: Include target.h.
13222 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
13223 OMP_CLAUSE_SAFELEN.
13224 (check_omp_nesting_restrictions): Same.
13226 (lower_rec_simd_input_clauses): New.
13227 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
13229 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
13230 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
13231 (expand_omp_build_assign): New.
13232 (expand_omp_for_init_counts): New.
13233 (expand_omp_for_init_vars): New.
13234 (extract_omp_for_update_vars): New.
13235 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
13236 and rewrite accordingly.
13237 (expand_omp_simd): New.
13238 (expand_omp_for): Use expand_omp_simd.
13239 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
13240 (lower_omp_for): Do not lower the body.
13241 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
13242 in their own loops.
13243 * tree-flow.h (find_omp_clause): Remove prototype.
13244 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
13245 forcing vectorization of the loop, or if flag_tree_vectorize.
13246 (gate_tree_if_conversion): Similarly.
13247 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
13248 gimple_build_omp_for.
13249 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
13250 * tree-parloops (create_parallel_loop): Pass kind argument to
13251 gimple_build_omp_for.
13252 * tree-pretty-print.c (dump_omp_clause): Add cases for
13253 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
13254 OMP_CLAUSE__SIMDUID_.
13255 (dump_generic_node): Handle OMP_SIMD.
13256 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
13257 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
13258 unroll OMP_SIMD loops here.
13259 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
13260 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
13262 (vect_analyze_data_refs): Handle simd loops.
13263 * tree-vect-loop.c (vectorizable_live_operation): Handle
13265 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
13266 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
13267 (vectorizable_load): Same.
13268 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
13269 (struct simduid_to_vf): New.
13270 (simduid_to_vf::hash): New.
13271 (simduid_to-vf::equal): New.
13272 (struct simd_array_to_simduid): New.
13273 (simd_array_to_simduid::hash): New.
13274 (simd_array_to_simduid::equal): New.
13275 (adjust_simduid_builtins): New.
13276 (struct note_simd_array_uses_struct): New.
13277 (note_simd_array_uses_cb): New.
13278 (note_simd_array_uses): New.
13279 (vectorize_loops): Handle simd hints and adjust simd builtins
13281 * tree-vectorizer.h (struct _stmt_vec_info): Add
13282 simd_lane_access_p field.
13283 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
13284 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
13285 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
13286 (omp_clause_code_name): Same.
13287 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
13288 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
13289 * tree.def (OMP_SIMD): New entry.
13290 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
13291 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
13292 (OMP_CLAUSE_DECL): Adjust range for new clauses.
13293 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
13294 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
13295 (OMP_CLAUSE_LINEAR_STEP): New.
13296 (OMP_CLAUSE_SAFELEN_EXPR): New.
13297 (OMP_CLAUSE__SIMDUID__DECL): New.
13298 (find_omp_clause): New prototype.
13300 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
13302 * config/i386/driver-i386.c (host_detect_local_cpu): Update
13303 Haswell processor detection.
13305 2013-08-27 Christian Widmer <shadow@umbrox.de>
13308 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
13309 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
13310 AVX2 capable processors.
13312 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
13314 * config/aarch64/arm_neon.h: Replace all inline asm implementations
13315 of vget_low_* with implementations in terms of other intrinsics.
13317 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
13319 PR middle-end/57219
13320 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
13321 values to -1, 0 and 1.
13323 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
13325 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
13326 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
13327 (<optab><mode>3_insn): Remove.
13328 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
13329 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
13330 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
13331 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
13332 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
13333 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
13334 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
13335 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
13336 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
13337 (ror<mode>3_insn): Likewise.
13338 * config/aarch64/predicates.md (aarch64_simd_register): New.
13340 2013-08-27 Richard Biener <rguenther@suse.de>
13342 PR tree-optimization/57521
13343 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
13344 one edge is non-critical.
13345 (find_phi_replacement_condition): Make sure to use a non-critical
13346 edge. Cleanup and remove old bug workarounds.
13347 (bb_postdominates_preds): Remove.
13348 (if_convertible_loop_p_1): Do not compute post-dominators.
13349 (combine_blocks): Do not free post-dominators.
13350 (main_tree_if_conversion): Likewise.
13351 (pass_data_if_conversion): Add TODO_verify_ssa.
13353 2013-08-27 DJ Delorie <dj@redhat.com>
13355 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
13357 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
13359 * function.c (assign_parm_find_data_types): Set passed_mode and
13360 nominal_mode to the TYPE_MODE of nominal_type for the built
13361 pointer type in case of the struct-pass-by-reference.
13363 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
13365 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
13366 (UINT16_TYPE): Change default to "unsigned int".
13368 * config/avr/avr.opt (mfract-convert-truncate): New option.
13369 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
13370 is set, round negative fractional integers according to n1169
13371 when converting to integer types.
13373 2013-08-26 Jan Hubicka <jh@suse.cz>
13375 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
13376 methods can not be called indirectly when their address is not taken.
13378 2013-08-26 Jan Hubicka <jh@suse.cz>
13380 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
13383 2013-08-26 Jan Hubicka <jh@suse.cz>
13385 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
13388 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
13390 * reload.c (find_valid_class): Allow classes that do not include
13391 FIRST_PSEUDO_REGISTER - 1.
13393 2013-08-26 Jan Hubicka <jh@suse.cz>
13395 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
13396 fix edge count/frequency when speculation failed; fix type check
13397 for the direct call.
13399 2013-08-26 Jan Hubicka <jh@suse.cz>
13401 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
13403 2013-08-26 Jan Hubicka <jh@suse.cz>
13405 * ipa-inline-transform.c (inline_transform): Be ready for basic block
13406 to be changed by edge redirection.
13408 2013-08-26 Jan Hubicka <jh@suse.cz>
13410 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
13411 formating; add sanity check.
13412 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
13413 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
13414 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
13415 (resolve_noninline_speculation): Update callee keys, too.
13417 2013-08-26 Jan Hubicka <jh@suse.cz>
13419 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
13420 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
13422 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
13424 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
13427 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
13429 * config/i386/i386.c (ix86_debug_options): Remove prototype.
13430 (x86_64_elf_select_section): Ditto.
13431 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
13433 (ix86_pass_by_reference): Ditto.
13434 (output_set_got): Ditto.
13435 (ix86_unary_operator_ok): Ditto.
13436 (ix86_expand_builtin): Ditto.
13438 2013-08-23 Jan Hubicka <jh@suse.cz>
13440 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
13442 2013-08-23 Jan Hubicka <jh@suse.cz>
13444 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
13445 (tree_decl_with_vis): Add FINAL field.
13447 2013-08-23 Jeff Law <law@redhat.com>
13449 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
13450 the debugging dump when the expression is fully redundant.
13452 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
13454 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
13455 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
13456 * pretty-print.c (pp_formatted_text_data): Likewise.
13457 (pp_write_text_to_stream): Likewise.
13458 (pp_write_text_as_dot_label_to_stream): Likewise.
13459 (pp_append_r): Likewise.
13460 (pp_format): Likewise.
13461 (pp_flush): Likewise.
13462 (pp_clear_output_area): Likewise.
13463 (pp_append_text): Likewise.
13464 (pp_formatted_text): Likewise.
13465 (pp_remaining_character_count_for_line): Likewise.
13466 (pp_newline): Likewise.
13467 (pp_character): Likewise.
13468 (output_buffer::~output_buffer): Define.
13469 (pretty_printer::~pretty_printer): Destruct output buffer.
13470 * pretty-print.h (output_buffer::~output_buffer): Declare.
13471 (pretty_printer::~pretty_printer): Declare virtual.
13473 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
13476 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
13477 HOST_WIDE_INT_M1U): New macros.
13478 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
13479 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
13480 unsigned -1 for lshift.
13481 * cse.c (cse_insn): Likewise.
13482 * double-int.c (rshift_double, lshift_double): Likewise.
13483 * builtins.c (fold_builtin_bitop): Likewise.
13484 * combine.c (force_to_mode): Likewise.
13485 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
13486 * simplify-rtx.c (simplify_const_unary_operation,
13487 simplify_const_binary_operation): Likewise.
13488 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
13489 check_va_list_escapes): Likewise.
13490 * rtlanal.c (nonzero_bits1): Likewise.
13491 * expmed.c (expand_smod_pow2): Likewise.
13492 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
13494 2013-08-23 Jan Hubicka <jh@suse.cz>
13496 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
13497 as having address taken.
13499 2013-08-23 Jan Hubicka <jh@suse.cz>
13501 * ipa-utils.h (method_class_type): Declare.
13502 * ipa-devirt.c (method_class_type): Export.
13504 * cgraphunit.c (analyze_functions): Do basic devirtualization;
13505 do not walk base classes of anonymous types.
13507 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
13509 PR rtl-optimization/58220
13510 PR regression/58221
13511 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
13512 handle SEQUENCE insns properly.
13514 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
13516 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
13518 (pp_newline_and_indent): Likewise.
13519 (pp_separate_with): Likewise.
13520 * pretty-print.c (pp_newline_and_flush): Define.
13521 (pp_newline_and_indent): Likewise.
13522 (pp_separate_with): Likewise.
13524 2013-08-23 Jakub Jelinek <jakub@redhat.com>
13527 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
13528 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
13530 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
13532 * config/i386/predicates.md (ext_sse_reg_operand): New.
13533 * config/i386/i386.md (*movti_internal): Use
13534 predicate to determine if EVEX is needed.
13535 (*movsi_internal): Ditto.
13536 (*movdf_internal): Ditto.
13537 (*movsf_internal): Ditto.
13538 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
13540 2013-08-23 Jakub Jelinek <jakub@redhat.com>
13542 PR tree-optimization/58209
13543 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
13544 (find_tail_calls): Give up for pointer result types if m is non-NULL.
13545 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
13546 emit POINTER_PLUS_EXPR.
13547 (create_tailcall_accumulator): For pointer result type accumulate in
13550 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
13552 * configure.ac: Add backslashes missing from the last change.
13553 * configure: Regenerate.
13555 2013-08-22 Jan Hubicka <jh@susue.cz>
13557 * ipa.c (function_and_variable_visibility): First remember function
13558 was global and then make it local.
13560 2013-08-22 Julian Brown <julian@codesourcery.com>
13562 * configure.ac: Add aarch64 to list of arches which use "nop" in
13564 * configure: Regenerate.
13566 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13568 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
13569 gnu_libc_has_function.
13570 * config/s390/tpf.h: Likewise.
13572 2013-08-22 Jan Hubicka <jh@susue.cz>
13574 * timevar.c (validate_phases): Add cast.
13576 2013-08-22 Jan Hubicka <jh@susue.cz>
13578 * timevar.c (validate_phases): Use size_t for memory.
13579 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
13581 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
13583 * pretty-print.h (output_buffer::output_buffer): Declare.
13584 (pretty_printer::pretty_printer): Likewise.
13585 (pp_construct): Remove.
13586 * pretty-print.c (output_buffer::output_buffer): Define.
13587 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
13588 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
13589 (print_gimple_expr): Likewise.
13590 (print_gimple_seq): Likewise.
13591 (gimple_dump_bb): Likewise.
13592 * sched-vis.c (dump_value_slim): Likewise.
13593 (dump_insn_slim): Likewise.
13594 (dump_rtl_slim): Likewise.
13595 (str_pattern_slim): Likewise.
13596 * tree-mudflap.c (mf_varname_tree): Likewise.
13597 * graph.c (print_graph_cfg): Likewise.
13598 (start_graph_dump): Likewise.
13599 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
13601 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
13602 pretty printer initialization.
13603 * coretypes.h (diagnostic_context): Remove superflous type alias
13605 (pretty_printer): Likewise. Declare directly as a class.
13606 (pretty_print_info): Remove declaration as class.
13607 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
13608 and pp_clear_output_area.
13609 (asan_add_global): Likewise.
13611 2013-08-22 Jan Hubicka <jh@suse.cz>
13613 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
13614 * ipa-utils.h (update_type_inheritance_graph): Declare.
13615 (possible_polymorphic_call_target_p): Declare.
13616 (possible_polymorphic_call_target_p): New.
13617 * ipa-devirt.c: Update toplevel comments.
13618 (cached_polymorphic_call_targets): Move up.
13619 (odr_type_d): Move ID down.
13620 (polymorphic_type_binfo_p): Update comment.
13621 (odr_hasher::remove): Likewise;
13622 (get_odr_type): Set anonymous_namespace.
13623 (dump_odr_type): Dump it.
13624 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
13625 (maybe_record_node): Record node in cached_polymorphic_call_targets.
13626 (record_binfo): Add comment.
13627 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
13629 (devirt_node_removal_hook): Do not iCE when cache is freed.
13630 (possible_polymorphic_call_target_p): New predicate.
13631 (update_type_inheritance_graph): New function.
13633 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
13634 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
13635 Sergey Lega <sergey.s.lega@intel.com>
13636 Anna Tikhonova <anna.tikhonova@intel.com>
13637 Ilya Tocar <ilya.tocar@intel.com>
13638 Andrey Turetskiy <andrey.turetskiy@intel.com>
13639 Ilya Verbin <ilya.verbin@intel.com>
13640 Kirill Yukhin <kirill.yukhin@intel.com>
13641 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
13643 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
13644 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
13645 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
13646 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
13647 (OPTION_MASK_ISA_AVX2_UNSET): Update.
13648 (OPTION_MASK_ISA_AVX512F_UNSET): New.
13649 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
13650 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
13651 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
13652 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
13653 OPT_mavx512pf, OPT_mavx512er cases.
13654 * config/i386/constraints.md (v): New constraint.
13655 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
13656 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
13657 (bit_AVX512CD): New.
13658 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
13659 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
13660 * config/i386/i386-c.c (ix86_target_macros_internal):
13661 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
13663 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
13664 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
13665 * config/i386/i386.c (regclass_map, dbx_register_map)
13666 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
13667 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
13668 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
13669 -mavx512pf options.
13670 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
13671 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
13672 -mavx512pf options. Fix formatting.
13673 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
13674 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
13675 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
13676 -mavx512cd, -mavx512pf options.
13677 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
13678 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
13679 (ix86_preferred_output_reload_class): Replace SSE_REGS with
13681 (ix86_hard_regno_mode_ok): Support 512-bit registers.
13682 (ix86_set_reg_reg_cost): Ditto.
13683 (x86_order_regs_for_local_alloc): Ditto.
13684 (MAX_VECT_LEN): Extend to 64-byte.
13685 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
13686 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
13687 (TARGET_AVX512ER, TARGET_AVX512CD): New.
13688 (BIGGEST_ALIGNMENT): Extend to 512-bits.
13689 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
13690 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
13691 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
13692 (SSE_REG_MODE_P): Support new modes.
13693 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
13694 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
13695 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
13696 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
13697 (REG_CLASS_CONTENTS): Add new registers.
13698 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
13699 (EXT_REX_SSE_REGNO_P): New.
13700 (HI_REGISTER_NAMES): Add new registers.
13701 * config/i386/i386.md: Define constants for new registers.
13702 (mode): Add new 512-bit modes.
13703 (prefix): Support evex prefix.
13704 (isa): Support avx512f, noavx512f, fma_avx512f.
13705 (ssemodesuffix): Add new 512-bit modes.
13707 (*movxi_internal_avx512f): Ditto.
13708 (*movdi_internal): Replace constraint "x" with the new constraint "v".
13710 (*movsi_internal): Likewise.
13711 (*movdf_internal): Likewise.
13712 (*movsf_internal): Likewise.
13713 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
13714 (<code><mode>3): Likewise.
13715 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
13717 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
13718 with the new constraint "v".
13719 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
13721 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
13722 with the new constraint "v".
13723 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
13724 (<sse2>_storedqu<avxsizesuffix>): Likewise.
13725 (*<plusminus_insn><mode>3): Likewise.
13726 (<sse>_vm<plusminus_insn><mode>3): Likewise.
13727 (*mul<mode>3): Likewise.
13728 (<sse>_vmmul<mode>3): Likewise.
13729 (<sse>_div<mode>3): Likewise.
13730 (<sse>_vmdiv<mode>3): Likewise.
13731 (<sse>_sqrt<mode>2): Likewise.
13732 (<sse>_vmsqrt<mode>2): Likewise.
13733 (*<code><mode>3_finite): Likewise.
13734 (*<code><mode>3) <smaxmin>: Likewise.
13735 (<sse>_vm<code><mode>3): Likewise.
13736 (*<code><mode>3) <any_logic>: Likewise.
13737 (*fma_fmadd_<mode>): Likewise.
13738 (*fma_fmsub_<mode>): Likewise.
13739 (*fma_fnmadd_<mode>): Likewise.
13740 (*fma_fnmsub_<mode>): Likewise.
13741 (*fma_fmaddsub_<mode>): Likewise.
13742 (*fma_fmsubadd_<mode>): Likewise.
13743 (*fmai_fmadd_<mode>): Likewise.
13744 (*fmai_fmsub_<mode>): Likewise.
13745 (*fmai_fnmadd_<mode>): Likewise.
13746 (*fmai_fnmsub_<mode>): Likewise.
13747 (sse_cvtsi2ss): Likewise.
13748 (sse_cvtsi2ssq): Likewise.
13749 (sse_cvtss2si): Likewise.
13750 (sse_cvtss2si_2): Likewise.
13751 (sse_cvtss2siq): Likewise.
13752 (sse_cvtss2siq_2): Likewise.
13753 (sse_cvttss2si): Likewise.
13754 (sse_cvtss2siq_2): Likewise.
13755 (float<sseintvecmodelower><mode>2): Likewise.
13756 (sse2_cvtsd2si_2): Likewise.
13757 (sse2_cvtsd2siq_2): Likewise.
13758 (*<plusminus_insn><mode>3): Likewise.
13759 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
13760 (*<sse4_1_avx2>_mul<mode>3): Likewise.
13761 (ashr<mode>3): Likewise.
13762 (<shift_insn><mode>3): Likewise.
13763 (avx2_<code><mode>3): Likewise.
13764 (*avx2_<code><mode>3): Likewise.
13765 (*andnot<mode>3): Likewise.
13766 (*<code><mode>3) <any_logic>: Likewise.
13767 (abs<mode>2): Likewise.
13768 (avx2_permvar<mode>): Likewise.
13769 (avx2_perm<mode>_1): Likewise.
13770 (*avx_vpermilp<mode>): Likewise.
13771 (avx_vpermilvar<mode>3): Likewise.
13772 (avx2_ashrv<mode>): Likewise.
13773 (avx2_<shift_insn>v<mode>): Likewise.
13774 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
13776 * doc/rtl.texi: Document XImode.
13778 2013-08-21 Jeff Law <law@redhat.com>
13780 * tree-flow.h (register_jump_thread): Pass vector of edges
13781 instead of each important edge.
13782 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
13783 thread path into a vector and pass that to register_jump_thread.
13784 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
13785 passed in edge vector to the current 3-edge form.
13788 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
13790 * dce.c (fini_dce): Call df_analyze again just in case
13791 delete_unmarked_insns removed anything.
13793 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
13795 * reload.h (struct reg_equivs): Rename to ..
13796 (struct reg_equivs_s): .. this.
13798 2013-08-20 Martin Liska <marxin.liska@gmail.com>
13800 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
13802 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
13804 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
13806 2013-08-21 Jeff Law <law@redhat.com>
13808 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
13809 simplify assignments too. If the RHS collapses to a singleton
13810 range, then return the value for the range.
13812 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
13814 * config/i386/sse.md (V16): Rename to...
13816 (mov<mode>): Update iterator name.
13817 (*mov<mode>_internal): Ditto.
13818 (push<mode>1): Ditto.
13819 (movmisalign<mode>): Ditto.
13821 2013-08-20 Jan Hubicka <jh@suse.cz>
13824 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
13825 entry for direct edges.
13826 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
13828 2013-08-20 David Malcolm <dmalcolm@redhat.com>
13830 Revert my last two changes, r201865 and r201864:
13833 2013-08-20 David Malcolm <dmalcolm@redhat.com>
13835 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
13836 instances can own GC refs.
13838 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
13839 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
13840 (gcc::context::gt_pch_nx): Likewise.
13841 (gcc::context::gt_pch_nx): Likewise.
13842 * ggc.h (gt_ggc_mx <T>): New.
13843 (gt_pch_nx_with_op <T>): New.
13844 (gt_pch_nx <T>): New.
13845 * passes.c (opt_pass::gt_ggc_mx): New.
13846 (opt_pass::gt_pch_nx): New.
13847 (opt_pass::gt_pch_nx_with_op): New.
13848 (pass_manager::gt_ggc_mx): New.
13849 (pass_manager::gt_pch_nx): New.
13850 (pass_manager::gt_pch_nx_with_op): New.
13851 (pass_manager::operator new): Use
13852 ggc_internal_cleared_alloc_stat rather than xcalloc.
13853 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
13854 (pass_manager::gt_ggc_mx): New.
13855 (pass_manager::gt_pch_nx): New.
13856 (pass_manager::gt_pch_nx_with_op): New.
13857 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
13858 (opt_pass::operator new): New.
13859 (opt_pass::gt_ggc_mx): New.
13860 (opt_pass::gt_pch_nx): New.
13861 (opt_pass::gt_pch_nx_with_op): New.
13864 2013-08-20 David Malcolm <dmalcolm@redhat.com>
13866 * Makefile.in (GTFILES): Add context.h.
13867 * context.c (gcc::context::operator new): New.
13868 (gcc::context::gt_ggc_mx): New.
13869 (gcc::context::gt_pch_nx): New.
13870 (gcc::context::gt_pch_nx): New.
13871 * context.h (gcc::context): Add GTY((user)) marking.
13872 (gcc::context::operator new): New.
13873 (gcc::context::gt_ggc_mx): New.
13874 (gcc::context::gt_pch_nx): New.
13875 (gcc::context::gt_pch_nx): New.
13876 (g): Add GTY marking.
13877 (gt_ggc_mx (gcc::context *)): New.
13878 (gt_pch_nx (gcc::context *)): New.
13879 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
13880 void *cookie)): New.
13881 * gengtype.c (open_base_files) <ifiles>: Add context.h.
13883 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
13885 * dce.c (fini_dce): Call df_analyze again just in case
13886 delete_unmarked_insns removed anything.
13888 2013-08-20 Teresa Johnson <tejohnson@google.com>
13890 PR rtl-optimizations/57451
13891 * final.c (reemit_insn_block_notes): Prevent lexical blocks
13892 from crossing split section boundaries.
13894 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
13896 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
13897 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
13898 with MULTLIB_DEFAULTS.
13900 2013-08-20 Nick Clifton <nickc@redhat.com>
13902 * target.def (narrow_volatile_bitfield): Note that the default
13903 value is false, not !TARGET_STRICT_ALIGN.
13904 * doc/tm.texi: Regenerate.
13906 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
13908 Fix LIB_SPEC for systems without libpthread.
13910 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
13911 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
13913 * config/i386/linux-common.h: Likewise.
13914 * config/mips/linux-common.h: Likewise.
13916 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
13918 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
13921 2013-08-20 David Malcolm <dmalcolm@redhat.com>
13923 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
13924 instances can own GC refs.
13926 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
13927 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
13928 (gcc::context::gt_pch_nx): Likewise.
13929 (gcc::context::gt_pch_nx): Likewise.
13930 * ggc.h (gt_ggc_mx <T>): New.
13931 (gt_pch_nx_with_op <T>): New.
13932 (gt_pch_nx <T>): New.
13933 * passes.c (opt_pass::gt_ggc_mx): New.
13934 (opt_pass::gt_pch_nx): New.
13935 (opt_pass::gt_pch_nx_with_op): New.
13936 (pass_manager::gt_ggc_mx): New.
13937 (pass_manager::gt_pch_nx): New.
13938 (pass_manager::gt_pch_nx_with_op): New.
13939 (pass_manager::operator new): Use
13940 ggc_internal_cleared_alloc_stat rather than xcalloc.
13941 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
13942 (pass_manager::gt_ggc_mx): New.
13943 (pass_manager::gt_pch_nx): New.
13944 (pass_manager::gt_pch_nx_with_op): New.
13945 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
13946 (opt_pass::operator new): New.
13947 (opt_pass::gt_ggc_mx): New.
13948 (opt_pass::gt_pch_nx): New.
13949 (opt_pass::gt_pch_nx_with_op): New.
13951 2013-08-20 David Malcolm <dmalcolm@redhat.com>
13953 * Makefile.in (GTFILES): Add context.h.
13954 * context.c (gcc::context::operator new): New.
13955 (gcc::context::gt_ggc_mx): New.
13956 (gcc::context::gt_pch_nx): New.
13957 (gcc::context::gt_pch_nx): New.
13958 * context.h (gcc::context): Add GTY((user)) marking.
13959 (gcc::context::operator new): New.
13960 (gcc::context::gt_ggc_mx): New.
13961 (gcc::context::gt_pch_nx): New.
13962 (gcc::context::gt_pch_nx): New.
13963 (g): Add GTY marking.
13964 (gt_ggc_mx (gcc::context *)): New.
13965 (gt_pch_nx (gcc::context *)): New.
13966 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
13967 void *cookie)): New.
13968 * gengtype.c (open_base_files) <ifiles>: Add context.h.
13970 2013-08-20 Alan Modra <amodra@gmail.com>
13973 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
13974 (rs6000_emit_epilogue): Likewise.
13976 2013-08-19 Dehao Chen <dehao@google.com>
13978 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
13980 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
13981 Jakub Jelinek <jakub@redhat.com>
13983 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
13984 (BUILT_IN_FABSD64): Likewise.
13985 (BUILT_IN_FABSD128): Likewise.
13986 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
13987 (fold_builtin_1): Likewise.
13988 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
13989 destination and source operands.
13990 (*abstd2_fpr): Likewise.
13991 (*nabstd2_fpr): Likewise.
13993 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
13995 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
13998 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
14000 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
14003 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
14005 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
14006 * builtins.c (default_libc_has_function): New.
14007 (gnu_libc_has_function): Ditto.
14008 (no_c99_libc_has_function): Ditto.
14009 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
14010 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
14011 (fold_builtin_sincos): Likewise.
14012 (fold_builtin_cexp): Likewise.
14013 * builtins.def (DEF_C94_BUILTIN): Likewise.
14014 (DEF_C99_BUILTIN): Likewise.
14015 (DEF_C99_C90RES_BUILTIN): Likewise.
14016 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
14017 definitions to using this define.
14018 * config/darwin-protos.h (darwin_libc_has_function): New.
14019 * config/darwin.c (darwin_libc_has_function): Ditto.
14020 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
14021 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
14022 * config/darwin.h: Ditto.
14023 * config/elfos.h: Ditto.
14024 * config/freebsd.h: Ditto.
14025 * config/i386/cygming.h: Ditto.
14026 * config/i386/djgpp.h: Ditto.
14027 * config/i386/i386-interix.h: Ditto.
14028 * config/microblaze/microblaze.h: Ditto.
14029 * config/mmix/mmix.h: Ditto.
14030 * config/gnu-user.h: Ditto.
14031 * config/ia64/hpux.h: Ditto.
14032 * config/pa/pa-hpux.h: Ditto.
14033 * config/pdp11/pdp11.h: Ditto.
14034 * config/picochip/picochip.h: Ditto.
14035 * config/linux.h: Ditto.
14036 * config/netbsd.h: Ditto.
14037 * config/openbsd.h: Ditto.
14038 * config/rs6000/aix43.h: Ditto.
14039 * config/rs6000/aix51.h: Ditto.
14040 * config/rs6000/aix52.h: Ditto.
14041 * config/rs6000/aix53.h: Ditto.
14042 * config/rs6000/aix61.h: Ditto.
14043 * config/rs6000/darwin.h: Ditto.
14044 * config/rs6000/linux.h: Ditto.
14045 * config/rs6000/linux64.h: Ditto.
14046 * config/s390/tpf.h: Ditto.
14047 * config/sol2-10.h: Ditto.
14048 * config/sol2.h: Ditto.
14049 * config/vms/vms.h: Ditto.
14050 * config/vxworks.h: Ditto.
14051 * config/linux-android.c (linux_android_libc_has_function):
14052 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
14053 * config/linux-protos.h (linux_android_libc_has_function):
14055 * config/i386/i386.c (ix86_libc_has_function): New.
14056 * config/i386/i386-protos.h
14057 (ix86_libc_has_function): New declaration.
14058 * config/i386/i386.md
14059 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
14060 ("isinf<mode>2): Likewise.
14061 * convert.c (convert_to_integer): Using new target hook
14062 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
14063 TARGET_C99_FUNCTIONS.
14064 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
14065 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
14066 * coretypes.h (function_class): New enum for different
14067 classes of functions.
14068 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
14069 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
14070 (TARGET_HAS_SINCOS): Likewise.
14071 (TARGET_LIBC_HAS_FUNCTION): New.
14072 * doc/tm.texi: Regenerated.
14073 * targhooks.h (default_libc_has_function): New declaration.
14074 (no_c99_libc_has_function): Ditto.
14075 (gnu_libc_has_function): Ditto.
14076 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
14077 and TARGET_HAS_SINCOS.
14079 2013-08-18 Jan Hubicka <jh@suse.cz>
14081 * Makeifle-in (ipa-devirt.o): New.
14082 (GTFILES): Add ipa-utils.h and ipa-devirt.c
14083 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
14084 (analyze_functions): Look into possible targets of polymorphic call.
14085 * dumpfile.c (dump_files): Add type-inheritance dump.
14086 * dumpfile.h (TDI_inheritance): New.
14087 * ipa-devirt.c: New file.
14088 * ipa-utils.h (odr_type_d): Forward declare.
14089 (odr_type): New type.
14090 (build_type_inheritance_graph): Declare.
14091 (possible_polymorphic_call_targets): Declare and introduce inline
14092 variant when only edge is pased.
14093 (dump_possible_polymorphic_call_targets): Likewise.
14094 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
14095 * tree.c (type_in_anonymous_namespace_p): Break out from ...
14096 (types_same_for_odr): ... here.
14097 * tree.h (type_in_anonymous_namespace_p): Declare.
14099 2013-08-18 Jakub Jelinek <jakub@redhat.com>
14101 PR tree-optimization/58006
14102 * tree-parloops.c (take_address_of): Don't ICE if get_name
14104 (eliminate_local_variables_stmt): Remove clobber stmts.
14106 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
14108 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
14109 error message is issued for an alias to undefined symbol.
14111 2013-08-18 Jan Hubicka <jh@suse.cz>
14113 * cgraph.c (cgraph_create_indirect_edge): Discover
14114 polymorphic calls and record basic info into indirect_info.
14115 * gimple-fold.c (gimple_fold_call): When doing BINFO based
14116 devirtualization, ignore objc function calls.
14117 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
14118 call with no parm index info.
14119 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
14120 * tree.c (virtual_method_call_p): New function.
14121 * tree.h (virtual_method_call_p): Declare.
14123 2013-08-16 Jan Hubicka <jh@suse.cz>
14125 PR middle-end/58179
14126 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
14128 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
14130 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
14133 2013-08-16 David Malcolm <dmalcolm@redhat.com>
14135 * gengtype.c (type_for_name): Add special-case support for
14136 locating types within the "gcc::" namespace.
14137 (open_base_files): Emit a "using namespace gcc" directive.
14139 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
14142 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
14143 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
14145 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
14146 array instead of each individual operand as a separate argument.
14147 (emit_fusion_gpr_load): Likewise.
14148 (expand_fusion_gpr_load): Add new function declaration.
14150 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
14151 signature to have the operands passed as an array, instead of as
14152 separate arguments. Allow ZERO_EXTEND to be in the memory
14153 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
14154 depend on the register live/dead flags when peepholes are run.
14155 (expand_fusion_gpr_load): New function to be called from the
14156 peephole2 pass, to change the register that addis sets to be the
14158 (emit_fusion_gpr_load): Change the calling signature to have the
14159 operands passed as an array, instead of as separate arguments.
14160 Allow ZERO_EXTEND to be in the memory address, and also
14161 SIGN_EXTEND if -mpower8-fusion-sign.
14163 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
14164 unspec enumeration.
14165 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
14166 adjust the register addis loads up in the peephole2 pass. Do not
14167 depend on the register live/dead state when the peephole pass is done.
14169 2013-08-16 David Malcolm <dmalcolm@redhat.com>
14171 * gengtype.c (create_user_defined_type): Ensure that the kind
14172 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
14173 declaration is seen before the GTY((user)) marking.
14175 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
14178 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
14180 2013-08-16 Jan Hubicka <jh@suse.cz>
14182 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
14183 arugment expected_type.
14184 (gimple_fold_call): Use it.
14185 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
14186 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
14187 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
14188 (try_make_edge_direct_virtual_call): Likewise.
14189 * tree.c (obj_type_ref_class): New.
14190 * tree.h (obj_type_ref_class): Use it.
14192 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
14194 * sched-vis.c (rtl_slim_pp_initialized): Remove.
14195 (rtl_slim_pp): Likewise.
14196 (init_rtl_slim_pretty_print): Likewise.
14197 (dump_value_slim): Don't call it. Use local pretty printer.
14198 (dump_insn_slim): Likewise.
14199 (dump_rtl_slim): Likewise.
14200 (str_pattern_slim): Likewise.
14201 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
14204 2013-08-16 Jakub Jelinek <jakub@redhat.com>
14206 PR tree-optimization/58164
14207 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
14208 walk gimple_goto_dest of GIMPLE_GOTO.
14210 PR tree-optimization/58165
14211 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
14212 bi_call must be the last stmt in a bb, don't split_block, instead
14213 use fallthru edge from it and give up if there is none.
14214 Release conds vector when returning early.
14216 2013-08-14 Xinliang David Li <davidxl@google.com>
14218 * config/i386/i386.c (ix86_option_override_internal):
14219 Remove unused variable and field.
14221 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14224 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
14225 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
14226 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
14227 and Linux, correct BLKmode alignment when 128-bit alignment is
14228 required and compatibility flag is not set.
14229 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
14230 for zero-size arguments when compatibility flag is not set.
14232 2013-08-14 Jakub Jelinek <jakub@redhat.com>
14234 PR tree-optimization/58145
14235 * tree-sra.c (build_ref_for_offset): If prev_base has
14236 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
14238 2013-08-14 Xinliang David Li <davidxl@google.com>
14240 * config/i386/i386.c (ix86_option_override_internal):
14241 Fix uninitialized variable error.
14243 2013-08-14 Xinliang David Li <davidxl@google.com>
14245 * config/i386/i386.opt: Define two new options.
14246 * config/i386/x86-tune.def: Add arch selector field in macros.
14247 * config/i386/i386.h: Adjust macro definition.
14248 * config/i386/i386.c (ix86_option_override_internal):
14250 (parse_mtune_ctrl_str): New function.
14251 (set_ix86_tune_features): New function.
14252 (ix86_function_specific_restore): Call the new helper function.
14254 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
14256 PR rtl-optimization/57662
14257 * sel-sched.c (code_motion_process_successors): When the current insn
14258 is removed after the recursive traversal, break from the loop.
14259 Add comments and debug printouts.
14261 2013-08-14 Jakub Jelinek <jakub@redhat.com>
14262 Alexandre Oliva <aoliva@redhat.com>
14265 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
14266 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
14267 there also UNSPEC_PLTOFF.
14269 2013-08-14 Marek Polacek <polacek@redhat.com>
14271 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
14272 NUM_CONDITIONS bit positions.
14274 2013-08-13 Cary Coutant <ccoutant@google.com>
14276 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
14277 (attr_checksum): Hash vector contents instead of pointer.
14278 (attr_checksum_ordered): Likewise.
14280 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
14282 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
14283 when Pmode != word_mode. Add length_address attribute.
14284 (sse3_monitor_<mode>): Merge from sse3_monitor and
14285 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
14286 Pmode != word_mode. Update insn length attribute.
14287 * config/i386/i386.c (ix86_option_override_internal): Update
14288 ix86_gen_monitor selection for merged sse3_monitor insn.
14290 2013-08-13 Julian Brown <julian@codesourcery.com>
14292 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
14293 perform invalid legitimization on greater-than-word-size modes for
14294 TARGET_E500_DOUBLE.
14296 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
14298 * ira.c (setup_class_translate_array): Use aclass instead of cl
14299 for classes not fully covered by allocno classes.
14301 2013-08-13 Jakub Jelinek <jakub@redhat.com>
14303 PR tree-optimization/57661
14304 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
14305 * tree-inline.c (tree_function_versioning): Initialize it.
14306 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
14307 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
14308 that is not being copied.
14311 * asan.c (instrument_strlen_call): Fix typo in comment.
14312 Use char * type even for the lhs of POINTER_PLUS_EXPR.
14314 2013-08-13 Steve Ellcey <sellcey@mips.com>
14316 * config/mips/mips.md (prefetch): Use lw instead of ld on
14317 loongson in 32bit mode.
14319 2013-08-13 Nick Clifton <nickc@redhat.com>
14321 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
14323 2013-08-13 Jan Hubicka <jh@suse.cz>
14325 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
14326 introduced edge; fix typo in sanity check.
14327 (cgraph_resolve_speculation): Export; improve diagnostic.
14328 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
14329 speculation at type mismatch.
14330 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
14331 (cgraph_resolve_speculation): Declare.
14332 (symtab_can_be_discarded): New function.
14333 * value-prof.c (gimple_ic_transform): Remove actual transform code.
14334 * ipa-inline-transform.c (speculation_removed): New global var.
14335 (clone_inlined_nodes): See if speculation can be removed.
14336 (inline_call): If speculations was removed, we growths may not match.
14337 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
14338 (speculation_useful_p): New function.
14339 (resolve_noninline_speculation): New function.
14340 (inline_small_functions): Resolve useless speculations.
14341 * ipa-inline.h (speculation_useful_p): Declare
14342 * ipa.c (can_replace_by_local_alias): Simplify.
14343 (ipa_profile): Produce speculative calls in non-lto, too;
14344 add simple cost model; produce local aliases.
14346 2013-08-13 David Malcolm <dmalcolm@redhat.com>
14348 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
14351 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
14353 * config/i386/i386.c (ix86_function_versions): Use error + inform.
14355 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
14357 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
14358 iterator instead of X87MODEF.
14360 2013-08-12 Perez Read <netfirewall@gmail.com>
14363 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
14364 operand 0 for intel asm alternative.
14365 (*movabs<mode>_2): Ditto for operand 1.
14367 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
14369 * config/aarch64/arm_none.h
14370 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
14372 2013-08-12 Nick Clifton <nickc@redhat.com>
14374 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
14376 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
14378 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
14379 expand for QI/HImode operand to produce more effictive code for
14380 unsigned char(short) --> float(double) conversion.
14382 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
14384 * doc/invoke.texi: Mention that -ftls-model does not force the final
14387 2013-08-12 Marek Polacek <polacek@redhat.com>
14388 Marc Glisse <marc.glisse@inria.fr>
14390 PR tree-optimization/57980
14391 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
14392 when creating -1 constant.
14394 2013-08-10 Jan Hubicka <jh@suse.cz>
14396 Workaround binutils PR14342.
14397 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
14398 (gimple_init_edge_profiler): Likewise.
14399 (gimple_gen_ic_func_profiler): Likewise.
14401 2013-08-09 Jan Hubicka <jh@suse.cz>
14403 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
14405 2013-08-09 Xinliang David Li <davidxl@google.com>
14407 * config/i386/stringop.def: New file.
14408 * config/i386/stringop.opt: New file.
14409 * config/i386/i386-opts.h: Include stringopt.def.
14410 * config/i386/i386.opt: Include stringopt.opt.
14411 * config/i386/i386.c (ix86_option_override_internal):
14412 Override default size based stringop inline strategies with options.
14413 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
14416 2013-08-09 Jan Hubicka <jh@suse.cz>
14418 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
14420 2013-08-09 Jan Hubicka <jh@suse.cz>
14422 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
14424 (dump_cgraph_node): Dump profile-id.
14425 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
14426 and common_target_probability.
14427 * lto-cgraph.c (lto_output_edge): Stream common targets.
14428 (lto_output_node): Stream profile ids.
14429 (input_node): Stream profile ids.
14430 (input_edge): Stream common targets.
14431 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
14432 * ipa.c: Include value-prof.h
14433 (ipa_profile_generate_summary): Turn indirect call statement histograms
14434 into common targets.
14435 (ipa_profile): Turn common targets into speculative edges.
14437 2013-08-09 Jan Hubicka <jh@suse.cz>
14439 * cgraph.h (cgraph_node): Add profile_id.
14440 * value-prof.c (cgraph_node_map): Turn into pointer_map.
14441 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
14442 (del_node_map): Update.
14443 (find_func_by_funcdef_no): Replace by ...
14444 (find_func_by_profile_id): ... this one.
14445 (gimple_ic_transform): Do not remove useful histograms when
14446 speculation is not done; dump info when indirect call removal
14448 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
14449 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
14450 (__gcov_indirect_call_profiler_v2): .. this one.
14451 * profile.h (init_node_map): Update.
14452 * coverage.c (coverage_compute_profile_id): New function.
14453 * coverage.h (coverage_compute_profile_id): Declare.
14454 * tree-profile.c (init_ic_make_global_vars): Make
14455 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
14456 (gimple_init_edge_profiler): Update prototype of
14457 __gcov_indirect_call_profiler.
14458 (gimple_gen_ic_func_profiler): Simplify.
14459 (tree_profiling): Use init_node_map
14461 2013-08-09 Jan Hubicka <jh@suse.cz>
14463 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
14464 non-speculative refs.
14465 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
14466 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
14467 (cgraph_set_call_stmt): Likewise.
14468 (cgraph_create_edge_1): Fix release checking compilatoin;
14469 clear lto_stmt_uid.
14470 (cgraph_free_edge): Free indirect info.
14471 (cgraph_turn_edge_to_speculative): New function.
14472 (cgraph_speculative_call_info): New function.
14473 (cgraph_make_edge_direct): Return direct edge; handle speculation.
14474 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
14475 (dump_cgraph_node): Dump speculation.
14476 (verify_edge_count_and_frequency): Accept speculative edges.
14477 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
14478 (verify_cgraph_node): Handle speculation.
14479 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
14480 (cgraph_set_call_stmt): Update prototype.
14481 (cgraph_make_edge_direct): Update prototype.
14482 (cgraph_speculative_call_info): Declare.
14483 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
14484 to change; update call of ipa_find_references.
14485 * ipa-ref.c (ipa_record_reference): Fix return value; clear
14486 lto_stmt_uid and speculative flags.
14487 (ipa_dump_references): Dump speculation.
14488 (ipa_clone_references): Clone speculative flag.
14489 (ipa_clone_referring): Likewise.
14490 (ipa_clone_ref): New function.
14491 (ipa_find_reference): Look into lto_stmt_uids
14492 (ipa_clear_stmts_in_references): Do not clear speculative calls.
14493 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
14494 (ipa_find_reference): Update declaration.
14495 (ipa_clone_ref): Declare.
14496 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
14497 stream speculative flag.
14498 (lto_output_ref): Stream statements uids and speculation.
14499 (input_ref): Likewise.
14500 (input_edge): Stream speuclation.
14501 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
14502 (cgraph_set_call_stmt_including_clones): Handle speculation.
14503 * ipa-inline.c (heap_edge_removal_hook): New function.
14504 (inline_small_functions): Register it.
14505 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
14506 also initialize refs.
14507 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
14509 (try_make_edge_direct_simple_call): Likewise.
14510 (try_make_edge_direct_simple_call): Likewise.
14511 (update_indirect_edges_after_inlining): Likewise.
14512 (remove_described_reference): Look proper lto_stmt_uid.
14513 (propagate_controlled_uses): Likewise.
14514 (propagate_controlled_uses): Liekwise.
14515 * tree-inline.c (copy_bb): Copy speculative edges.
14516 (redirect_all_calls): New function.
14517 (copy_cfg_body): Do redirection after loop info is updated.
14518 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
14520 2013-08-09 Jan Hubicka <jh@suse.cz>
14522 * lto-streamer-out.c (output_function): Renumber PHIs.
14523 * lto-streamer-in.c (input_function): Likewise.
14525 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
14527 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
14528 (get_lane_unsigned): Likewise.
14529 (dup_lane_scalar): Likewise.
14530 (get_lane): enable for VALL.
14531 * config/aarch64/aarch64-simd.md
14532 (aarch64_dup_lane_scalar<mode>): Remove.
14533 (aarch64_get_lane_signed<mode>): Likewise.
14534 (aarch64_get_lane_unsigned<mode>): Likewise.
14535 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
14536 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
14537 (aarch64_get_lane<mode>): Enable for all vector modes.
14538 (aarch64_get_lanedi): Remove misleading constraints.
14539 * config/aarch64/arm_neon.h
14540 (__aarch64_vget_lane_any): Define.
14541 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
14542 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
14543 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
14544 * config/aarch64/iterators.md (VDQQH): New.
14545 (VDQQHS): Likewise.
14546 (vwcore): Likewise.
14548 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
14550 * configure.ac: Add GAS check for LEON instructions on SPARC.
14551 * configure: Regenerate.
14552 * config.in: Likewise.
14553 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
14555 * config/sparc/sparc.opt (LEON, LEON3): New masks.
14556 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
14558 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
14559 (AS_LEON_FLAG): New macro.
14560 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
14561 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
14562 Deal with LEON and LEON3 for the memory model.
14563 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
14564 (atomic_compare_and_swap<mode>_1): Likewise.
14565 (*atomic_compare_and_swap<mode>_1): Likewise.
14567 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
14569 * config/arm/neon.md (vcond): Fix floating-point vector
14570 comparisons against 0.
14572 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
14574 * lra-constraints.c (emit_spill_move): Remove assert.
14575 (process_alt_operands): Add more debugging
14576 output. Increase reject for spilling into memory. Decrease
14577 reject for reloading scratch.
14578 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
14580 2013-08-08 Steve Ellcey <sellcey@mips.com>
14582 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
14583 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
14584 micromips incompatible. Add nan2008.
14585 (MULTILIB_DIRNAMES): Add nan2008.
14586 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
14587 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
14588 and micromips incompatible. Add nan2008.
14589 (MULTILIB_DIRNAMES): Add nan2008.
14590 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
14592 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
14594 PR rtl-optimization/58079
14595 * combine.c (combine_simplify_rtx): Avoid using SUBST if
14596 simplify_comparison has widened a comparison with an integer.
14598 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14600 * config/arm/neon.md (movmisalign<mode>): Disable when we
14601 don't allow unaligned accesses.
14602 (*movmisalign<mode>_neon_store): Likewise.
14603 (*movmisalign<mode>_neon_load): Likewise.
14604 (*movmisalign<mode>_neon_store): Likewise.
14605 (*movmisalign<mode>_neon_load): Likewise.
14607 2013-08-08 Jan Hubicka <jh@suse.cz>
14609 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
14610 (make_pass_rebuild_cgraph_edges): Also clear references.
14611 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
14612 * ipa-inline-transform.c (inline_transform): Remove all references
14614 * cgraphunit.c (expand_function): Remove all references after
14616 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
14617 (ipa_find_reference): Rewrite to iterator.
14618 (remove_stmt_references): Likewise.
14619 (ipa_clear_stmts_in_references): New function.
14620 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
14621 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
14623 * ipa-split.c (split_function): Remove references in split function.
14625 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
14628 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
14629 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
14631 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
14634 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
14635 suggested mode for the assignment isn't compatible with the
14636 registers required.
14638 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
14641 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
14643 2013-08-07 Xinliang David Li <davidxl@google.com>
14645 * config/i386/i386.opt: New option -mtune-ctrl=.
14646 * config/i386/x86-tune.def: New file.
14647 * config/i386/i386.h: include x86-tune.def.
14648 * config/i386/i386.c (ix86_option_override_internal):
14649 Parsing -mtune-ctrl= option and set tune features.
14651 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
14654 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
14655 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
14658 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
14660 * rtl.h (update_alignments): Declare.
14661 * final.c (grow_label_align): New function extracted from...
14662 (shorten_branches): ...here. Call it.
14663 (update_alignments): New function.
14664 * reorg.c (sibling_labels): New variable.
14665 (get_label_before): Add SIBLING parameter. If it is non-zero, push
14666 the new label along with it onto the sibling_labels vector.
14667 (fill_simple_delay_slots): Adjust call to get_label_before.
14668 (fill_slots_from_thread): Likewise.
14669 (relax_delay_slots): Likewise.
14670 (make_return_insns): Likewise.
14671 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
14673 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
14675 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
14676 document its semantics.
14677 (diagnostic_report_diagnostic): Adjust accordingly.
14679 2013-08-07 David Malcolm <dmalcolm@redhat.com>
14681 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
14682 (sparc_option_override): ...and port to new C++ pass API.
14683 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
14685 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
14687 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
14689 2013-08-06 Caroline Tice <cmtice@google.com>
14691 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
14692 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
14693 * tree-pass.h: Add pass_vtable_verify.
14694 * varasm.c (assemble_variable): Add code to properly set the comdat
14695 section and name for the .vtable_map_vars section.
14696 (assemble_vtyv_preinit_initializer): New function.
14697 (default_sectin_type_flags): Make sure .vtable_map_vars section has
14699 * output.h: Add function decl for assemble_vtv_preinit_initializer.
14700 * vtable-verify.c: New file.
14701 * vtable-verify.h: New file.
14702 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
14703 initialiation levels.
14704 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
14705 * passes.def: Insert pass_vtable_verify.
14706 * aclocal.m4: Reorder includes.
14707 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
14708 -fvtv-counts options.
14709 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
14710 as appropriate, if -fvtable-verify=... is used.
14711 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
14712 -fvtable-verify=... is used.
14713 * Makefile.in (OBJS): Add vtable-verify.o to list.
14714 (vtable-verify.o): Add new build rule.
14715 (GTFILES): Add vtable-verify.c to list.
14716 * common.opt (fvtable-verify=): New flag.
14717 (vtv_priority): Values for fvtable-verify= flag.
14718 (fvtv-counts): New flag.
14719 (fvtv-debug): New flag.
14720 * tree.h (save_vtable_map_decl): New extern function decl.
14722 2013-08-07 David Malcolm <dmalcolm@redhat.com>
14724 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
14725 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
14726 (pass_data_rl78_devirt): ...new pass_data instance and...
14727 (make_pass_rl78_devirt): ...new function.
14728 (rl78_asm_file_start): Port pass registration to new C++ API.
14730 2013-08-07 David Malcolm <dmalcolm@redhat.com>
14732 * coretypes.h (rtl_opt_pass): Add.
14733 (gcc::context): Add.
14734 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
14735 (epiphany_init): Port to new C++ pass API.
14736 (epiphany_optimize_mode_switching): Likewise.
14737 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
14738 (pass_manager::get_pass_mode_switching): New.
14739 (pass_manager::get_pass_peephole2): New.
14740 * mode-switching.c (pass_mode_switching): Add clone method.
14741 * recog.c (pass_peephole2): Add clone method.
14742 (pass_split_all_insns): Add clone method.
14744 2013-08-06 David Malcolm <dmalcolm@redhat.com>
14746 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
14747 (mips_option_override): ...here, porting to new C++ API for passes.
14749 2013-08-06 Jan Hubicka <jh@suse.cz>
14751 * cgraph.c (cgraph_get_body): New function based on lto.c
14753 * cgraph.h (cgraph_get_body): Declare.
14754 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
14756 * cgraphunit.c (expand_function): Get body prior expanding.
14757 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
14758 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
14760 * passes.c (do_per_function_toporder): Get body.
14761 * tree-inline.c (expand_call_inline): Get body prior inlining it.
14762 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
14764 2013-08-06 Martin Jambor <mjambor@suse.cz>
14767 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
14768 is not re-finalized. Rename second parameter to no_collect.
14770 2013-08-06 Martin Jambor <mjambor@suse.cz>
14772 PR middle-end/58041
14773 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
14774 MEM_REF has proper alignment information.
14776 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
14779 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
14781 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
14782 rtx (*) (rtx, ...) with insn_gen_fn.
14783 * genoutput.c (output_insn_data): Cast gen_? function pointers to
14784 insn_gen_fn::stored_funcptr. Add initializer braces.
14786 2013-08-05 David Malcolm <dmalcolm@redhat.com>
14788 Rewrite how instances of passes are cloned to remove assumptions
14789 about their sizes (thus allowing pass subclasses to have
14790 additional data fields, albeit non-GC-managed ones at this point).
14792 * passes.c (make_pass_instance): Now that passes have clone
14793 methods, rewrite this function to eliminate XNEW and memcpy
14794 calls that used hardcoded sizes. Since this function no longer
14795 creates pass instances, rename it to...
14796 (add_pass_instance): ...this. Document the old way that passes were
14797 numbered and flagged, and rework this function to continue using it.
14798 (next_pass_1): Add an initial_pass argument for use by
14800 (position_pass): When adding multiple instances of a pass, use
14801 the pass's clone method, rather than relying on the XNEW/memcpy
14802 within the former make_pass_instance (now add_pass_instance).
14803 (pass_manager::pass_manager): When invoking next_pass_1, also supply
14804 the initial instance of the current pass within the pass manager.
14806 2013-08-05 David Malcolm <dmalcolm@redhat.com>
14808 This is the automated part of the conversion of passes from C
14809 structs to C++ classes.
14811 Patch autogenerated by refactor_passes.py from
14812 https://github.com/davidmalcolm/gcc-refactoring-scripts
14813 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
14815 * asan.c (pass_asan): Convert from a global struct to a subclass of
14816 gimple_opt_pass along with...
14817 (pass_data_asan): ...new pass_data instance and...
14818 (make_pass_asan): ...new function.
14819 (pass_asan_O0): Convert from a global struct to a subclass of
14820 gimple_opt_pass along with...
14821 (pass_data_asan_O0): ...new pass_data instance and...
14822 (make_pass_asan_O0): ...new function.
14823 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
14824 subclass of rtl_opt_pass along with...
14825 (pass_data_inc_dec): ...new pass_data instance and...
14826 (make_pass_inc_dec): ...new function.
14827 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
14828 a subclass of rtl_opt_pass along with...
14829 (pass_data_reorder_blocks): ...new pass_data instance and...
14830 (make_pass_reorder_blocks): ...new function.
14831 (pass_duplicate_computed_gotos): Convert from a global struct to a
14832 subclass of rtl_opt_pass along with...
14833 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
14834 (make_pass_duplicate_computed_gotos): ...new function.
14835 (pass_partition_blocks): Convert from a global struct to a subclass of
14836 rtl_opt_pass along with...
14837 (pass_data_partition_blocks): ...new pass_data instance and...
14838 (make_pass_partition_blocks): ...new function.
14839 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
14840 struct to a subclass of rtl_opt_pass along with...
14841 (pass_data_branch_target_load_optimize1): ...new pass_data instance
14843 (make_pass_branch_target_load_optimize1): ...new function.
14844 (pass_branch_target_load_optimize2): Convert from a global struct to a
14845 subclass of rtl_opt_pass along with...
14846 (pass_data_branch_target_load_optimize2): ...new pass_data instance
14848 (make_pass_branch_target_load_optimize2): ...new function.
14849 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
14850 of rtl_opt_pass along with...
14851 (pass_data_jump): ...new pass_data instance and...
14852 (make_pass_jump): ...new function.
14853 (pass_jump2): Convert from a global struct to a subclass of
14854 rtl_opt_pass along with...
14855 (pass_data_jump2): ...new pass_data instance and...
14856 (make_pass_jump2): ...new function.
14857 * cfgexpand.c (pass_expand): Convert from a global struct to a
14858 subclass of rtl_opt_pass along with...
14859 (pass_data_expand): ...new pass_data instance and...
14860 (make_pass_expand): ...new function.
14861 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
14862 of rtl_opt_pass along with...
14863 (pass_data_free_cfg): ...new pass_data instance and...
14864 (make_pass_free_cfg): ...new function.
14865 (pass_into_cfg_layout_mode): Convert from a global struct to a
14866 subclass of rtl_opt_pass along with...
14867 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
14868 (make_pass_into_cfg_layout_mode): ...new function.
14869 (pass_outof_cfg_layout_mode): Convert from a global struct to a
14870 subclass of rtl_opt_pass along with...
14871 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
14872 (make_pass_outof_cfg_layout_mode): ...new function.
14873 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
14874 struct to a subclass of gimple_opt_pass along with...
14875 (pass_data_build_cgraph_edges): ...new pass_data instance and...
14876 (make_pass_build_cgraph_edges): ...new function.
14877 (pass_rebuild_cgraph_edges): Convert from a global struct to a
14878 subclass of gimple_opt_pass along with...
14879 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
14880 (make_pass_rebuild_cgraph_edges): ...new function.
14881 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
14882 subclass of gimple_opt_pass along with...
14883 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
14885 (make_pass_remove_cgraph_callee_edges): ...new function.
14886 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
14887 struct to a subclass of rtl_opt_pass along with...
14888 (pass_data_stack_adjustments): ...new pass_data instance and...
14889 (make_pass_stack_adjustments): ...new function.
14890 * combine.c (pass_combine): Convert from a global struct to a subclass
14891 of rtl_opt_pass along with...
14892 (pass_data_combine): ...new pass_data instance and...
14893 (make_pass_combine): ...new function.
14894 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
14895 global struct to a subclass of rtl_opt_pass along with...
14896 (pass_data_compare_elim_after_reload): ...new pass_data instance
14898 (make_pass_compare_elim_after_reload): ...new function.
14899 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
14900 of rtl_opt_pass along with...
14901 (pass_data_rtl_cprop): ...new pass_data instance and...
14902 (make_pass_rtl_cprop): ...new function.
14903 * cse.c (pass_cse): Convert from a global struct to a subclass of
14904 rtl_opt_pass along with...
14905 (pass_data_cse): ...new pass_data instance and...
14906 (make_pass_cse): ...new function.
14907 (pass_cse2): Convert from a global struct to a subclass of
14908 rtl_opt_pass along with...
14909 (pass_data_cse2): ...new pass_data instance and...
14910 (make_pass_cse2): ...new function.
14911 (pass_cse_after_global_opts): Convert from a global struct to a
14912 subclass of rtl_opt_pass along with...
14913 (pass_data_cse_after_global_opts): ...new pass_data instance and...
14914 (make_pass_cse_after_global_opts): ...new function.
14915 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
14916 of rtl_opt_pass along with...
14917 (pass_data_ud_rtl_dce): ...new pass_data instance and...
14918 (make_pass_ud_rtl_dce): ...new function.
14919 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
14920 rtl_opt_pass along with...
14921 (pass_data_fast_rtl_dce): ...new pass_data instance and...
14922 (make_pass_fast_rtl_dce): ...new function.
14923 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
14924 a subclass of rtl_opt_pass along with...
14925 (pass_data_df_initialize_opt): ...new pass_data instance and...
14926 (make_pass_df_initialize_opt): ...new function.
14927 (pass_df_initialize_no_opt): Convert from a global struct to a
14928 subclass of rtl_opt_pass along with...
14929 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
14930 (make_pass_df_initialize_no_opt): ...new function.
14931 (pass_df_finish): Convert from a global struct to a subclass of
14932 rtl_opt_pass along with...
14933 (pass_data_df_finish): ...new pass_data instance and...
14934 (make_pass_df_finish): ...new function.
14935 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
14936 rtl_opt_pass along with...
14937 (pass_data_rtl_dse1): ...new pass_data instance and...
14938 (make_pass_rtl_dse1): ...new function.
14939 (pass_rtl_dse2): Convert from a global struct to a subclass of
14940 rtl_opt_pass along with...
14941 (pass_data_rtl_dse2): ...new pass_data instance and...
14942 (make_pass_rtl_dse2): ...new function.
14943 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
14944 subclass of rtl_opt_pass along with...
14945 (pass_data_dwarf2_frame): ...new pass_data instance and...
14946 (make_pass_dwarf2_frame): ...new function.
14947 * except.c (pass_set_nothrow_function_flags): Convert from a global
14948 struct to a subclass of rtl_opt_pass along with...
14949 (pass_data_set_nothrow_function_flags): ...new pass_data instance
14951 (make_pass_set_nothrow_function_flags): ...new function.
14952 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
14953 subclass of rtl_opt_pass along with...
14954 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
14956 (make_pass_convert_to_eh_region_ranges): ...new function.
14957 * final.c (pass_compute_alignments): Convert from a global struct to a
14958 subclass of rtl_opt_pass along with...
14959 (pass_data_compute_alignments): ...new pass_data instance and...
14960 (make_pass_compute_alignments): ...new function.
14961 (pass_final): Convert from a global struct to a subclass of
14962 rtl_opt_pass along with...
14963 (pass_data_final): ...new pass_data instance and...
14964 (make_pass_final): ...new function.
14965 (pass_shorten_branches): Convert from a global struct to a subclass of
14966 rtl_opt_pass along with...
14967 (pass_data_shorten_branches): ...new pass_data instance and...
14968 (make_pass_shorten_branches): ...new function.
14969 (pass_clean_state): Convert from a global struct to a subclass of
14970 rtl_opt_pass along with...
14971 (pass_data_clean_state): ...new pass_data instance and...
14972 (make_pass_clean_state): ...new function.
14973 * function.c (pass_instantiate_virtual_regs): Convert from a global
14974 struct to a subclass of rtl_opt_pass along with...
14975 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
14976 (make_pass_instantiate_virtual_regs): ...new function.
14977 (pass_leaf_regs): Convert from a global struct to a subclass of
14978 rtl_opt_pass along with...
14979 (pass_data_leaf_regs): ...new pass_data instance and...
14980 (make_pass_leaf_regs): ...new function.
14981 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
14982 subclass of rtl_opt_pass along with...
14983 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
14985 (make_pass_thread_prologue_and_epilogue): ...new function.
14986 (pass_match_asm_constraints): Convert from a global struct to a
14987 subclass of rtl_opt_pass along with...
14988 (pass_data_match_asm_constraints): ...new pass_data instance and...
14989 (make_pass_match_asm_constraints): ...new function.
14990 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
14991 subclass of rtl_opt_pass along with...
14992 (pass_data_rtl_fwprop): ...new pass_data instance and...
14993 (make_pass_rtl_fwprop): ...new function.
14994 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
14995 rtl_opt_pass along with...
14996 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
14997 (make_pass_rtl_fwprop_addr): ...new function.
14998 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
14999 rtl_opt_pass along with...
15000 (pass_data_rtl_pre): ...new pass_data instance and...
15001 (make_pass_rtl_pre): ...new function.
15002 (pass_rtl_hoist): Convert from a global struct to a subclass of
15003 rtl_opt_pass along with...
15004 (pass_data_rtl_hoist): ...new pass_data instance and...
15005 (make_pass_rtl_hoist): ...new function.
15006 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
15007 subclass of gimple_opt_pass along with...
15008 (pass_data_lower_cf): ...new pass_data instance and...
15009 (make_pass_lower_cf): ...new function.
15010 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
15011 from a global struct to a subclass of gimple_opt_pass along with...
15012 (pass_data_strength_reduction): ...new pass_data instance and...
15013 (make_pass_strength_reduction): ...new function.
15014 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
15015 of rtl_opt_pass along with...
15016 (pass_data_rtl_ifcvt): ...new pass_data instance and...
15017 (make_pass_rtl_ifcvt): ...new function.
15018 (pass_if_after_combine): Convert from a global struct to a subclass of
15019 rtl_opt_pass along with...
15020 (pass_data_if_after_combine): ...new pass_data instance and...
15021 (make_pass_if_after_combine): ...new function.
15022 (pass_if_after_reload): Convert from a global struct to a subclass of
15023 rtl_opt_pass along with...
15024 (pass_data_if_after_reload): ...new pass_data instance and...
15025 (make_pass_if_after_reload): ...new function.
15026 * init-regs.c (pass_initialize_regs): Convert from a global struct to
15027 a subclass of rtl_opt_pass along with...
15028 (pass_data_initialize_regs): ...new pass_data instance and...
15029 (make_pass_initialize_regs): ...new function.
15030 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
15031 of ipa_opt_pass_d along with...
15032 (pass_data_ipa_cp): ...new pass_data instance and...
15033 (make_pass_ipa_cp): ...new function.
15034 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
15035 global struct to a subclass of gimple_opt_pass along with...
15036 (pass_data_inline_parameters): ...new pass_data instance and...
15037 (make_pass_inline_parameters): ...new function.
15038 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
15039 subclass of gimple_opt_pass along with...
15040 (pass_data_early_inline): ...new pass_data instance and...
15041 (make_pass_early_inline): ...new function.
15042 (pass_ipa_inline): Convert from a global struct to a subclass of
15043 ipa_opt_pass_d along with...
15044 (pass_data_ipa_inline): ...new pass_data instance and...
15045 (make_pass_ipa_inline): ...new function.
15046 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
15047 struct to a subclass of gimple_opt_pass along with...
15048 (pass_data_local_pure_const): ...new pass_data instance and...
15049 (make_pass_local_pure_const): ...new function.
15050 (pass_ipa_pure_const): Convert from a global struct to a subclass of
15051 ipa_opt_pass_d along with...
15052 (pass_data_ipa_pure_const): ...new pass_data instance and...
15053 (make_pass_ipa_pure_const): ...new function.
15054 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
15055 to a subclass of ipa_opt_pass_d along with...
15056 (pass_data_ipa_reference): ...new pass_data instance and...
15057 (make_pass_ipa_reference): ...new function.
15058 * ipa-split.c (pass_split_functions): Convert from a global struct to
15059 a subclass of gimple_opt_pass along with...
15060 (pass_data_split_functions): ...new pass_data instance and...
15061 (make_pass_split_functions): ...new function.
15062 (pass_feedback_split_functions): Convert from a global struct to a
15063 subclass of gimple_opt_pass along with...
15064 (pass_data_feedback_split_functions): ...new pass_data instance and...
15065 (make_pass_feedback_split_functions): ...new function.
15066 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
15067 global struct to a subclass of simple_ipa_opt_pass along with...
15068 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
15070 (make_pass_ipa_function_and_variable_visibility): ...new function.
15071 (pass_ipa_free_inline_summary): Convert from a global struct to a
15072 subclass of simple_ipa_opt_pass along with...
15073 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
15074 (make_pass_ipa_free_inline_summary): ...new function.
15075 (pass_ipa_whole_program_visibility): Convert from a global struct to a
15076 subclass of ipa_opt_pass_d along with...
15077 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
15079 (make_pass_ipa_whole_program_visibility): ...new function.
15080 (pass_ipa_profile): Convert from a global struct to a subclass of
15081 ipa_opt_pass_d along with...
15082 (pass_data_ipa_profile): ...new pass_data instance and...
15083 (make_pass_ipa_profile): ...new function.
15084 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
15085 ipa_opt_pass_d along with...
15086 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
15087 (make_pass_ipa_cdtor_merge): ...new function.
15088 * ira.c (pass_ira): Convert from a global struct to a subclass of
15089 rtl_opt_pass along with...
15090 (pass_data_ira): ...new pass_data instance and...
15091 (make_pass_ira): ...new function.
15092 (pass_reload): Convert from a global struct to a subclass of
15093 rtl_opt_pass along with...
15094 (pass_data_reload): ...new pass_data instance and...
15095 (make_pass_reload): ...new function.
15096 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
15097 subclass of rtl_opt_pass along with...
15098 (pass_data_cleanup_barriers): ...new pass_data instance and...
15099 (make_pass_cleanup_barriers): ...new function.
15100 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
15101 of rtl_opt_pass along with...
15102 (pass_data_loop2): ...new pass_data instance and...
15103 (make_pass_loop2): ...new function.
15104 (pass_rtl_loop_init): Convert from a global struct to a subclass of
15105 rtl_opt_pass along with...
15106 (pass_data_rtl_loop_init): ...new pass_data instance and...
15107 (make_pass_rtl_loop_init): ...new function.
15108 (pass_rtl_loop_done): Convert from a global struct to a subclass of
15109 rtl_opt_pass along with...
15110 (pass_data_rtl_loop_done): ...new pass_data instance and...
15111 (make_pass_rtl_loop_done): ...new function.
15112 (pass_rtl_move_loop_invariants): Convert from a global struct to a
15113 subclass of rtl_opt_pass along with...
15114 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
15115 (make_pass_rtl_move_loop_invariants): ...new function.
15116 (pass_rtl_unswitch): Convert from a global struct to a subclass of
15117 rtl_opt_pass along with...
15118 (pass_data_rtl_unswitch): ...new pass_data instance and...
15119 (make_pass_rtl_unswitch): ...new function.
15120 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
15121 subclass of rtl_opt_pass along with...
15122 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
15124 (make_pass_rtl_unroll_and_peel_loops): ...new function.
15125 (pass_rtl_doloop): Convert from a global struct to a subclass of
15126 rtl_opt_pass along with...
15127 (pass_data_rtl_doloop): ...new pass_data instance and...
15128 (make_pass_rtl_doloop): ...new function.
15129 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
15130 a subclass of rtl_opt_pass along with...
15131 (pass_data_lower_subreg): ...new pass_data instance and...
15132 (make_pass_lower_subreg): ...new function.
15133 (pass_lower_subreg2): Convert from a global struct to a subclass of
15134 rtl_opt_pass along with...
15135 (pass_data_lower_subreg2): ...new pass_data instance and...
15136 (make_pass_lower_subreg2): ...new function.
15137 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
15138 struct to a subclass of ipa_opt_pass_d along with...
15139 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
15140 (make_pass_ipa_lto_gimple_out): ...new function.
15141 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
15142 of ipa_opt_pass_d along with...
15143 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
15144 (make_pass_ipa_lto_finish_out): ...new function.
15145 * mode-switching.c (pass_mode_switching): Convert from a global struct
15146 to a subclass of rtl_opt_pass along with...
15147 (pass_data_mode_switching): ...new pass_data instance and...
15148 (make_pass_mode_switching): ...new function.
15149 * modulo-sched.c (pass_sms): Convert from a global struct to a
15150 subclass of rtl_opt_pass along with...
15151 (pass_data_sms): ...new pass_data instance and...
15152 (make_pass_sms): ...new function.
15153 * omp-low.c (pass_expand_omp): Convert from a global struct to a
15154 subclass of gimple_opt_pass along with...
15155 (pass_data_expand_omp): ...new pass_data instance and...
15156 (make_pass_expand_omp): ...new function.
15157 (pass_lower_omp): Convert from a global struct to a subclass of
15158 gimple_opt_pass along with...
15159 (pass_data_lower_omp): ...new pass_data instance and...
15160 (make_pass_lower_omp): ...new function.
15161 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
15162 of gimple_opt_pass along with...
15163 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
15164 (make_pass_diagnose_omp_blocks): ...new function.
15165 * passes.c (pass_early_local_passes): Convert from a global struct to
15166 a subclass of simple_ipa_opt_pass along with...
15167 (pass_data_early_local_passes): ...new pass_data instance and...
15168 (make_pass_early_local_passes): ...new function.
15169 (pass_all_early_optimizations): Convert from a global struct to a
15170 subclass of gimple_opt_pass along with...
15171 (pass_data_all_early_optimizations): ...new pass_data instance and...
15172 (make_pass_all_early_optimizations): ...new function.
15173 (pass_all_optimizations): Convert from a global struct to a subclass
15174 of gimple_opt_pass along with...
15175 (pass_data_all_optimizations): ...new pass_data instance and...
15176 (make_pass_all_optimizations): ...new function.
15177 (pass_all_optimizations_g): Convert from a global struct to a subclass
15178 of gimple_opt_pass along with...
15179 (pass_data_all_optimizations_g): ...new pass_data instance and...
15180 (make_pass_all_optimizations_g): ...new function.
15181 (pass_rest_of_compilation): Convert from a global struct to a subclass
15182 of rtl_opt_pass along with...
15183 (pass_data_rest_of_compilation): ...new pass_data instance and...
15184 (make_pass_rest_of_compilation): ...new function.
15185 (pass_postreload): Convert from a global struct to a subclass of
15186 rtl_opt_pass along with...
15187 (pass_data_postreload): ...new pass_data instance and...
15188 (make_pass_postreload): ...new function.
15189 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
15190 subclass of rtl_opt_pass along with...
15191 (pass_data_gcse2): ...new pass_data instance and...
15192 (make_pass_gcse2): ...new function.
15193 * postreload.c (pass_postreload_cse): Convert from a global struct to
15194 a subclass of rtl_opt_pass along with...
15195 (pass_data_postreload_cse): ...new pass_data instance and...
15196 (make_pass_postreload_cse): ...new function.
15197 * predict.c (pass_profile): Convert from a global struct to a subclass
15198 of gimple_opt_pass along with...
15199 (pass_data_profile): ...new pass_data instance and...
15200 (make_pass_profile): ...new function.
15201 (pass_strip_predict_hints): Convert from a global struct to a subclass
15202 of gimple_opt_pass along with...
15203 (pass_data_strip_predict_hints): ...new pass_data instance and...
15204 (make_pass_strip_predict_hints): ...new function.
15205 * recog.c (pass_peephole2): Convert from a global struct to a subclass
15206 of rtl_opt_pass along with...
15207 (pass_data_peephole2): ...new pass_data instance and...
15208 (make_pass_peephole2): ...new function.
15209 (pass_split_all_insns): Convert from a global struct to a subclass of
15210 rtl_opt_pass along with...
15211 (pass_data_split_all_insns): ...new pass_data instance and...
15212 (make_pass_split_all_insns): ...new function.
15213 (pass_split_after_reload): Convert from a global struct to a subclass
15214 of rtl_opt_pass along with...
15215 (pass_data_split_after_reload): ...new pass_data instance and...
15216 (make_pass_split_after_reload): ...new function.
15217 (pass_split_before_regstack): Convert from a global struct to a
15218 subclass of rtl_opt_pass along with...
15219 (pass_data_split_before_regstack): ...new pass_data instance and...
15220 (make_pass_split_before_regstack): ...new function.
15221 (pass_split_before_sched2): Convert from a global struct to a subclass
15222 of rtl_opt_pass along with...
15223 (pass_data_split_before_sched2): ...new pass_data instance and...
15224 (make_pass_split_before_sched2): ...new function.
15225 (pass_split_for_shorten_branches): Convert from a global struct to a
15226 subclass of rtl_opt_pass along with...
15227 (pass_data_split_for_shorten_branches): ...new pass_data instance
15229 (make_pass_split_for_shorten_branches): ...new function.
15230 * ree.c (pass_ree): Convert from a global struct to a subclass of
15231 rtl_opt_pass along with...
15232 (pass_data_ree): ...new pass_data instance and...
15233 (make_pass_ree): ...new function.
15234 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
15235 subclass of rtl_opt_pass along with...
15236 (pass_data_stack_regs): ...new pass_data instance and...
15237 (make_pass_stack_regs): ...new function.
15238 (pass_stack_regs_run): Convert from a global struct to a subclass of
15239 rtl_opt_pass along with...
15240 (pass_data_stack_regs_run): ...new pass_data instance and...
15241 (make_pass_stack_regs_run): ...new function.
15242 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
15243 subclass of rtl_opt_pass along with...
15244 (pass_data_cprop_hardreg): ...new pass_data instance and...
15245 (make_pass_cprop_hardreg): ...new function.
15246 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
15247 subclass of rtl_opt_pass along with...
15248 (pass_data_reginfo_init): ...new pass_data instance and...
15249 (make_pass_reginfo_init): ...new function.
15250 * regmove.c (pass_regmove): Convert from a global struct to a subclass
15251 of rtl_opt_pass along with...
15252 (pass_data_regmove): ...new pass_data instance and...
15253 (make_pass_regmove): ...new function.
15254 * regrename.c (pass_regrename): Convert from a global struct to a
15255 subclass of rtl_opt_pass along with...
15256 (pass_data_regrename): ...new pass_data instance and...
15257 (make_pass_regrename): ...new function.
15258 * reorg.c (pass_delay_slots): Convert from a global struct to a
15259 subclass of rtl_opt_pass along with...
15260 (pass_data_delay_slots): ...new pass_data instance and...
15261 (make_pass_delay_slots): ...new function.
15262 (pass_machine_reorg): Convert from a global struct to a subclass of
15263 rtl_opt_pass along with...
15264 (pass_data_machine_reorg): ...new pass_data instance and...
15265 (make_pass_machine_reorg): ...new function.
15266 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
15267 of rtl_opt_pass along with...
15268 (pass_data_sched): ...new pass_data instance and...
15269 (make_pass_sched): ...new function.
15270 (pass_sched2): Convert from a global struct to a subclass of
15271 rtl_opt_pass along with...
15272 (pass_data_sched2): ...new pass_data instance and...
15273 (make_pass_sched2): ...new function.
15274 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
15275 to a subclass of rtl_opt_pass along with...
15276 (pass_data_stack_ptr_mod): ...new pass_data instance and...
15277 (make_pass_stack_ptr_mod): ...new function.
15278 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
15279 to a subclass of rtl_opt_pass along with...
15280 (pass_data_rtl_store_motion): ...new pass_data instance and...
15281 (make_pass_rtl_store_motion): ...new function.
15282 * tracer.c (pass_tracer): Convert from a global struct to a subclass
15283 of gimple_opt_pass along with...
15284 (pass_data_tracer): ...new pass_data instance and...
15285 (make_pass_tracer): ...new function.
15286 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
15287 to a subclass of gimple_opt_pass along with...
15288 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
15289 (make_pass_diagnose_tm_blocks): ...new function.
15290 (pass_lower_tm): Convert from a global struct to a subclass of
15291 gimple_opt_pass along with...
15292 (pass_data_lower_tm): ...new pass_data instance and...
15293 (make_pass_lower_tm): ...new function.
15294 (pass_tm_init): Convert from a global struct to a subclass of
15295 gimple_opt_pass along with...
15296 (pass_data_tm_init): ...new pass_data instance and...
15297 (make_pass_tm_init): ...new function.
15298 (pass_tm_mark): Convert from a global struct to a subclass of
15299 gimple_opt_pass along with...
15300 (pass_data_tm_mark): ...new pass_data instance and...
15301 (make_pass_tm_mark): ...new function.
15302 (pass_tm_edges): Convert from a global struct to a subclass of
15303 gimple_opt_pass along with...
15304 (pass_data_tm_edges): ...new pass_data instance and...
15305 (make_pass_tm_edges): ...new function.
15306 (pass_tm_memopt): Convert from a global struct to a subclass of
15307 gimple_opt_pass along with...
15308 (pass_data_tm_memopt): ...new pass_data instance and...
15309 (make_pass_tm_memopt): ...new function.
15310 (pass_ipa_tm): Convert from a global struct to a subclass of
15311 simple_ipa_opt_pass along with...
15312 (pass_data_ipa_tm): ...new pass_data instance and...
15313 (make_pass_ipa_tm): ...new function.
15314 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
15315 subclass of gimple_opt_pass along with...
15316 (pass_data_call_cdce): ...new pass_data instance and...
15317 (make_pass_call_cdce): ...new function.
15318 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
15319 subclass of gimple_opt_pass along with...
15320 (pass_data_build_cfg): ...new pass_data instance and...
15321 (make_pass_build_cfg): ...new function.
15322 (pass_split_crit_edges): Convert from a global struct to a subclass of
15323 gimple_opt_pass along with...
15324 (pass_data_split_crit_edges): ...new pass_data instance and...
15325 (make_pass_split_crit_edges): ...new function.
15326 (pass_warn_function_return): Convert from a global struct to a
15327 subclass of gimple_opt_pass along with...
15328 (pass_data_warn_function_return): ...new pass_data instance and...
15329 (make_pass_warn_function_return): ...new function.
15330 (pass_warn_function_noreturn): Convert from a global struct to a
15331 subclass of gimple_opt_pass along with...
15332 (pass_data_warn_function_noreturn): ...new pass_data instance and...
15333 (make_pass_warn_function_noreturn): ...new function.
15334 (pass_warn_unused_result): Convert from a global struct to a subclass
15335 of gimple_opt_pass along with...
15336 (pass_data_warn_unused_result): ...new pass_data instance and...
15337 (make_pass_warn_unused_result): ...new function.
15338 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
15339 a subclass of gimple_opt_pass along with...
15340 (pass_data_merge_phi): ...new pass_data instance and...
15341 (make_pass_merge_phi): ...new function.
15342 * tree-complex.c (pass_lower_complex): Convert from a global struct to
15343 a subclass of gimple_opt_pass along with...
15344 (pass_data_lower_complex): ...new pass_data instance and...
15345 (make_pass_lower_complex): ...new function.
15346 (pass_lower_complex_O0): Convert from a global struct to a subclass of
15347 gimple_opt_pass along with...
15348 (pass_data_lower_complex_O0): ...new pass_data instance and...
15349 (make_pass_lower_complex_O0): ...new function.
15350 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
15351 subclass of gimple_opt_pass along with...
15352 (pass_data_lower_eh): ...new pass_data instance and...
15353 (make_pass_lower_eh): ...new function.
15354 (pass_refactor_eh): Convert from a global struct to a subclass of
15355 gimple_opt_pass along with...
15356 (pass_data_refactor_eh): ...new pass_data instance and...
15357 (make_pass_refactor_eh): ...new function.
15358 (pass_lower_resx): Convert from a global struct to a subclass of
15359 gimple_opt_pass along with...
15360 (pass_data_lower_resx): ...new pass_data instance and...
15361 (make_pass_lower_resx): ...new function.
15362 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
15363 of gimple_opt_pass along with...
15364 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
15365 (make_pass_lower_eh_dispatch): ...new function.
15366 (pass_cleanup_eh): Convert from a global struct to a subclass of
15367 gimple_opt_pass along with...
15368 (pass_data_cleanup_eh): ...new pass_data instance and...
15369 (make_pass_cleanup_eh): ...new function.
15370 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
15371 to a subclass of simple_ipa_opt_pass along with...
15372 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
15373 (make_pass_ipa_lower_emutls): ...new function.
15374 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
15375 a subclass of gimple_opt_pass along with...
15376 (pass_data_if_conversion): ...new pass_data instance and...
15377 (make_pass_if_conversion): ...new function.
15378 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
15379 subclass of gimple_opt_pass along with...
15380 (pass_data_build_ssa): ...new pass_data instance and...
15381 (make_pass_build_ssa): ...new function.
15382 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
15383 global struct to a subclass of gimple_opt_pass along with...
15384 (pass_data_loop_distribution): ...new pass_data instance and...
15385 (make_pass_loop_distribution): ...new function.
15386 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
15387 subclass of gimple_opt_pass along with...
15388 (pass_data_mudflap_1): ...new pass_data instance and...
15389 (make_pass_mudflap_1): ...new function.
15390 (pass_mudflap_2): Convert from a global struct to a subclass of
15391 gimple_opt_pass along with...
15392 (pass_data_mudflap_2): ...new pass_data instance and...
15393 (make_pass_mudflap_2): ...new function.
15394 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
15395 subclass of gimple_opt_pass along with...
15396 (pass_data_mudflap_1): ...new pass_data instance and...
15397 (make_pass_mudflap_1): ...new function.
15398 (pass_mudflap_2): Convert from a global struct to a subclass of
15399 gimple_opt_pass along with...
15400 (pass_data_mudflap_2): ...new pass_data instance and...
15401 (make_pass_mudflap_2): ...new function.
15402 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
15403 gimple_opt_pass along with...
15404 (pass_data_nrv): ...new pass_data instance and...
15405 (make_pass_nrv): ...new function.
15406 (pass_return_slot): Convert from a global struct to a subclass of
15407 gimple_opt_pass along with...
15408 (pass_data_return_slot): ...new pass_data instance and...
15409 (make_pass_return_slot): ...new function.
15410 * tree-object-size.c (pass_object_sizes): Convert from a global struct
15411 to a subclass of gimple_opt_pass along with...
15412 (pass_data_object_sizes): ...new pass_data instance and...
15413 (make_pass_object_sizes): ...new function.
15414 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
15415 global struct to a subclass of gimple_opt_pass along with...
15416 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
15418 (make_pass_cleanup_cfg_post_optimizing): ...new function.
15419 (pass_fixup_cfg): Convert from a global struct to a subclass of
15420 gimple_opt_pass along with...
15421 (pass_data_fixup_cfg): ...new pass_data instance and...
15422 (make_pass_fixup_cfg): ...new function.
15423 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
15424 (make_pass_mudflap_1): ...new function.
15425 (pass_mudflap_2): Replace declaration with that of...
15426 (make_pass_mudflap_2): ...new function.
15427 (pass_asan): Replace declaration with that of...
15428 (make_pass_asan): ...new function.
15429 (pass_asan_O0): Replace declaration with that of...
15430 (make_pass_asan_O0): ...new function.
15431 (pass_tsan): Replace declaration with that of...
15432 (make_pass_tsan): ...new function.
15433 (pass_tsan_O0): Replace declaration with that of...
15434 (make_pass_tsan_O0): ...new function.
15435 (pass_lower_cf): Replace declaration with that of...
15436 (make_pass_lower_cf): ...new function.
15437 (pass_refactor_eh): Replace declaration with that of...
15438 (make_pass_refactor_eh): ...new function.
15439 (pass_lower_eh): Replace declaration with that of...
15440 (make_pass_lower_eh): ...new function.
15441 (pass_lower_eh_dispatch): Replace declaration with that of...
15442 (make_pass_lower_eh_dispatch): ...new function.
15443 (pass_lower_resx): Replace declaration with that of...
15444 (make_pass_lower_resx): ...new function.
15445 (pass_build_cfg): Replace declaration with that of...
15446 (make_pass_build_cfg): ...new function.
15447 (pass_early_tree_profile): Replace declaration with that of...
15448 (make_pass_early_tree_profile): ...new function.
15449 (pass_cleanup_eh): Replace declaration with that of...
15450 (make_pass_cleanup_eh): ...new function.
15451 (pass_sra): Replace declaration with that of...
15452 (make_pass_sra): ...new function.
15453 (pass_sra_early): Replace declaration with that of...
15454 (make_pass_sra_early): ...new function.
15455 (pass_early_ipa_sra): Replace declaration with that of...
15456 (make_pass_early_ipa_sra): ...new function.
15457 (pass_tail_recursion): Replace declaration with that of...
15458 (make_pass_tail_recursion): ...new function.
15459 (pass_tail_calls): Replace declaration with that of...
15460 (make_pass_tail_calls): ...new function.
15461 (pass_tree_loop): Replace declaration with that of...
15462 (make_pass_tree_loop): ...new function.
15463 (pass_tree_loop_init): Replace declaration with that of...
15464 (make_pass_tree_loop_init): ...new function.
15465 (pass_lim): Replace declaration with that of...
15466 (make_pass_lim): ...new function.
15467 (pass_tree_unswitch): Replace declaration with that of...
15468 (make_pass_tree_unswitch): ...new function.
15469 (pass_predcom): Replace declaration with that of...
15470 (make_pass_predcom): ...new function.
15471 (pass_iv_canon): Replace declaration with that of...
15472 (make_pass_iv_canon): ...new function.
15473 (pass_scev_cprop): Replace declaration with that of...
15474 (make_pass_scev_cprop): ...new function.
15475 (pass_empty_loop): Replace declaration with that of...
15476 (make_pass_empty_loop): ...new function.
15477 (pass_record_bounds): Replace declaration with that of...
15478 (make_pass_record_bounds): ...new function.
15479 (pass_graphite): Replace declaration with that of...
15480 (make_pass_graphite): ...new function.
15481 (pass_graphite_transforms): Replace declaration with that of...
15482 (make_pass_graphite_transforms): ...new function.
15483 (pass_if_conversion): Replace declaration with that of...
15484 (make_pass_if_conversion): ...new function.
15485 (pass_loop_distribution): Replace declaration with that of...
15486 (make_pass_loop_distribution): ...new function.
15487 (pass_vectorize): Replace declaration with that of...
15488 (make_pass_vectorize): ...new function.
15489 (pass_slp_vectorize): Replace declaration with that of...
15490 (make_pass_slp_vectorize): ...new function.
15491 (pass_complete_unroll): Replace declaration with that of...
15492 (make_pass_complete_unroll): ...new function.
15493 (pass_complete_unrolli): Replace declaration with that of...
15494 (make_pass_complete_unrolli): ...new function.
15495 (pass_parallelize_loops): Replace declaration with that of...
15496 (make_pass_parallelize_loops): ...new function.
15497 (pass_loop_prefetch): Replace declaration with that of...
15498 (make_pass_loop_prefetch): ...new function.
15499 (pass_iv_optimize): Replace declaration with that of...
15500 (make_pass_iv_optimize): ...new function.
15501 (pass_tree_loop_done): Replace declaration with that of...
15502 (make_pass_tree_loop_done): ...new function.
15503 (pass_ch): Replace declaration with that of...
15504 (make_pass_ch): ...new function.
15505 (pass_ccp): Replace declaration with that of...
15506 (make_pass_ccp): ...new function.
15507 (pass_phi_only_cprop): Replace declaration with that of...
15508 (make_pass_phi_only_cprop): ...new function.
15509 (pass_build_ssa): Replace declaration with that of...
15510 (make_pass_build_ssa): ...new function.
15511 (pass_build_alias): Replace declaration with that of...
15512 (make_pass_build_alias): ...new function.
15513 (pass_build_ealias): Replace declaration with that of...
15514 (make_pass_build_ealias): ...new function.
15515 (pass_dominator): Replace declaration with that of...
15516 (make_pass_dominator): ...new function.
15517 (pass_dce): Replace declaration with that of...
15518 (make_pass_dce): ...new function.
15519 (pass_dce_loop): Replace declaration with that of...
15520 (make_pass_dce_loop): ...new function.
15521 (pass_cd_dce): Replace declaration with that of...
15522 (make_pass_cd_dce): ...new function.
15523 (pass_call_cdce): Replace declaration with that of...
15524 (make_pass_call_cdce): ...new function.
15525 (pass_merge_phi): Replace declaration with that of...
15526 (make_pass_merge_phi): ...new function.
15527 (pass_split_crit_edges): Replace declaration with that of...
15528 (make_pass_split_crit_edges): ...new function.
15529 (pass_pre): Replace declaration with that of...
15530 (make_pass_pre): ...new function.
15531 (pass_profile): Replace declaration with that of...
15532 (make_pass_profile): ...new function.
15533 (pass_strip_predict_hints): Replace declaration with that of...
15534 (make_pass_strip_predict_hints): ...new function.
15535 (pass_lower_complex_O0): Replace declaration with that of...
15536 (make_pass_lower_complex_O0): ...new function.
15537 (pass_lower_complex): Replace declaration with that of...
15538 (make_pass_lower_complex): ...new function.
15539 (pass_lower_vector): Replace declaration with that of...
15540 (make_pass_lower_vector): ...new function.
15541 (pass_lower_vector_ssa): Replace declaration with that of...
15542 (make_pass_lower_vector_ssa): ...new function.
15543 (pass_lower_omp): Replace declaration with that of...
15544 (make_pass_lower_omp): ...new function.
15545 (pass_diagnose_omp_blocks): Replace declaration with that of...
15546 (make_pass_diagnose_omp_blocks): ...new function.
15547 (pass_expand_omp): Replace declaration with that of...
15548 (make_pass_expand_omp): ...new function.
15549 (pass_expand_omp_ssa): Replace declaration with that of...
15550 (make_pass_expand_omp_ssa): ...new function.
15551 (pass_object_sizes): Replace declaration with that of...
15552 (make_pass_object_sizes): ...new function.
15553 (pass_strlen): Replace declaration with that of...
15554 (make_pass_strlen): ...new function.
15555 (pass_fold_builtins): Replace declaration with that of...
15556 (make_pass_fold_builtins): ...new function.
15557 (pass_stdarg): Replace declaration with that of...
15558 (make_pass_stdarg): ...new function.
15559 (pass_early_warn_uninitialized): Replace declaration with that of...
15560 (make_pass_early_warn_uninitialized): ...new function.
15561 (pass_late_warn_uninitialized): Replace declaration with that of...
15562 (make_pass_late_warn_uninitialized): ...new function.
15563 (pass_cse_reciprocals): Replace declaration with that of...
15564 (make_pass_cse_reciprocals): ...new function.
15565 (pass_cse_sincos): Replace declaration with that of...
15566 (make_pass_cse_sincos): ...new function.
15567 (pass_optimize_bswap): Replace declaration with that of...
15568 (make_pass_optimize_bswap): ...new function.
15569 (pass_optimize_widening_mul): Replace declaration with that of...
15570 (make_pass_optimize_widening_mul): ...new function.
15571 (pass_warn_function_return): Replace declaration with that of...
15572 (make_pass_warn_function_return): ...new function.
15573 (pass_warn_function_noreturn): Replace declaration with that of...
15574 (make_pass_warn_function_noreturn): ...new function.
15575 (pass_cselim): Replace declaration with that of...
15576 (make_pass_cselim): ...new function.
15577 (pass_phiopt): Replace declaration with that of...
15578 (make_pass_phiopt): ...new function.
15579 (pass_forwprop): Replace declaration with that of...
15580 (make_pass_forwprop): ...new function.
15581 (pass_phiprop): Replace declaration with that of...
15582 (make_pass_phiprop): ...new function.
15583 (pass_tree_ifcombine): Replace declaration with that of...
15584 (make_pass_tree_ifcombine): ...new function.
15585 (pass_dse): Replace declaration with that of...
15586 (make_pass_dse): ...new function.
15587 (pass_nrv): Replace declaration with that of...
15588 (make_pass_nrv): ...new function.
15589 (pass_rename_ssa_copies): Replace declaration with that of...
15590 (make_pass_rename_ssa_copies): ...new function.
15591 (pass_sink_code): Replace declaration with that of...
15592 (make_pass_sink_code): ...new function.
15593 (pass_fre): Replace declaration with that of...
15594 (make_pass_fre): ...new function.
15595 (pass_check_data_deps): Replace declaration with that of...
15596 (make_pass_check_data_deps): ...new function.
15597 (pass_copy_prop): Replace declaration with that of...
15598 (make_pass_copy_prop): ...new function.
15599 (pass_vrp): Replace declaration with that of...
15600 (make_pass_vrp): ...new function.
15601 (pass_uncprop): Replace declaration with that of...
15602 (make_pass_uncprop): ...new function.
15603 (pass_return_slot): Replace declaration with that of...
15604 (make_pass_return_slot): ...new function.
15605 (pass_reassoc): Replace declaration with that of...
15606 (make_pass_reassoc): ...new function.
15607 (pass_rebuild_cgraph_edges): Replace declaration with that of...
15608 (make_pass_rebuild_cgraph_edges): ...new function.
15609 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
15610 (make_pass_remove_cgraph_callee_edges): ...new function.
15611 (pass_build_cgraph_edges): Replace declaration with that of...
15612 (make_pass_build_cgraph_edges): ...new function.
15613 (pass_local_pure_const): Replace declaration with that of...
15614 (make_pass_local_pure_const): ...new function.
15615 (pass_tracer): Replace declaration with that of...
15616 (make_pass_tracer): ...new function.
15617 (pass_warn_unused_result): Replace declaration with that of...
15618 (make_pass_warn_unused_result): ...new function.
15619 (pass_diagnose_tm_blocks): Replace declaration with that of...
15620 (make_pass_diagnose_tm_blocks): ...new function.
15621 (pass_lower_tm): Replace declaration with that of...
15622 (make_pass_lower_tm): ...new function.
15623 (pass_tm_init): Replace declaration with that of...
15624 (make_pass_tm_init): ...new function.
15625 (pass_tm_mark): Replace declaration with that of...
15626 (make_pass_tm_mark): ...new function.
15627 (pass_tm_memopt): Replace declaration with that of...
15628 (make_pass_tm_memopt): ...new function.
15629 (pass_tm_edges): Replace declaration with that of...
15630 (make_pass_tm_edges): ...new function.
15631 (pass_split_functions): Replace declaration with that of...
15632 (make_pass_split_functions): ...new function.
15633 (pass_feedback_split_functions): Replace declaration with that of...
15634 (make_pass_feedback_split_functions): ...new function.
15635 (pass_strength_reduction): Replace declaration with that of...
15636 (make_pass_strength_reduction): ...new function.
15637 (pass_ipa_lower_emutls): Replace declaration with that of...
15638 (make_pass_ipa_lower_emutls): ...new function.
15639 (pass_ipa_function_and_variable_visibility): Replace declaration with
15641 (make_pass_ipa_function_and_variable_visibility): ...new function.
15642 (pass_ipa_tree_profile): Replace declaration with that of...
15643 (make_pass_ipa_tree_profile): ...new function.
15644 (pass_early_local_passes): Replace declaration with that of...
15645 (make_pass_early_local_passes): ...new function.
15646 (pass_ipa_whole_program_visibility): Replace declaration with that
15648 (make_pass_ipa_whole_program_visibility): ...new function.
15649 (pass_ipa_lto_gimple_out): Replace declaration with that of...
15650 (make_pass_ipa_lto_gimple_out): ...new function.
15651 (pass_ipa_increase_alignment): Replace declaration with that of...
15652 (make_pass_ipa_increase_alignment): ...new function.
15653 (pass_ipa_inline): Replace declaration with that of...
15654 (make_pass_ipa_inline): ...new function.
15655 (pass_ipa_free_lang_data): Replace declaration with that of...
15656 (make_pass_ipa_free_lang_data): ...new function.
15657 (pass_ipa_free_inline_summary): Replace declaration with that of...
15658 (make_pass_ipa_free_inline_summary): ...new function.
15659 (pass_ipa_cp): Replace declaration with that of...
15660 (make_pass_ipa_cp): ...new function.
15661 (pass_ipa_reference): Replace declaration with that of...
15662 (make_pass_ipa_reference): ...new function.
15663 (pass_ipa_pure_const): Replace declaration with that of...
15664 (make_pass_ipa_pure_const): ...new function.
15665 (pass_ipa_pta): Replace declaration with that of...
15666 (make_pass_ipa_pta): ...new function.
15667 (pass_ipa_lto_finish_out): Replace declaration with that of...
15668 (make_pass_ipa_lto_finish_out): ...new function.
15669 (pass_ipa_tm): Replace declaration with that of...
15670 (make_pass_ipa_tm): ...new function.
15671 (pass_ipa_profile): Replace declaration with that of...
15672 (make_pass_ipa_profile): ...new function.
15673 (pass_ipa_cdtor_merge): Replace declaration with that of...
15674 (make_pass_ipa_cdtor_merge): ...new function.
15675 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
15677 (make_pass_cleanup_cfg_post_optimizing): ...new function.
15678 (pass_init_datastructures): Replace declaration with that of...
15679 (make_pass_init_datastructures): ...new function.
15680 (pass_fixup_cfg): Replace declaration with that of...
15681 (make_pass_fixup_cfg): ...new function.
15682 (pass_expand): Replace declaration with that of...
15683 (make_pass_expand): ...new function.
15684 (pass_instantiate_virtual_regs): Replace declaration with that of...
15685 (make_pass_instantiate_virtual_regs): ...new function.
15686 (pass_rtl_fwprop): Replace declaration with that of...
15687 (make_pass_rtl_fwprop): ...new function.
15688 (pass_rtl_fwprop_addr): Replace declaration with that of...
15689 (make_pass_rtl_fwprop_addr): ...new function.
15690 (pass_jump): Replace declaration with that of...
15691 (make_pass_jump): ...new function.
15692 (pass_jump2): Replace declaration with that of...
15693 (make_pass_jump2): ...new function.
15694 (pass_lower_subreg): Replace declaration with that of...
15695 (make_pass_lower_subreg): ...new function.
15696 (pass_cse): Replace declaration with that of...
15697 (make_pass_cse): ...new function.
15698 (pass_fast_rtl_dce): Replace declaration with that of...
15699 (make_pass_fast_rtl_dce): ...new function.
15700 (pass_ud_rtl_dce): Replace declaration with that of...
15701 (make_pass_ud_rtl_dce): ...new function.
15702 (pass_rtl_dce): Replace declaration with that of...
15703 (make_pass_rtl_dce): ...new function.
15704 (pass_rtl_dse1): Replace declaration with that of...
15705 (make_pass_rtl_dse1): ...new function.
15706 (pass_rtl_dse2): Replace declaration with that of...
15707 (make_pass_rtl_dse2): ...new function.
15708 (pass_rtl_dse3): Replace declaration with that of...
15709 (make_pass_rtl_dse3): ...new function.
15710 (pass_rtl_cprop): Replace declaration with that of...
15711 (make_pass_rtl_cprop): ...new function.
15712 (pass_rtl_pre): Replace declaration with that of...
15713 (make_pass_rtl_pre): ...new function.
15714 (pass_rtl_hoist): Replace declaration with that of...
15715 (make_pass_rtl_hoist): ...new function.
15716 (pass_rtl_store_motion): Replace declaration with that of...
15717 (make_pass_rtl_store_motion): ...new function.
15718 (pass_cse_after_global_opts): Replace declaration with that of...
15719 (make_pass_cse_after_global_opts): ...new function.
15720 (pass_rtl_ifcvt): Replace declaration with that of...
15721 (make_pass_rtl_ifcvt): ...new function.
15722 (pass_into_cfg_layout_mode): Replace declaration with that of...
15723 (make_pass_into_cfg_layout_mode): ...new function.
15724 (pass_outof_cfg_layout_mode): Replace declaration with that of...
15725 (make_pass_outof_cfg_layout_mode): ...new function.
15726 (pass_loop2): Replace declaration with that of...
15727 (make_pass_loop2): ...new function.
15728 (pass_rtl_loop_init): Replace declaration with that of...
15729 (make_pass_rtl_loop_init): ...new function.
15730 (pass_rtl_move_loop_invariants): Replace declaration with that of...
15731 (make_pass_rtl_move_loop_invariants): ...new function.
15732 (pass_rtl_unswitch): Replace declaration with that of...
15733 (make_pass_rtl_unswitch): ...new function.
15734 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
15735 (make_pass_rtl_unroll_and_peel_loops): ...new function.
15736 (pass_rtl_doloop): Replace declaration with that of...
15737 (make_pass_rtl_doloop): ...new function.
15738 (pass_rtl_loop_done): Replace declaration with that of...
15739 (make_pass_rtl_loop_done): ...new function.
15740 (pass_web): Replace declaration with that of...
15741 (make_pass_web): ...new function.
15742 (pass_cse2): Replace declaration with that of...
15743 (make_pass_cse2): ...new function.
15744 (pass_df_initialize_opt): Replace declaration with that of...
15745 (make_pass_df_initialize_opt): ...new function.
15746 (pass_df_initialize_no_opt): Replace declaration with that of...
15747 (make_pass_df_initialize_no_opt): ...new function.
15748 (pass_reginfo_init): Replace declaration with that of...
15749 (make_pass_reginfo_init): ...new function.
15750 (pass_inc_dec): Replace declaration with that of...
15751 (make_pass_inc_dec): ...new function.
15752 (pass_stack_ptr_mod): Replace declaration with that of...
15753 (make_pass_stack_ptr_mod): ...new function.
15754 (pass_initialize_regs): Replace declaration with that of...
15755 (make_pass_initialize_regs): ...new function.
15756 (pass_combine): Replace declaration with that of...
15757 (make_pass_combine): ...new function.
15758 (pass_if_after_combine): Replace declaration with that of...
15759 (make_pass_if_after_combine): ...new function.
15760 (pass_ree): Replace declaration with that of...
15761 (make_pass_ree): ...new function.
15762 (pass_partition_blocks): Replace declaration with that of...
15763 (make_pass_partition_blocks): ...new function.
15764 (pass_match_asm_constraints): Replace declaration with that of...
15765 (make_pass_match_asm_constraints): ...new function.
15766 (pass_regmove): Replace declaration with that of...
15767 (make_pass_regmove): ...new function.
15768 (pass_split_all_insns): Replace declaration with that of...
15769 (make_pass_split_all_insns): ...new function.
15770 (pass_fast_rtl_byte_dce): Replace declaration with that of...
15771 (make_pass_fast_rtl_byte_dce): ...new function.
15772 (pass_lower_subreg2): Replace declaration with that of...
15773 (make_pass_lower_subreg2): ...new function.
15774 (pass_mode_switching): Replace declaration with that of...
15775 (make_pass_mode_switching): ...new function.
15776 (pass_sms): Replace declaration with that of...
15777 (make_pass_sms): ...new function.
15778 (pass_sched): Replace declaration with that of...
15779 (make_pass_sched): ...new function.
15780 (pass_ira): Replace declaration with that of...
15781 (make_pass_ira): ...new function.
15782 (pass_reload): Replace declaration with that of...
15783 (make_pass_reload): ...new function.
15784 (pass_clean_state): Replace declaration with that of...
15785 (make_pass_clean_state): ...new function.
15786 (pass_branch_prob): Replace declaration with that of...
15787 (make_pass_branch_prob): ...new function.
15788 (pass_value_profile_transformations): Replace declaration with that
15790 (make_pass_value_profile_transformations): ...new function.
15791 (pass_postreload_cse): Replace declaration with that of...
15792 (make_pass_postreload_cse): ...new function.
15793 (pass_gcse2): Replace declaration with that of...
15794 (make_pass_gcse2): ...new function.
15795 (pass_split_after_reload): Replace declaration with that of...
15796 (make_pass_split_after_reload): ...new function.
15797 (pass_branch_target_load_optimize1): Replace declaration with that
15799 (make_pass_branch_target_load_optimize1): ...new function.
15800 (pass_thread_prologue_and_epilogue): Replace declaration with that
15802 (make_pass_thread_prologue_and_epilogue): ...new function.
15803 (pass_stack_adjustments): Replace declaration with that of...
15804 (make_pass_stack_adjustments): ...new function.
15805 (pass_peephole2): Replace declaration with that of...
15806 (make_pass_peephole2): ...new function.
15807 (pass_if_after_reload): Replace declaration with that of...
15808 (make_pass_if_after_reload): ...new function.
15809 (pass_regrename): Replace declaration with that of...
15810 (make_pass_regrename): ...new function.
15811 (pass_cprop_hardreg): Replace declaration with that of...
15812 (make_pass_cprop_hardreg): ...new function.
15813 (pass_reorder_blocks): Replace declaration with that of...
15814 (make_pass_reorder_blocks): ...new function.
15815 (pass_branch_target_load_optimize2): Replace declaration with that
15817 (make_pass_branch_target_load_optimize2): ...new function.
15818 (pass_leaf_regs): Replace declaration with that of...
15819 (make_pass_leaf_regs): ...new function.
15820 (pass_split_before_sched2): Replace declaration with that of...
15821 (make_pass_split_before_sched2): ...new function.
15822 (pass_compare_elim_after_reload): Replace declaration with that of...
15823 (make_pass_compare_elim_after_reload): ...new function.
15824 (pass_sched2): Replace declaration with that of...
15825 (make_pass_sched2): ...new function.
15826 (pass_stack_regs): Replace declaration with that of...
15827 (make_pass_stack_regs): ...new function.
15828 (pass_stack_regs_run): Replace declaration with that of...
15829 (make_pass_stack_regs_run): ...new function.
15830 (pass_df_finish): Replace declaration with that of...
15831 (make_pass_df_finish): ...new function.
15832 (pass_compute_alignments): Replace declaration with that of...
15833 (make_pass_compute_alignments): ...new function.
15834 (pass_duplicate_computed_gotos): Replace declaration with that of...
15835 (make_pass_duplicate_computed_gotos): ...new function.
15836 (pass_variable_tracking): Replace declaration with that of...
15837 (make_pass_variable_tracking): ...new function.
15838 (pass_free_cfg): Replace declaration with that of...
15839 (make_pass_free_cfg): ...new function.
15840 (pass_machine_reorg): Replace declaration with that of...
15841 (make_pass_machine_reorg): ...new function.
15842 (pass_cleanup_barriers): Replace declaration with that of...
15843 (make_pass_cleanup_barriers): ...new function.
15844 (pass_delay_slots): Replace declaration with that of...
15845 (make_pass_delay_slots): ...new function.
15846 (pass_split_for_shorten_branches): Replace declaration with that of...
15847 (make_pass_split_for_shorten_branches): ...new function.
15848 (pass_split_before_regstack): Replace declaration with that of...
15849 (make_pass_split_before_regstack): ...new function.
15850 (pass_convert_to_eh_region_ranges): Replace declaration with that
15852 (make_pass_convert_to_eh_region_ranges): ...new function.
15853 (pass_shorten_branches): Replace declaration with that of...
15854 (make_pass_shorten_branches): ...new function.
15855 (pass_set_nothrow_function_flags): Replace declaration with that of...
15856 (make_pass_set_nothrow_function_flags): ...new function.
15857 (pass_dwarf2_frame): Replace declaration with that of...
15858 (make_pass_dwarf2_frame): ...new function.
15859 (pass_final): Replace declaration with that of...
15860 (make_pass_final): ...new function.
15861 (pass_rtl_seqabstr): Replace declaration with that of...
15862 (make_pass_rtl_seqabstr): ...new function.
15863 (pass_release_ssa_names): Replace declaration with that of...
15864 (make_pass_release_ssa_names): ...new function.
15865 (pass_early_inline): Replace declaration with that of...
15866 (make_pass_early_inline): ...new function.
15867 (pass_inline_parameters): Replace declaration with that of...
15868 (make_pass_inline_parameters): ...new function.
15869 (pass_update_address_taken): Replace declaration with that of...
15870 (make_pass_update_address_taken): ...new function.
15871 (pass_convert_switch): Replace declaration with that of...
15872 (make_pass_convert_switch): ...new function.
15873 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
15874 to a subclass of simple_ipa_opt_pass along with...
15875 (pass_data_ipa_tree_profile): ...new pass_data instance and...
15876 (make_pass_ipa_tree_profile): ...new function.
15877 * tree-sra.c (pass_sra_early): Convert from a global struct to a
15878 subclass of gimple_opt_pass along with...
15879 (pass_data_sra_early): ...new pass_data instance and...
15880 (make_pass_sra_early): ...new function.
15881 (pass_sra): Convert from a global struct to a subclass of
15882 gimple_opt_pass along with...
15883 (pass_data_sra): ...new pass_data instance and...
15884 (make_pass_sra): ...new function.
15885 (pass_early_ipa_sra): Convert from a global struct to a subclass of
15886 gimple_opt_pass along with...
15887 (pass_data_early_ipa_sra): ...new pass_data instance and...
15888 (make_pass_early_ipa_sra): ...new function.
15889 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
15890 subclass of gimple_opt_pass along with...
15891 (pass_data_ccp): ...new pass_data instance and...
15892 (make_pass_ccp): ...new function.
15893 (pass_fold_builtins): Convert from a global struct to a subclass of
15894 gimple_opt_pass along with...
15895 (pass_data_fold_builtins): ...new pass_data instance and...
15896 (make_pass_fold_builtins): ...new function.
15897 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
15898 subclass of gimple_opt_pass along with...
15899 (pass_data_copy_prop): ...new pass_data instance and...
15900 (make_pass_copy_prop): ...new function.
15901 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
15902 global struct to a subclass of gimple_opt_pass along with...
15903 (pass_data_rename_ssa_copies): ...new pass_data instance and...
15904 (make_pass_rename_ssa_copies): ...new function.
15905 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
15906 subclass of gimple_opt_pass along with...
15907 (pass_data_dce): ...new pass_data instance and...
15908 (make_pass_dce): ...new function.
15909 (pass_dce_loop): Convert from a global struct to a subclass of
15910 gimple_opt_pass along with...
15911 (pass_data_dce_loop): ...new pass_data instance and...
15912 (make_pass_dce_loop): ...new function.
15913 (pass_cd_dce): Convert from a global struct to a subclass of
15914 gimple_opt_pass along with...
15915 (pass_data_cd_dce): ...new pass_data instance and...
15916 (make_pass_cd_dce): ...new function.
15917 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
15918 subclass of gimple_opt_pass along with...
15919 (pass_data_dominator): ...new pass_data instance and...
15920 (make_pass_dominator): ...new function.
15921 (pass_phi_only_cprop): Convert from a global struct to a subclass of
15922 gimple_opt_pass along with...
15923 (pass_data_phi_only_cprop): ...new pass_data instance and...
15924 (make_pass_phi_only_cprop): ...new function.
15925 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
15926 subclass of gimple_opt_pass along with...
15927 (pass_data_dse): ...new pass_data instance and...
15928 (make_pass_dse): ...new function.
15929 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
15930 a subclass of gimple_opt_pass along with...
15931 (pass_data_forwprop): ...new pass_data instance and...
15932 (make_pass_forwprop): ...new function.
15933 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
15934 struct to a subclass of gimple_opt_pass along with...
15935 (pass_data_tree_ifcombine): ...new pass_data instance and...
15936 (make_pass_tree_ifcombine): ...new function.
15937 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
15938 subclass of gimple_opt_pass along with...
15939 (pass_data_ch): ...new pass_data instance and...
15940 (make_pass_ch): ...new function.
15941 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
15942 subclass of gimple_opt_pass along with...
15943 (pass_data_tree_loop): ...new pass_data instance and...
15944 (make_pass_tree_loop): ...new function.
15945 (pass_tree_loop_init): Convert from a global struct to a subclass of
15946 gimple_opt_pass along with...
15947 (pass_data_tree_loop_init): ...new pass_data instance and...
15948 (make_pass_tree_loop_init): ...new function.
15949 (pass_lim): Convert from a global struct to a subclass of
15950 gimple_opt_pass along with...
15951 (pass_data_lim): ...new pass_data instance and...
15952 (make_pass_lim): ...new function.
15953 (pass_tree_unswitch): Convert from a global struct to a subclass of
15954 gimple_opt_pass along with...
15955 (pass_data_tree_unswitch): ...new pass_data instance and...
15956 (make_pass_tree_unswitch): ...new function.
15957 (pass_predcom): Convert from a global struct to a subclass of
15958 gimple_opt_pass along with...
15959 (pass_data_predcom): ...new pass_data instance and...
15960 (make_pass_predcom): ...new function.
15961 (pass_vectorize): Convert from a global struct to a subclass of
15962 gimple_opt_pass along with...
15963 (pass_data_vectorize): ...new pass_data instance and...
15964 (make_pass_vectorize): ...new function.
15965 (pass_graphite): Convert from a global struct to a subclass of
15966 gimple_opt_pass along with...
15967 (pass_data_graphite): ...new pass_data instance and...
15968 (make_pass_graphite): ...new function.
15969 (pass_graphite_transforms): Convert from a global struct to a subclass
15970 of gimple_opt_pass along with...
15971 (pass_data_graphite_transforms): ...new pass_data instance and...
15972 (make_pass_graphite_transforms): ...new function.
15973 (pass_check_data_deps): Convert from a global struct to a subclass of
15974 gimple_opt_pass along with...
15975 (pass_data_check_data_deps): ...new pass_data instance and...
15976 (make_pass_check_data_deps): ...new function.
15977 (pass_iv_canon): Convert from a global struct to a subclass of
15978 gimple_opt_pass along with...
15979 (pass_data_iv_canon): ...new pass_data instance and...
15980 (make_pass_iv_canon): ...new function.
15981 (pass_scev_cprop): Convert from a global struct to a subclass of
15982 gimple_opt_pass along with...
15983 (pass_data_scev_cprop): ...new pass_data instance and...
15984 (make_pass_scev_cprop): ...new function.
15985 (pass_record_bounds): Convert from a global struct to a subclass of
15986 gimple_opt_pass along with...
15987 (pass_data_record_bounds): ...new pass_data instance and...
15988 (make_pass_record_bounds): ...new function.
15989 (pass_complete_unroll): Convert from a global struct to a subclass of
15990 gimple_opt_pass along with...
15991 (pass_data_complete_unroll): ...new pass_data instance and...
15992 (make_pass_complete_unroll): ...new function.
15993 (pass_complete_unrolli): Convert from a global struct to a subclass of
15994 gimple_opt_pass along with...
15995 (pass_data_complete_unrolli): ...new pass_data instance and...
15996 (make_pass_complete_unrolli): ...new function.
15997 (pass_parallelize_loops): Convert from a global struct to a subclass
15998 of gimple_opt_pass along with...
15999 (pass_data_parallelize_loops): ...new pass_data instance and...
16000 (make_pass_parallelize_loops): ...new function.
16001 (pass_loop_prefetch): Convert from a global struct to a subclass of
16002 gimple_opt_pass along with...
16003 (pass_data_loop_prefetch): ...new pass_data instance and...
16004 (make_pass_loop_prefetch): ...new function.
16005 (pass_iv_optimize): Convert from a global struct to a subclass of
16006 gimple_opt_pass along with...
16007 (pass_data_iv_optimize): ...new pass_data instance and...
16008 (make_pass_iv_optimize): ...new function.
16009 (pass_tree_loop_done): Convert from a global struct to a subclass of
16010 gimple_opt_pass along with...
16011 (pass_data_tree_loop_done): ...new pass_data instance and...
16012 (make_pass_tree_loop_done): ...new function.
16013 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
16014 struct to a subclass of gimple_opt_pass along with...
16015 (pass_data_cse_reciprocals): ...new pass_data instance and...
16016 (make_pass_cse_reciprocals): ...new function.
16017 (pass_cse_sincos): Convert from a global struct to a subclass of
16018 gimple_opt_pass along with...
16019 (pass_data_cse_sincos): ...new pass_data instance and...
16020 (make_pass_cse_sincos): ...new function.
16021 (pass_optimize_bswap): Convert from a global struct to a subclass of
16022 gimple_opt_pass along with...
16023 (pass_data_optimize_bswap): ...new pass_data instance and...
16024 (make_pass_optimize_bswap): ...new function.
16025 (pass_optimize_widening_mul): Convert from a global struct to a
16026 subclass of gimple_opt_pass along with...
16027 (pass_data_optimize_widening_mul): ...new pass_data instance and...
16028 (make_pass_optimize_widening_mul): ...new function.
16029 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
16030 subclass of gimple_opt_pass along with...
16031 (pass_data_phiopt): ...new pass_data instance and...
16032 (make_pass_phiopt): ...new function.
16033 (pass_cselim): Convert from a global struct to a subclass of
16034 gimple_opt_pass along with...
16035 (pass_data_cselim): ...new pass_data instance and...
16036 (make_pass_cselim): ...new function.
16037 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
16038 subclass of gimple_opt_pass along with...
16039 (pass_data_phiprop): ...new pass_data instance and...
16040 (make_pass_phiprop): ...new function.
16041 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
16042 subclass of gimple_opt_pass along with...
16043 (pass_data_pre): ...new pass_data instance and...
16044 (make_pass_pre): ...new function.
16045 (pass_fre): Convert from a global struct to a subclass of
16046 gimple_opt_pass along with...
16047 (pass_data_fre): ...new pass_data instance and...
16048 (make_pass_fre): ...new function.
16049 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
16050 subclass of gimple_opt_pass along with...
16051 (pass_data_reassoc): ...new pass_data instance and...
16052 (make_pass_reassoc): ...new function.
16053 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
16054 subclass of gimple_opt_pass along with...
16055 (pass_data_sink_code): ...new pass_data instance and...
16056 (make_pass_sink_code): ...new function.
16057 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
16058 subclass of gimple_opt_pass along with...
16059 (pass_data_strlen): ...new pass_data instance and...
16060 (make_pass_strlen): ...new function.
16061 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
16062 struct to a subclass of gimple_opt_pass along with...
16063 (pass_data_build_alias): ...new pass_data instance and...
16064 (make_pass_build_alias): ...new function.
16065 (pass_build_ealias): Convert from a global struct to a subclass of
16066 gimple_opt_pass along with...
16067 (pass_data_build_ealias): ...new pass_data instance and...
16068 (make_pass_build_ealias): ...new function.
16069 (pass_ipa_pta): Convert from a global struct to a subclass of
16070 simple_ipa_opt_pass along with...
16071 (pass_data_ipa_pta): ...new pass_data instance and...
16072 (make_pass_ipa_pta): ...new function.
16073 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
16074 subclass of gimple_opt_pass along with...
16075 (pass_data_uncprop): ...new pass_data instance and...
16076 (make_pass_uncprop): ...new function.
16077 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
16078 global struct to a subclass of gimple_opt_pass along with...
16079 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
16080 (make_pass_late_warn_uninitialized): ...new function.
16081 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
16082 to a subclass of gimple_opt_pass along with...
16083 (pass_data_init_datastructures): ...new pass_data instance and...
16084 (make_pass_init_datastructures): ...new function.
16085 (pass_early_warn_uninitialized): Convert from a global struct to a
16086 subclass of gimple_opt_pass along with...
16087 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
16088 (make_pass_early_warn_uninitialized): ...new function.
16089 (pass_update_address_taken): Convert from a global struct to a
16090 subclass of gimple_opt_pass along with...
16091 (pass_data_update_address_taken): ...new pass_data instance and...
16092 (make_pass_update_address_taken): ...new function.
16093 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
16094 struct to a subclass of gimple_opt_pass along with...
16095 (pass_data_release_ssa_names): ...new pass_data instance and...
16096 (make_pass_release_ssa_names): ...new function.
16097 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
16098 subclass of gimple_opt_pass along with...
16099 (pass_data_stdarg): ...new pass_data instance and...
16100 (make_pass_stdarg): ...new function.
16101 * tree-switch-conversion.c (pass_convert_switch): Convert from a
16102 global struct to a subclass of gimple_opt_pass along with...
16103 (pass_data_convert_switch): ...new pass_data instance and...
16104 (make_pass_convert_switch): ...new function.
16105 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
16106 to a subclass of gimple_opt_pass along with...
16107 (pass_data_tail_recursion): ...new pass_data instance and...
16108 (make_pass_tail_recursion): ...new function.
16109 (pass_tail_calls): Convert from a global struct to a subclass of
16110 gimple_opt_pass along with...
16111 (pass_data_tail_calls): ...new pass_data instance and...
16112 (make_pass_tail_calls): ...new function.
16113 * tree-vect-generic.c (pass_lower_vector): Convert from a global
16114 struct to a subclass of gimple_opt_pass along with...
16115 (pass_data_lower_vector): ...new pass_data instance and...
16116 (make_pass_lower_vector): ...new function.
16117 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
16118 gimple_opt_pass along with...
16119 (pass_data_lower_vector_ssa): ...new pass_data instance and...
16120 (make_pass_lower_vector_ssa): ...new function.
16121 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
16122 to a subclass of gimple_opt_pass along with...
16123 (pass_data_slp_vectorize): ...new pass_data instance and...
16124 (make_pass_slp_vectorize): ...new function.
16125 (pass_ipa_increase_alignment): Convert from a global struct to a
16126 subclass of simple_ipa_opt_pass along with...
16127 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
16128 (make_pass_ipa_increase_alignment): ...new function.
16129 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
16130 gimple_opt_pass along with...
16131 (pass_data_vrp): ...new pass_data instance and...
16132 (make_pass_vrp): ...new function.
16133 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
16134 subclass of simple_ipa_opt_pass along with...
16135 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
16136 (make_pass_ipa_free_lang_data): ...new function.
16137 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
16138 gimple_opt_pass along with...
16139 (pass_data_tsan): ...new pass_data instance and...
16140 (make_pass_tsan): ...new function.
16141 (pass_tsan_O0): Convert from a global struct to a subclass of
16142 gimple_opt_pass along with...
16143 (pass_data_tsan_O0): ...new pass_data instance and...
16144 (make_pass_tsan_O0): ...new function.
16145 * var-tracking.c (pass_variable_tracking): Convert from a global
16146 struct to a subclass of rtl_opt_pass along with...
16147 (pass_data_variable_tracking): ...new pass_data instance and...
16148 (make_pass_variable_tracking): ...new function.
16149 * web.c (pass_web): Convert from a global struct to a subclass of
16150 rtl_opt_pass along with...
16151 (pass_data_web): ...new pass_data instance and...
16152 (make_pass_web): ...new function.
16153 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
16154 declaration with that of...
16155 (make_pass_mode_switch_use): ...new function.
16156 (pass_resolve_sw_modes): Replace declaration with that of...
16157 (make_pass_resolve_sw_modes): ...new function.
16158 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
16159 from a global struct to a subclass of rtl_opt_pass along with...
16160 (pass_data_mode_switch_use): ...new pass_data instance and...
16161 (make_pass_mode_switch_use): ...new function.
16162 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
16163 from a global struct to a subclass of rtl_opt_pass along with...
16164 (pass_data_resolve_sw_modes): ...new pass_data instance and...
16165 (make_pass_resolve_sw_modes): ...new function.
16166 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
16167 struct to a subclass of rtl_opt_pass along with...
16168 (pass_data_insert_vzeroupper): ...new pass_data instance and...
16169 (make_pass_insert_vzeroupper): ...new function.
16170 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
16171 global struct to a subclass of rtl_opt_pass along with...
16172 (pass_data_work_around_errata): ...new pass_data instance and...
16173 (make_pass_work_around_errata): ...new function.
16174 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
16175 struct to a subclass of rtl_opt_pass along with...
16176 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
16177 (make_pass_mips_machine_reorg2): ...new function.
16179 2013-08-05 David Malcolm <dmalcolm@redhat.com>
16181 * passes.c (pass_manager::operator new): New.
16183 2013-08-05 David Malcolm <dmalcolm@redhat.com>
16185 Handwritten part of conversion of passes to C++ classes.
16187 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
16188 (toplev.o): Add dep on PASS_MANAGER_H.
16189 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
16190 of early local pases to reflect this moving from a global to a
16191 member of gcc::pass_manager.
16192 (cgraph_add_new_function): Likewise.
16193 * lto-cgraph.c (lto_output_node): Update for conversion of
16194 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
16195 * passes.c (opt_pass::clone): New.
16196 (opt_pass::gate): New.
16197 (opt_pass::execute): New.
16198 (opt_pass::opt_pass): New.
16199 (pass_manager::execute_early_local_passes): New.
16200 (pass_manager::execute_pass_mode_switching): new.
16201 (finish_optimization_passes): Convert to...
16202 (pass_manager::finish_optimization_passes): ...this.
16203 (finish_optimization_passes): Update for conversion of passes to
16205 (register_dump_files_1): Use has_gate since we cannot portably
16206 check a vtable entry against NULL.
16207 (dump_one_pass): Likewise.
16208 (ipa_write_summaries_2): Likewise.
16209 (ipa_write_optimization_summaries_1): Likewise.
16210 (ipa_read_summaries_1): Likewise.
16211 (ipa_read_optimization_summaries_1): Likewise.
16212 (execute_ipa_stmt_fixups): Likewise.
16213 (pass_manager::pass_manager): Rewrite pass-creation, invoking
16214 pass-creation functions rather than wiring up globals, and
16215 storing the results in fields of pass_manager generated using
16216 pass-instances.def.
16217 (pass_manager::dump_profile_report): Update for conversion of
16218 passes to C++ classes.
16219 (pass_manager::execute_ipa_summary_passes): Likewise.
16220 (execute_one_ipa_transform_pass): Likewise.
16221 (execute_one_pass): Use has_gate and has_execute since we cannot
16222 portably check a vtable entry against NULL.
16223 * pass_manager.h (pass_manager::finish_optimization_passes): New.
16224 (pass_manager): Use pass-instances.def to add fields for the
16225 various pass instances.
16226 * toplev.c (finalize): Update for move of
16227 finish_optimization_passes to a method of gcc::pass_manager.
16228 * toplev.h (finish_optimization_passes): Move to method of class
16230 * tree-pass.h (struct pass_data): New.
16231 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
16232 (opt_pass::gate): Convert to virtual function.
16233 (opt_pass::~opt_pass): New.
16234 (opt_pass::clone): New.
16235 (opt_pass::execute): Convert to virtual function.
16236 (opt_pass::opt_pass): New.
16237 (opt_pass::ctxt_): new.
16238 (gimple_opt_pass): Convert to subclass of opt_pass.
16239 (gimple_opt_pass::gimple_opt_pass): New.
16240 (rtl_opt_pass): Convert to subclass of opt_pass.
16241 (rtl_opt_pass::rtl_opt_pass): New.
16242 (ipa_opt_pass_d): Convert to subclass of opt_pass.
16243 (ipa_opt_pass_d::ipa_opt_pass_d): New.
16244 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
16245 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
16246 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
16247 invocation of pass_mode_switching to reflect this moving from a
16248 global to a member of gcc::pass_manager.
16249 (ix86_option_override): Rework how pass_insert_vzeroupper is
16250 added to the pass_manager to reflect autogenerated changes.
16251 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
16253 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
16255 PR rtl-optimization/57708
16256 * recog.c (peep2_find_free_register): Validate all regs in a
16259 2013-08-05 Jan Hubicka <jh@suse.cz>
16262 * cgraph.c (verify_cgraph_node): Accept local flags from other
16264 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
16265 (function_and_variable_visibility): Likewise.
16266 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
16268 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
16270 * graph.c (init_graph_slim_pretty_print): Remove.
16271 (print_graph_cfg): Do not call it. Use local pretty printer.
16272 (start_graph_dump): Likewise.
16274 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
16276 * gimple-pretty-print.c (buffer): Remove.
16277 (initialized): Likewise.
16278 (maybe_init_pretty_print): Likewise.
16279 (print_gimple_stmt): Do not call it. Use non-static local
16280 pretty_printer variable.
16281 (print_gimple_expr): Likewise.
16282 (print_gimple_seq): Likewise.
16283 (gimple_dump_bb): Likewise.
16285 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
16287 * asan.c (asan_pp): Remove.
16288 (asan_pp_initialized): Likewise.
16289 (asan_pp_initialize): Likewise.
16290 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
16291 (asan_emit_stack_protection): Tidy. Use local pretty printer.
16292 (asan_add_global): Likewise.
16294 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
16296 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
16297 * diagnostic.h (diagnostic_flush_buffer): Adjust.
16298 * pretty-print.c (pp_formatted_text_data): Likewise.
16299 (pp_indent): Rename from pp_base_indent.
16300 (pp_format): Rename from pp_base_format.
16301 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
16302 (pp_format_verbatim): Rename from pp_base_format_verbatim.
16303 (pp_flush): Rename from pp_base_flush.
16304 (pp_set_line_maximum_length): Rename from
16305 pp_base_set_line_maximum_length.
16306 (pp_clear_output_area): Rename from pp_base_clear_output_area.
16307 (pp_set_prefix): Rename from pp_base_set_prefix.
16308 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
16309 (pp_emit_prefix): Rename from pp_base_emit_prefix.
16310 (pp_append_text): Rename from pp_base_append_text.
16311 (pp_formatted_text): Rename from pp_base_formatted_text.
16312 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
16313 (pp_remaining_character_count_for_line): Rename from
16314 pp_base_remaining_character_count_for_line.
16315 (pp_newline): Rename from pp_base_newline.
16316 (pp_character): Rename from pp_base_character.
16317 (pp_string): Rename from pp_base_string.
16318 (pp_maybe_space): Rename from pp_base_maybe_space.
16319 * asan.c (asan_pp_string): Adjust.
16320 (asan_emit_stack_protection): Likewise.
16321 (asan_add_global): Likewise.
16322 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
16323 * tree-mudflap.c (mf_varname_tree): Likewise.
16324 * tree-pretty-print.c (pp_tree_identifier): Rename from
16325 pp_base_tree_identifier.
16326 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
16327 Declare as function.
16329 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
16331 * pretty-print.h (pp_bar_bar): New.
16332 (pp_ampersand_ampersand): Likewise.
16333 (pp_less_equal): Likewise.
16334 (pp_greater_equal): Likewise.
16335 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
16336 printer functions instead of pp_string or operators and punctuators.
16337 (dump_gimple_call): Likewise.
16338 (dump_gimple_omp_for): Likewise.
16339 (dump_gimple_transaction): Likewise.
16340 (dump_gimple_phi): Likewise.
16341 (pp_gimple_stmt_1): Likewise.
16342 * sched-vis.c (print_insn): Likewise.
16343 * tree-mudflap.c (mf_varname_tree): Likewise.
16344 * tree-pretty-print.c (dump_block_node): Likewise.
16345 (dump_generic_node): Likewise.
16347 2013-08-02 Jan Hubicka <jh@suse.cz>
16349 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
16351 * lto-streamer-out.c (tree_is_indexable): Results decls and
16352 parm decls are not indexable.
16353 (DFS_write_tree_body): Do not follow args and results.
16354 (hash_tree): Likewise.
16355 (output_functions): Rearrange so struct function is needed
16356 only when real body is output; be able to also ouptut abstract
16357 functions; output DECL_ARGUMENTS and DECL_RESULT.
16358 (lto_output): When not in WPA, ale store abstract functions.
16359 (write_symbol): Do not care about RESULT_DECL.
16360 (output_symbol_p): Handle correctly sbtract decls.
16361 * lto-streamer-in.c (input_function): Rearrange so struct
16362 function can be NULL at entry; allow streaming of
16363 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
16364 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
16365 sanity check during LTO.
16366 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
16367 RESULT_DECl and DECL_ARGUMENTS.
16368 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
16371 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
16373 * pretty-print.h (pp_underscore): New.
16375 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
16376 printer functions instead of pp_character.
16377 (dump_binary_rhs): Likewise.
16378 (dump_ternary_rhs): Likewise.
16379 (dump_gimple_call_args): Likewise.
16380 (pp_points_to_solution): Likewise.
16381 (dump_gimple_call): Likewise.
16382 (dump_gimple_switch): Likewise.
16383 (dump_gimple_cond): Likewise.
16384 (dump_gimple_bind): Likewise.
16385 (dump_gimple_try): Likewise.
16386 (dump_gimple_omp_for): Likewise.
16387 (dump_gimple_omp_continue): Likewise.
16388 (dump_gimple_omp_single): Likewise.
16389 (dump_gimple_omp_sections): Likewise.
16390 (dump_gimple_omp_block): Likewise.
16391 (dump_gimple_omp_critical): Likewise.
16392 (dump_gimple_transaction): Likewise.
16393 (dump_gimple_asm): Likewise.
16394 (dump_gimple_phi): Likewise.
16395 (dump_gimple_omp_parallel): Likewise.
16396 (dump_gimple_omp_task): Likewise.
16397 (dump_gimple_omp_atomic_load): Likewise.
16398 (dump_gimple_omp_atomic_store): Likewise.
16399 (dump_gimple_mem_ops): Likewise.
16400 (pp_gimple_stmt_1): Likewise.
16401 (pp_cfg_jump): Likewise.
16402 (dump_implicit_edges): Likewise.
16403 (gimple_dump_bb_for_graph): Likewise.
16404 * graph.c (draw_cfg_node): Likewise.
16405 * langhooks.c (lhd_print_error_function): Likewise.
16406 * sched-vis.c (print_exp): Likewise.
16407 (print_value): Likewise.
16408 (print_pattern): Likewise.
16409 (print_insn): Likewise.
16410 (rtl_dump_bb_for_graph): Likewise.
16411 * tree-pretty-print.c (dump_function_declaration): Likewise.
16412 (dump_array_domain): Likewise.
16413 (dump_omp_clause): Likewise.
16414 (dump_location): Likewise.
16415 (dump_generic_node): Likewise.
16416 (print_struct_decl): Likewise.
16417 * diagnostic.c (diagnostic_show_locus): Use pp_space.
16419 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
16421 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
16422 candidate table when replacing a candidate statement.
16423 (replace_rhs_if_not_dup): Likewise.
16424 (replace_one_candidate): Likewise.
16426 2013-08-02 Jan Hubicka <jh@suse.cz>
16427 Martin Liska <marxin.liska@gmail.com>
16429 * cgraphunit.c (add_new_function): Fix logic when adding from
16431 (assemble_thunk): Rename to ...
16432 (expand_thunk); .. this one; export; get it working with
16433 general functions; make produced gimple valid.
16434 * cgraph.h (expand_thunk): Declare.
16436 2013-08-02 Jan Hubicka <jh@suse.cz>
16438 * ipa-cp.c (gather_context_independent_values): Use
16439 ipa_get_param_move_cost.
16440 (get_replacement_map): Remove PARAM; move parameter folding
16442 (create_specialized_node): Update.
16443 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
16444 assert that we have gimple body; update move_cost.
16445 (count_formal_params): Assert that we have gimple body.
16446 (ipa_dump_param): New function.
16447 (ipa_alloc_node_params): Break out from ...
16448 (ipa_initialize_node_params): ... here.
16449 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
16450 (ipa_write_node_info): Stream move costs.
16451 (ipa_read_node_info): Read move costs.
16452 (ipa_update_after_lto_read): Do not recompute node params.
16453 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
16454 (ipa_get_param): Check we are not in WPA.
16455 (ipa_get_param_move_cost): New.
16456 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
16457 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
16458 parm numbers to be present.
16460 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
16462 PR rtl-optimization/58048
16463 * lra-constraints.c (process_alt_operands): Don't check asm
16464 operand on register.
16466 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
16468 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
16469 the implied StoreLoad barrier for atomic operations if before.
16471 2013-08-02 Jan Hubicka <jh@suse.cz>
16472 Martin Liska <marxin.liska@gmail.com>
16474 * cgraph.c (cgraph_function_body_availability): Do not check
16476 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
16477 symtab_node_availability): Declare.
16478 * ipa.c (can_replace_by_local_alias): New.
16479 (function_and_variable_visibility): Use it.
16480 * symtab.c (symtab_for_node_and_aliases,
16481 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
16483 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
16485 PR rtl-optimization/57963
16486 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
16487 (lra_constraints): Use them.
16489 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
16491 * config/arm/types.md (define_attr "type"): Add "load_acq"
16493 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
16495 (cortex_a53_store1): Likewise.
16497 2013-08-01 Jan Hubicka <jh@suse.cz>
16499 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
16500 partitions are not needed.
16502 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
16504 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
16505 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
16506 MAYBE_NON_Q_CLASS_P where appropriate.
16508 2013-08-01 Jan Hubicka <jh@suse.cz>
16510 * cgraph.h (release_function_body): Declare.
16511 * tree.c (free_lang_data_in_decl): Free, parameters and return values
16512 of unused delcarations.
16514 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16516 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
16517 RTL form when subtracting a constant.
16519 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16521 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
16522 Generate canonical plus rtx with negated immediate instead of minus
16524 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
16526 2013-08-01 Jan Hubicka <jh@suse.cz>
16528 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
16529 (cgraph_release_function_body): Likewise.
16530 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
16531 * cgraph.h (cgrpah_node): Rename abstract_and_needed
16532 to used_as_abstract_origin.
16533 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
16534 symbols used as abstract origins.
16535 * cgraphunit.c (analyze_functions): Update.
16536 * ipa.c (symtab_remove_unreachable_nodes): Recompute
16537 used_as_abstract_origin.
16538 * tree-inline.c (tree_function_versioning): Update
16539 used_as_abstract_origin; be ready for DECL_RESULT and
16540 DECL_ARGUMENTS to be NULL.
16542 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
16543 for abstract functions.
16544 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
16547 2013-08-01 Jan Hubicka <jh@suse.cz>
16549 * profile.c (compute_value_histograms): Fix thinko.
16551 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
16553 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
16554 aarch-common-protos.h to extra_headers.
16555 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
16556 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
16557 * config/aarch64/t-aarch64 (aarch-common.o): Define.
16559 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
16561 * config/aarch64/aarch64.md (define_attr "type"): Delete.
16562 Include "../arm/types.md". Define "type" attribute for all patterns.
16563 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
16566 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
16568 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
16569 to support power8 load fusion.
16570 (fusion_gpr_mem_load): Likewise.
16572 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
16574 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
16575 declarations for power8 load fusion.
16576 (emit_fusion_gpr_load): Likewise.
16578 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
16579 tuning for power8, turn on fusion mode by default. Turn on sign
16580 extending fusion mode if normal fusion mode is on, and we are at
16582 (fusion_gpr_load_p): New function, return true if we can fuse an
16583 addis instruction with a dependent load to a GPR.
16584 (emit_fusion_gpr_load): Emit the instructions for power8 load
16587 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
16588 (VSX load fusion peepholes): New peepholes to fuse together an
16589 addi instruction with a VSX load instruction.
16591 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
16592 peepholes to fuse an addis instruction with a load to a GPR base
16593 register. If we are supporting sign extending fusions, convert
16594 sign extending loads to zero extending loads and add an explicit
16597 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
16599 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
16600 aarch-common-protos.h to extra_headers.
16601 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
16602 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
16603 (arm_early_store_addr_dep): Likewise.
16604 (arm_no_early_alu_shift_dep): Likewise.
16605 (arm_no_early_alu_shift_value_dep): Likewise.
16606 (arm_no_early_mul_dep): Likewise.
16607 (arm_no_early_store_addr_dep): Likewise.
16608 (arm_mac_accumulator_is_mul_result): Likewise.
16609 (arm_mac_accumulator_is_result): Likewise.
16610 * config/arm/aarch-common.c: ... here. New file.
16611 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
16613 (arm_early_store_addr_dep): Likewise.
16614 (arm_no_early_alu_shift_dep): Likewise.
16615 (arm_no_early_alu_shift_value_dep): Likewise.
16616 (arm_no_early_mul_dep): Likewise.
16617 (arm_no_early_store_addr_dep): Likewise.
16618 (arm_mac_accumulator_is_mul_result): Likewise.
16619 (arm_mac_accumulator_is_result): Likewise.
16620 * config/arm/aarch-common-protos.h: ... here. New file.
16621 * config/arm/t-arm (aarch-common.o): Define.
16623 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
16625 * config/arm/arm.md: Include new file "types.md".
16626 (define_attr "type"): Move from here to ...
16627 (define_attr "mul32"): Likewise.
16628 (define_attr "mul64"): Likewise.
16629 * config/arm/types.md: ... here. New file.
16631 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
16633 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
16634 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
16636 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
16638 * gen-pass-instances.awk: Fix offset of substr().
16640 2013-07-31 David Malcolm <dmalcolm@redhat.com>
16642 * Makefile.in (pass-instances.def): New.
16643 (passes.o): Replace dependency on passes.def with one on
16646 * gen-pass-instances.awk: New.
16648 * passes.c (pass_manager::pass_manager): Use pass-instances.def
16649 rather than passes.def, updating local definition of NEXT_PASS
16650 macro to add an extra NUM parameter (currently unused).
16652 2013-07-30 David Malcolm <dmalcolm@redhat.com>
16654 * Makefile.in (PASS_MANAGER_H): New.
16655 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
16656 (passes.o): Likewise.
16657 (statistics.o): Likewise.
16658 (cgraphunit.o): Likewise.
16659 (context.o): Depend on PASS_MANAGER_H.
16661 * pass_manager.h: New.
16663 * cgraphunit.c (cgraph_add_new_function): Update for moves
16664 of globals to fields of pass_manager.
16665 (analyze_function): Likewise.
16666 (expand_function): Likewise.
16667 (ipa_passes): Likewise.
16668 (compile): Likewise.
16670 * context.c (context::context): New.
16671 * context.h (context::context): New.
16672 (context::get_passes): New.
16673 (context::passes_): New.
16675 * lto-cgraph.c (input_node): Update for moves of globals to
16676 fields of pass_manager.
16678 * passes.c (all_passes): Remove, in favor of a field of the
16679 same name within the new class pass_manager.
16680 (all_small_ipa_passes): Likewise.
16681 (all_lowering_passes): Likewise.
16682 (all_regular_ipa_passes): Likewise.
16683 (all_late_ipa_passes): Likewise.
16684 (all_lto_gen_passes): Likewise.
16685 (passes_by_id): Likewise.
16686 (passes_by_id_size): Likewise.
16687 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
16688 the new class pass_manager.
16689 (set_pass_for_id): Convert to...
16690 (pass_manager::set_pass_for_id): ...method.
16691 (get_pass_for_id): Convert to...
16692 (pass_manager::get_pass_for_id): ...method.
16693 (register_one_dump_file): Move body of implementation into...
16694 (pass_manager::register_one_dump_file): ...here.
16695 (register_dump_files_1): Convert to...
16696 (pass_manager::register_dump_files_1): ...method.
16697 (register_dump_files): Convert to...
16698 (pass_manager::register_dump_files): ...method.
16699 (create_pass_tab): Update for moves of globals to fields of
16701 (dump_passes): Move body of implementation into...
16702 (pass_manager::dump_passes): ...here.
16703 (register_pass): Move body of implementation into...
16704 (pass_manager::register_pass): ...here.
16705 (init_optimization_passes): Convert into...
16706 (pass_manager::pass_manager): ...constructor for new
16707 pass_manager class, and initialize the pass_lists array.
16708 (check_profile_consistency): Update for moves of globals to
16709 fields of pass_manager.
16710 (dump_profile_report): Move body of implementation into...
16711 (pass_manager::dump_profile_report): ...here.
16712 (ipa_write_summaries_1): Update for moves of pass lists from
16713 being globals to fields of pass_manager.
16714 (ipa_write_optimization_summaries): Likewise.
16715 (ipa_read_summaries): Likewise.
16716 (ipa_read_optimization_summaries): Likewise.
16717 (execute_all_ipa_stmt_fixups): Likewise.
16719 * statistics.c (statistics_fini): Update for moves of globals to
16720 fields of pass_manager.
16722 * toplev.c (general_init): Replace call to
16723 init_optimization_passes with construction of the pass_manager
16726 * tree-pass.h (all_passes): Remove, in favor of a field of the
16727 same name within the new class pass_manager.
16728 (all_small_ipa_passes): Likewise.
16729 (all_lowering_passes): Likewise.
16730 (all_regular_ipa_passes): Likewise.
16731 (all_lto_gen_passes): Likewise.
16732 (all_late_ipa_passes): Likewise.
16733 (passes_by_id): Likewise.
16734 (passes_by_id_size): Likewise.
16735 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
16736 the new class pass_manager.
16737 (get_pass_for_id): Remove.
16739 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
16741 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
16744 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
16746 * arm.md (mulhi3): New expand pattern.
16748 2013-07-30 Jan Hubicka <jh@suse.cz>
16749 Martin Liska <marxin.liska@gmail.com>
16751 * profile.c (compute_value_histograms): Do not ICE when
16752 there is mismatch only on some counters.
16754 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
16756 PR rtl-optimization/57637
16757 * function.c (move_insn_for_shrink_wrap): Also check the
16758 GEN set of the LIVE problem for the liveness analysis
16759 if it exists, otherwise give up.
16761 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
16763 PR tree-optimization/57993
16764 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
16765 replaced statement in the candidate table.
16766 (phi_add_costs): Return infinite cost when the hidden basis does
16767 not dominate all phis on which the candidate is dependent.
16768 (replace_one_candidate): Record replaced statement in the
16771 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
16773 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
16774 (ashlv2si3): New expander.
16775 (*ashlv2si3_i): New define_insn_and_split.
16776 * predicates.md (float_operation): Allow patterns with three
16777 basic sub-patterns.
16779 PR rtl-optimization/58021
16780 * mode-switching.c (create_pre_exit): Always split off preceding
16781 insns if we are not at the basic block head.
16783 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
16785 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
16786 (UCLIBC_DYNAMIC_LINKER): New macro.
16787 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
16789 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
16790 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
16792 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
16793 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
16794 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
16795 for SF and DF modes. Use ieee_quad_format for TF mode.
16796 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
16797 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
16798 (mips_option_override): Handle `-mnan=legacy'.
16799 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
16800 `-mabs=2008' and `-mnan=2008'.
16801 (OPTION_DEFAULT_SPECS): Add "nan" default.
16802 (ASM_SPEC): Handle `-mnan='.
16803 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
16804 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
16805 comment accordingly.
16806 (neg<mode>2): Likewise.
16807 * config/mips/mips.opt (mabs, mnan): New options.
16808 * doc/install.texi (Configuration): Document `--with-nan=' option.
16809 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
16811 (MIPS Options): Document them.
16812 * config.gcc <mips*-*-*>: Handle `--with-nan='.
16813 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
16814 * configure: Regenerate.
16815 * config.in: Regenerate.
16817 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
16819 * config/i386/i386.md (float post-reload splitters): Do not check
16820 for subregs of SSE registers.
16822 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
16823 H.J. Lu <hongjiu.lu@intel.com>
16827 * config/i386/i386.md (post-reload splitter
16828 to avoid partial SSE reg dependency stalls): New pattern.
16830 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
16832 * config/s390/s390.md ("movcc"): Swap load and store instructions.
16834 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
16836 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
16837 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
16839 2013-07-26 Cary Coutant <ccoutant@google.com>
16841 * dwarf2out.c (die_checksum_ordered): Don't include template
16842 instantiations in signature.
16843 (is_template_parameter): New function.
16844 (is_template_instantiation): New function.
16845 (generate_skeleton_bottom_up): Don't include template instantiations
16847 (generate_skeleton): Likewise.
16848 (break_out_comdat_types): Move recursive call to break out nested
16850 (prune_unused_types_mark_generic_parms_dies): Call
16851 is_template_parameter.
16853 2013-07-26 Ian Bolton <ian.bolton@arm.com>
16855 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
16856 uses vector registers.
16857 * config/aarch64/iterators.md: Add attributes rtn and vas.
16859 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16860 Richard Earnshaw <richard.earnshaw@arm.com>
16862 * combine.c (simplify_comparison): Re-canonicalize operands
16864 * config/arm/arm.md (movcond_addsi): New splitter.
16866 2013-07-25 Sterling Augustine <saugustine@google.com>
16868 * dwarf2out.c (size_of_pubnames): Move code to...
16869 (include_pubname_in_output): ...here. New.
16870 (want_pubnames): Rearrange.
16871 (output_pubnames): Call include_pubname_in_output. Move assertion.
16873 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
16875 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
16877 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
16880 * doc/extend.texi: Remove obsolete builtins. Fix
16881 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
16883 2013-07-25 Jan Hubicka <jh@suse.cz>
16885 * cgraph.c (release_function_body): Break out from ...
16886 (cgraph_release_function_body): ... this one; also release DECL_RESULT
16887 and DECL_ARGUMENTS.
16888 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
16889 old_tree in the map.
16890 (create_specialized_node): Update.
16891 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
16893 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
16894 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
16895 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
16896 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
16899 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16901 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
16902 addsi3_carryin_alt2_<optab>): Correct output template.
16904 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16906 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
16907 Adjust for arm_restrict_it.
16908 Remove trailing whitespace.
16910 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
16912 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
16913 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
16915 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
16917 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
16919 PR rtl-optimization/57960
16920 * lra-constraints.c (process_alt_operands): Use the right mode
16921 when checking strict_low.
16923 2013-07-25 Jan Hubicka <jh@suse.cz>
16925 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
16926 * cgraph.c (cgraph_remove_node): Do not release function body
16927 when in cgraph streaming.
16928 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
16929 in other partitions are not considered reachable; fix handling of
16932 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16934 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
16936 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16941 * config/arm/arm.md ("*sibcall_insn): Replace use of
16942 Ss with US. Adjust output for v5 and v4t.
16943 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
16945 * config/arm/constraints.md ("Ss"): Rename to US.
16947 2013-07-25 Terry Guo <terry.guo@arm.com>
16949 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
16950 shift_add/shift_sub0/shift_sub1 RTXs.
16952 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
16953 Anton Blanchard <anton@au1.ibm.com>
16955 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
16956 (altivec_vpks<VI_char>ss): Likewise.
16957 (altivec_vpks<VI_char>us): Likewise.
16958 (altivec_vpku<VI_char>us): Likewise.
16959 (altivec_vpku<VI_char>um): Likewise.
16961 2013-07-24 David Malcolm <dmalcolm@redhat.com>
16963 Introduce context class.
16965 * Makefile.in (CONTEXT_H): New.
16966 (OBJS): Add context.o.
16967 (toplev.o): Add CONTEXT_H to dependencies.
16970 * toplev.c (general_init): Create the singleton gcc::context instance.
16976 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
16978 PR rtl-optimization/57968
16979 * mode-switching.c (create_pre_exit): Allow instructions that
16980 don't set a return register to need a non-exit mode.
16982 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
16983 Anton Blanchard <anton@au1.ibm.com>
16985 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
16986 operands to vperm for little endian.
16987 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
16988 of lvsl to create the control mask for a vperm for little endian.
16990 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16991 Anton Blanchard <anton@au1.ibm.com>
16993 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
16994 two operands for little-endian.
16996 2013-07-23 Steve Ellcey <sellcey@mips.com>
16998 * config/mips/mips.c (mips_case_values_threshold): New.
16999 (TARGET_CASE_VALUES_THRESHOLD): Define.
17001 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17002 Anton Blanchard <anton@au1.ibm.com>
17004 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
17005 selection of field for vector splat in little endian mode.
17007 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
17009 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
17010 expanders to rs6000.md.
17011 (ior<mode>3): Likewise.
17012 (and<mode>3): Likewise.
17013 (one_cmpl<mode>2): Likewise.
17014 (nor<mode>3): Likewise.
17015 (andc<mode>3): Likewise.
17016 (eqv<mode>3): Likewise.
17017 (nand<mode>3): Likewise.
17018 (orc<mode>3): Likewise.
17020 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
17023 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
17024 to split multi-word logical operations.
17025 (rs6000_split_logical_di): Likewise.
17026 (rs6000_split_logical): Likewise.
17028 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
17029 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
17030 and allow TImode operations in 32-bit.
17031 (vsx_and<mode>3_64bit): Likewise.
17032 (vsx_ior<mode>3_32bit): Likewise.
17033 (vsx_ior<mode>3_64bit): Likewise.
17034 (vsx_xor<mode>3_32bit): Likewise.
17035 (vsx_xor<mode>3_64bit): Likewise.
17036 (vsx_one_cmpl<mode>2_32bit): Likewise.
17037 (vsx_one_cmpl<mode>2_64bit): Likewise.
17038 (vsx_nor<mode>3_32bit): Likewise.
17039 (vsx_nor<mode>3_64bit): Likewise.
17040 (vsx_andc<mode>3_32bit): Likewise.
17041 (vsx_andc<mode>3_64bit): Likewise.
17042 (vsx_eqv<mode>3_32bit): Likewise.
17043 (vsx_eqv<mode>3_64bit): Likewise.
17044 (vsx_nand<mode>3_32bit): Likewise.
17045 (vsx_nand<mode>3_64bit): Likewise.
17046 (vsx_orc<mode>3_32bit): Likewise.
17047 (vsx_orc<mode>3_64bit): Likewise.
17049 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
17050 logical types in GPRs.
17052 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
17053 logical insns to rs6000.md, and allow TImode operations in
17055 (altivec_ior<mode>3): Likewise.
17056 (altivec_xor<mode>3): Likewise.
17057 (altivec_one_cmpl<mode>2): Likewise.
17058 (altivec_nor<mode>3): Likewise.
17059 (altivec_andc<mode>3): Likewise.
17061 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
17062 attributes for moving the 128-bit logical operations into
17064 (BOOL_REGS_OUTPUT): Likewise.
17065 (BOOL_REGS_OP1): Likewise.
17066 (BOOL_REGS_OP2): Likewise.
17067 (BOOL_REGS_UNARY): Likewise.
17068 (BOOL_REGS_AND_CR0): Likewise.
17069 (one_cmpl<mode>2): Add support for DI logical operations on
17070 32-bit, splitting the operations to 32-bit.
17071 (anddi3): Likewise.
17072 (iordi3): Likewise.
17073 (xordi3): Likewise.
17074 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
17075 changes to combine the 32/64-bit code, allow logical operations on
17076 TI mode in 32-bit, and to use similar match_operator patterns like
17077 scalar mode uses. Combine the Altivec and VSX code for logical
17078 operations, and move it here.
17079 (ior<mode>3, 128-bit types): Likewise.
17080 (xor<mode>3, 128-bit types): Likewise.
17081 (one_cmpl<mode>3, 128-bit types): Likewise.
17082 (nor<mode>3, 128-bit types): Likewise.
17083 (andc<mode>3, 128-bit types): Likewise.
17084 (eqv<mode>3, 128-bit types): Likewise.
17085 (nand<mode>3, 128-bit types): Likewise.
17086 (orc<mode>3, 128-bit types): Likewise.
17087 (and<mode>3_internal): Likewise.
17088 (bool<mode>3_internal): Likewise.
17089 (boolc<mode>3_internal1): Likewise.
17090 (boolc<mode>3_internal2): Likewise.
17091 (boolcc<mode>3_internal1): Likewise.
17092 (boolcc<mode>3_internal2): Likewise.
17093 (eqv<mode>3_internal1): Likewise.
17094 (eqv<mode>3_internal2): Likewise.
17095 (one_cmpl1<mode>3_internal): Likewise.
17097 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
17099 * config/microblaze/microblaze.c (microblaze_expand_prologue):
17100 Rename flag_stack_usage to flag_stack_usage_info.
17102 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
17104 * config/microblaze/sync.md: New file.
17105 * config/microblaze/microblaze.md: Include sync.md
17106 * config/microblaze/microblaze.c: Add print_operand 'y'.
17107 * config/microblaze/constraints.md: Add memory_contraint
17108 'Q' which is a single register.
17110 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
17112 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
17114 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
17116 * reload.c (find_reloads): Exit loop once we find this operand
17117 cannot be reloaded somehow for this alternative.
17119 * reload.c (find_reloads): Exit loop once we find a hard register.
17121 * rtlanal.c (computed_jump_p): Exit loop once we find label
17124 * i386.c (ix86_pad_returns): Exit loop after setting replace.
17126 * cfgloopmanip.c (remove_path): Exit loop after setting
17129 * gensupport.c (subst_dup): Avoid loop if code is not
17130 MATCH_DUP nor MATCH_OP_DUP.
17132 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
17134 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
17136 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
17138 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
17139 true for SP_REGNUM if mode == ptr_mode.
17140 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
17141 with value R0_REGNUM + 31.
17143 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
17145 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
17146 pad pointer-typed argument downward.
17148 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
17150 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
17151 and __ILP32__ when the ILP32 model is in use.
17153 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
17155 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
17156 (aarch64_load_symref_appropriately): In the case of
17157 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
17158 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
17159 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
17160 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
17161 if 'mode' doesn't equal to 'ptr_mode'.
17162 (aarch64_output_mi_thunk): Add an assertion on the alignment of
17163 'vcall_offset'; change to call aarch64_emit_move differently depending
17164 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
17165 to calculate the upper bound of 'vcall_offset'.
17166 (aarch64_cannot_force_const_mem): Change to also return true if
17168 (aarch64_legitimize_reload_address): In the case of large
17169 displacements, add new local variable 'xmode' and an assertion
17170 based on it; change to use 'xmode' to generate the new rtx and
17172 (aarch64_asm_trampoline_template): Change to generate the template
17173 differently depending on TARGET_ILP32 or not; change to use
17174 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
17175 (aarch64_trampoline_size): Removed.
17176 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
17177 and replace immediate literals with it. Change to use 'ptr_mode'
17178 instead of 'DImode' and call convert_memory_address if the mode
17179 of 'fnaddr' doesn't equal to 'ptr_mode'.
17180 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
17182 (aarch64_elf_asm_destructor): Likewise.
17183 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
17184 on TARGET_ILP32 instead of aarch64_trampoline_size.
17185 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
17186 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
17187 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
17188 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
17189 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
17190 (storewb_pair<GPI:mode>_<P:mode>): ... this.
17191 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
17192 depending on the value of 'mode'.
17193 (add_losym_<mode>): New.
17194 (ldr_got_small_<mode>): New, based on ldr_got_small.
17195 (ldr_got_small): Remove.
17196 (ldr_got_small_sidi): New.
17197 * config/aarch64/iterators.md (P): New.
17198 (PTR): Change to 'ptr_mode' in the condition.
17200 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
17202 * config.gcc (aarch64*-*-*): Support --with-abi.
17203 (aarch64*-*-elf): Support --with-multilib-list.
17204 (aarch64*-*-linux*): Likewise.
17205 (supported_defaults): Add abi to aarch64*-*-*.
17206 * configure.ac: Mention AArch64 for --with-multilib-list.
17207 * configure: Re-generated.
17208 * config/aarch64/biarchilp32.h: New file.
17209 * config/aarch64/biarchlp64.h: New file.
17210 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
17212 (MULTILIB_DEFAULTS): Ditto.
17213 (DRIVER_SELF_SPECS): Ditto.
17214 (ASM_SPEC): Update to also substitute -mabi.
17215 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
17216 file whose name depends on -mabi= and -mbig-endian.
17217 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
17219 (POINTER_SIZE): New define.
17220 (POINTERS_EXTEND_UNSIGNED): Ditto.
17221 (enum aarch64_abi_type): New enumeration tag.
17222 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
17223 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
17224 (TARGET_ILP32): New define.
17225 * config/aarch64/aarch64.opt (mabi): New.
17226 (aarch64_abi): New.
17227 (ilp32, lp64): New values for -mabi.
17228 * config/aarch64/t-aarch64 (comma): New define.
17229 (MULTILIB_OPTIONS): Ditto.
17230 (MULTILIB_DIRNAMES): Ditto.
17231 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
17232 * doc/invoke.texi: Document -mabi for AArch64.
17234 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
17236 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
17238 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
17239 Anton Blanchard <anton@au1.ibm.com>
17241 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
17242 endianness when selecting field to splat.
17244 2013-07-22 Eric Christopher <echristo@gmail.com>
17246 * dwarf2out.c (die_odr_checksum): New function to use
17247 CHECKSUM_ macros and ULEB128 for DIE tag.
17248 (generate_type_signature): Use.
17250 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
17252 * config.gcc (sparc*-*-*): Accept leon3 processor.
17253 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
17254 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
17255 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
17256 * config/sparc/sparc.opt (enum processor_type): Add leon3.
17257 (mfix-ut699): Adjust comment.
17258 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
17259 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
17260 (CPP_CPU_SPEC): Likewise.
17261 (ASM_CPU_SPEC): Likewise.
17262 * config/sparc/sparc.c (leon3_cost): New constant.
17263 (sparc_option_override): Add leon3 support.
17264 (mem_ref): New function.
17265 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
17266 (sparc_do_work_around_errata): Look into the instruction in the delay
17267 slot and adjust accordingly. Add fix for the data cache nullify issues
17268 of the UT699. Change insertion position for the NOP.
17269 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
17270 (leon3_load): New reservation.
17271 (leon_store): Bump latency to 2.
17272 (grfpu): New automaton.
17273 (grfpu_alu): New unit.
17274 (grfpu_ds): Likewise.
17275 (leon_fp_alu): Adjust.
17276 (leon_fp_mult): Delete.
17277 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
17278 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
17279 * config/sparc/sparc.md (cpu): Add leon3.
17280 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
17281 (swapsi): Likewise.
17282 (atomic_test_and_set): Likewise.
17283 (ldstub): Likewise.
17285 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
17287 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
17288 default for R5900 targets.
17289 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
17290 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
17291 * config/mips/mips.c (mips_option_override): Report an error for
17292 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
17293 for -march=r5900 -mhard-float.
17295 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
17297 * df-problems.c (can_move_insns_across): Exit loop once we
17298 find a non-fixed, non-global register.
17300 * ipa-pure-const.c (propagate_nothrow): Exit loop after
17303 * omega.c (omega_eliminate_red): Break after setting red_found.
17304 (omega_problem_has_red_equations): Similarly after setting found.
17305 (omega_query_variable): Similarly after setting coupled.
17307 2013-07-22 Marek Polacek <polacek@redhat.com>
17309 * gimplify.c: Don't include gimple.h twice.
17311 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17313 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
17314 instead of TARGET_THUMB1.
17315 (Pz): New constraint.
17316 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
17318 (compare_negsi_si): Likewise.
17319 (compare_addsi2_op0): Likewise.
17320 (compare_addsi2_op1): Likewise.
17321 (addsi3_carryin_<optab>): Likewise.
17322 (addsi3_carryin_alt2_<optab>): Likewise.
17323 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
17324 for arm_restrict_it.
17325 (subsi3_carryin): Likewise.
17326 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
17327 (minmax_arithsi): Disable for arm_restrict_it.
17328 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
17329 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
17330 (satsi_<SAT:code>_shift): Likewise.
17331 (arm_shiftsi3): Add alternative for 16-bit encoding.
17332 (arm32_movhf): Disable for arm_restrict_it.
17333 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
17334 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
17336 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
17338 * config/arm/arm.md (attribute "insn"): Delete.
17339 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
17340 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
17341 (not_shiftsi): Update for attribute change.
17342 (not_shiftsi_compare0): Likewise.
17343 (not_shiftsi_compare0_scratch): Likewise.
17344 (arm_one_cmplsi2): Likewise.
17345 (thumb1_one_cmplsi2): Likewise.
17346 (notsi_compare0): Likewise.
17347 (notsi_compare0_scratch): Likewise.
17348 (thumb1_movdi_insn): Likewise.
17349 (arm_movsi_insn): Likewise.
17350 (movhi_insn_arch4): Likewise.
17351 (movhi_bytes): Likewise.
17352 (arm_movqi_insn): Likewise.
17353 (thumb1_movqi_insn): Likewise.
17354 (arm32_movhf): Likewise.
17355 (thumb1_movhf): Likewise.
17356 (arm_movsf_soft_insn): Likewise.
17357 (thumb1_movsf_insn): Likewise.
17358 (thumb_movdf_insn): Likewise.
17359 (movsicc_insn): Likewise.
17360 (movsfcc_soft_insn): Likewise.
17361 (and_scc): Likewise.
17362 (cond_move): Likewise.
17363 (if_move_not): Likewise.
17364 (if_not_move): Likewise.
17365 (if_shift_move): Likewise.
17366 (if_move_shift): Likewise.
17367 (if_shift_shift): Likewise.
17368 (if_not_arith): Likewise.
17369 (if_arith_not): Likewise.
17370 (cond_move_not): Likewise.
17371 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
17372 (neon_mov<mode>): Likewise.
17373 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
17374 (thumb2_movsi_vfp): Likewise.
17375 (movsf_vfp): Likewise.
17376 (thumb2_movsf_vfp): Likewise.
17377 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
17379 (cortexa7_older_only): Likewise.
17380 (cortexa7_younger): Likewise.
17381 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
17382 (1020alu_shift_op): Likewise.
17383 (1020alu_shift_reg_op): Likewise.
17384 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
17385 (alu_shift_op): Likewise.
17386 (alu_shift_reg_op): Likewise.
17387 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
17388 (11_alu_shift_op): Likewise.
17389 (11_alu_shift_reg_op): Likewise.
17390 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
17391 (9_alu_shift_reg_op): Likewise.
17392 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
17394 (cortex_a15_alu_shift): Likewise.
17395 (cortex_a15_alu_shift_reg): Likewise.
17396 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
17397 (cortex_a5_alu_shift): Likewise.
17398 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
17400 (cortex_a53_alu_shift): Likewise.
17401 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
17403 (cortex_a7_alu_reg): Likewise.
17404 (cortex_a7_alu_shift): Likewise.
17405 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
17406 (cortex_a8_alu_shift): Likewise.
17407 (cortex_a8_alu_shift_reg): Likewise.
17408 (cortex_a8_mov): Likewise.
17409 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
17410 (cortex_a9_dp_shift): Likewise.
17411 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
17412 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
17413 (cortex_r4_mov): Likewise.
17414 (cortex_r4_alu_shift): Likewise.
17415 (cortex_r4_alu_shift_reg): Likewise.
17416 * config/arm/fa526.md (526_alu_op): Update for attribute change.
17417 (526_alu_shift_op): Likewise.
17418 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
17419 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
17420 (626te_alu_shift_op): Likewise.
17421 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
17422 (726te_alu_op): Likewise.
17423 (726te_alu_shift_op): Likewise.
17424 (726te_alu_shift_reg_op): Likewise.
17425 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
17426 (mp626_alu_shift_op): Likewise.
17427 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
17428 (pj4_alu_e1_conds): Likewise.
17429 (pj4_alu): Likewise.
17430 (pj4_alu_conds): Likewise.
17431 (pj4_shift): Likewise.
17432 (pj4_shift_conds): Likewise.
17433 (pj4_alu_shift): Likewise.
17434 (pj4_alu_shift_conds): Likewise.
17436 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17438 * config/arm/predicates.md (shiftable_operator_strict_it):
17440 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
17441 Disable cond_exec version for arm_restrict_it.
17442 (thumb2_smaxsi3): Convert to generate cond_exec.
17443 (thumb2_sminsi3): Likewise.
17444 (thumb32_umaxsi3): Likewise.
17445 (thumb2_uminsi3): Likewise.
17446 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
17447 (thumb2_neg_abssi2): Likewise.
17448 (thumb2_mov_scc): Add alternative for 16-bit encoding.
17449 (thumb2_movsicc_insn): Adjust alternatives.
17450 (thumb2_mov_negscc): Disable for arm_restrict_it.
17451 (thumb2_mov_negscc_strict_it): New pattern.
17452 (thumb2_mov_notscc_strict_it): New pattern.
17453 (thumb2_mov_notscc): Disable for arm_restrict_it.
17454 (thumb2_ior_scc): Likewise.
17455 (thumb2_ior_scc_strict_it): New pattern.
17456 (thumb2_cond_move): Adjust for arm_restrict_it.
17457 (thumb2_cond_arith): Disable for arm_restrict_it.
17458 (thumb2_cond_arith_strict_it): New pattern.
17459 (thumb2_cond_sub): Adjust for arm_restrict_it.
17460 (thumb2_movcond): Likewise.
17461 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
17462 (thumb2_zero_extendhisi2_v6): Likewise.
17463 (thumb2_zero_extendqisi2_v6): Likewise.
17464 (orsi_notsi_si): Likewise.
17465 (orsi_not_shiftsi_si): Likewise.
17467 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
17469 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
17470 instruction sequence is 1 byte shorter.
17472 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
17474 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
17475 it is not needed after split.
17477 2013-07-20 Iain Sandoe <iain@codesourcery.com>
17480 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
17481 second label for nonlocal goto receivers. Don't output pic base labels
17482 unless we're producing PIC; mark that action unreachable().
17483 (ix86_save_reg): If the function contains a nonlocal label, save the
17485 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
17486 * config/darwin.c (emitted_pic_label_num): New GTY.
17487 (update_pic_label_number_if_needed): New.
17488 (machopic_output_function_base_name): Adjust for nonlocal receiver
17490 (machopic_should_output_picbase_label): New.
17491 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
17492 (nonlocal_goto_receiver): New insn and split.
17494 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
17496 * config/aarch64/aarch64-builtins.c
17497 (aarch64_fold_builtin): Fold abs in all modes.
17498 * config/aarch64/aarch64-simd-builtins.def
17499 (abs): Enable for all modes.
17500 * config/aarch64/arm_neon.h
17501 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
17502 (vabs_f64): Add missing intrinsic.
17504 2013-07-19 Ian Bolton <ian.bolton@arm.com>
17506 * config/aarch64/arm_neon.h (vabs_s64): New function
17508 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
17511 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
17512 * config/avr/avr.md (adjust_len): Add `round'.
17513 * config/avr/avr-protos.h (avr_out_round): New prototype.
17514 (avr_out_plus): Add `out_label' argument.
17515 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
17516 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
17517 Handle the case where `insn' is just a pattern.
17518 (avr_out_bitop): Handle the case where `insn' is just a pattern.
17519 (avr_out_round): New function.
17520 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
17522 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
17524 * config/microblaze/microblaze.c (microblaze_expand_prologue):
17525 Add check for flag_stack_usage to handle -fstack-usage support
17527 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
17529 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
17530 interaction for new Power8 flags and VSX.
17532 2013-07-18 Sriraman Tallam <tmsriram@google.com>
17534 PR middle-end/57698
17535 * tree-inline.c (expand_call_inline): Emit errors during
17536 early_inlining only if optimization is not turned on.
17538 2013-07-18 David Malcolm <dmalcolm@redhat.com>
17542 * passes.c (init_optimization_passes): Move the construction of
17543 the pass hierarchy into a new passes.def file.
17545 * Makefile.in (passes.o): Add dependency on passes.def.
17547 2013-07-18 David Malcolm <dmalcolm@redhat.com>
17549 * passes.c (init_optimization_passes): Introduce macros for
17550 constructing the tree of passes (INSERT_PASSES_AFTER,
17551 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
17552 TERMINATE_PASS_LIST).
17554 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
17555 Wei Mi <wmi@google.com>
17557 PR rtl-optimization/57878
17558 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
17560 (reload_pseudo_compare_func): Check nregs first for reload
17563 2013-07-18 David Malcolm <dmalcolm@redhat.com>
17565 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
17567 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
17569 * read-rtl.c (validate_const_int): Once an invalid character is
17570 seen, quit the loop.
17572 * gengtype.c (write_roots): Similarly once we find the "deletable"
17573 or "if_marked" option.
17575 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
17577 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
17578 "xtab" and "sat". Move value "clz" from here to ...
17579 (attriubte "type"): ... here.
17580 (satsi_<SAT:code>): Delete "insn" attribute.
17581 (satsi_<SAT:code>_shift): Likewise.
17582 (arm_zero_extendqisi2addsi): Likewise.
17583 (arm_extendqisi2addsi): Likewise.
17584 (clzsi2): Update for attribute changes.
17585 (rbitsi2): Likewise.
17586 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
17588 (arm_usatsihi): Likewise.
17589 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
17591 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
17593 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
17594 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
17595 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
17596 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
17597 in alphabetical order.
17598 (attribute "core_cycles"): Update for attribute changes.
17599 (arm_addsi3): Likewise.
17600 (addsi3_compare0): Likewise.
17601 (addsi3_compare0_scratch): Likewise.
17602 (addsi3_compare_op1): Likewise.
17603 (addsi3_compare_op2): Likewise.
17604 (compare_addsi2_op0): Likewise.
17605 (compare_addsi2_op1): Likewise.
17606 (addsi3_carryin_shift_<optab>): Likewise.
17607 (subsi3_carryin_shift): Likewise.
17608 (rsbsi3_carryin_shift): Likewise.
17609 (arm_subsi3_insn): Likewise.
17610 (subsi3_compare0): Likewise.
17611 (subsi3_compare): Likewise.
17612 (arm_andsi3_insn): Likewise.
17613 (thumb1_andsi3_insn): Likewise.
17614 (andsi3_compare0): Likewise.
17615 (andsi3_compare0_scratch): Likewise.
17616 (zeroextractsi_compare0_scratch
17617 (andsi_not_shiftsi_si): Likewise.
17618 (iorsi3_insn): Likewise.
17619 (iorsi3_compare0): Likewise.
17620 (iorsi3_compare0_scratch): Likewise.
17621 (arm_xorsi3): Likewise.
17622 (thumb1_xorsi3_insn): Likewise.
17623 (xorsi3_compare0): Likewise.
17624 (xorsi3_compare0_scratch): Likewise.
17625 (satsi_<SAT:code>_shift): Likewise.
17627 (arm_shiftsi3): Likewise.
17628 (shiftsi3_compare0): Likewise.
17629 (not_shiftsi): Likewise.
17630 (not_shiftsi_compare0): Likewise.
17631 (not_shiftsi_compare0_scratch): Likewise.
17632 (arm_one_cmplsi2): Likewise.
17633 (thumb_one_complsi2): Likewise.
17634 (notsi_compare0): Likewise.
17635 (notsi_compare0_scratch): Likewise.
17636 (thumb1_zero_extendhisi2): Likewise.
17637 (arm_zero_extendhisi2): Likewise.
17638 (arm_zero_extendhisi2_v6): Likewise.
17639 (arm_zero_extendhisi2addsi): Likewise.
17640 (thumb1_zero_extendqisi2): Likewise.
17641 (thumb1_zero_extendqisi2_v6): Likewise.
17642 (arm_zero_extendqisi2): Likewise.
17643 (arm_zero_extendqisi2_v6): Likewise.
17644 (arm_zero_extendqisi2addsi): Likewise.
17645 (thumb1_extendhisi2): Likewise.
17646 (arm_extendhisi2): Likewise.
17647 (arm_extendhisi2_v6): Likewise.
17648 (arm_extendqisi): Likewise.
17649 (arm_extendqisi_v6): Likewise.
17650 (arm_extendqisi2addsi): Likewise.
17651 (thumb1_extendqisi2): Likewise.
17652 (thumb1_movdi_insn): Likewise.
17653 (arm_movsi_insn): Likewise.
17654 (movsi_compare0): Likewise.
17655 (movhi_insn_arch4): Likewise.
17656 (movhi_bytes): Likewise.
17657 (arm_movqi_insn): Likewise.
17658 (thumb1_movqi_insn): Likewise.
17659 (arm32_movhf): Likewise.
17660 (thumb1_movhf): Likewise.
17661 (arm_movsf_soft_insn): Likewise.
17662 (thumb1_movsf_insn): Likewise.
17663 (movdf_soft_insn): Likewise.
17664 (thumb_movdf_insn): Likewise.
17665 (arm_cmpsi_insn): Likewise.
17666 (cmpsi_shiftsi): Likewise.
17667 (cmpsi_shiftsi_swp): Likewise.
17668 (arm_cmpsi_negshiftsi_si): Likewise.
17669 (movsicc_insn): Likewise.
17670 (movsfcc_soft_insn): Likewise.
17671 (arith_shiftsi): Likewise.
17672 (arith_shiftsi_compare0
17673 (arith_shiftsi_compare0_scratch
17674 (sub_shiftsi): Likewise.
17675 (sub_shiftsi_compare0
17676 (sub_shiftsi_compare0_scratch
17677 (and_scc): Likewise.
17678 (cond_move): Likewise.
17679 (if_plus_move): Likewise.
17680 (if_move_plus): Likewise.
17681 (if_move_not): Likewise.
17682 (if_not_move): Likewise.
17683 (if_shift_move): Likewise.
17684 (if_move_shift): Likewise.
17685 (if_shift_shift): Likewise.
17686 (if_not_arith): Likewise.
17687 (if_arith_not): Likewise.
17688 (cond_move_not): Likewise.
17689 (thumb1_ashlsi3): Set type attribute.
17690 (thumb1_ashrsi3): Likewise.
17691 (thumb1_lshrsi3): Likewise.
17692 (thumb1_rotrsi3): Likewise.
17693 (shiftsi3_compare0_scratch): Likewise.
17694 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
17695 (neon_mov<mode>): Likewise.
17696 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
17698 (thumb2_movsi_insn): Likewise.
17699 (thumb2_cmpsi_neg_shiftsi): Likewise.
17700 (thumb2_extendqisi_v6): Likewise.
17701 (thumb2_zero_extendhisi2_v6): Likewise.
17702 (thumb2_zero_extendqisi2_v6): Likewise.
17703 (thumb2_shiftsi3_short): Likewise.
17704 (thumb2_addsi3_compare0_scratch): Likewise.
17705 (orsi_not_shiftsi_si): Likewise.
17706 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
17707 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
17709 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
17710 (1020alu_shift_op): Likewise.
17711 (1020alu_shift_reg_op): Likewise.
17712 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
17713 (alu_shift_op): Likewise.
17714 (alu_shift_reg_op): Likewise.
17715 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
17716 (11_alu_shift_op): Likewise.
17717 (11_alu_shift_reg_op): Likewise.
17718 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
17719 (9_alu_shift_reg_op): Likewise.
17720 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
17722 (cortex_a15_alu_shift): Likewise.
17723 (cortex_a15_alu_shift_reg): Likewise.
17724 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
17726 (cortex_a5_alu_shift): Likewise.
17727 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
17729 (cortex_a53_alu_shift): Likewise.
17730 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
17732 (cortex_a7_alu_reg): Likewise.
17733 (cortex_a7_alu_shift): Likewise.
17734 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
17736 (cortex_a8_alu_shift): Likewise.
17737 (cortex_a8_alu_shift_reg): Likewise.
17738 (cortex_a8_mov): Likewise.
17739 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
17740 (cortex_a9_dp_shift): Likewise.
17741 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
17743 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
17745 (cortex_r4_mov): Likewise.
17746 (cortex_r4_alu_shift): Likewise.
17747 (cortex_r4_alu_shift_reg): Likewise.
17748 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
17749 (526_alu_shift_op): Likewise.
17750 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
17751 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
17752 (626te_alu_shift_op): Likewise.
17753 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
17754 (726te_alu_op): Likewise.
17755 (726te_alu_shift_op): Likewise.
17756 (726te_alu_shift_reg_op): Likewise.
17757 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
17758 (mp626_alu_shift_op): Likewise.
17759 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
17760 (pj4_alu_e1_conds): Likewise.
17761 (pj4_alu): Likewise.
17762 (pj4_alu_conds): Likewise.
17763 (pj4_shift): Likewise.
17764 (pj4_shift_conds): Likewise.
17765 (pj4_alu_shift): Likewise.
17766 (pj4_alu_shift_conds): Likewise.
17767 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
17769 (cortexa7_older_only): Likewise.
17770 (cortexa7_younger): Likewise.
17772 2013-07-18 David Malcolm <dmalcolm@redhat.com>
17774 * ipa-pure-const.c (generate_summary): Rename to...
17775 (pure_const_generate_summary): ... this.
17777 2013-07-17 Iain Sandoe <iain@codesourcery.com>
17779 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
17781 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
17784 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
17787 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17789 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
17790 enabled without -march=zEC12.
17791 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
17794 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
17796 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
17797 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
17798 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
17799 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
17800 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
17801 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
17802 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
17804 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
17805 and ISA_HAS_NMADD3_NMSUB3.
17806 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
17807 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
17808 (nmsub4<mode>, nmsub3<mode>): Likewise.
17809 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
17811 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
17813 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
17814 TARGET_MIPS5400 checking.
17816 2013-07-16 Jakub Jelinek <jakub@redhat.com>
17817 Peter Bergner <bergner@vnet.ibm.com>
17819 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
17820 registers in the comment.
17821 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
17822 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
17823 rather than FIRST_PSEUDO_REGISTERS.
17825 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
17827 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
17828 enable extra ISA flags with TARGET_HTM.
17830 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
17832 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
17835 2013-07-15 Cong Hou <congh@google.com>
17837 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
17838 in compare function for sorting.
17840 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
17842 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
17843 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
17844 * config/rs6000/rs6000.opt: Add -mhtm option.
17845 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
17846 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
17847 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
17848 __HTM__ if the HTM instructions are available.
17849 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
17850 htm_spr_reg_operand): New define_predicates.
17851 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
17852 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
17854 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
17855 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
17856 HTM builtin functions.
17857 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
17858 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
17859 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
17860 (rs6000_builtin_mask_calculate): Likewise.
17861 (rs6000_option_override_internal): Likewise.
17862 (bdesc_htm): Add new HTM builtin support.
17863 (htm_spr_num): New function.
17864 (htm_spr_regno): Likewise.
17865 (rs6000_htm_spr_icode): Likewise.
17866 (htm_expand_builtin): Likewise.
17867 (htm_init_builtins): Likewise.
17868 (rs6000_expand_builtin): Add support for HTM builtin functions.
17869 (rs6000_init_builtins): Likewise.
17870 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
17872 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
17873 (TARGET_HTM, MASK_HTM): Define macros.
17874 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
17875 (FIXED_REGISTERS): Likewise.
17876 (CALL_USED_REGISTERS): Likewise.
17877 (CALL_REALLY_USED_REGISTERS): Likewise.
17878 (REG_ALLOC_ORDER): Likewise.
17879 (enum reg_class): Likewise.
17880 (REG_CLASS_NAMES): Likewise.
17881 (REG_CLASS_CONTENTS): Likewise.
17882 (REGISTER_NAMES): Likewise.
17883 (ADDITIONAL_REGISTER_NAMES): Likewise.
17884 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
17885 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
17886 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
17887 * config/rs6000/htm.md: New file.
17888 * config/rs6000/htmintrin.h: New file.
17889 * config/rs6000/htmxlintrin.h: New file.
17891 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
17893 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
17894 Define SYMBOL_TINY_GOT, update comment.
17895 * config/aarch64/aarch64.c
17896 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
17897 (aarch64_expand_mov_immediate): Likewise.
17898 (aarch64_print_operand): Likewise.
17899 (aarch64_classify_symbol): Likewise.
17900 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
17901 (ldr_got_tiny): Define.
17903 2013-07-13 Tobias Grosser <tobias@grosser.es>
17905 PR tree-optimization/54094
17906 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
17907 scheduling dimension for the parallelism check from the polyhedral
17908 information in the AST.
17909 * graphite-dependences.c (carries_deps): Do not assume the schedule is
17912 2013-07-13 Jason Merrill <jason@redhat.com>
17914 * print-tree.c (debug_vec_tree): Use debug_raw.
17915 (debug_raw (vec<tree, va_gc> &)): New.
17916 (debug_raw (vec<tree, va_gc> *)): New.
17917 * tree.h: Declare them.
17919 2013-07-13 Bin Cheng <bin.cheng@arm.com>
17921 * ifcvt.c (ifcvt_after_combine): New static variable.
17922 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
17924 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
17925 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
17926 rest_of_handle_if_after_reload): Pass new argument for if_convert.
17928 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
17930 * config/mips/mips.c (mips_expand_call): Remove empty statement.
17932 2013-07-12 Michael Matz <matz@suse.de>
17934 PR middle-end/55771
17935 * convert.c (convert_to_real): Reject non-float inner types.
17937 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
17939 * config/aarch64/aarch64-protos.h
17940 (aarch64_simd_immediate_valid_for_move): Remove.
17941 * config/aarch64/aarch64.c (simd_immediate_info): New member.
17942 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
17944 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
17946 2013-07-11 Steve Ellcey <sellcey@mips.com>
17948 * config/mips/mips.c (mips_conditional_register_usage): Do not
17949 use t[0-7] registers in MIPS16 mode when optimizing for size.
17951 2013-07-11 Sriraman Tallam <tmsriram@google.com>
17953 * config/i386/i386.c (dispatch_function_versions): Fix array
17954 indexing of function_version_info to match actual_versions.
17956 2013-07-11 Teresa Johnson <tejohnson@google.com>
17958 * vec.h (struct va_gc): Move release out-of-line.
17959 (va_gc::release): Call ggc_free on released vec.
17961 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
17963 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
17964 Require GOT register as additional operand in UNSPEC.
17965 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
17966 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
17967 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
17968 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
17969 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
17970 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
17971 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
17973 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
17976 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
17977 name seen by assembler/linker rather if available.
17979 2013-07-11 Andreas Schwab <schwab@suse.de>
17981 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
17983 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
17985 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
17987 2013-07-10 Joseph Myers <joseph@codesourcery.com>
17989 * doc/tm.texi.in: Move hook documentation to ....
17990 * target.def: ... here.
17992 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
17993 text on @hook line.
17994 * doc/tm.texi: Regenerate.
17996 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
17999 * doc/invoke.texi: Document -Wconditionally-supported.
18001 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
18004 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
18007 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
18010 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
18011 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
18012 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
18013 Remove duplicate devices.
18014 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
18015 * config/avr/t-multilib: Regenerate.
18016 * config/avr/avr-tables.opt: Regenerate.
18017 * doc/avr-mmcu.texi: Regenerate.
18019 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
18022 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
18024 2013-07-10 Graham Stott <graham.stott@btinternet.com>
18026 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
18027 the cost of MULT when optimizing for size.
18029 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
18031 * config/cr16/cr16-protos.h: Don't include target.h.
18033 2013-07-09 Joseph Myers <joseph@codesourcery.com>
18035 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
18036 adjust register size for TDmode and TFmode for VSX registers.
18038 2013-07-08 Kai Tietz <ktietz@redhat.com>
18041 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
18042 hook_bool_const_tree_true.
18044 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18046 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
18047 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
18048 * config/s390/s390.md: Define FPR*_REGNUM constants.
18049 Fix FPR2_REGNUM constant (18 -> 17).
18050 ("*trunc<BFP:mode><DFP_ALL:mode>2")
18051 ("*trunc<DFP_ALL:mode><BFP:mode>2")
18052 ("trunc<BFP:mode><DFP_ALL:mode>2")
18053 ("trunc<DFP_ALL:mode><BFP:mode>2")
18054 ("*extend<BFP:mode><DFP_ALL:mode>2")
18055 ("*extend<DFP_ALL:mode><BFP:mode>2")
18056 ("extend<BFP:mode><DFP_ALL:mode>2")
18057 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
18060 2013-07-08 Graham Stott <graham.stott@btinternet.com>
18062 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
18064 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18066 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
18067 and cfun_fpr_bit_p to cfun_fpr_save_p.
18068 (s390_frame_area, s390_register_info, s390_frame_info)
18069 (s390_emit_prologue, s390_emit_epilogue)
18070 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
18072 * config/s390/s390.h: Define *_REGNUM macros for floating point
18075 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
18077 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
18079 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
18081 PR rtl-optimization/57786
18082 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
18083 and break out of the loop when it is set to false.
18085 2013-07-08 Jakub Jelinek <jakub@redhat.com>
18088 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
18089 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
18090 (const_int 63)) 0)).
18091 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
18092 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
18093 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
18095 PR rtl-optimization/57829
18096 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
18097 mask bits outside of mode are just sign-extension from mode to HWI.
18099 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
18101 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
18102 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
18103 adjust_address instead of change_address to keep info about alignment.
18104 (emit_strmov): Remove.
18105 (emit_memmov): New function.
18106 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
18107 (expand_movmem_epilogue): Likewise and return updated rtx for
18109 (expand_constant_movmem_prologue): Likewise and return updated rtx for
18110 destination and source.
18111 (decide_alignment): Refactor, handle vector_loop.
18112 (ix86_expand_movmem): Likewise.
18113 (ix86_expand_setmem): Likewise.
18114 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
18116 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
18118 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
18119 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
18121 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
18123 * config/i386/sse.md (sse_movlhps): Change alternative 3
18124 of operand 2 to "m".
18126 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
18129 * config/i386/sse.md (iptr): New mode attribute.
18130 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
18131 (<sse>_vm<plusminus_insn><mode>3): Ditto.
18132 (<sse>_vmmul<mode>3): Ditto.
18133 (<sse>_vmdiv<mode>3): Ditto.
18134 (sse_vmrcpv4sf2): Ditto.
18135 (<sse>_vmsqrt<mode>2): Ditto.
18136 (sse_vmrsqrtv4sf2): Ditto.
18137 (<sse>_vm<code><mode>3): Ditto.
18138 (avx_vmcmp<mode>3): Ditto.
18139 (<sse>_vmmaskcmp<mode>3): Ditto.
18140 (<sse>_comi): Ditto.
18141 (<sse>_ucomi): Ditto.
18142 (*xop_vmfrcz_<mode>): Ditto.
18143 (*fmai_fmadd_<mode>): Ditto.
18144 (*fmai_fmsub_<mode>): Ditto.
18145 (*fmai_fnmadd_<mode>): Ditto.
18146 (*fmai_fnmsub_<mode>): Ditto.
18147 (*fma4i_vmfmadd_<mode>): Ditto.
18148 (*fma4i_vmfmsub_<mode>): Ditto.
18149 (*fma4i_vmfnmadd_<mode>): Ditto.
18150 (*fma4i_vmfnmsub_<mode>): Ditto.
18151 (*xop_vmfrcz_<mode>): Ditto.
18152 (sse_cvtps2pi): Ditto.
18153 (sse_cvttps2pi): Ditto.
18154 (sse_cvtss2si): Ditto.
18155 (sse_cvtss2si_2): Ditto.
18156 (sse_cvtss2siq_2): Ditto.
18157 (sse_cvttss2si): Ditto.
18158 (sse_cvttss2siq): Ditto.
18159 (sse_cvtsd2si): Ditto.
18160 (sse_cvtsd2si_2): Ditto.
18161 (sse_cvtsd2siq_2): Ditto.
18162 (sse_cvttsd2si): Ditto.
18163 (sse_cvttsd2siq): Ditto.
18164 (sse_cvtsd2ss): Ditto.
18165 (sse_cvtss2sd): Ditto.
18166 (avx2_pbroadcast<mode>): Ditto.
18167 (avx2_pbroadcast<mode>_1): Ditto.
18168 (*avx_vperm_broadcast_v4sf): Ditto.
18170 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
18171 (sse_movlhps): Ditto.
18172 (sse_storehps): Ditto.
18173 (sse_loadhps): Ditto.
18174 (sse_storelps): Ditto.
18175 (sse_loadlps): Ditto.
18176 (*vec_concatv4sf): Ditto.
18177 (*vec_interleave_highv2df): Ditto.
18178 (*vec_interleave_lowv2df): Ditto.
18179 (*vec_extractv2df_1_sse): Ditto.
18180 (*vec_extractv2df_0_sse): Ditto.
18181 (sse2_storelpd): Ditto.
18182 (sse2_loadlpd): Ditto.
18183 (sse2_movsd): Ditto.
18184 (*vec_concatv4si): Ditto.
18185 (vec_concatv2di): Ditto.
18187 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
18188 for Intel asm dialect.
18189 (mmx_punpcklwd): Ditto.
18190 (mmx_punpckldq): Ditto.
18192 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
18193 for intel assembler dialect.
18195 2013-07-06 Jakub Jelinek <jakub@redhat.com>
18198 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
18199 for BUILT_IN_C{LZ,LRSB}*.
18200 * tree.h (CASE_INT_FN): Add FN##IMAX case.
18201 * tree-vrp.c (extract_range_basic): Handle
18202 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
18203 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
18204 fall thru to code calling set_value*.
18205 * builtins.c (expand_builtin): Remove *IMAX cases.
18206 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
18207 if width is bigger than 2*HWI.
18209 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
18211 PR rtl-optimization/55342
18212 * lra-int.h (lra_subreg_reload_pseudos): New.
18213 * lra.c: Add undoing optional reloads to the block diagram.
18214 (lra_subreg_reload_pseudos): New.
18215 (lra_optional_reload_pseudos): Change comments.
18216 (lra): Init and clear lra_subreg_reload_pseudos. Clear
18217 lra_optional_reload_pseudos after undo transformations.
18218 * lra-assigns.c (pseudo_prefix_title): New.
18219 (lra_setup_reg_renumber): Use it.
18220 (spill_for): Ditto. Check subreg reload pseudos too.
18221 (assign_by_spills): Consider subreg reload pseudos too.
18222 * lra-constraints.c (simplify_operand_subreg): Use
18223 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
18224 (curr_insn_transform): Recognize and do optional reloads.
18225 (undo_optional_reloads): New.
18226 (lra_undo_inheritance): Call undo_optional_reloads.
18228 2013-07-05 Thomas Quinot <quinot@adacore.com>
18230 * tree-complex.c (expand_complex_operations_1): Fix typo.
18232 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
18234 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
18235 (tune_params): New member 'const vec_costs'.
18236 * config/aarch64/aarch64.c (generic_vector_cost): New.
18237 (generic_tunings): New member 'generic_vector_cost'.
18238 (aarch64_builtin_vectorization_cost): New.
18239 (aarch64_add_stmt_cost): New.
18240 (TARGET_VECTORIZE_ADD_STMT_COST): New.
18241 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
18243 2013-07-03 Jakub Jelinek <jakub@redhat.com>
18246 * config/i386/predicates.md (vsib_address_operand): Disallow
18247 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
18249 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
18251 PR middle-end/55030
18252 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
18253 expand_builtin_setjmp_receiver.
18254 (expand_label): Adjust, call expand_builtin_setjmp_receiver
18255 with NULL for the label parameter.
18256 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
18257 the frame-pointer. Adjust comments.
18258 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
18259 only if LABEL is non-NULL.
18261 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
18263 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
18264 (ARM_ABI_AAPCS64): Ditto.
18266 (ARM_DEFAULT_ABI): Ditto.
18268 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
18270 * config/aarch64/aarch64-builtins.c
18271 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
18272 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
18274 * config/aarch64/aarch64-simd.md
18275 (aarch64_ld1<VALL:mode>): New.
18276 (aarch64_st1<VALL:mode>): Likewise.
18277 * config/aarch64/arm_neon.h
18278 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
18280 2013-07-02 Sriraman Tallam <tmsriram@google.com>
18282 * config/i386/i386.c (gate_insert_vzeroupper): Check if
18284 (ix86_option_override_internal):Turn on all -mavx target flags by
18285 default as they are dependent on AVX anyway.
18287 2013-07-02 Cary Coutant <ccoutant@google.com>
18289 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
18290 deterministic hash.
18291 (loc_checksum_ordered): Likewise.
18292 (hash_loc_operands): Remove inline keyword.
18294 2013-07-02 Jakub Jelinek <jakub@redhat.com>
18296 PR tree-optimization/57741
18297 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
18298 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
18299 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
18300 Allow REAL_CST step_exprs if flag_associative_math.
18301 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
18303 2013-07-02 Ian Bolton <ian.bolton@arm.com>
18305 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
18307 2013-07-02 Ian Bolton <ian.bolton@arm.com>
18309 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
18311 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18313 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
18315 (iorsi3_insn): Likewise.
18316 (arm_xorsi3): Likewise.
18318 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
18320 * arm.md (attribute "wtype"): Delete. Move attribute values from here
18322 (attribute "type"): ... here, and prefix with "wmmx_".
18323 (attribute "core_cycles"): Update for attribute changes.
18324 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
18325 (tbcstv4hi): Likewise.
18326 (tbcstv2si): Likewise.
18327 (iwmmxt_iordi3): Likewise.
18328 (iwmmxt_xordi3): Likewise.
18329 (iwmmxt_anddi3): Likewise.
18330 (iwmmxt_nanddi3): Likewise.
18331 (iwmmxt_arm_movdi): Likewise.
18332 (iwmmxt_movsi_insn): Likewise.
18333 (mov<mode>_internal): Likewise.
18334 (and<mode>3_iwmmxt): Likewise.
18335 (ior<mode>3_iwmmxt): Likewise.
18336 (xor<mode>3_iwmmxt): Likewise.
18337 (add<mode>3_iwmmxt): Likewise.
18338 (ssaddv8qi3): Likewise.
18339 (ssaddv4hi3): Likewise.
18340 (ssaddv2si3): Likewise.
18341 (usaddv8qi3): Likewise.
18342 (usaddv4hi3): Likewise.
18343 (usaddv2si3): Likewise.
18344 (sub<mode>3_iwmmxt): Likewise.
18345 (sssubv8qi3): Likewise.
18346 (sssubv4hi3): Likewise.
18347 (sssubv2si3): Likewise.
18348 (ussubv8qi3): Likewise.
18349 (ussubv4hi3): Likewise.
18350 (ussubv2si3): Likewise.
18351 (mulv4hi3_iwmmxt): Likewise.
18352 (smulv4hi3_highpart): Likewise.
18353 (umulv4hi3_highpart): Likewise.
18354 (iwmmxt_wmacs): Likewise.
18355 (iwmmxt_wmacsz): Likewise.
18356 (iwmmxt_wmacu): Likewise.
18357 (iwmmxt_wmacuz): Likewise.
18358 (iwmmxt_clrdi): Likewise.
18359 (iwmmxt_clrv8qi): Likewise.
18360 (iwmmxt_clr4hi): Likewise.
18361 (iwmmxt_clr2si): Likewise.
18362 (iwmmxt_uavgrndv8qi3): Likewise.
18363 (iwmmxt_uavgrndv4hi3): Likewise.
18364 (iwmmxt_uavgv8qi3): Likewise.
18365 (iwmmxt_uavgv4hi3): Likewise.
18366 (iwmmxt_tinsrb): Likewise.
18367 (iwmmxt_tinsrh): Likewise.
18368 (iwmmxt_tinsrw): Likewise.
18369 (iwmmxt_textrmub): Likewise.
18370 (iwmmxt_textrmsb): Likewise.
18371 (iwmmxt_textrmuh): Likewise.
18372 (iwmmxt_textrmsh): Likewise.
18373 (iwmmxt_textrmw): Likewise.
18374 (iwmxxt_wshufh): Likewise.
18375 (eqv8qi3): Likewise.
18376 (eqv4hi3): Likewise.
18377 (eqv2si3): Likewise.
18378 (gtuv8qi3): Likewise.
18379 (gtuv4hi3): Likewise.
18380 (gtuv2si3): Likewise.
18381 (gtv8qi3): Likewise.
18382 (gtv4hi3): Likewise.
18383 (gtv2si3): Likewise.
18384 (smax<mode>3_iwmmxt): Likewise.
18385 (umax<mode>3_iwmmxt): Likewise.
18386 (smin<mode>3_iwmmxt): Likewise.
18387 (umin<mode>3_iwmmxt): Likewise.
18388 (iwmmxt_wpackhss): Likewise.
18389 (iwmmxt_wpackwss): Likewise.
18390 (iwmmxt_wpackdss): Likewise.
18391 (iwmmxt_wpackhus): Likewise.
18392 (iwmmxt_wpackwus): Likewise.
18393 (iwmmxt_wpackdus): Likewise.
18394 (iwmmxt_wunpckihb): Likewise.
18395 (iwmmxt_wunpckihh): Likewise.
18396 (iwmmxt_wunpckihw): Likewise.
18397 (iwmmxt_wunpckilb): Likewise.
18398 (iwmmxt_wunpckilh): Likewise.
18399 (iwmmxt_wunpckilw): Likewise.
18400 (iwmmxt_wunpckehub): Likewise.
18401 (iwmmxt_wunpckehuh): Likewise.
18402 (iwmmxt_wunpckehuw): Likewise.
18403 (iwmmxt_wunpckehsb): Likewise.
18404 (iwmmxt_wunpckehsh): Likewise.
18405 (iwmmxt_wunpckehsw): Likewise.
18406 (iwmmxt_wunpckelub): Likewise.
18407 (iwmmxt_wunpckeluh): Likewise.
18408 (iwmmxt_wunpckeluw): Likewise.
18409 (iwmmxt_wunpckelsb): Likewise.
18410 (iwmmxt_wunpckelsh): Likewise.
18411 (iwmmxt_wunpckelsw): Likewise.
18412 (ror<mode>3): Likewise.
18413 (ashr<mode>3_iwmmxt): Likewise.
18414 (lshr<mode>3_iwmmxt): Likewise.
18415 (ashl<mode>3_iwmmxt): Likewise.
18416 (ror<mode>3_di): Likewise.
18417 (ashr<mode>3_di): Likewise.
18418 (lshr<mode>3_di): Likewise.
18419 (ashl<mode>3_di): Likewise.
18420 (iwmmxt_wmadds): Likewise.
18421 (iwmmxt_wmaddu): Likewise.
18422 (iwmmxt_tmia): Likewise.
18423 (iwmmxt_tmiaph): Likewise.
18424 (iwmmxt_tmiabb): Likewise.
18425 (iwmmxt_tmiatb): Likewise.
18426 (iwmmxt_tmiabt): Likewise.
18427 (iwmmxt_tmiatt): Likewise.
18428 (iwmmxt_tmovmskb): Likewise.
18429 (iwmmxt_tmovmskh): Likewise.
18430 (iwmmxt_tmovmskw): Likewise.
18431 (iwmmxt_waccb): Likewise.
18432 (iwmmxt_wacch): Likewise.
18433 (iwmmxt_waccw): Likewise.
18434 (iwmmxt_waligni): Likewise.
18435 (iwmmxt_walignr): Likewise.
18436 (iwmmxt_walignr0): Likewise.
18437 (iwmmxt_walignr1): Likewise.
18438 (iwmmxt_walignr2): Likewise.
18439 (iwmmxt_walignr3): Likewise.
18440 (iwmmxt_wsadb): Likewise.
18441 (iwmmxt_wsadh): Likewise.
18442 (iwmmxt_wsadbz): Likewise.
18443 (iwmmxt_wsadhz): Likewise.
18444 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
18445 (iwmmxt_wabsdiffb): Likewise.
18446 (iwmmxt_wabsdiffh): Likewise.
18447 (iwmmxt_wabsdiffw): Likewise.
18448 (iwmmxt_waddsubhx): Likewise
18449 (iwmmxt_wsubaddhx): Likewise.
18450 (addc<mode>3): Likewise.
18451 (iwmmxt_avg4): Likewise.
18452 (iwmmxt_avg4r): Likewise.
18453 (iwmmxt_wmaddsx): Likewise.
18454 (iwmmxt_wmaddux): Likewise.
18455 (iwmmxt_wmaddsn): Likewise.
18456 (iwmmxt_wmaddun): Likewise.
18457 (iwmmxt_wmulwsm): Likewise.
18458 (iwmmxt_wmulwum): Likewise.
18459 (iwmmxt_wmulsmr): Likewise.
18460 (iwmmxt_wmulumr): Likewise.
18461 (iwmmxt_wmulwsmr): Likewise.
18462 (iwmmxt_wmulwumr): Likewise.
18463 (iwmmxt_wmulwl): Likewise.
18464 (iwmmxt_wqmulm): Likewise.
18465 (iwmmxt_wqmulwm): Likewise.
18466 (iwmmxt_wqmulmr): Likewise.
18467 (iwmmxt_wqmulwmr): Likewise.
18468 (iwmmxt_waddbhusm): Likewise.
18469 (iwmmxt_waddbhusl): Likewise.
18470 (iwmmxt_wqmiabb): Likewise.
18471 (iwmmxt_wqmiabt): Likewise.
18472 (iwmmxt_wqmiatb): Likewise.
18473 (iwmmxt_wqmiatt): Likewise.
18474 (iwmmxt_wqmiabbn): Likewise.
18475 (iwmmxt_wqmiabtn): Likewise.
18476 (iwmmxt_wqmiatbn): Likewise.
18477 (iwmmxt_wqmiattn): Likewise.
18478 (iwmmxt_wmiabb): Likewise.
18479 (iwmmxt_wmiabt): Likewise.
18480 (iwmmxt_wmiatb): Likewise.
18481 (iwmmxt_wmiatt): Likewise.
18482 (iwmmxt_wmiabbn): Likewise.
18483 (iwmmxt_wmiabtn): Likewise.
18484 (iwmmxt_wmiatbn): Likewise.
18485 (iwmmxt_wmiattn): Likewise.
18486 (iwmmxt_wmiawbb): Likewise.
18487 (iwmmxt_wmiawbt): Likewise.
18488 (iwmmxt_wmiawtb): Likewise.
18489 (iwmmxt_wmiawtt): Likewise.
18490 (iwmmxt_wmiawbbn): Likewise.
18491 (iwmmxt_wmiawbtn): Likewise.
18492 (iwmmxt_wmiawtbn): Likewise.
18493 (iwmmxt_wmiawttn): Likewise.
18494 (iwmmxt_wmerge): Likewise.
18495 (iwmmxt_tandc<mode>3): Likewise.
18496 (iwmmxt_torc<mode>3): Likewise.
18497 (iwmmxt_torvsc<mode>3): Likewise.
18498 (iwmmxt_textrc<mode>3): Likewise.
18499 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
18500 (wmmxt_pack): Likewise.
18501 (wmmxt_mult_c1): Likewise.
18502 (wmmxt_mult_c2): Likewise.
18503 (wmmxt_alu_c1): Likewise.
18504 (wmmxt_alu_c2): Likewise.
18505 (wmmxt_alu_c3): Likewise.
18506 (wmmxt_transfer_c1): Likewise.
18507 (wmmxt_transfer_c2): Likewise.
18508 (wmmxt_transfer_c3): Likewise.
18509 (marvell_f_iwmmxt_wstr): Likewise.
18510 (marvell_f_iwmmxt_wldr): Likewise.
18512 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
18514 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
18516 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
18519 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
18520 * lra-constraints.c (need_for_split_p): Check call used hard regs
18521 living through calls.
18523 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
18524 call used regs for call insn.
18526 2013-06-28 Jakub Jelinek <jakub@redhat.com>
18529 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
18530 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
18532 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
18534 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
18536 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
18538 * lra-constraints.c (need_for_split_p): Check call used hard regs
18539 living through calls.
18541 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
18544 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
18545 to tie with any other modes. Eliminate Altivec vector mode tests,
18546 since these are a subset of ALTIVEC or VSX vector modes. Simplify
18547 code, to return 0 if testing MODE2 for a condition, if we've
18548 already tested MODE1 for the same condition.
18550 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
18552 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
18555 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
18557 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
18558 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
18560 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
18562 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
18564 (aarch64_symbolic_constant_p): Remove.
18565 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
18566 static. Fix line length and white space.
18567 (aarch64_symbolic_constant_p): Remove.
18568 * config/aarch64/predicates.md (aarch64_valid_symref):
18569 Use aarch64_classify_symbol_expression.
18571 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18573 * config/arm/constraints.md (Ts): New constraint.
18574 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
18576 (compare_scc): Use "Ts" constraint for operand 0.
18577 (ior_scc_scc): Likewise.
18578 (and_scc_scc): Likewise.
18579 (and_scc_scc_nodom): Likewise.
18580 (ior_scc_scc_cmp): Likewise for operand 7.
18581 (and_scc_scc_cmp): Likewise.
18582 * config/arm/thumb2.md (thumb2_movsi_insn):
18583 Add alternatives for 16-bit encodings.
18584 (thumb2_movhi_insn): Likewise.
18585 (thumb2_movsicc_insn): Likewise.
18586 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
18587 (thumb2_negscc): Use "Ts" constraint.
18588 Move mvn instruction outside cond_exec block.
18589 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
18590 for 16-bit encodings.
18592 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18594 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
18596 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
18597 (mulsi3subsi): Likewise.
18598 (mulsidi3adddi): Likewise.
18599 (mulsidi3_v6): Likewise.
18600 (umulsidi3_v6): Likewise.
18601 (umulsidi3adddi_v6): Likewise.
18602 (smulsi3_highpart_v6): Likewise.
18603 (umulsi3_highpart_v6): Likewise.
18604 (mulhisi3tb): Likewise.
18605 (mulhisi3bt): Likewise.
18606 (mulhisi3tt): Likewise.
18607 (maddhisi4): Likewise.
18608 (maddhisi4tb): Likewise.
18609 (maddhisi4tt): Likewise.
18610 (maddhidi4): Likewise.
18611 (maddhidi4tb): Likewise.
18612 (maddhidi4tt): Likewise.
18613 (zeroextractsi_compare0_scratch): Likewise.
18614 (insv_zero): Likewise.
18615 (insv_t2): Likewise.
18616 (anddi_notzesidi_di): Likewise.
18617 (anddi_notsesidi_di): Likewise.
18618 (andsi_notsi_si): Likewise.
18619 (iordi_zesidi_di): Likewise.
18620 (xordi_zesidi_di): Likewise.
18621 (andsi_iorsi3_notsi): Likewise.
18622 (smax_0): Likewise.
18623 (smax_m1): Likewise.
18624 (smin_0): Likewise.
18625 (not_shiftsi): Likewise.
18626 (unaligned_loadsi): Likewise.
18627 (unaligned_loadhis): Likewise.
18628 (unaligned_loadhiu): Likewise.
18629 (unaligned_storesi): Likewise.
18630 (unaligned_storehi): Likewise.
18631 (extv_reg): Likewise.
18632 (extzv_t2): Likewise.
18633 (divsi3): Likewise.
18634 (udivsi3): Likewise.
18635 (arm_zero_extendhisi2addsi): Likewise.
18636 (arm_zero_extendqisi2addsi): Likewise.
18637 (compareqi_eq0): Likewise.
18638 (arm_extendhisi2_v6): Likewise.
18639 (arm_extendqisi2addsi): Likewise.
18640 (arm_movt): Likewise.
18641 (thumb2_ldrd): Likewise.
18642 (thumb2_ldrd_base): Likewise.
18643 (thumb2_ldrd_base_neg): Likewise.
18644 (thumb2_strd): Likewise.
18645 (thumb2_strd_base): Likewise.
18646 (thumb2_strd_base_neg): Likewise.
18647 (arm_negsi2): Add alternative for 16-bit encoding.
18648 (arm_one_cmplsi2): Likewise.
18650 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18652 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
18653 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
18654 (movdfcc): Likewise.
18655 * config/arm/vfp.md (*thumb2_movsf_vfp):
18656 Disable predication for arm_restrict_it.
18657 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
18658 (*thumb2_movdfcc_vfp): Likewise.
18659 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
18660 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
18661 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
18662 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
18663 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
18664 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
18665 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
18666 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
18667 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
18668 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
18669 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
18670 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
18671 Disable predication for arm_restrict_it.
18673 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
18675 * config/i386/bmiintrin.h (_bextr_u32): New.
18676 (_bextr_u64): Ditto.
18678 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
18680 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
18681 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
18682 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
18683 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
18684 * config/mips/n32-elf.h: ...this new file.
18686 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
18689 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
18690 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
18692 2013-06-27 Catherine Moore <clm@codesourcery.com>
18694 * config/mips/mips-tables.opt: Regenerate.
18695 * config/mips/mips-cpus.def: Add m14ke and m14kec.
18696 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
18697 * doc/invoke.texi: Add -m14kc.
18699 2013-06-27 Jakub Jelinek <jakub@redhat.com>
18702 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
18703 constraints of operand 1 and 2.
18706 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
18707 to match RTL canonicalization. Swap predicates and
18708 constraints of operand 1 and 2.
18710 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
18712 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
18713 Process OP_INOUT regs for splitting too.
18715 2013-06-27 Jakub Jelinek <jakub@redhat.com>
18717 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
18718 decl before the loop, initialize to NULL.
18719 (vectorizable_load): Initialize ptr_incr to NULL.
18721 2013-06-27 Martin Jambor <mjambor@suse.cz>
18724 * ipa-ref.h (ipa_maybe_record_reference): Declare.
18725 * ipa-ref.c (ipa_maybe_record_reference): New function.
18726 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
18727 * ipa-cp.c (create_specialized_node): Record potential references from
18729 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
18731 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
18733 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
18734 parameter 'mode' of type 'enum machine_mode mode'; change to pass
18735 'mode' to force_reg.
18736 (aarch64_add_offset): Update calls to aarch64_force_temporary.
18737 (aarch64_expand_mov_immediate): Likewise.
18739 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
18741 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
18742 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
18744 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18746 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
18747 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
18748 (struct machine_function): Add tbegin_p.
18749 (s390_canonicalize_comparison): Fold CC mode compares to
18750 conditional jump if possible.
18751 (s390_emit_jump): Return the emitted jump.
18752 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
18753 Handle CCRAWmode compares.
18754 (s390_option_override): Default to -mhtm if available.
18755 (s390_reg_clobbered_rtx): Handle floating point regs as well.
18756 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
18757 FPRs instead of df_regs_ever_live_p.
18758 (s390_optimize_nonescaping_tx): New function.
18759 (s390_init_frame_layout): Extend clobbered_regs array to cover
18761 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
18762 (s390_expand_tbegin): New function.
18763 (enum s390_builtin): New enum definition.
18764 (code_for_builtin): New array definition.
18765 (s390_init_builtins): New function.
18766 (s390_expand_builtin): New function.
18767 (TARGET_INIT_BUILTINS): Define.
18768 (TARGET_EXPAND_BUILTIN): Define.
18769 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
18770 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
18771 (s390_alc_comparison): Likewise.
18772 * config/s390/s390-modes.def: Add CCRAWmode.
18773 * config/s390/s390.h (processor_flags): Add PF_TX.
18774 (TARGET_CPU_HTM): Define macro.
18775 (TARGET_HTM): Define macro.
18776 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
18777 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
18778 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
18779 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
18780 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
18781 (TBEGIN_MASK, TBEGINC_MASK): New constants.
18782 ("*cc_to_int"): Move up.
18783 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
18784 constants other than 0.
18785 ("*ccraw_to_int"): New insn and splitter definition.
18786 ("tbegin", "tbegin_nofloat", "tbegin_retry")
18787 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
18788 ("tx_assist"): New expander.
18789 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
18790 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
18791 * config/s390/s390.opt: Add -mhtm option.
18792 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
18793 * config/s390/htmxlintrin.h: New file.
18794 * config/s390/htmintrin.h: New file.
18795 * config/s390/s390intrin.h: New file.
18796 * doc/extend.texi: Document htm builtins.
18797 * config.gcc: Add the new header files to extra_headers.
18799 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
18801 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
18802 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
18804 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
18805 Pat Haugen <pthaugen@us.ibm.com>
18806 Peter Bergner <bergner@vnet.ibm.com>
18808 * config/rs6000/power8.md: New.
18809 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
18810 setting for power8 entry.
18811 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
18812 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
18813 test for Power4/Power5 only.
18814 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
18816 (force_new_group): Adjust comment.
18817 * config/rs6000/rs6000.md: Include power8.md.
18819 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
18821 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
18822 * config/arm/arm-protos.h (arm_max_conditional_execute): New
18824 (tune_params): Update comment.
18825 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
18826 (arm_max_conditional_execute): New function.
18827 (thumb2_final_prescan_insn): Use max_insn_skipped and
18828 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
18830 2013-06-25 Jakub Jelinek <jakub@redhat.com>
18832 PR tree-optimization/57705
18833 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
18834 SSA_NAME step, provided that it is not defined inside the loop.
18835 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
18836 (get_initial_def_for_induction): Handle SSA_NAME IV step.
18838 2013-06-25 Martin Jambor <mjambor@suse.cz>
18840 PR middle-end/57670
18841 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
18842 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
18844 (ipa_note_param_call): Initialize member_ptr flag.
18845 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
18846 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
18847 (ipa_write_indirect_edge_info): Stream member_ptr flag.
18848 (ipa_read_indirect_edge_info): Likewise.
18850 2013-06-25 Richard Biener <rguenther@suse.de>
18852 PR middle-end/56977
18853 * passes.c (init_optimization_passes): Move pass_fold_builtins
18854 and pass_dce earlier with -Og.
18856 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
18858 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
18859 <BIT_FIELD_REF>: Remove trailing TAB.
18860 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
18863 2013-06-24 Martin Jambor <mjambor@suse.cz>
18865 PR tree-optimization/57358
18866 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
18867 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
18868 (ipa_analyze_params_uses): Generate pessimistic info when true.
18870 2013-06-24 Martin Jambor <mjambor@suse.cz>
18872 PR tree-optimization/57539
18873 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
18874 global.inlined_to of the new node to it. All callers changed.
18875 * ipa-inline-transform.c (clone_inlined_nodes): New variable
18876 inlining_into, pass it to cgraph_clone_node.
18877 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
18878 ipa_free_edge_args_substructures.
18879 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
18880 rdesc linked list. Do not assert rdesc edges have inlined caller.
18881 Assert we have found an rdesc in the rdesc list.
18883 2013-06-24 Richard Biener <rguenther@suse.de>
18885 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
18886 (pointer_set_lookup): Declare.
18887 (class pointer_map): New template class implementing a
18888 generic pointer to T map.
18889 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
18890 pointer_map<T>::contains, pointer_map<T>::insert,
18891 pointer_map<T>::traverse): New functions.
18892 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
18893 (pointer_set_lookup): New function.
18894 (pointer_set_contains): Use pointer_set_lookup.
18895 (pointer_set_insert): Likewise.
18896 (insert_aux): Remove.
18897 (struct pointer_map_t): Embed a pointer_set_t.
18898 (pointer_map_create): Adjust.
18899 (pointer_map_destroy): Likewise.
18900 (pointer_map_contains): Likewise.
18901 (pointer_map_insert): Likewise.
18902 (pointer_map_traverse): Likewise.
18903 * tree-streamer.h (struct streamer_tree_cache_d): Use a
18904 pointer_map<unsigned> instead of a pointer_map_t.
18905 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
18906 (streamer_tree_cache_lookup): Likewise.
18907 (streamer_tree_cache_create): Likewise.
18908 (streamer_tree_cache_delete): Likewise.
18909 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
18910 pointer_map<unsigned> instead of a pointer_map_t.
18911 (lto_init_tree_ref_encoder): Adjust.
18912 (lto_destroy_tree_ref_encoder): Likewise.
18913 * lto-section-out.c (lto_output_decl_index): Likewise.
18914 (lto_record_function_out_decl_state): Likewise.
18915 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
18917 2013-06-24 Richard Biener <rguenther@suse.de>
18919 PR tree-optimization/57488
18920 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
18922 2013-06-24 Alan Modra <amodra@gmail.com>
18924 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
18925 (gen_easy_altivec_constant): Likewise.
18926 * config/rs6000/predicates.md (easy_vector_constant_add_self,
18927 easy_vector_constant_msb): Likewise.
18929 2013-06-23 Jakub Jelinek <jakub@redhat.com>
18932 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
18933 add missing return true.
18935 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
18938 * config/sh/predicates.md (general_extend_operand): Invoke
18939 general_movsrc_operand for memory operands.
18940 (general_movsrc_operand): Allow reg+reg addressing, do not use
18941 general_operand for memory operands.
18943 2013-06-23 Sriraman Tallam <tmsriram@google.com>
18945 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
18946 when current target options does not apply.
18947 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
18948 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
18949 * config/i386/bmiintrin.h: Pass appropriate target
18950 attributes to header.
18951 * config/i386/mmintrin.h: Ditto.
18952 * config/i386/nmmintrin.h: Ditto.
18953 * config/i386/avx2intrin.h: Ditto.
18954 * config/i386/fxsrintrin.h: Ditto.
18955 * config/i386/tbmintrin.h: Ditto.
18956 * config/i386/xsaveintrin.h: Ditto.
18957 * config/i386/f16cintrin.h: Ditto.
18958 * config/i386/xtestintrin.h: Ditto.
18959 * config/i386/xsaveoptintrin.h: Ditto.
18960 * config/i386/bmi2intrin.h: Ditto.
18961 * config/i386/lzcntintrin.h: Ditto.
18962 * config/i386/smmintrin.h: Ditto.
18963 * config/i386/wmmintrin.h: Ditto.
18964 * config/i386/x86intrin.h: Remove all header include guards.
18965 * config/i386/prfchwintrin.h: Ditto.
18966 * config/i386/pmmintrin.h: Ditto.
18967 * config/i386/tmmintrin.h: Ditto.
18968 * config/i386/xmmintrin.h: Ditto.
18969 * config/i386/popcntintrin.h: Ditto.
18970 * config/i386/rdseedintrin.h: Ditto.
18971 * config/i386/ammintrin.h: Ditto.
18972 * config/i386/emmintrin.h: Ditto.
18973 * config/i386/immintrin.h: Remove all header include guards.
18974 * config/i386/fma4intrin.h: Ditto.
18975 * config/i386/lwpintrin.h: Ditto.
18976 * config/i386/xopintrin.h: Ditto.
18977 * config/i386/ia32intrin.h: Ditto.
18978 * config/i386/avxintrin.h: Ditto.
18979 * config/i386/rtmintrin.h: Ditto.
18980 * config/i386/fmaintrin.h: Ditto.
18981 * config/i386/mm3dnow.h: Ditto.
18983 2013-06-22 Sriraman Tallam <tmsriram@google.com>
18985 * common/config/i386/i386-common.c: Handle LZCNT.
18987 2013-06-22 Andi Kleen <ak@linux.intel.com>
18989 * doc/extend.texi: Use __atomic_store_n instead of
18990 __atomic_store in HLE example.
18992 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
18994 * config/sh/sh.c: Remove <cstdlib> workaround.
18996 2013-06-21 Andi Kleen <ak@linux.intel.com>
18998 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
19000 2013-06-21 Andi Kleen <ak@linux.intel.com>
19002 * doc/extend.texi: Document that __atomic_clear and
19003 __atomic_test_and_set should only be used with bool.
19005 2013-06-20 Jan Hubicka <jh@suse.cz>
19007 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
19008 types_same_for_odr.
19009 * tree.c (decls_same_for_odr): New function.
19010 (same_for_edr): New function.
19011 (types_same_for_odr): New function.
19012 (get_binfo_at_offset): Use it.
19013 * tree.h (types_same_for_odr): Declare.
19015 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
19016 Jason Merrill <jason@redhat.com>
19018 * system.h: Include <cstdlib> as well as <stdlib.h>.
19020 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
19023 * config/i386/i386.c (construct_container): Report error if
19024 long double is used with disabled x87 float returns.
19026 2013-06-20 Jan Hubicka <jh@suse.cz>
19028 * lto-cgraph.c (input_symtab): Do not set cgraph state.
19030 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
19032 PR rtl-optimization/57425
19033 PR rtl-optimization/57569
19034 * alias.c (write_dependence_p): Remove parameters mem_mode and
19035 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
19036 Changed all callers.
19037 (canon_anti_dependence): Get comments and semantics in sync.
19038 Add parameter mem_canonicalized. Changed all callers.
19039 * rtl.h (canon_anti_dependence): Update prototype.
19041 2013-06-20 Richard Biener <rguenther@suse.de>
19043 * data-streamer-in.c (streamer_read_uhwi): Optimize single
19044 byte case, inline streamer_read_uchar and defer section
19047 2013-06-20 Richard Biener <rguenther@suse.de>
19049 PR tree-optimization/57584
19050 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
19051 SSA names into the expanded expression that take part in
19052 abnormal coalescing.
19054 2013-06-19 Sharad Singhai <singhai@google.com>
19056 * gcov.c (print_usage): Handle new option.
19057 (process_args): Ditto.
19058 (get_gcov_intermediate_filename): New function.
19059 (output_intermediate_file): New function.
19060 (output_gcov_file): New function
19061 (generate_results): Handle new option.
19062 (release_function): Relase demangled name.
19063 (read_graph_file): Handle demangled name.
19064 (output_lines): Ditto.
19065 * doc/gcov.texi: Document gcov intermediate format.
19067 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
19070 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
19071 (lra_emit_add): Use the functions. Add comment about Y as an
19074 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
19077 * collect2.c (collect_atexit): New.
19078 (collect_exit): Delete.
19079 (main): Register collect_atexit with atexit.
19080 (collect_wait): Change collect_exit to exit.
19082 * collect2.h (collect_exit): Delete.
19083 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
19085 2013-06-19 Wei Mi <wmi@google.com>
19087 PR rtl-optimization/57518
19088 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
19089 if regno is used in paradoxical subreg.
19090 (update_equiv_regs): Check pdx_subregs[regno] before
19091 set a reg to be equivalent with a mem.
19093 2013-06-19 Matthias Klose <doko@ubuntu.com>
19096 * file-find.h (find_a_file): Add a mode parameter.
19097 * file-find.c (find_a_file): Likewise.
19098 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
19099 with X_OK for the executables.
19100 * collect2.c (main): Call find_a_file with X_OK.
19102 2013-06-19 Steve Ellcey <sellcey@mips.com>
19105 * config/mips/mips.md (casesi_internal_mips16_<mode>):
19106 Use NEXT_INSN instead of next_real_insn.
19108 2013-06-19 Jan Hubicka <jh@suse.cz>
19110 * cgraph.h (const_value_known_p): Replace by ...
19111 (ctor_for_folding): .. this one.
19112 * cgraphunit.c (process_function_and_variable_attributes): Use it.
19113 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
19114 * expr.c (expand_expr_real_1): Likewise.
19115 (string_constant): Likewise.
19116 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
19117 * ipa.c (process_references): Likewise.
19118 (symtab_remove_unreachable_nodes): Likewise.
19119 * ipa-inline-analysis.c (param_change_prob): Likewise.
19120 * gimple-fold.c (canonicalize_constructor_val): Likewise.
19121 (get_base_constructor): Likwise.
19122 * varpool.c (varpool_remove_node): Likewise.
19123 (varpool_remove_initializer): LIkewise.
19124 (dump_varpool_node): LIkwise.
19125 (const_value_known_p): Rewrite to ...
19126 (ctor_for_folding): ... this one.
19128 2013-06-19 Jakub Jelinek <jakub@redhat.com>
19131 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
19132 PERSONALITY in $PATH derived prefixes.
19134 2013-06-19 Jeff Law <law@redhat.com>
19136 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
19139 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
19140 (simplify_bitwise_binary): Use it to simpify certain binary ops on
19143 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
19145 * config/arm/vfp.md: Move VFP instruction classification documentation
19147 * config/arm/arm.md: ... here. Update instruction classification
19150 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
19152 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
19153 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
19154 pattern. Use more efficient sequences on ARMv5 and Thumb2.
19156 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
19159 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
19160 with NEXT_INSN. Use tablejump_p to check for jump table data
19163 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
19166 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
19167 that now in C++ the value is correct per the C++ standards.
19169 2013-06-19 Richard Biener <rguenther@suse.de>
19171 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
19172 for global context.
19174 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19177 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19180 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
19181 with next_active_insn.
19183 2013-06-18 Sriraman Tallam <tmsriram@google.com>
19185 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
19186 functions are inlined during failures to flag an error.
19187 * tree-inline.c (expand_call_inline): Allow the error to be flagged
19188 in early inline pass.
19190 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
19192 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
19195 2013-06-18 Julian Brown <julian@codesourcery.com>
19197 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
19198 Permit virtual register pre-reload if !strict.
19199 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
19201 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
19203 * config/arm/neon.md (movmisalign<mode>): Use
19204 neon_perm_struct_or_reg_operand instead of
19205 neon_struct_or_register_operand.
19206 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
19207 neon_permissive_struct_operand instead of neon_struct_operand.
19208 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
19209 neon_vector_mem_operand.
19210 * config/arm/predicates.md (neon_struct_operand): Adjust call to
19211 neon_vector_mem_operand.
19212 (neon_permissive_struct_operand): New.
19213 (neon_struct_or_register_operand): Rename to...
19214 (neon_perm_struct_or_reg_operand): This. Adjust call to
19215 neon_vector_mem_operand.
19217 2013-06-18 Richard Biener <rguenther@suse.de>
19219 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
19220 * lto-streamer.h: Include pointer-set.h.
19221 (struct lto_decl_slot): Remove.
19222 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
19223 Remove next_index entry.
19224 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
19225 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
19226 (lto_init_tree_ref_encoder): Adjust.
19227 (lto_destroy_tree_ref_encoder): Likewise.
19228 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
19229 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
19230 (lto_output_decl_index): Adjust.
19231 (lto_new_out_decl_state): Likewise.
19232 (lto_record_function_out_decl_state): Likewise.
19233 * lto-streamer-out.c (copy_function): Likewise.
19235 2013-06-18 Richard Biener <rguenther@suse.de>
19237 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
19238 * cgraphunit.c: Include cfgloop.h.
19239 (init_lowered_empty_function): Initialize the loop tree.
19240 (assemble_thunk): Insert new BBs into loops.
19242 2013-06-18 Richard Biener <rguenther@suse.de>
19244 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
19245 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
19246 the map from cache entry to cache index optional.
19247 (streamer_tree_cache_replace_tree): Adjust accordingly.
19248 (streamer_tree_cache_append): Likewise.
19249 (streamer_tree_cache_delete): Likewise.
19250 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
19251 streamer cache map from cache entry to cache index.
19252 * lto-streamer-out.c (create_output_block): Adjust.
19254 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
19256 * config/arm/arm.md (attribute "insn"): Move multiplication and
19257 division attributes to...
19258 (attribute "type"): ... here. Remove mult.
19259 (attribute "mul32"): New attribute.
19260 (attribute "mul64"): Add umaal.
19261 (*arm_mulsi3): Update attributes.
19262 (*arm_mulsi3_v6): Likewise.
19263 (*thumb_mulsi3): Likewise.
19264 (*thumb_mulsi3_v6): Likewise.
19265 (*mulsi3_compare0): Likewise.
19266 (*mulsi3_compare0_v6): Likewise.
19267 (*mulsi_compare0_scratch): Likewise.
19268 (*mulsi_compare0_scratch_v6): Likewise.
19269 (*mulsi3addsi): Likewise.
19270 (*mulsi3addsi_v6): Likewise.
19271 (*mulsi3addsi_compare0): Likewise.
19272 (*mulsi3addsi_compare0_v6): Likewise.
19273 (*mulsi3addsi_compare0_scratch): Likewise.
19274 (*mulsi3addsi_compare0_scratch_v6): Likewise.
19275 (*mulsi3subsi): Likewise.
19276 (*mulsidi3adddi): Likewise.
19277 (*mulsi3addsi_v6): Likewise.
19278 (*mulsidi3adddi_v6): Likewise.
19279 (*mulsidi3_nov6): Likewise.
19280 (*mulsidi3_v6): Likewise.
19281 (*umulsidi3_nov6): Likewise.
19282 (*umulsidi3_v6): Likewise.
19283 (*umulsidi3adddi): Likewise.
19284 (*umulsidi3adddi_v6): Likewise.
19285 (*smulsi3_highpart_nov6): Likewise.
19286 (*smulsi3_highpart_v6): Likewise.
19287 (*umulsi3_highpart_nov6): Likewise.
19288 (*umulsi3_highpart_v6): Likewise.
19289 (mulhisi3): Likewise.
19290 (*mulhisi3tb): Likewise.
19291 (*mulhisi3bt): Likewise.
19292 (*mulhisi3tt): Likewise.
19293 (maddhisi4): Likewise.
19294 (*maddhisi4tb): Likewise.
19295 (*maddhisi4tt): Likewise.
19296 (maddhidi4): Likewise.
19297 (*maddhidi4tb): Likewise.
19298 (*maddhidi4tt): Likewise.
19299 (divsi3): Likewise.
19300 (udivsi3): Likewise.
19301 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
19302 (thumb2_mulsi_short_compare0): Likewise.
19303 (thumb2_mulsi_short_compare0_scratch): Likewise.
19304 * config/arm/arm1020e.md (1020mult1): Update attribute change.
19305 (1020mult2): Likewise.
19306 (1020mult3): Likewise.
19307 (1020mult4): Likewise.
19308 (1020mult5): Likewise.
19309 (1020mult6): Likewise.
19310 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
19312 (cortex_a15_mult64): Likewise.
19313 (cortex_a15_sdiv): Likewise.
19314 (cortex_a15_udiv): Likewise.
19315 * config/arm/arm1026ejs.md (mult1): Update attribute change.
19321 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
19322 (pj4_ir_div): Likewise.
19323 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
19324 (11_mult2): Likewise.
19325 (11_mult3): Likewise.
19326 (11_mult4): Likewise.
19327 (11_mult5): Likewise.
19328 (11_mult6): Likewise.
19329 (11_mult7): Likewise.
19330 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
19331 (cortex_a8_mla): Likewise.
19332 (cortex_a8_mull): Likewise.
19333 (cortex_a8_smulwy): Likewise.
19334 (cortex_a8_smlald): Likewise.
19335 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
19336 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
19337 (cortex_r4_mul_3): Likewise.
19338 (cortex_r4_mla_4): Likewise.
19339 (cortex_r4_mla_3): Likewise.
19340 (cortex_r4_smlald): Likewise.
19341 (cortex_r4_mull): Likewise.
19342 (cortex_r4_sdiv): Likewise.
19343 (cortex_r4_udiv): Likewise.
19344 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
19345 (cortex_a7_idiv): Likewise.
19346 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
19347 (9_mult2): Likewise.
19348 (9_mult3): Likewise.
19349 (9_mult4): Likewise.
19350 (9_mult5): Likewise.
19351 (9_mult6): Likewise.
19352 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
19353 (cortex_a53_sdiv): Likewise.
19354 (cortex_a53_udiv): Likewise.
19355 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
19356 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
19357 (mp626_mult2): Likewise.
19358 (mp626_mult3): Likewise.
19359 (mp626_mult4): Likewise.
19360 * config/arm/fa526.md (526_mult1): Update attribute change.
19361 (526_mult2): Likewise.
19362 * config/arm/arm-generic.md (mult): Update attribute change.
19363 (mult_ldsched_strongarm): Likewise.
19364 (mult_ldsched): Likewise.
19365 (multi_cycle): Likewise.
19366 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
19367 * config/arm/fa606te.md (606te_mult1): Update attribute change.
19368 (606te_mult2): Likewise.
19369 (606te_mult3): Likewise.
19370 (606te_mult4): Likewise.
19371 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
19372 (cortex_a9_mac16): Likewise.
19373 (cortex_a9_multiply): Likewise.
19374 (cortex_a9_mac): Likewise.
19375 (cortex_a9_multiply_long): Likewise.
19376 * config/arm/fa626te.md (626te_mult1): Update attribute change.
19377 (626te_mult2): Likewise.
19378 (626te_mult3): Likewise.
19379 (626te_mult4): Likewise.
19381 2013-06-18 Richard Biener <rguenther@suse.de>
19384 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
19386 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19389 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
19390 with next_active_insn.
19392 2013-06-18 Alan Modra <amodra@gmail.com>
19394 * config/rs6000/rs6000.h (enum data_align): New.
19395 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
19396 (DATA_ABI_ALIGNMENT): Define.
19397 (CONSTANT_ALIGNMENT): Correct comment.
19398 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
19399 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
19401 2013-06-17 David Malcolm <dmalcolm@redhat.com>
19403 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
19404 ATTRIBUTE_UNUSED marking.
19406 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
19408 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
19409 alternative and update.
19410 (aarch64_dup_lanedi): Delete.
19411 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
19412 * config/aarch64/aarch64-simd-builtins.def: Update.
19414 2013-06-17 Richard Biener <rguenther@suse.de>
19416 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
19417 (lto_input_scc): Declare.
19418 (lto_input_tree_1): Likewise.
19419 (struct lto_stats_d): Add num_tree_bodies_output and
19420 num_pickle_refs_output.
19421 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
19422 (lto_read_tree_1): Split out from ...
19423 (lto_read_tree): ... this.
19424 (lto_input_scc): New function.
19425 (lto_input_tree_1): Split out from ...
19426 (lto_input_tree): ... this. Handle LTO_tree_scc.
19427 (lto_data_in_create): Create the streamer cache without hashes.
19428 * lto-streamer-out.c (create_output_block): Create the streamer
19429 cache with hashes when not doing WPA.
19430 (lto_write_tree_1): Split out from ...
19431 (lto_write_tree): ... this.
19432 (get_symbol_initial_value): New function.
19433 (lto_output_tree_1): Split out from ...
19434 (lto_output_tree): ... this. Write trees as series of SCCs
19435 using a DFS walk via DFS_write_tree.
19436 (struct sccs, struct scc_entry): New types.
19437 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
19438 (DFS_write_tree_body): New function.
19439 (DFS_write_tree): Likewise.
19440 (hash_tree): Likewise.
19441 (scc_entry_compare): Likewise.
19442 (hash_scc): Likewise.
19443 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
19444 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
19445 TREE_CHAIN as regular reference.
19446 (streamer_read_integer_cst): Remove.
19447 (streamer_get_pickled_tree): Adjust.
19448 * tree-streamer-out.c (streamer_write_chain): Disable streaming
19449 of DECL_EXTERNALs in BLOCK_VARS for now.
19450 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
19452 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
19453 Add hash value argument and record that if hashes are recorded
19455 (streamer_tree_cache_insert_1): Adjust.
19456 (streamer_tree_cache_insert): Likewise.
19457 (streamer_tree_cache_insert_at): Rename to ...
19458 (streamer_tree_cache_replace_tree): ... this and adjust.
19459 (streamer_tree_cache_append): Adjust.
19460 (record_common_node): Likewise.
19461 (streamer_tree_cache_create): Add argument whether to
19462 record hash values together with trees.
19463 (streamer_tree_cache_delete): Adjust.
19464 * tree-streamer.h (struct streamer_tree_cache_d): Add
19466 (streamer_read_integer_cst): Remove.
19467 (streamer_tree_cache_insert): Adjust.
19468 (streamer_tree_cache_append): Likewise.
19469 (streamer_tree_cache_insert_at): Rename to ...
19470 (streamer_tree_cache_replace_tree): ... this and adjust.
19471 (streamer_tree_cache_create): Add argument whether to record hashes.
19472 (streamer_tree_cache_get): Rename to ...
19473 (streamer_tree_cache_get_tree): ... this.
19474 (streamer_tree_cache_get_hash): New function.
19475 * tree.c (cache_integer_cst): New function.
19476 * tree.h (cache_integer_cst): Declare.
19477 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
19478 * lto-symtab.c (lto_varpool_replace_node): Only release
19479 DECL_INITIAL of non-prevailing decls.
19480 * varpool.c (varpool_remove_initializer): Do not release
19481 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
19483 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
19485 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
19486 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
19487 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
19488 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
19489 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
19490 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
19491 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
19492 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
19493 instead of TARGET_64BIT.
19494 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
19495 Require ISA_HAS_<D>DIV.
19497 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
19499 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
19500 (mips*-*-linux*): Move default with_llsc setting to where other
19502 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
19504 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
19505 Likewise. Remove default with_tune setting. Move default float
19506 setting to its own block. Handle with_llsc in the same block as above.
19508 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
19510 PR rtl-optimization/57425
19511 PR rtl-optimization/57569
19512 * alias.c (write_dependence_p): Add new parameters mem_mode,
19513 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
19514 Changed all callers.
19515 (canon_anti_dependence): New function.
19516 * cse.c (check_dependence): Use canon_anti_dependence.
19517 * cselib.c (cselib_invalidate_mem): Likewise.
19518 * rtl.h (canon_anti_dependence): Declare.
19520 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
19522 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
19523 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
19524 ".set mips3" for 64-bit targets.
19526 2013-06-15 Dehao Chen <dehao@google.com>
19528 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
19529 * gimple-low.c (gimple_check_call_matching_types): Likewise.
19530 (gimple_check_call_args): Likewise.
19531 * value-prof.c (check_ic_target): Likewise.
19532 * ipa-inline.c (early_inliner): Likewise.
19533 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
19534 * cgraph.c (cgraph_create_edge_1): Likewise.
19535 (cgraph_make_edge_direct): Likewise.
19537 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
19540 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
19541 rs6000_output_move_128bit to handle emitting quad memory
19542 operations. Set attribute length to 8 bytes.
19544 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
19546 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
19548 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
19549 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
19550 (aarch64_<su>mlsl<mode>): Likewise.
19552 2013-06-14 Mike Stump <mikestump@comcast.net>
19554 * Makefile.in (TARGET_H): Add insn-codes.h.
19556 2013-06-14 Alan Modra <amodra@gmail.com>
19558 PR middle-end/57134
19559 PR middle-end/57586
19560 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
19561 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
19562 bitfield expansion when EXPAND_MEMORY.
19563 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
19565 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
19567 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
19568 test for clearing quad memory on 32-bit later.
19570 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
19572 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
19573 (fold_negate_expr): Likewise.
19574 (fold_real_zero_addition_p): Handle vectors.
19575 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
19577 2013-06-14 Alan Modra <amodra@gmail.com>
19579 * varasm.c (force_const_mem): Revert 2013-06-07 change.
19581 2013-06-13 Jan Hubicka <jh@suse.cz>
19583 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
19584 Local comdats are not externally visible.
19585 * symtab.c (dump_symtab_base): Dump externally visible.
19586 (verify_symtab_base): Verify back links in the symtab hash.
19588 2013-06-13 Bin Cheng <bin.cheng@arm.com>
19590 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
19591 CONVERT_EXPR as equal nodes.
19593 2013-06-13 Bin Cheng <bin.cheng@arm.com>
19595 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
19597 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
19599 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
19600 Generalize to complex and vector.
19601 * tree.c (build_all_ones_cst): New function.
19602 * tree.h (build_all_ones_cst): Declare it.
19604 2013-06-13 Alan Modra <amodra@gmail.com>
19606 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
19607 * config/rs6000/rs6000.md (signbittf2): New insn.
19608 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
19609 (abstf2_internal, cmptf_internal2): Likewise.
19610 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
19612 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
19613 Pat Haugen <pthaugen@us.ibm.com>
19614 Peter Bergner <bergner@vnet.ibm.com>
19616 * config/rs6000/rs6000.c (emit_load_locked): Add support for
19617 power8 byte, half-word, and quad-word atomic instructions.
19618 (emit_store_conditional): Likewise.
19619 (rs6000_expand_atomic_compare_and_swap): Likewise.
19620 (rs6000_expand_atomic_op): Likewise.
19622 * config/rs6000/sync.md (larx): Add new modes for power8.
19624 (AINT): New mode iterator to include TImode as well as normal
19625 integer modes on power8.
19626 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
19627 that VSX registers are not considered. Use AINT mode iterator
19628 instead of INT1 to allow inclusion of quad word atomic operations
19630 (load_locked<mode>): Likewise.
19631 (store_conditional<mode>): Likewise.
19632 (atomic_compare_and_swap<mode>): Likewise.
19633 (atomic_exchange<mode>): Likewise.
19634 (atomic_nand<mode>): Likewise.
19635 (atomic_fetch_<fetchop_name><mode>): Likewise.
19636 (atomic_nand_fetch<mode>): Likewise.
19637 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
19639 (ATOMIC): On power8, add QImode, HImode modes.
19640 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
19641 modes that promote to SImode.
19642 (load_lockedti): Convert TImode arguments to PTImode, so that we
19643 get a guaranteed even/odd register pair.
19644 (load_lockedpti): Likewise.
19645 (store_conditionalti): Likewise.
19646 (store_conditionalpti): Likewise.
19648 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
19649 atomic load/store instructions.
19652 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
19654 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
19656 (insn_count): New attribute, with most cases extracted from...
19657 (length): ...here. Redefine most cases in terms of insn_count.
19658 (single_insn): Delete.
19659 (can_delay): Use insn_count to check for single instructions.
19660 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
19661 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
19662 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
19663 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
19664 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
19665 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
19666 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
19667 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
19668 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
19669 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
19670 rather than "length".
19671 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
19672 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
19673 Use "insn_count" rather than "length".
19674 * config/mips/mips-dsp.md
19675 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
19676 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
19679 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
19681 PR tree-optimization/57361
19682 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
19684 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
19686 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
19688 (aarch64_simd_combine<mode>): New instruction expansion.
19689 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
19690 function prototype.
19691 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
19692 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
19694 2013-06-12 Jan Hubicka <jh@suse.cz>
19696 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
19697 decl has when in streaming stage.
19698 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
19699 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
19701 2013-06-12 Roland Stigge <stigge@antcom.de>
19704 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
19706 2013-06-12 Jakub Jelinek <jakub@redhat.com>
19708 PR tree-optimization/57537
19709 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
19710 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
19712 2013-06-12 Richard Biener <rguenther@suse.de>
19714 * data-streamer.h (streamer_write_char_stream): CSE
19715 obs->current_pointer.
19716 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
19717 streamer_write_char_stream manually and optimize the resulting loop.
19718 (streamer_write_hwi_stream): Likewise.
19720 2013-06-12 Jan Hubicka <jh@suse.cz>
19722 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
19723 * cgraph.h (varpool_create_empty_node): Declare.
19724 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
19726 * symtab.c (symtab_unregister_node): Be lax about missin entries
19728 (symtab_get_node): Update comment.
19729 * varpool.c (varpool_create_empty_node): Break out from ...
19730 (varpool_node_for_decl): ... here.
19731 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
19733 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
19735 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
19736 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
19737 part. Use straight-line flow at the end.
19738 <COMPONENT_REF>: Remove superfluous else.
19739 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
19741 2013-06-12 Jakub Jelinek <jakub@redhat.com>
19744 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
19745 target hook even for !TREE_PUBLIC decls. If no resolution info
19746 is available, return false for common and external decls.
19748 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
19750 * config/rl78/constraints.md (U): New constraint.
19751 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
19754 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
19757 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
19758 to allow returning address to AT_PLATFORM name.
19760 2013-06-11 Jan Hubicka <jh@suse.cz>
19762 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
19763 * cgraph.h (symtab_node_base): Add weakref flag.
19764 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
19765 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
19766 (output_weakrefs): Use weakref flag.
19767 * fold-const.c (simple_operand_p): Handle WEAK.
19768 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
19769 * ipa.c (varpool_externally_visible_p): Drop weakref.
19770 (function_and_variable_visibility): Update comment; fix weakref
19771 sanity checks; do not clear DECL_WEAK on them.
19772 * lto-cgraph.c (lto_output_node): update.
19773 (lto_output_varpool_node): Update.
19774 (input_overwrite_node): Update.
19775 (input_node): Update.
19776 (input_varpool_node): Update.
19777 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
19778 (lto_symtab_merge_symbols): Add sanity check.
19779 (lto_symtab_prevailing_decl): Do not special case weakrefs.
19780 * passes.c (rest_of_decl_compilation): Set static flag, too.
19781 * symtab.c (dump_symtab_base): Dump weakref.
19782 (verify_symtab_base): Sanity check weakrefs.
19783 (symtab_make_decl_local): Remove duplicated code.
19784 (symtab_alias_ultimate_target): Simplify.
19785 * varpool.c (varpool_create_variable_alias): Set weakref flag.
19787 2013-06-11 Tom de Vries <tom@codesourcery.com>
19789 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
19790 sequence_vect == NULL.
19792 2013-06-11 DJ Delorie <dj@redhat.com>
19794 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
19795 (rl78_unwind_word_mode): New.
19797 2013-06-11 David Malcolm <dmalcolm@redhat.com>
19799 * final.c (debug_prefix_maps): Make static.
19801 2013-06-11 David Malcolm <dmalcolm@redhat.com>
19803 * function.c (initial_trampoline): Remove stray copy.
19805 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
19807 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
19809 2013-06-11 Martin Jambor <mjambor@suse.cz>
19811 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
19812 within bounds at the beginning of the function.
19814 2013-06-11 Alan Modra <amodra@gmail.com>
19816 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
19818 (get_named_section): Don't NULL !DECL_P decl.
19820 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
19822 * doc/invoke.texi (core-avx2): Document.
19824 (atom): Updated with MOVBE.
19826 2013-06-11 Richard Biener <rguenther@suse.de>
19828 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
19830 2013-06-11 Anton Blanchard <anton@samba.org>
19832 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
19833 correct shift value in little-endian mode.
19835 2013-06-11 Jakub Jelinek <jakub@redhat.com>
19838 * varasm.c (get_variable_align): Move #endif to the right place.
19840 2013-06-10 Cary Coutant <ccoutant@google.com>
19842 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
19843 for hash so that hash table traversal order is deterministic.
19845 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
19846 Pat Haugen <pthaugen@us.ibm.com>
19847 Peter Bergner <bergner@vnet.ibm.com>
19849 * config/rs6000/vector.md (GPR move splitter): Do not split moves
19850 of vectors in GPRS if they are direct moves or quad word load or
19853 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
19855 (direct_move_p): Likewise.
19856 (quad_load_store_p): Likewise.
19858 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
19859 classes into bins based on the physical register type.
19860 (reg_class_to_reg_type): Likewise.
19861 (IS_STD_REG_TYPE): Likewise.
19862 (IS_FP_VECT_REG_TYPE): Likewise.
19863 (reload_fpr_gpr): Arrays to determine what insn to use if we can
19864 use direct move instructions.
19865 (reload_gpr_vsx): Likewise.
19866 (reload_vsx_gpr): Likewise.
19867 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
19868 information that is a simplification of register classes. Also
19869 precalculate direct move reload helpers.
19870 (direct_move_p): New function to return true if the operation can
19871 be done as a direct move instruciton.
19872 (quad_load_store_p): New function to return true if the operation
19873 is a quad memory operation.
19874 (rs6000_legitimize_address): If quad memory, only allow register
19875 indirect for TImode addresses.
19876 (rs6000_legitimate_address_p): Likewise.
19877 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
19878 (rs6000_reload_register_type): Likewise.
19879 (register_to_reg_type): Return register type.
19880 (rs6000_secondary_reload_simple_move): New helper function for
19881 secondary reload and secondary memory needed to identify anything
19882 that is a simple move, and does not need reloading.
19883 (rs6000_secondary_reload_direct_move): New helper function for
19884 secondary reload to identify cases that can be done with several
19885 instructions via the direct move instructions.
19886 (rs6000_secondary_reload_move): New helper function for secondary
19887 reload to identify moves between register types that can be done.
19888 (rs6000_secondary_reload): Add support for quad memory operations
19889 and for direct move.
19890 (rs6000_secondary_memory_needed): Likewise.
19891 (rs6000_debug_secondary_memory_needed): Change argument names.
19892 (rs6000_output_move_128bit): New function to return the move to
19893 use for 128-bit moves, including knowing about the various
19894 limitations of quad memory operations.
19896 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
19897 memory operations. call rs6000_output_move_128bit for the actual
19898 instruciton(s) to generate.
19899 (vsx_movti_64bit): Likewise.
19901 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
19902 (UNSPEC_P8V_MTVSRWZ): Likewise.
19903 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
19904 (UNSPEC_P8V_MTVSRD): Likewise.
19905 (UNSPEC_P8V_XXPERMDI): Likewise.
19906 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
19907 (UNSPEC_FUSION_GPR): Likewise.
19908 (FMOVE128_GPR): New iterator for direct move.
19909 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
19910 (f32_sv): Likewise.
19911 (f32_dm): Likewise.
19912 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
19913 loads and direct move instructions.
19914 (zero_extendsidi2_lfiwzx): Likewise.
19915 (extendsidi2_lfiwax): Likewise.
19916 (extendsidi2_nocell): Likewise.
19917 (floatsi<mode>2_lfiwax): Likewise.
19918 (lfiwax): Likewise.
19919 (floatunssi<mode>2_lfiwzx): Likewise.
19920 (lfiwzx): Likewise.
19921 (fix_trunc<mode>_stfiwx): Likewise.
19922 (fixuns_trunc<mode>_stfiwx): Likewise.
19923 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
19924 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
19925 (parity<mode>2_cmpb): Set length/type attr.
19926 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
19927 for 'mr.' to fast_compare.
19928 (bpermd_<mode>): Change type attr to popcnt.
19929 (p8_fmrgow_<mode>): New insns for power8 direct move support.
19930 (p8_mtvsrwz_1): Likewise.
19931 (p8_mtvsrwz_2): Likewise.
19932 (reload_fpr_from_gpr<mode>): Likewise.
19933 (p8_mtvsrd_1): Likewise.
19934 (p8_mtvsrd_2): Likewise.
19935 (p8_xxpermdi_<mode>): Likewise.
19936 (reload_vsx_from_gpr<mode>): Likewise.
19937 (reload_vsx_from_gprsf): Likewise.
19938 (p8_mfvsrd_3_<mode>): LIkewise.
19939 (reload_gpr_from_vsx<mode>): Likewise.
19940 (reload_gpr_from_vsxsf): Likewise.
19941 (p8_mfvsrd_4_disf): Likewise.
19942 (multi-word GPR splits): Do not split direct moves or quad memory
19945 2013-06-10 David Malcolm <dmalcolm@redhat.com>
19947 * tree-into-ssa.c (interesting_blocks): Make static.
19949 2013-06-10 Jakub Jelinek <jakub@redhat.com>
19952 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
19953 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
19954 Use DATA_ABI_ALIGNMENT for that case instead if defined.
19955 (get_variable_align): New function.
19956 (get_variable_section, emit_bss, emit_common,
19957 assemble_variable_contents, place_block_symbol): Use
19958 get_variable_align instead of DECL_ALIGN.
19959 (assemble_noswitch_variable): Add align argument, use it
19960 instead of DECL_ALIGN.
19961 (assemble_variable): Adjust caller. Use get_variable_align
19962 instead of DECL_ALIGN.
19963 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
19965 (DATA_ABI_ALIGNMENT): Define.
19966 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
19967 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
19968 opt is false, only return the psABI mandated alignment increase.
19969 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
19970 (DATA_ABI_ALIGNMENT): ... this.
19971 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
19972 (DATA_ABI_ALIGNMENT): ... this.
19973 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
19974 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
19975 (DATA_ABI_ALIGNMENT): ... this.
19976 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
19977 * doc/tm.texi: Regenerated.
19979 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
19981 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
19982 cmp_code to construct REG_EQUAL note.
19984 2013-06-09 Jakub Jelinek <jakub@redhat.com>
19987 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
19988 that operands[2] doesn't overlap with operands[0].
19990 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
19991 Jan Hubicka <jh@suse.cz>
19993 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
19994 hack to mark symbols as used.
19996 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
19998 PR rtl-optimization/57559
19999 * lra-constraints.c (process_alt_operands): Don't discourage
20000 memory with known offset for offsetable memory constraint.
20001 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
20003 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
20005 * varasm.c (struct oc_local_state): Reorder fields.
20006 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
20007 and adjust accordingly.
20008 (output_constructor): Reorder initialization code and adjust call to
20009 output_constructor_bitfield.
20011 2013-06-07 Jan Hubicka <jh@suse.cz>
20013 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
20015 2013-06-07 David Malcolm <dmalcolm@redhat.com>
20017 * tree-object-size.c (unknown): Make const.
20019 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20021 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
20022 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
20023 for last alternative in the cpu_facility attribute.
20025 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20028 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
20029 (xordi3): Change operand 2 constraint to arm_xordi_operand.
20030 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
20031 * config/arm/constraints.md (Dg): New constraint.
20032 * config/arm/neon.md (xordi3_neon): Remove.
20033 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
20034 * config/arm/predicates.md (arm_xordi_operand): New predicate.
20036 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20038 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
20039 Clean up alternatives.
20041 2013-06-07 Alan Modra <amodra@gmail.com>
20043 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
20046 2013-06-07 Alan Modra <amodra@gmail.com>
20048 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
20050 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20052 * config/arm/constraints.md (Df): New constraint.
20053 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
20054 Correct length attribute for last two alternatives.
20056 2013-06-07 Alan Modra <amodra@gmail.com>
20058 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
20059 override user -mfp-in-toc.
20060 (offsettable_ok_by_alignment): Consider just the current access
20061 rather than the whole object, unless BLKmode. Handle
20062 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
20063 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
20064 for -mcmodel=medium.
20065 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
20066 override user -mfp-in-toc or -msum-in-toc. Default to
20067 -mno-fp-in-toc for -mcmodel=medium.
20069 2013-06-06 DJ Delorie <dj@redhat.com>
20071 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
20072 TARGET_VALID_POINTER_MODE.
20074 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
20075 Pat Haugen <pthaugen@us.ibm.com>
20076 Peter Bergner <bergner@vnet.ibm.com>
20078 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
20079 Document new power8 builtins.
20081 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
20082 condition code register, to allow 128-bit logical operations to be
20083 done in the VSX or GPR registers.
20084 (nor<mode>3): Use the canonical form for nor.
20085 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
20086 vclz*, and vpopcnt* vector instructions.
20087 (nand<mode>3): Likewise.
20088 (orc<mode>3): Likewise.
20089 (clz<mode>2): LIkewise.
20090 (popcount<mode>2): Likewise.
20092 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
20093 that only the GPRs are recognized.
20095 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
20096 support for new power8 builtins.
20098 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
20100 (xscvdpspn): Likewise.
20106 (vpopcnt): Likewise.
20107 (vpopcntb): Likewise.
20108 (vpopcnth): Likewise.
20109 (vpopcntw): Likewise.
20110 (vpopcntd): Likewise.
20112 (vmrgew): Likewise.
20113 (vmrgow): Likewise.
20115 (eqv_v16qi3): Likewise.
20116 (eqv_v8hi3): Likewise.
20117 (eqv_v4si3): Likewise.
20118 (eqv_v2di3): Likewise.
20119 (eqv_v4sf3): Likewise.
20120 (eqv_v2df3): Likewise.
20122 (nand_v16qi3): Likewise.
20123 (nand_v8hi3): Likewise.
20124 (nand_v4si3): Likewise.
20125 (nand_v2di3): Likewise.
20126 (nand_v4sf3): Likewise.
20127 (nand_v2df3): Likewise.
20129 (orc_v16qi3): Likewise.
20130 (orc_v8hi3): Likewise.
20131 (orc_v4si3): Likewise.
20132 (orc_v2di3): Likewise.
20133 (orc_v4sf3): Likewise.
20134 (orc_v2df3): Likewise.
20136 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
20137 allow power8 quad mode in 64-bit.
20138 (rs6000_builtin_vectorized_function): Add support to vectorize
20139 ISA 2.07 count leading zeros, population count builtins.
20140 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
20141 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
20142 (builtin_function_type): Add vgbbd builtin function which takes an
20144 (altivec_expand_vec_perm_const): Add support for new power8 merge
20147 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
20148 that does not include TImdoe for use with 32-bit.
20149 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
20151 (UNSPEC_VSX_CVDPSPN): Likewise.
20152 (vsx_xscvdpspn): Likewise.
20153 (vsx_xscvspdpn): Likewise.
20154 (vsx_xscvdpspn_scalar): Likewise.
20155 (vsx_xscvspdpn_directmove): Likewise.
20156 (vsx_and<mode>3): Split logical operations into 32-bit and
20157 64-bit. Add support to do logical operations on TImode as well as
20158 VSX vector types. Allow logical operations to be done in either
20159 VSX registers or in general purpose registers in 64-bit mode. Add
20160 splitters if GPRs were used. For AND, add clobber of CCmode to
20161 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
20163 (vsx_and<mode>3_32bit): Likewise.
20164 (vsx_and<mode>3_64bit): Likewise.
20165 (vsx_ior<mode>3): Likewise.
20166 (vsx_ior<mode>3_32bit): Likewise.
20167 (vsx_ior<mode>3_64bit): Likewise.
20168 (vsx_xor<mode>3): Likewise.
20169 (vsx_xor<mode>3_32bit): Likewise.
20170 (vsx_xor<mode>3_64bit): Likewise.
20171 (vsx_one_cmpl<mode>2): Likewise.
20172 (vsx_one_cmpl<mode>2_32bit): Likewise.
20173 (vsx_one_cmpl<mode>2_64bit): Likewise.
20174 (vsx_nor<mode>3): Likewise.
20175 (vsx_nor<mode>3_32bit): Likewise.
20176 (vsx_nor<mode>3_64bit): Likewise.
20177 (vsx_andc<mode>3): Likewise.
20178 (vsx_andc<mode>3_32bit): Likewise.
20179 (vsx_andc<mode>3_64bit): Likewise.
20180 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
20181 and xxlorc instructions.
20182 (vsx_eqv<mode>3_64bit): Likewise.
20183 (vsx_nand<mode>3_32bit): Likewise.
20184 (vsx_nand<mode>3_64bit): Likewise.
20185 (vsx_orc<mode>3_32bit): Likewise.
20186 (vsx_orc<mode>3_64bit): Likewise.
20188 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
20190 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
20192 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
20193 (p8_vmrgow): Likewise.
20194 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
20195 GPRs to be split under VSX.
20196 (p8v_clz<mode>2): Add power8 count leading zero support.
20197 (p8v_popcount<mode>2): Add power8 population count support.
20198 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
20201 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
20204 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
20206 (vec_nand): Likewise.
20207 (vec_vclz): Likewise.
20208 (vec_vclzb): Likewise.
20209 (vec_vclzd): Likewise.
20210 (vec_vclzh): Likewise.
20211 (vec_vclzw): Likewise.
20212 (vec_vgbbd): Likewise.
20213 (vec_vmrgew): Likewise.
20214 (vec_vmrgow): Likewise.
20215 (vec_vpopcnt): Likewise.
20216 (vec_vpopcntb): Likewise.
20217 (vec_vpopcntd): Likewise.
20218 (vec_vpopcnth): Likewise.
20219 (vec_vpopcntw): Likewise.
20221 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
20223 PR rtl-optimization/57468
20224 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
20227 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
20229 PR rtl-optimization/57459
20230 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
20231 type when setting live regs.
20233 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
20235 * config/s390/s390.opt (mlra): New option.
20236 * config/s390/s390.c (s390_decompose_address): Check displacement
20237 for all registers for LRA.
20238 (s390_secondary_reload): Don't used secondary reloads for LRA.
20239 (s390_lra_p): New function.
20240 (TARGET_LRA_P): Define.
20241 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
20242 of attribute cpu_facility to zarch for the last alternative.
20243 (*cmpmem_short): Ditto.
20245 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
20247 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
20248 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
20249 (arm_expand_prologue): Likewise.
20251 2013-06-06 Teresa Johnson <tejohnson@google.com>
20254 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
20255 as this is now done by redirect_edge_and_branch_force.
20256 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
20257 barriers, and fix interaction with splitting.
20258 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
20259 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
20260 reflect changes made in the routine.
20261 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
20262 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
20263 since this is called in cfglayout mode, and replace partition fixup
20264 with assert as that is now done by force_nonfallthru_and_redirect.
20265 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
20266 already be marked with region crossing note.
20267 (insert_section_boundary_note): Make non-static, gate on flag
20268 has_bb_partition, rewrite to also check for multiple partitions.
20269 (rest_of_handle_reorder_blocks): Remove call to
20270 insert_section_boundary_note, now done later during free_cfg.
20271 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
20272 * bb-reorder.h (insert_section_boundary_note): Declare.
20273 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
20274 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
20275 invoke insert_section_boundary_note.
20276 (try_redirect_by_replacing_jump): Remove unnecessary
20277 check for region crossing note.
20278 (fixup_partition_crossing): New function.
20279 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
20280 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
20281 in non-cfglayout mode.
20282 (force_nonfallthru_and_redirect): Fixup partition boundaries,
20283 remove old code that tried to do this. Emit barrier correctly
20284 when we are in cfglayout mode.
20285 (last_bb_in_partition): New function.
20286 (rtl_split_edge): Correctly fixup partition boundaries.
20287 (commit_one_edge_insertion): Remove old code that tried to
20288 fixup region crossing edge since this is now handled in
20289 split_block, and set up insertion point correctly since
20290 block may now end in a jump.
20291 (verify_hot_cold_block_grouping): Guard against checking when not in
20292 linearized RTL mode.
20293 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
20295 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
20296 rtl_verify_flow_info, so not called in cfglayout mode.
20297 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
20298 (fixup_reorder_chain): Remove old code that attempted to fixup region
20299 crossing note as this is now handled in force_nonfallthru_and_redirect.
20300 (duplicate_insn_chain): Don't duplicate switch section notes.
20301 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
20303 * basic-block.h (emit_barrier_after_bb): Declare.
20305 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20307 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
20308 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
20309 arm_usatsihi): Adjust alternatives for arm_restrict_it.
20311 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20313 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
20315 * config/arm/ldmstm.md: Regenerate.
20317 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20319 * config/arm/sync.md (atomic_loaddi_1):
20320 Disable predication for arm_restrict_it.
20321 (arm_load_exclusive<mode>): Likewise.
20322 (arm_load_exclusivesi): Likewise.
20323 (arm_load_exclusivedi): Likewise.
20324 (arm_load_acquire_exclusive<mode>): Likewise.
20325 (arm_load_acquire_exclusivesi): Likewise.
20326 (arm_load_acquire_exclusivedi): Likewise.
20327 (arm_store_exclusive<mode>): Likewise.
20328 (arm_store_exclusive<mode>): Likewise.
20329 (arm_store_release_exclusivedi): Likewise.
20330 (arm_store_release_exclusive<mode>): Likewise.
20332 2013-06-06 Richard Biener <rguenther@suse.de>
20334 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
20336 (lto_tag_is_tree_code_p): Adjust.
20337 (lto_tag_is_gimple_code_p): Likewise.
20338 (lto_gimple_code_to_tag): Likewise.
20339 (lto_tag_to_gimple_code): Likewise.
20340 (lto_tree_code_to_tag): Likewise.
20341 (lto_tag_to_tree_code): Likewise.
20342 * data-streamer.h (streamer_write_hwi_in_range): Use
20343 uhwi streaming to stream the normalized range.
20344 (streamer_read_hwi_in_range): Likewise.
20346 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20348 * config/arm/arm.md (enabled_for_depr_it): New attribute.
20349 (predicable_short_it): Likewise.
20350 (predicated): Likewise.
20351 (enabled): Handle above.
20352 (define_cond_exec): Set predicated attribute to yes.
20354 2013-06-05 Mike Stump <mikestump@comcast.net>
20356 * gdbinit.in (__FUNCTION__): Add.
20358 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
20360 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
20361 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
20363 2013-06-05 Jan Hubicka <jh@suse.cz>
20365 * varasm.c (mark_decl_referenced): Revert the removal until targets
20368 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
20370 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
20371 instead of mark_decl_referenced.
20373 2013-06-05 Jan Hubicka <jh@suse.cz>
20375 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
20376 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
20377 and symtab_used_from_object_file_p.
20378 (cgraph_make_node_local_1): Clear forced_by_abi.
20379 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
20380 * cgraph.h (symtab_node_base): Add forced_by_abi.
20381 (decide_is_variable_needed): Remove.
20382 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
20383 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
20384 (decide_is_symbol_needed): ... this one; handle symbols in general;
20385 always analyze virtuals; honnor forced_by_abi.
20386 (cgraph_finalize_function): Update.
20387 (varpool_finalize_decl): Update.
20388 (symbol_defined_and_needed): Remove.
20389 (analyze_functions): Update.
20390 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
20391 output_refs, input_overwrite_node): Handle forced_by_abi.
20392 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
20393 (address_taken_from_non_vtable_p): ... this one.
20394 (comdat_can_be_unshared_p_1): New function.
20395 (cgraph_comdat_can_be_unshared_p): Rename to ...
20396 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
20397 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
20398 (function_and_variable_visibility): Clear forced_by_abi as needed.
20399 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
20400 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
20401 * symtab.c (dump_symtab_base): Dump forced_by_abi.
20402 * varpool.c (decide_is_variable_needed): Remove.
20404 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20406 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
20407 (arm_option_override): Override arm_restrict_it where appropriate.
20408 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
20409 * config/arm/arm.opt (mrestrict-it): New command-line option.
20410 * doc/invoke.texi: Document -mrestrict-it.
20412 2013-06-05 David Malcolm <dmalcolm@redhat.com>
20414 * tsan.c (tsan_atomic_table): Make const.
20416 2013-06-05 Richard Biener <rguenther@suse.de>
20418 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
20419 index associated with the tree we are supposed to replace.
20420 * tree-streamer-out.c (pack_ts_base_value_fields): Output
20421 TREE_ASM_WRITTEN as zero for everything but SSA names.
20423 2013-06-05 David Malcolm <dmalcolm@redhat.com>
20425 * tree-ssa-structalias.c (call_stmt_vars): Make static.
20427 2013-06-04 Jan Hubicka <jh@suse.cz>
20429 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
20430 (input_node, input_varpool_node): Handle correctly external same
20432 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
20433 nodes at ltrans stage.
20435 2013-06-04 Jan Hubicka <jh@suse.cz>
20437 * ipa-inline.c (update_caller_keys): Fix availability test.
20438 (update_callee_keys): Likewise.
20439 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
20440 to follow ELF standard.
20442 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
20444 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
20445 (mips64r5900el-*-elf*): New configurations.
20446 * config/mips/mips-cpus.def (r5900): New processor.
20447 * config/mips/mips-tables.opt: Regenerate.
20448 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
20449 (mips_issue_rate): Handle PROCESSOR_R5900.
20450 (mips_reorg_process_insns): Force reorder mode for the R5900.
20451 * config/mips/mips.h (TARGET_MIPS5900): Define.
20452 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
20454 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
20456 * config/mips/mips.md (processor): Add r5900.
20457 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
20459 2013-06-04 Ian Bolton <ian.bolton@arm.com>
20461 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
20462 into function to generate MOVI instruction.
20463 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
20464 (aarch64_preferred_simd_mode): Turn into wrapper.
20465 (aarch64_output_scalar_simd_mov_immediate): New function.
20466 * config/aarch64/aarch64-protos.h: Add prototype for above.
20468 2013-06-04 Ian Bolton <ian.bolton@arm.com>
20470 * config/aarch64/aarch64.c (simd_immediate_info): Remove
20471 element_char member.
20472 (sizetochar): Return signed char.
20473 (aarch64_simd_valid_immediate): Remove elchar and other
20474 unnecessary variables.
20475 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
20476 Calculate element_char as required.
20477 * config/aarch64/aarch64-protos.h: Update and move prototype
20478 for aarch64_output_simd_mov_immediate.
20479 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
20482 2013-06-04 Ian Bolton <ian.bolton@arm.com>
20484 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
20485 information completed by aarch64_simd_valid_immediate.
20486 (aarch64_legitimate_constant_p): Update arguments.
20487 (aarch64_simd_valid_immediate): Work with struct rather than many
20489 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
20490 (aarch64_simd_make_constant): Update arguments.
20491 (aarch64_output_simd_mov_immediate): Work with struct rather than
20492 many pointers. Output immediate directly rather than as operand.
20493 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
20495 * config/aarch64/constraints.md (Dn): Update arguments.
20497 2013-06-04 Ian Bolton <ian.bolton@arm.com>
20499 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
20501 (aarch64_simd_immediate_valid_for_move): Remove.
20502 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
20503 (aarch64_simd_make_constant): Update call.
20504 (aarch64_output_simd_mov_immediate): Update call.
20505 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
20507 * config/aarch64/constraints.md (Dn): Update call.
20509 2013-06-04 Ian Bolton <ian.bolton@arm.com>
20511 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
20512 return type to bool for prototype.
20513 (aarch64_legitimate_constant_p): Check for true instead of not -1.
20514 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
20515 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
20517 2013-06-04 Catherine Moore <clm@codesourcery.com>
20519 * config/mips/mips.opt (meva): New.
20520 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
20521 (ASM_SPEC): Handle -meva.
20522 * doc/invoke.texi (meva): Document.
20524 2013-06-04 Alan Modra <amodra@gmail.com>
20526 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
20529 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20531 * rtl.def: Add extra fourth optional field to define_cond_exec.
20532 * gensupport.c (process_one_cond_exec): Process attributes from
20534 * doc/md.texi: Document fourth field in define_cond_exec.
20536 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
20538 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
20539 out the processing order as in store_bit_field_1.
20541 2013-06-04 Jan Hubicka <jh@suse.cz>
20543 PR middle-end/57500
20544 * cgraphunit.c (cgraph_process_same_body_aliases): Create
20545 non-VAR_DECL node if it does not exist yet.
20547 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
20549 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
20550 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
20551 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
20552 target_cpu_default setting.
20554 2013-06-03 Teresa Johnson <tejohnson@google.com>
20556 * dumpfile.c (opt_info_switch_p): Change -fopt-info
20557 default to -fopt-info=optimized instead of all.
20558 * doc/invoke.texi: Ditto.
20559 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
20560 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
20561 (execute_vect_slp): Emit BB vectorization success under
20562 MSG_OPTIMIZED_LOCATIONS.
20563 * tree-vect-slp.c (vect_slp_transform_bb): Change
20564 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
20565 * tree-vect-loop.c (vect_transform_loop): Ditto.
20567 2013-06-03 Jason Merrill <jason@redhat.com>
20570 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
20571 Use TARGET_EXPR for C++.
20573 2013-06-03 Jakub Jelinek <jakub@redhat.com>
20575 PR rtl-optimization/57268
20576 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
20577 if DEBUG_INSN_P (insn).
20580 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
20582 PR rtl-optimization/57268
20583 * sched-deps.c (sched_analyze_2): Flush dependence lists if
20584 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
20586 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
20588 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
20589 (ix86_avoid_lea_for_addr): Likewise.
20590 (exact_dependency_1): Likewise.
20591 (ix86_adjust_cost): Likewise.
20592 (swap_top_of_ready_list): Fix formatting and !reload_completed check
20594 (do_reorder_for_imul): Fix typo, formatting and
20595 !reload_completed check removed.
20596 (ix86_sched_reorder): Fix typo and formatting.
20597 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
20600 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
20602 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
20604 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
20606 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
20607 <STRING_CST>: Likewise.
20608 <VECTOR_CST>: Likewise.
20610 2013-06-01 Janus Weil <janus@gcc.gnu.org>
20611 Mikael Morin <mikael@gcc.gnu.org>
20613 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
20614 * config.in: Regenerated.
20615 * configure: Regenerated.
20617 2013-06-01 Jan Hubicka <jh@suse.cz>
20619 PR middle-end/57366
20620 * cgraphunit.c (compile): When weakref is not supported,
20621 set up transparent aliases before final output pass.
20622 * varasm.c (assemble_alias): Do not try to do it here.
20624 2013-06-01 Jan Hubicka <jh@suse.cz>
20626 PR middle-end/57467
20627 * passes.c (for_per_function): Skip unanalyzed functions.
20629 2013-06-01 Jan Hubicka <jh@suse.cz>
20631 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
20632 (lto_symtab_merge_symbols_1): ... this one.
20633 (lto_symtab_merge_cgraph_nodes): Rename to ...
20634 (lto_symtab_merge_symbols): ... this one; simplify.
20635 * cgraph.c (same_body_aliases_done): Rename to ...
20636 (cpp_implicit_aliases_done): ... this one.
20637 (cgraph_create_function_alias): Update.
20638 (cgraph_same_body_alias): Update.
20639 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
20640 (verify_edge_corresponds_to_fndecl): Simplify.
20641 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
20642 (cgraph_node): Remove same_body_alias.
20643 (varpool_node): Remove alias_of and extra_name_alias.
20644 (same_body_aliases_done): Rename to ..
20645 (cpp_implicit_aliases_done): ... this one.
20646 (symtab_alias_ultimate_target): Add default parameter.
20647 (symtab_resolve_alias): New function.
20648 (fixup_same_cpp_alias_visibility): Declare.
20649 (cgraph_function_node): Add default parameter.
20650 (cgraph_node_asm_name): Likewise.
20651 (cgraph_function_or_thunk_node): Add default parameter; do
20652 not ICE when it is NULL.
20653 (varpool_variable_node): Likewise.
20654 * tree-emutls.c (create_emultls_var): Update.
20655 (ipa_lower_emutls): Update.
20656 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
20657 (cgraph_reset_node): Reset alias info.
20658 (cgraph_finalize_function): Update.
20659 (fixup_same_cpp_alias_visibility): Move to symtab.c.
20660 (analyze_function): Simplify.
20661 (cgraph_process_same_body_aliases): Simplify.
20662 (analyze_functions): Fixup same body aliases.
20663 (handle_alias_pairs): Simplify.
20664 (assemble_thunk): Update.
20665 (assemble_thunks_and_aliases): Update.
20666 (output_weakrefs): Rewrite.
20667 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
20668 (lto_output_varpool_node): Likewise.
20669 (compute_ltrans_boundary): Remve assert.
20670 (get_alias_symbol): New functoin.
20671 (input_node): Rewrite alias handling.
20672 (input_varpool_node): Likewise.
20673 * ipa-pure-const.c (propagate_pure_const): Fix formating.
20674 * ipa.c (process_references): Handle weakrefs correctly.
20675 (symtab_remove_unreachable_nodes): Likewise.
20676 * trans-mem.c (get_cg_data): Update.
20677 (ipa_tm_create_version_alias): Update.
20678 (ipa_tm_execute): Update.
20679 * symtab.c (dump_symtab_base): Dump aliases.
20680 (verify_symtab_base): Verify aliases.
20681 (symtab_node_availability): New function.
20682 (symtab_alias_ultimate_target): Simplify.
20683 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
20684 handle all the fixup cases.
20685 (symtab_resolve_alias): New function.
20686 * passes.c (ipa_write_summaries): Handle weakrefs.
20687 * varpool.c (varpool_analyze_node): Simplify.
20688 (assemble_aliases): Update.
20689 (varpool_create_variable_alias): Simplify.
20690 (varpool_extra_name_alias): Simplify.
20691 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
20692 (lto_symtab_merge_symbols): ... this one.
20694 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
20697 PR rtl-optimization/57268
20698 * sched-deps.c (sched_analyze_2): Flush dependence lists if
20699 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
20701 2013-06-01 Tobias Burnus <burnus@net-b.de>
20703 Partially reverted:
20704 2013-05-31 Tobias Burnus <burnus@net-b.de>
20706 PR middle-end/57073
20707 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
20710 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
20712 PR rtl-optimization/57268
20713 * sched-deps.c (sched_analyze_2): Flush dependence lists if
20714 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
20716 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
20718 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
20719 unordered comparison operators when -fno-trapping-math is in effect
20721 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
20722 and implement unordered comparison operators properly on the e500.
20724 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
20726 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
20727 for constant scalar integers.
20728 (simplify_relational_operation_1): Likewise.
20730 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
20732 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
20733 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
20736 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
20737 Igor Zamyatin <igor.zamyatin@intel.com>
20739 Silvermont (SLM) architecture performance tuning.
20740 * config/i386/i386.h (enum ix86_tune_indices): Add
20741 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
20742 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
20744 * config/i386/i386.c (initial_ix86_tune_features)
20745 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
20746 (ix86_lea_outperforms): Handle Silvermont tuning.
20747 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
20749 (ix86_use_lea_for_mov): Likewise.
20750 (ix86_avoid_lea_for_addr): Likewise.
20751 (ix86_lea_for_add_ok): Likewise.
20752 (exact_dependency_1): New function.
20753 (exact_store_load_dependency): Likewise.
20754 (ix86_adjust_cost): Handle Silvermont tuning.
20755 (do_reoder_for_imul): Likewise.
20756 (swap_top_of_ready_list): New function.
20757 (ix86_sched_reorder): Changed to handle Silvermont tuning.
20759 * config/i386/i386.md (peepholes that split memory operand in fp
20762 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
20764 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
20765 Remove un-necessary braces.
20767 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
20769 * config/aarch64/aarch64.c (aarch64_classify_symbol):
20770 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
20772 2013-05-31 Tobias Burnus <burnus@net-b.de>
20774 PR middle-end/57073
20775 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
20777 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20780 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
20781 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
20782 * config/arm/neon.md (iordi3_neon): Remove.
20783 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
20784 * config/arm/predicates.md (imm_for_neon_logic_operand):
20785 Move to earlier in the file.
20786 (neon_logic_op2): Likewise.
20787 (arm_iordi_operand_neon): New predicate.
20789 2013-05-31 Richard Biener <rguenther@suse.de>
20791 PR tree-optimization/57478
20792 PR tree-optimization/57453
20793 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
20796 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
20798 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
20799 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
20801 2013-05-30 Tobias Burnus <burnus@net-b.de>
20802 Thomas Koenig <tkoenig@gcc.gnu.org>
20804 PR middle-end/57073
20805 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
20806 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
20808 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
20810 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
20812 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
20814 * target.def (register_usage_leveling_p): New hook.
20815 * targhooks.c (default_register_usage_leveling_p): New.
20816 * targhooks.h (default_register_usage_leveling_p): New prototype.
20817 * lra-assigns.c (register_usage_leveling_p): Use the hook.
20818 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
20819 * doc/tm.texi: Update.
20820 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
20822 2013-05-30 Ian Bolton <ian.bolton@arm.com>
20824 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
20825 (*insv_reg<mode>): New define_insn.
20827 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
20829 PR rtl-optimization/57439
20830 * postreload.c (move2add_valid_value_p): Check that we have
20831 a zero subreg_regno_offset when accessing the register in
20832 the requested mode.
20834 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
20835 Igor Zamyatin <igor.zamyatin@intel.com>
20837 Silvermont (SLM) architecture pipeline model, tuning and
20839 * config.gcc: Add slm config options and target.
20841 * config/i386/slm.md: New.
20843 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
20845 * config/i386/i386-c.c (ix86_target_macros_internal): New case
20847 (ix86_target_macros_internal): Likewise.
20849 * config/i386/i386.c (slm_cost): New cost.
20850 (m_SLM): New macro flag.
20851 (initial_ix86_tune_features): Set m_SLM.
20852 (x86_accumulate_outgoing_args): Likewise.
20853 (x86_arch_always_fancy_math_387): Likewise.
20854 (processor_target_table): Add slm cost.
20855 (cpu_names): Add slm cpu name.
20856 (x86_option_override_internal): Set SLM ISA.
20857 (ix86_issue_rate): New case PROCESSOR_SLM.
20858 (ia32_multipass_dfa_lookahead): Likewise.
20859 (fold_builtin_cpu): Add slm.
20861 * config/i386/i386.h (TARGET_SLM): New target macro.
20862 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
20863 (processor_type): Add PROCESSOR_SLM.
20865 * config/i386/i386.md (cpu): Add new value "slm".
20866 (slm.md): Include slm.md.
20868 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
20869 Zhenqiang Chen <zhenqiang.chen@linaro.org>
20871 * config/arm/arm-protos.h: Add and update function protos.
20872 * config/arm/arm.c (use_simple_return_p): New added.
20873 (thumb2_expand_return): Check simple_return flag.
20874 * config/arm/arm.md: Add simple_return and conditional simple_return.
20875 * config/arm/iterators.md: Add iterator for return and simple_return.
20877 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
20879 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
20880 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
20881 (arm_emit_vfp_multi_reg_pop): Likewise.
20882 (thumb2_emit_ldrd_pop): Likewise.
20883 (arm_expand_epilogue): Add misc REG_CFA notes.
20884 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
20886 2013-05-29 Lawrence Crowl <crowl@google.com>
20888 * config/arm/t-arm: Update for below.
20890 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
20891 Change type to hash_table. Update dependent calls and types.
20893 * config/i386/t-cygming: Update for below.
20895 * config/i386/t-interix: Update for below.
20897 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
20898 Change type to hash_table. Update dependent calls and types.
20899 (i386_find_on_wrapper_list::wrappers): Likewise.
20901 * config/ia64/t-ia64: Update for below.
20903 * config/ia64/ia64.c (bundle_state_table):
20904 Change type to hash_table. Update dependent calls and types.
20906 * config/mips/mips.c (mips_reorg_process_insns::htab):
20907 Change type to hash_table. Update dependent calls and types.
20909 * config/sol2.c (solaris_comdat_htab):
20910 Change type to hash_table. Update dependent calls and types.
20912 * config/t-sol2: Update for above.
20914 2013-05-29 Teresa Johnson <tejohnson@google.com>
20916 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
20917 functions are not yet marked as defined.
20919 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
20920 Pat Haugen <pthaugen@us.ibm.com>
20921 Peter Bergner <bergner@vnet.ibm.com>
20923 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
20927 (vrotl<mode>3): Likewise.
20928 (vashl<mode>3): Likewise.
20929 (vlshr<mode>3): Likewise.
20930 (vashr<mode>3): Likewise.
20932 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
20933 support for power8 V2DI builtins.
20935 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
20936 power8 V2DI builtins.
20937 (vupkhsw): Likewise.
20938 (vupklsw): Likewise.
20939 (vaddudm): Likewise.
20940 (vminsd): Likewise.
20941 (vmaxsd): Likewise.
20942 (vminud): Likewise.
20943 (vmaxud): Likewise.
20944 (vpkudum): Likewise.
20945 (vpksdss): Likewise.
20946 (vpkudus): Likewise.
20947 (vpksdus): Likewise.
20952 (vsubudm): Likewise.
20953 (vcmpequd): Likewise.
20954 (vcmpgtsd): Likewise.
20955 (vcmpgtud): Likewise.
20956 (vcmpequd_p): Likewise.
20957 (vcmpgtsd_p): Likewise.
20958 (vcmpgtud_p): Likewise.
20959 (vupkhsw): Likewise.
20960 (vupklsw): Likewise.
20961 (vaddudm): Likewise.
20962 (vmaxsd): Likewise.
20963 (vmaxud): Likewise.
20964 (vminsd): Likewise.
20965 (vminud): Likewise.
20966 (vpksdss): Likewise.
20967 (vpksdus): Likewise.
20968 (vpkudum): Likewise.
20969 (vpkudus): Likewise.
20974 (vsubudm): Likewise.
20976 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
20977 support for power8 V2DI instructions.
20979 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
20980 power8 V2DI instructions. Combine pack and unpack insns to use an
20981 iterator for each mode. Check whether a particular mode supports
20982 Altivec instructions instead of just checking TARGET_ALTIVEC.
20983 (UNSPEC_VPKUWUM): Likewise.
20984 (UNSPEC_VPKSHSS): Likewise.
20985 (UNSPEC_VPKSWSS): Likewise.
20986 (UNSPEC_VPKUHUS): Likewise.
20987 (UNSPEC_VPKSHUS): Likewise.
20988 (UNSPEC_VPKUWUS): Likewise.
20989 (UNSPEC_VPKSWUS): Likewise.
20990 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
20991 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
20992 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
20993 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
20994 (UNSPEC_VUPKHSB): Likewise.
20995 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
20996 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
20997 (UNSPEC_VUPKHSH): Likewise.
20998 (UNSPEC_VUPKLSB): Likewise.
20999 (UNSPEC_VUPKLSH): Likewise.
21001 (VI_char): Likewise.
21002 (VI_scalar): Likewise.
21003 (VI_unit): Likewise.
21005 (VP_small): Likewise.
21006 (VP_small_lc): Likewise.
21007 (VU_char): Likewise.
21008 (add<mode>3): Likewise.
21009 (altivec_vaddcuw): Likewise.
21010 (altivec_vaddu<VI_char>s): Likewise.
21011 (altivec_vadds<VI_char>s): Likewise.
21012 (sub<mode>3): Likewise.
21013 (altivec_vsubcuw): Likewise.
21014 (altivec_vsubu<VI_char>s): Likewise.
21015 (altivec_vsubs<VI_char>s): Likewise.
21016 (altivec_vavgs<VI_char>): Likewise.
21017 (altivec_vcmpbfp): Likewise.
21018 (altivec_eq<mode>): Likewise.
21019 (altivec_gt<mode>): Likewise.
21020 (altivec_gtu<mode>): Likewise.
21021 (umax<mode>3): Likewise.
21022 (smax<mode>3): Likewise.
21023 (umin<mode>3): Likewise.
21024 (smin<mode>3): Likewise.
21025 (altivec_vpkuhum): Likewise.
21026 (altivec_vpkuwum): Likewise.
21027 (altivec_vpkshss): Likewise.
21028 (altivec_vpkswss): Likewise.
21029 (altivec_vpkuhus): Likewise.
21030 (altivec_vpkshus): Likewise.
21031 (altivec_vpkuwus): Likewise.
21032 (altivec_vpkswus): Likewise.
21033 (altivec_vpks<VI_char>ss): Likewise.
21034 (altivec_vpks<VI_char>us): Likewise.
21035 (altivec_vpku<VI_char>us): Likewise.
21036 (altivec_vpku<VI_char>um): Likewise.
21037 (altivec_vrl<VI_char>): Likewise.
21038 (altivec_vsl<VI_char>): Likewise.
21039 (altivec_vsr<VI_char>): Likewise.
21040 (altivec_vsra<VI_char>): Likewise.
21041 (altivec_vsldoi_<mode>): Likewise.
21042 (altivec_vupkhsb): Likewise.
21043 (altivec_vupkhs<VU_char>): Likewise.
21044 (altivec_vupkls<VU_char>): Likewise.
21045 (altivec_vupkhsh): Likewise.
21046 (altivec_vupklsb): Likewise.
21047 (altivec_vupklsh): Likewise.
21048 (altivec_vcmpequ<VI_char>_p): Likewise.
21049 (altivec_vcmpgts<VI_char>_p): Likewise.
21050 (altivec_vcmpgtu<VI_char>_p): Likewise.
21051 (abs<mode>2): Likewise.
21052 (vec_unpacks_hi_v16qi): Likewise.
21053 (vec_unpacks_hi_v8hi): Likewise.
21054 (vec_unpacks_lo_v16qi): Likewise.
21055 (vec_unpacks_hi_<VP_small_lc>): Likewise.
21056 (vec_unpacks_lo_v8hi): Likewise.
21057 (vec_unpacks_lo_<VP_small_lc>): Likewise.
21058 (vec_pack_trunc_v8h): Likewise.
21059 (vec_pack_trunc_v4si): Likewise.
21060 (vec_pack_trunc_<mode>): Likewise.
21062 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
21064 (vec_vmaxsd): Likewise.
21065 (vec_vmaxud): Likewise.
21066 (vec_vminsd): Likewise.
21067 (vec_vminud): Likewise.
21068 (vec_vpksdss): Likewise.
21069 (vec_vpksdus): Likewise.
21070 (vec_vpkudum): Likewise.
21071 (vec_vpkudus): Likewise.
21072 (vec_vrld): Likewise.
21073 (vec_vsld): Likewise.
21074 (vec_vsrad): Likewise.
21075 (vec_vsrd): Likewise.
21076 (vec_vsubudm): Likewise.
21077 (vec_vupkhsw): Likewise.
21078 (vec_vupklsw): Likewise.
21080 2013-05-29 Jan Hubicka <jh@suse.cz>
21082 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
21083 flags; reorder rest of fields in more consistent way.
21084 (varpool_node): Remove analyzed, finalized and alias.
21085 (cgraph_ndoe): Likewise.
21086 (symtab_alias_ultimate_target): New function.
21087 (cgraph_function_node): Move offline.
21088 (cgraph_reset_node): Declare.
21089 (cgraph_comdat_can_be_unshared_p): Remove.
21090 (varpool_remove_initializer): Declare.
21091 (varpool_first_defined_variable, varpool_next_defined_variable
21092 cgraph_first_defined_function, cgraph_next_defined_function): Update.
21093 (cgraph_function_with_gimple_body_p): Update.
21094 (varpool_all_refs_explicit_p): Update.
21095 (symtab_alias_target): New function.
21096 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
21097 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
21098 (cgraph_function_or_thunk_node): Simplify using
21099 symtab_alias_ultimate_target.
21100 (varpool_variable_node): Likewise.
21101 * cgraph.c (cgraph_create_function_alias): Update.
21102 (cgraph_add_thunk): Update.
21103 (cgraph_remove_node): Update.
21104 (dump_cgraph_node): Do not dump removed flags.
21105 (cgraph_function_body_availability): Update.
21106 (cgraph_propagate_frequency): Update.
21107 (verify_cgraph_node): Check sanity of local flag.
21108 (cgraph_function_node): Move here from cgraph.h; revamp for
21109 cgraph_function_or_thunk_node.
21110 * lto-symtab.c (lto_varpool_replace_node): Update.
21111 (lto_symtab_resolve_can_prevail_p): Update.
21112 (lto_symtab_merge_cgraph_nodes): Update.
21113 * ipa-cp.c (determine_versionability, initialize_node_lattices,
21114 propagate_constants_accross_call, devirtualization_time_bonus,
21115 ipcp_propagate_stage): Update.
21116 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
21117 * ipa-inline-transform.c (clone_inlined_nodes,
21118 preserve_function_body_p): Update.
21119 * ipa-reference.c (propagate): Update.
21120 (write_node_summary_p): Update.
21121 * toplev.c (wrapup_global_declaration_2): Update.
21122 * cgraphunit.c (cgraph_analyze_function): Rename to ...
21123 (analyze_function) ... this one.
21124 (cgraph_process_new_functions): Update.
21125 (cgraph_reset_node): Export.
21126 (cgraph_finalize_function): Update.
21127 (cgraph_add_new_function): Update.
21128 (process_function_and_variable_attributes): Update.
21129 (varpool_finalize_decl): Update.
21130 (symbol_finalized): Remove.
21131 (symbol_finalized_and_needed): Rename to ...
21132 (symbol_defined_and_needed): ... update.
21133 (cgraph_analyze_functions): Update.
21134 (handle_alias_pairs): Update.
21135 (mark_functions_to_output): Update.
21136 (assemble_thunk): Update.
21137 (output_in_order): Update.
21138 (output_weakrefs): Update.
21139 (finalize_compilation_unit): Update.
21140 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
21141 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
21142 input_node, input_varpool_node): Update.
21143 * dbxout.c (dbxout_expand_expr): Update.
21144 * cgraphclones.c (cgraph_clone_node): Update.
21145 (cgraph_copy_node_for_versioning): Update.
21146 (cgraph_materialize_clone): Update.
21147 (cgraph_materialize_all_clones): Update.
21148 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
21149 propagate_pure_const, propagate_nothrow): Update.
21150 * lto-streamer-out.c (lto_output, write_symbol): Update.
21151 * ipa-utils.c (ipa_reverse_postorder): Update.
21152 * ipa-inline.c (can_inline_edge_p): Update.
21153 (update_caller_keys, ipa_inline): Update.
21154 * dwarf2out.c (reference_to_unused,
21155 premark_types_used_by_global_vars_helper): Update.
21156 * tree-eh.c (tree_could_trap_p): Update.
21157 * ipa-split.c (consider_split, execute_split_functions): Update.
21158 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
21159 has_addr_references_p): Update; move ahead in file for better
21161 (process_references): Simplify.
21162 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
21163 bodies are removed.
21164 (cgraph_comdat_can_be_unshared_p): Make static.
21165 (cgraph_externally_visible_p): Update.
21166 (varpool_externally_visible_p): Update.
21167 (function_and_variable_visibility): Update.
21168 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
21169 ipa_tm_mark_force_output_node): Update.
21170 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
21171 estimate_edge_devirt_benefit, inline_generate_summary,
21172 inline_write_summary): Update.
21173 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
21174 * ipa-prop.c (ipa_compute_jump_functions): Update.
21175 (ipa_print_node_params, ipa_prop_read_section,
21176 ipa_update_after_lto_read, read_replacements_section): Update.
21177 * varasm.c (mark_decl_referenced): Update.
21178 (assemble_alias, dump_tm_clone_pairs): Update.
21179 * tree-inline.c (copy_bb): Update.
21180 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
21182 * symtab.c (dump_symtab_base): Print new flags.
21183 (verify_symtab_base): Verify new flags.
21184 (symtab_alias_ultimate_target): New function.
21185 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
21186 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
21188 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
21190 * i386.c (ix86_get_function_versions_dispatcher,
21191 ix86_generate_version_dispatcher_body): Update.
21192 (fold_builtin_cpu): Use varpool_add_new_variable.
21193 * varpool.c (varpool_remove_initializer): Break out from ...
21194 (varpool_remove_node): ... this one.
21195 (dump_varpool_node, varpool_node_for_asm,
21196 cgraph_variable_initializer_availability, varpool_analyze_node,
21197 varpool_assemble_decl, varpool_remove_unreferenced_decls,
21198 varpool_finalize_named_section_flags, varpool_create_variable_alias):
21201 2013-05-29 Jan Hubicka <jh@suse.cz>
21203 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
21205 2013-05-29 Easwaran Raman <eraman@google.com>
21207 PR tree-optimization/57442
21208 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
21209 when control exits the main loop.
21211 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
21213 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
21215 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
21216 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
21217 * rx/t-rx: Add rx100 under multi library matches option for nofpu
21220 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21222 PR tree-optimization/57441
21223 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
21224 Don't limit size of incr_vec to number of candidates.
21226 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
21228 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
21229 and mips16 directories.
21230 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
21231 (MULTILIB_DIRNAMES): Ditto.
21232 (MULTILIB_EXCEPTIONS): Add new exceptions.
21233 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
21234 (MULTILIB_DIRNAMES): Ditto.
21235 (MULTILIB_EXCEPTIONS): Add new exceptions.
21237 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
21238 Marcus Shawcroft <marcus.shawcroft@arm.com>
21240 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
21241 SYMBOL_TINY_ABSOLUTE.
21242 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
21243 SYMBOL_TINY_ABSOLUTE.
21244 (aarch64_expand_mov_immediate): Likewise.
21245 (aarch64_classify_symbol): Likewise.
21246 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
21247 Permit SYMBOL_TINY_ABSOLUTE.
21248 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
21250 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
21251 Marcus Shawcroft <marcus.shawcroft@arm.com>
21253 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
21254 Refactor if/switch. Replace gcc_assert with if.
21256 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
21258 * config/i386/i386.c (initial_ix86_tune_features): Enable
21259 FP Reassociation for AMD bdver1 and bdver2.
21261 2013-05-29 Martin Jambor <mjambor@suse.cz>
21263 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
21264 and IMAGPART_EXPR do not occur within other handled_components.
21266 2013-05-29 Richard Biener <rguenther@suse.de>
21268 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
21269 access on whether the use is in the BB we currently try to
21271 (vect_bb_vectorization_profitable_p): Pass the BB we currently
21272 vectorize to vect_bb_slp_scalar_cost.
21274 2013-05-29 Richard Biener <rguenther@suse.de>
21276 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
21277 computing scalar cost offsetted by stmts that are kept live
21279 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
21280 for computation of scalar cost.
21282 2013-05-28 Steve Ellcey <sellcey@mips.com>
21284 * config/mips/mips-cpus.def (mips32r2): Change processor type.
21286 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
21288 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
21289 array notation built-in reduction functions.
21290 * doc/passes.texi (Passes): Added documentation about changes done
21292 * doc/invoke.texi (C Dialect Options): Added documentation about
21293 the -fcilkplus flag.
21294 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
21295 (BUILTINS_DEF): Depend on cilkplus.def.
21296 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
21297 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
21298 * cilkplus.def: New file.
21300 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
21302 PR rtl-optimization/57439
21303 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
21305 2013-05-28 Easwaran Raman <eraman@google.com>
21307 PR tree-optimization/57337
21308 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
21309 (find_insert_point): Correctly identify the insertion point
21310 when two statements with the same UID is compared.
21312 2013-05-28 Richard Biener <rguenther@suse.de>
21314 PR tree-optimization/56787
21315 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
21316 from the list of data references.
21317 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
21319 (vect_analyze_loop_operations): Likewise.
21320 (vect_transform_loop): Remove clobbers.
21322 2013-05-28 Martin Jambor <mjambor@suse.cz>
21324 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
21325 and REALPART_EXPRs have scalar type.
21327 2013-05-28 Richard Biener <rguenther@suse.de>
21329 PR tree-optimization/57411
21330 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
21332 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
21333 virtual operand propagation.
21335 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
21337 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
21338 destination register for bmasksi_vis.
21339 (vector_init_bshuffle): Likewise.
21340 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
21342 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
21344 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
21345 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
21346 mode if the instruction isn't available in the original mode.
21347 * config/sparc/sparc.opt (mfix-ut699): New option.
21348 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
21349 (divdf3): Turn into expander.
21350 (divdf3_nofix): New insn.
21351 (divdf3_fix): Likewise.
21352 (divsf3): Disable if -mfix-ut699.
21353 (sqrtdf2): Turn into expander.
21354 (sqrtdf2_nofix): New insn.
21355 (sqrtdf2_fix): Likewise.
21356 (sqrtsf2): Disable if -mfix-ut699.
21358 2013-05-27 Richard Biener <rguenther@suse.de>
21360 PR middle-end/57412
21361 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
21362 block for the new loop.
21364 2013-05-27 Richard Biener <rguenther@suse.de>
21366 PR tree-optimization/57343
21367 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
21368 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
21369 (number_of_iterations_cond): Do not build the folded tree.
21371 2013-05-27 Richard Biener <rguenther@suse.de>
21374 PR middle-end/57381
21375 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
21376 OEP_CONSTANT_ADDRESS_OF retained.
21378 PR tree-optimization/57417
21379 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
21380 for unchanged base.
21381 (set_ssa_val_to): Compare addresses using
21382 get_addr_base_and_unit_offset.
21384 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
21386 PR rtl-optimization/56833
21387 * postreload.c (move2add_record_mode): New function.
21388 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
21389 (move2add_use_add2_insn): Use move2add_record_sym_value.
21390 (move2add_use_add3_insn): Likewise.
21391 (reload_cse_move2add): Use move2add_valid_value_p and
21392 move2add_record_mode. Invalidate call-clobbered and REG_INC
21393 affected regs by setting reg_mode to VOIDmode.
21394 (move2add_note_store): Don't pretend the inside of a SUBREG is
21395 the actual destination. Invalidate single/leading registers by
21396 setting reg_mode to VOIDmode.
21397 Use move2add_record_sym_value, move2add_valid_value_p and
21398 move2add_record_mode.
21400 2013-05-27 Richard Biener <rguenther@suse.de>
21402 PR tree-optimization/57396
21403 * tree-affine.c (double_int_constant_multiple_p): Properly
21404 return false for val == 0 and div != 0.
21406 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
21408 * config/mips/mips.h: Use #elif in preprocessor conditions.
21410 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
21413 * config/mips/constraints.md (kl): New constraint.
21414 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
21415 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
21416 constraint for operand 0. Split after CSE for MIPS16. Emit a move
21417 from LO for MIPS16.
21418 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
21420 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
21423 * config/mips/mips.c (mips_can_inline_p): New function.
21424 (TARGET_CAN_INLINE_P): Define.
21426 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
21428 * sched-int.h (ds_t, dw_t): Make unsigned int.
21429 Fix documentation that describes how all the ds_t bits are used.
21430 Reserve the last bit for delayed-branch scheduling.
21431 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
21432 (BITS_PER_DEP_WEAK): Fix definition and documentation.
21433 (gen_dep_weak_1): Remove prototype.
21434 * sched-deps.c (get_dep_weak_1): Make static.
21435 * target.def (speculate_insn, needs_block_p, gen_spec_check,
21436 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
21437 * doc/tm.texi: Regenerate.
21438 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
21440 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
21443 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
21445 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
21446 Sandra Loosemore <sandra@codesourcery.com>
21448 * config.gcc (powerpc-*): Allow native for with-cpu.
21450 2013-05-24 Jeff Law <law@redhat.com>
21452 PR tree-optimization/57124
21453 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
21454 conversion feeding a condition if the range has an overflow
21455 if -fstrict-overflow. Add warnings for when we do make the
21458 2013-05-24 Dehao Chen <dehao@google.com>
21460 * tree-cfg.c (locus_discrim_map): Fix the typo.
21461 (locus_discrim_hasher): Likewise.
21462 (locus_discrim_hasher::hash): Likewise.
21463 (locus_discrim_hasher::equal): Likewise.
21465 2013-05-24 Martin Jambor <mjambor@suse.cz>
21467 PR tree-optimization/57294
21468 * cgraph.h (ipa_record_stmt_references): Declare.
21469 * cgraphbuild.c (ipa_record_stmt_references): New function.
21470 (build_cgraph_edges): Use ipa_record_stmt_references.
21471 (rebuild_cgraph_edges): Likewise.
21472 (cgraph_rebuild_references): Likewise.
21473 * ipa-prop.c (ipa_modify_call_arguments): Discard references
21474 associated with the old statement and build references from the
21475 newly built statements.
21476 * ipa-ref.c (ipa_remove_stmt_references): New function.
21477 * ipa-ref.h (ipa_remove_stmt_references): Declare.
21479 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
21481 * lra-constraints.c (emit_spill_move): Use smaller mode for
21483 (check_and_process_move): Consider mem-reg moves for secondary
21485 (curr_insn_transform): Don't lose insns emitted before for
21486 secondary memory moves.
21487 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
21488 reg set up in the current insn.
21490 2013-05-24 Dehao Chen <dehao@google.com>
21492 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
21494 (locus_descrim_hasher::equal): Likewise.
21495 (build_gimple_cfg): New discriminator assignment algorithm.
21496 (make_edges): Likewise.
21497 (next_discriminator_for_locus): Likewise.
21498 (same_line_p): Likewise.
21499 (assign_discriminators): Likewise.
21500 (make_cond_expr_edges): Likewise.
21501 (make_gimple_switch_edges): Likewise.
21502 (make_goto_expr_edges): Likewise.
21503 (make_gimple_asm_edges): Likewise.
21505 2013-05-24 Ian Bolton <ian.bolton@arm.com>
21507 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
21508 X format specifier to only display bottom 16 bits.
21509 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
21510 immediate to match for operand 2, since it will be masked.
21512 2013-05-24 Richard Biener <rguenther@suse.de>
21514 PR tree-optimization/57287
21515 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
21516 all SSA names that occur in abnormal PHIs.
21518 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
21521 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
21522 that index is not negative.
21524 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
21526 PR rtl-optimization/55177
21527 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
21528 (simplify_byte_swapping_operation): New.
21529 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
21530 (simplify_relational_operation_1): Deal with BSWAP.
21532 2013-05-23 Richard Henderson <rth@redhat.com>
21535 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
21536 (ix86_reorg): Call it.
21538 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
21541 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
21542 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
21543 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
21545 2013-05-23 Christian Bruel <christian.bruel@st.com>
21548 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
21550 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
21551 Marcus Shawcroft <marcus.shawcroft@arm.com>
21553 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
21554 * config/aarch64/constraints.md (Usa): Remove.
21555 * doc/md.texi (AArch64 Usa): Remove.
21557 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
21558 Marcus Shawcroft <marcus.shawcroft@arm.com>
21560 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
21561 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
21562 * config/aarch64/predicates.md (aarch64_const_address): Remove.
21563 (aarch64_mov_operand): Use aarch64_mov_operand_p.
21565 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
21567 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
21568 instruction (AdvSIMD).
21569 * config/aarch64/aarch64-builtins.c
21570 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
21571 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
21573 2013-05-23 Martin Jambor <mjambor@suse.cz>
21575 PR middle-end/57347
21576 * tree.h (contains_bitfld_component_ref_p): Declare.
21577 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
21578 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
21580 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
21581 not access a bit-field. Assert all final offsets are byte-aligned.
21583 2013-05-23 Richard Biener <rguenther@suse.de>
21585 PR tree-optimization/57380
21586 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
21587 least one invariant or re-used load.
21588 * passes.c (init_optimization_passes): Move pass_phiprop before
21591 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
21593 * config/aarch64/aarch64-simd.md
21594 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
21596 2013-05-23 Richard Biener <rguenther@suse.de>
21598 PR middle-end/57381
21599 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
21600 OEP_CONSTANT_ADDRESS_OF retained.
21602 2013-05-23 Jakub Jelinek <jakub@redhat.com>
21604 PR middle-end/57344
21605 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
21606 don't lower unit. Handle unit not being always BITS_PER_WORD.
21608 2013-05-23 Richard Biener <rguenther@suse.de>
21610 PR rtl-optimization/57341
21611 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
21612 instead of true_dependence.
21614 2013-05-22 David Malcolm <dmalcolm@redhat.com>
21616 * bb-reorder.c (branch_threshold): Make const.
21617 (exec_threshold): Ditto.
21619 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
21620 Pat Haugen <pthaugen@us.ibm.com>
21621 Peter Bergner <bergner@vnet.ibm.com>
21623 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
21624 documentation for the power8 crypto builtins.
21626 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
21628 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
21629 macros for defining power8 builtin functions.
21630 (BU_P8V_AV_2): Likewise.
21631 (BU_P8V_AV_P): Likewise.
21632 (BU_P8V_VSX_1): Likewise.
21633 (BU_P8V_OVERLOAD_1): Likewise.
21634 (BU_P8V_OVERLOAD_2): Likewise.
21635 (BU_CRYPTO_1): Likewise.
21636 (BU_CRYPTO_2): Likewise.
21637 (BU_CRYPTO_3): Likewise.
21638 (BU_CRYPTO_OVERLOAD_1): Likewise.
21639 (BU_CRYPTO_OVERLOAD_2): Likewise.
21640 (XSCVSPDP): Fix typo, point to the correct instruction.
21641 (VCIPHER): Add power8 crypto builtins.
21642 (VCIPHERLAST): Likewise.
21643 (VNCIPHER): Likewise.
21644 (VNCIPHERLAST): Likewise.
21645 (VPMSUMB): Likewise.
21646 (VPMSUMH): Likewise.
21647 (VPMSUMW): Likewise.
21648 (VPERMXOR_V2DI): Likewise.
21649 (VPERMXOR_V4SI: Likewise.
21650 (VPERMXOR_V8HI: Likewise.
21651 (VPERMXOR_V16QI: Likewise.
21652 (VSHASIGMAW): Likewise.
21653 (VSHASIGMAD): Likewise.
21654 (VPMSUM): Likewise.
21655 (VPERMXOR): Likewise.
21656 (VSHASIGMA): Likewise.
21658 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
21659 __CRYPTO__ if the crypto instructions are available.
21660 (altivec_overloaded_builtins): Add support for overloaded power8
21663 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
21664 support for power8 crypto builtins.
21665 (builtin_function_type): Likewise.
21666 (altivec_init_builtins): Add support for builtins that take vector
21667 long long (V2DI) arguments.
21669 * config/rs6000/crypto.md: New file, define power8 crypto
21672 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
21673 Pat Haugen <pthaugen@us.ibm.com>
21674 Peter Bergner <bergner@vnet.ibm.com>
21676 * doc/invoke.texi (Option Summary): Add power8 options.
21677 (RS/6000 and PowerPC Options): Likewise.
21679 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
21680 constraints.md instead of rs6000.h. Reorder w* constraints. Add
21681 wm, wn, wr documentation.
21683 * config/rs6000/constraints.md (wm): New constraint for VSX
21684 registers if direct move instructions are enabled.
21685 (wn): New constraint for no registers.
21686 (wq): New constraint for quad word even GPR registers.
21687 (wr): New constraint if 64-bit instructions are enabled.
21688 (wv): New constraint if power8 vector instructions are enabled.
21689 (wQ): New constraint for quad word memory locations.
21691 * config/rs6000/predicates.md (const_0_to_15_operand): New
21692 constraint for 0..15 for crypto instructions.
21693 (gpc_reg_operand): If VSX allow registers in VSX registers as well
21694 as GPR and floating point registers.
21695 (int_reg_operand): New predicate to match only GPR registers.
21696 (base_reg_operand): New predicate to match base registers.
21697 (quad_int_reg_operand): New predicate to match even GPR registers
21698 for quad memory operations.
21699 (vsx_reg_or_cint_operand): New predicate to allow vector logical
21700 operations in both GPR and VSX registers.
21701 (quad_memory_operand): New predicate for quad memory operations.
21702 (reg_or_indexed_operand): New predicate for direct move support.
21704 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
21705 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
21706 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
21707 (POWERPC_MASKS): Add power8 options.
21708 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
21711 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
21712 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
21714 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
21715 (-mpower8-fusion): New power8 options.
21716 (-mpower8-fusion-sign): Likewise.
21717 (-mpower8-vector): Likewise.
21718 (-mcrypto): Likewise.
21719 (-mdirect-move): Likewise.
21720 (-mquad-memory): Likewise.
21722 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
21723 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
21725 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
21726 (rs6000_debug_vector_unit): Add p8_vector.
21727 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
21728 definitions. Also print fusion state.
21729 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
21730 (rs6000_builtin_mask_calculate): Add power8 builtin support.
21731 (rs6000_option_override_internal): Add support for power8.
21732 (rs6000_common_init_builtins): Add debugging for skipped builtins
21733 if -mdebug=builtin.
21734 (rs6000_adjust_cost): Add power8 support.
21735 (rs6000_issue_rate): Likewise.
21736 (insn_must_be_first_in_group): Likewise.
21737 (insn_must_be_last_in_group): Likewise.
21738 (force_new_group): Likewise.
21739 (rs6000_register_move_cost): Likewise.
21740 (rs6000_opt_masks): Likewise.
21742 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
21743 power8 capable assembler, default to power7 options.
21744 (TARGET_DIRECT_MOVE): Likewise.
21745 (TARGET_CRYPTO): Likewise.
21746 (TARGET_P8_VECTOR): Likewise.
21747 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
21748 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
21749 (VECTOR_MEM_P8_VECTOR_P): Likewise.
21750 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
21751 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
21752 (TARGET_XSCVDPSPN): Likewise.
21753 (TARGET_XSCVSPDPN): Likewsie.
21754 (TARGET_SYNC_HI_QI): Likewise.
21755 (TARGET_SYNC_TI): Likewise.
21756 (MASK_CRYPTO): Likewise.
21757 (MASK_DIRECT_MOVE): Likewise.
21758 (MASK_P8_FUSION): Likewise.
21759 (MASK_P8_VECTOR): Likewise.
21760 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
21761 temporary used by some of the direct move instructions to get two FP
21762 temporary registers does not force creation of a stack frame.
21763 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
21764 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
21765 that any VSX registers are tieable, even if they are also an
21766 Altivec vector mode.
21767 (r6000_reg_class_enum): Add wm, wr, wv constraints.
21768 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
21769 (RS6000_BTM_CRYPTO): Likewise.
21770 (RS6000_BTM_COMMON): Likewise.
21772 * config/rs6000/rs6000.md (cpu attribute): Add power8.
21773 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
21774 (enum rs6000_vector): Add power8 vector support.
21776 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
21780 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
21781 (any_sibcall_could_use_r3): this and handle indirect calls.
21782 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
21784 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21786 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
21788 2013-05-22 Richard Biener <rguenther@suse.de>
21790 PR middle-end/57349
21791 * profile.c (branch_prob): Do not split blocks that are
21792 abnormally receiving from ECF_RETURNS_TWICE functions.
21794 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
21796 * recog.c (offsettable_address_addr_space_p): Fix calculation of
21797 address mode. Move pointer mode initialization to the same place.
21799 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
21801 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
21802 while it has any effect.
21804 2013-05-21 Easwaran Raman <eraman@google.com>
21806 PR tree-optimization/57322
21807 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
21808 UID of the statement added to the BB to be 1.
21810 2013-05-21 Jakub Jelinek <jakub@redhat.com>
21812 PR tree-optimization/57331
21813 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
21814 of conversion from pointer type to integral type with integer.
21816 2013-05-21 Martin Jambor <mjambor@suse.cz>
21819 * ipa-prop.c (ipa_read_node_info): Process param_used and
21820 controlled_uses in the same order as when writing.
21822 2013-05-21 Magnus Granberg <baldrick@free.fr>
21825 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
21827 2013-05-21 Richard Biener <rguenther@suse.de>
21829 PR tree-optimization/57318
21830 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
21831 estimate stmts with side-effects as likely eliminated.
21833 2013-05-21 Richard Biener <rguenther@suse.de>
21835 PR tree-optimization/57330
21836 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
21837 preserve the call stmts fntype.
21839 2013-05-21 Richard Biener <rguenther@suse.de>
21841 PR tree-optimization/57303
21842 * tree-ssa-sink.c (statement_sink_location): Improve killing
21843 stmt detection and properly handle self-assignments.
21845 2013-05-21 Christian Bruel <christian.bruel@st.com>
21847 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
21848 spanning registers. LEAF_REG_REMAP is supported only for contiguous
21849 registers. Set register size out of the PARALLEL loop.
21851 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
21854 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
21855 (*fmasf4, *fmasf4_media): New insns.
21857 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
21859 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
21860 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
21861 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
21862 (mips_idiv_insns): Update the comments to say that the returned
21863 instruction counts are in units of BASE_INSN_LENGTH.
21864 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
21865 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
21866 using 2 rather than 4 as the length of indirect MIPS16 and
21867 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
21868 length of a NOP. Don't divide MIPS16 lengths by 2.
21869 (mips16_split_long_branches): Assume a branch is long if the
21870 length is greater than 4 rather than 8.
21871 * config/mips/mips.md (length): Give MIPS16 lengths directly,
21872 rather than multiplying them by 2. Multiply instruction counts
21873 by BASE_INSN_LENGTH rather than 4.
21874 (*jump_mips16, tls_get_tp_mips16_<mode>)
21875 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
21877 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
21879 * config/mips/mips.md (extended_mips16): Remove branch case.
21880 (length): Remove duplicated extended_mips16 test.
21882 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
21884 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
21886 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
21888 * recog.h (Recog_data): Rename to...
21889 (recog_data_d): ...this.
21890 (recog_data): Update accordingly.
21891 * recog.c (recog_data): Likewise.
21892 * reload.c (save_recog_data): Likewise.
21893 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
21894 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
21896 2013-05-17 Julian Brown <julian@codesourcery.com>
21898 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
21899 found in a REG_EQUAL note, invalidate it.
21901 2013-05-17 Easwaran Raman <eraman@google.com>
21903 * tree-ssa-reassoc.c (find_insert_point): New function.
21904 (insert_stmt_after): Likewise.
21905 (get_def_stmt): Likewise.
21906 (ensure_ops_are_available): Likewise.
21907 (not_dominated_by): Likewise.
21908 (rewrite_expr_tree): Do not move statements beyond what is
21909 necessary. Remove call to swap_ops_for_binary_stmt...
21910 (reassociate_bb): ... and move it here.
21911 (build_and_add_sum): Assign UIDs for new statements.
21912 (linearize_expr): Likewise.
21913 (do_reassoc): Renumber gimple statement UIDs.
21915 2013-05-17 Jan Hubicka <jh@suse.cz>
21917 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
21919 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
21920 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
21922 (output_weakrefs): Update.
21924 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
21925 Martin Jambor <mjambor@suse.cz>
21927 PR middle-end/57276
21928 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
21929 value that corresponds to the given aggval is found in values vector.
21931 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
21933 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
21934 sse, sse2, sse3, ssse3 and sse4a flags to options.
21936 2013-05-17 David Malcolm <dmalcolm@redhat.com>
21938 * gengtype-state.c: (s_expr_writer): New class, to handle
21939 prettifying of output layout of s-expressions.
21940 (state_writer): New class, to write out gtype.state.
21941 (state_written_type_count): Move this variable into member data of
21943 (s_expr_writer::s_expr_writer): New code: constructor for new class
21944 (state_writer::state_writer(): ditto
21945 (s_expr_writer::write_new_line): New function
21946 (s_expr_writer::write_any_indent): ditto
21947 (s_expr_writer::begin_s_expr): ditto
21948 (s_expr_writer::end_s_expr): ditto
21949 (write_state_fileloc): convert to method of state_writer...
21950 (state_writer:: write_state_fileloc): ...and use methods of
21951 s_expr_writer to write indentation into the gtype.state output file
21952 to visually represent the hierarchical structure of the list
21954 (write_state_fields): ditto, renaming to...
21955 (state_writer::write_state_fields)
21956 (write_state_a_string): ditto, renaming to...
21957 (state_writer::write_state_a_string)
21958 (write_state_string_option): ditto, renaming to...
21959 (state_writer::write_state_string_option)
21960 (write_state_type_option): ditto, renaming to...
21961 (state_writer::write_state_type_option)
21962 (write_state_nested_option): ditto, renaming to...
21963 (state_writer::write_state_nested_option)
21964 (write_state_option): ditto, renaming to...
21965 (state_writer::write_state_option)
21966 (write_state_options): ditto, renaming to...
21967 (state_writer::write_state_options)
21968 (write_state_lang_bitmap): ditto, renaming to...
21969 (state_writer::write_state_lang_bitmap)
21970 (write_state_version): ditto, renaming to...
21971 (state_writer::write_state_version)
21972 (write_state_scalar_type): ditto, renaming to...
21973 (state_writer::write_state_scalar_type)
21974 (write_state_string_type): ditto, renaming to...
21975 (state_writer::write_state_string_type)
21976 (write_state_undefined_type): ditto, renaming to...
21977 (state_writer::write_state_undefined_type)
21978 (write_state_struct_union_type): ditto, renaming to...
21979 (state_writer::write_state_struct_union_type)
21980 (write_state_struct_type): ditto, renaming to...
21981 (state_writer::write_state_struct_type)
21982 (write_state_user_struct_type): ditto, renaming to...
21983 (state_writer::write_state_user_struct_type)
21984 (write_state_lang_struct_type): ditto, renaming to...
21985 (state_writer::write_state_lang_struct_type)
21986 (write_state_param_struct_type): ditto, renaming to...
21987 (state_writer::write_state_param_struct_type)
21988 (write_state_pointer_type): ditto, renaming to...
21989 (state_writer::write_state_pointer_type)
21990 (write_state_array_type): ditto, renaming to...
21991 (state_writer::write_state_array_type)
21992 (write_state_gc_used): ditto, renaming to...
21993 (state_writer::write_state_gc_used)
21994 (write_state_common_type_content): ditto, renaming to...
21995 (state_writer::write_state_common_type_content)
21996 (write_state_type): ditto, renaming to...
21997 (state_writer::write_state_type)
21998 (write_state_pair_list): ditto, renaming to...
21999 (state_writer::write_state_pair_list)
22000 (write_state_pair): ditto, renaming to...
22001 (state_writer::write_state_pair)
22002 (write_state_typedefs): ditto, renaming to...
22003 (state_writer::write_state_typedefs)
22004 (write_state_structures): ditto, renaming to...
22005 (state_writer::write_state_structures)
22006 (write_state_param_structs): ditto, renaming to...
22007 (state_writer::write_state_param_structs)
22008 (write_state_variables): ditto, renaming to...
22009 (state_writer::write_state_variables)
22010 (write_state_srcdir): ditto, renaming to...
22011 (state_writer::write_state_srcdir)
22012 (write_state_files_list): ditto, renaming to...
22013 (state_writer::write_state_files_list)
22014 (write_state_languages): ditto, renaming to...
22015 (state_writer::write_state_languages)
22016 (write_state): create a state_writer instance and use it when
22017 writing out the state file
22019 2013-05-17 Mike Stump <mikestump@comcast.net>
22021 PR rtl-optimization/57304
22022 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
22023 accessing DF_REF_REAL_LOC.
22025 2013-05-17 Jakub Jelinek <jakub@redhat.com>
22027 PR rtl-optimization/57281
22028 PR rtl-optimization/57300
22029 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
22030 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
22031 what the other splitter did if the registers are dead.
22033 2013-05-17 Richard Biener <rguenther@suse.de>
22035 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
22038 2013-05-17 Jakub Jelinek <jakub@redhat.com>
22040 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
22043 2013-05-17 Marek Polacek <polacek@redhat.com>
22045 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
22046 length when doing non-zero store of storing '\0' to '\0'.
22048 2013-05-17 Jakub Jelinek <jakub@redhat.com>
22050 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
22051 vect_external_def oprnd1 with loop_vinfo, try to emit
22052 optional cast, negation and and stmts on the loop preheader
22053 edge instead of into the pattern def seq.
22055 PR tree-optimization/57051
22056 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
22057 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
22059 2013-05-16 Nick Clifton <nickc@redhat.com>
22061 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
22062 (rl78_is_naked_func): New function.
22063 (rl78_expand_prologue): Skip prologue generation for naked functions.
22064 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
22065 * doc/extend.texi (naked): Add RL78 to the list of processors
22066 that supports this attribute.
22068 2013-05-16 Jeff Law <law@redhat.com>
22070 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
22072 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
22074 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
22075 cache parameters using detect_caches_amd also for CYRIX,
22076 NSC and TM2 signatures.
22078 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
22079 Dzianis Kahanovich <mahatma@eu.by>
22083 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
22084 VIA/Centaur processors and determine their cache parameters
22085 using detect_caches_amd.
22087 2013-05-16 Teresa Johnson <tejohnson@google.com>
22089 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
22090 (rtl_verify_edges): New function.
22091 (rtl_verify_bb_insns): Ditto.
22092 (rtl_verify_bb_pointers): Ditto.
22093 (rtl_verify_bb_insn_chain): Ditto.
22094 (rtl_verify_fallthru): Ditto.
22095 (rtl_verify_bb_layout): Ditto.
22096 (rtl_verify_flow_info_1): Outline checks into new functions.
22097 (rtl_verify_flow_info): Ditto.
22099 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
22101 * cfghooks.c (copy_bbs): Add update_dominance argument.
22102 * cfghooks.h (copy_bbs): Update prototype.
22103 * tree-cfg.c (gimple_duplicate_sese_region):
22104 Add update_dominance argument.
22105 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
22106 * tree-ssa-loop-ch.c (copy_loop_headers): Update
22107 gimple_duplicate_sese_region call.
22108 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
22109 Update copy_bbs call.
22110 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
22111 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
22113 2013-05-16 Jakub Jelinek <jakub@redhat.com>
22115 * tree-vectorizer.h (NUM_PATTERNS): Increment.
22116 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
22117 vect_recog_rotate_pattern.
22118 (vect_recog_rotate_pattern): New function.
22120 2013-05-16 Jason Merrill <jason@redhat.com>
22122 * Makefile.in (LLINKER): New variable.
22123 (mostlyclean): Remove link mutex.
22124 * configure.ac: Handle --enable-link-mutex.
22125 * lock-and-run.sh: New script.
22127 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
22130 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
22133 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
22135 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
22137 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
22139 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
22140 * config/arm/arm.c (next_consecutive_mem): New function.
22141 (gen_movmem_ldrd_strd): Likewise.
22142 * config/arm/arm.md (movmemqi): Update condition and code.
22143 (unaligned_loaddi, unaligned_storedi): New patterns.
22145 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
22147 * config.gcc: Obsolete *-*-solaris2.9*.
22148 * doc/install.texi (Specific, *-*-solaris2*): Document it.
22150 2013-05-16 Richard Biener <rguenther@suse.de>
22152 * passes.c (init_optimization_passes): Move pass_parallelize_loops
22153 earlier, after GRAPHITE transforms and IV canonicalization.
22155 2013-05-16 Jakub Jelinek <jakub@redhat.com>
22157 * omp-low.c (extract_omp_for_data): For collapsed loops,
22158 if at least one of the loops is known at compile time to
22159 iterate zero times, set count to 0.
22160 (expand_omp_regimplify_p): New function.
22161 (expand_omp_for_generic): For collapsed loops, if at least
22162 one of the loops isn't known to iterate at least once,
22163 add runtime check with setting count to 0.
22164 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
22165 For unsigned types if it isn't known at compile time that
22166 the loop will iterate at least once, add runtime check to bypass
22167 the whole loop if initial condition isn't true.
22169 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
22171 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
22173 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
22175 PR middle-end/57286
22176 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
22177 transformations to avoid an infinite loop.
22179 2013-05-16 Marek Polacek <polacek@redhat.com>
22181 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
22183 2013-05-15 Leif Ekblad <leif@rdos.net>
22185 * config/i386/i386.c (ix86_decompose_address): Use
22186 DEFAULT_TLS_SEG_REG to access TLS segment register.
22187 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
22188 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
22189 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
22191 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
22194 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
22195 sibling calls to functions that would normally be lazily bound,
22196 unless $gp is call-clobbered.
22198 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
22200 * config/i386/i386.c (ix86_option_override_internal): Update
22201 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
22202 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
22203 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
22204 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
22206 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
22208 2013-05-15 Andreas Schwab <schwab@suse.de>
22210 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
22211 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
22214 2013-05-15 Teresa Johnson <tejohnson@google.com>
22216 * loop-unroll.c (report_unroll_peel): Check decision before
22217 emitting unroll/peel message.
22219 2013-05-15 Teresa Johnson <tejohnson@google.com>
22221 * function.h (has_bb_partition): New rtl_data flag.
22222 (bb_reorder_complete): Ditto.
22223 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
22224 instead of flag_reorder_blocks_and_partition.
22225 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
22226 with some enhancements.
22227 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
22228 * bb-reorder.c (connect_traces): Check for has_bb_partition
22229 instead of flag_reorder_blocks_and_partition.
22230 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
22231 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
22232 verify_hot_cold_block_grouping.
22233 (partition_hot_cold_basic_blocks): Set has_bb_partition.
22235 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
22238 * config/arm/predicates.md (call_insn_operand): New predicate.
22239 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
22240 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
22241 if insn is not a tail call.
22242 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
22244 * config/arm/arm.h (enum reg_class): New caller save register class.
22245 (REG_CLASS_NAMES): Likewise.
22246 (REG_CLASS_CONTENTS): Likewise.
22247 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
22250 2013-05-15 Richard Biener <rguenther@suse.de>
22252 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
22253 of MSG_OPTIMIZED_LOCATIONS.
22254 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
22255 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
22257 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
22258 of MSG_OPTIMIZED_LOCATIONS.
22259 (execute_vect_slp): Likewise.
22260 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
22261 (vect_create_cond_for_alias_checks): Likewise.
22262 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
22263 (vect_recog_widen_mult_pattern): Likewise.
22264 (vect_recog_widen_sum_pattern): Likewise.
22265 (vect_recog_over_widening_pattern): Likewise.
22266 (vect_recog_widen_shift_pattern): Likewise.
22267 (vect_recog_vector_vector_shift_pattern): Likewise.
22268 (vect_recog_divmod_pattern): Likewise.
22269 (vect_recog_mixed_size_cond_pattern): Likewise.
22270 (vect_recog_bool_pattern): Likewise.
22271 (vect_pattern_recog_1): Likewise.
22273 2013-05-15 Martin Jambor <mjambor@suse.cz>
22275 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
22276 non-functions to builtin_unreachable.
22277 * ipa-inline-transform.c (inline_call): Do not assert estimates were
22278 correct when new direct edges were discovered.
22280 2013-05-15 Martin Jambor <mjambor@suse.cz>
22282 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
22283 header, print symbol order instead of node uid, print more information
22284 about indirect edge targets.
22285 (ipa_make_edge_direct_to_target): Print symbol order instead of node
22287 (ipa_make_edge_direct_to_target): Likewise.
22288 (remove_described_reference): Likewise.
22289 (propagate_controlled_uses): Likewise.
22290 (ipa_print_node_params): Also print symbol order.
22291 (ipcp_transform_function): Print symbol order instead of node uids.
22292 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
22293 (cgraph_get_create_real_symbol_node): Likewise.
22294 * ipa-cp.c (print_lattice): Likewise.
22295 (print_all_lattices): Likewise.
22296 (determine_versionability): Likewise.
22297 (initialize_node_lattices): Likewise.
22298 (estimate_local_effects): Likewise.
22299 (update_profiling_info): Likewise.
22300 (create_specialized_node): Likewise.
22301 (perhaps_add_new_callers): Likewise.
22302 (decide_about_value): Likewise.
22303 (decide_whether_version_node): Likewise.
22304 (identify_dead_nodes): Likewise.
22305 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
22306 (dump_inline_summary): Likewise.
22307 (estimate_node_size_and_time): Likewise.
22308 (inline_analyze_function): Likewise.
22309 * ipa-inline.c (report_inline_failed_reason): Likewise.
22310 (want_early_inline_function_p): Likewise.
22311 (edge_badness): Likewise.
22312 (update_edge_key): Likewise.
22313 (inline_small_functions): Likewise. Add dumping of order to two other
22315 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
22316 instead of node uids.
22317 (propagate_pure_const): Likewise.
22318 (propagate_pure_const): Likewise.
22319 * ipa-utils.c (dump_cgraph_node_set): Likewise.
22320 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
22321 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
22323 * tree-pretty-print.c (dump_function_header): Likewise.
22324 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
22325 Print symbol order instead of node uids.
22327 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
22329 * config/s390/s390.c (s390_register_move_cost): Don't impose the
22330 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
22332 2013-05-15 Richard Biener <rguenther@suse.de>
22334 PR tree-optimization/57275
22335 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
22336 return value for fail to do runtime alias checks for gather loads.
22338 2013-05-15 Jan Hubicka <jh@suse.cz>
22342 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
22343 weakrefs are not external.
22344 (lto_symtab_merge_decls): Fix thinko when dealing with
22345 non-lto_symtab decls.
22346 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
22347 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
22348 * varpool.c (dump_varpool_node): Dump more flags.
22350 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
22352 * config/i386/i386.c (processor_alias_table): Add instruction
22353 FSGSBASE for AMD bdver3 architecture.
22355 2013-05-14 Jakub Jelinek <jakub@redhat.com>
22357 * tree.c (warn_deprecated_use): Print file:line using locus color.
22358 * diagnostic.c (diagnostic_report_current_module): Print file:line
22359 and file:line:column using locus color.
22361 2013-05-14 Mike Stump <mikestump@comcast.net>
22363 * gdbinit.in: Add __null.
22365 2013-05-14 Mike Stump <mikestump@comcast.net>
22367 * recog.h: Rename struct recog_data to Recog_data.
22368 * recog.c: Likewise.
22369 * reload.c (can_reload_into): Likewise.
22370 * config/picochip/picochip.c: Likewise.
22372 2013-05-14 Mike Stump <mikestump@comcast.net>
22374 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
22376 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
22378 * resource.h (struct resources): Remove unch_memory member.
22379 (CLEAR_RESOURCE): Don't clear unch_memory.
22380 * resource.c (mark_referenced_resources): Don't set it.
22381 (mark_set_resources): Likewise.
22382 (mark_target_live_regs): Don't clear it.
22383 (init_resource_info): Likewise.
22384 * reorg.c (resource_conflicts_p): Don't compare it.
22385 (redundant_insn): Don't set it.
22387 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
22389 * emit-rtl.c (next_label): Remove unused function.
22390 (skip_consecutive_labels, link_cc0_insns): Move to ...
22391 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
22392 only place where these functions are used, and make them static.
22394 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
22396 * fold-const.c (fold_negate_expr): Handle vectors.
22397 (fold_truth_not_expr): Make it static.
22398 (fold_invert_truthvalue): New static function.
22399 (invert_truthvalue_loc): Handle vectors. Do not call
22400 fold_truth_not_expr directly.
22401 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
22402 <TRUTH_NOT_EXPR>: Do not cast to boolean.
22403 (fold_comparison): Handle vector constants.
22404 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
22405 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
22406 * tree.h (fold_truth_not_expr): Remove declaration.
22408 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
22410 * config/aarch64/aarch64-simd.md
22411 (aarch64_vcond_internal<mode>): Rename to...
22412 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
22413 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
22414 float modes. Clarify all iterator modes.
22415 (vcond<mode><mode>): Use new name for vcond expanders.
22416 (vcond<v_cmp_result><mode>): Likewise.
22417 (vcondu<mode><mode>: Likewise.
22418 * config/aarch64/iterators.md (VDQF_COND): New.
22420 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
22423 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
22424 variable for the shift amount. Check that we shift by non-negative
22427 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
22430 * config/arm/arm.h (EPILOGUE_USES): Only return true
22431 for LR_REGNUM after epilogue_completed.
22433 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
22435 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
22436 is error_mark_node.
22438 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
22441 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
22442 and Solaris 11+/x86 with gld.
22443 * configure: Regenerate.
22445 2013-05-14 Jakub Jelinek <jakub@redhat.com>
22447 * expmed.c (expand_shift_1): Canonicalize rotates by
22448 constant bitsize / 2 to bitsize - 1.
22449 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
22450 case ROTATERT>: Likewise.
22453 2013-05-10 Jakub Jelinek <jakub@redhat.com>
22455 * config/i386/i386.md (rotateinv): New code attr.
22456 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
22457 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
22458 roll $31, %eax, etc.
22460 2013-05-14 Richard Biener <rguenther@suse.de>
22462 PR middle-end/57235
22463 * tree-eh.c (sink_clobbers): Give up for successors with
22464 multiple predecessors and no virtual uses.
22466 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
22468 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
22469 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
22471 2013-05-14 Jakub Jelinek <jakub@redhat.com>
22473 PR middle-end/57251
22474 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
22475 the case when both op0 and op1 have VOIDmode.
22477 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
22479 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
22480 in multiply-accumulate mode.
22482 2013-05-13 Guozhi Wei <carrot@google.com>
22484 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
22486 2013-05-13 Kai Tietz <ktietz@redhat.com>
22489 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
22490 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
22491 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
22492 * config/i386/i386.c (ix86_option_override_internal): Likewise.
22493 (ix86_expand_prologue): Likewise.
22494 (ix86_expand_split_stack_prologue): Likewise.
22495 (legitimate_pic_address_disp_p): Likewise.
22496 (legitimize_pic_address): Likewise.
22497 (legitimize_tls_address): Likewise.
22498 (legitimize_pe_coff_symbol): Likewise.
22499 (output_pic_addr_const): Likewise.
22500 (construct_plt_address): Likewise.
22501 (ix86_expand_call): Likewise.
22502 (x86_output_mi_thunk): Likewise.
22503 (x86_function_profiler): Likewise.
22505 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
22507 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
22508 similar switch cases.
22509 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
22510 (aarch64_simd_mov_to_<mode>low): Delete.
22511 (aarch64_simd_mov_to_<mode>high): Delete.
22512 (move_lo_quad_<mode>): Add w<-r alternative.
22513 (aarch64_simd_move_hi_quad_<mode>): Likewise.
22514 (aarch64_simd_mov_from_*): Update type attribute.
22515 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
22518 2013-05-13 Jan Hubicka <jh@suse.cz>
22520 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
22521 * config/i386/i386.c (ix86_compute_frame_layout,
22522 ix86_expand_epilogue, emit_i387_cw_initialization,
22523 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
22524 ix86_local_alignment): Fix use of size/speed predicates.
22526 2013-05-13 Jakub Jelinek <jakub@redhat.com>
22528 PR tree-optimization/45216
22529 PR tree-optimization/57157
22530 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
22531 the (-Y) & (B - 1) variant if OP is |.
22532 * expmed.c (expand_shift_1): For rotations by const0_rtx just
22533 return shifted. Use (-op1) & (prec - 1) as other_amount
22534 instead of prec - op1.
22536 2013-05-13 Martin Jambor <mjambor@suse.cz>
22538 PR middle-end/42371
22539 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
22540 (ipa_constant_data): New type.
22541 (ipa_jump_func): Use ipa_constant_data to hold information about
22542 constant jump functions.
22543 (ipa_get_jf_constant): Adjust to jump function type changes.
22544 (ipa_get_jf_constant_rdesc): New function.
22545 (ipa_param_descriptor): New field controlled_uses.
22546 (ipa_get_controlled_uses): New function.
22547 (ipa_set_controlled_uses): Likewise.
22548 * ipa-ref.h (ipa_find_reference): Declare.
22549 * ipa-prop.c (ipa_cst_ref_desc): New type.
22550 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
22552 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
22553 New parameter cs. Adjust all callers.
22554 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
22555 (remove_described_reference): New function.
22556 (jfunc_rdesc_usable): Likewise.
22557 (try_make_edge_direct_simple_call): Decrement controlled use count,
22558 attempt to remove reference if it hits zero.
22559 (combine_controlled_uses_counters): New function.
22560 (propagate_controlled_uses): Likewise.
22561 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
22562 (ipa_edge_duplication_hook): Duplicate reference descriptions.
22563 (ipa_print_node_params): Print described use counter.
22564 (ipa_write_jump_function): Adjust to jump function type changes.
22565 (ipa_read_jump_function): New parameter CS, pass it to
22566 ipa_set_jf_constant. Adjust caller.
22567 (ipa_write_node_info): Stream controlled use count
22568 (ipa_read_node_info): Likewise.
22569 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
22571 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
22572 count. Remove cloning-added reference if it reaches zero.
22573 * ipa-ref.c (ipa_find_reference): New function.
22575 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
22577 * config/i386/i386.c (processor_target_table): Modified default
22578 alignment values for AMD BD and BT architectures.
22580 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
22582 * tree-vect-generic.c (uniform_vector_p): Move ...
22583 * tree.c (uniform_vector_p): ... here.
22584 * tree.h (uniform_vector_p): Declare it.
22585 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
22588 2013-05-13 Jakub Jelinek <jakub@redhat.com>
22590 PR tree-optimization/57230
22591 * tree-ssa-strlen.c (handle_char_store): Record length for
22592 array store from STRING_CST.
22594 PR tree-optimization/57230
22595 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
22598 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
22600 * config/epiphany/epiphany.c (epiphany_init): Check size of
22601 NUM_MODES_FOR_MODE_SWITCHING.
22602 (epiphany_expand_prologue):
22603 Remove CONFIG_REGNUM initial value handling code.
22604 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
22605 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
22606 (emit_set_fp_mode, epiphany_mode_after): Likewise.
22607 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
22608 Don't return 1 for FP_MODE_NONE.
22609 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
22610 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
22611 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
22612 * config/epiphany/epiphany.md (save_config): New pattern.
22614 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
22616 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
22618 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
22620 * config/i386/i386.md (memory): Handle sseishft1.
22621 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
22622 (*vec_extractv2di_1): Ditto.
22624 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
22626 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
22629 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
22631 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
22632 Add mthumb/march=armv7-a multilib.
22633 Add mthumb/march=armv7-r multilib.
22634 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
22636 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
22638 * config/v850/t-rtems: Add more multilibs.
22640 2013-05-10 Richard Biener <rguenther@suse.de>
22642 PR tree-optimization/57214
22643 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
22644 not propagate from SSA names that occur in abnormal PHI nodes.
22646 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
22648 * stor-layout.c (element_precision): New function.
22649 * machmode.h (element_precision): Declare it.
22650 * tree.c (build_minus_one_cst): New function.
22651 (element_precision): Likewise.
22652 * tree.h (build_minus_one_cst): Declare new function.
22653 (element_precision): Likewise.
22654 * fold-const.c (operand_equal_p): Use element_precision.
22655 (fold_binary_loc): Handle vector types.
22656 * convert.c (convert_to_integer): Use element_precision.
22657 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
22660 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
22662 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
22663 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
22664 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
22665 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
22666 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
22667 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
22668 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
22669 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
22670 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
22671 (Uuw8): New constraints.
22672 (Usb4): Move into alphabetical order.
22673 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
22674 (sd8_operand, ub8_operand, uw8_operand): New predicates.
22675 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
22676 previously unnamed patterns.
22677 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
22678 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
22679 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
22680 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
22681 of set_attr_alternative/if_then_else. Use extended_mips16 instead
22682 of specific lengths.
22684 2013-05-10 Jakub Jelinek <jakub@redhat.com>
22686 * config/i386/i386.md (rotateinv): New code attr.
22687 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
22688 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
22689 roll $31, %eax, etc.
22691 PR tree-optimization/45216
22692 PR tree-optimization/57157
22693 * tree-ssa-forwprop.c (simplify_rotate): New function.
22694 (ssa_forward_propagate_and_combine): Call it.
22696 2013-05-10 Richard Biener <rguenther@suse.de>
22698 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
22699 disable peeling when we version for aliasing.
22700 (vector_alignment_reachable_p): Honor explicit user alignment.
22701 (vect_supportable_dr_alignment): Likewise.
22702 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
22703 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
22704 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
22705 then peeling to arrange for the cost-model check to come first.
22707 2013-05-10 Alan Modra <amodra@gmail.com>
22709 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
22710 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
22711 * configure: Regenerate.
22713 2013-05-10 Alan Modra <amodra@gmail.com>
22716 * varasm.c (default_elf_select_section): Move !DECL_P check..
22717 (get_named_section): ..to here before calling get_section_name.
22719 (default_section_type_flags): Add DECL_P check.
22720 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
22721 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
22723 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
22725 * config/epiphany/epiphany.c (epiphany_expand_prologue):
22726 When using gen_stack_adjust_str with a register offset, add a
22727 REG_FRAME_RELATED_EXPR note.
22729 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
22731 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
22732 (*vec_extractv4si_zext_mem): Ditto.
22733 (*vec_extractv2di): Add 0->x and x->x alternatives.
22734 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
22735 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
22737 2013-05-09 Jason Merrill <jason@redhat.com>
22739 N3639 C++1y VLA support
22740 * gimplify.c (gimplify_vla_decl): Don't touch an existing
22743 * tree.c (build_constructor_va): New.
22744 * tree.h: Declare it.
22746 2013-05-09 Martin Jambor <mjambor@suse.cz>
22749 * gimple-fold.c (canonicalize_constructor_val): Call
22750 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
22752 2013-05-09 Jan Hubicka <jh@suse.cz>
22753 Richard Biener <rguenther@suse.de>
22756 * symtab.c (symtab_make_decl_local): Do not add private names.
22758 2013-05-09 Jan Hubicka <jh@suse.cz>
22761 * symtab.c (insert_to_assembler_name_hash): Handle clones.
22762 (unlink_from_assembler_name_hash): Likewise.
22763 (symtab_prevail_in_asm_name_hash, symtab_register_node,
22764 symtab_unregister_node, symtab_initialize_asm_name_hash,
22765 change_decl_assembler_name): Update.
22767 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
22769 * config/aarch64/aarch64.md: New movtf split.
22770 (*movtf_aarch64): Update.
22771 (aarch64_movdi_tilow): Handle TF modes and rename to
22772 aarch64_movdi_<mode>low.
22773 (aarch64_movdi_tihigh): Handle TF modes and rename to
22774 aarch64_movdi_<mode>high
22775 (aarch64_movtihigh_di): Handle TF modes and rename to
22776 aarch64_mov<mode>high_di
22777 (aarch64_movtilow_di): Handle TF modes and rename to
22778 aarch64_mov<mode>low_di
22779 (aarch64_movtilow_tilow): Remove spurious whitespace.
22780 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
22782 (aarch64_print_operand): Update.
22784 2013-05-09 Alan Modra <amodra@gmail.com>
22786 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
22788 * configure: Regenerate.
22790 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
22792 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
22793 splitter preparation statements.
22794 * config/i386/sse.md (*vec_extract* splitters): Ditto.
22795 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
22798 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
22800 * gimple-ssa-strength-reduction.c (count_candidates): Change
22801 return value to int.
22802 (analyze_candidates_and_replace): Change type of length to int.
22804 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
22806 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
22807 (*vec_extract<mode>): Use VI12_128 mode iterator.
22808 (*vec_extract<mode>_mem): Ditto.
22809 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
22812 2013-05-08 Diego Novillo <dnovillo@google.com>
22817 2012-08-17 Diego Novillo <dnovillo@google.com>
22820 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
22821 * config.in: Regenerate.
22822 * configure: Regenerate.
22823 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
22825 2013-05-08 Jan Hubicka <jh@suse.cz>
22828 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
22829 * cgraph.h (symtab_node_base): Add unique_name.
22830 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
22831 input_overwrite_node, input_varpool_node): Stream unique_name.
22832 * cgraphclones.c (cgraph_create_virtual_clone,
22833 cgraph_function_versioning): Set unique_name.
22834 * ipa.c (function_and_variable_visibility): Set unique_name.
22836 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
22838 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
22839 (alloc_cand_and_find_basis): Restrict conditional candidate
22840 processing to CAND_MULTs.
22842 2013-05-08 Jan Hubicka <jh@suse.cz>
22845 lto-symtab.c (lto_symtab_symbol_p): New function.
22846 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
22847 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
22848 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
22849 Skip static symbols.
22851 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
22853 PR tree-optimization/57200
22854 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
22855 Only call inform if the preceding warning_at returns true.
22857 2013-05-07 Han Shen <shenhan@google.com>
22859 * cfgexpand.c (record_or_union_type_has_array_p): New function.
22860 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
22861 * common.opt (fstack-protector-strong): New option.
22862 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
22863 * doc/invoke.texi (Optimization Options): Document
22864 "-fstack-protector-strong".
22865 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
22867 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
22869 * config/mips/mips.c (mips_machine_reorg2): Return 0.
22871 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
22873 * ira.c (update_equiv_regs): Add insn having equiv memory even if
22874 it is not lhs of the insn.
22875 (setup_reg_equiv): Remove insn having equiv memory which it is not
22877 * lra-constraints.c (process_address): Try to improve generation
22878 code for address base + disp.
22879 (lra_constraints): Make correct the code for checking insn setting
22880 up backward equivalence. Remove insn only if it is in the init
22882 * lra-eliminations.c (update_reg_eliminate): Change return value.
22883 (lra_eliminate): Use the result.
22885 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
22887 * config/i386/sse.md (ssescalarnummask): New mode attribute.
22888 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
22889 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
22890 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
22891 register target operands.
22892 (*vec_extractv8hi_sse2): New pattern.
22893 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
22894 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
22895 (*vec_extract<mode>_mem): New insn and split pattern.
22897 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
22899 * config/arm/arm.c (arm_asan_shadow_offset): New function.
22900 (TARGET_ASAN_SHADOW_OFFSET): Define.
22901 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
22902 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
22904 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
22906 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
22907 (incr_vec_index): Return -1 if increment not found.
22908 (create_add_on_incoming_edge): Assert if increment not found.
22909 (record_increment): Limit number of increments recorded.
22910 (all_phi_incrs_profitable): Return false if an increment not found.
22911 (replace_profitable_candidates): Don't process increments that were
22913 (analyze_candidates_and_replace): Limit size of incr_vec.
22915 2013-05-07 Richard Biener <rguenther@suse.de>
22917 * calls.c (special_function_p): setjmp-like functions are leaf.
22918 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
22919 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
22921 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
22923 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
22924 (aarch64_simd_mov<mode>): New expander.
22925 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
22926 (aarch64_simd_mov_to_<mode>high): Likewise.
22927 (aarch64_simd_mov_from_<mode>low): Likewise.
22928 (aarch64_simd_mov_from_<mode>high): Likewise.
22929 (aarch64_dup_lane<mode>): Update.
22930 (aarch64_dup_lanedi): New instruction pattern.
22931 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
22932 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
22934 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
22936 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
22937 (replace_mult_candidate): Remove unnecessary argument; remove
22938 unnecessary parameter from call to introduce_cast_before_cand.
22939 (replace_unconditional_candidate): Remove unnecessary parameter
22940 from call to replace_mult_candidate.
22941 (replace_conditional_candidate): Likewise.
22942 (insert_initializers): Use make_temp_ssa_name.
22943 (introduce_cast_before_cand): Remove unnecessary argument; use
22944 make_temp_ssa_name.
22945 (replace_one_candidate): Remove unnecessary argument; remove
22946 unnecessary parameter from calls to introduce_cast_before_cand.
22947 (replace_profitable_candidates): Remove unnecessary parameters
22948 from calls to replace_one_candidate.
22950 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
22952 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
22953 phi def as possibly hiding a basis for a CAND_ADD whose operands
22954 have been commuted in the analysis.
22955 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
22957 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22959 * config/aarch64/aarch64.md
22960 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
22961 shift value between 0-4.
22963 2013-05-07 Richard Biener <rguenther@suse.de>
22965 * double-int.h (rshift): New overload.
22966 * double-int.c (rshift): New function.
22967 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
22968 (create_reference_ops_from_ref): Remove.
22969 (vn_reference_insert): Use shared ops for constructing the
22970 reference and copy it.
22972 2013-05-07 Richard Biener <rguenther@suse.de>
22974 PR middle-end/57190
22975 * tree-eh.c (sink_clobbers): Properly propagate
22976 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
22978 2013-05-07 Jakub Jelinek <jakub@redhat.com>
22980 PR tree-optimization/57149
22981 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
22982 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
22983 collect_phi_def_edges, execute_late_warn_uninitialized): Use
22984 uninit_undefined_value_p instead of ssa_undefined_value_p.
22987 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
22988 for modifier == EXPAND_INITIALIZER.
22990 2013-05-07 Anton Blanchard <anton@samba.org>
22992 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
22993 for powerpc64 little endian.
22994 * configure: Regenerate.
22996 2013-05-06 Graham Stott <grahams@btinternet.com>
22998 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
22999 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
23000 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
23001 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
23003 2013-05-06 Graham Stott <grahams@btinternet.com>
23005 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
23006 codes which allow non-lvalues.
23008 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
23010 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
23011 components are all 1s.
23012 (integer_minus_onep): New function.
23013 * tree.h (integer_minus_onep): Declare it.
23014 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
23015 integer_minus_onep instead of integer_all_onesp.
23017 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
23020 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
23021 variations of these patterns.
23023 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
23025 * config/i386/i386.md (isa): Add x64_sse4 member.
23026 (enabled): Handle x64_sse4.
23027 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
23028 instruction for 64bit SSE4_1 targets. Update insn attributes.
23029 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
23030 instruction for SSE4_1 targets. Update insn attributes.
23031 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
23032 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
23033 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
23035 (*vec_extractv4si): Rename from *sse4_1_pextrd.
23036 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
23037 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
23039 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
23042 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
23044 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
23046 * final.c (do_assembler_dialects): Don't handle curly braces and
23047 vertical bar escaped by % as dialect delimiters.
23048 (output_asm_insn): Print curly braces and vertical bar if escaped
23049 by % and ASSEMBLER_DIALECT defined.
23050 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
23051 * doc/tm.texi: Regenerated.
23053 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
23055 * config/mips/mips.c: Include tree-pass.h.
23056 (mips_reorg): Split in pre- and post-dbr_schedule parts.
23057 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
23058 (pass_mips_machine_reorg2): New machine specific pass.
23059 (insert_pass_mips_machine_reorg2): New pass plugin definition.
23060 (mips_option_override): Register the new pass.
23061 * rtl.h (cleanup_barriers): Remove prototype.
23062 (dbr_schedule): Likewise.
23063 * jump.c (cleanup_barriers): Make static.
23064 * reorg.c (dbr_schedule): Likewise.
23066 2013-05-06 Richard Biener <rguenther@suse.de>
23068 PR tree-optimization/57185
23069 * tree-parloops.c (add_field_for_reduction): Handle anonymous
23070 SSA names properly.
23072 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
23075 * config/i386/i386.c (add_parameter_dependencies): Add dependence
23076 between "first_arg" and "insn", not "last" and "insn".
23078 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
23080 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
23081 (find_candidates_in_block): Re-enable slsr_process_phi.
23082 (create_phi_basis): Fix double counting of candidate adjustment.
23084 2013-05-06 Richard Biener <rguenther@suse.de>
23086 PR middle-end/57147
23087 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
23088 the edge is also fallthru, preserve it and just clear the
23090 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
23091 also complex, preserve that and just clear the fallthru flag.
23092 * tree-inline.c (update_ssa_across_abnormal_edges): Also
23093 update virtual operands.
23095 2013-05-06 Alan Modra <amodra@gmail.com>
23097 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
23098 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
23099 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
23100 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
23101 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
23102 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
23104 2013-05-06 Alan Modra <amodra@gmail.com>
23106 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
23107 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
23108 (DEFAULT_ASM_ENDIAN): Define.
23109 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
23110 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
23111 Update -K PIC clause from sysv4.h.
23112 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
23113 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
23115 2013-05-06 Alan Modra <amodra@gmail.com>
23117 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
23118 twice for little-endian.
23119 (ashrdi3_no_power, ashrdi3): Support little-endian.
23121 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
23124 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
23125 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
23127 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
23128 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
23129 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
23132 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
23133 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23135 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
23136 * config/pa/pa.opt: Make mbig-switch a no-op.
23137 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
23138 (CASE_VECTOR_MODE): Always return SImode.
23139 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
23140 for the !TARGET_BIG_SWITCH case.
23141 * config/pa/pa-linux.h: Likewise.
23142 * config/pa/pa-openbsd.h: Likewise.
23143 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
23144 * config/pa/pa.md (short_jump): Remove define_insn.
23145 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
23146 (casesi0): Remove define_insn.
23147 (type): Remove btable_branch.
23148 (pa_combine_type): Likewise.
23149 (in_nullified_branch_delay): Likewise.
23150 (in_call_delay): Likewise.
23151 (define_delay): Likewise.
23152 (define_insn_reservation "Z3"): Likewise.
23153 (define_insn_reservation "Z4"): Likewise.
23154 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
23155 (pa_adjust_insn_length): Remove adjustment for btable branches.
23156 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
23159 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
23161 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
23162 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
23163 Add m->r,x alternatives.
23164 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
23165 splitters using SWI48x mode iterator.
23166 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
23167 TARGET_64BIT. Add m->x alternative.
23168 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
23169 Add o->x alternative. Enable for TARGET_SSE.
23170 (sse_storeq): Remove expander.
23171 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
23172 with memory input operand.
23173 (*vec_extractv2di_1 splitter): New.
23174 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
23175 * config/i386/i386.md (ssevecmodelower): New mode attribute.
23177 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
23179 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
23180 (INT_LOWPART): Delete.
23181 (extract_MB): Adjust.
23182 (extract_ME): Adjust.
23183 (print_operand): Adjust.
23185 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
23187 * config/rs6000/predicates.md (reg_or_add_cint_operand,
23188 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
23189 (reg_or_logical_cint_operand, easy_fp_constant,
23190 logical_const_operand): Delete "CONST_DOUBLE" case.
23191 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
23192 "HOST_BITS_PER_WIDE_INT == 64" test.
23193 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
23194 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
23195 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
23196 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
23198 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
23199 CONST_DOUBLE DImode/VOIDmode case.
23200 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
23201 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
23202 CONST_DOUBLE VOIDmode case.
23203 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
23204 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
23205 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
23206 Delete CONST_DOUBLE case.
23207 (splitters for mov FMOVE64 const_double): Delete
23208 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
23209 "HOST_BITS_PER_WIDE_INT >= 64" test.
23210 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
23212 (mov DI const_double): Delete.
23214 2013-05-04 Jakub Jelinek <jakub@redhat.com>
23216 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
23217 on op shows all bits zero in mode of a lowpart subreg, return zero.
23219 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
23222 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
23223 to save TFmode registers and DImode to save TImode registers for
23224 caller save operations.
23225 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
23226 mark being partially clobbered since they only use the first
23229 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
23230 and TDmode only use the upper 64-bits of each VSX register.
23232 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
23234 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
23235 (find_candidates_in_block): Disable slsr_process_phi.
23237 2013-05-03 Guozhi Wei <carrot@google.com>
23239 * coverage.c (coverage_obj_init): Move the construction of gcov
23241 (build_init_ctor): ... here.
23243 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
23245 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
23246 (slsr_cand_d): Redefine def_phi.
23247 (stride_status, phi_adjust_status, count_phis_status): New enums.
23248 (find_phi_def): New.
23249 (find_basis_for_base_expr): New.
23250 (find_basis_for_candidate): Handle hidden bases.
23251 (alloc_cand_and_find_basis): Handle phi candidates.
23252 (slsr_process_phi): New.
23253 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
23254 (create_mul_imm_cand): Likewise.
23255 (create_add_ssa_cand): Exclude phi base candidates.
23256 (create_add_imm_cand): Likewise.
23257 (slsr_process_cast): Likewise.
23258 (slsr_process_copy): Likewise.
23259 (find_candidates_in_block): Handle phi candidates.
23260 (dump_candidate): Likewise.
23261 (unconditional_cands): Delete.
23262 (unconditional_cands_with_known_stride_p): Delete.
23263 (phi_dependent_cand_p): New.
23264 (cand_increment): Handle phi-dependent candidates.
23265 (replace_dependent): Delete.
23266 (replace_mult_candidate): New.
23267 (replace_unconditional_candidate): New.
23268 (incr_vec_index): Move to avoid forward reference.
23269 (create_add_on_incoming_edge): New.
23270 (create_phi_basis): New.
23271 (replace_dependents): Delete.
23272 (replace_conditional_candidate): New.
23273 (phi_add_costs): New.
23274 (replace_uncond_cands_and_profitable_phis): New.
23275 (record_increment): Handle phi adjustments.
23276 (record_phi_increments): New.
23277 (record_increments): Handle phi adjustments.
23278 (phi_incr_cost): New.
23279 (lowest_cost_path): Handle phis.
23280 (total_savings): Likewise.
23281 (analyze_increments): Likewise.
23282 (ncd_with_phi): New.
23283 (ncd_of_cand_and_phis): New.
23284 (nearest_common_dominator_for_cands): Handle phi increments.
23285 (all_phi_incrs_profitable): New.
23286 (replace_profitable_candidates): Handle phi-dependent candidates.
23287 (analyze_candidates_and_replace): Likewise.
23289 2013-05-03 Teresa Johnson <tejohnson@google.com>
23292 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
23293 do not exceed REG_BR_PROB_BASE.
23295 2013-05-03 Jeff Law <law@redhat.com>
23297 PR tree-optimization/57144
23298 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
23299 operand of the condition will bit into the new type when eliminating
23300 a cast feeding a condition.
23302 2013-05-03 Jakub Jelinek <jakub@redhat.com>
23304 PR rtl-optimization/57130
23305 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
23306 of COMPARE as in_code to the recursive call if needed.
23308 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
23310 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
23311 (enabled): Handle new members.
23312 * config/i386/sse.md (*vec_concatv2si): Merge from
23313 *vec_concatv2si_sse2 and vec_concatv2si_sse.
23314 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
23316 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
23318 PR tree-optimization/57027
23319 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
23320 for fnms opportunity, check we got the prerequisite kind
23321 of tree / gimple before using accessor functions.
23323 2013-05-03 Richard Biener <rguenther@suse.de>
23325 * double-int.h (lshift): New overload without precision
23326 and arith argument.
23327 (operator *=, operator +=, operator -=): Move ...
23328 * double-int.c (operator *=, operator +=, operator -=): ... here
23329 and implement more efficiently.
23330 (mul_double_with_sign): Remove.
23331 (lshift_double): Adjust to take unsinged shift argument, push
23332 dispatching code to callers.
23333 (mul_double_wide_with_sign): Add early out for callers that
23334 are not interested in high parts or overflow.
23335 (lshift): New function.
23336 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
23337 dispatch code here.
23338 (lrotate, rrotate): Use logical shifts.
23339 * expr.c (get_inner_reference): Use lshift.
23340 * fixed-value.c (do_fixed_divide): Likewise.
23341 * tree-dfa.c (get_ref_base_and_extent): Likewise.
23342 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
23343 (indirect_refs_may_alias_p): Likewise.
23344 (stmt_kills_ref_p_1): Likewise.
23346 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
23348 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
23350 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
23352 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
23353 scalar form of FABD instruction.
23355 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
23357 * lra-constraints.c (process_alt_operands): Add checking alt
23358 number to choose the best alternative.
23360 2013-05-02 Richard Biener <rguenther@suse.de>
23362 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
23363 bitmap and its handling.
23364 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
23366 2013-05-02 Richard Biener <rguenther@suse.de>
23368 PR middle-end/57140
23369 * tree-inline.c (copy_loops): Properly handle removed loops.
23370 (copy_cfg_body): Mark destination loops for fixup if source
23371 loops needed fixup.
23373 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
23376 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
23377 generating simple_return for naked functions.
23379 2013-05-02 Martin Jambor <mjambor@suse.cz>
23381 PR middle-end/56988
23382 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
23383 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
23385 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
23386 ipa_agg_replacement_value structures.
23387 (known_aggs_to_agg_replacement_list): Likewise.
23388 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
23389 (read_agg_replacement_chain): Likewise.
23390 (ipcp_transform_function): Also check that by_ref flags match.
23392 2013-05-02 Richard Biener <rguenther@suse.de>
23394 * graphds.h (struct graph): Add obstack member.
23395 * graphds.c (new_graph): Initialize obstack and allocate
23397 (add_edge): Allocate edge from the obstack.
23398 (free_graph): Free the obstack instead of all edges and vertices.
23400 2013-05-02 Teresa Johnson <tejohnson@google.com>
23402 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
23404 * cfg.c (update_bb_profile_for_threading): Ditto.
23405 * tree-inline.c (copy_bb): Ditto.
23406 (copy_edges_for_bb): Ditto.
23407 (initialize_cfun): Ditto.
23408 (copy_cfg_body): Ditto.
23409 (expand_call_inline): Ditto.
23410 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
23411 (estimate_node_size_and_time): Ditto.
23412 (inline_merge_summary): Ditto.
23413 * cgraphclones.c (cgraph_clone_edge): Ditto.
23414 (cgraph_clone_node): Ditto.
23415 * sched-rgn.c (compute_dom_prob_ps): Ditto.
23416 (compute_trg_info): Ditto.
23418 2013-05-02 Ian Bolton <ian.bolton@arm.com>
23420 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
23421 S reg when fp attribute set.
23422 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
23424 2013-05-02 Ian Bolton <ian.bolton@arm.com>
23426 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
23428 (*and_one_cmplsi3_compare0_uxtw): Likewise.
23429 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
23430 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
23432 2013-05-02 Richard Biener <rguenther@suse.de>
23434 * tree-scalar-evolution.c (scev_info_hasher): Remove.
23435 (struct instantiate_cache_entry): New type.
23436 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
23437 (struct instantiate_cache_type): New type.
23438 (set_instantiated_value, get_instantiated_value): Remove.
23439 (get_instantiated_value_entry): New function.
23440 (instantiate_scev_name): Use the new cache and adjust.
23441 (instantiate_scev_poly): Adjust.
23442 (instantiate_scev_binary): Likewise.
23443 (instantiate_array_ref): Likewise.
23444 (instantiate_scev_convert): Likewise.
23445 (instantiate_scev_not): Likewise.
23446 (instantiate_scev_3): Likewise.
23447 (instantiate_scev_2): Likewise.
23448 (instantiate_scev_r): Likewise.
23449 (instantiate_scev): Likewise.
23450 (resolve_mixers): Likewise.
23452 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
23455 * lra-constraints.c (best_small_class_operands_num): Remove.
23456 (process_alt_operands): Remove small_class_operands_num. Take
23457 small classes operands into losers and only if the operand is not
23458 matched. Modify debugging output.
23459 (curr_insn_transform): Remove best_small_class_operands_num.
23462 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
23464 * config/aarch64/aarch64-builtins.c
23465 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
23466 * config/aarch64/aarch64-simd-builtins.def
23467 (reduc_splus_): Add new modes.
23468 (reduc_uplus_): New.
23469 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
23470 (reduc_uplus_v4sf): Likewise.
23471 (reduc_splus_v4sf): Likewise.
23472 (aarch64_addv<mode>): Likewise.
23473 (reduc_uplus_<mode>): Likewise.
23474 (reduc_splus_<mode>): Likewise.
23475 (aarch64_addvv2di): Likewise.
23476 (reduc_uplus_v2di): Likewise.
23477 (reduc_splus_v2di): Likewise.
23478 (aarch64_addvv2si): Likewise.
23479 (reduc_uplus_v2si): Likewise.
23480 (reduc_splus_v2si): Likewise.
23481 (reduc_<sur>plus_<mode>): New.
23482 (reduc_<sur>plus_v2di): Likewise.
23483 (reduc_<sur>plus_v2si): Likewise.
23484 (reduc_<sur>plus_v4sf): Likewise.
23485 (aarch64_addpv4sf): Likewise.
23486 * config/aarch64/arm_neon.h
23487 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
23488 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
23489 add UNSPEC_SADDV, UNSPEC_UADDV.
23491 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
23493 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
23495 * config/aarch64/arm_neon.h
23496 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
23498 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
23500 * config/aarch64/aarch64-builtins
23501 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
23503 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
23505 * config/aarch64/aarch64-simd-builtins.def
23506 (reduc_smax_): New.
23507 (reduc_smin_): Likewise.
23508 (reduc_umax_): Likewise.
23509 (reduc_umin_): Likewise.
23510 (reduc_smax_nan_): Likewise.
23511 (reduc_smin_nan_): Likewise.
23514 (smax): Update for V2SF, V4SF and V2DF modes.
23517 (smin_nan): Likewise.
23518 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
23519 (<su><maxmin><mode>3): ...This, refactor.
23520 (s<maxmin><mode>3): New.
23521 (<maxmin_uns><mode>3): Likewise.
23522 (reduc_<maxmin_uns>_<mode>): Refactor.
23523 (reduc_<maxmin_uns>_v4sf): Likewise.
23524 (reduc_<maxmin_uns>_v2si): Likewise.
23525 (aarch64_<fmaxmin><mode>: Remove.
23526 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
23528 (vmin<q>_f<32,64>): Likewise.
23529 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
23531 (su): Add mappings for smax, smin, umax, umin.
23533 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
23534 (FMAXMIN): Rename as...
23535 (FMAXMIN_UNS): ...This.
23537 (fmaxminv): Likewise.
23538 (fmaxmin): Likewise.
23540 (maxmin_uns_op): Likewise.
23542 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
23544 * config/aarch64/arm_neon.h
23545 (vac<ge, gt><sd>_f<32, 64>): Rename to...
23546 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
23547 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
23549 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
23551 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
23552 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
23554 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
23556 * config/aarch64/aarch64-simd.md
23557 (vcond<mode>_internal): Handle special cases for constant masks.
23558 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
23559 (vcondu<mode><mode>): Likewise.
23560 (vcond<v_cmp_result><mode>): New.
23562 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
23564 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
23565 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
23566 * config/aarch64/aarch64-simd-builtins.def
23567 (cmeq): Update to BUILTIN_VALLDI.
23572 * config/aarch64/arm_neon.h
23573 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
23574 to builtins or C as appropriate.
23576 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
23578 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
23580 (cmhi): Rename to...
23582 * config/aarch64/aarch64-simd.md
23583 (simd_mode): Add SF.
23584 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
23585 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
23586 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
23587 (cstore<mode>_neg): ...This.
23588 * config/aarch64/iterators.md
23590 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
23591 (COMPARISONS): New.
23592 (UCOMPARISONS): Likewise.
23593 (optab): Add missing comparisons.
23599 (VCMP_S): Likewise.
23600 (VCMP_U): Likewise.
23601 (V_cmp_result): Add DF, SF modes.
23602 (v_cmp_result): Likewise.
23604 (vmtype): Likewise.
23605 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
23607 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
23609 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
23610 define_insn to define_insn_and_split.
23611 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
23612 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
23613 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
23614 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
23615 (thumb2_negscc): Likewise.
23617 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
23619 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
23621 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
23623 * config/arm/thumb2.md: Remove trailing whitespaces.
23625 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
23627 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
23628 Use gen_int_mode rather than GEN_INT.
23630 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
23632 * value-prof.c (stream_in_histogram_value): Remove the strayed
23635 2013-04-30 Richard Biener <rguenther@suse.de>
23637 PR middle-end/57122
23638 * cfghooks.c (split_edge): Properly check for the loop latch edge.
23640 2013-04-30 Richard Biener <rguenther@suse.de>
23642 PR middle-end/57107
23643 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
23645 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
23647 PR rtl-optimization/56957
23648 PR rtl-optimization/57105
23649 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
23650 variable. Use just INSN_UID for determining whether an insn
23651 should be only disconnected from the insn stream.
23652 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
23654 2013-04-30 Jakub Jelinek <jakub@redhat.com>
23656 PR tree-optimization/57104
23657 * tsan.c (instrument_expr): Don't instrument accesses to
23658 DECL_HARD_REGISTER VAR_DECLs.
23660 2013-04-30 Richard Biener <rguenther@suse.de>
23662 * function.h (loops_for_fn): New inline function.
23663 (set_loops_for_fn): Likewise.
23664 * cfgloop.h (place_new_loop): Add struct function parameter.
23665 (get_loop): Likewise.
23666 (get_loops): Likewise.
23667 (number_of_loops): Likewise.
23668 (fel_next): Adjust.
23669 (fel_init): Likewise.
23670 * cfg.c (get_loop_copy): Adjust.
23671 * cfgloop.c (flow_loops_dump): Likewise.
23672 (record_loop_exits): Likewise.
23673 (verify_loop_structure): Likewise.
23674 * cfgloopanal.c (mark_irreducible_loops): Likewise.
23675 (estimate_reg_pressure_cost): Likewise.
23676 (mark_loop_exit_edges): Likewise.
23677 * cfgloopmanip.c (place_new_loop): Likewise.
23678 (add_loop): Likewise.
23679 (duplicate_loop): Likewise.
23680 * graph.c (draw_cfg_nodes): Likewise.
23681 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
23682 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
23683 (extract_affine_chrec): Likewise.
23684 (build_scop_iteration_domain): Likewise.
23685 * graphite.c (graphite_initialize): Likewise.
23686 * ira-build.c (create_loop_tree_nodes): Likewise.
23687 (more_one_region_p): Likewise.
23688 (rebuild_regno_allocno_maps): Likewise.
23689 (mark_loops_for_removal): Likewise.
23690 (mark_all_loops_for_removal): Likewise.
23691 (remove_unnecessary_regions): Likewise.
23692 (ira_build): Likewise.
23693 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
23694 * loop-init.c (fix_loop_structure): Likewise.
23695 (gate_rtl_move_loop_invariants): Likewise.
23696 (gate_rtl_unswitch): Likewise.
23697 (gate_rtl_unroll_and_peel_loops): Likewise.
23698 (rtl_doloop): Likewise.
23699 * lto-streamer-in.c (input_cfg): Likewise.
23700 * lto-streamer-out.c (output_cfg): Likewise.
23701 * modulo-sched.c (sms_schedule): Likewise.
23702 * predict.c (tree_estimate_probability): Likewise.
23703 (tree_estimate_probability_driver): Likewise.
23704 (estimate_loops): Likewise.
23705 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
23706 (move_sese_region_to_fn): Likewise.
23707 (debug_loop_num): Likewise.
23708 * tree-chrec.c (chrec_evaluate): Likewise.
23709 (hide_evolution_in_other_loops_than_loop): Likewise.
23710 (chrec_component_in_loop_num): Likewise.
23711 (reset_evolution_in_loop): Likewise.
23712 (evolution_function_is_invariant_rec_p): Likewise.
23713 * tree-if-conv.c (main_tree_if_conversion): Likewise.
23714 * tree-inline.c (copy_loops): Likewise.
23715 (copy_cfg_body): Likewise.
23716 (tree_function_versioning): Likewise.
23717 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
23718 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
23720 (add_to_evolution_1): Likewise.
23721 (scev_const_prop): Likewise.
23722 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
23723 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
23724 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
23725 (tree_ssa_lim_initialize): Likewise.
23726 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
23727 (verify_loop_closed_ssa): Likewise.
23728 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
23729 (tree_ssa_loop_im): Likewise.
23730 (tree_ssa_loop_unswitch): Likewise.
23731 (tree_vectorize): Likewise.
23732 (check_data_deps): Likewise.
23733 (tree_ssa_loop_ivcanon): Likewise.
23734 (tree_ssa_loop_bounds): Likewise.
23735 (tree_complete_unroll): Likewise.
23736 (tree_complete_unroll_inner): Likewise.
23737 (tree_parallelize_loops): Likewise.
23738 (tree_ssa_loop_prefetch): Likewise.
23739 (tree_ssa_loop_ivopts): Likewise.
23740 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
23741 * tree-vectorizer.c (vectorize_loops): Likewise.
23743 2013-04-29 Mike Frysinger <vapier@gentoo.org>
23745 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
23746 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
23747 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
23748 with EABI_LINK_SPEC.
23750 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
23753 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
23756 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
23759 * lra-constraints.c (process_alt_operands): Discourage a bit more
23760 using memory for pseudos. Print cost dump for alternatives.
23761 Modify cost values for conflicts with early clobbers.
23762 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
23764 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
23767 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
23769 2013-04-29 Ian Bolton <ian.bolton@arm.com>
23771 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
23772 from/to S register.
23773 (movdi_aarch64): Support LDR/STR from/to D register.
23775 2013-04-29 Ian Bolton <ian.bolton@arm.com>
23777 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
23778 or higher by default.
23780 2013-04-29 Richard Biener <rguenther@suse.de>
23782 PR middle-end/57075
23783 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
23784 even if not adding abnormal edges for calls that can make
23787 2013-04-29 Richard Biener <rguenther@suse.de>
23789 PR middle-end/57103
23790 * tree-cfg.c (move_stmt_op): Fix condition under which to update
23792 (move_stmt_r): Remove redundant checking.
23794 2013-04-29 Teresa Johnson <tejohnson@google.com>
23797 * basic-block.h (apply_scale): New function.
23798 (apply_probability): Use apply_scale.
23799 * gimple-streamer-in.c (input_bb): Ditto.
23800 * lto-streamer-in.c (input_cfg): Ditto.
23801 * lto-cgraph.c (merge_profile_summaries): Ditto.
23802 * tree-optimize.c (execute_fixup_cfg): Ditto.
23803 * tree-inline.c (copy_bb): Update comment to use apply_scale.
23804 (copy_edges_for_bb): Ditto.
23805 (copy_cfg_body): Ditto.
23807 2013-04-29 Tom de Vries <tom@codesourcery.com>
23809 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
23810 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
23811 (tail_merge_optimize): Handle current_loops == NULL.
23813 2013-04-26 Jeff Law <law@redhat.com>
23815 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
23816 (simplify_cond_using_ranges): Generalize code to simplify
23817 COND_EXPRs where one argument is a constant and the other
23818 is an SSA_NAME created by an integral type conversion.
23820 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23822 * config/arm/arm.md (store_minmaxsi): Use only when
23823 optimize_insn_for_size_p.
23825 2013-04-29 Christian Bruel <christian.bruel@st.com>
23828 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
23830 2013-04-29 Richard Biener <rguenther@suse.de>
23832 PR middle-end/57089
23833 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
23834 loop tree make sure to schedule a fixup for the child as well.
23835 (expand_omp_for_generic): Properly add loops.
23836 (expand_omp_for_static_nochunk): Likewise.
23837 (expand_omp_for_static_chunk): Likewise.
23838 (expand_omp_for): For the degenerate case fixup loops.
23839 (expand_omp_sections): Fix default bb placement in loops.
23840 (expand_omp_atomic_pipeline): Properly add loops.
23842 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23844 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
23846 2013-04-29 Tom de Vries <tom@codesourcery.com>
23848 * tree-ssa-tail-merge.c: Update header comment.
23850 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
23852 * config/aarch64/arm_neon.h
23853 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
23854 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
23855 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
23856 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
23857 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
23858 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
23859 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
23860 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
23862 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
23864 * config/aarch64/aarch64-simd.md
23865 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
23866 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
23867 fix_trunc, fixuns_trunc.
23868 (ftrunc<VDQF:mode>2): New.
23869 * config/aarch64/iterators.md (optab): Add fix, fixuns.
23870 (fix_trunc_optab): New.
23872 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
23874 * config/aarch64/aarch64-builtins.c
23875 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
23876 iceilf, lround, iroundf.
23878 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
23881 * config/i386/i386.h (enum ix86_tune_indices)
23882 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
23883 New, split from X86_TUNE_INTER_UNIT_MOVES.
23884 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
23885 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
23886 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
23887 (TARGET_INTER_UNIT_MOVES): Remove.
23888 * config/i386/i386.c (initial_ix86_tune_features): Update.
23889 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
23890 (ix86_expand_convert_uns_didf_sse): Use
23891 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
23892 (ix86_expand_vector_init_one_nonzero): Ditto.
23893 (ix86_expand_vector_init_interleave): Ditto.
23894 (inline_secondary_memory_needed): Return true for moves from SSE class
23895 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
23896 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
23897 * config/i386/constraints.md (Yi, Ym): Depend on
23898 TARGET_INTER_UNIT_MOVES_TO_VEC.
23899 (Yj, Yn): New constraints.
23900 * config/i386/i386.md (*movdi_internal): Change constraints of
23901 operand 1 from Yi to Yj and from Ym to Yn.
23902 (*movsi_internal): Ditto.
23903 (*movdf_internal): Ditto.
23904 (*movsf_internal): Ditto.
23905 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
23906 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
23907 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
23908 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
23909 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
23910 * config/i386/sse.md (movdi_to_sse): Ditto.
23911 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
23912 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
23913 TARGET_INTER_UNIT_MOVES.
23914 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
23915 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
23916 instead of TARGET_INTER_UNIT_MOVES.
23917 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
23918 operand 1 from Yi to Yj and from Ym to Yn.
23920 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
23922 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
23923 (float_truncate_hi_): Likewise.
23924 (float_extend_lo_): Likewise.
23925 (float_truncate_lo_): Likewise.
23926 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
23927 (aarch64_float_extend_lo_v2df): Likewise.
23928 (vec_unpacks_hi_v4sf): Likewise.
23929 (aarch64_float_truncate_lo_v2sf): Likewise.
23930 (aarch64_float_truncate_hi_v4sf): Likewise.
23931 (vec_pack_trunc_v2df): Likewise.
23932 (vec_pack_trunc_df): Likewise.
23934 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
23936 * config/aarch64/aarch64-builtins.c
23937 (aarch64_fold_builtin): Fold float conversions.
23938 * config/aarch64/aarch64-simd-builtins.def
23939 (floatv2si, floatv4si, floatv2di): New.
23940 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
23941 * config/aarch64/aarch64-simd.md
23942 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
23943 * config/aarch64/iterators.md (FLOATUORS): New.
23944 (optab): Add float, floatuns.
23945 (su_optab): Likewise.
23947 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
23949 * config/aarch64/aarch64-builtins.c
23950 (aarch64_builtin_vectorized_function): Use new names for
23952 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
23953 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
23954 (fcvtzu): Split as...
23955 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
23956 (fcvtas): Split as...
23957 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
23958 (fcvtau): Split as...
23959 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
23960 (fcvtps): Split as...
23961 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
23962 (fcvtpu): Split as...
23963 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
23964 (fcvtms): Split as...
23965 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
23966 (fcvtmu): Split as...
23967 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
23968 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
23969 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
23970 (lfrintnusf, lfrintnudf): Likewise.
23971 * config/aarch64/aarch64-simd.md
23972 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
23974 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
23975 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
23976 (fcvt_pattern): Likewise.
23978 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
23980 * config/aarch64/aarch64-simd.md
23981 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
23982 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
23984 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
23986 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
23987 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
23988 (vrnd<a,m,n,p>_f32): Implement using builtins.
23989 (vrnd<i,x><q>_f<32, 64>): New.
23991 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
23993 * config/aarch64/aarch64-builtins.c
23994 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
23995 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
23996 (frintz): Rename to...
23998 (frintp): Rename to...
24000 (frintm): Rename to...
24002 (frinti): Rename to...
24003 (nearbyint): ...this.
24004 (frintx): Rename to...
24006 (frinta): Rename to...
24008 * config/aarch64/aarch64-simd.md
24009 (aarch64_frint<frint_suffix><mode>): Delete.
24010 (<frint_pattern><mode>2): Convert to insn.
24011 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
24012 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
24013 (frint_pattern): Likewise.
24014 (frint_suffix): Likewise.
24016 2013-04-29 Richard Biener <rguenther@suse.de>
24018 PR tree-optimization/57081
24019 * loop-init.c: Include tree-flow.h.
24020 (loop_optimizer_finalize): Free number of iteration estimates.
24021 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
24023 2013-04-29 Jakub Jelinek <jakub@redhat.com>
24025 PR tree-optimization/57083
24026 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
24027 non-singleton shift count range, zero extend low_bound for uns case.
24029 * config/i386/predicates.md (general_vector_operand): New predicate.
24030 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
24031 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
24032 if they aren't nonimmediate operands. If their original values
24033 satisfy const_vector_equal_evenodd_p, don't shift them.
24034 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
24035 predicates. For the SSE4.1 case force operands[{1,2}] into registers
24036 if not nonimmediate_operand.
24037 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
24038 instead of register_operand.
24039 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
24041 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
24043 * stor-layout.c (finalize_size_functions): Allocate a structure and
24044 reset cfun before dumping the functions.
24046 2013-04-27 Jakub Jelinek <jakub@redhat.com>
24048 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
24051 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
24052 use xop_pmacsdqh if uns_p.
24053 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
24054 the immediate rotate count.
24056 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
24058 * rtl.h (struct rtx_def): Add comment for field jump.
24059 (LRA_SUBREG_P): New macro.
24060 * recog.c (register_operand): Check LRA_SUBREG_P.
24061 * lra.c (lra): Add note at the end of RTL code. Align non-empty
24063 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
24064 (lra_final_code_change): Skip subreg change for operators.
24065 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
24066 if there are no operand changes.
24067 * lra-constraints.c (curr_insn_set): New.
24068 (match_reload): Set LRA_SUBREG_P.
24069 (emit_spill_move): Ditto.
24070 (check_and_process_move): Use curr_insn_set. Process only single
24071 set insns. Don't initialize sec_mem_p and change_p.
24072 (simplify_operand_subreg): Use LRA_SUBREG_P.
24073 (reg_in_class_p): New function.
24074 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
24075 of #ifdef. Add code to remove cycling.
24076 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
24077 non-null disp. Reload inner instead of disp when base and index
24078 are null. Try to put lo_sum into register.
24079 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
24080 (check_and_process_move): Move code for move cost check to
24081 simple_move_p. Remove equiv_substitution.
24082 (simple_move_p): New function.
24083 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
24084 curr_insn_set. Call check_and_process_move only for single set
24085 insns. Use the new function. Move call of check_and_process_move
24086 after operand equiv substitution and address process.
24088 2013-04-26 Jakub Jelinek <jakub@redhat.com>
24091 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
24092 with nonlocal goto receivers or returns twice calls, ignore
24093 unininitialized values from abnormal edges to nl goto receiver
24094 or returns twice call.
24096 2013-04-26 Jakub Jelinek <jakub@redhat.com>
24098 PR tree-optimization/57051
24099 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
24100 and VEC_RSHIFT_EXPR if shift count is a multiple of element
24103 2013-04-26 Richard Biener <rguenther@suse.de>
24105 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
24106 (expand_omp_taskreg): Likewise. Mark loops for fixup.
24107 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
24108 (fixup_loop_arrays_after_move): New function.
24109 (move_sese_region_to_fn): Properly outline the loop tree parts
24110 of the SESE region.
24112 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
24114 * config/i386/i386.md (type, unit): Fix long lines.
24116 2013-04-26 Richard Biener <rguenther@suse.de>
24118 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
24119 (lto-streamer-out.o): Likewise.
24120 * cfgloop.c (init_loops_structure): Export, add struct function
24121 argument and adjust.
24122 (flow_loops_find): Adjust.
24123 * cfgloop.h (enum loop_estimation): Add EST_LAST.
24124 (init_loops_structure): Declare.
24125 * lto-streamer-in.c: Include cfgloop.h.
24126 (input_cfg): Input the loop tree.
24127 * lto-streamer-out.c: Include cfgloop.h.
24128 (output_cfg): Output the loop tree.
24129 (output_struct_function_base): Do not drop PROP_loops.
24131 2013-03-26 Richard Biener <rguenther@suse.de>
24133 * tree-cfg.c (execute_build_cfg): Build the loop tree.
24134 (pass_build_cfg): Provide PROP_loops.
24135 (move_sese_region_to_fn): Remove loops that are outlined into fn
24137 * tree-inline.c: Include cfgloop.h.
24138 (initialize_cfun): Do not drop PROP_loops.
24139 (copy_loops): New function.
24140 (copy_cfg_body): Copy loop structure.
24141 (tree_function_versioning): Initialize destination loop tree.
24142 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
24143 (pass_parallelize_loops): Do IL verification.
24144 * loop-init.c (loop_optimizer_init): Fixup loops if required.
24145 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
24146 the CFG make sure we fixup loops as well.
24147 * tree-ssa-tail-merge.c: Include cfgloop.h.
24148 (replace_block_by): When merging loop latches mark loops for fixup.
24149 * lto-streamer-out.c (output_struct_function_base): Drop
24150 PROP_loops for now.
24151 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
24152 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
24153 * ipa-split.c: Include cfgloop.h.
24154 (split_function): Add the new return block to the loop tree root.
24155 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
24156 whether we have removed the forwarder block.
24157 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
24158 * cfgloop.h (place_new_loop): Declare.
24159 * cfgloopmanip.c (place_new_loop): Export.
24160 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
24161 (tree-switch-conversion.o): Likewise.
24162 (tree-complex.o): Likewise.
24163 (tree-inline.o): Likewise.
24164 (tree-ssa-tailmerge.o): Likewise.
24165 (ipa-split.o): Likewise.
24166 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
24167 (tree-ssa-copy.o): Likewise.
24168 * tree-switch-conversion.c: Include cfgloop.h
24169 (process_switch): If we emit a bit-test cascade, schedule loops
24171 * tree-complex.c: Include cfgloop.h.
24172 (expand_complex_div_wide): Properly add new basic-blocks to loops.
24173 * asan.c: Include cfgloop.h.
24174 (create_cond_insert_point): Properly add new basic-blocks to
24175 loops, schedule loop fixup.
24176 * cfgloop.c (verify_loop_structure): Check that looks are not
24178 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
24180 (expand_omp_for_generic): Likewise.
24181 (expand_omp_sections): Likewise.
24182 (expand_omp_atomic_pipeline): Schedule loops for fixup.
24183 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
24184 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
24185 is initialized, not when loops are present.
24186 * tree-parloops.c (parallelize_loops): Remove checking here.
24187 * passes.c (init_optimization_passes): Schedule a copy-propagation
24188 pass before complete unrolling of inner loops.
24190 2013-04-26 Jakub Jelinek <jakub@redhat.com>
24192 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
24193 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
24194 (colorize_init): Add argument to _WIN32 version.
24195 * toplev.c: Include diagnostic-color.h.
24196 (process_options): Default to -fdiagnostics-color=auto if
24197 GCC_COLORS env var is in the environment.
24198 * common.opt (fdiagnostics-color=): Add Var and Init.
24199 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
24200 env var is in the environment, the default is auto rather than never.
24202 * diagnostic.h (file_name_as_prefix): Add context argument.
24203 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
24204 the string as locus.
24205 * langhooks.c (lhd_print_error_function): Adjust caller.
24207 2013-04-25 Lawrence Crowl <crowl@google.com>
24209 * var-tracking.c (shared_hash_def::htab):
24210 Change type to hash_table. Update dependent calls and types.
24212 2013-04-25 Lawrence Crowl <crowl@google.com>
24214 * Makefile.in: Update as needed below.
24216 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
24217 Move declaration to after the type's method definitons.
24219 * attribs.c (htab_t scoped_attributes::attribute_hash):
24220 Change type to hash_table. Update dependent calls and types.
24222 * bitmap.c (htab_t bitmap_desc_hash):
24223 Change type to hash_table. Update dependent calls and types.
24225 * cselib.c (htab_t cselib_hash_table):
24226 Change type to hash_table. Update dependent calls and types.
24228 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
24229 (hash_string_slot_node): Move implementation into lto-streamer.h
24230 struct string_slot_hasher.
24231 (eq_string_slot_node): Likewise.
24233 * data-streamer-out.c: Update output_block::string_hash_table
24234 dependent calls and types.
24236 * dwarf2cfi.c (htab_t trace_index):
24237 Change type to hash_table. Update dependent calls and types.
24239 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
24240 Change type to hash_table. Update dependent calls and types.
24241 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
24242 (htab_t optimize_external_refs::map): Likewise.
24243 (htab_t output_comp_unit::extern_map): Likewise.
24244 (htab_t output_comdat_type_unit::extern_map): Likewise.
24245 (htab_t output_macinfo::macinfo_htab): Likewise.
24246 (htab_t optimize_location_lists::htab): Likewise.
24247 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
24249 * except.c (htab_t ehspec_hash_type):
24250 Change type to hash_table. Update dependent calls and types.
24251 (assign_filter_values::ttypes): Likewise.
24252 (assign_filter_values::ehspec): Likewise.
24253 (sjlj_assign_call_site_values::ar_hash): Likewise.
24254 (convert_to_eh_region_ranges::ar_hash): Likewise.
24256 * gcse.c (htab_t pre_ldst_table):
24257 Change type to hash_table. Update dependent calls and types.
24259 * ggc-common.c (htab_t saving_htab):
24260 Change type to hash_table. Update dependent calls and types.
24261 (htab_t loc_hash): Likewise.
24262 (htab_t ptr_hash): Likewise.
24263 (call_count): Rename ggc_call_count.
24264 (call_alloc): Rename ggc_call_alloc.
24265 (loc_descriptor): Rename make_loc_descriptor.
24266 (add_statistics): Rename ggc_add_statistics.
24268 * ggc-common.c (saving_htab):
24269 Change type to hash_table. Update dependent calls and types.
24271 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
24272 (push_gimplify_context): Likewise.
24273 (pop_gimplify_context): Likewise.
24274 (struct gimple_temp_hash_elt): Added.
24275 (struct gimplify_hasher): Likewise.
24276 (struct gimplify_ctx.temp_htab):
24277 Change type to hash_table. Update dependent calls and types.
24279 * gimple-fold.c: Include gimplify-ctx.h.
24281 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
24282 Change type to hash_table. Update dependent calls and types.
24283 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
24284 avoid potential global name collision.
24286 * gimplify.c: Include gimplify-ctx.h.
24287 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
24288 (htab_t gimplify_ctx::temp_htab):
24289 Update dependent calls and types for new type hash_table.
24290 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
24291 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
24293 * gimplify-ctx.h: New.
24294 (struct gimple_temp_hash_elt): Move from gimplify.c.
24295 (class gimplify_hasher): New.
24296 (struct gimplify_ctx): Move from gimple.h.
24297 (htab_t gimplify_ctx::temp_htab):
24298 Change type to hash_table. Update dependent calls and types.
24300 * graphite-clast-to-gimple.c: Include graphite-htab.h.
24301 (htab_t ivs_params::newivs_index):
24302 Change type to hash_table. Update dependent calls and types.
24303 (htab_t ivs_params::params_index): Likewise.
24304 (htab_t print_generated_program::params_index): Likewise.
24305 (htab_t gloog::newivs_index): Likewise.
24306 (htab_t gloog::params_index): Likewise.
24308 * graphite.c: Include graphite-htab.h.
24309 4htab_t graphite_transform_loops::bb_pbb_mapping):
24310 Change type to hash_table. Update dependent calls and types.
24312 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
24313 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
24314 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
24316 * graphite-dependences.c: Include graphite-htab.h.
24317 (loop_is_parallel_p): Change hash table type of parameter.
24319 * graphite-htab.h: New.
24320 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
24321 (extern find_pbb_via_hash): Move from graphite-poly.h.
24322 (extern loop_is_parallel_p): Move from graphite-poly.h.
24323 (extern get_loop_body_pbbs): Move from graphite-poly.h.
24325 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
24326 (extern loop_is_parallel_p): Move to graphite-htab.h.
24327 (extern get_loop_body_pbbs): Move to graphite-htab.h.
24329 * haifa-sched.c (htab_t delay_htab):
24330 Change type to hash_table. Update dependent calls and types.
24331 (htab_t delay_htab_i2): Likewise.
24333 * ira-color.c (htab_t allocno_hard_regs_htab):
24334 Change type to hash_table. Update dependent calls and types.
24336 * ira-costs.c (htab_t cost_classes_htab):
24337 Change type to hash_table. Update dependent calls and types.
24339 * loop-invariant.c (htab_t merge_identical_invariants::eq):
24340 Change type to hash_table. Update dependent calls and types.
24342 * loop-iv.c (htab_t bivs):
24343 Change type to hash_table. Update dependent calls and types.
24345 * loop-unroll.c (htab_t opt_info::insns_to_split):
24346 Change type to hash_table. Update dependent calls and types.
24347 (htab_t opt_info::insns_with_var_to_expand): Likewise.
24349 * lto-streamer.h (struct string_slot): Move from data-streamer.h
24350 (struct string_slot_hasher): New.
24351 (htab_t output_block::string_hash_table):
24352 Change type to hash_table. Update dependent calls and types.
24354 * lto-streamer-in.c (freeing_string_slot_hasher): New.
24355 (htab_t file_name_hash_table):
24356 Change type to hash_table. Update dependent calls and types.
24358 * lto-streamer-out.c: Update output_block::string_hash_table dependent
24361 * lto-streamer.c (htab_t tree_htab):
24362 Change type to hash_table. Update dependent calls and types.
24364 * omp-low.c: Include gimplify-ctx.h.
24366 * passes.c (htab_t name_to_pass_map):
24367 Change type to hash_table. Update dependent calls and types.
24368 (pass_traverse): Rename to passes_pass_traverse.
24370 * plugin.c (htab_t event_tab):
24371 Change type to hash_table. Update dependent calls and types.
24373 * postreload-gcse.c (htab_t expr_table):
24374 Change type to hash_table. Update dependent calls and types.
24375 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
24377 * sese.c (debug_rename_map_1): Make extern.
24378 (htab_t copy_bb_and_scalar_dependences::rename_map):
24379 Change type to hash_table. Update dependent calls and types.
24381 * sese.h (extern debug_rename_map): Move to .c file.
24383 * store-motion.c (htab_t store_motion_mems_table):
24384 Change type to hash_table. Update dependent calls and types.
24386 * trans-mem.c (htab_t tm_new_mem_hash):
24387 Change type to hash_table. Update dependent calls and types.
24389 * tree-browser.c (htab_t TB_up_ht):
24390 Change type to hash_table. Update dependent calls and types.
24392 * tree-cfg.c (htab_t discriminator_per_locus):
24393 Change type to hash_table. Update dependent calls and types.
24395 * tree-complex.c: Include tree-hasher.h
24396 (htab_t complex_variable_components):
24397 Change type to hash_table. Update dependent calls and types.
24399 * tree-eh.c (htab_t finally_tree):
24400 Change type to hash_table. Update dependent calls and types.
24402 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
24403 struct int_tree_hasher.
24404 (extern int_tree_map_eq): Likewise.
24405 (uid_decl_map_hash): Removed.
24406 (extern decl_tree_map_eq): Likewise.
24408 * tree-hasher.h: New.
24409 (struct int_tree_hasher): New.
24410 (typedef int_tree_htab_type): New.
24412 * tree-inline.c: Include gimplify-ctx.h.
24414 * tree-mudflap.c: Include gimplify-ctx.h.
24416 * tree-parloops.c: Include tree-hasher.h.
24417 (htab_t eliminate_local_variables_stmt::decl_address):
24418 Change type to hash_table. Update dependent calls and types.
24419 (htab_t separate_decls_in_region::decl_copies): Likewise.
24421 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
24422 Change type to hash_table. Update dependent calls and types.
24424 * tree-sra.c (candidates):
24425 Change type to hash_table. Update dependent calls and types.
24427 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
24429 (int_tree_map_hash): Likewise.
24431 * tree-ssa-dom.c (htab_t avail_exprs):
24432 Change type to hash_table. Update dependent calls and types.
24434 * tree-ssa-live.c (var_map_base_init::tree_to_index):
24435 Change type to hash_table. Update dependent calls and types.
24437 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
24438 Change type to hash_table. Update dependent calls and types.
24440 * tree-ssa-phiopt.c (seen_ssa_names):
24441 Change type to hash_table. Update dependent calls and types.
24443 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
24444 Change type to hash_table. Update dependent calls and types.
24446 * tree-ssa-uncprop.c (equiv):
24447 Change type to hash_table. Update dependent calls and types.
24449 2013-04-25 Jakub Jelinek <jakub@redhat.com>
24451 PR rtl-optimization/57003
24452 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
24453 call note_stores with kill_clobbered_value callback again after
24454 killing regs_invalidated_by_call.
24456 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
24458 * config/aarch64/aarch64-simd.md
24459 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
24460 (aarch64_simd_bsl<mode>): Likewise.
24461 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
24463 2013-04-25 Marek Polacek <polacek@redhat.com>
24465 PR tree-optimization/57066
24466 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
24468 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
24470 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
24472 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
24474 * config/aarch64/aarch64-builtins.c
24475 (aarch64_fold_builtin): New.
24476 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
24477 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
24478 * config/aarch64/aarch64-simd-builtins.def (abs): New.
24479 * config/aarch64/arm_neon.h
24480 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
24482 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
24483 Tejas Belagod <tejas.belagod@arm.com>
24485 * config/aarch64/aarch64-builtins.c
24486 (aarch64_gimple_fold_builtin): New.
24487 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
24488 * config/aarch64/aarch64-simd-builtins.def (addv): New.
24489 * config/aarch64/aarch64-simd.md (addpv4sf): New.
24490 (addvv4sf): Update.
24491 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
24493 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
24495 * config/aarch64/aarch64.md
24496 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
24498 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
24500 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
24501 (*ngcsi_uxtw): New pattern.
24503 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24504 Julian Brown <julian@codesourcery.com>
24506 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
24507 (TB_DREG): Add T_V4HF.
24508 (v4hf_UP): New macro.
24509 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
24510 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
24511 Handle initialisation of V4HF. Adjust initialisation of reinterpret
24513 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
24514 (arm_vector_mode_supported_p): Handle V4HF.
24515 (arm_mangle_map): Handle V4HFmode.
24516 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
24517 * config/arm/arm_neon_builtins.def: Add entries for
24518 vcvtv4hfv4sf, vcvtv4sfv4hf.
24519 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
24520 (neon_vcvtv4hfv4sf): Likewise.
24521 * config/arm/neon-gen.ml: Handle half-precision floating point
24523 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
24524 * config/arm/arm_neon.h: Regenerate.
24525 * config/arm/neon.ml (type elts): Add F16.
24526 (type vectype): Add T_float16x4, T_floatHF.
24527 (type vecmode): Add V4HF.
24528 (type features): Add Requires_FP_bit feature.
24529 (elt_width): Handle F16.
24530 (elt_class): Likewise.
24531 (elt_of_class_width): Likewise.
24532 (mode_of_elt): Refactor.
24533 (type_for_elt): Handle F16, fix error messages.
24534 (vectype_size): Handle T_float16x4.
24535 (vcvt_sh): New function.
24536 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
24537 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
24538 (string_of_mode): Handle V4HF.
24539 * doc/arm-neon-intrinsics.texi: Regenerate.
24541 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
24543 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
24544 format specifier in 'X' case.
24546 2013-04-25 Alan Modra <amodra@gmail.com>
24549 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
24550 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
24551 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
24552 Repeat for many other rotate/shift and mask patterns using subregs.
24553 Name lshiftrt insns.
24554 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
24555 on WORDS_BIG_ENDIAN.
24557 2013-04-25 Alan Modra <amodra@gmail.com>
24559 * config.gcc: Support little-endian powerpc-linux targets.
24560 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
24561 (LINK_OS_LINUX_SPEC): Define.
24562 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
24563 Preserve MASK_LITTLE_ENDIAN.
24564 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
24565 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
24566 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
24567 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
24568 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
24569 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
24570 Correct fp word order for little-endian. Don't shift toc entries
24571 smaller than a word for little-endian.
24572 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
24573 (bswapdi2 splits): Correct low-part subreg for little-endian.
24574 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
24575 low/high where such is correct only for be.
24576 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
24577 little-endian for -mcall-aixdesc.
24579 2013-04-25 Alan Modra <amodra@gmail.com>
24581 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
24582 replace_equiv_address_nv.
24584 2013-04-25 Alan Modra <amodra@gmail.com>
24586 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
24588 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
24591 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
24592 * rtl.h (struct rtx_def): ...
24594 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
24596 PR rtl-optimizations/57046
24597 * lra-constraints (split_reg): Set up lra_risky_transformations_p
24598 for multi-reg splits.
24600 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
24602 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
24604 2013-04-24 Sterling Augustine <saugustine@google.com>
24606 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
24607 (comp_dir_string, debug_str_dwo_section): New.
24608 (DEBUG_STR_DWO_SECTION): Rename to ...
24609 (DEBUG_DWO_STR_SECTION): ... this.
24610 (DEBUG_NORM_STR_SECTION): Delete.
24611 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
24612 (DEBUG_STR_DWO_SECTION_FLAGS): New.
24613 (find_AT_string): Move most logic to ...
24614 (find_AT_string_in_table): ... here. New.
24615 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
24616 add_skeleton_AT_string. Delete logic.
24617 (output_skeleton_debug_sections): Remove call to
24618 add_top_level_skeleton_die_attrs.
24619 (add_comp_dir_attribute): Move logic to comp_dir_string.
24620 (dwarf2out_init): Initialize debug_str_dwo_section.
24621 (output_indirect_string): Call find_string_form.
24622 (output_indirect_strings): Rewrite.
24623 (prune_unused_types): Empty skeleton_debug_str_hash.
24624 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
24625 (dwarf2out_finish): Call output_indirect_strings.
24627 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
24629 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
24631 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
24633 * rtl.h (struct rtx_def): Add comment for field jump.
24634 (LRA_SUBREG_P): New macro.
24635 * recog.c (register_operand): Check LRA_SUBREG_P.
24636 * lra.c (lra): Add note at the end of RTL code. Align non-empty
24638 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
24639 (lra_final_code_change): Skip subreg change for operators.
24640 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
24641 if there are no operand changes.
24642 * lra-constraints.c (curr_insn_set): New.
24643 (match_reload): Set LRA_SUBREG_P.
24644 (emit_spill_move): Ditto.
24645 (check_and_process_move): Use curr_insn_set. Process only single
24646 set insns. Don't initialize sec_mem_p and change_p.
24647 (simplify_operand_subreg): Use LRA_SUBREG_P.
24648 (reg_in_class_p): New function.
24649 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
24650 of #ifdef. Add code to remove cycling.
24651 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
24652 non-null disp. Reload inner instead of disp when base and index
24653 are null. Try to put lo_sum into register.
24654 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
24655 (check_and_process_move): Move code for move cost check to
24656 simple_move_p. Remove equiv_substitution.
24657 (simple_move_p): New function.
24658 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
24659 curr_insn_set. Call check_and_process_move only for single set
24660 insns. Use the new function. Move call of check_and_process_move
24661 after operand equiv substitution and address process.
24663 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
24665 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
24666 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
24667 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
24669 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
24671 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
24673 2013-04-24 Marek Polacek <polacek@redhat.com>
24675 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
24676 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
24677 (select_loops_exit_conditions): Likewise.
24678 (number_of_iterations_for_all_loops): Likewise.
24679 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
24680 (scev_analysis): Likewise.
24682 2013-04-02 Catherine Moore <clm@codesourcery.com>
24683 Chao-ying Fu <fu@mips.com>
24685 * config/mips/micromips.md (jraddiusp): New pattern.
24686 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
24687 instruction if possible.
24689 2013-04-24 Alan Modra <amodra@gmail.com>
24691 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
24693 2013-04-24 Julian Brown <julian@codesourcery.com>
24694 Chung-Lin Tang <cltang@codesourcery.com>
24696 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
24697 dependency behavior in enumeration type DIE generation. Add TODO note
24698 to comments about future DW_FORM_sdata/udata re-work of related code.
24700 2013-04-23 Lawrence Crowl <crowl@google.com>
24702 * Makefile.in: Update as needed below.
24704 * hash-table.h (class hash_table):
24705 Correct many methods with parameter types compare_type to the correct
24706 value_type. (Correct code was unlikely to notice the change.)
24707 (hash_table::elements_with_deleted) New.
24708 (class hashtable::iterator): New.
24709 (hashtable::begin()): New.
24710 (hashtable::end()): New.
24711 (FOR_EACH_HASH_TABLE_ELEMENT): New.
24713 * statistics.c (statistics_hashes):
24714 Change type to hash_table. Update dependent calls and types.
24716 * tree-into-ssa.c (var_infos):
24717 Change type to hash_table. Update dependent calls and types.
24719 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
24720 Change type to hash_table. Update dependent calls and types.
24722 * tree-ssa-loop-im.c (struct mem_ref.refs):
24723 Change type to hash_table. Update dependent calls and types.
24725 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
24726 Change type to hash_table. Update dependent calls and types.
24728 * tree-ssa-sccvn.c (vn_tables_s::nary):
24729 Change type to hash_table. Update dependent calls and types.
24730 (vn_tables_s::phis): Likewise.
24731 (vn_tables_s::references): Likewise.
24733 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
24734 (vn_reference_eq): Update parameter and return types.
24736 * tree-ssa-structalias.c (pointer_equiv_class_table):
24737 Change type to hash_table. Update dependent calls and types.
24738 (location_equiv_class_table): Likewise.
24740 * tree-vect-data-refs.c: Consequential changes for making
24741 peeling a hash_table.
24743 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
24744 (destroy_loop_vec_info): Dependent hash_table update.
24746 * tree-vectorizer.h (peeling_htab):
24747 Change type to hash_table. Update dependent calls and types.
24749 2013-04-23 Shiva Chen <shiva0217@gmail.com>
24751 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
24752 to check the register content is equal or not.
24753 * lra-constraints.c (match_reload): Use lra_assign_reg_val
24754 to assign register content record.
24755 * lra-eliminations.c (update_reg_eliminate): Use
24756 lra_update_reg_val_offset to update register content offset.
24757 * lra-int.h (struct lra_reg): Add offset member.
24758 (lra_reg_val_equal_p): New static inline function.
24759 (lra_update_reg_val_offset): New static inline function.
24760 (lra_assign_reg_val): New static inline function.
24761 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
24762 to assign register content record.
24763 (initialize_lra_reg_info_element): Initial offset to zero.
24765 2013-04-23 Catherine Moore <clm@codesourcery.com>
24767 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
24768 operands. Record compression.
24770 2013-04-23 Xinliang David Li <davidxl@google.com>
24772 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
24774 2013-04-23 Richard Biener <rguenther@suse.de>
24776 PR middle-end/57036
24777 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
24778 parameter, only add abnormal goto edges from the copied body
24779 if the call could perform abnormal gotos.
24780 (copy_cfg_body): Adjust.
24782 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
24784 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
24786 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
24788 * coretypes.h (gimple_stmt_iterator): Add struct to make
24791 2013-04-23 Richard Biener <rguenther@suse.de>
24793 PR tree-optimization/57026
24794 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
24795 from SSA names occuring in abnormal PHI nodes.
24797 2013-04-22 Andi Kleen <ak@linux.intel.com>
24799 * lto/lto.c (print_lto_report_1): Fix LTO report names.
24801 2013-04-22 Andi Kleen <ak@linux.intel.com>
24803 * lto/lto.c (print_lto_report_1): Declare early.
24804 (read_cgraph_and_symbols): Call print_lto_report_1 early.
24806 2013-04-22 Andi Kleen <ak@linux.intel.com>
24808 * common.opt (-flto-report-wpa): Add.
24809 * doc/invoke.texi (-flto-report-wpa): Add.
24810 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
24813 2013-04-22 Xinliang David Li <davidxl@google.com>
24815 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
24816 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
24817 * Makefile.in: New dependency
24819 David Daney <ddaney.cavm@gmail.com>
24821 * configure.ac (gcc_cv_as_micromips_support): Use the
24822 --fatal-warnings option.
24823 * configure: Regenerate.
24825 2013-04-22 Marek Polacek <polacek@redhat.com>
24828 * tsan.c (instrument_expr): Don't instrument expression
24829 in case its size is zero.
24831 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
24835 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
24837 * config/alpha/alpha.c (TARGET_LRA_P): New define.
24839 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
24841 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
24842 (gimple_stmt_iterator): New typedef.
24843 * gimple.h (gimple_stmt_iterator): Rename to...
24844 (gimple_stmt_iterator_d): ... This.
24845 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
24846 trees be valid for GIMPLE and GENERIC.
24847 (TARGET_GIMPLE_FOLD_BUILTIN): New.
24848 * gimple-fold.c (gimple_fold_call): Call target hook
24849 gimple_fold_builtin.
24850 * hooks.c (hook_bool_gsiptr_false): New.
24851 * hooks.h (hook_bool_gsiptr_false): New.
24852 * target.def (fold_stmt): New.
24853 * doc/tm.texi: Regenerate.
24855 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
24858 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
24859 a set sp if no stack realignment.
24861 2013-04-22 Nick Clifton <nickc@redhat.com>
24863 * config.gcc (tilegx-linux): Extend extra_objs rather than
24865 (tilepro-linux): Likewise.
24867 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
24869 * config/aarch64/aarch64-builtins.c
24871 (CF0, CF1, CF2, CF3, CF4, CF10): New.
24872 (VAR<1-12>): Add MAP parameter.
24873 (BUILTIN_*): Likewise.
24874 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
24875 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
24876 (aarch64_ushl_n<mode>): Likewise.
24877 (aarch64_sshr_n<mode>): Likewise.
24878 (aarch64_ushr_n<mode>): Likewise.
24879 (aarch64_<maxmin><mode>): Likewise.
24880 (aarch64_sqrt<mode>): Likewise.
24881 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
24882 (vshr<q>_n_*): Likewise.
24884 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
24886 * config/aarch64/aarch64-builtins.c
24887 (aarch64_simd_builtin_type_mode): Handle SF types.
24889 (BUILTIN_GPF): Define.
24890 (aarch64_init_simd_builtins): Handle SF types.
24891 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
24892 (frecps): Likewise.
24893 (frecpx): Likewise.
24894 * config/aarch64/aarch64-simd.md
24895 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
24896 (aarch64_frecpe<mode>): New.
24897 (aarch64_frecps<mode>): Likewise.
24898 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
24899 (v8type): Add frecp<esx>.
24900 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
24901 (aarch64_frecps<mode>): Likewise.
24902 * config/aarch64/iterators.md (FRECP): New.
24903 (frecp_suffix): Likewise.
24904 * config/aarch64/arm_neon.h
24905 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
24907 2013-04-22 Christian Bruel <christian.bruel@st.com>
24910 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
24911 (REG_CLASS_NAMES): Idem.
24912 (REG_CLASS_CONTENTS): Idem.
24913 (REGCLASS_HAS_FP_REG): Idem.
24914 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
24915 (sh_conditional_register_usage): Idem.
24917 2013-04-21 Jeff Law <law@redhat.com>
24919 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
24920 (ssa_forward_propagate_and_combine): Use it.
24922 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
24924 * lra.c: Update the flow chart diagram.
24926 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
24928 PR rtl-optimization/56847
24929 * lra-constraints.c (process_alt_operands): Discourage alternative
24930 with non-matche doffsettable memory constraint fro memory with
24933 2013-04-19 Richard Biener <rguenther@suse.de>
24935 PR tree-optimization/56982
24936 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
24938 * gimplify.c (gimplify_call_expr): Notice special calls.
24939 (gimplify_modify_expr): Likewise.
24940 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
24941 abnormal control flow receivers.
24942 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
24943 in the same way as cfun->has_nonlocal_labels.
24944 (gimple_purge_dead_abnormal_call_edges): Likewise.
24945 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
24946 receivers start a basic-block.
24948 2013-04-19 Richard Biener <rguenther@suse.de>
24950 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
24952 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
24953 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
24954 (SLP_TREE_LOAD_PERMUTATION): Add.
24955 (vect_transform_slp_perm_load): Adjust prototype.
24956 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
24957 (vect_free_slp_instance): Likewise.
24958 (vect_create_new_slp_node): Likewise.
24959 (vect_supported_slp_permutation_p): Remove.
24960 (vect_slp_rearrange_stmts): Adjust.
24961 (vect_supported_load_permutation_p): Likewise. Inline
24962 vect_supported_slp_permutation_p here.
24963 (vect_analyze_slp_instance): Compute load permutations per
24964 slp node instead of per instance.
24965 (vect_get_slp_defs): Adjust.
24966 (vect_transform_slp_perm_load): Likewise.
24967 (vect_schedule_slp_instance): Remove redundant code.
24968 (vect_schedule_slp): Remove hack for PR56270, add it ...
24969 * tree-vect-stmts.c (vectorizable_load): ... here, do not
24970 CSE loads for SLP. Adjust.
24972 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
24974 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
24975 spelling in two comments.
24977 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
24980 * config/arm/arm.c (load_multiple_sequence): Require SP
24981 as base register for loads if SP is in the register list.
24983 2013-04-19 Martin Jambor <mjambor@suse.cz>
24985 PR tree-optimization/56718
24986 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
24987 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
24988 and made public. Adjusted all callers.
24989 (ipa_intraprocedural_devirtualization): New function.
24990 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
24991 (ipa_intraprocedural_devirtualization): Likewise.
24992 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
24994 2013-04-19 Richard Biener <rguenther@suse.de>
24996 PR tree-optimization/57000
24997 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
24999 2013-04-19 Terry Guo <terry.guo@arm.com>
25001 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
25003 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
25004 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
25005 (cortex_m4_fmacs): Use new reservations.
25006 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
25008 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
25010 PR rtl-optimization/56999
25011 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
25013 (lra_coalesce): Remove split_origin_bitmap and related code.
25014 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
25015 ranges if necessary.
25017 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
25019 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
25021 (ix86_expand_call): Remove clobbered_registers array and use
25022 x86_64_ms_sysv_extra_clobbered_registers instead.
25023 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
25025 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
25027 * config/i386/i386.md (*call_rex64_ms_sysv): Use
25028 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
25029 (*call_value_rex64_ms_sysv): Ditto.
25031 2013-04-18 Cary Coutant <ccoutant@google.com>
25033 * dwarf2out.c (output_pubnames): Check die_perennial_p of
25034 parent instead of die_mark.
25036 2013-04-18 Diego Novillo <dnovillo@google.com>
25038 * gimple.c (create_gimple_tmp): New.
25039 (get_expr_type): New.
25040 (build_assign): New.
25041 (build_type_cast): New.
25042 * gimple.h (enum ssa_mode): Define.
25043 (gimple_seq_set_location): New.
25044 * asan.c (build_check_stmt): Change some gimple_build_* calls
25045 to use build_assign and build_type_cast.
25047 2013-04-18 Richard Biener <rguenther@suse.de>
25049 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
25050 handle negative step. Remove redundant checks.
25051 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
25052 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
25053 for negative step and grouped loads fail to vectorize.
25055 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
25057 * emit-rtl.c (reset_insn_used_flags): New function.
25058 (reset_all_used_flags): Use it.
25059 (verify_insn_sharing): New function.
25060 (verify_rtl_sharing): Fix verification for SEQUENCEs.
25062 2013-04-18 Jakub Jelinek <jakub@redhat.com>
25064 PR tree-optimization/56984
25065 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
25066 and (x >> M) >= N don't register any assertion if N << M is the
25069 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
25071 * lower-subreg.c (resolve_simple_move): If called self-recursive,
25072 do not delete_insn insns that have not yet been emitted, only
25073 unlink them with remove_insn.
25074 * df-scan.c (df_insn_delete): Revert r197492.
25076 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
25078 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
25079 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
25081 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
25083 * config/arm/arm.md (movsicc_insn): Convert define_insn into
25084 define_insn_and_split.
25085 (and_scc,ior_scc,negscc): Likewise.
25086 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
25088 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
25090 * config/arm/arm.c (use_return_insn): Return 0 for targets that
25091 can benefit from using a sequence of LDRD instructions in epilogue
25092 instead of a single LDM instruction.
25094 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
25097 * doc/extend.texi: Fix typo.
25099 2013-04-17 Richard Biener <rguenther@suse.de>
25101 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
25102 (vect_build_slp_tree): ... here.
25103 (vect_build_slp_tree_1): Compute which stmts of the SLP group
25104 match. Remove special-casing of mismatched complex loads.
25105 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
25106 re-try the match with swapped commutative operands.
25107 (vect_supported_load_permutation_p): Remove special-casing of
25108 mismatched complex loads.
25109 (vect_analyze_slp_instance): Adjust.
25111 2013-04-17 Richard Biener <rguenther@suse.de>
25113 PR rtl-optimization/56921
25114 * cfgloop.h (struct loop): Add simple_loop_desc member.
25115 (struct niter_desc): Mark with GTY(()).
25116 (simple_loop_desc): Do not use aux field but simple_loop_desc.
25117 * loop-iv.c (get_simple_loop_desc): Likewise.
25118 (free_simple_loop_desc): Likewise.
25121 2013-04-16 Richard Biener <rguenther@suse.de>
25123 PR rtl-optimization/56921
25124 * loop-init.c (pass_rtl_move_loop_invariants): Add
25125 TODO_do_not_ggc_collect to todo_flags_finish.
25126 (pass_rtl_unswitch): Same.
25127 (pass_rtl_unroll_and_peel_loops): Same.
25128 (pass_rtl_doloop): Same.
25130 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
25132 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
25133 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
25134 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
25136 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
25137 * tree-streamer.c (record_common_node): Adjust reference in comment.
25139 2013-04-17 Terry Guo <terry.guo@arm.com>
25141 * config/arm/cortex-m4.md: Add a new bypass.
25143 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25145 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
25147 (*subs_<optab><mode>_multp2): New pattern.
25148 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
25149 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
25151 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25153 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
25154 (*subs_mul_imm_<mode>): New pattern.
25156 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
25159 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
25160 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
25161 (vsx_movti_32bit): Change j->wa to O->wa.
25163 2013-04-16 Richard Biener <rguenther@suse.de>
25165 PR rtl-optimization/56921
25166 * loop-init.c (pass_rtl_move_loop_invariants): Add
25167 TODO_do_not_ggc_collect to todo_flags_finish.
25168 (pass_rtl_unswitch): Same.
25169 (pass_rtl_unroll_and_peel_loops): Same.
25170 (pass_rtl_doloop): Same.
25172 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
25174 * config/arm/arm.c (emit_multi_reg_push): New declaration
25175 for an existing function.
25176 (arm_emit_strd_push): New function.
25177 (arm_expand_prologue): Used here.
25178 (arm_emit_ldrd_pop): New function.
25179 (arm_expand_epilogue): Used here.
25180 (arm_get_frame_offsets): Update condition.
25181 (arm_emit_multi_reg_pop): Add a special case for load of a single
25182 register with writeback.
25184 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
25186 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
25189 2013-04-16 Richard Biener <rguenther@suse.de>
25191 PR tree-optimization/56756
25192 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
25193 (first_mem_ref_loc): New.
25194 (execute_sm): Place the load temporarily before a previous
25195 access instead of in the latch edge to ensure its SSA dependencies
25196 are defined at points dominating the load.
25198 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
25200 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
25201 correct fix by moving header and footer insn to the footer of
25202 the merged basic block. Clear BB_END of the merged-away block.
25204 PR middle-end/43631
25205 * emit-rtl.c (make_note_raw): New function.
25206 (link_insn_into_chain): New static inline function.
25207 (add_insn): Use it.
25208 (add_insn_before, add_insn_after): Factor insn chain linking code...
25209 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
25210 using link_insn_into_chain.
25211 (note_outside_basic_block_p): New helper function for emit_note_after
25212 and emit_note_before.
25213 (emit_note_after): Use nobb variant of add_insn_after if the note
25214 should not be contained in a basic block.
25215 (emit_note_before): Use nobb variant of add_insn_before if the note
25216 should not be contained in a basic block.
25217 (emit_note_copy): Use make_note_raw.
25218 (emit_note): Likewise.
25219 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
25220 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
25221 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
25222 the moved barrier the tail of the basic block it follows.
25223 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
25225 2013-04-15 Jakub Jelinek <jakub@redhat.com>
25227 PR tree-optimization/56962
25228 * gimple-ssa-strength-reduction.c (record_increment): Only set
25229 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
25230 either rhs1 or rhs2 is equal to c->base_expr.
25232 2013-04-15 Richard Biener <rguenther@suse.de>
25234 PR tree-optimization/56933
25235 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
25237 (GROUP_READ_WRITE_DEPENDENCE): Remove.
25238 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
25239 * tree-vect-data-refs.c (vect_analyze_group_access): Move
25240 dependence check ...
25241 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
25243 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
25244 GROUP_READ_WRITE_DEPENDENCE.
25246 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
25248 * emit-rtl.c (reset_all_used_flags): New function.
25249 (verify_rtl_sharing): Call reset_all_used_flags before and after
25250 performing the checks.
25252 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
25254 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
25255 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
25256 * config/arm/constraints.md (De): New constraint.
25257 * config/arm/neon.md (anddi3_neon): Delete.
25258 (neon_vand<mode>): Expand to standard anddi3 pattern.
25259 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
25260 Move earlier in the file.
25261 (neon_inv_logic_op2): Likewise.
25262 (arm_anddi_operand_neon): New predicate.
25264 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
25266 * configure.ac (gcc_cv_ld_as_needed): Set
25267 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
25268 Use -z ignore, -z record on *-*-solaris2*.
25269 (HAVE_LD_AS_NEEDED): Update comment.
25270 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
25271 * configure: Regenerate.
25272 * config.in: Regenerate.
25273 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
25274 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
25275 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
25276 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
25277 equivalents. Fix markup.
25278 * doc/tm.texi: Regenerate.
25280 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
25282 * config/i386/i386.opt: New option mstack-protector-guard=.
25283 * config/i386/i386-opts.h: Add enum stack_protector_guard.
25284 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
25285 TARGET_SSP_TLS_GUARD.
25286 * config/i386/i386.c (ix86_option_override_internal): Set
25287 ix86_stack_protector_guard.
25288 * config/i386/i386.md (stack_protect_set): Enable for
25289 TARGET_SSP_TLS_GUARD only.
25290 (stack_protect_set_<mode>): Ditto.
25291 (stack_protect_test): Ditto.
25292 (stack_protect_test_<mode>): Ditto.
25293 * doc/invoke.texi (i386 Option): Document.
25295 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
25298 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
25299 (S_MODES): Set H_MODE bit.
25300 (SF_MODES): Set only S_MODE and SF_MODE bits.
25301 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
25302 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
25303 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
25304 <MODE_FLOAT>: Likewise.
25306 2013-04-15 Joey Ye <joey.ye@arm.com>
25308 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
25310 2013-04-15 Joey Ye <joey.ye@arm.com>
25312 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
25314 (thumb_far_jump_used_p): Count instruction size and set
25317 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
25319 * reorg.c (fill_simple_delay_slots): Reindent block of code.
25320 * resource.c (mark_target_live_regs): Reformat conditional block.
25322 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
25324 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
25325 notes, they are emitted only just before final.
25326 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
25328 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
25330 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
25331 * cfgrtl.c (delete_insn): Call it here instead.
25332 * lra-spills.c (lra_final_code_change): Use delete_insn.
25333 * haifa-sched.c (sched_remove_insn): Likewise.
25334 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
25335 returning to the nop pool.
25336 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
25337 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
25339 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
25341 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
25342 * doc/tm.texi: Regenerated.
25344 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
25346 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
25349 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
25351 * df-core.c (df_find_def): Compare register numbers.
25352 (df_find_use): Likewise.
25354 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
25357 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
25358 lra_in_progress for return.
25360 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
25362 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
25363 define_insn into define_insn_and_split and emit movsicc patterns.
25365 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
25367 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
25369 2013-04-12 Richard Biener <rguenther@suse.de>
25371 * tree-pass.h (TODO_do_not_ggc_collect): New.
25372 * passes.c (execute_one_ipa_transform_pass): Honor
25373 TODO_do_not_ggc_collect.
25374 (execute_one_pass): Likewise.
25377 2013-04-10 Richard Biener <rguenther@suse.de>
25379 * passes.c (init_optimization_passes): Remove reload pass.
25380 * ira.c (do_reload): Merge into ...
25382 (rest_of_handle_reload): Remove.
25383 (pass_reload): Likewise.
25384 * config/i386/i386.c (ix86_option_override): Refer to ira instead
25385 of reload for vzeroupper pass placement.
25387 2013-04-12 Jakub Jelinek <jakub@redhat.com>
25389 PR tree-optimization/56918
25390 PR tree-optimization/56920
25391 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
25392 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
25393 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
25394 use wide_mul_with_sign method.
25396 2013-04-12 Richard Biener <rguenther@suse.de>
25398 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
25399 not be considered a gimple constant.
25401 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
25403 * fold-const.c (const_binop): Handle vector shifts by a scalar.
25404 (fold_binary_loc): Call const_binop also for mixed vector-scalar
25407 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
25408 Jakub Jelinek <jakub@redhat.com>
25410 * opts.c: Include diagnostic-color.h.
25411 (common_handle_option): Handle OPT_fdiagnostics_color_.
25412 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
25413 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
25414 (diagnostic-color.o): New.
25415 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
25416 (diagnostic_color_rule): New enum.
25417 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
25418 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
25419 the location string.
25420 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
25421 either NULL, or color kind.
25422 * diagnostic-color.c: New file.
25423 * diagnostic-color.h: New file.
25424 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
25426 * doc/invoke.texi (-fdiagnostics-color): Document.
25427 * pretty-print.h (pp_show_color): Define.
25428 (struct pretty_print_info): Add show_color field.
25429 * diagnostic.c: Include diagnostic-color.h.
25430 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
25431 macros. Colorize error:, warning: etc. strings and also the location
25433 (diagnostic_show_locus): Colorize the caret line.
25434 * pretty-print.c: Include diagnostic-color.h.
25435 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
25436 inside of %< %> quotes or quoted through q format modifier.
25438 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
25440 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
25442 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25444 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
25445 code in CC_NZ mode.
25446 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
25449 2013-04-11 Marek Polacek <polacek@redhat.com>
25451 PR tree-optimization/48184
25452 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
25454 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
25456 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
25457 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
25458 (skip_simple_arithmetic): Tidy up.
25459 * tree.h (skip_simple_constant_arithmetic): Declare.
25461 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25463 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
25465 2013-04-11 Richard Biener <rguenther@suse.de>
25467 * tree-vect-loop.c (get_initial_def_for_induction): Properly
25468 generate vector constants.
25470 2013-04-11 Richard Biener <rguenther@suse.de>
25472 PR tree-optimization/56878
25473 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
25474 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
25476 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
25477 Prefer to align the DR with the most invariant base address.
25479 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
25481 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
25484 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
25486 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
25487 floating-point vector comparisons against 0.
25489 2013-04-11 Jakub Jelinek <jakub@redhat.com>
25491 PR tree-optimization/56899
25492 * fold-const.c (extract_muldiv_1): Apply distributive law
25493 only if TYPE_OVERFLOW_WRAPS (ctype).
25495 2013-04-11 Bin Cheng <bin.cheng@arm.com>
25498 * ira-costs.c (scan_one_insn): Check whether the source rtx of
25499 loading has side effect.
25501 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
25503 * config/sparc/sparc.c: Include tree-pass.h.
25504 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
25505 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
25506 head of file. Change return type. Split off gate function.
25507 (sparc_gate_work_around_errata): New function.
25508 (pass_work_around_errata): New pass definition.
25509 (insert_pass_work_around_errata) New pass insert definition to
25510 insert pass_work_around_errata just after delayed-branch scheduling.
25511 (sparc_option_override): Insert the pass.
25512 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
25514 2013-04-10 David S. Miller <davem@davemloft.net>
25516 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
25517 or -mcpu=hypersparc.
25519 * target.def (cstore_mode): New hook.
25520 * target.h: Include insn-codes.h
25521 * targhooks.c: Likewise.
25522 (default_cstore_mode): New function.
25523 * targhooks.h: Declare it.
25524 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
25525 * doc/tm.texi: Rebuild.
25526 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
25527 target hook, rather than inspecting the insn_data.
25528 * config/sparc/sparc.c (sparc_cstore_mode): New function.
25529 (TARGET_CSTORE_MODE): Redefine.
25530 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
25532 * config/sparc/predicates.md (cstore_result_operand): New special
25534 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
25535 Use it for operand 0.
25536 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
25537 (*snesi_special): Likewise.
25538 (*snesi_zero): Likewise.
25539 (*seqsi_zero): Likewise.
25540 (*sltu_insn): Likewise.
25541 (*sgeu_insn): Likewise.
25542 (*seqdi_special): Make operand 0 and comparison operation be of
25544 (*snedi_special): Likewise.
25545 (*snedi_special_vis3): Likewise.
25546 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
25547 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
25548 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
25549 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
25550 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
25551 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
25552 (*sltu_extend_sp64): Likewise.
25553 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
25554 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
25555 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
25556 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
25557 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
25559 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
25561 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
25562 (aarch64_start_file): Use the new function.
25564 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
25565 Jason Merrill <jason@redhat.com>
25567 * common.opt: Add -gdwarf.
25568 * opts.c (common_handle_option): Handle it.
25569 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
25571 2013-04-10 Richard Biener <rguenther@suse.de>
25573 * passes.c (execute_todo): Do not call ggc_collect conditional here.
25574 (execute_one_ipa_transform_pass): But unconditionally here.
25575 (execute_one_pass): And here.
25576 (init_optimization_passes): Remove reload pass.
25577 * tree-pass.h (TODO_ggc_collect): Remove.
25578 (pass_reload): Likewise.
25579 * ira.c (do_reload): Merge into ...
25581 (rest_of_handle_reload): Remove.
25582 (pass_reload): Likewise.
25583 * config/i386/i386.c (ix86_option_override): Refer to ira instead
25584 of reload for vzeroupper pass placement.
25585 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
25586 and todo_flags_finish of all passes.
25588 2013-04-10 Richard Biener <rguenther@suse.de>
25590 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
25591 first_const_oprnd field, rename first_def_type to first_op_type.
25592 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
25593 (vect_get_and_check_slp_defs): Always use the type of the
25594 operand. Allow mixed vect_external_def, vect_constant_def types.
25595 (vect_get_constant_vectors): Handle mixed vect_external_def,
25596 vect_constant_def types.
25598 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
25600 PR tree-optimization/55524
25601 * tree-ssa-math-opts.c
25602 (convert_mult_to_fma): Don't use an fms construct
25603 when we don't have an fms operation, but fnma, and it looks
25604 likely that we'll be able to use the latter.
25606 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
25608 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
25610 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
25611 inline fail caused by overwritable functions.
25613 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
25615 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
25616 unnecessary bits in the constant power of two case.
25618 2013-04-10 Richard Biener <rguenther@suse.de>
25620 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
25621 broken code swapping operands.
25622 (vect_build_slp_tree): Do not compute load permutations here.
25623 (vect_analyze_slp_instance): Compute load permutations here,
25624 after building the SLP tree.
25626 2013-04-09 Christian Bruel <christian.bruel@st.com>
25628 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
25629 of next/prev_real_insn.
25631 2013-04-09 Jan Hubicka <jh@suse.cz>
25633 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
25634 Drop aliased parameter.
25635 (function_and_variable_visibility): Do not handle alias pairs.
25636 * cgraph.c (varpool_externally_visible_p): Update prototype.
25637 * varpool.c (varpool_add_new_variable): Update.
25639 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
25641 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
25643 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
25645 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
25647 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
25649 2013-04-09 Marek Polacek <polacek@redhat.com>
25651 PR tree-optimization/48762
25652 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
25654 2013-04-09 Richard Biener <rguenther@suse.de>
25656 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
25658 (vect_build_slp_tree): Likewise.
25659 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
25660 calculating the cost of a SLP instance.
25661 (vect_analyze_slp_instance): Use it from here, after building
25664 2013-04-09 Jakub Jelinek <jakub@redhat.com>
25666 PR middle-end/56883
25667 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
25668 expand_omp_for_static_chunk): Use simple_p = true in
25669 force_gimple_operand_gsi calls when assigning to addressable decls.
25671 2013-04-09 Jeff Law <law@redhat.com>
25673 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
25674 when the boolean was created by converting a wider object which
25675 had a boolean range.
25677 2013-04-09 Richard Biener <rguenther@suse.de>
25679 * tree-vectorizer.h (slp_void_p): Remove.
25680 (slp_tree): Typedef before _slp_tree declaration.
25681 (struct _slp_tree): Use a vector of slp_tree as children.
25682 (vect_get_place_in_interleaving_chain): Remove.
25683 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
25685 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
25687 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
25688 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
25689 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
25690 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
25691 Use slp_node instead of slp_void_p and adjust.
25693 2013-04-09 Richard Biener <rguenther@suse.de>
25695 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
25696 work that is not necessary.
25698 2013-04-09 Jakub Jelinek <jakub@redhat.com>
25700 PR tree-optimization/56854
25701 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
25702 forward into clobber stmts if it would change MEM_REF lhs into
25705 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
25707 * tree.c (type_hash_lookup, type_hash_add): Make static.
25708 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
25710 2013-04-09 Richard Biener <rguenther@suse.de>
25712 * tree.h (unsave_expr_now): Remove.
25713 * tree-inline.c (mark_local_for_remap_r): Remove.
25714 (unsave_expr_1): Likewise.
25715 (unsave_r): Likewise.
25716 (unsave_expr_now): Likewise.
25717 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
25718 (propagate_tree_value): Likewise.
25720 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
25722 * doc/rtl.texi (sequence): Rewrite documentation to match the
25723 current use of SEQUENCE rtl objects.
25724 * rtl.def (SEQUENCE): Likewise.
25726 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
25727 Update documentation.
25728 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
25729 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
25731 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
25733 2013-04-08 Teresa Johnson <tejohnson@google.com>
25735 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
25736 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
25738 (estimate_edge_size_and_time): Add comment to suggest using rounding
25740 (estimate_node_size_and_time): Ditto.
25741 (remap_edge_change_prob): Use helper rounding divide methods.
25742 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
25743 (gimple_mod_pow2_value_transform): Ditto.
25744 (gimple_mod_subtract_transform): Ditto.
25745 (gimple_ic_transform): Ditto.
25746 (gimple_stringops_transform): Ditto.
25747 * stmt.c (conditional_probability): Ditto.
25748 (emit_case_dispatch_table): Ditto.
25749 * lto-cgraph.c (merge_profile_summaries): Ditto.
25750 * tree-optimize.c (execute_fixup_cfg): Ditto.
25751 * cfgcleanup.c (try_forward_edges): Ditto.
25752 * cfgloopmanip.c (scale_loop_profile): Ditto.
25754 (duplicate_loop_to_header_edge): Ditto.
25755 (lv_adjust_loop_entry_edge): Ditto.
25756 * tree-vect-loop.c (vect_transform_loop): Ditto.
25757 * profile.c (compute_branch_probabilities): Ditto.
25758 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
25759 * lto-streamer-in.c (input_cfg): Ditto.
25760 * gimple-streamer-in.c (input_bb): Ditto.
25761 * ipa-cp.c (update_profiling_info): Ditto.
25762 (update_specialized_profile): Ditto.
25763 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
25764 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
25766 * sched-rgn.c (compute_dom_prob_ps): Ditto.
25767 (compute_trg_info): Ditto.
25768 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
25769 (purge_dead_edges): Ditto.
25770 * loop-unswitch.c (unswitch_loop): Ditto.
25771 * cgraphclones.c (cgraph_clone_edge): Ditto.
25772 (cgraph_clone_node): Ditto.
25773 * tree-inline.c (copy_bb): Ditto.
25774 (copy_edges_for_bb): Ditto.
25775 (initialize_cfun): Ditto.
25776 (copy_cfg_body): Ditto.
25777 (expand_call_inline): Ditto.
25779 2013-04-08 Kai Tietz <ktietz@redhat.com>
25781 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
25782 TARGET_CYGWIN64 by TARGET_64BIT.
25784 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
25786 * config/epiphany/epiphany.md (GPR_1): New constant.
25787 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
25788 * config/epiphany/epiphany.c (gen_compare_reg):
25789 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
25790 is already in place.
25791 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
25792 Don't require being called during rtl expansion; If y operlaps r0,
25794 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
25795 (epiphany_expand_epilogue): Likewise.
25797 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
25798 Don't use CC_FPmode for ORDERED / UNORDERED.
25799 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
25801 * config/epiphany/constraints.md (CnL): New constraint.
25802 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
25803 * config/epiphany/predicates.md (add_operand): Allow 1024.
25805 * config/epiphany/epiphany.md (logical_op): New code iterator.
25806 (op_mnc): New code attribute.
25807 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
25808 (mov_f+1, mov_f+2): New peephole2 patterns.
25810 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
25811 (cstoresi4): Also allow re-use of zero result when doing a NE
25812 comparison to a non-zero operand.
25813 Use (clobber (scratch)) for first insn if the gpr output is not needed.
25815 * config/epiphany/epiphany.md (<insn_opname>v2si3):
25816 Use gen_addsi3_i / gen_subsi3_i.
25818 2013-04-08 Jakub Jelinek <jakub@redhat.com>
25822 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
25823 contain anything but clobbers, at most one __builtin_stack_restore,
25824 optionally debug stmts and final resx, and if it has at least one
25825 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
25826 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
25827 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
25828 which isn't defaut definition, remove them.
25829 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
25830 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
25831 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
25832 with MEM_REF LHS with SSA_NAME address.
25834 2013-04-08 Jeff Law <law@redhat.com>
25836 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
25838 2013-04-08 Richard Biener <rguenther@suse.de>
25840 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
25842 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
25843 determined vector type.
25844 (vect_analyze_data_refs): Likewise.
25845 (vect_get_new_vect_var): Adjust.
25846 (vect_create_destination_var): Preserve SSA name versions.
25847 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
25848 not dump anything here.
25850 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
25852 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
25853 Add member lr_slot_known.
25854 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
25856 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
25857 Remove code that sets lr_slot_offset according to what a previous
25858 version of epiphany_emit_save_restore used to do.
25859 (epiphany_emit_save_restore): When doing an lr save or restore,
25860 set/verify lr_slot_known and lr_slot_offset.
25862 2013-04-08 Xinyu Qi <xyqi@marvell.com>
25865 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
25868 2013-04-08 Richard Biener <rguenther@suse.de>
25870 * alias.c (find_base_term): Fix thinko in previous change.
25872 2013-04-08 Jakub Jelinek <jakub@redhat.com>
25874 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
25875 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
25876 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
25877 if possible to compute val.
25878 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
25879 For QImode integers don't require anything about precision. Use
25880 const_with_all_bytes_same to find out if the constant doesn't have
25881 repeated bytes in it.
25883 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
25885 * config/s390/s390.c (s390_expand_insv): Only accept insertions
25888 2013-04-08 Marek Polacek <polacek@redhat.com>
25890 PR rtl-optimization/48182
25891 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
25894 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25897 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
25898 nuses, make sure we have a label.
25900 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
25903 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
25904 (rs6000_emit_swdiv_low_precision): Remove.
25905 (rs6000_emit_swdiv): Rewrite to handle between one and four
25906 iterations of Newton-Raphson generally; modify required number of
25907 iterations for some cases.
25908 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
25910 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
25912 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
25913 set-but-unused variable.
25915 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
25916 basic blocks of released function bodies garbage-collectable.
25918 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
25919 (struct rtl_opt_pass): Add TODO_df_finish.
25921 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
25923 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
25925 * config/arm/constraints.md (q): New constraint.
25926 * config/arm/ldrdstrd.md: New file.
25927 * config/arm/arm.md (ldrdstrd.md) New include.
25928 (arm_movdi): Use "q" instead of "r" constraint
25929 for double-word memory access.
25930 (movdf_soft_insn): Likewise.
25931 * config/arm/vfp.md (movdi_vfp): Likewise.
25932 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
25933 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
25934 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
25935 (mem_ok_for_ldrd_strd): Likewise.
25936 (output_move_double): Update assertion.
25938 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
25940 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
25942 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
25944 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
25945 define_insn_and_split.
25946 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
25948 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
25950 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
25951 define_insn_and_split.
25952 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
25953 (shiftsi3_compare): New pattern.
25954 (rrx): New pattern.
25955 * config/arm/unspecs.md (UNSPEC_RRX): New.
25957 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
25959 * config/arm/arm.md (negdi_extendsidi): New pattern.
25960 (negdi_zero_extendsidi): Likewise.
25962 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
25964 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
25965 define_insn_and_split.
25966 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
25967 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
25969 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
25971 * config/arm/arm.md (arm_subdi3): Convert define_insn into
25972 define_insn_and_split.
25973 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
25974 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
25976 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
25978 * config/arm/arm.md (subsi3_carryin): New pattern.
25979 (subsi3_carryin_const): Likewise.
25980 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
25981 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
25983 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
25985 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
25987 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
25989 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
25990 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
25992 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
25994 * config/arm/arm.c (arm_expand_builtin): Change fcode
25995 type to unsigned int.
25997 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
25999 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
26001 2013-04-04 Ian Lance Taylor <iant@google.com>
26003 * doc/standards.texi (Standards): The Go frontend supports the Go 1
26006 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
26008 PR middle-end/56729
26009 * df-scan.c (df_insn_delete): Disable failing assert.
26011 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
26013 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
26014 New function prototype.
26015 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
26016 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
26017 (arm_builtin_vectorized_function): New function.
26019 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
26021 * config/arm/arm_neon_builtins.def: New file.
26022 * config/arm/arm.c (neon_builtin_data): Move contents to
26023 arm_neon_builtins.def.
26024 (enum arm_builtins): Include neon builtin definitions.
26025 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
26026 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
26028 2013-04-04 Marek Polacek <polacek@redhat.com>
26030 PR tree-optimization/48186
26031 * predict.c (maybe_hot_frequency_p): Return false if
26032 HOT_BB_FREQUENCY_FRACTION is 0.
26033 (cgraph_maybe_hot_edge_p): Likewise.
26035 2013-04-04 Richard Biener <rguenther@suse.de>
26037 PR tree-optimization/56826
26038 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
26041 2013-04-04 Richard Biener <rguenther@suse.de>
26043 PR tree-optimization/56213
26044 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
26045 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
26047 2013-04-04 Richard Biener <rguenther@suse.de>
26049 PR tree-optimization/56837
26050 * tree-loop-distribution.c (classify_partition): For non-zero
26051 values require that the value has the same precision as its
26052 mode to be useful as memset value.
26054 2013-04-03 Nick Clifton <nickc@redhat.com>
26056 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
26057 (fmssf4): Use fmsf.s on E3V5 architectures.
26058 (fnmasf4): Use fnmaf.s on E3V5 architectures.
26059 (fnmssf4): Use fnmsf.s on E3V5 architectures.
26061 2013-04-03 Jeff Law <law@redhat.com>
26063 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
26064 (lra-eliminations.o): Likewise.
26066 2013-04-03 Teresa Johnson <tejohnson@google.com>
26068 * gcov-io.c (compute_working_sets): Moved most of body of old
26069 compute_working_sets here from profile.c.
26070 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
26071 (gcov_working_set_t): Moved typedef here from basic-block.h
26072 (compute_working_set): Declare.
26073 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
26074 (get_working_sets): Renamed from compute_working_set,
26075 replace most of body with call to new compute_working_sets.
26076 (get_exec_counts): Replace call to compute_working_sets
26077 to get_working_sets.
26078 * profile.h (get_working_sets): Renamed from compute_working_set.
26079 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
26080 to get_working_sets.
26081 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
26082 * gcov-dump.c (dump_working_sets): New function.
26084 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
26086 * hwint.c (sext_hwi, zext_hwi): New functions.
26087 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
26088 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
26089 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
26090 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
26091 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
26092 (sext_hwi, zext_hwi): New functions.
26094 2013-04-03 Jeff Law <law@redhat.com>
26096 PR tree-optimization/56799
26097 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
26098 back test for widening conversion erroneously dropped in prior change.
26100 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
26103 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
26104 instead of next_real_insn.
26106 2013-04-03 Marek Polacek <polacek@redhat.com>
26109 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
26111 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
26114 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
26116 (thumb1_output_casesi): Likewise.
26117 (thumb2_output_casesi): Likewise.
26119 2013-04-03 Richard Biener <rguenther@suse.de>
26121 PR tree-optimization/56817
26122 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
26124 (tree_unroll_loops_completely_1): ... new function to manually
26125 walk the loop tree, properly defering outer loops of unrolled
26126 loops to later iterations.
26128 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
26130 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
26131 (vectorizable_load): Likewise.
26132 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
26133 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
26135 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
26137 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
26140 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
26142 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
26144 2013-04-03 Bin Cheng <bin.cheng@arm.com>
26146 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
26148 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
26150 PR tree-optimization/56790
26151 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
26154 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
26156 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
26158 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
26159 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
26162 2013-04-03 Jakub Jelinek <jakub@redhat.com>
26165 * tree.h (force_folding_builtin_constant_p): New decl.
26166 * builtins.c (force_folding_builtin_constant_p): New variable.
26167 (fold_builtin_constant_p): Fold immediately also if
26168 force_folding_builtin_constant_p.
26170 2013-04-03 Richard Biener <rguenther@suse.de>
26172 PR tree-optimization/56812
26173 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
26174 DRs of the same interleaving chain are independent.
26176 2013-04-02 Jason Merrill <jason@redhat.com>
26178 * gdbinit.in (pbb): Use debug fn.
26180 2013-04-02 Lawrence Crowl <crowl@google.com>
26182 * sese.h (struct ivtype_map_elt_s): Remove unused.
26183 (extern debug_ivtype_map): Remove unused.
26184 (extern eq_ivtype_map_elts): Remove unused.
26185 * sese.c (debug_ivtype_map): Removed unused.
26186 (debug_ivtype_map_1): Removed unused.
26187 (debug_ivtype_elt): Remove unused.
26188 (eq_ivtype_map_elts): Remove unused.
26190 2013-04-02 Kai Tietz <ktietz@redhat.com>
26193 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
26194 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
26195 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
26197 (legitimize_pe_coff_symbol): Likewise.
26198 (is_imported_p): New helper-function.
26199 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
26200 for Windows x64 targets.
26201 (ix86_expand_prologue): Optimize for pe-coff targets.
26202 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
26203 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
26204 medium/large code-model.
26205 (legitimize_pic_address): Likewise.
26206 (legitimize_tls_address): Likewise.
26207 (ix86_expand_call): Likewise.
26208 (x86_output_mi_thunk): Likewise.
26209 (get_dllimport_decl): Add new beimport argument.
26210 (construct_plt_address): Don't assert for x64 pe-coff targets.
26211 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
26213 (SYMBOL_FLAG_STUBVAR): New macro.
26214 (SYMBOL_REF_STUBVAR_P): Likewise.
26215 * config/i386/winnt.c (stub_list): New structure.
26216 (stub_head): New local variable.
26217 (i386_pe_record_stub): New function.
26218 (i386_pe_file_end): Emit refptr-stubs.
26220 2013-04-02 Jakub Jelinek <jakub@redhat.com>
26222 PR rtl-optimization/56745
26223 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
26224 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
26227 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
26228 and both of them are MEM_REFs, just compare first argument for
26229 equality and attempt to deal even with differing offsets.
26232 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
26233 of gimple_clobber_p to be MEM_REF.
26234 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
26235 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
26236 after gimplification.
26237 * asan.c (get_mem_ref_of_assignment): Don't instrument
26238 gimple_clobber_p stmts.
26239 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
26240 gimple_clobber_p stmt if they have MEM_REF lhs and
26241 are dead because of another gimple_clobber_p stmt.
26242 * tree-ssa-live.c (clear_unused_block_pointer): Treat
26243 gimple_clobber_p stmts like debug stmts.
26244 (remove_unused_locals): Remove clobbers with MEM_REF lhs
26245 that refer to unused VAR_DECLs or uninitialized values.
26246 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
26247 gimple_clobber_p stmts if they refer to removed parameters.
26248 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
26251 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
26253 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
26254 using SWI48 mode attribute.
26256 2013-04-02 Wei Mi <wmi@google.com>
26258 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
26259 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
26260 *<rotate_insn><mode>3_mask in i386.md.
26262 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
26264 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
26266 2013-04-02 Richard Biener <rguenther@suse.de>
26268 PR tree-optimization/56778
26269 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
26270 Runtime alias tests are not supported for gather loads.
26271 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
26272 stmts referenced from SSA operands before updating SSA form.
26274 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
26275 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
26277 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
26278 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
26279 * config/arm/cortex-a53.md: New file.
26280 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
26281 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
26282 * config/arm/arm.c (arm_issue_rate): Likewise.
26283 * config/arm/arm-tune.md: Regenerate
26284 * config/arm/arm-tables.opt: Regenerate.
26285 * config/arm/arm-cores.def: Add cortex-a53.
26287 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
26289 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
26292 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
26294 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
26295 scalar load/store operations using B/H registers.
26296 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
26298 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
26300 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
26302 * config/aarch64/aarch64.c
26303 (aarch64_simd_scalar_immediate_valid_for_move): New.
26304 * config/aarch64/aarch64-protos.h
26305 (aarch64_simd_scalar_immediate_valid_for_move): New.
26306 * config/aarch64/constraints.md (Dh, Dq): New.
26307 * config/aarch64/iterators.md (hq): New.
26309 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
26311 * reorg.c (get_branch_condition): Deal with conditional returns.
26312 (fill_simple_delay_slots): Remove dead code dealing with jumps.
26314 2013-04-01 Wei Mi <wmi@google.com>
26316 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
26317 Truncate operand 2 using %b asm operand modifier.
26318 (*<shift_insn><mode>3_mask): Ditto.
26319 (*<rotate_insn><mode>3_mask): Ditto.
26321 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
26323 PR middle-end/56798
26324 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
26326 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
26328 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
26330 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
26332 2013-03-30 Lawrence Crowl <crowl@google.com>
26334 * dse.c (clear_alias_sets): Remove never set.
26335 (disqualified_clear_alias_sets): Remove never set.
26336 (clear_alias_mode_pool): Remove never set.
26337 (dse_step0): Remove condition that is never true.
26338 (canon_address): Remove condition that is never true.
26339 (dse_step7): Remove condition that is never true.
26340 (rest_of_handle_dse): Remove condition that is never true.
26341 (rest_of_handle_dse::did_global): Remove never read from above.
26342 (dse_step2_spill): Remove never called from above.
26343 (dse_step5_spill): Remove never called from above.
26345 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
26347 * doc/md.texi (Standard Names) <casesi>: Update documentation for
26348 JUMP_TABLE_DATA changes.
26349 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
26350 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
26351 (Insns) <jump_table_data>: New entry.
26352 * doc/tm.texi: Regenerate.
26354 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
26356 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
26357 for table jump at the end of a basic block using tablejump_p.
26358 * targhooks.c (default_invalid_within_doloop): Likewise.
26359 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
26360 target hook implementation that is identical to the default hook.
26361 (rs6000_invalid_within_doloop): Remove.
26363 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
26364 unused variable from tablejump_p call.
26366 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
26367 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
26368 (INSN_DELETED_P): Likewise.
26369 (emit_jump_table_data): New prototype.
26370 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
26371 after 4th as unused.
26372 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
26373 * sched-vis.c (print_insn): Likewise.
26374 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
26375 insn for compatibility with back ends that use next_active_insn to
26376 identify jump table data.
26377 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
26378 (remove_insn): Likewise.
26379 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
26381 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
26382 (emit_jump_table_data): New function.
26384 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
26385 basic block, a JUMP_TABLE_DATA never is.
26386 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
26387 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
26388 off from code handling real insns.
26389 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
26390 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
26391 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
26392 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
26393 is not a NONDEBUG_INSN_P.
26394 * ira-costs.c (scan_one_insn): Likewise.
26395 * jump.c (mark_all_labels): Likewise.
26396 (mark_jump_label_1): Likewise.
26397 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
26398 * lra.c (get_insn_freq): Expect all insns reaching here to be in
26400 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
26401 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
26402 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
26403 JUMP_TABLE_DATA_P insns.
26404 (calculate_elim_costs_all_insns): Likewise.
26405 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
26406 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
26407 (delete_output_reload): Code style fixups.
26408 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
26409 insn flags on this non-insn.
26410 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
26411 as scheduling barriers, for pre-change compatibility.
26412 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
26413 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
26415 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
26416 redundant JUMP_TABLE_DATA_P test.
26417 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
26418 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
26419 (frv_for_each_packet): Likewise.
26420 * config/i386/i386.c (min_insn_size): Likewise.
26421 (ix86_avoid_jump_mispredicts): Likewise.
26422 * config/m32r/m32r.c (m32r_is_insn): Likewise.
26423 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
26424 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
26425 (mips16_insn_length): Robustify.
26426 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
26427 (mips16_split_long_branches): Likewise.
26428 * config/pa/pa.c (pa_combine_instructions): Likewise.
26429 * config/rs6000/rs6000.c (get_next_active_insn): Treat
26430 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
26431 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
26432 as contributing to pool range lengths.
26433 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
26434 Remove redundant JUMP_TABLE_DATA_P test.
26435 (sh_loop_align): Likewise.
26436 (split_branches): Likewise.
26437 (sh_insn_length_adjustment): Likewise.
26438 * config/spu/spu.c (get_branch_target): Likewise.
26440 2013-03-29 Jan Hubicka <jh@suse.cz>
26442 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
26443 gcov streaming; stream hot bb threshold to ltrans.
26444 * predict.c (get_hot_bb_threshold): Break out from ....
26445 (maybe_hot_count_p): ... here.
26446 (set_hot_bb_threshold): New function.
26447 * lto-section-in.c (lto_section_name): Add profile.
26448 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
26449 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
26450 and data-streamer.h
26451 (histogram_entry): New structure.
26452 (histogram, histogram_pool): New global vars.
26453 (histogram_hash): New structure.
26454 (histogram_hash::hash): New method.
26455 (histogram_hash::equal): Likewise.
26456 (account_time_size): New function.
26457 (cmp_counts): New function.
26458 (dump_histogram): New function.
26459 (ipa_profile_generate_summary): New function.
26460 (ipa_profile_write_summary): New function.
26461 (ipa_profile_read_summary): New function.
26462 (ipa_profile): Decide on threshold.
26463 (pass_ipa_profile): Add ipa_profile_write_summary and
26464 ipa_profile_read_summary.
26465 * Makefile.in (ipa.o): Update dependencies.
26466 * lto-streamer.h (LTO_section_ipa_profile): New section.
26468 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
26470 * tree.h (VAR_P): New.
26472 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
26475 * doc/invoke.texi ([-fwhole-program]): Fix typo.
26477 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
26479 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
26480 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
26481 (control_flow_insn_p): Likewise.
26482 * cfgrtl.c (duplicate_insn_chain): Likewise.
26483 * final.c (get_attr_length_1): Likewise.
26484 (shorten_branches): Likewise.
26485 (final_scan_insn): Likewise.
26486 * function.c (instantiate_virtual_regs): Likewise.
26487 * gcse.c (insert_insn_end_basic_block): Likewise.
26488 * ira-costs.c (scan_one_insn): Likewise.
26489 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
26490 * lra.c (check_rtl): Likewise.
26491 * reload1.c (elimination_costs_in_insn): Likewise.
26492 * reorg.c (follow_jumps): Likewise.
26494 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
26495 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
26496 (thumb_far_jump_used_p): Likewise.
26497 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
26498 (workaround_speculation): Likewise.
26499 (add_sched_insns_for_speculation): Likewise.
26500 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
26501 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
26502 (frv_for_each_packet): Likewise.
26503 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
26504 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
26505 (final_emit_insn_group_barriers): Likewise.
26506 * config/m32r/m32r.c (m32r_is_insn): Likewise.
26507 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
26508 (mips16_insn_length): Likewise.
26509 * config/pa/pa.c (pa_reorg): Likewise.
26510 (pa_combine_instructions): Likewise.
26511 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
26512 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
26513 (sh_reorg): Likewise.
26514 (split_branches): Likewise.
26515 * config/spu/spu.c (get_branch_target): Likewise.
26517 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
26520 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
26522 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
26523 Fix declaration name.
26525 2013-03-28 Lawrence Crowl <crowl@google.com>
26527 * graphds.h (struct graph.indicies): Remove unused.
26528 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
26529 (SCOP_ORIGINAL_PDDRS): Remove unused.
26530 * sese.h (extern insert_loop_close_phis): Removed unused.
26531 (extern insert_guard_phis): Removed unused.
26532 (extern ivtype_map_elt_info): Removed unused.
26533 (new_ivtype_map_elt): Removed unused.
26534 * sese.c (ivtype_map_elt_info): Removed unused.
26536 2013-03-28 Lawrence Crowl <crowl@google.com>
26538 * Makefile.in: Add several missing include dependences.
26540 (test-dump.o): New. This object is not added to any executable,
26541 but is present for ad-hoc testing.
26543 (debug (const bitmap_head_def &)): New.
26544 (debug (const bitmap_head_def *)): New.
26546 (extern debug (const bitmap_head_def &)): New.
26547 (extern debug (const bitmap_head_def *)): New.
26549 (debug (edge_def &)): New.
26550 (debug (edge_def *)): New.
26552 (debug (basic_block_def &)): New.
26553 (debug (basic_block_def *)): New.
26555 (dump_node (const_tree, int, FILE *)): Correct source file.
26557 (debug (die_struct &)): New.
26558 (debug (die_struct *)): New.
26560 (extern debug (die_struct &)): New.
26561 (extern debug (die_struct *)): New.
26562 * gimple-pretty-print.c
26563 (debug (gimple_statement_d &)): New.
26564 (debug (gimple_statement_d *)): New.
26565 * gimple-pretty-print.h
26566 (extern debug (gimple_statement_d &)): New.
26567 (extern debug (gimple_statement_d *)): New.
26569 (debug (ira_allocno_copy &)): New.
26570 (debug (ira_allocno_copy *)): New.
26571 (debug (ira_allocno &)): New.
26572 (debug (ira_allocno *)): New.
26574 (extern debug (ira_allocno_copy &)): New.
26575 (extern debug (ira_allocno_copy *)): New.
26576 (extern debug (ira_allocno &)): New.
26577 (extern debug (ira_allocno *)): New.
26579 (debug (live_range &)): New.
26580 (debug (live_range *)): New.
26582 (debug (lra_live_range &)): New.
26583 (debug (lra_live_range *)): New.
26585 (debug (lra_live_range &)): New.
26586 (debug (lra_live_range *)): New.
26588 (debug (omega_pb_d &)): New.
26589 (debug (omega_pb_d *)): New.
26591 (extern debug (omega_pb_d &)): New.
26592 (extern debug (omega_pb_d *)): New.
26594 (debug (const rtx_def &)): New.
26595 (debug (const rtx_def *)): New.
26597 (debug_tree (tree): Move within file.
26598 (debug_raw (const tree_node &)): New.
26599 (debug_raw (const tree_node *)): New.
26600 (dump_tree_via_hooks (const tree_node *, int)): New.
26601 (debug (const tree_node &)): New.
26602 (debug (const tree_node *)): New.
26603 (debug_verbose (const tree_node &)): New.
26604 (debug_verbose (const tree_node *)): New.
26605 (debug_head (const tree_node &)): New.
26606 (debug_head (const tree_node *)): New.
26607 (debug_body (const tree_node &)): New.
26608 (debug_body (const tree_node *)): New.
26609 (debug_vec_tree (tree): Move and reimplement in terms of dump.
26610 (debug (vec<tree, va_gc> &)): New.
26611 (debug (vec<tree, va_gc> *)): New.
26613 (extern debug (const rtx_def &)): New.
26614 (extern debug (const rtx_def *)): New.
26616 (debug_raw (simple_bitmap_def &)): New.
26617 (debug_raw (simple_bitmap_def *)): New.
26618 (debug (simple_bitmap_def &)): New.
26619 (debug (simple_bitmap_def *)): New.
26621 (extern debug (simple_bitmap_def &)): New.
26622 (extern debug (simple_bitmap_def *)): New.
26623 (extern debug_raw (simple_bitmap_def &)): New.
26624 (extern debug_raw (simple_bitmap_def *)): New.
26626 (debug (vinsn_def &)): New.
26627 (debug (vinsn_def *)): New.
26628 (debug_verbose (vinsn_def &)): New.
26629 (debug_verbose (vinsn_def *)): New.
26630 (debug (expr_def &)): New.
26631 (debug (expr_def *)): New.
26632 (debug_verbose (expr_def &)): New.
26633 (debug_verbose (expr_def *)): New.
26634 (debug (vec<rtx> &)): New.
26635 (debug (vec<rtx> *)): New.
26637 (extern debug (vinsn_def &)): New.
26638 (extern debug (vinsn_def *)): New.
26639 (extern debug_verbose (vinsn_def &)): New.
26640 (extern debug_verbose (vinsn_def *)): New.
26641 (extern debug (expr_def &)): New.
26642 (extern debug (expr_def *)): New.
26643 (extern debug_verbose (expr_def &)): New.
26644 (extern debug_verbose (expr_def *)): New.
26645 (extern debug (vec<rtx> &)): New.
26646 (extern debug (vec<rtx> *)): New.
26648 (_list_iter_cond_expr): Make inline instead of static.
26650 (debug (sreal &)): New.
26651 (debug (sreal *)): New.
26653 (extern debug (sreal &)): New.
26654 (extern debug (sreal *)): New.
26656 (extern debug_raw (const tree_node &)): New.
26657 (extern debug_raw (const tree_node *)): New.
26658 (extern debug (const tree_node &)): New.
26659 (extern debug (const tree_node *)): New.
26660 (extern debug_verbose (const tree_node &)): New.
26661 (extern debug_verbose (const tree_node *)): New.
26662 (extern debug_head (const tree_node &)): New.
26663 (extern debug_head (const tree_node *)): New.
26664 (extern debug_body (const tree_node &)): New.
26665 (extern debug_body (const tree_node *)): New.
26666 (extern debug (vec<tree, va_gc> &)): New.
26667 (extern debug (vec<tree, va_gc> *)): New.
26669 (debug (struct loop &)): New.
26670 (debug (struct loop *)): New.
26671 (debug_verbose (struct loop &)): New.
26672 (debug_verbose (struct loop *)): New.
26673 * tree-dump.c: Add header dependence.
26675 (extern debug (struct loop &)): New.
26676 (extern debug (struct loop *)): New.
26677 (extern debug_verbose (struct loop &)): New.
26678 (extern debug_verbose (struct loop *)): New.
26680 (debug (data_reference &)): New.
26681 (debug (data_reference *)): New.
26682 (debug (vec<data_reference_p> &)): New.
26683 (debug (vec<data_reference_p> *)): New.
26684 (debug (vec<ddr_p> &)): New.
26685 (debug (vec<ddr_p> *)): New.
26687 (extern debug (data_reference &)): New.
26688 (extern debug (data_reference *)): New.
26689 (extern debug (vec<data_reference_p> &)): New.
26690 (extern debug (vec<data_reference_p> *)): New.
26691 (extern debug (vec<ddr_p> &)): New.
26692 (extern debug (vec<ddr_p> *)): New.
26694 (debug (pt_solution &)): New.
26695 (debug (pt_solution *)): New.
26697 (extern debug (pt_solution &)): New.
26698 (extern debug (pt_solution *)): New.
26700 (debug (_var_map &)): New.
26701 (debug (_var_map *)): New.
26702 (debug (tree_live_info_d &)): New.
26703 (debug (tree_live_info_d *)): New.
26705 (extern debug (_var_map &)): New.
26706 (extern debug (_var_map *)): New.
26707 (extern debug (tree_live_info_d &)): New.
26708 (extern debug (tree_live_info_d *)): New.
26710 2013-03-28 Jan Hubicka <jh@suse.cz>
26712 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
26714 2013-03-28 Ian Bolton <ian.bolton@arm.com>
26716 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
26717 record only when desired or required.
26719 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
26721 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
26722 *vec_extractv2di_1_rex64. Use x64 isa attribute.
26724 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
26726 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
26727 (*andsi3_compare0_uxtw): New pattern.
26728 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
26729 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
26731 2013-03-28 Jan Hubicka <jh@suse.cz>
26733 * data-streamer-in.c (streamer_read_gcov_count): New function.
26734 * gimple-streamer-out.c: Include value-prof.h.
26735 (output_gimple_stmt): Output histogram.
26736 (output_bb): Use streamer_write_gcov_count.
26737 * value-prof.c: Include data-streamer.h
26738 (dump_histogram_value): Add HIST_TYPE_MAX.
26739 (stream_out_histogram_value): New function.
26740 (stream_in_histogram_value): New function.
26741 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
26742 (stream_out_histogram_value, stream_in_histogram_value): Declare.
26743 * data-streamer-out.c (streamer_write_gcov_count): New function.
26744 (streamer_write_gcov_count_stream): New function.
26745 * lto-cgraph.c (lto_output_edge): Update counter streaming.
26746 (lto_output_node): Likewise.
26747 (input_node, input_edge): Likewise.
26748 * lto-streamer-out.c (output_cfg): Update streaming.
26749 * lto-streamer-in.c (input_cfg): Likewise.
26750 * data-streamer.h (streamer_write_gcov_count,
26751 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
26752 * gimple-streamer-in.c: Include value-prof.h
26753 (input_gimple_stmt): Input histograms.
26754 (input_bb): Update profile streaming.
26756 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
26758 * genmodes.c (emit_max_int): New function.
26759 (emit_insn_modes_h): Added call to emit_max_function.
26760 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
26762 * machmode.def: Fixed comment.
26764 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
26766 * combine.c (try_combine): Removed useless assert.
26767 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
26769 2013-03-28 Marek Polacek <polacek@redhat.com>
26770 Richard Biener <rguenther@suse.de>
26772 PR tree-optimization/56695
26773 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
26774 build signed result of a vector comparison.
26775 * tree-cfg.c (verify_gimple_comparison): Check that a result
26776 of a vector comparison has signed type.
26778 2013-03-28 Richard Biener <rguenther@suse.de>
26780 PR tree-optimization/37021
26781 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
26782 do not restrict gaps between groups.
26783 * tree-vect-stmts.c (vectorizable_load): Properly account for
26784 a gap between groups.
26786 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
26788 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
26789 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
26792 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
26794 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
26795 * doc/extend.texi (Named Address Spaces): Ditto.
26796 (Variable Attributes): Ditto.
26798 2013-03-27 Kai Tietz <ktietz@redhat.com>
26800 * config.build: Add support for cygwin x64 target.
26801 * config.gcc: Likewise.
26802 * config.host: Likewise.
26803 * configure.ac: Likewise
26804 * configure: Regenerated.
26806 2013-03-27 Kai Tietz <ktietz@redhat.com>
26808 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
26809 * config/i386/t-cygwin-w64: New file.
26810 * config/i386/cygwin-w64.h: New file.
26811 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
26812 and add support for x64-cygwin target.
26813 (CPP_SPEC): Likewise.
26814 (CXX_WRAP_SPEC_LIST): Undefine before define.
26815 (LIBGCJ_SONAME): Use 15 as version.
26817 2013-03-27 Richard Biener <rguenther@suse.de>
26819 PR tree-optimization/56716
26820 * tree-ssa-structalias.c (perform_var_substitution): Adjust
26821 dumping for ref nodes.
26823 2013-03-27 Martin Jambor <mjambor@suse.cz>
26825 PR tree-optimization/55334
26826 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
26827 restricted pointers to arrays.
26829 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
26831 * Makefile.in (.SUFFIXES): Add .cc.
26832 (.c.o): Apply same recipe for implicit rule .cc.o.
26834 2013-03-27 Richard Biener <rguenther@suse.de>
26836 PR tree-optimization/37021
26837 * tree-vect-data-refs.c (vect_check_strided_load): Allow
26838 REALPART/IMAGPART_EXPRs around the supported refs.
26839 * tree-ssa-structalias.c (find_func_aliases): Assume that
26840 floating-point values are not used to transfer pointers.
26842 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
26844 * target.def (TARGET_HAS_IFUNC_P): New target hook.
26845 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
26846 * doc/tm.texi: Regenerate.
26847 * targhooks.h (default_has_ifunc_p): New.
26848 * targhooks.c (default_has_ifunc_p): Ditto.
26849 * config/linux-protos.h: New file.
26850 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
26851 hook for linux which disables support of indirect functions in android.
26852 * config/linux-android.c: New file.
26853 * config/t-linux-android.c: Ditto.
26854 * config.gcc: Added new object file linux-android.o.
26855 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
26856 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
26857 * varasm.c (do_assemble_alias): Likewise.
26858 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
26859 doesn't support indirect functions.
26860 * configure: Regenerate.
26862 2013-03-27 Bin Cheng <bin.cheng@arm.com>
26865 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
26866 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
26869 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26871 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
26873 2013-03-27 Terry Guo <terry.guo@arm.com>
26875 * config/arm/arm-cores.def: Added core cortex-r7.
26876 * config/arm/arm-tune.md: Regenerated.
26877 * config/arm/arm-tables.opt: Regenerated.
26878 * doc/invoke.texi: Added entry for core cortex-r7.
26880 2013-03-27 Walter Lee <walt@tilera.com>
26882 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
26883 double-decrement of next_scratch_regno.
26885 2013-03-27 Walter Lee <walt@tilera.com>
26887 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
26889 (insn_v1mulus): Ditto.
26890 (insn_v2muls): Ditto.
26892 2013-03-27 Walter Lee <walt@tilera.com>
26894 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
26895 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
26897 2013-03-27 Walter Lee <walt@tilera.com>
26899 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
26900 (*sibcall_value): Ditto.
26902 2013-03-27 Walter Lee <walt@tilera.com>
26904 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
26905 (insn_mnz_v8qi): ... this ...
26906 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
26908 (insn_v<n>mnz): Replaced by ...
26909 (insn_v1mnz): ... this ...
26910 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
26912 (insn_mz_<mode>): Replaced by ...
26913 (insn_mz_v8qi): ... this ...
26914 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
26916 (insn_v<n>mz): Replaced by ...
26917 (insn_v1mz): ... this ...
26918 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
26921 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
26923 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
26925 2013-03-26 Roland McGrath <mcgrathr@google.com>
26927 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
26928 than fprintf with a non-constant, non-format string.
26930 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
26932 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
26933 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
26934 operand 0 predicate.
26935 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
26936 attribute. Use general_x64nomem_operand as operand 1 predicate.
26937 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
26938 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
26939 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
26940 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
26941 (mov<mode>_insv_1): Remove expander. Merge insn with
26942 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
26943 Use general_x64nomem_operand as operand 1 predicate.
26944 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
26945 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
26946 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
26947 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
26948 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
26949 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
26950 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
26951 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
26952 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
26953 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
26954 (general_x64nomem_operand): Ditto.
26956 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
26958 * config/rtems.opt: Add -pthread option.
26960 2013-03-26 Richard Biener <rguenther@suse.de>
26962 * alias.c (find_base_term): Avoid redundant and not used recursion.
26963 (base_alias_check): Get the initial base term from the caller.
26964 (true_dependence_1): Compute and pass base terms to base_alias_check.
26965 (write_dependence_p): Likewise.
26966 (may_alias_p): Likewise.
26968 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
26970 * config/aarch64/aarch64.c (aarch64_classify_address): Support
26971 PC-relative load in SI modes and above only.
26973 2013-03-26 Xinyu Qi <xyqi@marvell.com>
26975 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
26976 * config/arm/iwmmxt.md (WCGR0): Update.
26977 (WCGR1, WCGR2, WCGR3): Likewise.
26979 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
26981 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
26982 Use x64 and nox64 isa attributes.
26984 2013-03-26 Richard Biener <rguenther@suse.de>
26986 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
26987 alignment computations and rely on get_object_alignment_1
26988 for the !TYPE_P case.
26989 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
26991 2013-03-26 Walter Lee <walt@tilera.com>
26993 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
26994 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
26996 2013-03-25 Jeff Law <law@redhat.com>
26998 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
26999 check for INTEGRAL_TYPE_P that was missing due to checking in
27000 wrong version of prior patch.
27002 2013-03-25 Walter Lee <walt@tilera.com>
27004 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
27005 TILEGX_INSN_SHUFFLEBYTES1.
27006 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
27008 (tilegx_builtins): Ditto.
27009 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
27011 2013-03-25 Walter Lee <walt@tilera.com>
27013 * config/tilegx/tilegx.md (floatsisf2): New pattern.
27014 (floatunssisf2): New pattern.
27015 (floatsidf2): New pattern.
27016 (floatunssidf2): New pattern.
27018 2013-03-25 Walter Lee <walt@tilera.com>
27020 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
27021 tests for constraint J, K, N, P.
27023 2013-03-25 Walter Lee <walt@tilera.com>
27025 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
27026 Use indirect/pcrel encoding.
27027 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
27030 2013-03-25 Steve Ellcey <sellcey@mips.com>
27032 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
27033 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
27034 * config/mips/mips.c (mips_option_override): Set IMADD default.
27035 * config/mips/mips.h (PTF_AVOID_IMADD): New.
27036 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
27037 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
27038 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
27039 * doc/invoke.texi (-mimadd/-mno-imadd): New.
27041 2013-03-25 Jeff Law <law@redhat.com>
27043 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
27044 slightly to avoid creating and folding useless trees. Simplify
27045 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
27047 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
27049 * config/i386/i386.md (*zero_extendsidi2): Merge with
27050 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
27051 * config/i386/predicates.md (x86_64_zext_operand): Rename from
27052 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
27053 targets. Clarify comment.
27055 2013-03-25 Martin Jambor <mjambor@suse.cz>
27057 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
27058 pass-through jump functions differently.
27059 (ipa_read_jump_function): Likewise. Also use setter functions to set
27062 2013-03-25 Martin Jambor <mjambor@suse.cz>
27064 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
27065 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
27067 (ipa_get_indirect_edge_target): New function.
27068 (devirtualization_time_bonus): New parameter known_aggs, pass it to
27069 ipa_get_indirect_edge_target. Update all callers.
27070 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
27071 ipa_get_indirect_edge_target_1 instead of calling
27072 ipa_get_indirect_edge_target.
27073 (create_specialized_node): Pass aggvlas to
27074 ipcp_discover_new_direct_edges.
27076 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
27078 * config/arm/arm.md (f_sels, f_seld): New types.
27079 (*cmov<mode>): New pattern.
27080 * config/arm/predicates.md (arm_vsel_comparison_operator): New
27083 2013-03-25 Kai Tietz <ktietz@redhat.com>
27085 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
27086 POSIX-printf for mingw-hosted builds.
27088 2013-03-25 Richard Biener <rguenther@suse.de>
27090 PR middle-end/56694
27091 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
27092 must-not-throw stmt location.
27094 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
27096 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
27097 Emit load-acquire versions when acq is true.
27098 (arm_emit_store_exclusive): Add rel parameter.
27099 Emit store-release versions when rel is true.
27100 (arm_split_compare_and_swap): Use acquire-release instructions
27102 of barriers when appropriate.
27103 (arm_split_atomic_op): Likewise.
27104 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
27105 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
27106 (VUNSPEC_SLX): Likewise.
27107 (VUNSPEC_LDA): Likewise.
27108 (VUNSPEC_STL): Likewise.
27109 * config/arm/sync.md (atomic_load<mode>): New pattern.
27110 (atomic_store<mode>): Likewise.
27111 (arm_load_acquire_exclusive<mode>): Likewise.
27112 (arm_load_acquire_exclusivesi): Likewise.
27113 (arm_load_acquire_exclusivedi): Likewise.
27114 (arm_store_release_exclusive<mode>): Likewise.
27116 2013-03-25 Catherine Moore <clm@codesourcery.com>
27118 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
27119 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
27120 * config/mip/predicates.md (lwsp_swsp_operand,
27121 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
27122 sb16_operand, db4_operand, db7_operand, ib3_operand,
27123 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
27124 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
27125 andi16_operand): New predicates.
27126 * config/mips/mips.md (compression): New attribute.
27127 (enabled): New attribute.
27128 (length): Consider compression in computing length.
27129 (shift_compression): New code attribute.
27130 (*add<mode>3): New operands. Record compression.
27131 (sub<mode>3): Likewise.
27132 (one_cmpl<mode>2): Likewise.
27133 (*and<mode>3): Likewise.
27134 (*ior<mode>3): Likewise.
27135 (unnamed pattern for xor): Likewise.
27136 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
27137 (*<optab><mode>3): Likewise.
27138 (*mov<mode>_internal: Likewise.
27139 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
27140 (mips_unsigned_immediate_p): New.
27141 (umips_lwsp_swsp_address_p): New.
27142 (m16_based_address_p): New.
27143 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
27144 (mips_unsigned_immediate_p): New prototype.
27145 (lwsp_swsp_address_p): New prototype.
27146 (m16_based_address_p): New prototype.
27147 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
27148 (mips_signed_immediate_p): New function.
27149 (m16_based_address_p): New function.
27150 (lwsp_swsp_address_p): New function.
27151 (mips_print_operand_punctuation): Recognize short delay slot insns
27152 for microMIPS.add<mode>3"
27154 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
27157 * config/arm/iterators.md (v_cmp_result): New mode attribute.
27158 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
27160 2013-03-25 Richard Biener <rguenther@suse.de>
27162 PR tree-optimization/56689
27163 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
27166 2013-03-25 Richard Biener <rguenther@suse.de>
27168 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
27170 (memory_references): Likewise.
27171 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
27172 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
27173 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
27174 (gather_mem_refs_in_loops): Fold into ...
27175 (analyze_memory_references): ... this. Move initialization
27176 to tree_ssa_lim_initialize.
27177 (fill_always_executed_in): Rename to ...
27178 (fill_always_executed_in_1): ... this.
27179 (fill_always_executed_in): Move contains_call computation to
27180 this new function from ...
27181 (tree_ssa_lim_initialize): ... here.
27182 (tree_ssa_lim): Call fill_always_executed_in.
27184 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
27186 * postreload.c (reload_combine): Fix code detecting returns.
27188 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
27190 * function.c (emit_use_return_register_into_block): On cc0 targets,
27191 do not emit the sequence between cc0 setter and user.
27193 2013-03-25 Kai Tietz <ktietz@redhat.com>
27195 * config/i386/predicates.md (local_symbolic_operand): Interpret
27196 dll-imported symbols as none-local.
27198 2013-03-25 Richard Biener <rguenther@suse.de>
27200 * tree-ssa-loop-im.c (struct depend): Remove.
27201 (struct lim_aux_data): Make depends a vec of gimples.
27202 (free_lim_aux_data): Adjust.
27203 (add_dependency): Likewise.
27204 (set_level): Likewise.
27206 2013-03-25 Richard Biener <rguenther@suse.de>
27208 PR middle-end/56434
27209 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
27210 the pointer returned by calls with ECF_MALLOC set.
27212 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
27214 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
27216 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
27218 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
27219 using MMXMODE mode iterator.
27220 (*move<mode>_internal): Merge with *movv2sf_internal and
27221 *movv2sf_internal_rex64 using MMXMODE mode iterator.
27223 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
27225 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
27226 (record_last_mem_set_info): Likewise.
27228 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
27229 of XNEWVEC followed by memset.
27230 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
27232 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
27234 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
27235 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
27236 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
27237 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
27238 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
27239 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
27240 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
27241 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
27242 BARRIER_P instead of GET_CODE.
27244 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
27246 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
27247 inaccuracy in the probing code.
27249 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
27250 (ctrapdi4): Likewise.
27252 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
27254 * calls.c (expand_call): Add missing guard to code handling return
27255 of non-BLKmode structures in MSB.
27256 * function.c (expand_function_end): Likewise.
27258 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
27260 * combine.c (try_combine): Adjust comment. Do not add the set of
27261 insn #0 if the destination indirectly is set or dies in insn #2.
27262 Tidy up code to distribute a new note.
27264 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
27266 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
27267 also for alternatives 16 and 17.
27269 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
27271 * config/i386/sse.md (*mov<mode>_internal): Merge with
27272 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
27273 Emit insn template depending on type attribute. Use
27274 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
27275 movd instead of movq mnemonic for interunit moves. Rewrite mode
27276 attribute calculation. Remove unit attribute calculation.
27277 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
27278 Set prefix_data16 attribute for DImode ssemov types.
27279 Use Ym instead of y for SSE-MMX conversion alternatives.
27280 Reorder operand constraints.
27282 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
27284 * df.h (df_insn_delete): Adjust prototype.
27285 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
27286 and let it decide whether mark the basic block dirty.
27287 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
27288 * df-scan.c (df_insn_info_delete): New helper function, split
27289 off from df_insn_delete.
27290 (df_scan_free_bb_info): Use it.
27291 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
27293 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
27294 that the insn is actually an insn and it has a non-NULL basic block.
27295 Do not mark basic block dirty if only deleting a DEBUG_INSN.
27297 2013-03-22 Richard Biener <rguenther@suse.de>
27299 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
27301 (mem_ref_alloc): Do not allocate them.
27302 (refs_independent_p): Do not query or maintain a cache.
27304 2013-03-22 Richard Biener <rguenther@suse.de>
27306 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
27307 (gather_mem_refs_in_loops): Do not compute it.
27308 (analyze_memory_references): Do not allocate it.
27309 (tree_ssa_lim_finalize): Do not free it.
27310 (for_all_locs_in_loop): Do not query all_refs_in_loop.
27312 2013-03-22 Richard Biener <rguenther@suse.de>
27314 * is-a.h (as_a): Use gcc_checking_assert.
27316 2013-03-22 Ian Bolton <ian.bolton@arm.com>
27318 * config/aarch64/aarch64.c (aarch64_print_operand): New
27319 format specifier for printing a constant in hex.
27320 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
27321 format specifier for printing second operand.
27323 2013-03-22 Richard Biener <rguenther@suse.de>
27325 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
27327 (gather_mem_refs_in_loops): Perform store accumulation here.
27328 (create_vop_ref_mapping_loop): Remove.
27329 (create_vop_ref_mapping): Likewise.
27330 (analyze_memory_references): Initialize refs_stored_in_loop.
27331 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
27332 (record_indep_loop): Remove.
27333 (record_dep_loop): New function.
27334 (ref_indep_loop_p_1): Adjust to only walk over references
27335 in the loop, not its subloops.
27336 (ref_indep_loop_p): Rename to ...
27337 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
27338 maintaining a more fine-grained cache.
27339 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
27340 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
27342 2013-03-22 Richard Biener <rguenther@suse.de>
27344 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
27345 (struct mem_ref): Make accesses_in_loop a vec of a vec of
27346 aggregate mem_ref_loc.
27347 (free_mem_ref_locs): Inline into ...
27348 (memref_free): ... this and adjust.
27349 (mem_ref_alloc): Adjust.
27350 (mem_ref_locs_alloc): Remove.
27351 (record_mem_ref_loc): Adjust.
27352 (get_all_locs_in_loop): Rewrite into ...
27353 (for_all_locs_in_loop): ... this iterator.
27354 (rewrite_mem_ref_loc): New functor.
27355 (rewrite_mem_refs): Use for_all_locs_in_loop.
27356 (sm_set_flag_if_changed): New functor.
27357 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
27358 (ref_always_accessed): New functor.
27359 (ref_always_accessed_p): Use for_all_locs_in_loop.
27361 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
27363 * tree-pass.h (PROP_gimple_lvec): New.
27364 * passes.c (dump_properties): Handle PROP_gimple_lvec.
27365 (init_optimization_passes): Move pass_lower_vector.
27366 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
27368 (pass_lower_vector): Provide PROP_gimple_lvec.
27369 (pass_lower_vector_ssa): Likewise.
27370 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
27372 2013-03-21 Mark Wielaard <mjw@redhat.com>
27374 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
27376 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
27378 * config/i386/i386.md (*movdi_internal): Disparage slightly
27379 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
27380 conversion alternatives.
27382 2013-03-21 Jakub Jelinek <jakub@redhat.com>
27384 PR middle-end/48087
27385 * diagnostic.def (DK_WERROR): New kind.
27386 * diagnostic.h (werrorcount): Define.
27387 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
27388 promoted to DK_ERROR, increment DK_WERROR counter instead of
27390 * toplev.c (toplev_main): Call print_ignored_options even if
27391 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
27392 even if just werrorcount is non-zero.
27395 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
27397 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
27398 (string_cst_pool_decl): New function.
27399 (optimize_one_addr_into_implicit_ptr): New function.
27400 (resolve_addr_in_expr): Optimize DWARF location expression
27401 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
27402 which doesn't live in memory, but has DW_AT_location or
27403 DW_AT_const_value, or refers to a string literal, into
27404 DW_OP_GNU_implicit_pointer.
27405 (optimize_location_into_implicit_ptr): New function.
27406 (resolve_addr): If removing DW_AT_location of a variable because
27407 it was DW_OP_addr of address of the variable, but the variable doesn't
27408 live in memory, try to emit const value attribute for the initializer.
27410 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
27412 * tree.h (VECTOR_TYPE_P): New macro.
27413 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
27414 TYPE_MODE): Use it.
27415 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
27416 VEC_COND_EXPR cannot be lvalues.
27417 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
27419 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
27421 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
27422 Restrict the transformation to equal modes.
27424 2013-03-21 Richard Biener <rguenther@suse.de>
27426 PR tree-optimization/39326
27427 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
27428 (MEM_ANALYZABLE): Adjust.
27429 (record_mem_ref_loc): Move bitmap ops ...
27430 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
27431 unanalyzable refs, do not record locations for it.
27432 (analyze_memory_references): Allocate ref zero as shared
27434 (refs_independent_p): Do not test for unanalyzed mems here.
27435 (ref_indep_loop_p_1): Special-case disambiguation against
27436 the unanalyzed ref.
27437 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
27439 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
27441 * config/arm/arm-protos.h (tune_params): Add
27442 prefer_neon_for_64bits field.
27443 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
27444 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
27445 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
27446 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
27447 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
27448 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
27449 (arm_option_override): Handle -mneon-for-64bits new option.
27450 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
27451 (prefer_neon_for_64bits): Declare new variable.
27452 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
27453 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
27454 (arch_enabled): Handle new arch types. Remove support for onlya8
27456 (one_cmpldi2): Use new arch names.
27457 (zero_extend<mode>di2, extend<mode>di2): Ditto.
27458 * config/arm/arm.opt (mneon-for-64bits): Add option.
27459 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
27460 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
27461 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
27463 * doc/invoke.texi (-mneon-for-64bits): Document.
27465 2013-03-21 Richard Biener <rguenther@suse.de>
27467 PR tree-optimization/39326
27468 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
27469 (sort_bbs_in_loop_postorder_cmp): New function.
27470 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
27472 2013-03-21 Richard Biener <rguenther@suse.de>
27474 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
27475 (vect_insert_into_interleaving_chain): Likewise.
27476 (vect_drs_dependent_in_basic_block): Inline ...
27477 (vect_slp_analyze_data_ref_dependence): ... here. New function,
27479 (vect_analyze_data_ref_dependence): ... here. Simplify.
27480 (vect_check_interleaving): Simplify.
27481 (vect_analyze_data_ref_dependences): Likewise. Split out ...
27482 (vect_slp_analyze_data_ref_dependences): ... this new function.
27483 (dr_group_sort_cmp): New function.
27484 (vect_analyze_data_ref_accesses): Compute data-reference groups
27485 here instead of in vect_analyze_data_ref_dependence. Use
27486 a more efficient algorithm.
27487 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
27488 vect_slp_analyze_data_ref_dependences. Call
27489 vect_analyze_data_ref_accesses earlier.
27490 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
27491 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
27492 (vect_slp_analyze_data_ref_dependences): New prototype.
27494 2013-03-21 Richard Biener <rguenther@suse.de>
27496 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
27497 ref is stored in the loop.
27498 (find_refs_for_sm): Walk only over all stores.
27499 (store_motion_loop): Allocate from lim_bitmap_obstack.
27500 (store_motion): Likewise.
27502 2013-03-21 Richard Biener <rguenther@suse.de>
27504 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
27505 Update virtual SSA form.
27507 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
27509 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
27510 * configure: Regenerate.
27511 * config.in: Regenerate.
27512 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
27513 if !HAVE_LD_EH_FRAME_CIEV3.
27515 2013-03-21 Richard Biener <rguenther@suse.de>
27517 * tree-cfg.c (verify_expr_no_block): New function.
27518 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
27519 nor DECL_VALUE_EXPR have locations with associated blocks.
27520 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
27521 (clear_unused_block_pointer): Remove code dealing with
27522 blocks in DECL_DEBUG_EXPR locations.
27524 2013-03-21 Richard Biener <rguenther@suse.de>
27526 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
27527 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
27528 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
27529 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
27530 instead of DECL_DEBUG_EXPR_IS_FROM.
27531 * gimplify.c (gimplify_modify_expr): Likewise.
27532 * tree-cfg.c (verify_expr_location_1): Likewise.
27533 * tree-complex.c (create_one_component_var): Likewise.
27534 * tree-sra.c (create_access_replacement): Likewise.
27535 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
27536 (clear_unused_block_pointer): Likewise.
27537 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
27538 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
27539 * var-tracking.c (var_debug_decl): Likewise.
27540 (track_expr_p): Likewise.
27541 * tree-inline.c (add_local_variables): Likewise. Set
27542 DECL_HAS_DEBUG_EXPR_P after copying it.
27543 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
27544 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
27546 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
27549 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
27550 * configure: Regenerate.
27551 * config.in: Regenerate.
27552 * config/i386/i386.md (*movdf_internal): Use
27553 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
27554 movd instead of movq mnemonic for interunit moves.
27555 (*movdi_internal): Ditto.
27557 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
27559 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
27560 (abd<mode>_3): New pattern.
27561 (aba<mode>_3): New pattern.
27562 (fabd<mode>_3): New pattern.
27564 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
27566 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
27567 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
27568 occurrence of REGISTER_PREFIX as its empty string.
27570 2013-03-20 Jeff Law <law@redhat.com>
27572 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
27573 addititional equivalences for equality comparisons between an SSA_NAME
27574 and a constant where the SSA_NAME was set from a widening conversion.
27576 2013-03-20 Walter Lee <walt@tilera.com>
27578 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
27580 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
27582 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
27583 depending on type attribute.
27584 (*movti_internal): Ditto.
27585 (*movtf_internal): Ditto.
27586 (*movxf_internal): Ditto.
27587 (*movdf_internal): Ditto.
27588 (*movsf_internal): Ditto.
27590 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
27592 * config/i386/i386.md (*movti_internal): Set prefix attribute to
27593 maybe_vex for sselog1 and ssemov types.
27594 (*movdi_internal): Reorder operand constraints.
27595 (*movsi_internal): Ditto. Set prefix attribute to
27596 maybe_vex for sselog1 and ssemov types.
27597 (*movtf_internal): Set prefix attribute to maybe_vex
27598 for sselog1 and ssemov types.
27599 (*movdf_internal): Ditto. Set prefix_data16 attribute for
27600 DImode ssemov types. Reorder operand constraints.
27601 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
27602 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
27603 attribute for SImode ssemov types. Reorder operand constraints.
27605 2013-03-20 Martin Jambor <mjambor@suse.cz>
27607 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
27608 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
27610 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
27612 * config/rs6000/predicates.md (indexed_address, update_address_mem
27613 update_indexed_address_mem): New predicates.
27614 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
27615 attribute for load/store instructions.
27616 * config/rs6000/dfp.md (movsd_store): Likewise.
27617 (movsd_load): Likewise.
27618 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
27619 (unnamed HI->DI extend define_insn): Likewise.
27620 (unnamed SI->DI extend define_insn): Likewise.
27621 (unnamed QI->SI extend define_insn): Likewise.
27622 (unnamed QI->HI extend define_insn): Likewise.
27623 (unnamed HI->SI extend define_insn): Likewise.
27624 (unnamed HI->SI extend define_insn): Likewise.
27625 (extendsfdf2_fpr): Likewise.
27626 (movsi_internal1): Likewise.
27627 (movsi_internal1_single): Likewise.
27628 (movhi_internal): Likewise.
27629 (movqi_internal): Likewise.
27630 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
27631 attribute for load/store instructions.
27632 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
27634 (mov<mode>_softfloat): Likewise.
27635 (mov<mode>_hardfloat32): Likewise.
27636 (mov<mode>_hardfloat64): Likewise.
27637 (mov<mode>_softfloat64): Likewise.
27638 (movdi_internal32): Likewise.
27639 (movdi_internal64): Likewise.
27640 (probe_stack_<mode>): Likewise.
27642 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
27644 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
27645 floating point, and decimal floating point to reload iterator.
27647 * config/rs6000/constraints.md (wl constraint): New constraints to
27648 return FLOAT_REGS if certain options are used to reduce the number
27649 of separate patterns that exist in the file.
27650 (wx constraint): Likewise.
27651 (wz constraint): Likewise.
27653 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
27654 -mdebug=reg, print wg, wl, wx, and wz constraints.
27655 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
27656 Initialize the reload functions for 64-bit binary/decimal floating
27658 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
27659 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
27660 create the buffer on the stack to overcome not having a 32-bit
27662 (rs6000_emit_move): Likewise.
27663 (rs6000_secondary_memory_needed_rtx): Likewise.
27664 (rs6000_alloc_sdmode_stack_slot): Likewise.
27665 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
27666 via xxlxor, just like DFmode 0.0.
27668 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
27669 define as 1 if we are running on a power7 or newer.
27670 (enum r6000_reg_class_enum): Add new constraints.
27672 * config/rs6000/dfp.md (movsd): Delete, combine with binary
27673 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
27674 with other moves by using conditional constraits (wg). Use LFIWZX
27675 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
27676 (movsd splitter): Likewise.
27677 (movsd_hardfloat): Likewise.
27678 (movsd_softfloat): Likewise.
27680 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
27681 binary and decimal floating point moves.
27682 (fmove_ok): New attributes to combine binary and decimal floating
27683 point moves, and to combine power6x (mfpgpr) moves along normal
27685 (real_value_to_target): Likewise.
27686 (f32_lr): Likewise.
27687 (f32_lm): Likewise.
27688 (f32_li): Likewise.
27689 (f32_sr): Likewise.
27690 (f32_sm): Likewise.
27691 (f32_si): Likewise.
27692 (movsf): Combine binary and decimal floating point moves. Combine
27693 power6x (mfpgpr) moves with other moves by using conditional
27694 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
27695 (mov<mode> for SFmode/SDmode); Likewise.
27696 (SFmode/SDmode splitters): Likewise.
27697 (movsf_hardfloat): Likewise.
27698 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
27699 (movsf_softfloat): Likewise.
27700 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
27702 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
27703 wx and wz constraints.
27705 * config/rs6000/constraints.md (wg constraint): New constraint to
27706 return FLOAT_REGS if -mmfpgpr (power6x) was used.
27708 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
27711 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
27712 -mdebug=reg, print wg, wl, wx, and wz constraints.
27713 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
27714 Initialize the reload functions for 64-bit binary/decimal floating
27716 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
27717 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
27718 create the buffer on the stack to overcome not having a 32-bit
27720 (rs6000_emit_move): Likewise.
27721 (rs6000_secondary_memory_needed_rtx): Likewise.
27722 (rs6000_alloc_sdmode_stack_slot): Likewise.
27723 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
27724 via xxlxor, just like DFmode 0.0.
27726 * config/rs6000/dfp.md (movdd): Delete, combine with binary
27727 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
27728 with other moves by using conditional constraits (wg). Use LFIWZX
27729 and STFIWX for loading SDmode on power7.
27730 (movdd splitters): Likewise.
27731 (movdd_hardfloat32): Likewise.
27732 (movdd_softfloat32): Likewise.
27733 (movdd_hardfloat64_mfpgpr): Likewise.
27734 (movdd_hardfloat64): Likewise.
27735 (movdd_softfloat64): Likewise.
27737 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
27738 64-bit binary and decimal floating point moves.
27739 (FMOVE64X): Likewise.
27740 (movdf): Combine 64-bit binary and decimal floating point moves.
27741 Combine power6x (mfpgpr) moves with other moves by using
27742 conditional constraits (wg).
27743 (mov<mode> for DFmode/DDmode): Likewise.
27744 (DFmode/DDmode splitters): Likewise.
27745 (movdf_hardfloat32): Likewise.
27746 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
27747 (movdf_softfloat32): Likewise.
27748 (movdf_hardfloat64_mfpgpr): Likewise.
27749 (movdf_hardfloat64): Likewise.
27750 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
27751 (movdf_softfloat64): Likewise.
27752 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
27753 (reload_<mode>_load): Move to later in the file so they aren't in
27754 the middle of the floating point move insns.
27755 (reload_<mode>_store): Likewise.
27757 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
27760 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
27761 constraint if -mdebug=reg.
27762 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
27763 Enable using dd reload support if needed.
27765 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
27766 binary and decimal floating point moves in rs6000.md.
27767 (movtd_internal): Likewise.
27769 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
27770 decimal floating point moves.
27772 (movtf_internal): Likewise.
27773 (mov<mode>_internal, TDmode/TFmode): Likewise.
27774 (movtf_softfloat): Likewise.
27775 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
27777 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
27778 movdi_internal64, using wg constraint for move direct operations.
27779 (movdi_internal64): Likewise.
27781 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
27782 MODES_TIEABLE_P for selected modes. Print the numerical value of
27783 the various virtual registers. Use GPR/FPR first/last values,
27784 instead of hard coding the register numbers. Print which modes
27785 have reload functions registered.
27786 (rs6000_option_override_internal): If -mdebug=reg, trace the options
27787 settings before/after setting cpu, target and subtarget settings.
27788 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
27789 and for secondary reload failures in rs6000_secondary_reload_inner.
27790 (rs6000_secondary_reload_fail): Likewise.
27791 (rs6000_secondary_reload_inner): Likewise.
27793 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
27794 macros for first/last GPR and FPR registers.
27795 (LAST_GPR_REGNO): Likewise.
27796 (FIRST_FPR_REGNO): Likewise.
27797 (LAST_FPR_REGNO): Likewise.
27799 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
27800 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
27801 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
27802 (vcond<mode><mode>): Likewise.
27803 (vcondu<mode><mode>): Likewise.
27804 (vector_gtu<mode>): Likewise.
27805 (vector_gte<mode>): Likewise.
27806 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
27807 to prevent the compiler from converting DImode operations to TImode.
27808 (ior<mode>3): Likewise.
27809 (and<mode>3): Likewise.
27810 (one_cmpl<mode>2): Likewise.
27811 (nor<mode>3): Likewise.
27812 (andc<mode>3): Likewise.
27814 * config/rs6000/constraints.md (wt constraint): New constraint
27815 that returns VSX_REGS if TImode is allowed in VSX registers.
27817 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
27818 constant under VSX.
27820 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
27821 similar to TImode, but it is restricted to being in the GPRs.
27823 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
27824 TImode to occupy a single VSX register.
27826 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
27827 -mvsx-timode for power7/power8.
27828 (power7 cpu): Likewise.
27829 (power8 cpu): Likewise.
27831 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
27832 sure that TFmode/TDmode take up two registers if they are ever
27833 allowed in the upper VSX registers.
27834 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
27836 (rs6000_init_hard_regno_mode_ok): Likewise.
27837 (rs6000_debug_reg_global): Add debugging for PTImode and wt
27838 constraint. Print if LRA is turned on.
27839 (rs6000_option_override_internal): Give an error if -mvsx-timode
27840 and VSX is not enabled.
27841 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
27842 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
27843 to reg+offset addressing. Use PTImode when checking offset
27844 addresses for validity.
27845 (reg_offset_addressing_ok_p): Likewise.
27846 (rs6000_legitimate_offset_address_p): Likewise.
27847 (rs6000_legitimize_address): Likewise.
27848 (rs6000_legitimize_reload_address): Likewise.
27849 (rs6000_legitimate_address_p): Likewise.
27850 (rs6000_eliminate_indexed_memrefs): Likewise.
27851 (rs6000_emit_move): Likewise.
27852 (rs6000_secondary_reload): Likewise.
27853 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
27854 reloads to fpr registers to continue to use reg+offset addressing,
27855 but 64-bit reloads to altivec registers need reg+reg addressing.
27856 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
27857 it. Treat LO_SUM like a PLUS operation.
27858 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
27859 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
27860 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
27861 registers to share a register with a smaller sized type, since VSX
27862 puts scalars in the upper 64-bits.
27863 (print_operand): Add support for PTImode.
27864 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
27865 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
27866 registers, but don't have arithmetic support.
27867 (rs6000_memory_move_cost): Add test for VSX.
27868 (rs6000_opt_masks): Add -mvsx-timode.
27870 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
27873 (VSr): Use wt constraint for TImode.
27874 (VSv): Drop TImode support.
27875 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
27876 (vsx_movti_64bit): Likewise.
27877 (vsx_movti_32bit): Likewise.
27878 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
27879 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
27880 one '?' on the appropriate output constraint. Do not allow TImode
27881 logical operations on 32-bit systems.
27882 (vsx_ior<mode>3): Likewise.
27883 (vsx_xor<mode>3): Likewise.
27884 (vsx_one_cmpl<mode>2): Likewise.
27885 (vsx_nor<mode>3): Likewise.
27886 (vsx_andc<mode>3): Likewise.
27887 (vsx_concat_<mode>): Likewise.
27888 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
27890 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
27891 OPTION_MASK_VSX_TIMODE.
27892 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
27893 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
27895 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
27896 (TI2 iterator): New iterator for TImode, PTImode.
27897 (wd mode attribute): Add values for vector types.
27898 (movti_string): Replace TI move operations with operations for TImode
27899 and PTImode. Add support for TImode being allowed in VSX registers.
27900 (mov<mode>_string, TImode/PTImode): Likewise.
27901 (movti_ppc64): Likewise.
27902 (mov<mode>_ppc64, TImode/PTImode): Likewise.
27903 (TI mode splitters): Likewise.
27905 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
27908 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
27910 PR tree-optimization/56355
27911 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
27912 Also handle integers with undefined overflow.
27914 2013-03-20 Catherine Moore <clm@codesourcery.com>
27915 Maciej W. Rozycki <macro@codesourcery.com>
27916 Tom de Vries <tom@codesourcery.com>
27917 Nathan Sidwell <nathan@codesourcery.com>
27918 Iain Sandoe <iain@codesourcery.com>
27919 Nathan Froyd <froydnj@codesourcery.com>
27920 Chao-ying Fu <fu@mips.com>
27922 * doc/extend.texi: (micromips, nomicromips, nocompression):
27923 Document new function attributes.
27924 * doc/invoke.texi (minterlink-compressed, mmicromips,
27925 m14k, m14ke, m14kec): Document new options.
27926 (minterlink-mips16): Update documentation.
27927 * doc/md.texi (ZC, ZD): Document new constraints.
27928 * configure.ac (gcc_cv_as_micromips): Check if linker
27929 supports the .set micromips directive.
27930 * configure: Regenerate.
27931 * config.in: Regenerate.
27932 * config/mips/mips-tables.opt: Regenerate.
27933 * config/mips/micromips.md: New file.
27934 * constraints.md (ZC, ZD): New constraints.
27935 * config/mips/predicates.md (movep_src_register): New predicate.
27936 (movep_src_operand): New predicate.
27937 (non_volatile_mem_operand): New predicate.
27938 * config/mips/mips.md (multimem): New type.
27939 (length): Differentiate between 17-bit and 18-bit branch offsets.
27940 (MOVEP1, MOVEP2): New mode iterator.
27941 (mov_<load>l): Use ZC constraint.
27942 (mov_<load>r): Likewise.
27943 (mov_<store>l): Likewise.
27944 (mov_<store>r): Likewise.
27945 (*branch_equality<mode>_inverted): Add microMIPS support.
27946 (*branch_equality<mode>): Likewise.
27947 (*jump_absolute): Likewise.
27948 (indirect_jump_<mode>): Likewise.
27949 (tablejump_<mode>): Likewise.
27950 (<optab>_internal): Likewise.
27951 (sibcall_internal): Likewise.
27952 (sibcall_value_internal): Likewise.
27953 (prefetch): Use constraint ZD.
27954 * config/mips/mips.opt (minterlink-compressed): New option.
27955 (minterlink-mips16): Now an alias for minterlink-compressed.
27956 (mmicromips): New option.
27957 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
27958 (compare_and_swap_12): Likewise.
27959 (sync_add<mode>): Likewise.
27960 (sync_<optab>_12): Likewise.
27961 (sync_old_<optab>_12): Likewise.
27962 (sync_new_<optab>_12): Likewise.
27963 (sync_nand_12): Likewise.
27964 (sync_old_nand_12): Likewise.
27965 (sync_new_nand_12): Likewise.
27966 (sync_sub<mode>): Likewise.
27967 (sync_old_add<mode>): Likewise.
27968 (sync_old_sub<mode>): Likewise.
27969 (sync_new_add<mode>): Likewise.
27970 (sync_new_sub<mode>): Likewise.
27971 (sync_<optab><mode>): Likewise.
27972 (sync_old_<optab><mode>): Likewise.
27973 (sync_new_<optab><mode>): Likewise.
27974 (sync_nand<mode>): Likewise.
27975 (sync_old_nand<mode>): Likewise.
27976 (sync_new_nand<mode>): Likewise.
27977 (sync_lock_test_and_set<mode>): Likewise.
27978 (test_and_set_12): Likewise.
27979 (atomic_compare_and_swap<mode>): Likewise.
27980 (atomic_exchange<mode>_llsc): Likewise.
27981 (atomic_fetch_add<mode>_llsc): Likewise.
27982 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
27983 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
27984 (umips_save_restore_pattern_p): Likewise.
27985 (umips_load_store_pair_p): Likewise.
27986 (umips_output_load_store_pair): Likewise.
27987 (umips_movep_target_p): Likewise.
27988 (umips_12bit_offset_address_p): Likewise.
27989 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
27990 (mips_base_mips16): Rename this...
27991 (mips_base_compression_flags): ...to this. Update all uses.
27992 (mips_attribute_table): Add micromips, nomicromips and nocompression.
27993 (mips_mips16_decl_p): Delete.
27994 (mips_nomips16_decl_p): Delete.
27995 (mips_get_compress_on_flags): New function.
27996 (mips_get_compress_off_flags): New function.
27997 (mips_get_compress_mode): New function.
27998 (mips_get_compress_on_name): New function.
27999 (mips_get_compress_off_name): New function.
28000 (mips_insert_attributes): Support multiple compression types.
28001 (mips_merge_decl_attributes): Likewise.
28002 (umips_12bit_offset_address_p): New function.
28003 (mips_start_function_definition): Emit .set micromips directive.
28004 (mips_call_may_need_jalx_p): New function.
28005 (mips_function_ok_for_sibcall): Add microMIPS support.
28006 (mips_print_operand_punctuation): Support short delay slots and
28008 (umips_swm_mask, umips_swm_encoding): New.
28009 (umips_build_save_restore): New function.
28010 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
28011 (was_mips16_p): Remove.
28012 (old_compression_mode): New.
28013 (mips_set_compression_mode): New function.
28014 (mips_set_current_function): Add microMIPS support.
28015 (mips_option_override): Likewise.
28016 (umips_save_restore_pattern_p): New function.
28017 (umips_output_save_restore): New function.
28018 (umips_load_store_pair_p_1): New function.
28019 (umips_load_store_pair_p): New function.
28020 (umips_output_load_store_pair_1): New function.
28021 (umips_output_load_store_pair): New function.
28022 (umips_movep_target_p) New function.
28023 (mips_prepare_pch_save): Add microMIPS support.
28024 * config/mips/mips.h (TARGET_COMPRESSION): New.
28025 (TARGET_CPU_CPP_BUILTINS): Update macro
28026 to use new compression flags and to support microMIPS.
28027 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
28028 (MIPS_ARCH_FLOAT_SPEC): Likewise.
28029 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
28030 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
28031 (ASM_SPEC): Support mmicromips and mno-micromips.
28032 (M16STORE_REG_P): New macro.
28033 (MIPS_CALL): Support TARGET_MICROMIPS.
28034 (MICROMIPS_J): New macro.
28035 (mips_base_mips16): Rename this...
28036 (mips_base_compression_flags): ...to this.
28037 (UMIPS_12BIT_OFFSET_P): New macro.
28038 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
28039 (MULTILIB_DIRNAMES): Likewise.
28040 2013-03-20 Richard Biener <rguenther@suse.de>
28042 PR tree-optimization/56661
28043 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
28044 the result does not have to be distinct.
28046 2013-03-20 Richard Biener <rguenther@suse.de>
28048 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
28051 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
28052 Steven Bosscher <steven@gcc.gnu.org>
28054 PR rtl-optimization/56605
28055 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
28057 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
28060 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
28061 that require movd instead of movq.
28063 2013-03-20 Richard Biener <rguenther@suse.de>
28065 * tree-ssa-structalias.c (struct variable_info): Add pointer
28066 to the first field of an aggregate with sub-vars. Make
28067 this and the pointer to the next subfield its ID.
28068 (vi_next): New function.
28069 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
28070 storedanything_id, integer_id): Increment by one.
28071 (new_var_info, get_call_vi, lookup_call_clobber_vi,
28072 get_call_clobber_vi): Adjust.
28073 (solution_set_expand): Simplify and speedup.
28074 (solution_set_add): Inline into ...
28075 (set_union_with_increment): ... this. Adjust accordingly.
28076 (do_sd_constraint): Likewise.
28077 (do_ds_constraint): Likewise.
28078 (do_complex_constraint): Simplify.
28079 (build_pred_graph): Adjust.
28080 (solve_graph): Likewise. Simplify and speedup.
28081 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
28082 get_constraint_for_component_ref, get_constraint_for_1,
28083 first_vi_for_offset, first_or_preceding_vi_for_offset,
28084 create_function_info_for, create_variable_info_for_1,
28085 create_variable_info_for, intra_create_variable_infos): Adjust.
28086 (init_base_vars): Push NULL for ID zero.
28087 (compute_points_to_sets): Adjust.
28089 2013-03-20 Richard Biener <rguenther@suse.de>
28091 * cfgloop.c (verify_loop_structure): Streamline and avoid
28092 ICEing on corrupt loop tree.
28093 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
28096 2013-03-20 Richard Biener <rguenther@suse.de>
28098 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
28099 check whether an SSA update is needed.
28101 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
28103 * config/mips/constraints.md (T): Rename to...
28107 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
28108 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
28110 2013-03-19 Ian Bolton <ian.bolton@arm.com>
28112 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
28113 (*subsi3_carryin_uxtw): Likewise.
28115 2013-03-19 Ian Bolton <ian.bolton@arm.com>
28117 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
28118 (*rorsi3_insn_uxtw): Likewise.
28120 2013-03-19 Ian Bolton <ian.bolton@arm.com>
28122 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
28123 (*extrsi5_insn_uxtw): Likewise.
28125 2013-03-19 Richard Biener <rguenther@suse.de>
28127 PR tree-optimization/56273
28128 * passes.c (init_optimization_passes): Move second VRP after DOM.
28130 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
28132 * config/i386/i386.md (*movti_internal): Merge from
28133 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
28134 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
28135 nox64 isa attributes.
28137 2013-03-18 Richard Biener <rguenther@suse.de>
28139 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
28141 (merge_node_constraints): Likewise.
28142 (build_succ_graph): Likewise.
28143 (valid_graph_edge): Inline into single caller.
28144 (unify_nodes): Likewise. Use bitmap_set_bit return value
28146 (scc_visit): Fix formatting and variable use.
28147 (do_sd_constraint): Use gcc_checking_assert.
28148 (do_ds_constraint): Likewise.
28149 (do_complex_constraint): Likewise.
28150 (condense_visit): Likewise. Cleanup.
28151 (dump_pred_graph): New function.
28152 (perform_var_substitution): Dump the pred-graph before
28153 variable substitution.
28154 (find_equivalent_node): Use gcc_checking_assert.
28155 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
28157 2013-03-18 Richard Biener <rguenther@suse.de>
28159 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
28160 Remove cond_expr_stmt_list argument and do not gimplify the
28162 (vect_loop_versioning): Adjust.
28163 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
28164 Cleanup to use less temporaries.
28165 (vect_create_data_ref_ptr): Cleanup.
28167 2013-03-18 Jakub Jelinek <jakub@redhat.com>
28169 PR tree-optimization/56635
28170 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
28171 require types_compatible_p types.
28173 2013-03-18 Nick Clifton <nickc@redhat.com>
28175 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
28176 spurious backslash.
28178 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
28179 Add missing line to comment describing function.
28181 2013-03-18 Richard Biener <rguenther@suse.de>
28183 PR tree-optimization/56210
28184 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
28185 Handle string / character search functions.
28186 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
28188 2013-03-18 Richard Biener <rguenther@suse.de>
28190 PR middle-end/56483
28191 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
28192 and implement properly.
28193 * gimple.h (gimple_cond_single_var_p): Remove.
28195 2013-03-18 Richard Biener <rguenther@suse.de>
28197 * tree-data-ref.h (find_data_references_in_loop): Declare.
28198 * tree-data-ref.c (get_references_in_stmt): Use a stack
28199 vector pre-allocated in the callers.
28200 (find_data_references_in_stmt): Adjust.
28201 (graphite_find_data_references_in_stmt): Likewise.
28202 (create_rdg_vertices): Likewise.
28203 (find_data_references_in_loop): Export.
28204 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
28205 Compute dependences here...
28206 (vect_analyze_data_refs): ...not here. When we encounter
28207 a non-vectorizable data reference in basic-block vectorization
28208 truncate the data reference vector. Do not bother to
28209 fixup data-dependence information for gather loads.
28210 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
28211 of data references, as reported.
28213 2013-03-18 Richard Biener <rguenther@suse.de>
28215 PR tree-optimization/3713
28216 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
28217 has_constants and expr.
28218 (stmt_has_constants): Properly valueize SSA names when deciding
28219 whether the stmt has constants.
28221 2013-03-18 Richard Biener <rguenther@suse.de>
28223 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
28224 whole function when there is nothing to do.
28225 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
28226 * tree-vectorizer.c (vectorize_loops): Update virtual and
28227 loop-closed SSA once.
28228 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
28230 2013-03-18 Richard Biener <rguenther@suse.de>
28232 PR middle-end/56113
28233 * domwalk.c (bb_postorder): New global static.
28234 (cmp_bb_postorder): New function.
28235 (walk_dominator_tree): Replace scheme imposing an order for
28236 visiting dominator sons by one sorting them at the time they
28237 are pushed on the stack.
28239 2013-03-18 Richard Biener <rguenther@suse.de>
28241 PR tree-optimization/39326
28242 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
28243 (struct mem_ref): Replace mem member with ao_ref typed member.
28244 (MEM_ANALYZABLE): Adjust.
28245 (memref_eq): Likewise.
28246 (mem_ref_alloc): Likewise.
28247 (gather_mem_refs_stmt): Likewise.
28248 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
28249 (execute_sm_if_changed_flag_set): Adjust.
28250 (execute_sm): Likewise.
28251 (ref_always_accessed_p): Likewise.
28252 (refs_independent_p): Likewise.
28253 (can_sm_ref_p): Likewise.
28255 2013-03-18 Jakub Jelinek <jakub@redhat.com>
28258 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
28259 return 1 even for !unsignedp.
28261 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
28263 * config/i386/i386.md (isa): Add x64 and nox64.
28264 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
28265 (*pushtf): Enable *roF alternative for x64 isa only.
28266 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
28267 mode attribute of integer alternatives to DImode for TARGET_64BIT.
28268 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
28269 (*movtf_internal): Merge from *movtf_internal_rex64 and
28270 *movtf_internal_sse. Use x64 and nox64 isa attributes.
28271 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
28272 nox64 isa attributes.
28273 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
28274 nox64 isa attributes.
28275 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
28277 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
28279 * config/alpha/alpha.c (TARGET_LRA_P): New define.
28281 2013-03-17 Jakub Jelinek <jakub@redhat.com>
28284 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
28285 class names. Remove trailing comma after "ALL_REGS".
28287 2013-03-16 Jan Hubicka <jh@suse.cz>
28289 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
28290 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
28291 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
28292 of cgraph_get_create_node.
28293 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
28295 2013-03-16 Jason Merrill <jason@redhat.com>
28298 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
28299 with DW_AT_default_value.
28301 2013-03-16 Jakub Jelinek <jakub@redhat.com>
28303 * BASE-VER: Set to 4.9.0.
28305 2013-03-14 Andi Kleen <ak@linux.intel.com>
28308 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
28309 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
28310 Document _x* TSX intrinsics.
28312 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
28313 David Holsgrove <david.holsgrove@xilinx.com>
28315 * configure.ac: Add MicroBlaze TLS support detection.
28316 * configure: Regenerate.
28317 * config/microblaze/microblaze-protos.h
28318 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
28319 symbol_mentioned_p, label_mentioned_p): Add prototypes.
28320 * config/microblaze/microblaze.c (microblaze_address_type): Add
28321 ADDRESS_TLS and tls_reloc address types.
28322 (microblaze_address_info): Add tls_reloc.
28323 (TARGET_HAVE_TLS): Define.
28324 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
28325 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
28326 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
28327 load_tls_operand, microblaze_call_tls_get_addr,
28328 microblaze_legitimize_tls_address): New functions.
28329 (microblaze_classify_unspec): Handle UNSPEC_TLS.
28330 (get_base_reg): Use microblaze_tls_symbol_p.
28331 (microblaze_classify_address): Handle TLS.
28332 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
28333 label_mentioned_p and microblaze_tls_referenced_p.
28334 (microblaze_legitimize_address): Handle TLS.
28335 (microblaze_address_insns): Handle ADDRESS_TLS.
28336 (pic_address_needs_scratch): Handle TLS.
28337 (print_operand_address): Handle TLS.
28338 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
28339 (microblaze_expand_move): Handle TLS.
28340 (microblaze_legitimate_constant_p): Check
28341 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
28342 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
28343 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
28344 (PIC_OFFSET_TABLE_REGNUM): Set.
28345 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
28346 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
28347 (addsi3, movsi_internal2, movdf_internal): Update constraints
28348 * config/microblaze/predicates.md (arith_plus_operand): Define
28349 (move_operand): Redefine as move_src_operand,
28350 check microblaze_tls_referenced_p.
28352 2013-03-14 Ian Bolton <ian.bolton@arm.com>
28354 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
28355 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
28357 2013-03-14 Ian Bolton <ian.bolton@arm.com>
28359 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
28362 2013-03-14 Jakub Jelinek <jakub@redhat.com>
28364 PR tree-optimization/53265
28365 * common.opt (Waggressive-loop-optimizations): New option.
28366 * tree-ssa-loop-niter.c: Include tree-pass.h.
28367 (do_warn_aggressive_loop_optimizations): New function.
28368 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
28369 if number_of_latch_executions returned constant.
28370 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
28371 early. If number_of_latch_executions returned constant, set
28372 nb_iterations_upper_bound back to it.
28373 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
28375 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
28376 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
28378 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
28379 (MULTILIB_OSDIRNAMES): Set.
28380 * genmultilib: If defaultosdirname doesn't start with :: , set
28381 defaultosdirname2 instead, clear it and emit two . multilib_raw
28382 entries instead of just one.
28384 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
28386 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
28387 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
28388 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
28389 (SUBTARGET_OVERRIDE_OPTIONS): New.
28391 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
28394 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
28395 (musermode): Convert to Var(TARGET_USERMODE).
28396 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
28397 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
28398 * config/sh/sh.c (sh_option_override): Use
28399 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
28400 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
28402 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
28404 (udivsi3_i4_single, divsi3_i4_single): Use
28405 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
28407 2013-03-13 Dave Korn <dave.korn.cygwin@gmail.com>
28409 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
28412 2013-03-13 Richard Biener <rguenther@suse.de>
28414 PR tree-optimization/56608
28415 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
28416 calls when vectorizing basic-blocks.
28418 2013-03-13 Jakub Jelinek <jakub@redhat.com>
28421 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
28424 2013-03-12 Jakub Jelinek <jakub@redhat.com>
28426 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
28428 2013-03-11 Jan Hubicka <jh@suse.cz>
28431 * lto-streamer-out.c (output_symbol_p): Skip references from
28432 constructors of external variables.
28434 2013-03-11 Jan Hubicka <jh@suse.cz>
28436 PR middle-end/56571
28437 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
28439 * emit-rtl.c (verify_rtx_sharing): Likewise.
28440 (copy_insn_1): Likewise.
28441 * rtl.c (copy_rtx): Likewise.
28443 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
28446 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
28447 output_operand_lossage message.
28449 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
28452 * arm.c (shift_op): Validate RTL pattern on the fly.
28453 (arm_print_operand, case 'S'): Don't use shift_operator to validate
28456 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
28459 * config/pa/pa.md (call_value): Check for calls to powf and direct to
28460 new call patterns that clobber %fr12.
28461 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
28462 split and postreload patterns.
28463 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
28464 registers %fr12 and %fr12R as call used.
28466 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
28468 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
28469 (canon_address, record_store, replace_read, check_mem_read_rtx,
28470 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
28471 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
28472 rest_of_handle_dse): Likewise.
28474 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
28476 PR middle-end/56524
28477 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
28479 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
28480 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
28481 (save_optabs_if_changed): Replace with...
28482 (init_tree_optimization_optabs): ...this.
28483 * optabs.c (save_optabs_if_changed): Rename to...
28484 (init_tree_optimization_optabs): ...this. Take the optimization node
28485 as argument. Do nothing if the base optabs are already correct.
28486 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
28487 to recompute optabs.
28488 * function.h (function): Remove optabs field.
28489 * function.c (invoke_set_current_function_hook): Call
28490 init_tree_optimization_optabs. Use the result to initialize
28493 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
28495 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
28496 if GTMA_HAS_NO_INSTRUMENTATION.
28497 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
28498 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
28499 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
28500 * gimple-pretty-print.c (dump_gimple_transaction): Handle
28501 GTMA_HAS_NO_INSTRUMENTATION.
28503 2013-03-08 Jakub Jelinek <jakub@redhat.com>
28505 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
28508 2013-03-08 Marek Polacek <polacek@redhat.com>
28509 Jakub Jelinek <jakub@redhat.com>
28511 PR tree-optimization/56478
28512 * predict.c (is_comparison_with_loop_invariant_p): Change the
28513 type of loop_step to tree.
28514 (predict_loops): Adjust.
28515 (predict_iv_comparison): Perform the computations on double_ints.
28517 2013-03-08 Richard Biener <rguenther@suse.de>
28519 PR tree-optimization/56570
28520 * tree-cfg.c (verify_expr_location_1): Verify locations for
28522 * tree-sra.c (create_access_replacement): Strip locations
28523 from DECL_DEBUG_EXPRs.
28525 2013-03-08 Richard Biener <rguenther@suse.de>
28527 * tree-inline.c (expand_call_inline): Do not associate
28528 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
28529 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
28531 2013-03-08 Richard Biener <rguenther@suse.de>
28533 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
28534 or block changes with -Og. Fix for location / block encoding
28535 changes and PHI arguments with locations.
28537 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
28539 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
28541 (struct output_info): Likewise.
28542 (register_overhead): Remove bad gcc_assert.
28543 (bitmap_find_bit): If there is only a single bitmap element, do not
28544 count a miss as a search.
28545 (print_statistics): Update for counter type changes.
28546 (dump_bitmap_statistics): Likewise. Print headers such that they
28547 are properly lined up with the printed counters.
28549 2013-03-07 Jakub Jelinek <jakub@redhat.com>
28551 PR tree-optimization/56559
28552 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
28553 check that it has only a single use.
28555 2013-03-07 Richard Biener <rguenther@suse.de>
28557 * doc/invoke.texi (fwhole-program): Discourage use in combination
28560 2013-03-06 Jakub Jelinek <jakub@redhat.com>
28562 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
28564 PR tree-optimization/56539
28565 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
28566 instead of GSI_CONTINUE_LINKING as last argument to
28567 force_gimple_operand_gsi. Adjust function comment.
28569 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
28572 PR middle-end/56548
28573 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
28574 promoted mode, convert the result back to the original mode.
28576 2013-03-06 Richard Biener <rguenther@suse.de>
28578 PR middle-end/56294
28579 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
28580 (insert_updated_phi_nodes_compare_uids): New function.
28581 (update_ssa): Sort symbols_to_rename after UID before
28582 traversing it to insert PHI nodes.
28584 2013-03-06 Richard Biener <rguenther@suse.de>
28586 PR middle-end/50494
28587 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
28588 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
28591 2013-02-13 Richard Biener <rguenther@suse.de>
28594 * varasm.c (output_constant_def_1): Get the decl representing
28595 the constant as argument.
28596 (output_constant_def): Wrap output_constant_def_1.
28597 (make_decl_rtl): Use output_constant_def_1 with the decl
28598 representing the constant.
28599 (build_constant_desc): Optionally re-use a decl already
28600 representing the constant.
28601 (tree_output_constant_def): Adjust.
28603 2013-03-06 Joey Ye <joey.ye@arm.com>
28606 * gcc.c (convert_white_space): New function.
28607 (main): Handles white space in function name.
28609 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
28612 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
28613 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
28614 to SH_DIV_CALL_TABLE for TARGET_SH2.
28615 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
28617 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
28618 call-table options.
28620 2013-03-05 Sterling Augustine <saugustine@google.com>
28621 Cary Coutant <ccoutant@google.com>
28624 * dwarf2out.c (resolve_addr): Don't call
28625 remove_loc_list_addr_table_entries a second time for the same
28628 2013-03-05 Jakub Jelinek <jakub@redhat.com>
28631 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
28632 (avoid_complex_debug_insns): New function.
28633 (expand_debug_locations): Call it.
28635 PR rtl-optimization/56484
28636 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
28637 lifetimes of hard registers on small register class machines.
28639 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
28641 * config/microblaze/microblaze-protos.h: Rename
28642 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
28643 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
28645 (microblaze_fast_interrupt_function_p): New function.
28646 (microblaze_is_interrupt_handler): Rename to
28647 microblaze_is_interrupt_variant and add fast_interrupt check.
28648 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
28649 (save_restore_insns): Likewise.
28650 (compute_frame_size): Likewise.
28651 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
28652 (microblaze_globalize_label): Likewise.
28653 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
28654 * config/microblaze/microblaze.md: Use wrapper
28655 microblaze_is_interrupt_variant.
28657 2013-03-05 Kai Tietz <ktietz@redhat.com>
28659 * sdbout.c (sdbout_one_type): Switch to current function's section
28660 supporting cold/hot.
28662 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
28664 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
28667 2013-03-05 Jakub Jelinek <jakub@redhat.com>
28669 PR middle-end/56461
28670 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
28671 if VALGRIND_GET_VBITS is defined, temporarily make object
28672 memory all defined, and restore previous valgrind addressability
28673 and definability afterwards. Free this_object at the end.
28675 PR middle-end/56461
28676 * lra.c (lra): Call lra_clear_live_ranges if live_p,
28677 right before calling lra_create_live_ranges, also call it
28678 when clearing live_p. Only call lra_clear_live_ranges
28679 at the end if live_p.
28681 PR middle-end/56461
28682 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
28684 2013-03-05 Richard Biener <rguenther@suse.de>
28686 PR tree-optimization/56521
28687 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
28690 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
28693 * except.h (remove_unreachable_eh_regions): New prototype.
28694 * except.c (remove_eh_handler_splicer): New function, split out
28695 of remove_eh_handler.
28696 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
28697 warning about running it on many EH regions one at a time.
28698 (remove_unreachable_eh_regions_worker): New function, walk the
28699 EH tree in depth-first order and remove non-marked regions.
28700 (remove_unreachable_eh_regions): New function.
28701 * tree-eh.c (mark_reachable_handlers): New function, split out
28702 from remove_unreachable_handlers.
28703 (remove_unreachable_handlers): Use mark_reachable_handlers and
28704 remove_unreachable_eh_regions.
28705 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
28706 and remove_unreachable_eh_regions.
28708 2013-03-05 Richard Biener <rguenther@suse.de>
28710 PR middle-end/56525
28711 * loop-init.c (fix_loop_structure): Remove loops in two stages,
28712 not freeing them until the end.
28714 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
28716 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
28718 2013-03-05 Richard Biener <rguenther@suse.de>
28720 PR tree-optimization/56270
28721 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
28722 of loads after scheduling an SLP instance.
28724 2013-03-05 Jakub Jelinek <jakub@redhat.com>
28726 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
28728 (check_gcc_parallelize): Run guality.exp as a separate job from
28729 vect.exp with unsorted.exp and $(dg_target_exps) separately from
28730 struct-layout-1.exp with stackalign.exp.
28732 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
28734 PR middle-end/56461
28735 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
28736 load_index sbitmap even if some bit in it isn't set.
28738 PR middle-end/56461
28739 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
28740 (discover_iteration_bound_by_body_walk): Change queues to
28741 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
28742 spelling in comment. Call safe_push on queues[bound_index] directly.
28743 Release queues[queue_index] in every iteration unconditionally.
28744 Release bounds vector.
28746 PR middle-end/56461
28747 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
28748 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
28749 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
28752 2013-03-05 Richard Biener <rguenther@suse.de>
28755 * tree-inline.c (remap_blocks_to_null): New function.
28756 (expand_call_inline): When expanding a call stmt without
28757 an associated block inline remap all callee blocks to NULL.
28759 2013-03-05 Jakub Jelinek <jakub@redhat.com>
28761 PR rtl-optimization/56494
28762 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
28763 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
28764 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
28766 PR middle-end/56461
28767 * sel-sched-ir.c (free_sched_pools): Release
28768 succs_info_pool.stack[succs_info_pool.max_top] vectors too
28769 if succs_info_pool.max_top isn't -1.
28772 * opts.c (opts_obstack, opts_concat): Moved to...
28773 * opts-common.c (opts_obstack, opts_concat): ... here.
28775 2013-03-04 Jakub Jelinek <jakub@redhat.com>
28777 PR middle-end/56461
28778 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
28780 2013-03-04 Martin Jambor <mjambor@suse.cz>
28782 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
28783 all appropriate places.
28785 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
28787 PR tree-optimization/56424
28788 * ipa-split.c (split_function): Do not set the RSO flag if result is
28789 not by reference and its type is a register type.
28791 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
28793 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
28794 (microblaze_legitimate_pic_operand): Likewise
28795 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
28796 new function microblaze_legitimate_pic_operand
28797 * config/microblaze/microblaze-protos.h
28798 (microblaze_legitimate_pic_operand): Declare.
28800 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28802 * config/microblaze/predicates.md (call_insn_simple_operand):
28803 New predicate for supported rtx code types.
28804 * config/microblaze/microblaze.md (call_internal1): Use
28805 call_insn_simple_operand predicate.
28807 2013-03-04 Jakub Jelinek <jakub@redhat.com>
28809 PR middle-end/56461
28810 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
28811 partitions.ordered_remove.
28813 PR middle-end/56461
28814 * tree-vect-stmts.c (vectorizable_conversion): Don't call
28815 vec_oprnds0.create (1) for modifier == NONE.
28817 PR middle-end/56461
28818 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
28819 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
28820 vec_oprnds1 right before pushing anything to it for
28823 PR middle-end/56461
28824 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
28825 set nbbs to 0 instead of having separate code path.
28826 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
28827 instead of false as last argument if returning NULL.
28829 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
28831 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
28832 the attribute is now called "target" instead of "option".
28833 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
28834 * doc/tm.texi.in (Target Attributes): Likewise document the correct
28835 attribute/pragma name for TARGET_OPTION_VALID_P and
28836 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
28837 * doc/tm.texi: Regenerated.
28839 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
28841 * config/microblaze/microblaze.c:
28842 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
28843 * config/microblaze/microblaze.h: Add -mxl-reorder to
28845 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
28846 instructions emitted if TARGET_REORDER.
28847 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
28848 or 0 for -m/-mno case, but initialises as 2 to detect default use case
28851 2013-03-01 Xinliang David Li <davidxl@google.com>
28853 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
28856 2013-03-01 Jakub Jelinek <jakub@redhat.com>
28858 PR middle-end/56461
28859 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
28860 vector even when returning true. Fix up function comment formatting.
28862 PR middle-end/56461
28863 * ira-build.c (ira_loop_nodes_count): New variable.
28864 (create_loop_tree_nodes): Initialize it.
28865 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
28867 PR middle-end/56461
28868 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
28869 method on dr_chain and result_chain.
28870 * tree-vect-stmts.c (vectorizable_store): Only call
28871 result_chain.create if j == 0.
28873 PR middle-end/56461
28874 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
28875 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
28876 before overwriting it.
28878 2013-03-01 Tobias Burnus <burnus@net-b.de>
28880 * doc/extended.texi (C Extensions): Change order in @menu
28882 (Other MIPS Built-in Functions): Move last MIPS entry before
28883 "picoChip Built-in Functions".
28884 (SH Built-in Functions): Move after RX Built-in Functions.
28885 * doc/gcc.texi (Introduction): Change order in @menu
28887 * doc/md.texi (Constraints): Ditto.
28888 * gty.texi (Type Information): Ditto.
28889 (User-provided marking routines for template types): Make
28891 * doc/invoke.texi (AArch64 Options): Move before
28892 "Adapteva Epiphany Options".
28894 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
28895 Jakub Jelinek <jakub@redhat.com>
28898 * asan.c (gate_asan): Lookup no_sanitize_address instead of
28899 no_address_safety_analysis attribute.
28900 * doc/extend.texi (no_address_safety_attribute): Rename to
28901 no_sanitize_address attribute, mention no_address_safety_analysis
28902 attribute as deprecated alias.
28904 2013-02-28 Jakub Jelinek <jakub@redhat.com>
28906 PR middle-end/56461
28907 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
28908 type to vec<vec<tree> > *.
28909 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
28910 to be vec<tree> instead of vec<tree> *, set vec_defs
28911 to vNULL and call vec_defs.create (number_of_vects), adjust other
28913 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
28914 vectorizable_condition): Adjust vect_get_slp_defs callers.
28916 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
28918 * config/aarch64/aarch64.c
28919 (aarch64_float_const_representable): Remove unused variable.
28921 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
28923 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
28925 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
28927 * config/aarch64/aarch64-builtins.c
28928 (aarch64_init_simd_builtins): Make static.
28930 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
28932 * config/aarch64/aarch64.c
28933 (aarch64_simd_make_constant): Make static.
28935 2013-02-28 Martin Jambor <mjambor@suse.cz>
28937 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
28938 with no initialization to the RHS of debug statements.
28940 2013-02-28 Martin Jambor <mjambor@suse.cz>
28942 PR tree-optimization/56294
28943 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
28945 (get_access_replacement): Do not call create_access_replacement.
28946 Assert a replacement exists.
28947 (get_repl_default_def_ssa_name): Create the replacement declaration
28950 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
28952 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
28953 final_end_function.
28955 2013-02-28 Marek Polacek <polacek@redhat.com>
28957 PR rtl-optimization/56466
28958 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
28959 if we're changing a loop.
28960 (peel_loops_completely): Likewise.
28962 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
28965 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
28967 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
28970 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
28971 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
28972 INTX_FTYPE_FX, FX_FTYPE_INTX.
28973 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
28975 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
28977 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
28978 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
28979 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
28980 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
28981 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
28982 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
28983 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
28984 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
28985 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
28986 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
28987 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
28988 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
28989 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
28990 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
28991 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
28992 (avrxmega6): Increase max flash segments from 5 to 6.
28993 * config/avr/t-multilib: Regenerate.
28994 * config/avr/avr-tables.opt: Regenerate.
28995 * doc/avr-mmcu.texi: Regenerate.
28997 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
28999 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
29000 (avr_device_to_arch): Rename to avr_device_to_ld.
29001 (avr_device_to_as): New prototype.
29002 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
29003 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
29004 * config/avr/driver-avr.c (avr_device_to_as): New.
29005 (avr_device_to_arch): Rename to avr_device_to_ld.
29007 2013-02-27 Jakub Jelinek <jakub@redhat.com>
29009 PR middle-end/56461
29010 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
29011 method on dr_chain and result_chain.
29013 PR middle-end/56461
29014 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
29015 pointer_set_destroy on not_executed_last_iteration.
29017 PR middle-end/56461
29018 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
29020 PR middle-end/56461
29021 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
29022 FOR_EACH_DEFINED_FUNCTION when freeing state.
29024 PR middle-end/56461
29025 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
29027 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
29030 PR middle-end/56461
29031 * ipa-cp.c (decide_whether_version_node): Call vec_free on
29032 known_aggs[i].items and release known_aggs vector.
29034 PR middle-end/56461
29035 * ipa-reference.c (propagate): Free node_info even for alias nodes.
29037 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
29039 * config/microblaze/microblaze.c (microblaze_emit_compare):
29040 Use xor for EQ/NE comparisions.
29041 * config/microblaze/microblaze.md (cstoresf4): Add constraints
29042 (cbranchsf4): Adjust operator to comparison_operator.
29044 2013-02-27 Jakub Jelinek <jakub@redhat.com>
29046 PR middle-end/56461
29047 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
29049 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
29050 vec_safe_push, always update *slot.
29051 (redirect_edge_var_map_clear): Use vec_free.
29052 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
29053 (free_var_map_entry): Use vec_free.
29054 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
29055 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
29057 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
29059 PR middle-end/45472
29060 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
29061 when the may_trap_p bit of the exprs being merged differs.
29062 Reorder tests for speculativeness in the logical and operator.
29064 2013-02-27 Jakub Jelinek <jakub@redhat.com>
29066 * incpath.c (add_standard_paths): Use reconcat instead of concat
29067 where appropriate and avoid leaking memory.
29069 * opts.h: Include obstack.h.
29070 (opts_concat): New prototype.
29071 (opts_obstack): New declaration.
29072 * opts.c (opts_concat): New function.
29073 (opts_obstack): New variable.
29074 (init_options_struct): Call gcc_init_obstack on opts_obstack.
29075 (finish_options): Use opts_concat instead of concat
29076 and XOBNEWVEC instead of XNEWVEC.
29077 * opts-common.c (generate_canonical_option, decode_cmdline_option,
29078 generate_option): Likewise.
29079 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
29080 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
29083 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
29084 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
29086 2013-02-26 Jakub Jelinek <jakub@redhat.com>
29088 PR middle-end/56461
29089 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
29091 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
29093 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
29094 (arm_block_move_unaligned_straight): Likewise.
29095 (arm_adjust_block_mem): Likewise.
29097 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
29100 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
29101 temp, cond and label.
29102 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
29105 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
29106 * config/c6x/c6x.h (dbx_register_map): Update declaration.
29109 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
29110 of prologue/epilogue functions.
29113 * config/tilegx/tilegx.c (tilegx_expand_prologue):
29114 Remove unused variable cfa_offset.
29115 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
29118 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
29119 type promotion to unsigned.
29122 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
29123 for HOST_WIDE_INT of 32 bit / same size as int.
29124 (arm_block_move_unaligned_straight): Likewise.
29125 (arm_adjust_block_mem): Likewise.
29128 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
29131 2013-02-26 Marek Polacek <polacek@redhat.com>
29133 PR tree-optimization/56426
29134 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
29136 2013-02-26 Richard Biener <rguenther@suse.de>
29139 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
29140 unused variable loops.
29142 2013-02-26 Jakub Jelinek <jakub@redhat.com>
29144 PR tree-optimization/56448
29145 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
29146 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
29147 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
29148 later operands of the references, or even first operand for
29149 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
29151 PR tree-optimization/56443
29152 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
29153 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
29154 to type_for_mode langhook.
29156 2013-02-25 Matt Turner <mattst88@gmail.com>
29158 * doc/invoke.texi: Document r4700.
29160 2013-02-25 Richard Biener <rguenther@suse.de>
29162 PR tree-optimization/56175
29163 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
29165 (simplify_bitwise_binary): ... here. Also guard the conversion
29166 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
29168 2013-02-25 Catherine Moore <clm@codesourcery.com>
29171 2013-02-24 Catherine Moore <clm@codesourcery.com>
29172 Maciej W. Rozycki <macro@codesourcery.com>
29173 Tom de Vries <tom@codesourcery.com>
29174 Nathan Sidwell <nathan@codesourcery.com>
29175 Iain Sandoe <iain@codesourcery.com>
29176 Nathan Froyd <froydnj@codesourcery.com>
29177 Chao-ying Fu <fu@mips.com>
29179 * doc/extend.texi: (micromips, nomicromips, nocompression):
29180 Document new function attributes.
29181 * doc/invoke.texi (minterlink-compressed, mmicromips,
29182 m14k, m14ke, m14kec): Document new options.
29183 (minterlink-mips16): Update documentation.
29184 * doc/md.texi (ZC, ZD): Document new constraints.
29185 * configure.ac (gcc_cv_as_micromips): Check if linker
29186 supports the .set micromips directive.
29187 * configure: Regenerate.
29188 * config.in: Regenerate.
29189 * config/mips/mips-tables.opt: Regenerate.
29190 * config/mips/micromips.md: New file.
29191 * constraints.md (ZC, AD): New constraints.
29192 * config/mips/predicates.md (movep_src_register): New predicate.
29193 (movep_src_operand): New predicate.
29194 (non_volatile_mem_operand): New predicate.
29195 * config/mips/mips.md (multimem): New type.
29196 (length): Differentiate between 17-bit and 18-bit branch offsets.
29197 (MOVEP1, MOVEP2): New mode iterator.
29198 (mov_<load>l): Use ZC constraint.
29199 (mov_<load>r): Likewise.
29200 (mov_<store>l): Likewise.
29201 (mov_<store>r): Likewise.
29202 (*branch_equality<mode>_inverted): Add microMIPS support.
29203 (*branch_equality<mode>): Likewise.
29204 (*jump_absolute): Likewise.
29205 (indirect_jump_<mode>): Likewise.
29206 (tablejump_<mode>): Likewise.
29207 (<optab>_internal): Likewise.
29208 (sibcall_internal): Likewise.
29209 (sibcall_value_internal): Likewise.
29210 (prefetch): Use constraint ZD.
29211 * config/mips/mips.opt (minterlink-compressed): New option.
29212 (minterlink-mips16): Now an alias for minterlink-compressed.
29213 (mmicromips): New option.
29214 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
29215 (compare_and_swap_12): Likewise.
29216 (sync_add<mode>): Likewise.
29217 (sync_<optab>_12): Likewise.
29218 (sync_old_<optab>_12): Likewise.
29219 (sync_new_<optab>_12): Likewise.
29220 (sync_nand_12): Likewise.
29221 (sync_old_nand_12): Likewise.
29222 (sync_new_nand_12): Likewise.
29223 (sync_sub<mode>): Likewise.
29224 (sync_old_add<mode>): Likewise.
29225 (sync_old_sub<mode>): Likewise.
29226 (sync_new_add<mode>): Likewise.
29227 (sync_new_sub<mode>): Likewise.
29228 (sync_<optab><mode>): Likewise.
29229 (sync_old_<optab><mode>): Likewise.
29230 (sync_new_<optab><mode>): Likewise.
29231 (sync_nand<mode>): Likewise.
29232 (sync_old_nand<mode>): Likewise.
29233 (sync_new_nand<mode>): Likewise.
29234 (sync_lock_test_and_set<mode>): Likewise.
29235 (test_and_set_12): Likewise.
29236 (atomic_compare_and_swap<mode>): Likewise.
29237 (atomic_exchange<mode>_llsc): Likewise.
29238 (atomic_fetch_add<mode>_llsc): Likewise.
29239 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
29240 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
29241 (umips_save_restore_pattern_p): Likewise.
29242 (umips_load_store_pair_p): Likewise.
29243 (umips_output_load_store_pair): Likewise.
29244 (umips_movep_target_p): Likewise.
29245 (umips_12bit_offset_address_p): Likewise.
29246 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
29247 (mips_base_mips16): Rename this...
29248 (mips_base_compression_flags): ...to this. Update all uses.
29249 (mips_attribute_table): Add micromips, nomicromips and nocompression.
29250 (mips_mips16_decl_p): Delete.
29251 (mips_nomips16_decl_p): Delete.
29252 (mips_get_compress_on_flags): New function.
29253 (mips_get_compress_off_flags): New function.
29254 (mips_get_compress_mode): New function.
29255 (mips_get_compress_on_name): New function.
29256 (mips_get_compress_off_name): New function.
29257 (mips_insert_attributes): Support multiple compression types.
29258 (mips_merge_decl_attributes): Likewise.
29259 (umips_12bit_offset_address_p): New function.
29260 (mips_start_function_definition): Emit .set micromips directive.
29261 (mips_call_may_need_jalx_p): New function.
29262 (mips_function_ok_for_sibcall): Add microMIPS support.
29263 (mips_print_operand_punctuation): Support short delay slots and
29265 (umips_swm_mask, umips_swm_encoding): New.
29266 (umips_build_save_restore): New function.
29267 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
29268 (was_mips16_p): Remove.
29269 (old_compression_mode): New.
29270 (mips_set_compression_mode): New function.
29271 (mips_set_current_function): Add microMIPS support.
29272 (mips_option_override): Likewise.
29273 (umips_save_restore_pattern_p): New function.
29274 (umips_output_save_restore): New function.
29275 (umips_load_store_pair_p_1): New function.
29276 (umips_load_store_pair_p): New function.
29277 (umips_output_load_store_pair_1): New function.
29278 (umips_output_load_store_pair): New function.
29279 (umips_movep_target_p) New function.
29280 (mips_prepare_pch_save): Add microMIPS support.
29281 * config/mips/mips.h (TARGET_COMPRESSION): New.
29282 (TARGET_CPU_CPP_BUILTINS): Update macro
29283 to use new compression flags and to support microMIPS.
29284 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
29285 (MIPS_ARCH_FLOAT_SPEC): Likewise.
29286 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
29287 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
29288 (ASM_SPEC): Support mmicromips and mno-micromips.
29289 (M16STORE_REG_P): New macro.
29290 (MIPS_CALL): Support TARGET_MICROMIPS.
29291 (MICROMIPS_J): New macro.
29292 (mips_base_mips16): Rename this...
29293 (mips_base_compression_flags): ...to this.
29294 (UMIPS_12BIT_OFFSET_P): New macro.
29295 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
29296 (MULTILIB_DIRNAMES): Likewise.
29298 2013-02-25 Tom de Vries <tom@codesourcery.com>
29300 PR rtl-optimization/56131
29301 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
29302 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
29303 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
29305 2013-02-25 Tobias Burnus <burnus@net-b.de>
29307 * doc/invoke.texi (-fsanitize=): Move from optimization
29308 to debugging options.
29310 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
29312 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
29314 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
29315 Alexander Monakov <amonakov@ispras.ru>
29317 PR middle-end/56077
29318 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
29319 flush pending lists also on non-jumps. Adjust comment.
29321 2013-02-24 Catherine Moore <clm@codesourcery.com>
29322 Maciej W. Rozycki <macro@codesourcery.com>
29323 Tom de Vries <tom@codesourcery.com>
29324 Nathan Sidwell <nathan@codesourcery.com>
29325 Iain Sandoe <iain@codesourcery.com>
29326 Nathan Froyd <froydnj@codesourcery.com>
29327 Chao-ying Fu <fu@mips.com>
29329 * doc/extend.texi: (micromips, nomicromips, nocompression):
29330 Document new function attributes.
29331 * doc/invoke.texi (minterlink-compressed, mmicromips,
29332 m14k, m14ke, m14kec): Document new options.
29333 (minterlink-mips16): Update documentation.
29334 * doc/md.texi (ZC, ZD): Document new constraints.
29335 * configure.ac (gcc_cv_as_micromips): Check if linker
29336 supports the .set micromips directive.
29337 * configure: Regenerate.
29338 * config.in: Regenerate.
29339 * config/mips/mips-tables.opt: Regenerate.
29340 * config/mips/micromips.md: New file.
29341 * constraints.md (ZC, AD): New constraints.
29342 * config/mips/predicates.md (movep_src_register): New predicate.
29343 (movep_src_operand): New predicate.
29344 (non_volatile_mem_operand): New predicate.
29345 * config/mips/mips.md (multimem): New type.
29346 (length): Differentiate between 17-bit and 18-bit branch offsets.
29347 (MOVEP1, MOVEP2): New mode iterator.
29348 (mov_<load>l): Use ZC constraint.
29349 (mov_<load>r): Likewise.
29350 (mov_<store>l): Likewise.
29351 (mov_<store>r): Likewise.
29352 (*branch_equality<mode>_inverted): Add microMIPS support.
29353 (*branch_equality<mode>): Likewise.
29354 (*jump_absolute): Likewise.
29355 (indirect_jump_<mode>): Likewise.
29356 (tablejump_<mode>): Likewise.
29357 (<optab>_internal): Likewise.
29358 (sibcall_internal): Likewise.
29359 (sibcall_value_internal): Likewise.
29360 (prefetch): Use constraint ZD.
29361 * config/mips/mips.opt (minterlink-compressed): New option.
29362 (minterlink-mips16): Now an alias for minterlink-compressed.
29363 (mmicromips): New option.
29364 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
29365 (compare_and_swap_12): Likewise.
29366 (sync_add<mode>): Likewise.
29367 (sync_<optab>_12): Likewise.
29368 (sync_old_<optab>_12): Likewise.
29369 (sync_new_<optab>_12): Likewise.
29370 (sync_nand_12): Likewise.
29371 (sync_old_nand_12): Likewise.
29372 (sync_new_nand_12): Likewise.
29373 (sync_sub<mode>): Likewise.
29374 (sync_old_add<mode>): Likewise.
29375 (sync_old_sub<mode>): Likewise.
29376 (sync_new_add<mode>): Likewise.
29377 (sync_new_sub<mode>): Likewise.
29378 (sync_<optab><mode>): Likewise.
29379 (sync_old_<optab><mode>): Likewise.
29380 (sync_new_<optab><mode>): Likewise.
29381 (sync_nand<mode>): Likewise.
29382 (sync_old_nand<mode>): Likewise.
29383 (sync_new_nand<mode>): Likewise.
29384 (sync_lock_test_and_set<mode>): Likewise.
29385 (test_and_set_12): Likewise.
29386 (atomic_compare_and_swap<mode>): Likewise.
29387 (atomic_exchange<mode>_llsc): Likewise.
29388 (atomic_fetch_add<mode>_llsc): Likewise.
29389 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
29390 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
29391 (umips_save_restore_pattern_p): Likewise.
29392 (umips_load_store_pair_p): Likewise.
29393 (umips_output_load_store_pair): Likewise.
29394 (umips_movep_target_p): Likewise.
29395 (umips_12bit_offset_address_p): Likewise.
29396 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
29397 (mips_base_mips16): Rename this...
29398 (mips_base_compression_flags): ...to this. Update all uses.
29399 (mips_attribute_table): Add micromips, nomicromips and nocompression.
29400 (mips_mips16_decl_p): Delete.
29401 (mips_nomips16_decl_p): Delete.
29402 (mips_get_compress_on_flags): New function.
29403 (mips_get_compress_off_flags): New function.
29404 (mips_get_compress_mode): New function.
29405 (mips_get_compress_on_name): New function.
29406 (mips_get_compress_off_name): New function.
29407 (mips_insert_attributes): Support multiple compression types.
29408 (mips_merge_decl_attributes): Likewise.
29409 (umips_12bit_offset_address_p): New function.
29410 (mips_start_function_definition): Emit .set micromips directive.
29411 (mips_call_may_need_jalx_p): New function.
29412 (mips_function_ok_for_sibcall): Add microMIPS support.
29413 (mips_print_operand_punctuation): Support short delay slots and
29415 (umips_swm_mask, umips_swm_encoding): New.
29416 (umips_build_save_restore): New function.
29417 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
29418 (was_mips16_p): Remove.
29419 (old_compression_mode): New.
29420 (mips_set_compression_mode): New function.
29421 (mips_set_current_function): Add microMIPS support.
29422 (mips_option_override): Likewise.
29423 (umips_save_restore_pattern_p): New function.
29424 (umips_output_save_restore): New function.
29425 (umips_load_store_pair_p_1): New function.
29426 (umips_load_store_pair_p): New function.
29427 (umips_output_load_store_pair_1): New function.
29428 (umips_output_load_store_pair): New function.
29429 (umips_movep_target_p) New function.
29430 (mips_prepare_pch_save): Add microMIPS support.
29431 * config/mips/mips.h (TARGET_COMPRESSION): New.
29432 (TARGET_CPU_CPP_BUILTINS): Update macro
29433 to use new compression flags and to support microMIPS.
29434 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
29435 (MIPS_ARCH_FLOAT_SPEC): Likewise.
29436 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
29437 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
29438 (ASM_SPEC): Support mmicromips and mno-micromips.
29439 (M16STORE_REG_P): New macro.
29440 (MIPS_CALL): Support TARGET_MICROMIPS.
29441 (MICROMIPS_J): New macro.
29442 (mips_base_mips16): Rename this...
29443 (mips_base_compression_flags): ...to this.
29444 (UMIPS_12BIT_OFFSET_P): New macro.
29445 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
29446 (MULTILIB_DIRNAMES): Likewise.
29448 2013-02-24 Jakub Jelinek <jakub@redhat.com>
29451 * target-globals.c (save_target_globals): For init_reg_sets and
29452 target_reinit remporarily set this_fn_optabs to this_target_optabs.
29454 2013-02-22 James Greenhalgh <james.greenhalgh@arm.com>
29456 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
29457 * config/aarch64/t-aarch64
29458 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
29460 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
29462 PR inline-asm/56148
29463 * lra-constraints.c (process_alt_operands): Reload operand
29464 conflicting with earlier clobber only if no more other conflicting
29467 2013-02-22 Jakub Jelinek <jakub@redhat.com>
29470 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
29471 if not linking a shared library.
29473 2013-02-22 Seth LaForge <sethml@google.com>
29475 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
29477 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
29479 * config/arm/arm.md (split for extendsidi): Update condition.
29480 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
29481 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
29482 (qhs_zextenddi_cstr): Likewise.
29484 2013-02-21 Jakub Jelinek <jakub@redhat.com>
29486 PR middle-end/56420
29487 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
29488 avoid signed wrapping.
29489 (expand_mult): Handle properly multiplication by
29490 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
29491 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
29492 in the compiler if coeff is HOST_WIDE_INT_MIN.
29493 (expand_divmod): Don't make ext_op1 static, change it's type to
29494 uhwi. Avoid undefined behavior in -INTVAL (op1).
29496 PR rtl-optimization/50339
29497 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
29499 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
29500 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
29501 into splitting_ashiftrt field.
29502 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
29504 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
29507 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
29509 PR middle-end/56108
29510 * trans-mem.c (execute_tm_mark): Do not expand transactions that
29511 are sure to go irrevocable.
29513 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
29515 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
29516 scalars are valid operands.
29518 2013-02-21 Martin Jambor <mjambor@suse.cz>
29520 PR tree-optimization/56310
29521 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
29522 only matching indices and non-negative final offsets.
29523 (intersect_aggregates_with_edge): Pass src_idx to
29524 agg_replacements_to_vector. Pass src_idx insstead of index to
29525 intersect_with_agg_replacements.
29527 2013-02-21 Martin Jambor <mjambor@suse.cz>
29529 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
29530 instead of hard-wired defaults.
29532 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
29534 * doc/invoke.texi (MIPS Options): Update documentation of the
29535 floating-point multiply-accumulate instruction restrictions.
29537 2013-02-21 Kostya Serebryany <kcc@google.com>
29539 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
29540 asan_shadow_offset on x86_64 linux.
29542 2013-02-21 Richard Biener <rguenther@suse.de>
29544 PR tree-optimization/56415
29546 2013-02-11 Richard Biener <rguenther@suse.de>
29548 PR tree-optimization/56273
29549 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
29552 2013-02-21 Jakub Jelinek <jakub@redhat.com>
29555 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
29558 PR inline-asm/56405
29559 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
29560 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
29562 2013-02-20 Jan Hubicka <jh@suse.cz>
29564 PR tree-optimization/56265
29565 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
29566 when target is referenced for first time.
29568 2013-02-20 Richard Biener <rguenther@suse.de>
29570 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
29571 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
29572 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
29573 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
29574 not return anything.
29575 (rename_ssa_copies): Do not remove unused locals.
29576 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
29577 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
29578 * passes.c (execute_function_todo): Do not schedule unused locals
29579 removal if cleanup_tree_cfg did something.
29580 * tree-ssa-live.c (remove_unused_locals): Dump statistics
29581 about the number of removed locals.
29583 2013-02-20 Richard Biener <rguenther@suse.de>
29585 PR tree-optimization/56398
29586 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
29588 2013-02-20 Martin Jambor <mjambor@suse.cz>
29590 PR tree-optimization/55334
29591 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
29592 restricted pointers to arrays.
29594 2013-02-20 Richard Biener <rguenther@suse.de>
29595 Jakub Jelinek <jakub@redhat.com>
29597 PR tree-optimization/56396
29598 * tree-ssa-ccp.c (n_const_val): New static variable.
29599 (get_value): Return NULL for SSA names we don't have a lattice
29601 (ccp_initialize): Initialize n_const_val.
29602 * tree-ssa-copy.c (n_copy_of): New static variable.
29603 (init_copy_prop): Initialize n_copy_of.
29604 (get_value): Return NULL_TREE for SSA names we don't have a
29607 2013-02-20 Martin Jambor <mjambor@suse.cz>
29609 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
29611 2013-02-20 Richard Biener <rguenther@suse.de>
29613 * genpreds.c (write_lookup_constraint): Do not compare first
29614 letter of the constraint again.
29616 2013-02-20 Richard Biener <rguenther@suse.de>
29618 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
29620 (get_use_iv_cost): Terminate hashtable walk when coming across
29623 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
29625 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
29626 reassociation for avx2 targets.
29628 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
29630 * config/microblaze/microblaze.c: microblaze_has_clz = 0
29631 Add version check for v8.10.a to enable microblaze_has_clz
29632 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
29633 version and TARGET_PATTERN_COMPARE check
29634 * config/microblaze/microblaze.md: New clzsi2 instruction
29636 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
29638 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
29639 function before branching.
29641 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
29643 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
29645 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
29647 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
29649 PR middle-end/55889
29650 * sel-sched.c: Include ira.h.
29651 (implicit_clobber_conflict_p): New function.
29652 (moveup_expr): Use it.
29653 * Makefile.in (sel-sched.o): Depend on ira.h.
29655 2013-02-19 Richard Biener <rguenther@suse.de>
29657 PR tree-optimization/56384
29658 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
29659 (vn_hash_type): Split out from ...
29660 (vn_hash_constant_with_type): ... here.
29661 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
29662 (vn_phi_eq): Compare types from vn_phi_s structure.
29663 (vn_phi_lookup): Populate vn_phi_s type.
29664 (vn_phi_insert): Likewise.
29666 2013-02-19 Jakub Jelinek <jakub@redhat.com>
29668 PR tree-optimization/56350
29669 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
29670 if haven't found reduction or nested cycle operand, rather than
29671 asserting we must find it.
29673 PR tree-optimization/56381
29674 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
29677 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
29678 Jakub Jelinek <jakub@redhat.com>
29681 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
29682 (swap_optab_enable): Same.
29683 (init_all_optabs): Use argument instead of global.
29684 * tree.h (struct tree_optimization_option): New field target_optabs.
29685 * expr.h (init_all_optabs): Add argument to prototype.
29686 (TREE_OPTIMIZATION_OPTABS): New.
29687 (save_optabs_if_changed): Protoize.
29688 * optabs.h: Declare this_fn_optabs.
29689 * optabs.c (save_optabs_if_changed): New.
29690 Declare this_fn_optabs.
29691 (init_optabs): Add argument to init_all_optabs() call.
29692 * function.c (invoke_set_current_function_hook): Handle per
29694 * function.h (struct function): New field optabs.
29695 * config/mips/mips.c (mips_set_mips16_mode): Handle when
29696 optimization_current_node has changed.
29697 * target-globals.h (save_target_globals_default_opts): Protoize.
29698 * target-globals.c (save_target_globals_default_opts): New.
29700 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29703 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
29704 registers %fr12 and %fr12R as call used.
29707 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
29708 and HImode, require all displacements to be an integer multiple of
29710 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
29711 only allow QImode and HImode when reload is in progress and strict is
29712 true. Likewise for symbolic addresses. Use base14_operand to check
29713 displacements in REG+BASE addresses.
29715 2013-02-18 Richard Biener <rguenther@suse.de>
29717 PR tree-optimization/56366
29718 * tree-vect-loop.c (get_initial_def_for_induction): Properly
29719 handle sign-conversion of outer-loop initial induction value.
29721 2013-02-18 Richard Biener <rguenther@suse.de>
29723 PR middle-end/56349
29724 * cfghooks.c (merge_blocks): If we merge a latch into another
29725 block adjust references to it.
29726 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
29727 (verify_loop_structure): Verify that a recorded latch is in fact
29730 2013-02-18 Richard Biener <rguenther@suse.de>
29732 PR tree-optimization/56321
29733 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
29734 order SSA name release and virtual operand unlinking.
29736 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
29738 * config/microblaze/microblaze.md (save_stack_block): Define.
29739 (restore_stack_block): Likewise.
29741 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
29743 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
29744 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
29745 * config/microblaze/microblaze.c (microblaze_option_override):
29746 Bail out early for PIC modes when target does not support PIC.
29748 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
29750 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
29751 Replace with a microblaze version.
29752 (microblaze_trampoline_init): Adapt for microblaze.
29753 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
29756 2013-02-16 Jakub Jelinek <jakub@redhat.com>
29757 Dodji Seketeli <dodji@redhat.com>
29760 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
29761 (instrument_mem_region_access): Do not forget to always put
29762 instrumentation of the of 'base' and 'base + len' in a "if (len !=
29763 0) statement, even for cases where either 'base' or 'base + len'
29764 are not instrumented -- because they have been previously
29765 instrumented. Simplify the logic by putting all the statements
29766 instrument 'base + len' inside a sequence, and then insert that
29767 sequence right before the current insertion point. Then, to
29768 instrument 'base + len', just get an iterator on that statement.
29769 And do not forget to update the pointer to iterator the function
29770 received as argument.
29772 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
29774 PR rtl-optimization/56348
29775 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
29777 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
29779 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
29780 (clean_graph_dump_file): Pass base to start_graph_dump.
29782 2013-02-14 Richard Henderson <rth@redhat.com>
29785 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
29787 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
29789 * collect2-aix.h: Define F_LOADONLY.
29791 2013-02-14 Richard Biener <rguenther@suse.de>
29794 * varasm.c (output_constant_def_1): Get the decl representing
29795 the constant as argument.
29796 (output_constant_def): Wrap output_constant_def_1.
29797 (make_decl_rtl): Use output_constant_def_1 with the decl
29798 representing the constant.
29799 (build_constant_desc): Optionally re-use a decl already
29800 representing the constant.
29801 (tree_output_constant_def): Adjust.
29803 2013-02-14 Dodji Seketeli <dodji@redhat.com>
29806 * asan.c (instrument_builtin_call): Really put the length of the
29807 second source argument into src1_len.
29809 2013-02-13 Jakub Jelinek <jakub@redhat.com>
29811 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
29812 argument. If it is false, don't create edge from then_bb to
29814 (insert_if_then_before_iter): Pass true to it.
29815 (build_check_stmt): Pass false to it.
29816 (transform_statements): Flush hash table only on extended basic
29817 block boundaries, rather than at the beginning of every bb.
29818 Don't flush hash table on nonfreeing_call_p calls.
29819 * tree-flow.h (nonfreeing_call_p): New prototype.
29820 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
29822 2013-02-13 David S. Miller <davem@davemloft.net>
29824 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
29826 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
29829 * ira.c (max_regno_before_ira): Move from ...
29831 (fix_reg_equiv_init): Use max_regno_before_ira instead of
29834 2013-02-13 Jakub Jelinek <jakub@redhat.com>
29836 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
29838 2013-02-13 Richard Biener <rguenther@suse.de>
29841 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
29842 globals in MEM_REFs.
29844 2013-02-13 Richard Biener <rguenther@suse.de>
29846 * loop-init.c (loop_optimizer_init): Clear loop state when
29847 re-initializing preserved loops.
29848 * loop-unswitch.c (unswitch_single_loop): Return whether
29849 we unswitched the loop. Do not verify loop state here.
29850 (unswitch_loops): When we unswitched a loop discover new loops.
29852 2013-02-13 Kostya Serebryany <kcc@google.com>
29854 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
29856 * sanitizer.def: Rename __asan_init to __asan_init_v1.
29858 2013-02-12 Dodji Seketeli <dodji@redhat.com>
29860 Avoid instrumenting duplicated memory access in the same basic block
29861 * Makefile.in (asan.o): Add new dependency on hash-table.h
29862 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
29863 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
29864 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
29865 (free_mem_ref_resources, has_mem_ref_been_instrumented)
29866 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
29867 (get_mem_ref_of_assignment): New functions.
29868 (get_mem_refs_of_builtin_call): Extract from
29869 instrument_builtin_call and tweak a little bit to make it fit with
29871 (instrument_builtin_call): Use the new
29872 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
29873 of is_gimple_builtin_call.
29874 (instrument_derefs, instrument_mem_region_access): Insert the
29875 instrumented memory reference into the hash table.
29876 (maybe_instrument_assignment): Renamed instrument_assignment into
29877 this, and change it to advance the iterator when instrumentation
29878 actually happened and return true in that case. This makes it
29879 homogeneous with maybe_instrument_assignment, and thus give a
29880 chance to callers to be more 'regular'.
29881 (transform_statements): Clear the memory reference hash table
29882 whenever we enter a new BB, when we cross a function call, or when
29883 we are done transforming statements. Use
29884 maybe_instrument_assignment instead of instrumentation. No more
29885 need to special case maybe_instrument_assignment and advance the
29886 iterator after calling it; it's now handled just like
29887 maybe_instrument_call. Update comment.
29889 2013-02-13 Richard Biener <rguenther@suse.de>
29891 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
29892 Fix loop discovery code.
29894 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
29896 PR inline-asm/56148
29897 * lra-constraints.c (process_alt_operands): Match early clobber
29898 operand with itself. Check conflicts with earlyclobber only if
29899 the operand is not reloaded. Prefer to reload conflicting operand
29900 if earlyclobber and matching operands are the same.
29902 2013-02-12 Richard Biener <rguenther@suse.de>
29905 * lto-streamer-out.c (write_symbol): Do not output symbols
29906 for hard register variables.
29908 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
29911 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
29912 (umulsidi3_insn, mulsidi3_insn): New insns.
29914 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
29916 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
29917 (struct tune_params): Add vec_costs field.
29918 * config/arm/arm.c (arm_builtin_vectorization_cost)
29919 (arm_add_stmt_cost): New functions.
29920 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
29921 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
29922 (arm_default_vec_cost): New struct of type cpu_vec_costs.
29923 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
29924 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
29925 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
29926 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
29928 2013-02-12 Richard Biener <rguenther@suse.de>
29931 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
29932 decls again if possible.
29934 2013-02-12 Richard Biener <rguenther@suse.de>
29936 PR middle-end/56288
29937 * tree-ssa.c (verify_ssa_name): Fix check, move
29938 SSA_NAME_IN_FREE_LIST check up.
29940 2013-02-12 Jakub Jelinek <jakub@redhat.com>
29941 Steven Bosscher <steven@gcc.gnu.org>
29943 PR rtl-optimization/56151
29944 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
29945 equal to op0 or op1, and last_insn pattern is CODE operation
29946 with MEM dest and one of the operands matches that MEM.
29948 2013-02-11 Sriraman Tallam <tmsriram@google.com>
29950 * doc/extend.texi: Document Function Multiversioning and "default"
29951 parameter string to target attribute.
29952 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
29953 target attribute parameter is "default".
29954 (ix86_compare_version_priority): Remove checks for target attribute.
29955 (ix86_mangle_function_version_assembler_name): Change error to sorry.
29956 Remove check for target attribute equal to NULL. Add assert.
29957 (ix86_generate_version_dispatcher_body): Change error to sorry.
29959 2013-02-11 Iain Sandoe <iain@codesourcery.com>
29960 Jack Howarth <howarth@bromo.med.uc.edu>
29961 Patrick Marlier <patrick.marlier@gmail.com>
29964 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
29965 define ENDFILE_SPEC as TM_DESTRUCTOR.
29966 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
29968 2013-02-11 Alexander Potapenko <glider@google.com>
29969 Jack Howarth <howarth@bromo.med.uc.edu>
29970 Jakub Jelinek <jakub@redhat.com>
29973 * config/darwin.c (cdtor_record): Rename ctor_record.
29974 (sort_cdtor_records): Rename sort_ctor_records.
29975 (finalize_dtors): New routine to sort destructors by
29976 priority before use in assemble_integer.
29977 (machopic_asm_out_destructor): Use finalize_dtors if needed.
29979 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
29981 PR rtl-optimization/56275
29982 * simplify-rtx.c (avoid_constant_pool_reference): Check that
29983 offset is non-negative and less than cmode size before
29984 calling simplify_subreg.
29986 2013-02-11 Richard Biener <rguenther@suse.de>
29988 PR tree-optimization/56264
29989 * cfgloop.h (fix_loop_structure): Adjust prototype.
29990 * loop-init.c (fix_loop_structure): Return the number of
29991 newly discovered loops.
29992 * tree-cfgcleanup.c (repair_loop_structures): When new loops
29993 are discovered, do a full loop-closed SSA rewrite.
29995 2013-02-11 Richard Biener <rguenther@suse.de>
29997 PR tree-optimization/56273
29998 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
30000 (check_array_ref): Fix missing newline in dumps.
30001 (search_for_addr_array): Likewise.
30003 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
30005 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
30007 2013-02-09 Jakub Jelinek <jakub@redhat.com>
30010 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
30012 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
30014 PR rtl-optimization/56246
30015 * lra-constraints.c (simplify_operand_subreg): Try to reuse
30017 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
30018 constraints are satisfied.
30020 2013-02-08 Jeff Law <law@redhat.com>
30023 * emit-rtl.c (reg_is_parm_p): New function.
30024 * regs.h (reg_is_parm_p): New prototype.
30025 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
30026 callee-clobbered registers.
30028 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
30031 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
30032 If there is no implicit builtin declaration, just return NULL.
30034 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
30036 * config/i386/sse.md (FMAMODEM): New mode iterator.
30037 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
30038 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
30040 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
30042 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
30043 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
30044 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
30046 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
30048 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
30049 (microblaze*-*-elf): Likewise.
30050 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
30052 * config/microblaze/microblaze-c.c: Add builtin defines for
30053 _LITTLE_ENDIAN and _BIG_ENDIAN.
30054 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
30055 add to TARGET_DEFAULT flags.
30056 Expand ASM_SPEC and LINK_SPEC.
30057 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
30058 * config/microblaze/microblaze.md: Update extendsidi2 and
30059 movdi_internal instructions to use low-order / high-order reg
30061 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
30062 options and inversemask / mask of LITTLE_ENDIAN.
30063 * config/microblaze/t-microblaze: Expand multilib options to
30064 include mlittle-endian (le) and update exceptions patterns.
30066 2013-02-08 Jakub Jelinek <jakub@redhat.com>
30068 PR rtl-optimization/56195
30069 * lra-constraints.c (get_reload_reg): Don't reuse regs
30070 if they have smaller mode than requested, if they have
30071 wider mode than requested, try to return a SUBREG.
30073 PR tree-optimization/56250
30074 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
30075 if type is unsigned and code isn't MULT_EXPR.
30077 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
30079 PR tree-optimization/56064
30080 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
30081 bits according to mode.
30082 * fixed-value.h (fixed_from_double_int)
30083 (const_fixed_from_double_int): Adjust comments.
30085 2013-02-08 Richard Biener <rguenther@suse.de>
30088 * lto-streamer.h (struct data_in): Remove current_file, current_line
30089 and current_col members.
30090 * lto-streamer-out.c (lto_output_location): Stream changed bits
30091 en-block for efficiency.
30092 * lto-streamer-in.c (clear_line_info): Remove.
30093 (lto_input_location): Cache current file, line and column
30094 globally via local statics. Read changed bits en-block.
30095 (input_function): Do not call clear_line_info.
30096 (lto_read_body): Likewise.
30097 (lto_input_toplevel_asms): Likewise.
30099 2013-02-08 Michael Matz <matz@suse.de>
30101 PR tree-optimization/52448
30102 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
30103 (nt_call_phase): New static.
30104 (add_or_mark_expr): Only mark accesses with newer phase than any
30106 (nonfreeing_call_p): New.
30107 (nt_init_block): Update nt_call_phase, mark blocks as visited.
30108 (nt_fini_block): Keep blocks marked as visited.
30109 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
30111 2013-02-08 Richard Biener <rguenther@suse.de>
30113 * ira.c (ira): Free broken dominator information.
30115 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
30117 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
30119 2013-02-08 Marek Polacek <polacek@redhat.com>
30121 * cfgloop.c (verify_loop_structure): Add more checking of headers.
30123 2013-02-08 Richard Biener <rguenther@suse.de>
30125 PR middle-end/56181
30126 * cfgloop.h (flow_loops_find): Adjust.
30127 (bb_loop_header_p): Declare.
30128 * cfgloop.c (bb_loop_header_p): New function split out from ...
30129 (flow_loops_find): ... here. Adjust function signature,
30130 support incremental loop structure update.
30131 (verify_loop_structure): Cleanup. Verify a loop is a loop.
30132 * cfgloopmanip.c (fix_loop_structure): Move ...
30133 * loop-init.c (fix_loop_structure): ... here.
30134 (apply_loop_flags): Split out from ...
30135 (loop_optimizer_init): ... here.
30136 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
30137 in incremental mode, only remove dead loops here.
30139 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
30142 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
30143 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
30144 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
30145 (*round<mode>3.libgcc): New insns for fixed-modes.
30146 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
30147 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
30148 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
30149 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
30150 implementations. Define to __builtin_avr_absFX,
30151 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
30152 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
30153 __builtin_avr_countlsFX, respectively.
30154 * config/avr/avr-c.c (target.h): Include it.
30155 (enum avr_builtin_id): New enum.
30156 (avr_resolve_overloaded_builtin): New static function.
30157 (avr_register_target_pragmas): Use it to set
30158 targetm.resolve_overloaded_builtin.
30159 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
30160 tree nodes used by DEF_BUILTIN.
30161 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
30162 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
30163 <AVR_BUILTIN_xxBITS>: Same.
30165 2013-02-08 Richard Biener <rguenther@suse.de>
30167 * cfgloop.c (verify_loop_structure): Properly handle
30168 a loop exiting to another loop header.
30169 * ira-int.h (ira_loops): Remove.
30170 * ira.c (ira_loops): Remove.
30171 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
30172 (do_reload): Use loop_optimizer_finalize.
30173 * ira-build.c (create_loop_tree_nodes): Use get_loops and
30174 number_of_loops to access the loop tree.
30175 (more_one_region_p): Likewise.
30176 (finish_loop_tree_nodes): Likewise.
30177 (rebuild_regno_allocno_maps): Likewise.
30178 (mark_loops_for_removal): Likewise.
30179 (mark_all_loops_for_removal): Likewise.
30180 (remove_unnecessary_regions): Likewise.
30181 (ira_build): Likewise.
30182 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
30184 2013-02-08 Richard Biener <rguenther@suse.de>
30186 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
30187 * ipa-pure-const.c (analyze_function): Avoid calling
30188 mark_irreducible_loops twice.
30189 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
30191 2013-02-07 David S. Miller <davem@davemloft.net>
30193 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
30195 * var-tracking.c (vt_add_function_parameter): Test the presence of
30196 HAVE_window_save properly and do not remap argument registers when
30197 we have a leaf function.
30199 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
30202 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
30204 * config/i386/i386.c (ix86_print_operand): Ditto.
30206 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
30208 * lra-constraints.c (process_alt_operands): Fix recently added comment.
30210 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
30212 PR rtl-optimization/56225
30213 * lra-constraints.c (process_alt_operands): Check that reload hard
30214 reg can hold value for strict_low_part.
30216 2013-02-07 Jakub Jelinek <jakub@redhat.com>
30219 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
30220 dwarf2out_end_function.
30221 (in_first_function_p, maybe_at_text_label_p,
30222 first_loclabel_num_not_at_text_label): New variables.
30223 (dwarf2out_var_location): In the first function find out
30224 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
30225 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
30228 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
30230 PR rtl-optimization/56178
30231 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
30232 SUBREG of a register. Tidy up related block of code.
30233 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
30234 note if the source is a register or a SUBREG of a register.
30236 2013-02-07 Jakub Jelinek <jakub@redhat.com>
30239 * config/rs6000/rs6000.md (ptrm): New mode attr.
30240 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
30241 call_value_indirect_aix<pttrsize>,
30242 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
30245 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
30247 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
30248 if -bnortl. Convert to strcmp and strncmp.
30250 2013-02-07 Alan Modra <amodra@gmail.com>
30253 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
30254 addresses won't wrap when offsetting.
30255 (rs6000_secondary_reload): Provide secondary reloads needed for
30256 wrapping LO_SUM addresses.
30258 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
30260 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
30261 MACH, just __MACH__.
30263 2013-02-06 Richard Biener <rguenther@suse.de>
30265 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
30266 instead of calling fix_loop_structure.
30268 2013-02-06 Jakub Jelinek <jakub@redhat.com>
30270 PR middle-end/56217
30271 * omp-low.c (use_pointer_for_field): Return false if
30272 lower_send_shared_vars doesn't generate any copy-out code.
30274 2013-02-06 Tom de Vries <tom@codesourcery.com>
30276 PR rtl-optimization/56131
30277 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
30278 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
30279 of the label is NULL. Add comment.
30281 2013-02-05 Jakub Jelinek <jakub@redhat.com>
30283 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
30286 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
30287 (STATIC_LIBTSAN_LIBS): Likewise.
30288 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
30289 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
30290 is defined, don't add anything else beyond that.
30291 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
30292 (LINK_COMMAND_SPEC): Use them.
30294 PR tree-optimization/56205
30295 * tree-stdarg.c (check_all_va_list_escapes): Return true if
30296 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
30297 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
30299 2013-02-05 Richard Biener <rguenther@suse.de>
30301 PR tree-optimization/53342
30302 PR tree-optimization/53185
30303 * tree-vectorizer.h (vect_check_strided_load): Remove.
30304 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
30305 not disallow peeling for vectorized strided loads.
30306 (vect_check_strided_load): Make static and simplify.
30307 (vect_analyze_data_refs): Adjust.
30308 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
30309 correctly when vectorizing strided loads.
30311 2013-02-05 Richard Biener <rguenther@suse.de>
30313 * doc/install.texi: Refer to ISL, not PPL.
30315 2013-02-05 Jan Hubicka <jh@suse.cz>
30317 PR tree-optimization/55789
30318 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
30320 2013-02-05 Jan Hubicka <jh@suse.cz>
30322 PR tree-optimization/55789
30323 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
30324 the dead call anyway.
30326 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
30329 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
30331 2013-02-04 Alexander Potapenko <glider@google.com>
30332 Jack Howarth <howarth@bromo.med.uc.edu>
30333 Jakub Jelinek <jakub@redhat.com>
30336 * config/darwin.c (sort_ctor_records): Stabilized qsort
30337 on constructor priority by using original position.
30338 (finalize_ctors): New routine to sort constructors by
30339 priority before use in assemble_integer.
30340 (machopic_asm_out_constructor): Use finalize_ctors if needed.
30342 2013-02-04 Jakub Jelinek <jakub@redhat.com>
30345 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
30346 about visibility on artificial decls.
30347 * config/sol2.c (solaris_assemble_visibility): Likewise.
30349 2013-02-04 Kai Tietz <ktietz@redhat.com>
30352 * config/i386/i386.c (function_value_ms_64): Add additional valtype
30353 argument and improve checking of return-argument types for 16-byte
30355 (ix86_function_value_1): Add additional valtype argument on call
30356 of function_value_64.
30357 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
30358 handling infunction_value_64 function.
30360 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
30362 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
30364 2013-02-04 Richard Biener <rguenther@suse.de>
30366 PR tree-optimization/56188
30367 * tree-ssa-structalias.c (label_visit): Consider case with
30368 initially non-empty points-to set.
30369 (perform_var_substitution): Dump node mapping and clean up.
30371 2013-02-04 Richard Guenther <rguenther@suse.de>
30374 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
30375 node prevail as last resort.
30376 (lto_symtab_merge_decls): Remove guard on LTRANS here.
30377 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
30379 2013-02-04 Richard Biener <rguenther@suse.de>
30381 PR tree-optimization/56113
30382 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
30384 (equiv_class_lookup_or_add): ... this.
30385 (label_visit): Adjust and fix error in previous patch.
30386 (perform_var_substitution): Adjust.
30388 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
30390 * config/sh/divtab.c: Fix formatting and comments throughout the file.
30391 * config/sh/sh4-300.md: Likewise.
30392 * config/sh/sh4a.md: Likewise.
30393 * config/sh/constraints.md: Likewise.
30394 * config/sh/sh.md: Likewise.
30395 * config/sh/netbsd-elf.h: Likewise.
30396 * config/sh/predicates.md: Likewise.
30397 * config/sh/sh-protos.h: Likewise.
30398 * config/sh/ushmedia.h: Likewise.
30399 * config/sh/linux.h: Likewise.
30400 * config/sh/sh.c: Likewise.
30401 * config/sh/superh.h: Likewise.
30402 * config/sh/elf.h: Likewise.
30403 * config/sh/sh4.md: Likewise.
30404 * config/sh/sh.h: Likewise.
30406 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
30408 * config/pa/constraints.md: Adjust unused letters. Change "T"
30409 constraint to match_test floating_point_store_memory_operand().
30410 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
30411 (base14_operand): New.
30412 (floating_point_store_memory_operand): New.
30413 (integer_store_memory_operand): Revise to use base14_operand and
30414 reg_plus_base_memory_operand.
30415 (move_dest_operand): Allow symbolic_memory_operands.
30416 (symbolic_memory_operand): Check for LO_SOM.
30417 (symbolic_operand): Change default case to break.
30418 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
30419 CONST_DOUBLE values to be reloaded by putting them into memory when
30420 the destination is a floating point register.
30421 (movdf): Remove code to handle CONST_DOUBLE.
30423 (reload_indf_r1): New.
30424 (reload_insf_r1): New.
30425 Consistently use "Q" and "T" constraints with integer and floating
30426 point move instructions, respectively.
30427 (movdi): Remove FAIL.
30428 Change predicate for source operand unamed DImode move from
30429 general_operand to move_src_operand.
30430 (umulsidi3): Change predicate for destination operand to
30432 Likewise for similar unamed patterns.
30433 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
30434 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
30435 (hppa_legitimize_address): Simplify mask calculation.
30436 (pa_emit_move_sequence): Revised handling of secondary reloads from
30437 REG+D addresses for floating point loads and stores. Directly handle
30438 loading CONST0_RTX (mode) to a floating point register.
30439 (pa_secondary_reload): Handle reloading DF and SFmode constant values
30440 to floating point registers. Don't restrict secondary reloads to
30441 floating point registers to integer modes. Revise some comments and
30443 (TARGET_LEGITIMATE_ADDRESS_P): Define.
30444 (pa_legitimate_address_p): New.
30445 (pa_legitimize_reload_address): New.
30446 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
30447 (STRICT_REG_OK_FOR_BASE_P): New.
30448 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
30449 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
30451 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
30452 Andrew Dixie <andrewd@gentrack.com>
30454 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
30457 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
30459 * expmed.c (extract_bit_field_1): Pass the full width of the
30460 structure to get_best_reg_extraction_insn.
30462 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
30465 * configure.ac (use_cxa_atexit): Add AIX.
30466 * configure: Regenerate.
30468 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
30470 2013-02-01 Jakub Jelinek <jakub@redhat.com>
30473 * final.c (need_profile_function): New variable.
30474 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
30475 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
30476 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
30477 notes, targetm.asm_out.function_prologue doesn't emit anything,
30478 HAVE_prologue and profiler should be emitted before prologue,
30479 set need_profile_function instead of emitting it.
30480 (final_scan_insn): If need_profile_function, emit
30481 profile_function on the first NOTE_INSN_BASIC_BLOCK or
30482 NOTE_INSN_FUNCTION_BEG note.
30484 2013-02-01 Richard Henderson <rth@redhat.com>
30486 * config/rs6000/rs6000.md (smulditi3): New.
30489 * config/alpha/alpha.md (umulditi3): New.
30491 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
30493 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
30494 (ASM_OUTPUT_ALIGNED_LOCAL): New.
30496 2013-02-01 Richard Biener <rguenther@suse.de>
30498 PR tree-optimization/56113
30499 * tree-ssa-structalias.c (label_visit): Reduce work for
30500 single-predecessor nodes.
30502 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
30504 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
30505 range isn't testing for zero.
30507 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
30509 PR middle-end/56113
30510 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
30512 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
30513 Nick Clifton <nickc@redhat.com>
30515 * config/v850/constraints.md (Q): Define as a memory constraint.
30516 * config/v850/predicates.md (label_ref_operand): New predicate.
30517 (e3v5_shift_operand): New predicate.
30518 (ior_operator): New predicate.
30519 * config/v850/t-v850: Add e3v5 multilib.
30520 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
30521 (v850_gen_movdi): Prototype.
30522 * config/v850/v850.c: Add support for e3v5 architecture.
30523 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
30525 (construct_save_jarl): Add e3v5 long JARL support.
30526 (v850_adjust_insn_length): New function. Adjust length of call
30527 insns when using e3v5 instructions.
30528 (v850_gen_movdi): New function: Generate instructions to move a
30530 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
30531 (CPP_SPEC): Define __v850e3v5__ as appropriate.
30532 (TARGET_USE_FPU): Enable for e3v5.
30533 (CONST_OK_FOR_W): New macro.
30534 (ADJUST_INSN_LENGTH): Define.
30535 * config/v850/v850.md (UNSPEC_LOOP): Define.
30536 (attr cpu): Add v850e3v5.
30537 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
30538 (movdi): New pattern.
30539 (movdi_internal): New pattern.
30540 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
30541 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
30542 (cstoresf4): Likewise.
30543 (cstoredf4): Likewise.
30544 (insv): New pattern.
30545 (rotlso3_a): New pattern.
30546 (rotlsi3_b): New pattern
30547 (rotlsi3_v850e3v5): New pattern.
30548 (doloop_begin): New pattern.
30549 (fix_loop_counter): New pattern.
30550 (doloop_end): New pattern.
30551 (branch_normal): Add e3v5 long branch support.
30552 (branch_invert): Likewise.
30553 (branch_z_normal): Likewise.
30554 (branch_z_invert): Likewise.
30555 (branch_nz_normal): Likewise.
30556 (branch_nz_invert): Likewise.
30557 (call_internal_short): Add e3v5 register-indirect JARL support.
30558 (call_internal_long): Likewise.
30559 (call_value_internal_short): Likewise.
30560 (call_value_internal_long): Likewise.
30561 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
30562 (mloop): New option.
30563 * config.gcc: Add support for configuring v840e3v5 target.
30564 * doc/invoke.texi: Document new v850 specific command line options.
30566 2013-01-31 Paul Koning <ni1d@arrl.net>
30570 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
30571 children if parent is a class.
30572 (prune_unused_types_prune): Don't add DW_AT_declaration.
30574 2013-01-31 Richard Biener <rguenther@suse.de>
30576 PR tree-optimization/56157
30577 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
30578 match up operand with SLP child.
30580 2013-01-31 Jason Merrill <jason@redhat.com>
30583 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
30584 parameters the first time.
30585 (gen_scheduled_generic_parms_dies): Check completeness here.
30587 2013-01-31 Richard Biener <rguenther@suse.de>
30589 PR middle-end/53073
30590 * common.opt (faggressive-loop-optimizations): New flag,
30591 enabled by default.
30592 * doc/invoke.texi (faggressive-loop-optimizations): Document.
30593 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
30594 infer_loop_bounds_from_undefined by it.
30596 2013-01-31 Richard Biener <rguenther@suse.de>
30598 PR tree-optimization/56150
30599 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
30600 visit virtual operands.
30601 (find_uses_to_rename_bb): Likewise.
30603 2013-01-31 Richard Biener <rguenther@suse.de>
30605 PR tree-optimization/56150
30606 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
30607 mixed store non-store stmts.
30609 2013-01-30 Jakub Jelinek <jakub@redhat.com>
30612 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
30613 LIBASAN_EARLY_SPEC is defined.
30614 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
30615 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
30617 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
30620 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
30621 invalid args instead of ICEing on it.
30622 (ix86_valid_target_attribute_tree): Return error_mark_node if
30623 ix86_valid_target_attribute_inner_p failed.
30624 (ix86_valid_target_attribute_p): Return false only if
30625 ix86_valid_target_attribute_tree returned error_mark_node. Allow
30626 target("default") attribute.
30627 (sorted_attr_string): Change argument from const char * to tree,
30628 merge in all target attribute arguments rather than just one.
30629 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
30630 instead of free. Avoid using strcat.
30631 (ix86_mangle_function_version_assembler_name): Mangle
30632 target("default") as if no target attribute is present. Adjust
30633 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
30634 instead of xmalloc and XDELETEVEC instead of free.
30635 (ix86_function_versions): Don't return true if one of the decls
30636 doesn't have target attribute. If they don't and one of the decls
30637 is DECL_FUNCTION_VERSIONED, report an error. Adjust
30638 sorted_attr_string caller. Use XDELETEVEC instead of free.
30639 (ix86_supports_function_versions): Remove.
30640 (make_name): Fix up formatting.
30641 (make_dispatcher_decl): Remove resolver_name and its initialization.
30642 Avoid leaking memory.
30643 (is_function_default_version): Return true if there is
30644 target("default") attribute rather than no target attribute at all.
30645 (make_resolver_func): Avoid leaking memory.
30646 (ix86_generate_version_dispatcher_body): Likewise.
30647 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
30648 * target.def (supports_function_versions): Remove.
30649 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
30650 * doc/tm.texi: Regenerated.
30652 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
30654 PR rtl-optimization/56144
30655 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
30656 for values with side effects.
30658 2013-01-30 Richard Biener <rguenther@suse.de>
30660 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
30661 (sparseset_pop): Likewise.
30662 * cfganal.c (compute_idf): Likewise. Increase work-stack size
30663 to be able to use quick_push in the worker loop.
30665 2013-01-30 Marek Polacek <polacek@redhat.com>
30667 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
30669 2013-01-30 Richard Biener <rguenther@suse.de>
30672 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
30674 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
30676 PR tree-optimization/56064
30677 * fixed-value.c (fixed_from_double_int): New function.
30678 * fixed-value.h (fixed_from_double_int): New prototype.
30679 (const_fixed_from_double_int): New static inline function.
30680 * fold-const.c (native_interpret_fixed): New static function.
30681 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
30682 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
30683 (native_encode_fixed): New static function.
30684 (native_encode_expr) <FIXED_CST>: Use it.
30685 (native_interpret_int): Move double_int worker code to...
30686 * double-int.c (double_int::from_buffer): ...this new static method.
30687 * double-int.h (double_int::from_buffer): Prototype it.
30689 2013-01-30 Richard Biener <rguenther@suse.de>
30691 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
30692 New pointer-map and obstack.
30693 (init_alias_vars): Allocate pointer-map and obstack.
30694 (delete_points_to_sets): Free them.
30695 (find_what_var_points_to): Cache result.
30696 (find_what_p_points_to): Adjust for changed interface of
30697 find_what_var_points_to.
30698 (compute_points_to_sets): Likewise.
30699 (ipa_pta_execute): Likewise.
30701 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
30703 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
30704 * configure: Regenerate.
30705 * config.in: Regenerate.
30706 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
30707 #nobits/#progbits if supported.
30709 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
30712 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
30713 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
30714 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
30716 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
30718 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
30719 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
30721 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
30723 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
30724 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
30726 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
30728 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
30730 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
30731 * config/arm/cortex-a7.md: New bypasses using
30732 arm_mac_accumulator_is_result.
30734 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
30736 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
30737 (cortex_a7_neon_mla): Likewise.
30738 (cortex_a7_fpfmad): New reservation.
30739 (cortex_a7_fpmacs): Use ffmas and update required units.
30740 (cortex_a7_fpmuld): Update required units and latency.
30741 (cortex_a7_fpmacd): Likewise.
30742 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
30743 (cortex_a7_neon). Likewise.
30744 (bypass) Update participating units.
30746 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
30748 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
30749 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
30751 * config/arm/vfp11.md (vfp_farith): Use ffmas.
30752 (vfp_fmul): Use ffmad.
30753 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
30754 (cortex_r4_fmacd): Use ffmad.
30755 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
30756 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
30757 (cortex_a9_fmacd): Use ffmad.
30758 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
30759 (cortex_a8_vfp_macd): Use ffmad.
30760 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
30761 (cortex_a5_fpmacd): Use ffmad.
30762 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
30763 (cortex_a15_vfp_macd): Use ffmad.
30764 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
30766 2013-01-29 Jason Merrill <jason@redhat.com>
30769 * varasm.c (default_assemble_visibility): Don't warn about
30770 visibility on artificial decls.
30772 2013-01-29 Richard Biener <rguenther@suse.de>
30774 PR tree-optimization/56113
30775 * tree-ssa-structalias.c (equiv_class_lookup): Also return
30777 (label_visit): Free duplicate bitmaps and record the leader instead.
30778 (perform_var_substitution): Adjust.
30780 2013-01-29 Richard Biener <rguenther@suse.de>
30782 PR tree-optimization/55270
30783 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
30784 the CFG, schedule loops for fixup.
30786 2013-01-29 Nick Clifton <nickc@redhat.com>
30788 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
30791 2013-01-28 Leif Ekblad <leif@rdos.net>
30793 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
30794 * config/i386/i386.h (TARGET_RDOS): New macro.
30795 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
30796 * config/i386/i386.c (ix86_option_override_internal): For 64bit
30797 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
30798 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
30799 DEFAULT_LARGE_SECTION_THRESHOLD.
30800 * config/i386/i386.md (R14_REG, R15_REG): New constants.
30801 * config/i386/rdos.h: New file.
30802 * config/i386/rdos64.h: New file.
30804 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
30807 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
30810 2013-01-28 Jakub Jelinek <jakub@redhat.com>
30812 PR rtl-optimization/56117
30813 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
30814 call cselib_lookup_from_insn on the MEM before calling
30815 add_insn_mem_dependence.
30817 2013-01-28 Richard Biener <rguenther@suse.de>
30819 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
30820 to a stmt that didn't have one.
30821 (copy_phis_for_bb): Likewise for PHI arguments.
30822 (copy_debug_stmt): Likewise for debug stmts.
30824 2013-01-28 Richard Biener <rguenther@suse.de>
30826 PR tree-optimization/56034
30827 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
30828 (partition_builtin_p): Adjust.
30829 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
30830 it is the last partition.
30831 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
30832 up the vertex for the definition.
30833 (classify_partition): Classify whether a partition is a
30834 PKIND_REDUCTION, thus has uses outside of the loop.
30835 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
30836 Merge all PKIND_REDUCTION partitions into the last partition.
30837 (tree_loop_distribution): Seed partitions from reductions as well.
30839 2013-01-28 Jakub Jelinek <jakub@redhat.com>
30841 PR tree-optimization/56125
30842 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
30843 pow(x,c) into sqrt(x) * powi(x, n/2) or
30844 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
30845 optimizing for size.
30846 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
30847 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
30850 PR tree-optimization/56094
30851 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
30852 to UNKNOWN_LOCATION while gimplifying expr.
30854 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
30857 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
30858 operand 0 in movabs insn template for -masm=intel asm alternative.
30859 (*movabs<mode>_2): Ditto for operand 1.
30861 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
30864 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
30867 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
30869 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
30872 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
30874 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
30875 Undef to avoid warning.
30877 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
30879 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
30880 * configure: Regenerate.
30882 2013-01-25 Jakub Jelinek <jakub@redhat.com>
30884 PR tree-optimization/56098
30885 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
30886 for stmts with volatile ops.
30887 (cond_store_replacement): Don't optimize if assign has volatile ops.
30888 (cond_if_else_store_replacement_1): Don't optimize if either
30889 then_assign or else_assign have volatile ops.
30890 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
30893 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
30895 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
30897 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
30899 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
30900 missing ':' in asm example.
30902 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
30904 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
30905 entries into lane and laneq entries.
30906 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
30907 Remove AdvSIMD scalar modes.
30908 (aarch64_sq<r>dmulh_laneq<mode>): New.
30909 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
30911 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
30912 builtin implementations to relfect changes in RTL in aarch64-simd.md.
30913 * config/aarch64/iterators.md (VCOND): New.
30916 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
30919 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
30920 Add NULL LIBNAME argument to existing definitions.
30921 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
30922 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
30923 * config/avr/avr.c (DEF_BUILTIN): Same.
30924 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
30925 (avr_expand_builtin): Expand to a vanilla call if a libgcc
30926 implementation is available (DECL_ASSEMBLER_NAME is set).
30927 (avr_fold_absfx): New static function.
30928 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
30929 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
30930 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
30931 AVR_BUILTIN_ABSLLK.
30932 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
30933 (abshk, absk, abslk, absllk): Provide as static inline functions.
30935 2013-01-25 Marek Polacek <polacek@redhat.com>
30937 PR tree-optimization/56035
30938 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
30940 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
30942 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
30943 (*movtf_internal_rex64): Add (!o,C) alternative
30944 (*movxf_internal_rex64): Ditto.
30945 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
30947 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
30949 * doc/invoke.texi: fix typo.
30950 * doc/objc.texi: fix typo.
30952 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
30954 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
30955 for the first two alternatives.
30957 2013-01-24 Diego Novillo <dnovillo@google.com>
30959 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
30960 (ggc-zone.o): Remove.
30961 * configure.ac: Remove option --with-gc.
30962 * configure: Re-generate.
30963 * doc/install.texi: Remove documentation for --with-gc.
30964 * gengtype.c (write_enum_defn): Remove. Update all users.
30965 (write_Types_process_field): Remove generation of gt_e_* argument.
30966 (output_type_enum): Remove. Update all users.
30967 (write_enum_defn): Remove. Update all users.
30968 (enum alloc_zone): Remove. Update all users.
30969 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
30970 * ggc-common.c (ggc_splay_alloc): Remove first argument.
30971 Update all callers.
30972 (struct ptr_data): Remove field TYPE. Update all users.
30973 (gt_pch_note_object): Remove argument TYPE. Update all users.
30974 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
30976 * ggc-none.c (ggc_alloc_typed_stat): Remove.
30977 (struct alloc_zone): Remove.
30978 (ggc_internal_alloc_zone_stat): Remove.
30979 (ggc_internal_cleared_alloc_zone_stat): Remove.
30980 * ggc-page.c (ggc_alloc_typed_stat): Remove.
30981 (ggc_pch_count_object): Remove last argument. Update all users.
30982 (ggc_pch_alloc_object): Remove last argument. Update all users.
30983 (struct alloc_zone): Remove.
30984 * ggc-zone.c: Remove.
30985 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
30986 (struct alloc_zone): Remove.
30987 (ggc_alloc_typed_stat): Remove.
30988 (ggc_alloc_typed): Remove.
30989 (ggc_splay_alloc): Remove first argument.
30990 (rtl_zone): Remove. Update all users.
30991 (tree_zone): Remove. Update all users.
30992 (tree_id_zone): Remove. Update all users.
30993 (ggc_internal_zone_alloc_stat): Remove. Update all users.
30994 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
30995 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
30996 * tree-ssanames.c: Remove references to zone allocator in comments.
30998 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
31000 * config/avr/avr.c (avr_out_fract): Make register numbers that
31001 might be outside of source operand signed.
31003 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
31005 * config/i386/constraints.md (Yf): New constraint.
31006 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
31007 of f constraint to conditionaly disable x87 register preferences.
31008 (*movdf_internal): Ditto.
31009 (*movsf_internal): Ditto.
31011 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
31013 PR inline-asm/55934
31014 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
31015 that have operands with impossible constraints.
31016 Add a FIXME for a speed-up opportunity.
31017 * lra-constraints.c (process_alt_operands): Verify that a class
31018 selected from constraints on asms is valid for the operand mode.
31019 (curr_insn_transform): Remove incorrect comment.
31021 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
31023 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
31024 TOC operand is a valid symbol ref in the constant pool.
31026 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
31028 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
31030 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
31033 * config/avr/stdfix.h: New file.
31034 * t-avr (stdfix-gcc.h): New rule to build it.
31035 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
31037 2013-01-23 Kostya Serebryany <kcc@google.com>
31039 * config/darwin.h: remove dependency on
31040 CoreFoundation (asan on Mac OS).
31042 2013-01-23 Jakub Jelinek <jakub@redhat.com>
31045 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
31046 instead of cmpdi_operand for first comparison operand.
31047 Don't assert that comparison operands aren't both constants.
31049 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
31051 * doc/install.texi (Downloading the Source): Update references to
31052 downloading separate components.
31054 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
31056 * doc/extend.texi (__int128): Improve grammar.
31058 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
31061 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
31062 alternative to (o,r).
31063 (*movdi_internal_rex64): Remove (!o,n) alternative.
31064 (DImode immediate->memory splitter): Remove.
31065 (DImode immediate->memory peephole2): Remove.
31066 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
31067 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
31068 alternative to (!o,*r).
31069 (*movtf_internal_sse): New pattern.
31070 (*movxf_internal_rex64): New pattern.
31071 (*movxf_internal): Disable for TARGET_64BIT.
31072 (*movdf_internal_rex64): Remove (!o,F) alternative.
31074 2013-01-22 Jakub Jelinek <jakub@redhat.com>
31076 PR middle-end/56074
31077 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
31078 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
31079 * tree-vect-loop-manip.c (find_loop_location): Also ignore
31080 stmt locations where LOCATION_LOCUS of the stmt location is
31081 UNKNOWN_LOCATION or BUILTINS_LOCATION.
31084 * config/i386/i386.md (UNSPEC_STOS): New.
31085 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
31086 *strsetqi_1): Add UNSPEC_STOS.
31088 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
31091 * doc/invoke.texi: Remove left over -Wsynth example.
31093 2013-01-21 Jakub Jelinek <jakub@redhat.com>
31095 PR tree-optimization/56051
31096 * fold-const.c (fold_binary_loc): Don't fold
31097 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
31098 a narrowing conversion, or widening conversion from signed
31101 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
31103 PR rtl-optimization/56023
31104 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
31105 dependent on debug instruction.
31107 2013-01-21 Martin Jambor <mjambor@suse.cz>
31109 PR middle-end/56022
31110 * function.c (allocate_struct_function): Call
31111 invoke_set_current_function_hook earlier.
31113 2013-01-21 Jakub Jelinek <jakub@redhat.com>
31115 * reload1.c (init_reload): Only initialize reload_obstack
31116 during the first call.
31118 2013-01-21 Marek Polacek <polacek@redhat.com>
31120 * cfgloop.c (verify_loop_structure): Fix up grammar.
31122 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
31124 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
31125 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
31127 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
31130 * config/arm/marvell-pj4.md: Update copyright year.
31131 Fix up use of alu to alu_reg and simple_alu_imm.
31133 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
31135 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
31137 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
31140 * lra-constraints.c (curr_insn_transform): Don't reuse original
31141 insn for secondary memory move when memory mode should be different.
31143 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
31145 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
31146 atomic_storedi_1): New patterns.
31148 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
31150 btver2 pipeline descriptions.
31151 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
31153 * config/i386/i386.md (btver2_decode): New type attributes.
31154 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
31156 * config/i386/btver2.md: New file describing btver2 pipelines.
31158 2013-01-19 Andrew Pinski <apinski@cavium.com>
31160 PR tree-optimization/52631
31161 * tree-ssa-sccvn (visit_use): Before looking up the original
31162 statement, try looking up the simplified expression.
31164 2013-01-19 Anthony Green <green@moxielogic.com>
31166 * config/moxie/moxie.c (moxie_expand_prologue): Set
31167 current_function_static_stack_size.
31169 2013-01-18 Jakub Jelinek <jakub@redhat.com>
31171 PR tree-optimization/56029
31172 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
31173 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
31175 2013-01-18 Sharad Singhai <singhai@google.com>
31177 PR tree-optimization/55995
31178 * dumpfile.c (dump_loc): Print location only if available.
31179 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
31181 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
31184 * lra-constraints.c (curr_insn_transform): Reuse original insn for
31185 secondary memory move.
31186 (inherit_reload_reg): Use rclass instead of cl for
31187 check_secondary_memory_needed_p.
31189 2013-01-18 Jakub Jelinek <jakub@redhat.com>
31191 PR middle-end/56015
31192 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
31193 the case where writing real complex part of target modifies op1.
31195 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
31197 * config/aarch64/aarch64-simd.md
31198 (aarch64_vcond_internal<mode>): Handle unordered cases.
31199 * config/aarch64/iterators.md (v_cmp_result): New.
31201 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
31202 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
31204 * config/arm/marvell-pj4.md: New file.
31205 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
31206 * config/arm/arm.md (generic_sched): Add marvell_pj4.
31207 (generic_vfp): Likewise.
31208 * config/arm/arm-cores.def: Add marvell-pj4.
31209 * config/arm/arm-tune.md: Regenerate.
31210 * config/arm/arm-tables.opt: Regenerate.
31211 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
31212 * doc/invoke.texi: Document marvell-pj4.
31214 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
31216 * config/aarch64/arm_neon.h: Map scalar types to standard types.
31218 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
31223 * var-tracking.c (negative_power_of_two_p): New.
31224 (global_get_addr_cache, local_get_addr_cache): New.
31225 (get_addr_from_global_cache, get_addr_from_local_cache): New.
31226 (vt_canonicalize_addr): Rewrite using the above. Adjust the
31228 (vt_stack_offset_p): Remove.
31229 (vt_canon_true_dep): Always canonicalize loc's address.
31230 (clobber_overlapping_mems): Make sure we have a MEM.
31231 (local_get_addr_clear_given_value): New.
31232 (val_reset): Clear local cached entries.
31233 (compute_bb_dataflow): Create and release the local cache.
31234 Disable duplicate MEMs clobbering.
31235 (emit_notes_in_bb): Clobber MEMs likewise.
31236 (vt_emit_notes): Create and release the local cache.
31237 (vt_initialize, vt_finalize): Create and release the global
31238 cache, respectively.
31239 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
31241 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
31243 PR libmudflap/53359
31244 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
31245 not found in the symtab.
31247 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
31250 PR rtl-optimization/55547
31251 PR rtl-optimization/53827
31254 * alias.c (offset_overlap_p): New, factored out of...
31255 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
31256 the conservative special case for symbolic constants. Don't
31257 adjust zero sizes on alignment.
31259 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
31261 PR rtl-optimization/52573
31262 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
31263 REG_UNUSED for the same register.
31265 2013-01-17 Richard Biener <rguenther@suse.de>
31266 Marek Polacek <polacek@redhat.com>
31268 PR rtl-optimization/55833
31269 * loop-unswitch.c (unswitch_loops): Move loop verification...
31270 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
31271 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
31272 Set it to true when we're removing a loop from hierarchy tree in
31273 an irreducible region.
31274 (fix_bb_placements): Adjust caller.
31275 (fix_loop_placements): Likewise.
31277 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
31279 * config/avr/builtins.def (DEF_BUILTIN): Factor out
31280 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
31281 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
31282 Remove ID. Adjust comments.
31283 * config/avr/avr-c.c (avr_builtin_name): Remove.
31284 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
31285 * config/avr/avr.c (avr_tolower): New static function.
31286 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
31287 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
31288 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
31291 2013-01-17 Jan Hubicka <jh@suse.cz>
31293 PR tree-optimization/55273
31294 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
31296 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
31299 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
31300 store through atomic_store<mode>_1.
31301 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
31303 2013-01-17 Martin Jambor <mjambor@suse.cz>
31305 PR tree-optimizations/55264
31306 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
31307 for virtual methods.
31308 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
31309 virtual methods before inlining is over.
31310 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
31312 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
31315 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
31317 PR rtl-optimization/56005
31318 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
31319 pending reads for prefetch.
31321 2013-01-16 Ian Bolton <ian.bolton@arm.com>
31323 * config/aarch64/aarch64.md
31324 (*cstoresi_neg_uxtw): New pattern.
31325 (*cmovsi_insn_uxtw): New pattern.
31326 (*<optab>si3_uxtw): New pattern.
31327 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
31328 (*<optab>si3_insn_uxtw): New pattern.
31329 (*bswapsi2_uxtw): New pattern.
31331 2013-01-16 Richard Biener <rguenther@suse.de>
31333 * tree-inline.c (tree_function_versioning): Remove set but
31334 never used variable.
31336 2013-01-16 Richard Biener <rguenther@suse.de>
31338 PR tree-optimization/55964
31339 * tree-flow.h (rename_variables_in_loop): Remove.
31340 (rename_variables_in_bb): Likewise.
31341 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
31342 (copy_loop_before): Adjust and delete update-ssa status.
31343 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
31344 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
31345 (rename_variables_in_loop): Remove.
31346 (slpeel_update_phis_for_duplicate_loop): Likewise.
31347 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
31348 use available cfg machinery instead of duplicating it.
31349 Update PHI nodes and perform poor-mans SSA update here.
31350 (slpeel_tree_peel_loop_to_edge): Adjust.
31352 2013-01-16 Richard Biener <rguenther@suse.de>
31354 PR tree-optimization/54767
31355 PR tree-optimization/53465
31356 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
31357 (vrp_visit_phi_node): For PHI arguments coming via backedges
31358 drop all symbolical range information.
31359 (execute_vrp): Compute backedges.
31361 2013-01-16 Richard Biener <rguenther@suse.de>
31363 * doc/install.texi: Update CLooG and ISL requirements to
31366 2013-01-16 Christian Bruel <christian.bruel@st.com>
31369 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
31370 (broken_move): Handle UNSPECV_SP_SWITCH_B.
31371 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
31373 2013-01-16 DJ Delorie <dj@redhat.com>
31375 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
31376 (UNSPECV_SP_SWITCH_E): New.
31377 (sp_switch_1): Change to an unspec.
31378 (sp_switch_2): Change to an unspec. Don't use post-inc when we
31381 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
31383 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
31384 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
31385 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
31386 (expand_mem_thread_fence): Ditto.
31387 (expand_mem_signal_fence): Ditto.
31388 (expand_atomic_load): Ditto.
31389 (expand_atomic_store): Ditto.
31391 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
31393 PR rtl-optimization/55547
31394 PR rtl-optimization/53827
31397 * alias.c (memrefs_conflict_p): Set sizes to negative after
31400 2013-01-15 Jakub Jelinek <jakub@redhat.com>
31403 * function.c (thread_prologue_and_epilogue_insns): Always
31404 add crtl->drap_reg to set_up_by_prologue.set, even if
31405 stack_realign_drap is false.
31407 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
31409 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
31410 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
31411 *call): Fix indention.
31413 2013-01-15 Tom de Vries <tom@codesourcery.com>
31416 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
31419 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
31421 PR rtl-optimization/55153
31422 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
31424 2013-01-15 Martin Jambor <mjambor@suse.cz>
31426 PR tree-optimization/55920
31427 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
31428 accesses as grp_to_be_debug_replaced.
31430 2013-01-15 Jakub Jelinek <jakub@redhat.com>
31432 PR tree-optimization/55920
31433 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
31434 there is non-useless type conversion needed from debug rhs to lhs,
31435 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
31437 2013-01-15 Joseph Myers <joseph@codesourcery.com>
31438 Mikael Pettersson <mikpe@it.uu.se>
31441 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
31443 (ASM_OUTPUT_CASE_LABEL): Remove.
31444 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
31445 * final.c (shorten_branches): Update alignment of labels before
31446 jump tables if CASE_VECTOR_SHORTEN_MODE.
31448 2013-01-15 Richard Biener <rguenther@suse.de>
31451 * system.h: Do not include gmp.h for building host tools.
31453 2013-01-15 Richard Biener <rguenther@suse.de>
31455 PR middle-end/55882
31456 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
31457 account for bitpos when computing alignment.
31459 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
31461 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
31462 (ix86_target_macros_internal): Likewise.
31464 * config/i386/i386.c (m_CORE2I7): Removed.
31465 (m_CORE_HASWELL): New macro.
31466 (m_CORE_ALL): Likewise.
31467 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
31468 (initial_ix86_arch_features): Likewise.
31469 (processor_target_table): Initializations for Core avx2.
31470 (cpu_names): New names "core-avx2".
31471 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
31472 PROCESSOR_CORE_HASWELL.
31473 (ix86_issue_rate): New case.
31474 (ia32_multipass_dfa_lookahead): Likewise.
31475 (ix86_sched_init_global): Likewise.
31477 * config/i386/i386.h (TARGET_HASWELL): New macro.
31478 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
31479 (processor_type): New PROCESSOR_HASWELL.
31481 2013-01-15 Jakub Jelinek <jakub@redhat.com>
31483 PR tree-optimization/55955
31484 * tree-vect-loop.c (vectorizable_reduction): Give up early on
31485 *SHIFT_EXPR and *ROTATE_EXPR codes.
31487 PR tree-optimization/48766
31488 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
31489 -ftrapv disable -fwrapv.
31491 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
31494 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
31495 etc. to 1 and not to __flash.
31496 Use LL suffix for __INT24_MAX__ with -mint8.
31497 Use ULL suffix for __UINT24_MAX__ with -mint8.
31499 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
31501 * config/avr/avr-arch.h
31502 (struct base_arch_s): Use typedef avr_arch_t instead.
31503 (struct arch_info_s): Use typedef avr_arch_info_t instead.
31504 (struct mcu_type_s): Use typedef avr_mcu_t instead.
31505 * config/avr/avr.c: Same.
31506 * config/avr/avr-devices.c: Same.
31507 * config/avr/driver-avr.c: Same.
31508 * config/avr/gen-avr-mmcu-texi.c: Same.
31509 * config/avr/avr-mcus.def: Adjust comment.
31511 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
31513 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
31514 * config/aarch64/iterators.md (VALLDI): New.
31516 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
31517 Andi Kleen <ak@linux.intel.com>
31520 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
31521 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
31524 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
31526 * config/avr/avr-stdint.h: Remove trailing blanks.
31527 * config/avr/avr-log.h: Same.
31528 * config/avr/avr-arch.h: Same.
31529 * config/avr/avr-devices.c: Same.
31530 * config/avr/avr-dimode.md: Same.
31531 * config/avr/predicates.md: Same.
31532 * config/avr/avr-c.c: Same. And fix typo.
31534 * config/avr/avr-protos.h: Same. And:
31535 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
31536 (init_cumulative_args): Rename to avr_init_cumulative_args.
31537 (expand_prologue): Rename to avr_expand_prologue.
31538 (expand_epilogue): Rename to avr_expand_epilogue.
31539 (adjust_insn_length): Rename to avr_adjust_insn_length.
31540 (notice_update_cc): Rename to avr_notice_update_cc.
31541 (final_prescan_insn): Rename to avr_final_prescan_insn.
31542 * config/avr/avr.c: Same.
31543 * config/avr/avr.h: Same.
31544 * config/avr/avr.md: Remove trailing blanks.
31545 (prologue): Use avr_expand_prologue.
31546 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
31548 2013-01-14 Richard Biener <rguenther@suse.de>
31550 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
31551 verify_location, collect_subblocks): New functions.
31552 (verify_gimple_in_cfg): Verify that locations only reference
31553 BLOCKs in the functions BLOCK tree.
31555 2013-01-14 Richard Biener <rguenther@suse.de>
31557 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
31559 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
31561 (insert_out_of_ssa_copy_on_edge): Likewise.
31562 (rewrite_close_phi_out_of_ssa): Likewise.
31563 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
31565 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
31566 propagated constants.
31567 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
31570 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
31572 * config/avr/avr-modes.def: Add GPL copyright notice.
31574 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
31576 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
31577 MEMMODEL_MASK to determine memory model.
31578 (atomic_store<mode>): Ditto from operands[2].
31579 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
31581 2013-01-13 Jakub Jelinek <jakub@redhat.com>
31584 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
31585 (fold_gimple_assign): Don't call unshare_expr here.
31586 (fold_ctor_reference): Call unshare_expr.
31588 2013-01-13 Terry Guo <terry.guo@arm.com>
31590 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
31591 * doc/fragments.texi: Document MULTILIB_REUSE.
31592 * gcc.c (multilib_reuse): New internal spec.
31593 (set_multilib_dir): Also search multilib from multilib_reuse.
31594 * genmultilib (tmpmultilib3): Refactor code.
31595 (tmpmultilib4): Ditto.
31596 (multilib_reuse): New multilib argument.
31598 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
31600 * Makefile.in: Update copyright.
31602 2013-01-12 Tom de Vries <tom@codesourcery.com>
31604 PR middle-end/55890
31605 * calls.c (expand_call): Check if arg_nr is valid.
31607 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
31609 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
31610 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
31611 documentation. Add missing '__' in front of
31612 __builtin_ia32_packssdw256.
31614 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
31617 * config/s390/s390.c (s390_preferred_reload_class): Do not return
31618 NO_REGS for larl operands.
31619 (s390_reload_larl_operand): Use s390_load_address instead of
31622 2013-01-11 Richard Biener <rguenther@suse.de>
31624 * tree-cfg.c (verify_node_sharing_1): Split out from ...
31625 (verify_node_sharing): ... here.
31626 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
31628 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
31630 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
31631 Substitute TREECHECKING.
31632 * configure: Regenerate.
31633 * Makefile.in (TREECHECKING): New.
31635 2013-01-11 Richard Guenther <rguenther@suse.de>
31637 PR tree-optimization/44061
31638 * tree-vrp.c (extract_range_basic): Compute zero as
31639 value-range for __builtin_constant_p of function parameters.
31641 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
31643 Update copyright years.
31645 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
31647 PR rtl-optimization/55672
31648 * lra-eliminations.c (mark_not_eliminable): Permit addition with
31649 const to be eliminable.
31651 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
31653 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
31654 * configure: Regenerate.
31656 2013-01-10 Richard Biener <rguenther@suse.de>
31658 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
31660 2013-01-10 Richard Biener <rguenther@suse.de>
31663 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
31664 locations for virtual PHI arguments.
31665 (rewrite_update_phi_arguments): Likewise.
31667 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
31669 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
31672 2013-01-10 Jakub Jelinek <jakub@redhat.com>
31674 PR tree-optimization/55921
31675 * tree-complex.c (expand_complex_asm): New function.
31676 (expand_complex_operations_1): Call it for GIMPLE_ASM.
31678 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
31681 * config/s390/s390.c (s390_symref_operand_p)
31682 (s390_loadrelative_operand_p): Merge the two functions.
31683 (s390_check_qrst_address, print_operand_address): Add parameters
31684 to s390_loadrelative_operand_p invokation.
31685 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
31686 (s390_reload_larl_operand, s390_secondary_reload): Use
31687 s390_loadrelative_operand_p instead of s390_symref_operand_p.
31688 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
31690 2013-01-09 Mike Stump <mikestump@comcast.net>
31692 * dse.c (record_store): Remove unnecessary assert.
31694 2013-01-09 Jan Hubicka <jh@suse.cz>
31696 PR tree-optimization/55569
31697 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
31698 * cfgloop.h (scale_loop_profile): Likewise.
31700 2013-01-09 Jan Hubicka <jh@suse.cz>
31703 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
31705 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
31707 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
31709 PR middle-end/55114
31710 * expr.h (maybe_emit_group_store): Declare.
31711 * expr.c (maybe_emit_group_store): New function.
31712 * builtins.c (expand_builtin_int_roundingfn): Call it.
31713 (expand_builtin_int_roundingfn_2): Likewise.
31715 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
31717 PR rtl-optimization/55829
31718 * lra-constraints.c (match_reload): Add code for absent output.
31719 (curr_insn_transform): Add code for reloads of matched inputs
31722 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
31724 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
31725 attribute of movddup insn to DF.
31726 (*vec_interleave_lowv2df): Ditto.
31727 (vec_dupv2df): Ditto.
31729 2013-01-09 Jan Hubicka <jh@suse.cz>
31731 PR tree-optimiation/55875
31732 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
31733 EVERY_ITERATION parameter.
31734 (number_of_iterations_exit): Check if exit is executed every iteration.
31735 (idx_infer_loop_bounds): Similarly here.
31736 (n_of_executions_at_most): Simplify
31737 to only test for cases where statement is dominated by the
31738 particular bound; handle correctly the "postdominance" test.
31739 (scev_probably_wraps_p): Use max loop iterations info
31740 as a global bound first.
31742 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
31743 Nick Clifton <nickc@redhat.com>
31745 * config/v850/v850.md (cbranchsf4): New pattern.
31746 (cstoresf4): New pattern.
31747 (cbranchdf4): New pattern.
31748 (cstoredf4): New pattern.
31749 (movsicc): Disallow floating point comparisons.
31750 (cmpsf_le_insn): Fix order of operators.
31751 (cmpsf_lt_insn): Likewise.
31752 (cmpsf_eq_insn): Likewise.
31753 (cmpdf_le_insn): Likewise.
31754 (cmpdf_lt_insn): Likewise.
31755 (cmpdf_eq_insn): Likewise.
31756 (cmpsf_ge_insn): Use LE comparison.
31757 (cmpdf_ge_insn): Likewise.
31758 (cmpsf_gt_insn): Use LT comparison.
31759 (cmpdf_gt_insn): Likewise.
31760 (cmpsf_ne_insn): Delete pattern.
31761 (cmpdf_ne_insn): Delete pattern.
31762 * config/v850/v850.c (v850_gen_float_compare): Use
31763 gen_cmpdf_eq_insn for NE comparison.
31764 (v850_float_z_comparison_operator)
31765 (v850_float_nz_comparison_operator): Move from here ...
31766 * config/v850/predicates.md: ... to here. Move GT and GE
31767 comparisons into v850_float_z_comparison_operator.
31768 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
31770 (v850_float_nz_comparison_operator): Likewise.
31772 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
31774 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
31775 with calls to gen_insvsi/gen_insvdi.
31777 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
31779 * config/i386/i386.c (initial_ix86_tune_features): Set up
31780 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
31782 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
31783 Jakub Jelinek <jakub@redhat.com>
31785 PR tree-optimization/48189
31786 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
31787 If nitercst is 0, don't predict the exit edge.
31789 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
31791 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
31792 in asm_fprintf with reg_names.
31793 (aarch64_print_operand_address): Likewise.
31794 (aarch64_return_addr): Likewise.
31795 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
31797 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
31799 * config/pa/pa.h (VAL_U6_BITS_P): Define.
31800 (INT_U6_BITS): Likewise.
31801 * config/pa/predicates.md (uint6_operand): New predicate.
31802 (shift5_operand, shift6_operand): Likewise.
31803 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
31805 (lshrdi3): Use shift6_operand.
31806 (shrpsi4, shrpdi4): New insn patterns.
31807 (extzv): Delete expander.
31808 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
31809 predicates in unamed zero extract patterns. Tighten common constraint.
31810 (extv): Delete expander.
31811 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
31812 predicates in unamed sign extract patterns. Tighten common constraint.
31813 (insv): Delete expander.
31814 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
31815 predicates in unamed insert patterns. Tighten common constraint.
31816 Change uint32_operand predicate to uint6_operand predicate in unamed
31817 DImode pattern to insert constant values of type 1...1xxxx.
31819 2013-01-04 Jan Hubicka <jh@suse.cz>
31821 PR tree-optimization/55823
31822 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
31825 2013-01-08 Jakub Jelinek <jakub@redhat.com>
31826 Uros Bizjak <ubizjak@gmail.com>
31828 PR rtl-optimization/55845
31829 * df-problems.c (can_move_insns_across): Stop scanning at
31830 volatile_insn_p source instruction or give up if
31831 across_from .. across_to range contains any volatile_insn_p
31834 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
31836 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
31837 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
31839 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
31840 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
31842 2013-01-08 Jakub Jelinek <jakub@redhat.com>
31845 * asan.c (asan_clear_shadow): New function.
31846 (asan_emit_stack_protection): Use it.
31848 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
31850 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
31851 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
31852 with tab instead of space.
31854 2013-01-08 Nick Clifton <nickc@redhat.com>
31856 * config/rl78/rl78.c (rl78_expand_prologue): Always select
31857 register bank 0 at the start of an interrupt handler.
31858 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
31861 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
31863 * config/aarch64/aarch64-simd.md
31864 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
31865 (aarch64_simd_bsl): Likewise.
31866 (aarch64_vcond_internal<mode>): Likewise.
31867 (vcond<mode><mode>): Likewise.
31868 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
31869 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
31871 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
31873 * config/aarch64/aarch64-builtins.c
31874 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
31876 2013-01-08 Martin Jambor <mjambor@suse.cz>
31879 * tree-sra.c (analyze_access_subtree): Return true also after
31880 potentially creating a debug-only replacement.
31882 2013-01-08 Jakub Jelinek <jakub@redhat.com>
31884 PR middle-end/55890
31885 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
31887 PR tree-optimization/54120
31888 * tree-vrp.c (range_fits_type_p): Don't allow
31889 src_precision < precision from signed vr to unsigned_p
31890 if vr->min or vr->max is negative.
31891 (simplify_float_conversion_using_ranges): Test can_float_p
31892 against CODE_FOR_nothing.
31894 2013-01-08 Jakub Jelinek <jakub@redhat.com>
31895 Richard Biener <rguenther@suse.de>
31897 PR middle-end/55851
31898 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
31899 types instead of just INTEGER_TYPE types.
31901 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
31903 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
31906 2013-01-07 Steve Ellcey <sellcey@mips.com>
31909 * config/mips/mips.opt: Change mad to mmad to match documentation.
31911 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
31914 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
31915 .progmemx.data now.
31917 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
31920 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
31921 (avr_addrspace_t): Add .section_name field.
31922 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
31924 (avr_addrspace): Same. Initialize .section_name. Remove last
31925 NULL entry. Put __memx into .progmemx.data.
31926 (progmem_section_prefix): Remove.
31927 (avr_asm_init_sections): No need to initialize progmem_section.
31928 (avr_asm_named_section): Use avr_addrspace[].section_name to get
31929 section name prefix.
31930 (avr_asm_select_section): Ditto. And use get_unnamed_section to
31931 retrieve the progmem section.
31932 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
31933 boundary to run over avr_addrspace[].
31934 (avr_register_target_pragmas): Ditto.
31936 2013-01-06 Jakub Jelinek <jakub@redhat.com>
31938 * varasm.c (output_constant_def_contents): For asan_protect_global
31939 protected strings, adjust DECL_ALIGN if needed, before testing for
31941 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
31942 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
31944 (output_object_block): For asan protected decls, emit asan padding
31945 after their contents.
31946 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
31947 (asan_finish_file): Test it here instead.
31949 2013-01-07 Nick Clifton <nickc@redhat.com>
31950 Matthias Klose <doko@debian.org>
31951 Doug Kwan <dougkwan@google.com>
31952 H.J. Lu <hongjiu.lu@intel.com>
31955 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
31957 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
31959 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
31961 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
31963 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
31965 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
31968 * doc/install.texi (Cross-Compiler-Specific Options): Document
31971 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
31973 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
31974 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
31975 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
31976 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
31977 vqmovun_high_s64): Fix source operand number and update copyright.
31979 2013-01-07 Richard Biener <rguenther@suse.de>
31981 PR middle-end/55890
31982 * gimple.h (gimple_call_builtin_p): New overload.
31983 * gimple.c (validate_call): New function.
31984 (gimple_call_builtin_p): Likewise.
31985 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
31986 Use gimple_call_builtin_p.
31987 (find_func_clobbers): Likewise.
31988 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
31989 (strlen_optimize_stmt): Likewise.
31991 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
31993 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
31994 (vld1q_dup_*): Likewise.
31995 (vld1_*): Likewise.
31996 (vld1q_*): Likewise.
31997 (vld1_lane_*): Likewise.
31998 (vld1q_lane_*): Likewise.
32000 2013-01-07 Richard Biener <rguenther@suse.de>
32002 * lto-streamer.h (LTO_minor_version): Bump to 2.
32004 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
32006 * config/aarch64/aarch64-protos.h
32007 (aarch64_const_double_zero_rtx_p): Rename to...
32008 (aarch64_float_const_zero_rtx_p): ...this.
32009 (aarch64_float_const_representable_p): New.
32010 (aarch64_output_simd_mov_immediate): Likewise.
32011 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
32012 move immediate case.
32013 * config/aarch64/aarch64.c
32014 (aarch64_const_double_zero_rtx_p): Rename to...
32015 (aarch64_float_const_zero_rtx_p): ...this.
32016 (aarch64_print_operand): Allow printing of new constants.
32017 (aarch64_valid_floating_const): New.
32018 (aarch64_legitimate_constant_p): Check for valid floating-point
32020 (aarch64_simd_valid_immediate): Likewise.
32021 (aarch64_vect_float_const_representable_p): New.
32022 (aarch64_float_const_representable_p): Likewise.
32023 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
32024 (aarch64_output_simd_mov_immediate): New.
32025 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
32026 (*movdf_aarch64): Likewise.
32027 * config/aarch64/constraints.md (Ufc): New.
32028 (Y): call aarch64_float_const_zero_rtx.
32029 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
32031 2013-01-07 Richard Biener <rguenther@suse.de>
32033 PR tree-optimization/55888
32034 PR tree-optimization/55862
32035 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
32036 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
32037 not if it is contained therein.
32039 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
32041 * config/avr/t-avr: Typo.
32043 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
32046 * config/avr/t-avr: Don't automatically rebuild
32047 $(srcdir)/config/avr/t-multilib
32048 $(srcdir)/config/avr/avr-tables.opt
32049 $(srcdir)/doc/avr-mmcu.texi
32050 (avr-mcus): New phony target to build them on request.
32051 (s-avr-mlib, s-avr-mmcu-texi): Remove.
32052 * avr/avr-mcus.def: Adjust comments.
32054 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
32056 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
32058 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
32060 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
32062 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
32064 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
32066 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
32068 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
32069 to generate profiling.
32070 * config/rs6000/aix64.h (LIB_SPEC): Same.
32072 2013-01-04 Andrew Pinski <apinski@cavium.com>
32074 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
32076 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
32078 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
32080 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
32082 (ix86_expand_move): Ditto.
32083 (ix86_zero_extend_to_Pmode): Ditto.
32084 (ix86_expand_call): Ditto.
32085 (ix86_expand_special_args_builtin): Ditto.
32086 (ix86_expand_builtin): Ditto.
32088 2013-01-04 Richard Biener <rguenther@suse.de>
32090 PR tree-optimization/55862
32091 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
32092 translating them through PHI nodes.
32094 2013-01-04 Martin Jambor <mjambor@suse.cz>
32096 PR tree-optimization/55755
32097 * tree-sra.c (sra_modify_assign): Do not check that an access has no
32098 children when trying to avoid producing a VIEW_CONVERT_EXPR.
32100 2013-01-04 Marek Polacek <polacek@redhat.com>
32102 PR middle-end/55859
32103 * opts.c (default_options_optimization): Clarify error message.
32105 2013-01-04 Richard Biener <rguenther@suse.de>
32107 PR middle-end/55863
32108 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
32111 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
32114 * config/pa/pa.md (movsi): Revert previous change.
32115 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
32118 2013-01-03 Richard Henderson <rth@redhat.com>
32120 * config/i386/i386.c (ix86_expand_move): Always assign to op1
32121 after eliminating TLS symbols.
32123 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
32126 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
32127 * graphite-poly.c (debug_gmp_value): Likewise.
32129 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
32132 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
32133 selected code model, define __code_mode_small__, __code_model_medium__,
32134 __code_model_large__, __code_model_32__ or __code_model_kernel__.
32135 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
32136 xchg temporary register with %k. Declare temporary register as
32138 [__x86_64__]: For medium and large code models, preserve %rbx register.
32140 2013-01-03 Richard Biener <rguenther@suse.de>
32142 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
32143 (dump_subscript): Adjust.
32144 (finalize_ddr_dependent): Do not dump redundant info.
32145 (analyze_siv_subscript): Adjust.
32146 (subscript_dependence_tester): Likewise.
32147 (compute_affine_dependence): Likewise.
32149 2013-01-03 Richard Biener <rguenther@suse.de>
32152 2013-01-03 Richard Biener <rguenther@suse.de>
32154 PR tree-optimization/55857
32155 * tree-vect-stmts.c (vectorizable_load): Do not setup
32156 re-alignment for invariant loads.
32158 2013-01-02 Richard Biener <rguenther@suse.de>
32160 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
32161 invariant load do not generate a vector load from the scalar location.
32163 2013-01-03 Richard Biener <rguenther@suse.de>
32165 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
32166 for not vectorizing.
32167 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
32168 not build INDIRECT_REFs, call get_name once only.
32169 (vect_create_data_ref_ptr): Likewise. Dump base object kind
32170 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
32172 2013-01-03 Richard Biener <rguenther@suse.de>
32174 PR tree-optimization/55857
32175 * tree-vect-stmts.c (vectorizable_load): Do not setup
32176 re-alignment for invariant loads.
32178 2013-01-03 Richard Biener <rguenther@suse.de>
32181 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
32182 prefer a built-in decl.
32184 2013-01-03 Jakub Jelinek <jakub@redhat.com>
32186 * gcc.c (process_command): Update copyright notice dates.
32187 * gcov.c (print_version): Likewise.
32188 * gcov-dump.c (print_version): Likewise.
32190 PR rtl-optimization/55838
32191 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
32192 iv0.step, iv1.step and step.
32194 2013-01-03 Jakub Jelinek <jakub@redhat.com>
32195 Marc Glisse <marc.glisse@inria.fr>
32197 PR tree-optimization/55832
32198 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
32199 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
32200 integer_{one,zero}_node.
32202 2013-01-03 Jakub Jelinek <jakub@redhat.com>
32205 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
32206 * var-tracking.c (reverse_op): Don't add reverse ops to
32207 VALUEs that have already
32208 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
32210 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
32212 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
32214 2013-01-02 Teresa Johnson <tejohnson@google.com>
32216 * dumpfile.c (dump_loc): Print filename with location.
32217 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
32218 new location_t parameter to emit complete unroll message with
32219 new dump framework.
32220 (canonicalize_loop_induction_variables): Compute loops location
32221 and pass to try_unroll_loop_completely.
32222 * loop-unroll.c (report_unroll_peel): New function.
32223 (peel_loops_completely): Use new dump format with location
32224 for main dumpfile message, and invoke report_unroll_peel on success.
32225 (decide_unrolling_and_peeling): Ditto.
32226 (decide_peel_once_rolling): Remove old dumpfile message subsumed
32227 by report_unroll_peel.
32228 (decide_peel_completely): Ditto.
32229 (decide_unroll_constant_iterations): Ditto.
32230 (decide_unroll_runtime_iterations): Ditto.
32231 (decide_peel_simple): Ditto.
32232 (decide_unroll_stupid): Ditto.
32233 * cfgloop.c (get_loop_location): New function.
32234 * cfgloop.h (get_loop_location): Declare.
32236 2013-01-02 Sriraman Tallam <tmsriram@google.com>
32238 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
32241 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
32243 PR middle-end/55198
32244 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
32245 BLKmode objects when EXPAND_MEMORY is specified.
32247 2013-01-02 Sriraman Tallam <tmsriram@google.com>
32249 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
32251 (fold_builtin_cpu): Do not share cpu model decls across statements.
32253 2013-01-02 Jason Merrill <jason@redhat.com>
32256 * tree.c (build_array_type_1): Revert earlier change.
32258 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
32260 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
32262 * config/aarch64/aarch64-tune.md: Re-generate.
32264 2013-01-02 Richard Biener <rguenther@suse.de>
32266 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
32267 invariant load do not generate a vector load from the scalar location.
32269 2013-01-02 Richard Biener <rguenther@suse.de>
32272 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
32273 * configure: Regenerate.
32275 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
32277 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
32278 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
32279 (expand_builtin_int_roundingfn_2): Keep the original target around
32280 for the fallback case.
32282 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
32284 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
32285 to be clear for sign changes.
32287 2013-01-01 Jan Hubicka <jh@suse.cz>
32289 * ipa-inline-analysis.c: Fix formatting.
32291 2013-01-01 Jakub Jelinek <jakub@redhat.com>
32293 PR tree-optimization/55831
32294 * tree-vect-loop.c (get_initial_def_for_induction): Use
32295 gsi_after_labels instead of gsi_start_bb.
32297 Copyright (C) 2013 Free Software Foundation, Inc.
32299 Copying and distribution of this file, with or without modification,
32300 are permitted in any medium without royalty provided the copyright
32301 notice and this notice are preserved.