1 2013-11-18 Richard Biener <rguenther@suse.de>
3 PR tree-optimization/59125
4 PR tree-optimization/54570
5 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): When inlining
6 is not complete do not treat component-references with offset zero
7 but different fields as equal.
8 * tree-object-size.c: Include tree-phinodes.h and ssa-iterators.h.
9 (compute_object_sizes): Apply TLC. Propagate the constant
10 results into all uses and fold their stmts.
11 * passes.def (pass_all_optimizations): Move pass_object_sizes
12 after the first pass_forwprop and before pass_fre.
14 2013-11-18 Richard Sandiford <rdsandiford@googlemail.com>
16 * tree.h (tree_to_uhwi): Return an unsigned HOST_WIDE_INT.
17 * tree.c (tree_to_uhwi): Return an unsigned HOST_WIDE_INT.
18 (tree_ctz): Remove cast to unsigned type.
19 * builtins.c (fold_builtin_memory_op): Likewise.
20 * dwarf2out.c (descr_info_loc): Likewise.
21 * godump.c (go_output_typedef): Likewise.
22 * omp-low.c (expand_omp_simd): Likewise.
23 * stor-layout.c (excess_unit_span): Likewise.
24 * tree-object-size.c (addr_object_size): Likewise.
25 * tree-sra.c (analyze_all_variable_accesses): Likewise.
26 * tree-ssa-forwprop.c (simplify_builtin_call): Likewise.
27 (simplify_rotate): Likewise.
28 * tree-ssa-strlen.c (adjust_last_stmt, handle_builtin_memcpy)
29 (handle_pointer_plus): Likewise.
30 * tree-switch-conversion.c (check_range): Likewise.
31 * tree-vect-patterns.c (vect_recog_rotate_pattern): Likewise.
32 * tsan.c (instrument_builtin_call): Likewise.
33 * cfgexpand.c (defer_stack_allocation): Add cast to HOST_WIDE_INT.
34 * trans-mem.c (tm_log_add): Likewise.
35 * config/aarch64/aarch64.c (aapcs_vfp_sub_candidate): Likewise.
36 * config/arm/arm.c (aapcs_vfp_sub_candidate): Likewise.
37 * config/rs6000/rs6000.c (rs6000_aggregate_candidate): Likewise.
38 * config/mips/mips.c (r10k_safe_mem_expr_p): Make offset unsigned.
40 2013-11-18 Richard Sandiford <rdsandiford@googlemail.com>
42 * tree.h (host_integerp, tree_low_cst): Delete.
43 * tree.c (host_integerp, tree_low_cst): Delete.
45 2013-11-18 Richard Sandiford <rdsandiford@googlemail.com>
47 * expr.h: Update comments to refer to tree_to_[su]hwi rather
49 * fold-const.c (fold_binary_loc): Likewise.
50 * expr.c (store_constructor): Use tree_to_uhwi rather than
52 * ipa-utils.h (possible_polymorphic_call_target_p): Likewise.
53 * stmt.c (emit_case_dispatch_table): Likewise.
54 * tree-switch-conversion.c (emit_case_bit_tests): Likewise.
56 2013-11-18 Richard Sandiford <rdsandiford@googlemail.com>
58 * alias.c, asan.c, builtins.c, cfgexpand.c, cgraph.c,
59 config/aarch64/aarch64.c, config/alpha/predicates.md,
60 config/arm/arm.c, config/darwin.c, config/epiphany/epiphany.c,
61 config/i386/i386.c, config/iq2000/iq2000.c, config/m32c/m32c-pragma.c,
62 config/mep/mep-pragma.c, config/mips/mips.c,
63 config/picochip/picochip.c, config/rs6000/rs6000.c, cppbuiltin.c,
64 dbxout.c, dwarf2out.c, emit-rtl.c, except.c, expr.c, fold-const.c,
65 function.c, gimple-fold.c, godump.c, ipa-cp.c, ipa-prop.c, omp-low.c,
66 predict.c, sdbout.c, stor-layout.c, trans-mem.c, tree-object-size.c,
67 tree-sra.c, tree-ssa-ccp.c, tree-ssa-forwprop.c,
68 tree-ssa-loop-ivcanon.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-niter.c,
69 tree-ssa-loop-prefetch.c, tree-ssa-strlen.c, tree-stdarg.c,
70 tree-switch-conversion.c, tree-vect-generic.c, tree-vect-loop.c,
71 tree-vect-patterns.c, tree-vrp.c, tree.c, tsan.c, ubsan.c, varasm.c:
72 Replace tree_low_cst (..., 1) with tree_to_uhwi throughout.
74 2013-11-18 Richard Sandiford <rdsandiford@googlemail.com>
76 * builtins.c, cilk-common.c, config/aarch64/aarch64.c,
77 config/alpha/alpha.c, config/arm/arm.c, config/c6x/predicates.md,
78 config/i386/i386.c, config/ia64/predicates.md, config/s390/s390.c,
79 coverage.c, dbxout.c, dwarf2out.c, except.c, explow.c, expr.c, expr.h,
80 fold-const.c, gimple-fold.c, godump.c, ipa-prop.c, omp-low.c,
81 predict.c, rtlanal.c, sdbout.c, stmt.c, stor-layout.c, targhooks.c,
82 tree-cfg.c, tree-data-ref.c, tree-inline.c, tree-ssa-forwprop.c,
83 tree-ssa-loop-prefetch.c, tree-ssa-phiopt.c, tree-ssa-sccvn.c,
84 tree-ssa-strlen.c, tree-stdarg.c, tree-vect-data-refs.c,
85 tree-vect-patterns.c, tree.c, tree.h, var-tracking.c, varasm.c:
86 Replace tree_low_cst (..., 0) with tree_to_shwi throughout.
88 2013-11-18 Richard Sandiford <rdsandiford@googlemail.com>
90 * tree.h (tree_to_shwi, tree_to_uhwi): Declare, with inline expansions.
91 * tree.c (tree_to_shwi, tree_to_uhwi): New functions.
93 2013-11-18 Richard Sandiford <rdsandiford@googlemail.com>
95 * expr.h: Update comments to refer to tree_fits_[su]hwi_p rather
98 2013-11-18 Richard Sandiford <rdsandiford@googlemail.com>
100 * builtins.c, config/alpha/alpha.c, config/iq2000/iq2000.c,
101 config/mips/mips.c, dbxout.c, dwarf2out.c, expr.c, fold-const.c,
102 gimple-fold.c, godump.c, omp-low.c, predict.c, sdbout.c, stor-layout.c,
103 tree-dfa.c, tree-sra.c, tree-ssa-forwprop.c, tree-ssa-loop-prefetch.c,
104 tree-ssa-phiopt.c, tree-ssa-sccvn.c, tree-ssa-strlen.c,
105 tree-ssa-structalias.c, tree-vect-data-refs.c, tree-vect-patterns.c,
106 tree.c, varasm.c, alias.c, cfgexpand.c, config/aarch64/aarch64.c,
107 config/arm/arm.c, config/epiphany/epiphany.c, config/i386/i386.c,
108 config/m32c/m32c-pragma.c, config/mep/mep-pragma.c,
109 config/rs6000/rs6000.c, config/sparc/sparc.c, emit-rtl.c, function.c,
110 gimplify.c, ipa-prop.c, stmt.c, trans-mem.c, tree-cfg.c,
111 tree-object-size.c, tree-ssa-ccp.c, tree-ssa-loop-ivcanon.c,
112 tree-stdarg.c, tree-switch-conversion.c, tree-vect-generic.c,
113 tree-vrp.c, tsan.c, ubsan.c: Replace host_integerp (..., 1) with
114 tree_fits_uhwi_p throughout.
116 2013-11-18 Richard Sandiford <rdsandiford@googlemail.com>
118 * builtins.c, config/alpha/alpha.c, config/c6x/predicates.md,
119 config/ia64/predicates.md, config/iq2000/iq2000.c, config/mips/mips.c,
120 config/s390/s390.c, dbxout.c, dwarf2out.c, except.c, explow.c, expr.c,
121 expr.h, fold-const.c, gimple-fold.c, gimple-ssa-strength-reduction.c,
122 gimple.c, godump.c, graphite-scop-detection.c, graphite-sese-to-poly.c,
123 omp-low.c, predict.c, rtlanal.c, sdbout.c, simplify-rtx.c,
124 stor-layout.c, tree-data-ref.c, tree-dfa.c, tree-pretty-print.c,
125 tree-sra.c, tree-ssa-alias.c, tree-ssa-forwprop.c,
126 tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c, tree-ssa-math-opts.c,
127 tree-ssa-phiopt.c, tree-ssa-reassoc.c, tree-ssa-sccvn.c,
128 tree-ssa-strlen.c, tree-ssa-structalias.c, tree-vect-data-refs.c,
129 tree-vect-patterns.c, tree-vectorizer.h, tree.c, var-tracking.c,
130 varasm.c: Replace host_integerp (..., 0) with tree_fits_shwi_p
133 2013-11-18 Richard Sandiford <rdsandiford@googlemail.com>
135 * tree.h (tree_fits_shwi_p, tree_fits_uhwi_p): Declare.
136 * tree.c (tree_fits_shwi_p, tree_fits_uhwi_p): Define.
138 2013-11-18 Kirill Yukhin <kirill.yukhin@intel.com>
140 * gcc/config/ia64/ia64.c (ia64_split_tmode_move): Mark
141 load with `dead' flag if it kills address, not its
144 2013-11-18 Ilya Enkovich <ilya.enkovich@intel.com>
146 * builtin-types.def (BT_FN_PTR_CONST_PTR_VAR): New.
147 * chkp-builtins.def (BUILT_IN_CHKP_BIND_BOUNDS): New.
148 * cfgexpand.c (expand_call_stmt): Expand BUILT_IN_CHKP_BIND_BOUNDS.
149 * gimple.c (gimple_call_get_nobnd_arg_index): Remove.
150 * gimple.h (gf_mask): Add GF_CALL_WITH_BOUNDS.
151 (gimple_call_with_bounds_p): New.
152 (gimple_call_set_with_bounds): New.
153 (gimple_call_num_nobnd_args): Remove.
154 (gimple_call_nobnd_arg): Remove.
155 * tree.h (CALL_WITH_BOUNDS_P): New.
156 * rtl.h (CALL_EXPR_WITH_BOUNDS_P): New.
158 2013-11-18 Trevor Saunders <tsaunders@mozilla.com>
160 * cgraph.h (symtab_node_asm_name): Rename to symtab_node::asm_name.
161 (symtab_node_name): Rename to symtab_node::name.
162 (cgraph_node_asm_name): Remove.
163 (varpool_node_asm_name): Remove.
164 * cgraph.c cgraphclones.c cgraphunit.c ipa-cp.c ipa-devirt.c
165 ipa-inline-analysis.c ipa-inline-transform.c ipa-inline.c
166 ipa-profile.c ipa-prop.c ipa-pure-const.c ipa-ref.c ipa-reference.c
167 ipa-utils.c ipa.c symtab.c tree-inline.c tree-sra.c
168 tree-ssa-structalias.c value-prof.c varpool.c Adjust.
170 2013-11-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
172 * config/arm/aarch-cost-tables.h (cortexa53_extra_costs): New table.
173 * config/arm/arm.c (arm_cortex_a53_tune): New.
174 * config/arm/arm-cores.def (cortex-a53): Use cortex_a53 tuning struct.
176 2013-11-12 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
178 * config.gcc (i[34567]86-*-linux* | ...): Add bdver4.
179 (case ${target}): Add bdver4.
180 * config/i386/bdver3.md: Add bdver4.
181 * config/i386/driver-i386.c: (host_detect_local_cpu): Let
182 -march=native recognize bdver4 processors.
183 * config/i386/i386-c.c (ix86_target_macros_internal): Add
185 * config/i386/i386.c (struct processor_costs bdver4_cost): New.
186 (m_BDVER4): New definition.
187 (m_AMD_MULTIPLE): Includes m_BDVER4.
188 (processor_target_table): Add bdver4 entry.
189 (static const char *const cpu_names): Add bdver4 entry.
190 (software_prefetching_beneficial_p): Add bdver3.
191 (ix86_option_override_internal): Add bdver4 instruction sets.
192 (ix86_issue_rate): Add bdver4.
193 (ix86_adjust_cost): Add bdver4.
194 (ia32_multipass_dfa_lookahead): Add bdver4.
195 (enum processor_model): Add M_AMDFAM15H_BDVER4.
196 (struct _arch_names_table): Add M_AMDFAM15H_BDVER4.
197 (has_dispatch): Add bdver4.
198 * config/i386/i386.h (TARGET_BDVER4): New definition.
199 (enum target_cpu_default): Add TARGET_CPU_DEFAULT_bdver4.
200 (enum processor_type): Add PROCESSOR_BDVER4.
201 * config/i386/i386.md (define_attr "cpu"): Add bdver4.
202 * config/i386/i386.opt (flag_dispatch_scheduler): Add bdver4.
203 * gcc/doc/extend.texi: Add details about bdver4.
204 * gcc/doc/invoke.texi: Add details about bdver4. Add
205 fma4 and fsgsbase for bdver3. Add fma4 for bdver2.
207 2013-11-17 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
209 * config/rs6000/rs6000.c (rs6000_emit_move): Use low word of
210 sdmode_stack_slot also in little-endian mode.
212 2013-11-17 Jan Hubicka <jh@suse.cz>
214 * doc/md.texi (setmem, movstr): Update documentation.
215 * builtins.c (determine_block_size): New function.
216 (expand_builtin_memcpy): Use it and pass it to
217 emit_block_move_hints.
218 (expand_builtin_memset_args): Use it and pass it to
219 set_storage_via_setmem.
220 * expr.c (emit_block_move_via_movmem): Add min_size/max_size parameters;
221 update call to expander.
222 (emit_block_move_hints): Add min_size/max_size parameters.
223 (clear_storage_hints): Likewise.
224 (set_storage_via_setmem): Likewise.
225 (clear_storage): Update.
226 * expr.h (emit_block_move_hints, clear_storage_hints,
227 set_storage_via_setmem): Update prototype.
228 * i386.c (ix86_expand_set_or_movmem): Add bounds; export.
229 (ix86_expand_movmem, ix86_expand_setmem): Remove.
230 (ix86_expand_movmem, ix86_expand_setmem): Remove.
231 * i386.md (movmem, setmem): Pass parameters.
233 2013-11-17 Uros Bizjak <ubizjak@gmail.com>
236 * config/i386/i386.c (ix86_address_subreg_operand): Do not
237 reject non-integer subregs.
238 (ix86_decompose_address): Do not reject invalid CONST_INT RTXes.
239 Move check for invalid x32 constant addresses ...
240 (ix86_legitimate_address_p): ... here.
242 2011-11-17 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
244 * config/rs6000/rs6000.c (rs6000_frame_related): Add split_reg
245 parameter and use it in REG_FRAME_RELATED_EXPR note.
246 (emit_frame_save): Call rs6000_frame_related with extra NULL_RTX
248 (rs6000_emit_prologue): Likewise, but for little endian VSX
249 stores, pass the source register of the store instead.
251 2013-11-17 Andrew MacLeod <amacleod@redhat.com>
253 * gimple.h: Reorder prototypes to match .c declaration order, and
254 remove protyotypes for functions not in gimple.c.
255 (LABEL): Move to tree-into-ssa.c.
256 * gimple.c: Remove unused prototypes.
257 (get_base_address): Move to tree.c.
258 * tree.c (get_base_address): Relocate from gimple.c.
259 * builtins.h (validate_gimple_arglist): Add prototype.
260 * trans-mem.h (compute_transaction_bits, is_tm_ending): Add prototype.
261 * cfgexpand.h: New File.
262 (gimple_assign_rhs_to_tree, estimated_stack_frame_size): Add protoype.
263 * tree.h (build_addr): Move to tree-nested.h.
264 * tree-nested.h: New File.
265 (build_addr, lower_nested_functions, insert_field_into_struct): Add
267 * tree-inline.h (estimated_stack_frame_size): Remove prototype.
268 * ipa-inline-analysis.c: Include cfgexpand.h.
269 * cgraphunit.c: Include tree-nested.h.
270 * omp-low.c: Likewise.
271 * tree-parloops.c: Likewise.
272 * gimple-low.h: Likewise.
273 * tree-profile.h: Likewise.
274 * expr.c: Include cfgexpand.h.
275 * tree-affine.c: Likewise.
276 * tree-ssa.c: Likewise.
277 * tree-ssa-loop-im.c: Include trans-mem.h.
278 * tree-ssa-tail-merge.c: Likewise.
279 * value-prof.c: Include builtins.h and tree-nested.h.
280 * tree-into-ssa.c (LABEL): Define here.
282 2013-11-16 Joern Rennecke <joern.rennecke@embecosm.com>
284 * config/arc/arc.c (arc_predicate_delay_insns): New function.
285 (pass_data_arc_predicate_delay_insns): New pass_data instance.
286 (pass_arc_predicate_delay_insns): New subclass of rtl_opt_class.
287 (make_pass_arc_predicate_delay_insns): New function.
288 (arc_init): Register pass_arc_predicate_delay_insns if
289 flag_delayed_branch is active.
291 2013-11-16 Joern Rennecke <joern.rennecke@embecosm.com>
293 * config/arc/constraints.md (Rcq): Simplify register number test.
295 2013-11-15 Aldy Hernandez <aldyh@redhat.com>
297 * gimple.h (enum gf_mask): Change the ordering of GF_OMP_* bits.
299 2013-11-15 Kaz Kojima <kkojima@gcc.gnu.org>
301 * config/sh/sh.c (barrier_align): Return 0 when barrier_or_label
304 2013-11-15 Aldy Hernandez <aldyh@redhat.com>
306 * Makefile.in (C_COMMON_OBJS): Depend on c-cilkplus.o.
307 * gimple-pretty-print.c (dump_omp_for): Add case for
308 GF_OMP_FOR_KIND_CILKSIMD.
309 * gimple.h (enum gf_mask): Restructure entries to add
310 GF_OMP_FOR_KIND_CILKSIMD.
311 * gimplify.c (is_gimple_stmt): Add case for CILK_SIMD.
312 (gimplify_omp_for): Handle CILK_SIMD.
313 (gimplify_expr): Add ccase for CILK_SIMD.
314 * omp-low.c (extract_omp_for_data): Handle CILK_SIMD.
315 (build_outer_var_ref): Same.
316 (check_omp_nesting_restrictions): Same.
317 (lower_rec_input_clauses): Same.
318 (lower_lastprivate_clauses): Same.
319 (expand_omp_for): Same.
320 (execute_expand_omp): Check flag_enable_cilkplus.
321 (execute_lower_omp): Same.
322 (diagnose_sb_0): Handle CILK_SIMD.
323 (diagnose_omp_structured_block_errors): Check flag_enable_cilkplus.
324 (setjmp_or_longjmp_p): New.
325 (scan_omp_1_stmt): Error on setjmp/longjmp in a simd construct.
326 * tree-pretty-print.c (dump_generic_node): Add case for CILK_SIMD.
327 * tree.def: Add tree code for CILK_SIMD.
329 2013-11-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
331 * config/rs6000/altivec.md (UNSPEC_VPERM_X, UNSPEC_VPERM_UNS_X):
333 (altivec_vperm_<mode>): Revert earlier little endian change.
334 (*altivec_vperm_<mode>_internal): Remove.
335 (altivec_vperm_<mode>_uns): Revert earlier little endian change.
336 (*altivec_vperm_<mode>_uns_internal): Remove.
337 * config/rs6000/vector.md (vec_realign_load_<mode>): Revise commentary.
339 2013-11-15 Jeff Law <law@redhat.com>
341 * basic-block.h (has_abnormal_or_eh_outgoing_edge): Renamed from
342 has_abnormal_or_outgoing_edge. Check for EH edges as well.
343 * gimple-ssa-isolate-paths.c
344 (find_implicit_erroneous_behaviour): Corresponding changes.
345 Do not check stmt_ends_bb_p or GIMPLE_RETURN anymore.
346 (find_explicit_erroneous_behaviour): Likewise.
348 2013-11-15 Jeff Law <law@redhat.com>
350 * ifcvt.c (find_cond_trap): Properly handle case where
353 2013-11-15 Andreas Schwab <schwab@linux-m68k.org>
355 * configure: Regenerate.
357 2013-11-15 James Greenhalgh <james.greenhalgh@arm.com>
359 * config/aarch64/aarch64-simd.md: Remove simd_type from all patterns.
360 * config/aarch64/aarch64.md: Likewise, correct "type" attribute
361 where it is incorrect or missing.
363 2013-11-15 Richard Sandiford <rdsandiford@googlemail.com>
365 * dwarf2out.c (gen_enumeration_type_die): Remove unnecessary
367 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
368 Use TREE_INT_CST_LOW rather than tree_low_cst when reading the
370 * fold-const.c (fold_binary_loc): Replace a host_integerp/tree_low_cst
371 pair with a TREE_CODE test and TREE_INT_CST_LOW.
372 * tree-vect-generic.c (expand_vector_divmod): Likewise.
374 2013-11-15 Richard Biener <rguenther@suse.de>
376 PR tree-optimization/50262
377 * tree-ssa-alias.h (struct pt_solution): Split
378 vars_contains_global into vars_contains_nonlocal,
379 vars_contains_escaped and vars_contains_escaped_heap.
380 * tree-ssa-structalias.c (label_visit): Expand comment.
381 (handle_lhs_call): Adjust comment.
382 (set_uids_in_ptset): Set the new flags appropriately.
383 (pt_solution_set): Adjust.
384 (pt_solution_set_var): Likewise.
385 (pt_solution_ior_into): Likewise.
386 (pt_solution_includes_global): Likewise.
387 (pt_solutions_intersect_1): Optimize escaped handling.
388 (compute_points_to_sets): Remove heap variable globalization.
389 (ipa_escaped_pt): Adjust initializer.
390 (pass_data_ipa_pta): Do not run TODO_update_ssa.
391 * gimple-pretty-print.c (pp_points_to_solution): Print split flags.
392 * tree-ssa-alias.c (dump_points_to_solution): Likewise.
394 2013-11-15 Richard Biener <rguenther@suse.de>
396 * tree-loop-distribution.c (tree_loop_distribution): Make sure
397 to distribute all stores.
399 2013-11-15 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
401 * doc/invoke.texi (-mabi=elfv1, -mabi=elfv2): Document.
403 2013-11-15 Joseph Myers <joseph@codesourcery.com>
405 * acinclude.m4 (GCC_GLIBC_VERSION_GTE_IFELSE): New configure macro.
406 * configure.ac: Determine target_header_dir earlier.
407 (--with-glibc-version): New configure option.
408 Use GCC_GLIBC_VERSION_GTE_IFELSE in enable_gnu_unique_object,
409 gcc_cv_libc_provides_ssp and gcc_cv_target_ldbl128 tests.
410 * configure: Regenerate.
411 * doc/install.texi (--enable-gnu-unique-object): Don't refer to
412 native toolchains for default.
413 (--with-glibc-version): Document.
415 2013-11-15 Eric Botcazou <ebotcazou@adacore.com>
417 * fold-const.c (fold_binary_loc) <comparisons>: Reuse local variable.
419 2013-11-15 Uros Bizjak <ubizjak@gmail.com>
421 * lto-streamer-in.c (input function): Call cgraph_create_node if
422 cgraph_get_node failed.
424 2013-11-14 Olivier Hainque <hainque@adacore.com>
426 * cfgexpand.c (defer_stack_allocation): When optimization is enabled,
427 defer allocation of DECL_IGNORED_P variables at toplevel unless really
428 small. Factorize size threshold computation from the existing one.
429 (expand_used_vars): Refine comment.
431 2013-11-14 Cong Hou <congh@google.com>
433 * tree-vectorizer.h (struct dr_with_seg_len): Remove the base
434 address field as it can be obtained from dr. Rename the struct.
435 * tree-vect-data-refs.c (comp_dr_with_seg_len_pair): Consider
436 steps of data references during sort.
437 (vect_prune_runtime_alias_test_list): Adjust with the change to
438 struct dr_with_seg_len.
439 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
440 Adjust with the change to struct dr_with_seg_len.
442 2013-11-14 Jeff Law <law@redhat.com>
445 * basic-block.h (has_abnormal_outgoing_edge_p): Moved here from...
446 * tree-inline.c (has_abnormal_outgoing_edge_p): Remove.
447 * gimple-ssa-isolate-paths.c: Include tree-cfg.h.
448 (find_implicit_erroneous_behaviour): If a block has abnormal outgoing
449 edges, then ignore it. If the statement exhibiting erroneous
450 behaviour ends basic blocks, with the exception of GIMPLE_RETURNs,
451 then we can not optimize.
452 (find_explicit_erroneous_behaviour): Likewise.
454 2013-11-14 Andrew MacLeod <amacleod@redhat.com>
456 * gimplify-me.h: New file. Add prototypes.
457 * gimplify.h: Don't include gimple.h.
458 (struct gimplify_hasher, struct gimplify_ctx, is_gimple_sizepos):
459 Relocate from gimple.h.
460 * gimple.h (struct gimplify_hasher, struct gimplify_ctx,
461 is_gimple_sizepos): Move to gimplify.h.
462 (gimplify_hasher::hash, gimplify_hasher::equal): Move to gimplify.c.
463 (enum gsi_iterator_update): Move to gimple-iterator.h.
464 * gimple-iterator.h (enum gsi_iterator_update): Relocate from gimple.h.
465 * gimplify-me.c: New File.
466 (force_gimple_operand_1, force_gimple_operand,
467 force_gimple_operand_gsi_1, force_gimple_operand_gsi,
468 gimple_regimplify_operands): Relocate from gimplify.c.
469 * gimplify.c (force_gimple_operand_1, force_gimple_operand,
470 force_gimple_operand_gsi_1, force_gimple_operand_gsi,
471 gimple_regimplify_operands): Move to gimplify-me.c.
472 (gimplify_hasher::hash, gimplify_hasher::equal): Relocate
474 * Makefile.in (OBJS): Add gimplify-me.o
475 * asan.c: Include only gimplify.h, gimplify-me.h, and/or gimple.h as
477 * cfgloopmanip.c: Likewise.
478 * cgraphunit.c: Likewise.
479 * cilk-common.c: Likewise.
480 * fold-const.c: Likewise.
481 * function.c: Likewise.
482 * gimple-expr.c: Likewise.
483 * gimple-fold.c: Likewise.
484 * gimple-ssa-strength-reduction.c: Likewise.
485 * gimple.c: Likewise.
486 * graphite-clast-to-gimple.c: Likewise.
487 * graphite-sese-to-poly.c: Likewise.
488 * ipa-prop.c: Likewise.
489 * ipa-split.c: Likewise.
491 * langhooks.c: Likewise.
492 * omp-low.c: Likewise.
494 * stor-layout.c: Likewise.
495 * targhooks.c: Likewise.
496 * trans-mem.c: Likewise.
497 * tree-affine.c: Likewise.
498 * tree-cfg.c: Likewise.
499 * tree-cfgcleanup.c: Likewise.
500 * tree-complex.c: Likewise.
501 * tree-if-conv.c: Likewise.
502 * tree-inline.c: Likewise.
503 * tree-loop-distribution.c: Likewise.
504 * tree-nested.c: Likewise.
505 * tree-parloops.c: Likewise.
506 * tree-predcom.c: Likewise.
507 * tree-profile.c: Likewise.
508 * tree-scalar-evolution.c: Likewise.
509 * tree-sra.c: Likewise.
510 * tree-ssa-address.c: Likewise.
511 * tree-ssa-ccp.c: Likewise.
512 * tree-ssa-dce.c: Likewise.
513 * tree-ssa-forwprop.c: Likewise.
514 * tree-ssa-ifcombine.c: Likewise.
515 * tree-ssa-loop-im.c: Likewise.
516 * tree-ssa-loop-ivopts.c: Likewise.
517 * tree-ssa-loop-manip.c: Likewise.
518 * tree-ssa-loop-niter.c: Likewise.
519 * tree-ssa-loop-prefetch.c: Likewise.
520 * tree-ssa-loop-unswitch.c: Likewise.
521 * tree-ssa-math-opts.c: Likewise.
522 * tree-ssa-phiopt.c: Likewise.
523 * tree-ssa-phiprop.c: Likewise.
524 * tree-ssa-pre.c: Likewise.
525 * tree-ssa-propagate.c: Likewise.
526 * tree-ssa-reassoc.c: Likewise.
527 * tree-ssa-sccvn.c: Likewise.
528 * tree-ssa-strlen.c: Likewise.
529 * tree-ssa.c: Likewise.
530 * tree-switch-conversion.c: Likewise.
531 * tree-tailcall.c: Likewise.
532 * tree-vect-data-refs.c: Likewise.
533 * tree-vect-generic.c: Likewise.
534 * tree-vect-loop-manip.c: Likewise.
535 * tree-vect-loop.c: Likewise.
536 * tree-vect-patterns.c: Likewise.
537 * tree-vect-stmts.c: Likewise.
540 * value-prof.c: Likewise.
541 * config/aarch64/aarch64.c: Likewise.
542 * config/alpha/alpha.c: Likewise.
543 * config/darwin.c: Likewise.
544 * config/i386/i386.c: Likewise.
545 * config/ia64/ia64.c: Likewise.
546 * config/mep/mep.c: Likewise.
547 * config/mips/mips.c: Likewise.
548 * config/rs6000/rs6000.c: Likewise.
549 * config/s390/s390.c: Likewise.
550 * config/sh/sh.c: Likewise.
551 * config/sparc/sparc.c: Likewise.
552 * config/spu/spu.c: Likewise.
553 * config/stormy16/stormy16.c: Likewise.
554 * config/tilegx/tilegx.c: Likewise.
555 * config/tilepro/tilepro.c: Likewise.
556 * config/xtensa/xtensa.c: Likewise.
558 2013-11-14 Joern Rennecke <joern.rennecke@embecosm.com>
560 * config/arc/arc.md (doloop_begin_i): Remove extra alignment;
563 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
565 * config/rs6000/sysv4le.h (LINUX64_DEFAULT_ABI_ELFv2): Define.
567 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
568 Alan Modra <amodra@gmail.com>
570 * config/rs6000/rs6000.h (RS6000_SAVE_AREA): Handle ABI_ELFv2.
571 (RS6000_SAVE_TOC): Remove.
572 (RS6000_TOC_SAVE_SLOT): New macro.
573 * config/rs6000/rs6000.c (rs6000_parm_offset): New function.
574 (rs6000_parm_start): Use it.
575 (rs6000_function_arg_advance_1): Likewise.
576 (rs6000_emit_prologue): Use RS6000_TOC_SAVE_SLOT.
577 (rs6000_emit_epilogue): Likewise.
578 (rs6000_call_aix): Likewise.
579 (rs6000_output_function_prologue): Do not save/restore r11
580 around calling _mcount for ABI_ELFv2.
582 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
583 Alan Modra <amodra@gmail.com>
585 * config/rs6000/rs6000-protos.h (rs6000_reg_parm_stack_space):
587 * config/rs6000/rs6000.h (RS6000_REG_SAVE): Remove.
588 (REG_PARM_STACK_SPACE): Call rs6000_reg_parm_stack_space.
589 * config/rs6000/rs6000.c (rs6000_parm_needs_stack): New function.
590 (rs6000_function_parms_need_stack): Likewise.
591 (rs6000_reg_parm_stack_space): Likewise.
592 (rs6000_function_arg): Do not replace BLKmode by Pmode when
593 returning a register argument.
595 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
596 Michael Gschwind <mkg@us.ibm.com>
598 * config/rs6000/rs6000.h (FP_ARG_MAX_RETURN): New macro.
599 (ALTIVEC_ARG_MAX_RETURN): Likewise.
600 (FUNCTION_VALUE_REGNO_P): Use them.
601 * config/rs6000/rs6000.c (TARGET_RETURN_IN_MSB): Define.
602 (rs6000_return_in_msb): New function.
603 (rs6000_return_in_memory): Handle ELFv2 homogeneous aggregates.
604 Handle aggregates of up to 16 bytes for ELFv2.
605 (rs6000_function_value): Handle ELFv2 homogeneous aggregates.
607 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
608 Michael Gschwind <mkg@us.ibm.com>
610 * config/rs6000/rs6000.h (AGGR_ARG_NUM_REG): Define.
611 * config/rs6000/rs6000.c (rs6000_aggregate_candidate): New function.
612 (rs6000_discover_homogeneous_aggregate): Likewise.
613 (rs6000_function_arg_boundary): Handle homogeneous aggregates.
614 (rs6000_function_arg_advance_1): Likewise.
615 (rs6000_function_arg): Likewise.
616 (rs6000_arg_partial_bytes): Likewise.
617 (rs6000_psave_function_arg): Handle BLKmode arguments.
619 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
620 Michael Gschwind <mkg@us.ibm.com>
622 * config/rs6000/rs6000.h (AGGR_ARG_NUM_REG): Define.
623 * config/rs6000/rs6000.c (rs6000_aggregate_candidate): New function.
624 (rs6000_discover_homogeneous_aggregate): Likewise.
625 (rs6000_function_arg_boundary): Handle homogeneous aggregates.
626 (rs6000_function_arg_advance_1): Likewise.
627 (rs6000_function_arg): Likewise.
628 (rs6000_arg_partial_bytes): Likewise.
629 (rs6000_psave_function_arg): Handle BLKmode arguments.
631 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
633 * config/rs6000/rs6000.c (machine_function): New member
635 (rs6000_emit_prologue): Set r2_setup_needed if necessary.
636 (rs6000_output_mi_thunk): Set r2_setup_needed.
637 (rs6000_output_function_prologue): Output global entry point
638 prologue and local entry point marker if needed for ABI_ELFv2.
639 Output -mprofile-kernel code here.
640 (output_function_profiler): Do not output -mprofile-kernel
641 code here; moved to rs6000_output_function_prologue.
642 (rs6000_file_start): Output ".abiversion 2" for ABI_ELFv2.
644 (rs6000_emit_move): Do not handle dot symbols for ABI_ELFv2.
645 (rs6000_output_function_entry): Likewise.
646 (rs6000_assemble_integer): Likewise.
647 (rs6000_elf_encode_section_info): Likewise.
648 (rs6000_elf_declare_function_name): Do not create dot symbols
649 or .opd section for ABI_ELFv2.
651 (rs6000_trampoline_size): Update for ABI_ELFv2 trampolines.
652 (rs6000_trampoline_init): Likewise.
653 (rs6000_elf_file_end): Call file_end_indicate_exec_stack for ABI_ELFv2.
655 (rs6000_call_aix): Handle ELFv2 indirect calls. Do not check
656 for function descriptors in ABI_ELFv2.
658 * config/rs6000/rs6000.md ("*call_indirect_aix<mode>"): Support
659 on ABI_AIX only, not ABI_ELFv2.
660 ("*call_value_indirect_aix<mode>"): Likewise.
661 ("*call_indirect_elfv2<mode>"): New pattern.
662 ("*call_value_indirect_elfv2<mode>"): Likewise.
664 * config/rs6000/predicates.md ("symbol_ref_operand"): Do not
665 check for function descriptors in ABI_ELFv2.
666 ("current_file_function_operand"): Likewise.
668 * config/rs6000/ppc-asm.h [__powerpc64__ && _CALL_ELF == 2]:
670 (FUNC_NAME): Define ELFv2 variant.
671 (JUMP_TARGET): Likewise.
672 (FUNC_START): Likewise.
673 (HIDDEN_FUNC): Likewise.
674 (FUNC_END): Likeiwse.
676 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
678 * config.gcc [powerpc*-*-* | rs6000-*-*]: Support --with-abi=elfv1
679 and --with-abi=elfv2.
680 * config/rs6000/option-defaults.h (OPTION_DEFAULT_SPECS): Add "abi".
681 * config/rs6000/rs6000.opt (mabi=elfv1): New option.
682 (mabi=elfv2): Likewise.
683 * config/rs6000/rs6000-opts.h (enum rs6000_abi): Add ABI_ELFv2.
684 * config/rs6000/linux64.h (DEFAULT_ABI): Do not hard-code to AIX_ABI
686 (ELFv2_ABI_CHECK): New macro.
687 (SUBSUBTARGET_OVERRIDE_OPTIONS): Use it to decide whether to set
688 rs6000_current_abi to ABI_AIX or ABI_ELFv2.
689 (GLIBC_DYNAMIC_LINKER64): Support ELFv2 ld.so version.
690 * config/rs6000/rs6000-c.c (rs6000_cpu_cpp_builtins): Predefine
691 _CALL_ELF and __STRUCT_PARM_ALIGN__ if appropriate.
693 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Handle ABI_ELFv2.
694 (debug_stack_info): Likewise.
695 (rs6000_file_start): Treat ABI_ELFv2 the same as ABI_AIX.
696 (rs6000_legitimize_tls_address): Likewise.
697 (rs6000_conditional_register_usage): Likewise.
698 (rs6000_emit_move): Likewise.
699 (init_cumulative_args): Likewise.
700 (rs6000_function_arg_advance_1): Likewise.
701 (rs6000_function_arg): Likewise.
702 (rs6000_arg_partial_bytes): Likewise.
703 (rs6000_output_function_entry): Likewise.
704 (rs6000_assemble_integer): Likewise.
705 (rs6000_savres_strategy): Likewise.
706 (rs6000_stack_info): Likewise.
707 (rs6000_function_ok_for_sibcall): Likewise.
708 (rs6000_emit_load_toc_table): Likewise.
709 (rs6000_savres_routine_name): Likewise.
710 (ptr_regno_for_savres): Likewise.
711 (rs6000_emit_prologue): Likewise.
712 (rs6000_emit_epilogue): Likewise.
713 (rs6000_output_function_epilogue): Likewise.
714 (output_profile_hook): Likewise.
715 (output_function_profiler): Likewise.
716 (rs6000_trampoline_size): Likewise.
717 (rs6000_trampoline_init): Likewise.
718 (rs6000_elf_output_toc_section_asm_op): Likewise.
719 (rs6000_elf_encode_section_info): Likewise.
720 (rs6000_elf_reloc_rw_mask): Likewise.
721 (rs6000_elf_declare_function_name): Likewise.
722 (rs6000_function_arg_boundary): Treat ABI_ELFv2 the same as ABI_AIX,
723 except that rs6000_compat_align_parm is always assumed false.
724 (rs6000_gimplify_va_arg): Likewise.
725 (rs6000_call_aix): Update comment.
726 (rs6000_sibcall_aix): Likewise.
727 * config/rs6000/rs6000.md ("tls_gd_aix<TLSmode:tls_abi_suffix>"):
728 Treat ABI_ELFv2 the same as ABI_AIX.
729 ("*tls_gd_call_aix<TLSmode:tls_abi_suffix>"): Likewise.
730 ("tls_ld_aix<TLSmode:tls_abi_suffix>"): Likewise.
731 ("*tls_ld_call_aix<TLSmode:tls_abi_suffix>"): Likewise.
732 ("load_toc_aix_si"): Likewise.
733 ("load_toc_aix_di"): Likewise.
735 ("call_value"): Likewise.
736 ("*call_local_aix<mode>"): Likewise.
737 ("*call_value_local_aix<mode>"): Likewise.
738 ("*call_nonlocal_aix<mode>"): Likewise.
739 ("*call_value_nonlocal_aix<mode>"): Likewise.
740 ("*call_indirect_aix<mode>"): Likewise.
741 ("*call_value_indirect_aix<mode>"): Likewise.
742 ("sibcall"): Likewise.
743 ("sibcall_value"): Likewise.
744 ("*sibcall_aix<mode>"): Likewise.
745 ("*sibcall_value_aix<mode>"): Likewise.
746 * config/rs6000/predicates.md ("symbol_ref_operand"): Likewise.
747 ("current_file_function_operand"): Likewise.
749 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
751 * config/rs6000/rs6000.c (rs6000_arg_partial_bytes): Simplify logic
752 by making use of the fact that for vector / floating point arguments
753 passed both in VRs/FPRs and in the fixed parameter area, the partial
754 bytes mechanism is in fact not used.
756 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
758 * config/rs6000/rs6000.c (rs6000_psave_function_arg): New function.
759 (rs6000_finish_function_arg): Likewise.
760 (rs6000_function_arg): Use rs6000_psave_function_arg and
761 rs6000_finish_function_arg to handle both vector and floating
762 point arguments that are also passed in GPRs / the stack.
764 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
766 * config/rs6000/rs6000.c (USE_FP_FOR_ARG_P): Remove TYPE argument.
767 (USE_ALTIVEC_FOR_ARG_P): Likewise.
768 (rs6000_darwin64_record_arg_advance_recurse): Update uses.
769 (rs6000_function_arg_advance_1):Likewise.
770 (rs6000_darwin64_record_arg_recurse): Likewise.
771 (rs6000_function_arg): Likewise.
772 (rs6000_arg_partial_bytes): Likewise.
774 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
776 * config/rs6000/rs6000.c (rs6000_option_override_internal): Replace
777 "DEFAULT_ABI != ABI_AIX" test by testing for ABI_V4 or ABI_DARWIN.
778 (rs6000_savres_strategy): Likewise.
779 (rs6000_return_addr): Likewise.
780 (rs6000_emit_load_toc_table): Replace "DEFAULT_ABI != ABI_AIX" by
781 testing for ABI_V4 (since ABI_DARWIN is impossible here).
782 (rs6000_emit_prologue): Likewise.
783 (legitimate_lo_sum_address_p): Simplify DEFAULT_ABI test.
784 (rs6000_elf_declare_function_name): Remove duplicated test.
785 * config/rs6000/rs6000.md ("load_toc_v4_PIC_1"): Explicitly test
786 for ABI_V4 (instead of "DEFAULT_ABI != ABI_AIX" test).
787 ("load_toc_v4_PIC_1_normal"): Likewise.
788 ("load_toc_v4_PIC_1_476"): Likewise.
789 ("load_toc_v4_PIC_1b"): Likewise.
790 ("load_toc_v4_PIC_1b_normal"): Likewise.
791 ("load_toc_v4_PIC_1b_476"): Likewise.
792 ("load_toc_v4_PIC_2"): Likewise.
793 ("load_toc_v4_PIC_3b"): Likewise.
794 ("load_toc_v4_PIC_3c"): Likewise.
795 * config/rs6000/rs6000.h (RS6000_REG_SAVE): Simplify DEFAULT_ABI test.
796 (RS6000_SAVE_AREA): Likewise.
797 (FP_ARG_MAX_REG): Likewise.
798 (RETURN_ADDRESS_OFFSET): Likewise.
799 * config/rs6000/sysv.h (TARGET_TOC): Test for ABI_V4 instead
801 (SUBTARGET_OVERRIDE_OPTIONS): Likewise.
802 (MINIMAL_TOC_SECTION_ASM_OP): Likewise.
804 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
806 * config/rs6000/rs6000.c (rs6000_call_indirect_aix): Rename to ...
807 (rs6000_call_aix): ... this. Handle both direct and indirect calls.
808 Create call insn directly instead of via various gen_... routines.
809 Mention special registers used by the call in CALL_INSN_FUNCTION_USAGE.
810 (rs6000_sibcall_aix): New function.
811 * config/rs6000/rs6000.md (TOC_SAVE_OFFSET_32BIT): Remove.
812 (TOC_SAVE_OFFSET_64BIT): Likewise.
813 (AIX_FUNC_DESC_TOC_32BIT): Likewise.
814 (AIX_FUNC_DESC_TOC_64BIT): Likewise.
815 (AIX_FUNC_DESC_SC_32BIT): Likewise.
816 (AIX_FUNC_DESC_SC_64BIT): Likewise.
817 ("call" expander): Call rs6000_call_aix.
818 ("call_value" expander): Likewise.
819 ("call_indirect_aix<ptrsize>"): Replace this pattern ...
820 ("call_indirect_aix<ptrsize>_nor11"): ... and this pattern ...
821 ("*call_indirect_aix<mode>"): ... by this insn pattern.
822 ("call_value_indirect_aix<ptrsize>"): Replace this pattern ...
823 ("call_value_indirect_aix<ptrsize>_nor11"): ... and this pattern ...
824 ("*call_value_indirect_aix<mode>"): ... by this insn pattern.
825 ("*call_nonlocal_aix32", "*call_nonlocal_aix64"): Replace by ...
826 ("*call_nonlocal_aix<mode>"): ... this pattern.
827 ("*call_value_nonlocal_aix32", "*call_value_nonlocal_aix64"): Replace
828 ("*call_value_nonlocal_aix<mode>"): ... by this pattern.
829 ("*call_local_aix<mode>"): New insn pattern.
830 ("*call_value_local_aix<mode>"): Likewise.
831 ("sibcall" expander): Call rs6000_sibcall_aix.
832 ("sibcall_value" expander): Likewise. Move earlier in file.
833 ("*sibcall_nonlocal_aix<mode>"): Replace by ...
834 ("*sibcall_aix<mode>"): ... this pattern.
835 ("*sibcall_value_nonlocal_aix<mode>"): Replace by ...
836 ("*sibcall_value_aix<mode>"): ... this pattern.
837 * config/rs6000/rs6000-protos.h (rs6000_call_indirect_aix): Remove.
838 (rs6000_call_aix): Add prototype.
839 (rs6000_sibcall_aix): Likewise.
841 2013-11-14 Jakub Jelinek <jakub@redhat.com>
844 * asan.c (asan_emit_stack_protection): Ensure -fsection-anchors
845 isn't confused by the artificial decl.
847 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
849 * config/rs6000/rs6000.c (rs6000_emit_prologue): Do not place a
850 RTX_FRAME_RELATED_P marker on the UNSPEC_MOVESI_FROM_CR insn.
851 Instead, add USEs of all modified call-saved CR fields to the
852 insn storing the result to the stack slot, and provide an
853 appropriate REG_FRAME_RELATED_EXPR for that insn.
854 * config/rs6000/rs6000.md ("*crsave"): New insn pattern.
855 * config/rs6000/predicates.md ("crsave_operation"): New predicate.
857 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
858 Alan Modra <amodra@gmail.com>
860 * function.c (assign_parms): Use all.reg_parm_stack_space instead
861 of re-evaluating REG_PARM_STACK_SPACE target macro.
862 (locate_and_pad_parm): New parameter REG_PARM_STACK_SPACE. Use it
863 instead of evaluating target macro REG_PARM_STACK_SPACE every time.
864 (assign_parm_find_entry_rtl): Update call.
865 * calls.c (initialize_argument_information): Update call.
866 (emit_library_call_value_1): Likewise.
867 * expr.h (locate_and_pad_parm): Update prototype.
869 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
871 * calls.c (store_unaligned_arguments_into_pseudos): Skip PARALLEL
874 2013-11-14 DJ Delorie <dj@redhat.com>
876 * config/rx/rx.c (rx_mode_dependent_address_p): Allow offsets up
879 2013-11-14 Jeff Law <law@redhat.com>
881 * tree-ssa-threadedge.c (thread_through_normal_block): Only push the
882 EDGE_START_JUMP_THREAD marker if the jump threading path is empty.
884 2013-11-14 James Greenhalgh <james.greenhalgh@arm.com>
886 * doc/invoke.texi: Update documentation for AArch64's -mcpu
889 2013-11-14 James Greenhalgh <james.greenhalgh@arm.com>
891 * config/aarch64/aarch64-cores.def (example-1): Remove.
892 (example-2): Likewise.
893 * config/aarch64/aarch64-tune.md: Regenerate.
894 * config/aarch64/aarch64.md: Do not include "large.md" or "small.md".
895 (generic_sched): Remove "large", "small".
896 * config/aarch64/large.md: Delete.
897 * config/aarch64/small.md: Delete.
899 2013-11-14 James Greenhalgh <james.greenhalgh@arm.com>
901 * config/aarch64/aarch64-cores.def (cortex-a57): Tune for cortexa15.
902 * config/aarch64/aarch64-tune.md: Regenerate.
903 * config/aarch64/aarch64.md: Include cortex-a15 pipeline model.
904 (generic_sched): "no" if we are tuning for cortexa15.
905 * config/arm/cortex-a15.md: Include cortex-a15-neon.md by
908 2013-11-14 James Greenhalgh <james.greenhalgh@arm.com>
910 * config/aarch64/aarch64-arches.def (armv8-a): Tune for cortex-a53.
911 * config/aarch64/aarch64.md: Do not include aarch64-generic.md.
912 * config/aarch64/aarch64.c (aarch64_tune): Initialize to cortexa53.
913 (all_cores): Use cortexa53 when tuning for "generic".
914 (aarch64_override_options): Fix comment.
915 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Set to cortexa53.
916 * config/aarch64/aarch64-generic.md: Delete.
918 2013-11-14 James Greenhalgh <james.greenhalgh@arm.com>
920 * config/aarch64/aarch64.c (all_architectures): Remove "generic".
922 2013-11-14 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
924 * config/aarch64/aarch64.c: Include aarch-cost-tables.h.
925 (generic_rtx_cost_table): Remove.
926 (aarch64_rtx_costs): Use fields from cpu_cost_table.
927 * config/aarch64/aarch64-protos.h (tune_params): Use cpu_cost_table for
929 (cpu_rtx_cost_table): Remove.
931 2013-11-14 Julian Brown <julian@codesourcery.com>
932 Joey Ye <joey.ye@arm.com>
934 * config/arm/arm.c (arm_cortex_m_branch_cost): New.
936 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune, arm_9e_tune,
937 arm_v6t2_tune, arm_cortex_tune, arm_cortex_a15_tune,
938 arm_cortex_a5_tune, arm_v6m_tune): Add comments for Sched adj cost.
939 * config/arm/arm-cores.def (cortex-m4, cortex-m3): Use arm_v7m_tune.
941 2013-11-14 Kirill Yukhin <kirill.yukhin@intel.com>
944 * config/ia64/ia64.c (ia64_split_tmode_move): Relax `dead'
947 2013-11-14 Jakub Jelinek <jakub@redhat.com>
948 Uros Bizjak <ubizjak@gmail.com>
951 * config/i386/i386.md (*anddi_2): Only allow CCZmode if
952 operands[2] satisfies_constraint_Z that might have bit 31 set.
954 2013-11-13 Jeff Law <law@redhat.com>
956 PR tree-optimization/59102
957 * gimple-ssa-isolate-paths.c
958 (insert_trap_and_remove_trailing_statments): Ensure STMT is a
959 gimple assignment before looking at gimple_assign_lhs.
961 2013-11-13 Vladimir Makarov <vmakarov@redhat.com>
963 * ira.c: Add comment about threads at the top of file.
965 2013-11-13 Vladimir Makarov <vmakarov@redhat.com>
967 * ira-color.c (coalesce_allocnos): Don't allocate and free
970 2013-11-14 Tom de Vries <tom@codesourcery.com>
972 * tree-ssa-tail-merge.c (gimple_equal_p): Add test for structural
973 equality for GIMPLE_ASSIGN.
975 2013-11-14 Tom de Vries <tom@codesourcery.com>
977 * tree-ssa-tail-merge.c (gimple_operand_equal_value_p): Factor new
979 (gimple_equal_p): ... here.
981 2013-11-14 Tom de Vries <tom@codesourcery.com>
983 * trans-mem.c (is_tm_ending): New function.
984 * gimple.h (is_tm_ending): Declare.
985 * tree-ssa-tail-merge.c (gimple_equal_p): Remove test on
987 (find_duplicate): Use is_tm_ending instead of is_tm_ending_fndecl.
989 2013-11-14 Tom de Vries <tom@codesourcery.com>
991 * tree-ssa-tail-merge.c (gimple_equal_p): Remove equal variable.
993 2013-11-13 Andrew MacLeod <amacleod@redhat.com>
995 * gimple-walk.h: New File. Relocate prototypes from gimple.h.
996 (struct walk_stmt_info): Relocate here from gimple.h.
997 * gimple-iterator.h: New File. Relocate prototypes from gimple.h.
998 (struct gimple_stmt_iterator_d): Relocate here from gimple.h.
999 (gsi_start_1, gsi_none, gsi_start_bb, gsi_last_1, gsi_last_bb,
1000 gsi_end_p, gsi_one_before_end_p, gsi_next, gsi_prev, gsi_stmt,
1001 gsi_after_labels, gsi_next_nondebug, gsi_prev_nondebug,
1002 gsi_start_nondebug_bb, gsi_start_nondebug_after_labels_bb,
1003 gsi_last_nondebug_bb, gsi_bb, gsi_seq): Relocate here from gimple.h.
1004 * gimple.h (struct gimple_stmt_iterator_d): Move to gimple-iterator.h.
1005 (gsi_start_1, gsi_none, gsi_start_bb, gsi_last_1, gsi_last_bb,
1006 gsi_end_p, gsi_one_before_end_p, gsi_next, gsi_prev, gsi_stmt,
1007 gsi_after_labels, gsi_next_nondebug, gsi_prev_nondebug,
1008 gsi_start_nondebug_bb, gsi_start_nondebug_after_labels_bb,
1009 gsi_last_nondebug_bb, gsi_bb, gsi_seq): Move to gimple-iterator.h.
1010 (struct walk_stmt_info): Move to gimple-walk.h.
1011 (gimple_seq_set_location): Move to gimple.c
1012 * gimple-walk.c: New File.
1013 (walk_gimple_seq_mod, walk_gimple_seq, walk_gimple_asm, walk_gimple_op,
1014 walk_gimple_stmt, get_base_loadstore, walk_stmt_load_store_addr_ops,
1015 walk_stmt_load_store_ops): Relocate here from gimple.c.
1016 * gimple-iterator.c: Include gimple-iterator.h.
1017 * gimple.c (walk_gimple_seq_mod, walk_gimple_seq, walk_gimple_asm,
1018 walk_gimple_op, walk_gimple_stmt, get_base_loadstore,
1019 walk_stmt_load_store_addr_ops, walk_stmt_load_store_ops): Move to
1021 (gimple_seq_set_location): Relocate from gimple.h.
1022 * tree-phinodes.h (set_phi_nodes): Move to tree-phinodes.c.
1023 * tree-phinodes.c (set_phi_nodes): Relocate from tree-phinodes.h.
1024 * gengtype.c (open_base_files): Add gimple-iterator.h to include list.
1025 * Makefile.in (OBJS): Add gimple-walk.o
1026 * asan.c: Update Include list as required for gimple-iterator.h and
1028 * cfgexpand.c: Likewise.
1029 * cfgloop.c: Likewise.
1030 * cfgloopmanip.c: Likewise.
1031 * cgraph.c: Likewise.
1032 * cgraphbuild.c: Likewise.
1033 * cgraphunit.c: Likewise.
1034 * gimple-fold.c: Likewise.
1035 * gimple-low.c: Likewise.
1036 * gimple-pretty-print.c: Likewise.
1037 * gimple-ssa-isolate-paths.c: Likewise.
1038 * gimple-ssa-strength-reduction.c: Likewise.
1039 * gimple-streamer-in.c: Likewise.
1040 * gimple-streamer-out.c: Likewise.
1041 * gimplify.c: Likewise.
1042 * graphite-blocking.c: Likewise.
1043 * graphite-clast-to-gimple.c: Likewise.
1044 * graphite-dependences.c: Likewise.
1045 * graphite-interchange.c: Likewise.
1046 * graphite-optimize-isl.c: Likewise.
1047 * graphite-poly.c: Likewise.
1048 * graphite-scop-detection.c: Likewise.
1049 * graphite-sese-to-poly.c: Likewise.
1050 * graphite.c: Likewise.
1051 * ipa-inline-analysis.c: Likewise.
1052 * ipa-profile.c: Likewise.
1053 * ipa-prop.c: Likewise.
1054 * ipa-pure-const.c: Likewise.
1055 * ipa-split.c: Likewise.
1056 * lto-streamer-in.c: Likewise.
1057 * lto-streamer-out.c: Likewise.
1058 * omp-low.c: Likewise.
1059 * predict.c: Likewise.
1060 * profile.c: Likewise.
1062 * tracer.c: Likewise.
1063 * trans-mem.c: Likewise.
1064 * tree-call-cdce.c: Likewise.
1065 * tree-cfg.c: Likewise.
1066 * tree-cfgcleanup.c: Likewise.
1067 * tree-complex.c: Likewise.
1068 * tree-data-ref.c: Likewise.
1069 * tree-dfa.c: Likewise.
1070 * tree-eh.c: Likewise.
1071 * tree-emutls.c: Likewise.
1072 * tree-if-conv.c: Likewise.
1073 * tree-inline.c: Likewise.
1074 * tree-into-ssa.c: Likewise.
1075 * tree-loop-distribution.c: Likewise.
1076 * tree-nested.c: Likewise.
1077 * tree-nrv.c: Likewise.
1078 * tree-object-size.c: Likewise.
1079 * tree-outof-ssa.c: Likewise.
1080 * tree-parloops.c: Likewise.
1081 * tree-predcom.c: Likewise.
1082 * tree-profile.c: Likewise.
1083 * tree-scalar-evolution.c: Likewise.
1084 * tree-sra.c: Likewise.
1085 * tree-ssa-ccp.c: Likewise.
1086 * tree-ssa-coalesce.c: Likewise.
1087 * tree-ssa-copy.c: Likewise.
1088 * tree-ssa-copyrename.c: Likewise.
1089 * tree-ssa-dce.c: Likewise.
1090 * tree-ssa-dom.c: Likewise.
1091 * tree-ssa-dse.c: Likewise.
1092 * tree-ssa-forwprop.c: Likewise.
1093 * tree-ssa-ifcombine.c: Likewise.
1094 * tree-ssa-live.c: Likewise.
1095 * tree-ssa-loop-ch.c: Likewise.
1096 * tree-ssa-loop-im.c: Likewise.
1097 * tree-ssa-loop-ivcanon.c: Likewise.
1098 * tree-ssa-loop-ivopts.c: Likewise.
1099 * tree-ssa-loop-manip.c: Likewise.
1100 * tree-ssa-loop-niter.c: Likewise.
1101 * tree-ssa-loop-prefetch.c: Likewise.
1102 * tree-ssa-loop.c: Likewise.
1103 * tree-ssa-math-opts.c: Likewise.
1104 * tree-ssa-phiopt.c: Likewise.
1105 * tree-ssa-phiprop.c: Likewise.
1106 * tree-ssa-pre.c: Likewise.
1107 * tree-ssa-propagate.c: Likewise.
1108 * tree-ssa-reassoc.c: Likewise.
1109 * tree-ssa-sink.c: Likewise.
1110 * tree-ssa-strlen.c: Likewise.
1111 * tree-ssa-structalias.c: Likewise.
1112 * tree-ssa-tail-merge.c: Likewise.
1113 * tree-ssa-ter.c: Likewise.
1114 * tree-ssa-threadedge.c: Likewise.
1115 * tree-ssa-threadupdate.c: Likewise.
1116 * tree-ssa-uncprop.c: Likewise.
1117 * tree-ssa-uninit.c: Likewise.
1118 * tree-ssa.c: Likewise.
1119 * tree-stdarg.c: Likewise.
1120 * tree-switch-conversion.c: Likewise.
1121 * tree-tailcall.c: Likewise.
1122 * tree-vect-data-refs.c: Likewise.
1123 * tree-vect-generic.c: Likewise.
1124 * tree-vect-loop-manip.c: Likewise.
1125 * tree-vect-loop.c: Likewise.
1126 * tree-vect-patterns.c: Likewise.
1127 * tree-vect-slp.c: Likewise.
1128 * tree-vect-stmts.c: Likewise.
1129 * tree-vectorizer.c: Likewise.
1130 * tree-vrp.c: Likewise.
1133 * value-prof.c: Likewise.
1134 * vtable-verify.c: Likewise.
1136 2013-11-13 Steven Bosscher <steven@gcc.gnu.org>
1138 * gimple-ssa-isolate-paths.c (pass_isolate_erroneous_paths): Comment
1141 2013-11-13 Jeff Law <law@redhat.com>
1143 * PR middle-end/59119
1144 * gimple-ssa-isolate-paths.c (find_implicit_erroneous_behaviour): New
1145 function, extracted from gimple_ssa_isolate_erroneous_paths.
1146 (find_explicit_erroneous_behaviour): Similarly.
1147 (insert_trap_and_remove_trailing_statements): Remove statements
1150 2013-11-13 Steven Bosscher <steven@gcc.gnu.org>
1152 * cfgrtl.c (can_fallthru): Reorder code to move tablejump check up.
1153 Make that check explicit. BB_HEAD cannot be NULL, remove check for it.
1154 * haifa-sched.c (ready_remove_first_dispatch): Check INSN_P before
1155 looking at INSN_CODE.
1156 * reload1.c (delete_dead_insn) Do not expect JUMP_TABLE_DATA to be an
1157 active_insn_p object, respect basic block boundaries.
1158 * reorg.c (follow_jumps): Use invariant that JUMP_TABLE_DATA always
1159 follows immediately after the jump table data label.
1160 * config/nds32/nds32.c (nds32_output_casesi_pc_relative): Likewise.
1161 * config/sh/sh.c (barrier_align): Likewise. Rearrange code such
1162 that JUMP_TABLE_DATA is not expected to be an active_insn_p object.
1164 2013-11-13 Teresa Johnson <tejohnson@google.com>
1167 * predict.c (drop_profile): Error is currently too strict.
1168 (handle_missing_profiles): Pass call_count to drop_profile.
1170 2013-11-13 Teresa Johnson <tejohnson@google.com>
1173 * ipa-inline.c (edge_badness): Fix overflow.
1175 2013-11-13 Vladimir Makarov <vmakarov@redhat.com>
1177 PR rtl-optimization/59036
1178 * ira-color.c (struct allocno_color_data): Add new members
1179 first_thread_allocno, next_thread_allocno, thread_freq.
1180 (sorted_copies): New static var.
1181 (allocnos_conflict_by_live_ranges_p, copy_freq_compare_func): Move up.
1182 (allocno_thread_conflict_p, merge_threads)
1183 (form_threads_from_copies, form_threads_from_bucket)
1184 (form_threads_from_colorable_allocno, init_allocno_threads): New
1186 (bucket_allocno_compare_func): Add comparison by thread frequency
1188 (add_allocno_to_ordered_bucket): Rename to
1189 add_allocno_to_ordered_colorable_bucket. Remove parameter.
1190 (push_only_colorable): Call form_threads_from_bucket.
1191 (color_pass): Call init_allocno_threads. Use
1192 consideration_allocno_bitmap instead of coloring_allocno_bitmap
1193 for nuillify allocno color data.
1194 (ira_initiate_assign, ira_finish_assign): Allocate/free sorted_copies.
1195 (coalesce_allocnos): Use static sorted copies.
1197 2013-11-13 Jakub Jelinek <jakub@redhat.com>
1199 * passes.c (execute_todo): Don't call do_per_function if
1201 (execute_one_ipa_transform_pass, execute_one_pass): Don't call
1202 execute_function_dump if dump_file is NULL.
1204 2013-11-13 Martin Jambor <mjambor@suse.cz>
1206 * cgraph.c (cgraph_get_create_node): Do what
1207 cgraph_get_create_real_symbol_node used to do.
1208 (cgraph_get_create_real_symbol_node): Removed. Changed all users to
1209 call cgraph_get_create_node.
1210 * cgraph.h (cgraph_get_create_real_symbol_node): Removed.
1211 * lto-streamer-in.c (input_function): Call cgraph_get_node instead of
1212 cgraph_get_create_node. Assert we get a node.
1214 2013-11-13 Tejas Belagod <tejas.belagod@arm.com>
1216 * config/aarch64/aarch64-simd.md (vec_extract): New.
1218 2013-11-13 Tejas Belagod <tejas.belagod@arm.com>
1220 * config/aarch64/aarch64-simd.md (vec_set<mode>): Add w -> w option to
1223 2013-11-13 Eric Botcazou <ebotcazou@adacore.com>
1225 * cfgexpand.c (expand_used_vars): Allocate space for partitions based
1226 on PARM_DECLs or RESULT_DECLs only if they are ignored for debug info
1227 or if optimization is enabled.
1228 * tree-ssa-coalesce.c (coalesce_ssa_name): If optimization is disabled,
1229 require that all the names based on a PARM_DECL or a RESULT_DECL that
1230 isn't ignored for debug info be coalesced.
1232 2013-11-13 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1234 * config/c6x/c6x.c: Include "gimple-expr.h".
1236 2013-11-13 Richard Biener <rguenther@suse.de>
1238 * gimple-streamer-out.c (output_gimple_stmt): Also wrap
1239 decls in ADDR_EXPR operands inside a MEM_REF and optimize that.
1240 * gimple-streamer-in.c (input_gimple_stmt): Remove now dead code
1241 dealing with type mismatches inside component reference chains.
1243 2013-11-13 Marc Glisse <marc.glisse@inria.fr>
1245 PR tree-optimization/59077
1246 * ipa-pure-const.c (better_state): Update *state.
1248 2013-11-13 Christophe Lyon <christophe.lyon@linaro.org>
1250 * config/aarch64/aarch64.h (FRAME_GROWS_DOWNWARD): Define to 1.
1251 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
1252 Update offset calculations.
1254 2013-11-13 Eric Botcazou <ebotcazou@adacore.com>
1257 * dwarf2out.c (add_byte_size_attribute): Also use int_size_in_bytes
1258 for fields. Do not add the attribute if the size is negative.
1260 2013-11-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1262 * config/arm/arm.c: Include aarch-cost-tables.h.
1263 (generic_extra_costs): Move from here...
1264 * config/arm/aarch-cost-tables.h: ... To here. New file.
1266 2013-11-13 Alexander Ivchenko <alexander.ivchenko@intel.com>
1267 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1268 Sergey Lega <sergey.s.lega@intel.com>
1269 Anna Tikhonova <anna.tikhonova@intel.com>
1270 Ilya Tocar <ilya.tocar@intel.com>
1271 Andrey Turetskiy <andrey.turetskiy@intel.com>
1272 Ilya Verbin <ilya.verbin@intel.com>
1273 Kirill Yukhin <kirill.yukhin@intel.com>
1274 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1276 * config/i386/i386.c (ix86_print_operand): Support z-masking.
1277 * config/i386/predicate.md (const_0_to_4_operand): New.
1278 (const_0_to_5_operand): Ditto.
1279 * config/i386/sse.md (UNSPEC_COMPRESS): New.
1280 (UNSPEC_COMPRESS_STORE): Ditto.
1281 (UNSPEC_EXPAND): Ditto.
1282 (UNSPEC_EMBEDDED_ROUNDING): Ditto.
1283 (define_mode_attr ssescalarsize): Ditto.
1284 (avx512f_load<mode>_mask): Ditto.
1285 (avx512f_store<mode>_mask): Ditto.
1286 (avx512f_storedqu<mode>_mask): Ditto.
1287 (avx512f_vmcmp<mode>3_mask): Ditto.
1288 (avx512f_fmadd_<mode>_mask): Ditto.
1289 (avx512f_fmadd_<mode>_mask3): Ditto.
1290 (avx512f_fmsub_<mode>_mask): Ditto.
1291 (avx512f_fmsub_<mode>_mask3): Ditto.
1292 (avx512f_fnmadd_<mode>_mask): Ditto.
1293 (avx512f_fnmadd_<mode>_mask3): Ditto.
1294 (avx512f_fnmsub_<mode>_mask): Ditto.
1295 (avx512f_fnmsub_<mode>_mask3): Ditto.
1296 (avx512f_fmaddsub_<mode>_mask): Ditto.
1297 (avx512f_fmaddsub_<mode>_mask3): Ditto.
1298 (avx512f_fmsubadd_<mode>_mask): Ditto.
1299 (avx512f_fmsubadd_<mode>_mask3): Ditto.
1300 (vec_unpacku_float_lo_v16si): Ditto.
1301 (avx512f_vextract<shuffletype>32x4_mask): Ditto.
1302 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
1303 (avx512f_vextract<shuffletype>64x4_mask): Ditto.
1304 (vec_extract_lo_<mode>_maskm): Ditto.
1305 (vec_extract_hi_<mode>_maskm): Ditto.
1306 (avx512f_vternlog<mode>_mask): Ditto.
1307 (avx512f_shufps512_mask): Ditto.
1308 (avx512f_fixupimm<mode>_mask): Ditto.
1309 (avx512f_shufpd512_mask): Ditto.
1310 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
1311 (avx512f_<code>v8div16qi2_mask/trunc): Ditto.
1312 (*avx512f_<code>v8div16qi2_store_mask): Ditto.
1313 (ashr<mode>3<mask_name>): Ditto.
1314 (avx512f_vinsert<shuffletype>32x4_mask): Ditto.
1315 (avx512f_vinsert<shuffletype>64x4_mask): Ditto.
1316 (avx512f_shuf_<shuffletype>64x2_mask): Ditto.
1317 (avx512f_shuf_<shuffletype>32x4_mask): Ditto.
1318 (avx512f_pshufdv3_mask): Ditto.
1319 (avx512f_perm<mode>_mask): Ditto.
1320 (avx512f_vpermi2var<mode>3_mask): Ditto.
1321 (avx512f_vpermt2var<mode>3_mask): Ditto.
1322 (avx512f_compress<mode>_mask): Ditto.
1323 (avx512f_compressstore<mode>_mask): Ditto.
1324 (avx512f_expand<mode>_mask): Ditto.
1325 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>): Extend
1327 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
1328 (<plusminus_insn><mode>3<mask_name>): Ditto.
1329 (*<plusminus_insn><mode>3<mask_name>): Ditto.
1330 (mul<mode>3<mask_name>): Ditto.
1331 (*mul<mode>3<mask_name>): Ditto.
1332 (<sse>_div<mode>3<mask_name>): Ditto.
1333 (<mask_codefor>rcp14<mode><mask_name>): Ditto.
1334 (<sse>_sqrt<mode>2<mask_name>): Ditto.
1335 (<mask_codefor>rsqrt14<mode><mask_name>): Ditto.
1336 (<code><mode>3<mask_name>/smaxmin): Ditto.
1337 (*<code><mode>3_finite<mask_name>/smaxmin): Ditto.
1338 (*<code><mode>3<mask_name>/smaxmin): Ditto.
1339 (float<sseintvecmodelower><mode>2<mask_name>): Ditto.
1340 (ufloatv16siv16sf2<mask_name>): Ditto.
1341 (<mask_codefor>avx512f_fix_notruncv16sfv16si<mask_name>): Ditto.
1342 (<mask_codefor>avx512f_ufix_notruncv16sfv16si<mask_name>): Ditto.
1343 (<fixsuffix>fix_truncv16sfv16si2<mask_name>): Ditto.
1344 (float<si2dfmodelower><mode>2<mask_name>): Ditto.
1345 (ufloatv8siv8df<mask_name>): Ditto.
1346 (<mask_codefor>avx512f_cvtpd2dq512<mask_name>): Ditto.
1347 (avx512f_ufix_notruncv8dfv8si<mask_name>): Ditto.
1348 (<fixsuffix>fix_truncv8dfv8si2<mask_name>): Ditto.
1349 (<mask_codefor>avx512f_cvtpd2ps512<mask_name>): Ditto.
1350 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix><mask_name>): Ditto.
1351 (<mask_codefor>avx512f_unpckhps512<mask_name>): Ditto.
1352 (<mask_codefor>avx512f_unpcklps512<mask_name>): Ditto.
1353 (<mask_codefor>avx512f_movshdup512<mask_name>): Ditto.
1354 (<mask_codefor>avx512f_movsldup512<mask_name>): Ditto.
1355 (<mask_codefor>avx512f_vextract<shuffletype>32x4_1<mask_name>): Ditto.
1356 (vec_extract_lo_<mode><mask_name>): Ditto.
1357 (vec_extract_hi_<mode><mask_name>): Ditto.
1358 (<mask_codefor>avx512f_unpckhpd512<mask_name>): Ditto.
1359 (avx512f_movddup512<mask_name>): Ditto.
1360 (avx512f_unpcklpd512<mask_name>): Ditto.
1361 (*avx512f_unpcklpd512<mask_name>): Ditto.
1362 (*avx512f_vmscalef<mode>): Ditto.
1363 (avx512f_scalef<mode><mask_name>): Ditto.
1364 (avx512f_getexp<mode><mask_name>): Ditto.
1365 (<mask_codefor>avx512f_align<mode><mask_name>): Ditto.
1366 (avx512f_rndscale<mode><mask_name>): Ditto.
1367 (avx512f_shufps512_1<mask_name>): Ditto.
1368 (avx512f_shufpd512_1<mask_name>): Ditto.
1369 (<plusminus_insn><mode>3<mask_name>): Ditto.
1370 (*<plusminus_insn><mode>3<mask_name>): Ditto.
1371 (vec_widen_umult_even_v16si<mask_name>): Ditto.
1372 (*vec_widen_umult_even_v16si<mask_name>): Ditto.
1373 (vec_widen_smult_even_v16si<mask_name>): Ditto.
1374 (*vec_widen_smult_even_v16si<mask_name>): Ditto.
1375 (mul<mode>3<mask_name>): Ditto.
1376 (*<sse4_1_avx2>_mul<mode>3<mask_name>): Ditto.
1377 (<shift_insn><mode>3<mask_name>): Ditto.
1378 (avx512f_<rotate>v<mode><mask_name>/rotate): Ditto.
1379 (avx512f_<rotate><mode><mask_name>): Ditto.
1380 (<code><mode>3<mask_name>/maxmin): Ditto.
1381 (*avx2_<code><mode>3<mask_name>/maxmin): Ditto.
1382 (<sse2_avx2>_andnot<mode>3<mask_name>): Ditto.
1383 (*andnot<mode>3<mask_name>): Ditto.
1384 (<mask_codefor><code><mode>3<mask_name>/any_logic): Ditto.
1385 (<mask_codefor>avx512f_interleave_highv16si<mask_name>): Ditto.
1386 (<mask_codefor>avx512f_interleave_lowv16si<mask_name>): Ditto.
1387 (<mask_codefor>avx512f_vinsert<shuffletype>32x4_1<mask_name>): Ditto.
1388 (vec_set_lo_<mode><mask_name>): Ditto.
1389 (vec_set_hi_<mode><mask_name>): Ditto.
1390 (avx512f_shuf_<shuffletype>64x2_1<mask_name>): Ditto.
1391 (avx512f_shuf_<shuffletype>32x4_1<mask_name>): Ditto.
1392 (avx512f_pshufd_1<mask_name>): Ditto.
1393 (<mask_codefor>abs<mode>2<mask_name>): Ditto.
1394 (<mask_codefor>avx512f_<code>v16qiv16si2<mask_name>): Ditto.
1395 (avx512f_<code>v16hiv16si2<mask_name>/any_extend): Ditto.
1396 (avx512f_<code>v8qiv8di2<mask_name>/any_extend): Ditto.
1397 (avx512f_<code>v8hiv8di2<mask_name>/any_extend): Ditto.
1398 (avx512f_<code>v8siv8di2<mask_name>/any_extend): Ditto.
1399 (avx512er_exp2<mode><mask_name>): Ditto.
1400 (<mask_codefor>avx512er_rcp28<mode><mask_name>): Ditto.
1401 (<mask_codefor>avx512er_rsqrt28<mode><mask_name>): Ditto.
1402 (<avx2_avx512f>_permvar<mode><mask_name>): Ditto.
1403 (<avx2_avx512f>_perm<mode>_1<mask_name>): Ditto.
1404 (<mask_codefor>avx512f_vec_dup<mode><mask_name>): Ditto.
1405 (<mask_codefor>avx512f_broadcast<mode><mask_name>/V16FI): Ditto.
1406 (<mask_codefor>avx512f_broadcast<mode><mask_name>/V8FI): Ditto.
1407 (<mask_codefor>avx512f_vec_dup_gpr<mode><mask_name>): Ditto.
1408 (<mask_codefor>avx512f_vec_dup_mem<mode><mask_name>): Ditto.
1409 (<sse2_avx_avx512f>_vpermil<mode><mask_name>/VF2): Ditto.
1410 (<sse2_avx_avx512f>_vpermil<mode><mask_name>/VF1): Ditto.
1411 (*<sse2_avx_avx512f>_vpermilp<mode><mask_name>): Ditto.
1412 (<sse2_avx_avx512f>_vpermilvar<mode>3<mask_name>): Ditto.
1413 (<avx2_avx512f>_ashrv<mode><mask_name>): Ditto.
1414 (<avx2_avx512f>_<shift_insn>v<mode><mask_name>): Ditto.
1415 (<mask_codefor>avx512f_vcvtph2ps512<mask_name>): Ditto.
1416 (<mask_codefor>avx512f_vcvtps2ph512<mask_name>): Ditto.
1417 (avx512f_getmant<mode><mask_name>): Ditto.
1418 (clz<mode>2<mask_name>): Ditto.
1419 (<mask_codefor>conflict<mode><mask_name>): Ditto.
1420 (*srcp14<mode>): Remove visibility.
1421 (*rsqrt14<mode>): Ditto.
1422 (*fma_fmsub_<mode>): Ditto.
1423 (*fma_fnmadd_<mode>): Ditto.
1424 (*avx512f_rndscale<mode>): Ditto.
1425 * config/i386/subst.md: New file.
1427 2013-11-13 Joseph Myers <joseph@codesourcery.com>
1429 * doc/extend.texi (Statement Exprs, Typeof): Discuss __auto_type.
1430 * ginclude/stdatomic.h (kill_dependency, atomic_store_explicit)
1431 (atomic_load_explicit, atomic_exchange_explicit)
1432 (atomic_compare_exchange_strong_explicit)
1433 (atomic_compare_exchange_weak_explicit): Use __auto_type to
1434 declare variable initialized with PTR argument.
1436 2013-11-12 Jeff Law <law@redhat.com>
1438 * tree-ssa-threadedge.c (thread_around_empty_blocks): New argument
1439 backedge_seen_p. Set, use and pass it to children appropriately.
1440 (thread_through_normal_block): Similarly.
1441 (thread_across_edge): Similarly.
1443 * gimple-ssa-isolate-paths.c (check_loadstore): Mark discovered
1444 memory references as volatile.
1445 (insert_trap_and_remove_trailing_statements): Fix comment.
1447 2013-11-12 Vladimir Makarov <vmakarov@redhat.com>
1450 * ira-costs.c (record_operand_costs): Check operands number for
1453 2013-11-12 Michael Meissner <meissner@linux.vnet.ibm.com>
1456 * config/rs6000/rs6000.md (movdi_internal32): Eliminate
1457 constraints that would allow DImode into the traditional Altivec
1458 registers, but cause undesirable code generation when loading 0 as
1460 (movdi_internal64): Likewise.
1461 (cmp<mode>_fpr): Do not use %x for CR register output.
1462 (extendsfdf2_fpr): Fix constraints when -mallow-upper-df and
1463 -mallow-upper-sf debug switches are used.
1465 2013-11-12 Andrew MacLeod <amacleod@redhat.com>
1467 * gimple-expr.h (create_tmp_var_name, create_tmp_var_raw,
1468 create_tmp_var, create_tmp_reg, mark_addressable, is_gimple_reg_rhs):
1469 Relocate prototypes from gimple.h.
1470 * gimplify.h: New File. Relocate some prototypes from gimple.h here.
1471 (gimple_predicate, enum fallback, enum gimplify_status): Relocate
1473 * gimple.h: Move some prototypes to gimplify.h.
1474 (gimple_predicate, enum fallback, enum gimplify_status): Move to
1476 (gimple_do_not_emit_location_p, gimple_set_do_not_emit_location):
1477 Relocate from gimpify.c.
1478 * gimple-expr.c (remove_suffix, tmp_var_id_num, create_tmp_var_name,
1479 create_tmp_var_raw, create_tmp_var, create_tmp_reg, mark_addressable,
1480 is_gimple_reg_rhs) Relocate from gimplify.c.
1481 * gimplify.c (mark_addressable): Move to gimple-expr.c.
1482 (gimple_seq_add_stmt_without_update): Move to gimple.c.
1483 (remove_suffix, tmp_var_id_num, create_tmp_var_name,
1484 create_tmp_var_raw, create_tmp_var, create_tmp_reg,
1485 is_gimple_reg_rhs): Move to gimple-expr.c.
1486 (should_carry_location_p): Move to gimple.c.
1487 (gimple_do_not_emit_location_p, gimple_set_do_not_emit_location): Move
1489 (annotate_one_with_location, annotate_all_with_location_after,
1490 annotate_all_with_location): Move to gimple.c.
1491 (compare_case_labels, sort_case_labels,
1492 preprocess_case_label_vec_for_gimple): Move to gimple.c.
1493 (rhs_predicate_for): Make static.
1494 (gimplify_assign): Relocate from gimple.c.
1495 * gimple.c (gimplify_assign): Move to gimplify.c.
1496 (gimple_seq_add_stmt_without_update, should_carry_location_p,
1497 annotate_one_with_location, annotate_all_with_location_after,
1498 annotate_all_with_location, compare_case_labels, sort_case_labels,
1499 preprocess_case_label_vec_for_gimple): Relocate from gimplify.c.
1500 * tree.h (unshare_expr, unshare_expr_without_location,
1501 mark_addressable): Move prototypes to gimplify.h.
1502 * Makefile.in (GTFILES): gimple-expr.c now has the GTY tag for
1504 * asan.c: Include gimplify.h rather than gimple.h.
1505 * cfgloopmanip.c: Likewise.
1506 * cgraphunit.c: Likewise.
1507 * cilk-common.c: Likewise.
1508 * dwarf2out.c: Dont include gimple.h.
1509 * fold-const.c: Include gimplify.h rather than gimple.h.
1510 * function.c: Likewise.
1511 * gimple-fold.c: Likewise.
1512 * gimple-ssa-strength-reduction.c: Likewise.
1513 * graphite-clast-to-gimple.c: Likewise.
1514 * graphite-sese-to-poly.c: Likewise.
1515 * ipa-prop.c: Likewise.
1516 * ipa-split.c: Likewise.
1518 * langhooks.c: Dont include gimple.h.
1519 * loop-init.c: Include gimplify.h rather than gimple.h.
1520 * omp-low.c: Likewise.
1522 * stor-layout.c: Likewise.
1523 * targhooks.c: Likewise.
1524 * trans-mem.c: Likewise.
1525 * tree-affine.c: Likewise.
1526 * tree-cfg.c: Likewise.
1527 * tree-cfgcleanup.c: Likewise.
1528 * tree-complex.c: Likewise.
1529 * tree-if-conv.c: Likewise.
1530 * tree-inline.c: Likewise.
1531 * tree-iterator.c: Likewise.
1532 * tree-loop-distribution.c: Likewise.
1533 * tree-nested.c: Likewise.
1534 * tree-parloops.c: Likewise.
1535 * tree-predcom.c: Likewise.
1536 * tree-profile.c: Likewise.
1537 * tree-scalar-evolution.c: Likewise.
1538 * tree-sra.c: Likewise.
1539 * tree-ssa-address.c: Likewise.
1540 * tree-ssa-ccp.c: Likewise.
1541 * tree-ssa-dce.c: Likewise.
1542 * tree-ssa-forwprop.c: Likewise.
1543 * tree-ssa-ifcombine.c: Likewise.
1544 * tree-ssa-loop-im.c: Likewise.
1545 * tree-ssa-loop-ivopts.c: Likewise.
1546 * tree-ssa-loop-manip.c: Likewise.
1547 * tree-ssa-loop-niter.c: Likewise.
1548 * tree-ssa-loop-prefetch.c: Likewise.
1549 * tree-ssa-loop-unswitch.c: Likewise.
1550 * tree-ssa-math-opts.c: Likewise.
1551 * tree-ssa-phiopt.c: Likewise.
1552 * tree-ssa-phiprop.c: Likewise.
1553 * tree-ssa-pre.c: Likewise.
1554 * tree-ssa-propagate.c: Likewise.
1555 * tree-ssa-reassoc.c: Likewise.
1556 * tree-ssa-sccvn.c: Likewise.
1557 * tree-ssa-strlen.c: Likewise.
1558 * tree-ssa.c: Likewise.
1559 * tree-switch-conversio: Likewise.n.c
1560 * tree-tailcall.c: Likewise.
1561 * tree-vect-data-refs.c: Likewise.
1562 * tree-vect-generic.c: Likewise.
1563 * tree-vect-loop-manip.c: Likewise.
1564 * tree-vect-loop.c: Likewise.
1565 * tree-vect-patterns.c: Likewise.
1566 * tree-vect-stmts.c: Likewise.
1568 * value-prof.c: Likewise.
1569 * config/aarch64/aarch64.c: Include gimplify.h instead of gimple.h.
1570 * config/alpha/alpha.c: Likewise.
1571 * config/darwin.c: Likewise.
1572 * config/i386/i386.c: Likewise.
1573 * config/ia64/ia64.c: Likewise.
1574 * config/mep/mep.c: Likewise.
1575 * config/mips/mips.c: Likewise.
1576 * config/rs6000/rs6000.c: Likewise.
1577 * config/s390/s390.c: Likewise.
1578 * config/sh/sh.c: Likewise.
1579 * config/sparc/sparc.c: Likewise.
1580 * config/spu/spu.c: Likewise.
1581 * config/stormy16/stormy16.c: Likewise.
1582 * config/tilegx/tilegx.c: Likewise.
1583 * config/tilepro/tilepro.c: Likewise.
1584 * config/xtensa/xtensa.c: Likewise.
1586 2013-11-12 Adam Butcher <adam@jessamine.co.uk>
1588 * tree.c (grow_tree_vec_stat): New function ...
1589 * tree.h (grow_tree_vec_stat) (grow_tree_vec): ... and its declaration
1590 and macro front-end.
1592 2013-11-12 Marek Polacek <polacek@redhat.com>
1594 * final.c (update_alignments): Initialize label to NULL_RTX.
1596 2013-11-12 Jeff Law <law@redhat.com>
1598 * gimple-ssa-isolate-paths.c (check_loadstore): New function.
1599 (insert_trap_and_remove_trailing_statements): New argument OP which
1600 is the NULL pointer. Emit the trap after the load/store through
1601 the NULL pointer. Simplify the RHS of a store through a NULL pointer
1602 when trivial to do so.
1603 (isolate_path): Corresponding changes.
1604 (gimple_ssa_isolate_erroneous_path): Likewise.
1606 2013-11-12 Teresa Johnson <tejohnson@google.com>
1607 Jan Hubicka <jh@suse.cz>
1609 * predict.c (drop_profile): New function.
1610 (handle_missing_profiles): Ditto.
1611 (counts_to_freqs): Don't overwrite estimated frequencies
1612 when function has no profile counts.
1613 * predict.h (handle_missing_profiles): Declare.
1614 * tree-inline.c (freqs_to_counts): New function.
1615 (copy_cfg_body): Invoke freqs_to_counts as needed.
1616 * tree-profile.c (tree_profiling): Invoke handle_missing_profiles.
1618 2013-11-12 H.J. Lu <hongjiu.lu@intel.com>
1621 * config/i386/x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL):
1623 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Set for m_HASWELL.
1625 2013-11-12 H.J. Lu <hongjiu.lu@intel.com>
1628 * config/i386/i386.c (ix86_option_override_internal): Check
1629 X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL and
1630 X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL for
1631 MASK_AVX256_SPLIT_UNALIGNED_LOAD and
1632 MASK_AVX256_SPLIT_UNALIGNED_STORE.
1634 * config/i386/x86-tune.def (X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL):
1635 Clear m_COREI7_AVX and update comments.
1636 (X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL): Likewise.
1638 2013-11-12 Martin Jambor <mjambor@suse.cz>
1640 PR rtl-optimization/10474
1641 * ira.c (interesting_dest_for_shprep): New function.
1642 (split_live_ranges_for_shrink_wrap): Likewise.
1643 (find_moveable_pseudos): Move calculation of dominance info,
1644 df_analysios and the final anlyses to...
1645 (ira): ...here, call split_live_ranges_for_shrink_wrap.
1647 2013-11-12 Bin Cheng <bin.cheng@arm.com>
1649 * tree-ssa-loop-ivopts.c (force_expr_to_var_cost): Refactor the code.
1650 Handle type conversion.
1652 2013-11-11 Martin Liska <marxin.liska@gmail.com>
1653 Jan Hubicka <jh@suse.cz>
1655 * cgraph.c (dump_cgraph_node): Profile dump added.
1656 * cgraph.h (struct cgraph_node): New time profile variable added.
1657 * cgraphclones.c (cgraph_clone_node): Time profile is cloned.
1658 * gcov-io.h (gcov_type): New profiler type introduced.
1659 * ipa-profile.c (lto_output_node): Streaming for time profile added.
1660 (input_node): Time profiler is read from LTO stream.
1661 * predict.c (maybe_hot_count_p): Hot prediction changed.
1662 * profile.c (instrument_values): New case for time profiler added.
1663 (compute_value_histograms): Read of time profile.
1664 * tree-pretty-print.c (dump_function_header): Time profiler is dumped.
1665 * tree-profile.c (init_ic_make_global_vars): Time profiler
1667 (gimple_init_edge_profiler): TP function instrumentation.
1668 (gimple_gen_time_profiler): New.
1669 * value-prof.c (gimple_add_histogram_value): Support for time profiler
1671 (dump_histogram_value): TP type added to dumps.
1672 (visit_hist): More sensitive check that takes TP into account.
1673 (gimple_find_values_to_profile): TP instrumentation.
1674 * value-prof.h (hist_type): New histogram type added.
1675 (struct histogram_value_t): Pointer to struct function added.
1676 * libgcc/Makefile.in: New GCOV merge function for TP added.
1677 * libgcov.c: function_counter variable introduced.
1678 (_gcov_merge_time_profile): New.
1679 (_gcov_time_profiler): New.
1681 2013-11-11 Marc Glisse <marc.glisse@inria.fr>
1682 Jeff Law <law@redhat.com>
1684 * tree-ssa-alias.c (stmt_kills_ref_p_1): Use
1685 ao_ref_init_from_ptr_and_size for builtins.
1687 2013-11-11 Uros Bizjak <ubizjak@gmail.com>
1688 H.J. Lu <hongjiu.lu@intel.com>
1691 * config/i386/x86-tune.def
1692 (X86_TUNE_MISALIGNED_MOVE_STRING_PRO_EPILOGUES): Rename from
1693 TARGET_MISALIGNED_MOVE_STRING_PROLOGUES.
1694 * config/i386/i386.h
1695 (TARGET_MISALIGNED_MOVE_STRING_PRO_EPILOGUES): Rename from
1696 TARGET_MISALIGNED_MOVE_STRING_PROLOGUES_EPILOGUES. Update for renamed
1697 X86_TUNE_MISALIGNED_MOVE_STRING_PRO_EPILOGUES.
1698 * config/i386/i386.c (ix86_expand_set_or_movmem): Use
1699 TARGET_MISALIGNED_MOVE_STRING_PRO_EPILOGUES to calculate
1700 misaligned_prologue_used. Check that
1701 desired_aling <= epilogue_size_needed.
1703 2013-11-11 Cong Hou <congh@google.com>
1705 PR tree-optimization/59050
1706 * tree-vect-data-refs.c (comp_dr_addr_with_seg_len_pair): Bug fix.
1708 2013-11-11 Joern Rennecke <joern.rennecke@embecosm.com>
1711 * expmed.c (emit_store_flag): Fail for const-const comparison.
1713 2013-11-11 Tristan Gingold <gingold@adacore.com>
1714 Eric Botcazou <ebotcazou@adacore.com>
1716 * tree.h (CONSTRUCTOR_NO_CLEARING): Define.
1717 * tree-core.h (CONSTRUCTOR_NO_CLEARING): Document it.
1718 * tree.def (CONSTRUCTOR): Likewise.
1719 * doc/generic.texi (CONSTRUCTOR): Likewise. Update description.
1720 * gimplify.c (gimplify_init_constructor): Do not clear the object when
1721 the constructor is incomplete and CONSTRUCTOR_NO_CLEARING is set.
1723 2013-11-11 Basile Starynkevitch <basile@starynkevitch.net>
1725 * toplev.c (toplev_main): Move PLUGIN_FINISH invocation before
1728 2013-11-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1730 * config/arm/arm.c (arm_new_rtx_costs): Return after handling
1733 2013-11-11 Joern Rennecke <joern.rennecke@embecosm.com>
1735 * config/arc/arc.h (LOGICAL_OP_NON_SHORT_CIRCUIT): Define.
1737 2013-11-08 Jeff Law <law@redhat.com>
1739 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
1740 threading paths first, then perform PHI node checks if applicable.
1742 2013-11-10 Karlson2k <k2k@narod.ru>
1743 Kai Tietz <ktietz@redhat.com>
1746 * configure.ac: Adding for exported symbols check
1747 and for rdynamic-check executable-extension.
1748 * configure: Regenerated.
1750 2013-11-10 Uros Bizjak <ubizjak@gmail.com>
1752 * mode-switching.c (optimize_mode_switching): Mark block as
1753 nontransparent, if last_mode at block exit is different from no_mode.
1755 2013-11-09 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1757 * function.c (NAME__MAIN): Move to...
1758 * cfgexpand.c (NAME__MAIN): ...here.
1760 2013-11-09 Richard Sandiford <rdsandiford@googlemail.com>
1762 * target.def (can_use_doloop_p): New hook.
1763 * doc/tm.texi.in (TARGET_CAN_USE_DOLOOP_P): Add.
1764 * doc/tm.texi: Regenerate.
1765 * doc/md.texi (doloop_begin, doloop_end): Update documentation.
1766 * hooks.h (hook_bool_dint_dint_uint_true): Declare.
1767 * hooks.c (hook_bool_dint_dint_uint_true): New function.
1768 * targhooks.h (can_use_doloop_if_innermost): Declare.
1769 * targhooks.c (can_use_doloop_if_innermost): New function.
1770 * target.h: Include double-int.h.
1771 * loop-doloop.c (doloop_optimize): Call targetm.can_use_doloop_p.
1772 Remove iteration count, maximum iteration count, loop depth and
1773 enter-at-top inputs from doloop_begin and doloop_end.
1774 * config/arc/arc.md (doloop_begin, doloop_end): Update for new
1776 * config/arc/arc.c (arc_can_use_doloop_p): New function.
1777 (TARGET_CAN_USE_DOLOOP_P): Define.
1778 * config/arm/thumb2.md (doloop_end): Update for new interface.
1779 * config/arm/arm.c (TARGET_CAN_USE_DOLOOP_P): Define.
1780 * config/bfin/bfin.md (doloop_end): Update for new interface.
1781 * config/bfin/bfin.c (bfin_can_use_doloop_p): New function.
1782 (TARGET_CAN_USE_DOLOOP_P): Define.
1783 * config/c6x/c6x.md (doloop_end): Update for new interface.
1784 * config/ia64/ia64.md (doloop_end): Update for new interface.
1785 * config/ia64/ia64.c (TARGET_CAN_USE_DOLOOP_P): Define.
1786 * config/mep/mep.md (doloop_begin, doloop_end): Update for new
1788 * config/mep/mep.c (mep_emit_doloop): Likewise.
1789 (TARGET_CAN_USE_DOLOOP_P): Define.
1790 * config/rs6000/rs6000.md (doloop_end): Update for new interface.
1791 * config/rs6000/rs6000.c (TARGET_CAN_USE_DOLOOP_P): Define.
1792 * config/s390/s390.md (doloop_end): Update for new interface.
1793 * config/sh/sh.md (doloop_end): Likewise.
1794 * config/spu/spu.md (doloop_end): Likewise.
1795 * config/spu/spu.c (TARGET_CAN_USE_DOLOOP_P): Define.
1796 * config/tilegx/tilegx.md (doloop_end): Update for new interface.
1797 * config/tilegx/tilegx.c (TARGET_CAN_USE_DOLOOP_P): Define.
1798 * config/tilepro/tilepro.md (doloop_end): Update for new interface.
1799 * config/tilepro/tilepro.c (TARGET_CAN_USE_DOLOOP_P): Define.
1800 * config/v850/v850.md (doloop_begin, doloop_end): Update for new
1802 * config/v850/v850.c (TARGET_CAN_USE_DOLOOP_P): Define.
1804 2013-11-08 H.J. Lu <hongjiu.lu@intel.com>
1807 * doc/extend.texi: Move Cilk Plus Builtins node before Other
1810 2013-11-08 Andrew MacLeod <amacleod@redhat.com>
1811 Joseph Myers <joseph@codesourcery.com>
1813 * ginclude/stdatomic.h: New file.
1814 * Makefile.in (USER_H): Add stdatomic.h.
1816 2013-11-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1818 * config/arm/arm.c (arm_new_rtx_costs): Break after handling
1821 2013-11-08 Jeff Law <law@redhat.com>
1823 * tree-ssa-threadupdate.h (delete_thread_path): Declare.
1824 * tree-ssa-threadupdate.c (delete_thread_path): New function.
1825 (ssa_redirect_edges, thread_block_1): Use it.
1826 (thread_through_loop_header, mark_threaded_blocks): Likewise.
1827 (thread_through_all_blocks, register_jump_thread): Likewise.
1828 * tree-ssa-threadedge.c (thread_across_edge): Likewise.
1830 2013-11-08 James Greenhalgh <james.greenhalgh@arm.com>
1832 * config/arm/aarch-common.c
1833 (search_term): New typedef.
1834 (shift_rtx_costs): New array.
1835 (arm_rtx_shift_left_p): New.
1836 (arm_find_sub_rtx_with_search_term): Likewise.
1837 (arm_find_sub_rtx_with_code): Likewise.
1838 (arm_early_load_addr_dep): Add sanity checking.
1839 (arm_no_early_alu_shift_dep): Likewise.
1840 (arm_no_early_alu_shift_value_dep): Likewise.
1841 (arm_no_early_mul_dep): Likewise.
1842 (arm_no_early_store_addr_dep): Likewise.
1844 2013-11-08 Richard Biener <rguenther@suse.de>
1846 PR tree-optimization/59047
1847 * tree-predcom.c (ref_at_iteration): Handle bitfield accesses properly.
1849 2013-11-08 Ilya Enkovich <ilya.enkovich@intel.com>
1851 * common.opt (fcheck-pointer-bounds): Move to ...
1852 * c-family/c.opt: ... here.
1853 * langhooks-def.h (LANG_HOOKS_CHKP_SUPPORTED): Remove.
1854 (LANG_HOOKS_INITIALIZER): Remove LANG_HOOKS_CHKP_SUPPORTED.
1855 * langhooks.h (lang_hooks): Remove chkp_supported field.
1856 * toplev.c (process_options): Remove chkp_supported check.
1858 2013-11-08 Richard Biener <rguenther@suse.de>
1860 PR tree-optimization/59038
1861 PR tree-optimization/58955
1862 * tree-loop-distribution.c (pg_add_dependence_edges): Revert
1863 previous change. Handle known dependences correctly.
1865 2013-11-08 Tom de Vries <tom@codesourcery.com>
1867 * config/rs6000/t-xilinx: Remove duplicate contents.
1869 2013-11-07 Andrew MacLeod <amacleod@redhat.com>
1870 Joseph Myers <joseph@codesourcery.com>
1872 * tree-core.h (enum cv_qualifier): Add TYPE_QUAL_ATOMIC.
1873 (enum tree_index): Add TI_ATOMICQI_TYPE, TI_ATOMICHI_TYPE,
1874 TI_ATOMICSI_TYPE, TI_ATOMICDI_TYPE and TI_ATOMICTI_TYPE.
1875 (struct tree_base): Add atomic_flag field.
1876 * tree.h (TYPE_ATOMIC): New accessor macro.
1877 (TYPE_QUALS, TYPE_QUALS_NO_ADDR_SPACE): Add TYPE_QUAL_ATOMIC.
1878 (TYPE_QUALS_NO_ADDR_SPACE_NO_ATOMIC): New macro.
1879 (atomicQI_type_node, atomicHI_type_node, atomicSI_type_node)
1880 (atomicDI_type_node, atomicTI_type_node): New macros for type nodes.
1881 * tree.c (set_type_quals): Set TYPE_ATOMIC.
1882 (find_atomic_core_type): New function.
1883 (build_qualified_type): Adjust alignment for qualified types.
1884 (build_atomic_base): New function
1885 (build_common_tree_nodes): Build atomicQI_type_node,
1886 atomicHI_type_node, atomicSI_type_node, atomicDI_type_node and
1888 * print-tree.c (print_node): Print atomic qualifier.
1889 * tree-pretty-print.c (dump_generic_node): Print atomic type attribute.
1890 * target.def (atomic_assign_expand_fenv): New hook.
1891 * doc/tm.texi.in (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): New @hook.
1892 * doc/tm.texi: Regenerate.
1893 * targhooks.c (default_atomic_assign_expand_fenv): New function.
1894 * targhooks.h (default_atomic_assign_expand_fenv): Declare.
1895 * sync-builtins.def (__atomic_feraiseexcept): New built-in function.
1896 * config/i386/i386-builtin-types.def (VOID_FTYPE_PUSHORT): New
1898 * config/i386/i386.c (enum ix86_builtins): Add
1899 IX86_BUILTIN_FNSTENV, IX86_BUILTIN_FLDENV, IX86_BUILTIN_FNSTSW and
1900 IX86_BUILTIN_FNCLEX.
1901 (bdesc_special_args): Add __builtin_ia32_fnstenv,
1902 __builtin_ia32_fldenv, __builtin_ia32_fnstsw and __builtin_ia32_fnclex.
1903 (ix86_expand_builtin): Handle the new built-in functions.
1904 (ix86_atomic_assign_expand_fenv): New function.
1905 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): New macro.
1906 * config/i386/i386.md (UNSPECV_FNSTENV, UNSPECV_FLDENV)
1907 (UNSPECV_FNSTSW, UNSPECV_FNCLEX): New unspecs.
1908 (fnstenv, fldenv, fnstsw, fnclex): New insns.
1910 2013-11-07 Steve Ellcey <sellcey@mips.com>
1912 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add fp64 directory.
1913 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add -mfp64 flag.
1914 (MULTILIB_DIRNAMES): Add fp64 directory.
1915 (MULTILIB_EXCEPTIONS): Add new exclusions.
1917 2013-11-07 Aldy Hernandez <aldyh@redhat.com>
1919 * gimplify.c (gimple_regimplify_operands): Do not set
1921 * graphite-sese-to-poly.c (remove_simple_copy_phi): Same.
1922 (rewrite_close_phi_out_of_ssa): Same.
1923 (rewrite_phi_out_of_ssa): Same.
1924 (rewrite_degenerate_phi): Same.
1925 (handle_scalar_deps_crossing_scop_limits): Same.
1926 * tree-if-conv.c (predicate_scalar_phi): Same.
1927 * tree-parloops.c (create_loads_for_reductions): Same.
1928 (create_final_loads_for_reduction): Same.
1929 (create_loads_and_stores_for_name): Same.
1930 (transform_to_exit_first_loop): Same.
1931 (create_parallel_loop): Same.
1932 * tree-ssa-loop-im.c
1933 (move_computations_dom_walker::before_dom_children): Same.
1934 * tree-ssa-loop-manip.c (rewrite_phi_with_iv): Same.
1935 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children): Same.
1936 * tree-ssa-propagate.c (substitute_and_fold): Same.
1937 * tree-vect-loop.c (vect_finalize_reduction): Same.
1938 * tree-vect-stmts.c (vectorizable_call): Same.
1940 2013-11-07 Mike Stump <mikestump@comcast.net>
1942 * config/pdp11/pdp11.c: Include dbxout.h.
1943 * config/picochip/picochip.c: Likewise.
1945 2013-11-07 Cong Hou <congh@google.com>
1947 PR tree-optimization/56764
1948 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
1949 Combine alias checks if it is possible to amortize the runtime
1950 overhead. Return the number of alias checks after merging.
1951 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
1952 Use the function vect_create_cond_for_alias_checks () to check
1953 the number of alias checks.
1955 2013-11-07 Jeff Law <law@redhat.com>
1957 * varpool.c (ctor_for_folding): Fix typo in comment.
1959 2013-11-07 Joern Rennecke <joern.rennecke@embecosm.com>
1961 * config/arc/arc.c (arc_ifcvt): Use commutativity, e.g.:
1962 reg_a := reg_b + reg_a ==> reg_a := reg_a + reg_b
1964 2013-11-07 Jeff Law <law@redhat.com>
1966 * doc/invoke.texi (-fisolate-erroneous-paths): Document.
1968 * gimple-ssa-isolate-paths.c (gate_isolate_erroneous_paths):
1969 No longer check if we have __builtin_trap, assume it's available.
1971 2013-11-07 Diego Novillo <dnovillo@google.com>
1973 * attribs.c (lookup_scoped_attribute_spec): Make static.
1974 (get_attribute_namespace): Likewise.
1975 * builtins.c (more_const_call_expr_args_p): Move from tree.h.
1976 (validate_arglist): Move earlier in the file. Make static.
1977 (expand_stack_restore): Move from stmt.c
1978 (expand_stack_save): Move from stmt.c
1979 (rewrite_call_expr_array): Move earlier in the file.
1980 (rewrite_call_expr_valist): Likewise.
1981 * cfgexpand.c: Include hard-reg-set.h before tree.h
1984 (expand_asm_loc): Move from stmt.c.
1985 (n_occurrences): Move from stmt.c.
1986 (check_operand_nalternatives): Move from stmt.c.
1987 (tree_conflicts_with_clobbers_p): Move from stmt.c.
1988 (expand_asm_operands): Move from stmt.c
1989 (expand_asm_stmt): Move from stmt.c
1990 (expand_computed_goto): Move from stmt.c
1991 (expand_goto): Move from stmt.c
1992 (expand_null_return_1): Move from stmt.c
1993 (expand_null_return): Move from stmt.c
1994 (expand_value_return): Move from stmt.c
1995 (expand_return): Move from stmt.c
1996 (expand_main_function): Move from function.c
1997 (stack_protect_prologue): Move from function.c
1998 * cgraphclones.c (build_function_type_skip_args): Move from tree.c.
1999 (build_function_decl_skip_args): Move from tree.c.
2000 * explow.c (tree_expr_size): Move from tree.c.
2001 * expr.c (addr_expr_of_non_mem_decl_p): Remove.
2002 (fields_length): Move from tree.c.
2003 * fold-const.c (size_low_cst): Move from tree.c.
2004 (tree_expr_nonzero_warnv_p): Make static. Move earlier in the file.
2005 (tree_expr_nonzero_p): Make static. Move earlier in the file.
2006 (fold_build3_initializer_loc): Remove.
2007 (tree_invalid_nonnegative_warnv_p): Make static.
2008 * function.c (expand_main_function): Move to cfgexpand.c.
2009 (stack_protect_prologue): Move to cfgexpand.c.
2010 (set_insn_locations): Move earlier in the file.
2011 * gimple-fold.c: Include langhooks.h.
2012 (truth_type_for): Move from tree.c.
2013 * print-tree.c (print_vec_tree): Remove.
2014 * stmt.c (expand_computed_goto): Move to cfgexpand.c.
2015 (expand_goto): Move to cfgexpand.c.
2016 (n_occurrences): Move to cfgexpand.c.
2017 (expand_asm_loc): Move to cfgexpand.c
2018 (tree_conflicts_with_clobbers_p): Move to cfgexpand.c.
2019 (expand_asm_operands): Move to cfgexpand.c.
2020 (expand_asm_stmt): Move to cfgexpand.c.
2021 (check_operand_nalternatives): Move to cfgexpand.c
2022 (expand_null_return): Move to cfgexpand.c.
2023 (expand_value_return): Move to cfgexpand.c.
2024 (expand_null_return_1): Move to cfgexpand.c.
2025 (expand_return): Move to cfgexpand.c.
2026 (expand_stack_save): Move to builtins.c.
2027 (expand_stack_restore): Move to builtins.c
2028 * symtab.c: Include output.h.
2029 (decl_assembler_name_hash): Move from tree.c.
2030 (decl_assembler_name_equal): Move from tree.c.
2031 * trans-mem.c (is_tm_safe_or_pure): Move from tree.h.
2032 * tree-eh.c (in_array_bounds_p): Move from tree.c.
2033 (range_in_array_bounds_p): Move from tree.c.
2034 * tree-object-size.c (fini_object_sizes): Make static.
2035 * tree-ssa-dom.c (iterative_hash_exprs_commutative): Move from tree.h.
2036 * tree-vrp.c (ssa_name_nonnegative_p): Remove.
2037 * tree.c (decl_assembler_name_equal): Move to symtab.c.
2038 (tree_expr_size): Move to explow.c.
2039 (decl_assembler_name_hash): Move to symtab.c.
2040 (real_twop): Remove.
2041 (tree_expr_size): Move to explow.c.
2042 (stabilize_reference_1): Move earlier in the file. Make static.
2043 (omp_remove_redundant_declare_simd_attrs): Remove.
2044 (simple_cst_list_equal): Move earlier in the file. Make static.
2045 (size_low_cst): Move to fold-const.c.
2046 (build_type_no_quals): Remove.
2047 (build_function_type_skip_args): Move to cgraphclones.c.
2048 (build_function_decl_skip_args): Move to cgraphclones.c.
2049 (in_array_bounds_p): Move to tree-eh.c.
2050 (range_in_array_bounds_p): Move to tree-eh.c.
2051 (truth_type_for): Move to gimple-fold.c.
2052 (list_equal_p): Remove.
2053 * tree.h (decl_assembler_name_equal): Remove.
2054 (decl_assembler_name_hash): Remove.
2055 (truth_type_for): Remove.
2056 (build_type_no_quals): Remove.
2057 (build_function_decl_skip_args): Remove.
2058 (in_array_bounds_p): Remove.
2059 (range_in_array_bounds_p): Remove.
2060 (size_low_cst): Remove.
2061 (omp_remove_redundant_declare_simd_attrs): Remove.
2062 (tree_expr_size): Remove.
2063 (fields_length): Remove.
2064 (stabilize_reference_1): Remove.
2065 (expand_goto): Remove.
2066 (expand_stack_save): Remove.
2067 (expand_stack_restore): Remove.
2068 (expand_return): Remove.
2069 (fold_build3_initializer_loc): Remove.
2070 (tree_expr_nonzero_p): Remove.
2071 (tree_invalid_nonnegative_warnv_p): Remove.
2072 (tree_expr_nonzero_warnv_p): Remove.
2073 (fold_builtin_snprintf_chk): Remove.
2074 (validate_arglist): Remove.
2075 (iterative_hash_exprs_commutative): Move to tree-ssa-dom.c.
2076 (simple_cst_list_equal): Remove.
2077 (real_twop): Remove.
2078 (expand_main_function): Remove.
2079 (stack_protect_prologue): Remove.
2080 (print_vec_tree): Remove.
2081 (lookup_scoped_attribute_spec): Remove.
2082 (get_attribute_namespace): Remove.
2083 (expand_computed_goto): Remove.
2084 (expand_asm_stmt): Remove.
2085 (list_equal_p): Remove.
2086 (ssa_name_nonnegative_p): Remove.
2087 (fini_object_sizes): Remove.
2088 (addr_expr_of_non_mem_decl_p): Remove.
2089 (is_tm_safe_or_pure): Move to trans-mem.c.
2090 (more_const_call_expr_args_p): Remove.
2091 (save_vtable_map_decl): Remove.
2093 2013-11-07 Thomas Schwinge <thomas@codesourcery.com>
2095 * doc/sourcebuild.texi (Top Level) <lto-plugin>: GNU ld can use
2096 linker plugins, too.
2098 * config/arc/arc.h (LINK_COMMAND_SPEC): For -ftree-parallelize-loops=*,
2099 link to libgomp and its dependencies.
2100 * config/ia64/hpux.h (LIB_SPEC): Likewise.
2101 * config/pa/pa-hpux11.h (LIB_SPEC): Likewise.
2102 * config/pa/pa64-hpux.h (LIB_SPEC): Likewise.
2103 * gcc.c (GOMP_SELF_SPECS): Update comment about libgomp's dependencies.
2105 2013-11-07 Jakub Jelinek <jakub@redhat.com>
2107 * tree-ssa-loop-niter.c: Include tree-ssanames.h.
2108 (determine_value_range): Add loop argument. Use get_range_info to
2110 (bound_difference): Adjust caller.
2112 2013-11-07 Richard Biener <rguenther@suse.de>
2113 Jakub Jelinek <jakub@redhat.com>
2115 * tree-vrp.c (find_assert_locations): Pre-seed live bitmaps for loop
2116 latches from header PHI arguments from the latch edge.
2118 2013-11-07 Paolo Carlini <paolo.carlini@oracle.com>
2121 * varasm.c (output_constant): Handle NULLPTR_TYPE.
2123 2013-11-07 H.J. Lu <hongjiu.lu@intel.com>
2125 * config/i386/i386.c (ix86_expand_set_or_movmem): Don't set
2126 misaligned_prologue_used when it has been set.
2128 2013-11-07 Yury Gribov <y.gribov@samsung.com>
2129 Jakub Jelinek <jakub@redhat.com>
2132 * asan.c (get_mem_refs_of_builtin_call): Allow
2133 integer literals as addresses in instrumented builtins.
2135 2013-11-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2137 * config/aarch64/aarch64.c (aarch64_legitimize_reload_address):
2138 Explain why plus_constant is not used.
2140 2013-11-07 Richard Biener <rguenther@suse.de>
2142 * tree-ssa-ccp.c (canonicalize_float_value): Rename to ...
2143 (canonicalize_value): ... this. Also handle stripping of
2145 (get_value, set_lattice_value, get_value_for_expr): Adjust.
2146 * gimple-fold.c (canonicalize_constructor_val): Strip TREE_OVERFLOW.
2147 * tree-ssa-threadedge.c (set_ssa_name_value): Likewise.
2149 2013-11-07 Richard Biener <rguenther@suse.de>
2151 * tree-dfa.c (get_ref_base_and_extent): Fix casting.
2153 2013-11-07 H.J. Lu <hongjiu.lu@intel.com>
2156 * config/i386/i386.md (push peepholer/splitter): Use Pmode
2157 with stack_pointer_rtx.
2159 2013-11-07 Bin Cheng <bin.cheng@arm.com>
2161 * tree-ssa-loop-ivopts.c (get_shiftadd_cost): Check equality
2162 using operand_equal_p.
2164 2013-11-07 Bin Cheng <bin.cheng@arm.com>
2166 * tree-ssa-loop-ivopts.c (alloc_iv): Lower address expressions.
2167 * tree-affine.c (get_inner_reference_aff): Return base.
2168 * tree-affine.h (get_inner_reference_aff): Change prototype.
2170 2013-11-06 Tobias Burnus <burnus@net-b.de>
2172 * doc/invoke.texi (Wdate-time): Fix typo.
2174 2013-11-06 Oleg Endo <olegendo@gcc.gnu.org>
2176 * config/sh/sh.md (addsf3, divsf3, divsf3_i, rsqrtsf2, cmpgtdf_t,
2177 cmpeqdf_t, *ieee_ccmpeqdf_t, negdf2, sqrtdf2, absdf2): Use
2178 fp_arith_reg_operand instead of arith_reg_operand.
2180 2013-11-06 Oleg Endo <olegendo@gcc.gnu.org>
2182 * config/sh/sh.md (adddi3): Remove empty constraints.
2183 Remove can_create_pseudo_p and arith_reg_operand check.
2184 (adddi3_compact, subdi3_compact, *negdi2): Remove constraints.
2185 Split before reload.
2187 2013-11-06 Jeff Law <law@redhat.com>
2188 Tom Tromey <tromey@redhat.com>
2190 * gdbinit.in: Disable strict type checking.
2192 2013-11-06 Vladimir Makarov <vmakarov@redhat.com>
2194 * tree-pass.h (make_pass_live_range_shrinkage): New external.
2195 * timevar.def (TV_LIVE_RANGE_SHRINKAGE): New.
2196 * sched-rgn.c (gate_handle_live_range_shrinkage): New.
2197 (rest_of_handle_live_range_shrinkage): Ditto
2198 (class pass_live_range_shrinkage): Ditto.
2199 (pass_data_live_range_shrinkage): Ditto.
2200 (make_pass_live_range_shrinkage): Ditto.
2201 * sched-int.h (initialize_live_range_shrinkage): New prototype.
2202 (finish_live_range_shrinkage): Ditto.
2203 * sched-deps.c (create_insn_reg_set): Make void return value.
2204 * passes.def: Add pass_live_range_shrinkage.
2205 * ira.c (update_equiv_regs): Don't move if flag_live_range_shrinkage.
2206 * haifa-sched.c (live_range_shrinkage_p): New.
2207 (initialize_live_range_shrinkage, finish_live_range_shrinkage):
2209 (rank_for_schedule): Add code for pressure relief through live
2211 (schedule_insn): Print more debug info.
2212 (sched_init): Setup SCHED_PRESSURE_WEIGHTED for pressure relief
2213 through live range shrinkage.
2214 * doc/invoke.texi (-flive-range-shrinkage): New.
2215 * common.opt (flive-range-shrinkage): New.
2217 2013-11-06 Uros Bizjak <ubizjak@gmail.com>
2220 * config/i386/i386.c (ix86_avx_u128_mode_needed): Require
2221 AVX_U128_DIRTY mode for call_insn RTXes that use AVX256 registers.
2222 (ix86_avx_u128_mode_needed): Return AVX_U128_DIRTY mode for call_insn
2223 RTXes that return in AVX256 register.
2225 2013-11-06 Richard Biener <rguenther@suse.de>
2227 PR tree-optimization/58653
2228 * tree-predcom.c (ref_at_iteration): Rewrite to generate a MEM_REF.
2229 (prepare_initializers_chain): Adjust.
2231 2013-11-06 Andrew MacLeod <amacleod@redhat.com>
2233 * gimple.h (block_in_transaction): Move to basic-block.h and rename.
2234 (gimple_in_transaction): Use bb_in_transaction.
2235 * basic-block.h (bb_in_transaction): Relocate here and rename.
2236 * tree-ssa-loop-im.c (execute_sm): Use bb_in_transaction.
2238 2013-11-06 Richard Biener <rguenther@suse.de>
2240 * tree.c (drop_tree_overflow): New function.
2241 * tree.h (drop_tree_overflow): Declare.
2242 * gimplify.c (gimplify_expr): Drop TREE_OVERFLOW.
2243 * tree-vrp.c (range_int_cst_singleton_p): Use
2244 is_overflow_infinity instead of testing TREE_OVERFLOW.
2245 (extract_range_from_assert): Likewise.
2246 (zero_nonzero_bits_from_vr): Likewise.
2247 (extract_range_basic): Likewise.
2248 (register_new_assert_for): Use drop_tree_overflow.
2249 (vrp_visit_phi_node): Likewise.
2251 2013-11-06 Eric Botcazou <ebotcazou@adacore.com>
2253 * config/i386/i386.c (ix86_expand_prologue): Optimize stack
2254 checking for leaf functions without dynamic stack allocation.
2255 * config/ia64/ia64.c (ia64_emit_probe_stack_range): Adjust.
2256 (ia64_expand_prologue): Likewise.
2257 * config/mips/mips.c (mips_expand_prologue): Likewise.
2258 * config/rs6000/rs6000.c (rs6000_emit_prologue): Likewise.
2259 * config/sparc/sparc.c (sparc_expand_prologue): Likewise.
2260 (sparc_flat_expand_prologue): Likewise.
2262 2013-11-06 James Greenhalgh <james.greenhalgh@arm.com>
2264 * config/aarch64/arm_neon.h (__ST2_LANE_FUNC): Better model data size.
2265 (__ST3_LANE_FUNC): Likewise.
2266 (__ST4_LANE_FUNC): Likewise.
2268 2013-11-06 Nick Clifton <nickc@redhat.com>
2270 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define the
2271 name returned by msp430_mcu_name.
2272 (LIB_SPEC): If a -T option has not been specified then set a
2273 default, mcu-specific, linker script.
2274 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more mcu names.
2275 * config/msp430/msp430.c (msp430x_names): Likewise.
2276 Alpha sort the names for ease of comparison.
2277 (msp430_mcu_name): New function: Returns a string suitable for
2278 use as a C preprocessor symbol based upon the name of the MCU
2280 (msp430_option_override): Accept msp430x and msp430xv2 as generic
2282 * config/msp430/msp430-protos.h (msp430_mcu_name): Prototype.
2284 * gcc.c (do_spec_1): Do not insert a space after a %* substitution
2285 unless it is the last part of a spec substring.
2286 * doc/invoke.texi (Spec Files): Document space insertion
2289 2013-11-06 Christian Bruel <christian.bruel@st.com>
2291 * config/sh/sh-mem.cc (sh_expand_cmpnstr, sh_expand_cmpstr):
2292 Factorize probabilities, Use adjust_address instead of
2293 adjust_automodify_address when possible. Enable for optimize.
2294 (sh_expand_strlen): New function.
2295 * config/sh/sh-protos.h (sh_expand_strlen): Declare.
2296 * config/sh/sh.md (strlensi): New pattern.
2297 (UNSPEC_BUILTIN_STRLEN): Define.
2299 2013-11-06 Jakub Jelinek <jakub@redhat.com>
2302 * expr.c (get_bit_range): Handle *offset == NULL_TREE.
2303 (expand_assignment): If *bitpos is negative, set *offset
2304 and adjust *bitpos, so that it is not negative.
2306 2013-11-06 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
2308 * config/i386/bdver3.md : Added two additional decoder units
2309 to support issue rate of 4 and remodeled vector unit.
2310 * config/i386/i386.c (ix86_issue_rate): Issue rate for BD
2311 architectures is set to 4.
2312 * config/i386/i386.c (ia32_multipass_dfa_lookahead): DFA
2313 lookahead is set to 4 for BD architectures.
2315 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2317 * config/rs6000/rs6000.c (rs6000_option_override_internal):
2318 Remove restriction against use of VSX instructions when generating
2319 code for little endian mode.
2321 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2323 * config/rs6000/altivec.md (mulv4si3): Ensure we generate vmulouh
2324 for both big and little endian.
2325 (mulv8hi3): Swap input operands for merge high and merge low
2326 instructions for little endian.
2328 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2330 * config/rs6000/altivec.md (vec_widen_umult_even_v16qi): Change
2331 define_insn to define_expand that uses even patterns for big
2332 endian and odd patterns for little endian.
2333 (vec_widen_smult_even_v16qi): Likewise.
2334 (vec_widen_umult_even_v8hi): Likewise.
2335 (vec_widen_smult_even_v8hi): Likewise.
2336 (vec_widen_umult_odd_v16qi): Likewise.
2337 (vec_widen_smult_odd_v16qi): Likewise.
2338 (vec_widen_umult_odd_v8hi): Likewise.
2339 (vec_widen_smult_odd_v8hi): Likewise.
2340 (altivec_vmuleub): New define_insn.
2341 (altivec_vmuloub): Likewise.
2342 (altivec_vmulesb): Likewise.
2343 (altivec_vmulosb): Likewise.
2344 (altivec_vmuleuh): Likewise.
2345 (altivec_vmulouh): Likewise.
2346 (altivec_vmulesh): Likewise.
2347 (altivec_vmulosh): Likewise.
2349 2013-11-05 Mike Stump <mikestump@comcast.net>
2351 * Makefile.in (mostlyclean): Remove c-family objects.
2353 2013-11-05 Ian Lance Taylor <iant@google.com>
2355 * config/i386/sync.md (atomic_compare_and_swap<dwi>_doubleword):
2356 If possible, add .cfi directives to record change to bx.
2357 * config/i386/i386.c (ix86_emit_cfi): New function.
2358 * config/i386/i386-protos.h (ix86_emit_cfi): Declare.
2360 2013-11-05 Steven Bosscher <steven@gcc.gnu.org>
2362 * rtlanal.c (tablejump_p): Expect a JUMP_TABLE_DATA to always follow
2363 immediately after a label for a tablejump pattern.
2365 * config/arm/arm.c (is_jump_table): Remove.
2366 (create_fix_barrier): Use tablejump_p instead.
2367 (arm_reorg): Likewise.
2368 (thumb1_output_casesi): Expect JUMP_TABLE_DATA to always be NEXT_INSN.
2369 (thumb2_output_casesi): Likewise.
2370 * config/aarch64/aarch64.c (aarch64_output_casesi): Likewise.
2371 * config/sh/sh.md (casesi_worker_1, casesi_worker_2,
2372 casesi_shift_media, casesi_load_media): Likewise.
2373 * config/iq2000/iq2000.md: Likewise (in anonymous define_insn).
2374 * config/microblaze/microblaze.md: Likewise.
2376 2013-11-05 Tobias Burnus <burnus@net-b.de>
2378 * doc/invoke.texi (-Wdate-time): Document.
2380 2013-11-05 Richard Sandiford <rdsandiford@googlemail.com>
2382 * double-int.c (lshift_double, rshift_double): Remove
2383 SHIFT_COUNT_TRUNCATED handling.
2385 2013-11-05 Jeff Law <law@redhat.com>
2387 * Makefile.in (OBJS): Add gimple-ssa-isolate-paths.o
2388 * common.opt (-fisolate-erroneous-paths): Add option and documentation.
2389 * gimple-ssa-isolate-paths.c: New file.
2390 * gimple.c (check_loadstore): New function.
2391 (infer_nonnull_range): Moved into gimple.c from tree-vrp.c
2392 Verify OP is in the argument list and the argument corresponding
2393 to OP is a pointer type. Use operand_equal_p rather than
2394 pointer equality when testing if OP is on the nonnull list.
2395 Use check_loadstore rather than count_ptr_derefs. Handle
2396 GIMPLE_RETURN statements.
2397 * tree-vrp.c (infer_nonnull_range): Remove.
2398 * gimple.h (infer_nonnull_range): Declare.
2399 * opts.c (default_options_table): Add OPT_fisolate_erroneous_paths.
2400 * passes.def: Add pass_isolate_erroneous_paths.
2401 * timevar.def (TV_ISOLATE_ERRONEOUS_PATHS): New timevar.
2402 * tree-pass.h (make_pass_isolate_erroneous_paths): Declare.
2403 * tree-ssa.c (struct count_ptr_d): Remove.
2404 (count_ptr_derefs, count_uses_and_derefs): Remove.
2405 * tree-ssa.h (count_uses_and_derefs): Remove.
2407 2013-11-05 Jakub Jelinek <jakub@redhat.com>
2409 PR rtl-optimization/58997
2410 * loop-iv.c (iv_subreg): For IV_UNKNOWN_EXTEND, expect
2411 get_iv_value to be in iv->mode rather than iv->extend_mode.
2412 (iv_extend): Likewise. Otherwise, if iv->extend != extend,
2413 use lowpart_subreg on get_iv_value before calling simplify_gen_unary.
2414 * loop-unswitch.c (may_unswitch_on): Make sure op[i] is in the right
2417 2013-11-05 Andrew MacLeod <amacleod@redhat.com>
2419 * gimple.h: Move some prototypes to gimple-expr.h and add to include
2421 (extract_ops_from_tree, gimple_call_addr_fndecl, is_gimple_reg_type):
2422 Move to gimple-expr.h.
2423 * gimple-expr.h: New file. Relocate some prototypes from gimple.h.
2424 (types_compatible_p, is_gimple_reg_type, is_gimple_variable,
2425 is_gimple_id, virtual_operand_p, is_gimple_addressable,
2426 is_gimple_constant, extract_ops_from_tree, gimple_call_addr_fndecl):
2428 * gimple.c (extract_ops_from_tree_1, gimple_cond_get_ops_from_tree,
2429 gimple_set_body, gimple_body, gimple_has_body_p, is_gimple_lvalue,
2430 is_gimple_condexpr, is_gimple_addressable, is_gimple_constant,
2431 is_gimple_address, is_gimple_invariant_address,
2432 is_gimple_ip_invariant_address, is_gimple_min_invariant,
2433 is_gimple_ip_invariant, is_gimple_variable, is_gimple_id,
2434 virtual_operand_p, is_gimple_reg, is_gimple_val, is_gimple_asm_val,
2435 is_gimple_min_lval, is_gimple_call_addr, is_gimple_mem_ref_addr,
2436 gimple_decl_printable_name, useless_type_conversion_p,
2437 types_compatible_p, gimple_can_coalesce_p, copy_var_decl): Move to
2439 * gimple-expr.c: New File.
2440 (useless_type_conversion_p, gimple_set_body, gimple_body,
2441 gimple_has_body_p, gimple_decl_printable_name, copy_var_decl,
2442 gimple_can_coalesce_p, extract_ops_from_tree_1,
2443 gimple_cond_get_ops_from_tree, is_gimple_lvalue, is_gimple_condexpr,
2444 is_gimple_address, is_gimple_invariant_address,
2445 is_gimple_ip_invariant_address, is_gimple_min_invariant,
2446 is_gimple_ip_invariant, is_gimple_reg, is_gimple_val,
2447 is_gimple_asm_val, is_gimple_min_lval, is_gimple_call_addr,
2448 is_gimple_mem_ref_addr): Relocate here.
2449 * Makefile.in (OBJS): Add gimple-expr.o.
2451 2013-11-05 David Malcolm <dmalcolm@redhat.com>
2453 * gengtype-parse.c (struct_field_seq): Support empty structs.
2455 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
2457 * config/i386/t-rtems (MULTILIB_MATCHES): Fix option typos.
2459 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
2461 * config/i386/i386-c.c (ix86_target_macros): Define _SOFT_FLOAT
2463 * config/i386/rtemself.h (TARGET_OS_CPP_BUILTINS): Do not define
2465 (LONG_DOUBLE_TYPE_SIZE): New define.
2466 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Ditto.
2468 2013-11-05 Paolo Carlini <paolo.carlini@oracle.com>
2471 * doc/extend.texi [visibility ("visibility_type")]: Add example
2472 about visibility attribute on namespace declaration.
2474 2013-11-05 Richard Biener <rguenther@suse.de>
2477 * passes.def (all_passes): Start with pass_fixup_cfg again.
2479 2013-11-05 Richard Biener <rguenther@suse.de>
2481 PR tree-optimization/58955
2482 * tree-loop-distribution.c (pg_add_dependence_edges): Fix
2485 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2487 * config/rs6000/vector.md (vec_pack_sfix_trunc_v2df): Adjust for
2489 (vec_pack_ufix_trunc_v2df): Likewise.
2491 2013-11-05 H.J. Lu <hongjiu.lu@intel.com>
2494 * doc/md.texi (@code{movmem@var{m}}): Specify Pmode as mode of
2495 pattern, instead of word_mode.
2497 * expr.c (emit_block_move_via_movmem): Don't use mode wider than
2499 (set_storage_via_setmem): Likewise.
2501 2013-11-05 Andrew MacLeod <amacleod@redhat.com>
2503 * tree-outof-ssa.c (queue_phi_copy_p): Combine phi_ssa_name_p from
2504 gimple.h and the rest of the condition in eliminate_build.
2505 (eliminate_build): Call new routine.
2506 * gimple.h (phi_ssa_name_p): Delete.
2508 2013-11-05 Trevor Saunders <tsaunders@mozilla.com>
2510 * vec.c (vec_prefix::calculate_allocation): Don't try to handle the
2511 case of no prefix and reserving zero slots, because when that's the
2512 case we'll never get here.
2513 * vec.h (va_heap::reserve): Don't try and handle
2514 vec_prefix::calculate_allocation returning zero because that should
2517 2013-11-05 Richard Biener <rguenther@suse.de>
2520 * tree-dfa.c (get_ref_base_and_extent): Merge common code
2521 in MEM_REF and TARGET_MEM_REF handling. Make sure to
2522 process trailing array detection before diving into the
2523 view-converted object (and possibly apply some extra offset).
2525 2013-11-05 Joseph Myers <joseph@codesourcery.com>
2527 * config/i386/i386.c (ix86_float_exceptions_rounding_supported_p):
2529 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): Define.
2531 2013-11-05 Marc Glisse <marc.glisse@inria.fr>
2533 PR tree-optimization/58958
2534 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Use
2535 get_addr_base_and_unit_offset instead of get_ref_base_and_extent.
2537 2013-11-05 Marc Glisse <marc.glisse@inria.fr>
2539 * tree-ssa-alias.h (ranges_overlap_p): Handle negative offsets.
2540 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Likewise.
2542 2013-11-05 Jakub Jelinek <jakub@redhat.com>
2544 PR tree-optimization/58984
2545 * ipa-prop.c (ipa_load_from_parm_agg_1): Add SIZE_P argument,
2546 set *SIZE_P if non-NULL on success.
2547 (ipa_load_from_parm_agg, ipa_analyze_indirect_call_uses): Adjust
2549 (ipcp_transform_function): Likewise. Punt if size of access
2550 is different from TYPE_SIZE on v->value's type.
2552 2013-11-05 Tobias Burnus <burnus@net-b.de>
2554 * doc/invoke.texi (-fopenmp-simd): Document new option.
2555 * gimplify.c (gimplify_body): Accept -fopenmp-simd.
2556 * omp-low.c (execute_expand_omp, execute_lower_omp): Ditto.
2557 * tree.c (attribute_value_equal): Ditto.
2559 2013-11-04 Wei Mi <wmi@google.com>
2561 * sched-rgn.c (add_branch_dependences): Keep insns in
2562 a SCHED_GROUP at the end of BB to remain their location.
2564 2013-11-04 Wei Mi <wmi@google.com>
2566 * config/i386/i386.c (memory_address_length): Extract a part
2567 of code to rip_relative_addr_p.
2568 (rip_relative_addr_p): New Function.
2569 (ix86_macro_fusion_p): Ditto.
2570 (ix86_macro_fusion_pair_p): Ditto.
2571 * config/i386/i386.h: Add new tune features about macro-fusion.
2572 * config/i386/x86-tune.def (DEF_TUNE): Ditto.
2573 * doc/tm.texi: Generated.
2574 * doc/tm.texi.in: Ditto.
2575 * haifa-sched.c (try_group_insn): New Function.
2576 (group_insns_for_macro_fusion): Ditto.
2577 (sched_init): Call group_insns_for_macro_fusion.
2578 * target.def: Add two hooks: macro_fusion_p and
2579 macro_fusion_pair_p.
2581 2013-11-04 Kostya Serebryany <kcc@google.com>
2583 Update to match the changed asan API.
2584 * asan.c (asan_function_start): New function.
2585 (asan_emit_stack_protection): Update the string stored in the
2586 stack red zone to match new API. Store the PC of the current
2587 function in the red zone.
2588 (asan_global_struct): Update the __asan_global definition to match
2590 (asan_add_global): Ditto.
2591 * asan.h (asan_function_start): New prototype.
2592 * final.c (final_start_function): Call asan_function_start.
2593 * sanitizer.def (BUILT_IN_ASAN_INIT): Rename __asan_init_v1
2596 2013-11-04 Wei Mi <wmi@google.com>
2598 * config/i386/i386-c.c (ix86_target_macros_internal): Separate
2599 PROCESSOR_COREI7_AVX out from PROCESSOR_COREI7.
2600 * config/i386/i386.c (ix86_option_override_internal): Ditto.
2601 (ix86_issue_rate): Ditto.
2602 (ix86_adjust_cost): Ditto.
2603 (ia32_multipass_dfa_lookahead): Ditto.
2604 (ix86_sched_init_global): Ditto.
2605 (get_builtin_code_for_version): Ditto.
2606 * config/i386/i386.h (enum target_cpu_default): Ditto.
2607 (enum processor_type): Ditto.
2608 * config/i386/x86-tune.def (DEF_TUNE): Ditto.
2610 2013-11-04 Vladimir Makarov <vmakarov@redhat.com>
2612 PR rtl-optimization/58967
2613 * config/rs6000/rs6000.c (legitimate_lo_sum_address_p): Remove
2614 !lra_in_progress for mode sizes bigger word.
2616 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2618 * config/rs6000/altivec.md (vec_widen_umult_hi_v16qi): Swap
2619 arguments to merge instruction for little endian.
2620 (vec_widen_umult_lo_v16qi): Likewise.
2621 (vec_widen_smult_hi_v16qi): Likewise.
2622 (vec_widen_smult_lo_v16qi): Likewise.
2623 (vec_widen_umult_hi_v8hi): Likewise.
2624 (vec_widen_umult_lo_v8hi): Likewise.
2625 (vec_widen_smult_hi_v8hi): Likewise.
2626 (vec_widen_smult_lo_v8hi): Likewise.
2628 2013-11-04 Ian Lance Taylor <iant@google.com>
2630 * builtins.def (ATTR_NOTHROWCALL_LEAF_LIST): Define.
2631 * sync-builtins.def: Use ATTR_NOTHROWCALL_LEAF_LIST for all sync
2632 builtins that take pointers.
2633 * lto-opts.c (lto_write_options): Write -fnon-call-exceptions if set.
2634 * lto-wrapper.c (merge_and_complain): Collect OPT_fnon_call_exceptions.
2635 (run_gcc): Pass -fnon-call-exceptions.
2637 2013-11-04 Jakub Jelinek <jakub@redhat.com>
2639 * optabs.c (expand_vec_perm): Revert one incorrect line from
2642 PR tree-optimization/58978
2643 * tree-vrp.c (all_imm_uses_in_stmt_or_feed_cond): Don't modify
2644 use_stmt by single_imm_use directly. Only call single_imm_use
2647 2013-11-04 Vladimir Makarov <vmakarov@redhat.com>
2649 PR rtl-optimization/58968
2650 * lra-spills.c (return_regno_p): New function.
2651 (lra_final_code_change): Use it.
2653 2013-11-04 Joseph Myers <joseph@codesourcery.com>
2655 * doc/cpp.texi (__GCC_IEC_559, __GCC_IEC_559_COMPLEX): Document macros.
2656 * target.def (float_exceptions_rounding_supported_p): New hook.
2657 * targhooks.c (default_float_exceptions_rounding_supported_p): New
2659 * targhooks.h (default_float_exceptions_rounding_supported_p): Declare.
2660 * doc/tm.texi.in (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P):
2662 * doc/tm.texi: Regenerate.
2663 * config.gcc (powerpc*-*-linux*): Set extra_objs.
2664 * config/rs6000/rs6000-linux.c: New file.
2665 * config/rs6000/rs6000-protos.h
2666 (rs6000_linux_float_exceptions_rounding_supported_p): Declare.
2667 * config/rs6000/linux.h
2668 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): New macro.
2669 * config/rs6000/linux64.h
2670 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): Likewise.
2671 * config/rs6000/t-linux (rs6000-linux.o): New rule.
2672 * config/rs6000/t-linux64 (rs6000-linux.o): Likewise.
2674 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2676 * config/rs6000/vsx.md (*vsx_le_perm_store_<mode> for VSX_D):
2677 Replace the define_insn_and_split with a define_insn and two
2678 define_splits, with the split after reload re-permuting the source
2679 register to its original value.
2680 (*vsx_le_perm_store_<mode> for VSX_W): Likewise.
2681 (*vsx_le_perm_store_v8hi): Likewise.
2682 (*vsx_le_perm_store_v16qi): Likewise.
2684 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2686 * config/rs6000/vector.md (vec_pack_trunc_v2df): Adjust for
2689 2013-11-04 Jakub Jelinek <jakub@redhat.com>
2691 PR tree-optimization/58946
2692 * tree-ssa-reassoc.c (maybe_optimize_range_tests): Update all
2693 bbs with bbinfo[idx].op != NULL before all blocks with
2694 bbinfo[idx].op == NULL.
2696 2013-11-04 Richard Sandiford <rdsandiford@googlemail.com>
2698 * config/avr/avr-log.c (avr_double_int_pop_digit): Delete.
2699 (avr_dump_double_int_hex): Likewise.
2700 (avr_log_vadump): Remove %D and %X handling.
2701 * config/avr/avr.c (avr_double_int_push_digit): Delete.
2702 (avr_map_op_t): Change map from double_int to unsigned int.
2703 (avr_map_op): Update accordingly.
2704 (avr_map, avr_map_metric, avr_has_nibble_0xf, avr_map_decompose)
2705 (avr_move_bits, avr_out_insert_bits, avr_fold_builtin): Operate on
2706 unsigned ints rather than double_ints.
2708 2013-11-03 Marek Polacek <polacek@redhat.com>
2710 Implement -fsanitize=vla-bound.
2711 * opts.c (common_handle_option): Handle vla-bound.
2712 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE): Define.
2713 * flag-types.h (enum sanitize_code): Add SANITIZE_VLA.
2714 * asan.c (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR.
2716 2013-11-02 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
2718 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Adjust for
2721 2013-11-02 Uros Bizjak <ubizjak@gmail.com>
2723 * config/i386/constraints.md (Ts, Tv): New address constrains.
2724 * config/i386/i386.md (*lea<mode>, *<mode>_<bndcheck>): Use Ts
2725 constraint for address_no_seg_operand.
2726 * config/i386/sse.md (*avx512pf_gatherpf<mode>_mask)
2727 (*avx512pf_gatherpf<mode>, *avx512pf_scatterpf<mode>_mask)
2728 (*avx512pf_scatterpf<mode>, *avx2_gathersi<mode>)
2729 (*avx2_gathersi<mode>_2, *avx2_gatherdi<mode>, *avx2_gatherdi<mode>_2)
2730 (*avx2_gatherdi<mode>_3, *avx2_gatherdi<mode>_4)
2731 (*avx512f_gathersi<mode>, *avx512f_gathersi<mode>_2)
2732 (*avx512f_gatherdi<mode>, *avx512f_gatherdi<mode>_2)
2733 (*avx512f_scattersi<mode> *avx512f_scatterdi<mode>): Use Tv
2734 constraint for vsib_address_operand.
2736 2013-11-02 Steven Bosscher <steven@gcc.gnu.org>
2738 * gcse.c (pre_delete): Remove references to regmove from comments.
2739 * recog.c: (validate_replace_rtx_1): Likewise.
2740 * config/rl78/rl78.c: Likewise.
2741 * config/v850/v850.h: Likewise, and remove unused ENABLE_REGMOVE_PASS.
2742 * common/config/m32r/m32r-common.c: Don't manipulate OPT_fregmove.
2743 * common/config/mmix/mmix-common.c: Likewise.
2745 2013-11-01 Trevor Saunders <tsaunders@mozilla.com>
2747 * function.c (reorder_blocks): Convert block_stack to a stack_vec.
2748 * gimplify.c (gimplify_compound_lval): Likewise.
2749 * graphite-clast-to-gimple.c (gloog): Likewise.
2750 * graphite-dependences.c (loop_is_parallel_p): Likewise.
2751 * graphite-scop-detection.c (scopdet_basic_block_info): Likewise.
2752 (limit_scop); Likewise.
2753 (build_scops): Likewise.
2754 (dot_scop): Likewise.
2755 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
2756 (build_scop_drs): Likewise.
2757 (insert_stmts): Likewise.
2758 (insert_out_of_ssa_copy): Likewise.
2759 (remove_phi): Likewise.
2760 (rewrite_commutative_reductions_out_of_ssa_close_phi): Likewise.
2761 * hw-doloop.c (discover_loop): Likewise.
2762 * tree-call-cdce.c (shrink_wrap_one_built_in_call): Likewise.
2763 * tree-dfa.c (dump_enumerated_decls): Likewise.
2764 * tree-if-conv.c (if_convertable_loop_p): Likewise.
2765 * tree-inline.c (tree_function_versioning): Likewise.
2766 * tree-loop-distribution.c (build_rdg): Likewise.
2767 (rdg_flag_vertex_and_dependent): Likewise.
2768 (distribute_loop): Likewise.
2769 * tree-parloops.c (loop_parallel_p): Likewise.
2770 (eliminate_local_variables): Likewise.
2771 (separate_decls_in_region): Likewise.
2772 * tree-predcom.c (tree_predictive_commoning_loop): Likewise.
2773 * tree-ssa-phiopt.c (cond_if_else_store_replacement): Likewise.
2774 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
2775 * tree-vect-loop.c (vect_analyze_scaler_cycles_1): Likewise.
2776 * tree-vect-patterns.c (vect_pattern_recog): Likewise.
2777 * tree-vect-stmts.c (vect_mark_stmts_to_be_vectorized): Likewise.
2778 (vectorizable_condition): Likewise.
2780 2013-11-01 Uros Bizjak <ubizjak@gmail.com>
2782 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): Always define as 0/1.
2783 * configure: Regenerate.
2784 * config/i386/i386.md (*movdi_internal): Change
2785 HAVE_AS_IX86_INTERUNIT_MOVQ to runtime check.
2786 (*movdf_internal): Ditto.
2787 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
2788 * config/i386/sse.md (vec_concatv2di): Output interunit movq
2789 for HAVE_AS_IX86_INTERUNIT_MOVQ targets.
2791 2013-10-31 Robert Suchanek <Robert.Suchanek@imgtec.com>
2793 * lra-spills.c (assign_spill_hard_regs): Remove statement terminator
2796 2013-10-31 David Malcolm <dmalcolm@redhat.com>
2798 Automated part of renaming of symtab_node_base to symtab_node.
2800 Patch autogenerated by rename_symtab.py from
2801 https://github.com/davidmalcolm/gcc-refactoring-scripts
2802 revision 58bb219cc090b2f4516a9297d868c245495ee622
2803 with ChangeLog entry fixed up by hand.
2805 * cgraph.c (x_cgraph_nodes_queue): Rename symtab_node_base to
2807 (cgraph_node_for_asm): Likewise.
2808 * cgraph.h (symtab_node_base): Likewise.
2809 (cgraph_node): Likewise.
2810 (varpool_node): Likewise.
2811 (is_a_helper <cgraph_node>::test): Likewise.
2812 (is_a_helper <varpool_node>::test): Likewise.
2813 (symtab_nodes): Likewise.
2814 (symtab_register_node): Likewise.
2815 (symtab_unregister_node): Likewise.
2816 (symtab_remove_node): Likewise.
2817 (symtab_get_node): Likewise.
2818 (symtab_node_for_asm): Likewise.
2819 (symtab_node_asm_name): Likewise.
2820 (symtab_node_name): Likewise.
2821 (symtab_insert_node_to_hashtable): Likewise.
2822 (symtab_add_to_same_comdat_group): Likewise.
2823 (symtab_dissolve_same_comdat_group_list): Likewise.
2824 (dump_symtab_node): Likewise.
2825 (debug_symtab_node): Likewise.
2826 (dump_symtab_base): Likewise.
2827 (verify_symtab_node): Likewise.
2828 (verify_symtab_base): Likewise.
2829 (symtab_used_from_object_file_p): Likewise.
2830 (symtab_alias_ultimate_target): Likewise.
2831 (symtab_resolve_alias): Likewise.
2832 (fixup_same_cpp_alias_visibility): Likewise.
2833 (symtab_for_node_and_aliases): Likewise.
2834 (symtab_nonoverwritable_alias): Likewise.
2835 (availability symtab_node_availability): Likewise.
2836 (symtab_semantically_equivalent_p): Likewise.
2837 (fixup_same_cpp_alias_visibility): Likewise.
2838 (symtab_prevail_in_asm_name_hash): Likewise.
2840 (varpool): Likewise.
2841 (varpool_first_variable): Likewise.
2842 (varpool_next_variable): Likewise.
2843 (varpool_first_static_initializer): Likewise.
2844 (varpool_next_static_initializer): Likewise.
2845 (varpool_first_defined_variable): Likewise.
2846 (varpool_next_defined_variable): Likewise.
2847 (cgraph_first_defined_function): Likewise.
2848 (cgraph_next_defined_function): Likewise.
2849 (cgraph_first_function): Likewise.
2850 (cgraph_next_function): Likewise.
2851 (cgraph_first_function_with_gimple_body): Likewise.
2852 (cgraph_next_function_with_gimple_body): Likewise.
2853 (symtab_alias_target): Likewise.
2854 (symtab_real_symbol_p): Likewise.
2855 (symtab_can_be_discarded): Likewise.
2856 * cgraphbuild.c (mark_address): Likewise.
2857 (mark_load): Likewise.
2858 (mark_store): Likewise.
2859 * cgraphunit.c (decide_is_symbol_needed): Likewise.
2861 (enqueue_node): Likewise.
2862 (referred_to_p): Likewise.
2863 (cgraph_process_same_body_aliases): Likewise.
2864 (analyze_functions): Likewise.
2865 (handle_alias_pairs): Likewise.
2866 (output_weakrefs): Likewise.
2867 (compile): Likewise.
2868 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
2869 * ipa-inline-analysis.c (inline_write_summary): Likewise.
2870 * ipa-prop.c (remove_described_reference): Likewise.
2871 (try_decrement_rdesc_refcount): Likewise.
2872 (ipa_edge_duplication_hook): Likewise.
2873 * ipa-ref.c (ipa_record_reference): Likewise.
2874 (ipa_maybe_record_reference): Likewise.
2875 (ipa_clone_ref): Likewise.
2876 (ipa_clone_references): Likewise.
2877 (ipa_clone_referring): Likewise.
2878 (ipa_find_reference): Likewise.
2879 (ipa_remove_stmt_references): Likewise.
2880 (ipa_clear_stmts_in_references): Likewise.
2881 * ipa-ref.h (symtab_node_base): Likewise.
2882 (ipa_ref): Likewise.
2883 (ipa_record_reference): Likewise.
2884 (ipa_maybe_record_reference): Likewise.
2885 (ipa_clone_references): Likewise.
2886 (ipa_clone_referring): Likewise.
2887 (ipa_clone_ref): Likewise.
2888 (ipa_find_reference): Likewise.
2889 (ipa_remove_stmt_references): Likewise.
2890 (ipa_clear_stmts_in_references): Likewise.
2891 * ipa-reference.c (ipa_reference_write_optimization_summary):
2893 * ipa.c (enqueue_node): Likewise.
2894 (process_references): Likewise.
2895 (walk_polymorphic_call_targets): Likewise.
2896 (symtab_remove_unreachable_nodes): Likewise.
2897 (address_taken_from_non_vtable_p): Likewise.
2898 (comdat_can_be_unshared_p_1): Likewise.
2899 (comdat_can_be_unshared_p): Likewise.
2900 (can_replace_by_local_alias): Likewise.
2901 (function_and_variable_visibility): Likewise.
2902 * is-a.h: Likewise (within example in comment).
2903 * lto-cgraph.c (input_cgraph_opt_summary): Likewise.
2904 (lto_symtab_encoder_encode): Likewise.
2905 (lto_symtab_encoder_delete_node): Likewise.
2906 (lto_symtab_encoder_in_partition_p): Likewise.
2907 (lto_set_symtab_encoder_in_partition): Likewise.
2908 (output_refs): Likewise.
2909 (compute_ltrans_boundary): Likewise.
2910 (output_symtab): Likewise.
2911 (input_node): Likewise.
2912 (input_ref): Likewise.
2913 (input_edge): Likewise.
2914 (input_cgraph_1): Likewise.
2915 (input_refs): Likewise.
2916 (output_cgraph_opt_summary): Likewise.
2917 (input_node_opt_summary): Likewise.
2918 (input_cgraph_opt_section): Likewise.
2919 * lto-section-in.c (lto_free_function_in_decl_state_for_node):
2921 * lto-streamer-out.c (lto_output): Likewise.
2922 (output_symbol_p): Likewise.
2923 (produce_symtab): Likewise.
2924 * lto-streamer.h (lto_encoder_entry): Likewise.
2925 (lto_free_function_in_decl_state_for_node): Likewise.
2926 (lto_symtab_encoder_encode): Likewise.
2927 (lto_symtab_encoder_delete_node): Likewise.
2928 (lto_symtab_encoder_in_partition_p): Likewise.
2929 (lto_set_symtab_encoder_in_partition): Likewise.
2930 (lto_symtab_encoder_lookup): Likewise.
2931 (lsei_node): Likewise.
2932 (lto_symtab_encoder_deref): Likewise.
2933 * symtab.c (symtab_hash): Likewise.
2934 (assembler_name_hash): Likewise.
2935 (symtab_nodes): Likewise.
2936 (hash_node): Likewise.
2937 (eq_node): Likewise.
2938 (hash_node_by_assembler_name): Likewise.
2939 (eq_assembler_name): Likewise.
2940 (insert_to_assembler_name_hash): Likewise.
2941 (unlink_from_assembler_name_hash): Likewise.
2942 (symtab_prevail_in_asm_name_hash): Likewise.
2943 (symtab_register_node): Likewise.
2944 (symtab_insert_node_to_hashtable): Likewise.
2945 (symtab_unregister_node): Likewise.
2946 (symtab_get_node): Likewise.
2947 (symtab_remove_node): Likewise.
2948 (symtab_initialize_asm_name_hash): Likewise.
2949 (symtab_node_for_asm): Likewise.
2950 (symtab_add_to_same_comdat_group): Likewise.
2951 (symtab_dissolve_same_comdat_group_list): Likewise.
2952 (symtab_node_asm_name): Likewise.
2953 (symtab_node_name): Likewise.
2954 (dump_symtab_base): Likewise.
2955 (dump_symtab_node): Likewise.
2956 (dump_symtab): Likewise.
2957 (debug_symtab_node): Likewise.
2958 (verify_symtab_base): Likewise.
2959 (verify_symtab_node): Likewise.
2960 (verify_symtab): Likewise.
2961 (symtab_used_from_object_file_p): Likewise.
2962 (symtab_node_availability): Likewise.
2963 (symtab_alias_ultimate_target): Likewise.
2964 (fixup_same_cpp_alias_visibility): Likewise.
2965 (symtab_resolve_alias): Likewise.
2966 (symtab_for_node_and_aliases): Likewise.
2967 (symtab_for_node_and_aliases): Likewise.
2968 (symtab_nonoverwritable_alias_1): Likewise.
2969 (symtab_nonoverwritable_alias): Likewise.
2970 (symtab_semantically_equivalent_p): Likewise.
2971 * value-prof.c (init_node_map): Likewise.
2972 * varasm.c (find_decl): Likewise.
2973 * varpool.c (varpool_node_for_asm): Likewise.
2974 (varpool_remove_unreferenced_decls): Likewise.
2976 2013-10-31 David Malcolm <dmalcolm@redhat.com>
2978 Manual part of renaming of symtab_node_base to symtab_node.
2980 * ipa-ref.h (symtab_node): Remove typedef to pointer type, as it
2981 clashes with the preferred name for the base class.
2982 (const_symtab_node): Remove redundant typedef.
2984 2013-10-31 Jakub Jelinek <jakub@redhat.com>
2986 * optabs.c (expand_vec_perm): Avoid vector mode punning
2987 SUBREGs in SET_DEST.
2988 * expmed.c (store_bit_field_1): Likewise.
2989 * config/i386/sse.md (movdi_to_sse, vec_pack_sfix_trunc_v2df,
2990 vec_pack_sfix_v2df, vec_shl_<mode>, vec_shr_<mode>,
2991 vec_interleave_high<mode>, vec_interleave_low<mode>): Likewise.
2992 * config/i386/i386.c (ix86_expand_vector_move_misalign,
2993 ix86_expand_sse_movcc, ix86_expand_int_vcond, ix86_expand_vec_perm,
2994 ix86_expand_sse_unpack, ix86_expand_args_builtin,
2995 ix86_expand_vector_init_duplicate, ix86_expand_vector_set,
2996 emit_reduc_half, expand_vec_perm_blend, expand_vec_perm_pshufb,
2997 expand_vec_perm_interleave2, expand_vec_perm_pshufb2,
2998 expand_vec_perm_vpshufb2_vpermq,
2999 expand_vec_perm_vpshufb2_vpermq_even_odd, expand_vec_perm_even_odd_1,
3000 expand_vec_perm_broadcast_1, expand_vec_perm_vpshufb4_vpermq2,
3001 ix86_expand_sse2_mulv4si3, ix86_expand_pinsr): Likewise.
3002 (expand_vec_perm_palignr): Likewise. Modify a copy of *d rather
3005 2013-10-31 Uros Bizjak <ubizjak@gmail.com>
3007 * config/i386/i386.c (ix86_expand_sse2_abs): Rename function arguments.
3008 Use gcc_unreachable for unhandled modes. Do not check results of
3009 expand_simple_binop. If not expanded to target, move the result.
3011 2013-10-31 Chung-Ju Wu <jasonwucj@gmail.com>
3012 Shiva Chen <shiva0217@gmail.com>
3014 * config.gcc (nds32*-*-*): Add nds32 target.
3015 * config/nds32/nds32.c: New file.
3016 * config/nds32/nds32.h: New file.
3017 * config/nds32/nds32.md: New file.
3018 * config/nds32/constants.md: New file.
3019 * config/nds32/constraints.md: New file.
3020 * config/nds32/iterators.md: New file.
3021 * config/nds32/nds32-doubleword.md: New file.
3022 * config/nds32/nds32-intrinsic.md: New file.
3023 * config/nds32/nds32_intrinsic.h: New file.
3024 * config/nds32/nds32-modes.def: New file.
3025 * config/nds32/nds32-multiple.md: New file.
3026 * config/nds32/nds32.opt: New file.
3027 * config/nds32/nds32-opts.h: New file.
3028 * config/nds32/nds32-protos.h: New file.
3029 * config/nds32/nds32-peephole2.md: New file.
3030 * config/nds32/pipelines.md: New file.
3031 * config/nds32/predicates.md: New file.
3032 * config/nds32/t-mlibs: New file.
3033 * common/config/nds32: New directory and files.
3035 * doc/invoke.texi (NDS32 options): Document nds32 specific options.
3036 * doc/md.texi (NDS32 family): Document nds32 specific constraints.
3037 * doc/install.texi (Cross-Compiler-Specific Options): Document
3038 --with-nds32-lib for nds32 target.
3039 * doc/extend.texi (Function Attributes, Target Builtins): Document
3040 nds32 specific attributes.
3042 2013-10-31 Vladimir Makarov <vmakarov@redhat.com>
3044 * lra-constraints (process_alt_operands): Use the result
3045 elimination register for operand when matching constraints.
3047 2013-10-31 Jakub Jelinek <jakub@redhat.com>
3049 * tree-vrp.c (maybe_set_nonzero_bits): New function.
3050 (remove_range_assertions): Call it.
3052 * tree.c (tree_ctz): New function.
3053 * tree.h (tree_ctz): New prototype.
3054 * tree-ssanames.h (get_range_info, get_nonzero_bits): Change
3055 first argument from tree to const_tree.
3056 * tree-ssanames.c (get_range_info, get_nonzero_bits): Likewise.
3057 * tree-vectorizer.h (vect_generate_tmps_on_preheader): New prototype.
3058 * tree-vect-loop-manip.c (vect_generate_tmps_on_preheader): No longer
3060 * expr.c (highest_pow2_factor): Reimplemented using tree_ctz.
3061 * tree-vect-loop.c (vect_analyze_loop_operations,
3062 vect_transform_loop): Don't force scalar loop for bound just because
3063 number of iterations is unknown, only do it if it is not known to be
3064 a multiple of vectorization_factor.
3065 * builtins.c (get_object_alignment_2): Use tree_ctz on offset.
3067 * gimple-pretty-print.c (dump_ssaname_info): Print newline also
3068 in case of VR_VARYING. Print get_nonzero_bits if not all ones.
3069 * tree-ssanames.h (struct range_info_def): Add nonzero_bits field.
3070 (set_nonzero_bits, get_nonzero_bits): New prototypes.
3071 * tree-ssa-ccp.c (get_default_value): Use get_range_info to see if
3072 a default def isn't partially constant.
3073 (ccp_finalize): If after IPA, set_range_info if integral SSA_NAME
3074 is known to be partially zero.
3075 (evaluate_stmt): If we'd return otherwise VARYING, use get_range_info
3076 to see if a default def isn't partially constant.
3077 * tree-ssanames.c (set_range_info): Initialize nonzero_bits upon
3078 creation of a range, if VR_RANGE, try to improve nonzero_bits from
3080 (set_nonzero_bits, get_nonzero_bits): New functions.
3082 * tree-cfg.c (assert_unreachable_fallthru_edge_p): New function.
3083 * tree-cfg.h (assert_unreachable_fallthru_edge_p): New prototype.
3084 * tree-vrp.c (all_imm_uses_in_stmt_or_feed_cond): New function.
3085 (remove_range_assertions): If ASSERT_EXPR_VAR has no other immediate
3086 uses but in the condition and ASSERT_EXPR and the other successor of
3087 the predecessor bb is __builtin_unreachable (), set_range_info of the
3088 ASSERT_EXPR_VAR to the range info of the ASSERT_EXPR's lhs.
3090 2013-10-31 Martin Jambor <mjambor@suse.cz>
3092 PR rtl-optimization/58934
3094 2013-10-30 Martin Jambor <mjambor@suse.cz>
3095 PR rtl-optimization/10474
3096 * ira.c (find_moveable_pseudos): Do not calculate dominance info
3098 (interesting_dest_for_shprep): New function.
3099 (split_live_ranges_for_shrink_wrap): Likewise.
3100 (ira): Calculate dominance info and df analysis. Call
3101 split_live_ranges_for_shrink_wrap.
3103 2013-10-31 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
3104 Yury Gribov <y.gribov@samsung.com>
3107 * asan.c (asan_clear_shadow): Allocate a new vreg for temporary
3108 shadow pointer to avoid clobbering the main one.
3110 2013-10-31 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3112 * lower-subreg.c (resolve_simple_move): Copy REG_INC note.
3114 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
3117 * ira-color.c (update_costs_from_copies): Add new parameter. Use
3118 it for calling update_costs_from_allocno.
3119 (assign_hard_reg): Call restore_costs_from_copies only for
3120 !retry_p. Pass new argument to update_costs_from_copies.
3121 (color_pass): Pass new argument to update_costs_from_copies.
3122 (ira_mark_allocation_change): Ditto.
3124 2013-10-30 Sharad Singhai <singhai@google.com>
3127 * opts.c (common_handle_option): Remove deprecated option
3128 -ftree-vectorizer-verbose.
3129 * doc/invoke.texi (Debugging Options): Ditto.
3130 * opts-global.c (handle_common_deferred_options): Ditto.
3131 (dump_remap_tree_vectorizer_verbose): Delete.
3132 * common.opt: Set -ftree-vectorizer-verbose as an ignored option.
3134 2013-10-30 DJ Delorie <dj@redhat.com>
3136 * config/rx/rx.c (ADD_RX_BUILTIN0): New macro, used for builtins
3137 that take no arguments.
3139 2013-10-30 Joern Rennecke <joern.rennecke@embecosm.com>
3142 * reload1.c (update_eliminables_and_spill): New function, broken
3144 (reload): Use it. Check for frame size change after frame size
3145 alignment, and call update_eliminables_and_spill first if continue-ing.
3147 2013-10-30 Cong Hou <congh@google.com>
3150 * config/i386/i386-protos.h (ix86_expand_sse2_abs): New function.
3151 * config/i386/i386.c (ix86_expand_sse2_abs): New function.
3152 * config/i386/sse.md: Add SSE2 support to abs (8/16/32-bit-int).
3154 2013-10-18 Mikael Pettersson <mikpelinux@gmail.com>
3156 PR rtl-optimization/58369
3157 * reload1.c (compute_reload_subreg_offset): New function.
3158 (choose_reload_regs): Use it to pass endian-correct
3159 offset to subreg_regno_offset.
3161 2013-10-30 Tobias Burnus <burnus@net-b.de>
3164 * tree-cfg.c (replace_loop_annotate): Replace warning by
3167 2013-10-30 Jason Merrill <jason@redhat.com>
3169 * configure.ac (loose_warn): Add -Wno-format if
3170 --disable-build-format-warnings.
3172 2013-10-30 David Malcolm <dmalcolm@redhat.com>
3174 * cgraphunit.c (analyze_functions): Split symtab_node declarations
3175 onto multiple lines to make things easier for rename_symtab.py.
3177 * symtab.c (symtab_dissolve_same_comdat_group_list): Likewise.
3178 (symtab_semantically_equivalent_p): Likewise.
3180 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
3183 * lra.c (check_rtl): Remove address check before LRA work.
3185 2013-10-30 Marc Glisse <marc.glisse@inria.fr>
3187 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Look for a
3188 POINTER_PLUS_EXPR in the defining statement.
3190 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
3192 * regmove.c: Remove.
3193 * tree-pass.h (make_pass_regmove): Remove.
3194 * timevar.def (TV_REGMOVE): Remove.
3195 * passes.def (pass_regmove): Remove.
3196 * opts.c (default_options_table): Remove entry for regmove.
3197 * doc/passes.texi: Remove regmove pass description.
3198 * doc/invoke.texi (-foptimize-register-move, -fregmove): Remove
3200 (-fdump-rtl-regmove): Ditto.
3201 * common.opt (foptimize-register-move, fregmove): Ignore.
3202 * Makefile.in (OBJS): Remove regmove.o.
3203 * regmove.c: Remove.
3204 * ira-int.h (struct ira_allocno_pref, ira_pref_t): New structure
3206 (struct ira_allocno) New member allocno_prefs.
3207 (ALLOCNO_PREFS): New macro.
3208 (ira_prefs, ira_prefs_num): New external vars.
3209 (ira_setup_alts, ira_get_dup_out_num, ira_debug_pref): New prototypes.
3210 (ira_debug_prefs, ira_debug_allocno_prefs, ira_create_pref): Ditto.
3211 (ira_add_allocno_pref, ira_remove_pref, ira_remove_allocno_prefs):
3213 (ira_add_allocno_copy_to_list): Remove prototype.
3214 (ira_swap_allocno_copy_ends_if_necessary): Ditto.
3215 (ira_pref_iterator): New type.
3216 (ira_pref_iter_init, ira_pref_iter_cond): New functions.
3217 (FOR_EACH_PREF): New macro.
3218 * ira.c (commutative_constraint_p): Move from ira-conflicts.c.
3219 (ira_get_dup_out_num): Ditto. Rename from get_dup_num. Modify the
3221 (ira_setup_alts): New function.
3222 (decrease_live_ranges_number): New function.
3223 (ira): Call the above function.
3224 * ira-build.c (ira_prefs, ira_prefs_num): New global vars.
3225 (ira_create_allocno): Initialize allocno prefs.
3226 (pref_pool, pref_vec): New static vars.
3227 (initiate_prefs, find_allocno_pref, ira_create_pref): New
3229 (add_allocno_pref_to_list, ira_add_allocno_pref, print_pref): Ditto.
3230 (ira_debug_pref, print_prefs, ira_debug_prefs): Ditto.
3231 (print_allocno_prefs, ira_debug_allocno_prefs, finish_pref): Ditto.
3232 (ira_remove_pref, ira_remove_allocno_prefs, finish_prefs): Ditto.
3233 (ira_add_allocno_copy_to_list): Make static. Rename to
3234 add_allocno_copy_to_list.
3235 (ira_swap_allocno_copy_ends_if_necessary): Make static. Rename to
3236 swap_allocno_copy_ends_if_necessary.
3237 (remove_unnecessary_allocnos, remove_low_level_allocnos): Call
3238 ira_remove_allocno_prefs.
3239 (ira_flattening): Ditto.
3240 (ira_build): Call initiate_prefs, print_prefs.
3241 (ira_destroy): Call finish_prefs.
3242 * ira-color.c (struct update_cost_record): New.
3243 (struct allocno_color_data): Add new member update_cost_records.
3244 (update_cost_record_pool): New static var.
3245 (init_update_cost_records, get_update_cost_record): New functions.
3246 (free_update_cost_record_list, finish_update_cost_records): Ditto.
3247 (struct update_cost_queue_elem): Add member from.
3248 (initiate_cost_update): Call init_update_cost_records.
3249 (finish_cost_update): Call finish_update_cost_records.
3250 (queue_update_cost, get_next_update_cost): Add new param from.
3251 (Update_allocno_cost, update_costs_from_allocno): New functions.
3252 (update_costs_from_prefs): Ditto.
3253 (update_copy_costs): Rename to update_costs_from_copies.
3254 (restore_costs_from_copies): New function.
3255 (update_conflict_hard_regno_costs): Don't go back.
3256 (assign_hard_reg): Call restore_costs_from_copies. Add printing
3258 (pop_allocnos): Add priniting more debug info.
3259 (color_allocnos): Remove prefs for conflicting hard regs.
3260 Call update_costs_from_prefs.
3261 * ira-conflicts.c (commutative_constraint_p): Move to ira.c
3262 (get_dup_num): Rename, modify, and move to ira.c
3263 (process_regs_for_copy): Add prefs.
3264 (add_insn_allocno_copies): Put src as first arg of
3265 process_regs_for_copy. Remove dead code. Call ira_setup_alts.
3266 * ira-costs.c (record_reg_classes): Modify and move code into
3267 record_operands_costs.
3268 (find_costs_and_classes): Create prefs for the hard reg of small
3270 (process_bb_node_for_hard_reg_moves): Add prefs.
3272 2013-10-30 Richard Biener <rguenther@suse.de>
3275 * basic-block.h (pre_and_rev_post_order_compute_fn): New function.
3276 * cfganal.c (pre_and_rev_post_order_compute_fn): New function
3278 (pre_and_rev_post_order_compute): ... which now wraps it.
3279 * graph.c (draw_cfg_nodes_no_loops): Use
3280 pre_and_rev_post_order_compute_fn to avoid ICEing and dependence
3283 2013-10-30 Christian Bruel <christian.bruel@st.com>
3285 * config/sh/sh-mem.cc (sh_expand_cmpnstr): New function.
3286 (sh_expand_cmpstr): Handle known align and schedule improvements.
3287 * config/sh/sh-protos.h (sh_expand_cmpstrn): Declare.
3288 * config/sh/sh.md (cmpstrnsi): New pattern.
3290 2013-10-30 Martin Jambor <mjambor@suse.cz>
3292 PR rtl-optimization/10474
3293 * ira.c (find_moveable_pseudos): Do not calculate dominance info
3295 (interesting_dest_for_shprep): New function.
3296 (split_live_ranges_for_shrink_wrap): Likewise.
3297 (ira): Calculate dominance info and df analysis. Call
3298 split_live_ranges_for_shrink_wrap.
3300 2013-10-30 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3303 * config/arm/arm.c (arm_expand_epilogue_apcs_frame): Emit blockage.
3305 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
3307 * tree-core.h (tree_index): Add TI_POINTER_BOUNDS_TYPE.
3308 * tree.h (POINTER_BOUNDS_P): New.
3309 (BOUNDED_TYPE_P): New.
3311 (pointer_bounds_type_node): New.
3312 * tree.c (build_common_tree_nodes): Initialize
3313 pointer_bounds_type_node.
3314 * gimple.h (gimple_call_get_nobnd_arg_index): New.
3315 (gimple_call_num_nobnd_args): New.
3316 (gimple_call_nobnd_arg): New.
3317 (gimple_return_retbnd): New.
3318 (gimple_return_set_retbnd): New
3319 * gimple.c (gimple_build_return): Increase number of ops
3320 for return statement.
3321 (gimple_call_get_nobnd_arg_index): New.
3322 * gimple-pretty-print.c (dump_gimple_return): Print second op.
3324 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
3326 * ipa.c (cgraph_build_static_cdtor_1): Support contructors
3327 with "chkp ctor" and "bnd_legacy" attributes.
3328 * gimplify.c (gimplify_init_constructor): Avoid infinite
3329 loop during gimplification of bounds initializer.
3331 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
3333 * c-family/c-common.c (handle_bnd_variable_size_attribute): New.
3334 (handle_bnd_legacy): New.
3335 (c_common_attribute_table): Add bnd_variable_size and bnd_legacy.
3336 * doc/extend.texi: Document bnd_variable_size and bnd_legacy
3339 2013-10-29 Ilya Enkovich <ilya.enkovich@intel.com>
3341 * builtin-types.def (BT_FN_VOID_CONST_PTR): New.
3342 (BT_FN_PTR_CONST_PTR): New.
3343 (BT_FN_CONST_PTR_CONST_PTR): New.
3344 (BT_FN_PTR_CONST_PTR_SIZE): New.
3345 (BT_FN_PTR_CONST_PTR_CONST_PTR): New.
3346 (BT_FN_VOID_PTRPTR_CONST_PTR): New.
3347 (BT_FN_VOID_CONST_PTR_SIZE): New.
3348 (BT_FN_PTR_CONST_PTR_CONST_PTR_SIZE): New.
3349 * chkp-builtins.def: New.
3350 * builtins.def: include chkp-builtins.def.
3351 (DEF_CHKP_BUILTIN): New.
3352 * builtins.c (expand_builtin): Support BUILT_IN_CHKP_INIT_PTR_BOUNDS,
3353 BUILT_IN_CHKP_NULL_PTR_BOUNDS, BUILT_IN_CHKP_COPY_PTR_BOUNDS,
3354 BUILT_IN_CHKP_CHECK_PTR_LBOUNDS, BUILT_IN_CHKP_CHECK_PTR_UBOUNDS,
3355 BUILT_IN_CHKP_CHECK_PTR_BOUNDS, BUILT_IN_CHKP_SET_PTR_BOUNDS,
3356 BUILT_IN_CHKP_NARROW_PTR_BOUNDS, BUILT_IN_CHKP_STORE_PTR_BOUNDS,
3357 BUILT_IN_CHKP_GET_PTR_LBOUND, BUILT_IN_CHKP_GET_PTR_UBOUND,
3358 BUILT_IN_CHKP_BNDMK, BUILT_IN_CHKP_BNDSTX, BUILT_IN_CHKP_BNDCL,
3359 BUILT_IN_CHKP_BNDCU, BUILT_IN_CHKP_BNDLDX, BUILT_IN_CHKP_BNDRET,
3360 BUILT_IN_CHKP_INTERSECT, BUILT_IN_CHKP_ARG_BND, BUILT_IN_CHKP_NARROW,
3361 BUILT_IN_CHKP_EXTRACT_LOWER, BUILT_IN_CHKP_EXTRACT_UPPER.
3362 * common.opt (fcheck-pointer-bounds): New.
3363 * toplev.c (process_options): Check Pointer Bounds Checker is
3365 * doc/extend.texi: Document Pointer Bounds Checker built-in functions.
3367 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
3369 * target.def (builtin_chkp_function): New.
3370 (chkp_bound_type): New.
3371 (chkp_bound_mode): New.
3372 (fn_abi_va_list_bounds_size): New.
3373 (load_bounds_for_arg): New.
3374 (store_bounds_for_arg): New.
3375 * targhooks.h (default_load_bounds_for_arg): New.
3376 (default_store_bounds_for_arg): New.
3377 (default_fn_abi_va_list_bounds_size): New.
3378 (default_chkp_bound_type): New.
3379 (default_chkp_bound_mode): New.
3380 (default_builtin_chkp_function): New.
3381 * targhooks.c (default_load_bounds_for_arg): New.
3382 (default_store_bounds_for_arg): New.
3383 (default_fn_abi_va_list_bounds_size): New.
3384 (default_chkp_bound_type): New.
3385 (default_chkp_bound_mode); New.
3386 (default_builtin_chkp_function): New.
3387 * doc/tm.texi.in (TARGET_FN_ABI_VA_LIST_BOUNDS_SIZE): New.
3388 (TARGET_LOAD_BOUNDS_FOR_ARG): New.
3389 (TARGET_STORE_BOUNDS_FOR_ARG): New.
3390 (TARGET_BUILTIN_CHKP_FUNCTION): New.
3391 (TARGET_CHKP_BOUND_TYPE): New.
3392 (TARGET_CHKP_BOUND_MODE): New.
3393 * doc/tm.texi: Regenerated.
3394 * langhooks.h (lang_hooks): Add chkp_supported field.
3395 * langhooks-def.h (LANG_HOOKS_CHKP_SUPPORTED): New.
3396 (LANG_HOOKS_INITIALIZER); Add LANG_HOOKS_CHKP_SUPPORTED.
3398 2013-10-29 Andrew Pinski <apinski@cavium.com>
3400 * tree-ssa-ifcombine.c: Include rtl.h and tm_p.h.
3401 (ifcombine_ifandif): Handle cases where maybe_fold_and_comparisons
3402 fails, combining the branches anyways.
3403 (tree_ssa_ifcombine): Inverse the order of the basic block walk,
3404 increases the number of combinings.
3405 * gimple.h (gsi_start_nondebug_after_labels_bb): New function.
3407 2013-10-29 Mike Stump <mikestump@comcast.net>
3409 * machmode.def (PARTIAL_INT_MODE): Add precision and name.
3410 * genmodes.c (PARTIAL_INT_MODE): Add precision and name.
3411 (make_vector_mode): Increase namebuf to 16.
3412 (emit_insn_modes_h): When processing BImode, don't
3413 also match partial int modes.
3414 (emit_class_narrowest_mode): Likewise.
3416 * config/bfin/bfin-modes.def: Add precision to PDI.
3417 * config/m32c/m32c-modes.def: Add precision to PSI.
3418 * config/msp430/msp430-modes.def: Add precision to PSI.
3419 * config/rs6000/rs6000-modes.def: Add precision to PTI.
3420 * config/sh/sh-modes.def: Add precision to PSI and PDI.
3422 2013-10-29 Oleg Endo <olegendo@gcc.gnu.org>
3425 * config/sh/sh.md (*addc): Rename existing variations to ...
3426 (*addc_r_r_1, *addc_2r_1, *addc_r_1): ... these.
3427 (*addc_r_lsb, *addc_r_r_lsb, *addc_r_lsb_r, *addc_2r_lsb, *addc_r_msb,
3428 *addc_r_r_msb, *addc_2r_msb): New insn_and_split patterns.
3429 * config/sh/sh.c (addsubcosts): Handle some addc special cases.
3431 2013-10-29 Teresa Johnson <tejohnson@google.com>
3434 * tree-ssa-tail-merge.c (replace_block_by): Tolerate profile
3435 insanities when updating probabilities.
3437 2013-10-29 David Malcolm <dmalcolm@redhat.com>
3439 * gdbhooks.py (CGraphNodePrinter.to_string): Update gdb
3440 prettyprinter for cgraph_node to reflect the conversion of the
3441 symtable types to a C++ class hierarchy: it now *is* a
3442 symtab_node_base, rather than having one (named "symbol").
3444 2013-10-29 Balaji V. Iyer <balaji.v.iyer@intel.com>
3446 * builtins.c (is_builtin_name): Added a check for __cilkrts_detach and
3447 __cilkrts_pop_frame. If matched, then return true for built-in
3449 (expand_builtin): Added BUILT_IN_CILK_DETACH and
3450 BUILT_IN_CILK_POP_FRAME case.
3451 * langhooks-def.h (lhd_install_body_with_frame_cleanup): New prototype.
3452 (lhs_cilk_detect_spawn): Likewise.
3453 (LANG_HOOKS_DECLS): Added LANG_HOOKS_CILKPLUS.
3454 (LANG_HOOKS_CILKPLUS_DETECT_SPAWN_AND_UNWRAP): New #define.
3455 (LANG_HOOKS_CILKPLUS_FRAME_CLEANUP): Likewise.
3456 (LANG_HOOKS_CILKPLUS_GIMPLIFY_SPAWN): Likewise.
3457 (LANG_HOOKS_CILKPLUS): Likewise.
3458 * tree.h (CILK_SPAWN_FN): Likewise.
3459 * builtin.def (DEF_CILK_BUILTIN_STUB): Likewise.
3460 * Makefile.in (C_COMMON_OBJS): Added c-family/cilk.o.
3461 (OBJS): Added cilk-common.o.
3462 (BUILTINS_DEF): Added cilk-builtins.def.
3463 * langhooks.c (lhd_install_body_with_frame_cleanup): New function.
3464 (lhd_cilk_detect_spawn): Likewise.
3465 * langhooks.h (lang_hooks_for_cilkplus): New struct.
3466 (struct lang_hooks): Added new field called "cilkplus."
3467 * cilk-common.c: New file.
3469 * cilk-builtins.def: Likewise.
3470 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Added
3471 "__cilk" macro and set it to 200.
3472 * function.h (struct function::cilk_frame_decl): New field.
3473 (struct function::is_cilk_function): Likewise.
3474 (struct function::calls_cilk_spawn): Likewise.
3475 * gimplify.c (gimplify_call_expr): Added a check if the function call
3476 being gimplified is a spawn detach point. If so, then add pop_frame
3477 and detach function calls.
3478 (gimplify_expr): Added a CILK_SPAWN_STMT and CILK_SYNC_STMT case
3479 for gimplifying _Cilk_spawn and _Cilk_sync statements.
3480 (gimplify_return_expr): Added a check for _Cilk_spawn usage in
3481 function. If so, added a _Cilk_sync and gimplified it.
3482 (gimplify_modify_expr): Added a check for _Cilk_spawn in MODIFY and
3483 INIT_EXPRs. If so, then call gimplify_cilk_spawn.
3484 * ipa-inline-analysis (initialize_inline_failed): Prevent inlining of
3486 (can_inline_edge_p): Prevent inling of spawnee function.
3487 * ira.c (ira_setup_eliminable_regset): Force usage of frame pointer
3488 for functions that use Cilk keywords.
3489 * tree-inline.h (struct copy_body_data::remap_var_for_cilk): New field.
3490 * tree-pretty-print.c (dump_generic_node): Added CILK_SPAWN_STMT and
3491 CILK_SYNC_STMT cases.
3492 * tree.def (DEFTREECODE): Added CILK_SPAWN_STMT and CILK_SYNC_STMT
3494 * generic.texi (CILK_SPAWN_STMT): Added documentation for _Cilk_spawn.
3495 (CILK_SYNC_STMT): Added documentation for _Cilk_sync.
3496 * passes.texi (Cilk Keywords): New section that describes the compiler
3497 code changes for handling Cilk Keywords.
3499 2013-10-29 David Malcolm <dmalcolm@redhat.com>
3501 Patch autogenerated by refactor_symtab.py from
3502 https://github.com/davidmalcolm/gcc-refactoring-scripts
3503 revision 58bb219cc090b2f4516a9297d868c245495ee622
3505 * asan.c (asan_finish_file): Update for conversion of symtab types to
3506 a true class hierarchy.
3507 * cfgexpand.c (estimated_stack_frame_size): Likewise.
3508 * cgraph.c (cgraph_get_body): Likewise.
3509 (cgraph_get_create_real_symbol_node): Likewise.
3510 (verify_cgraph_node): Likewise.
3511 (verify_edge_corresponds_to_fndecl): Likewise.
3512 (verify_edge_count_and_frequency): Likewise.
3513 (cgraph_will_be_removed_from_program_if_no_direct_calls): Likewise.
3514 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
3515 (cgraph_can_remove_if_no_direct_calls_and_refs_p): Likewise.
3516 (cgraph_node_cannot_return): Likewise.
3517 (cgraph_set_pure_flag_1): Likewise.
3518 (cgraph_set_const_flag_1): Likewise.
3519 (cgraph_set_nothrow_flag_1): Likewise.
3520 (cgraph_make_node_local_1): Likewise.
3521 (cgraph_for_node_and_aliases): Likewise.
3522 (cgraph_for_node_thunks_and_aliases): Likewise.
3523 (cgraph_node_can_be_local_p): Likewise.
3524 (cgraph_node_cannot_be_local_p_1): Likewise.
3525 (cgraph_function_body_availability): Likewise.
3526 (dump_cgraph_node): Likewise.
3527 (cgraph_rtl_info): Likewise.
3528 (cgraph_mark_address_taken_node): Likewise.
3529 (cgraph_remove_node): Likewise.
3530 (cgraph_release_function_body): Likewise.
3531 (cgraph_update_edges_for_call_stmt_node): Likewise.
3532 (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
3533 (cgraph_make_edge_direct): Likewise.
3534 (cgraph_resolve_speculation): Likewise.
3535 (cgraph_speculative_call_info): Likewise.
3536 (cgraph_turn_edge_to_speculative): Likewise.
3537 (cgraph_create_edge_1): Likewise.
3538 (cgraph_set_call_stmt): Likewise.
3539 (cgraph_node_for_asm): Likewise.
3540 (cgraph_add_thunk): Likewise.
3541 (cgraph_same_body_alias): Likewise.
3542 (cgraph_create_function_alias): Likewise.
3543 (cgraph_create_node): Likewise.
3544 (cgraph_create_empty_node): Likewise.
3545 (record_function_versions): Likewise.
3546 (used_from_object_file_p): Likewise.
3547 * cgraph.h (symtab_can_be_discarded): Likewise.
3548 (symtab_real_symbol_p): Likewise.
3549 (cgraph_mark_force_output_node): Likewise.
3550 (cgraph_edge_recursive_p): Likewise.
3551 (symtab_alias_target): Likewise.
3552 (varpool_all_refs_explicit_p): Likewise.
3553 (varpool_can_remove_if_no_refs): Likewise.
3554 (cgraph_only_called_directly_or_aliased_p): Likewise.
3555 (cgraph_next_function_with_gimple_body): Likewise.
3556 (cgraph_first_function_with_gimple_body): Likewise.
3557 (cgraph_function_with_gimple_body_p): Likewise.
3558 (cgraph_next_function): Likewise.
3559 (cgraph_first_function): Likewise.
3560 (cgraph_next_defined_function): Likewise.
3561 (cgraph_first_defined_function): Likewise.
3562 (varpool_next_defined_variable): Likewise.
3563 (varpool_first_defined_variable): Likewise.
3564 (varpool_next_static_initializer): Likewise.
3565 (varpool_first_static_initializer): Likewise.
3566 (varpool_next_variable): Likewise.
3567 (varpool_first_variable): Likewise.
3568 (varpool_node_name): Likewise.
3569 (varpool): Likewise.
3571 (is_a_helper <varpool_node>::test): Likewise.
3572 (is_a_helper <cgraph_node>::test): Likewise.
3573 (varpool_variable_node): Likewise.
3574 (cgraph_function_or_thunk_node): Likewise.
3575 (varpool_alias_target): Likewise.
3576 (cgraph_alias_target): Likewise.
3577 (cgraph_node_name): Likewise.
3578 (varpool_node_asm_name): Likewise.
3579 (cgraph_node_asm_name): Likewise.
3580 * cgraphbuild.c (remove_cgraph_callee_edges): Likewise.
3581 (cgraph_rebuild_references): Likewise.
3582 (rebuild_cgraph_edges): Likewise.
3583 (record_eh_tables): Likewise.
3584 (build_cgraph_edges): Likewise.
3585 (mark_store): Likewise.
3586 (mark_load): Likewise.
3587 (mark_address): Likewise.
3588 (record_type_list): Likewise.
3589 (record_reference): Likewise.
3590 * cgraphclones.c (cgraph_materialize_all_clones): Likewise.
3591 (cgraph_materialize_clone): Likewise.
3592 (cgraph_function_versioning): Likewise.
3593 (cgraph_copy_node_for_versioning): Likewise.
3594 (update_call_expr): Likewise.
3595 (cgraph_find_replacement_node): Likewise.
3596 (cgraph_create_virtual_clone): Likewise.
3597 (cgraph_clone_node): Likewise.
3598 * cgraphunit.c (compile): Likewise.
3599 (output_weakrefs): Likewise.
3600 (output_in_order): Likewise.
3601 (expand_function): Likewise.
3602 (assemble_thunks_and_aliases): Likewise.
3603 (expand_thunk): Likewise.
3604 (mark_functions_to_output): Likewise.
3605 (handle_alias_pairs): Likewise.
3606 (analyze_functions): Likewise.
3607 (walk_polymorphic_call_targets): Likewise.
3608 (varpool_finalize_decl): Likewise.
3609 (process_function_and_variable_attributes): Likewise.
3610 (cgraph_process_same_body_aliases): Likewise.
3611 (analyze_function): Likewise.
3612 (cgraph_add_new_function): Likewise.
3613 (cgraph_finalize_function): Likewise.
3614 (referred_to_p): Likewise.
3615 (cgraph_reset_node): Likewise.
3616 (cgraph_process_new_functions): Likewise.
3617 (enqueue_node): Likewise.
3618 (decide_is_symbol_needed): Likewise.
3619 * coverage.c (coverage_compute_profile_id): Likewise.
3620 * dbxout.c (dbxout_expand_expr): Likewise.
3621 * dwarf2out.c (premark_types_used_by_global_vars_helper): Likewise.
3622 (reference_to_unused): Likewise.
3623 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
3624 * gimplify.c (unvisit_body): Likewise.
3625 (unshare_body): Likewise.
3626 * ipa-cp.c (ipcp_generate_summary): Likewise.
3627 (ipcp_decision_stage): Likewise.
3628 (identify_dead_nodes): Likewise.
3629 (decide_whether_version_node): Likewise.
3630 (decide_about_value): Likewise.
3631 (perhaps_add_new_callers): Likewise.
3632 (create_specialized_node): Likewise.
3633 (update_profiling_info): Likewise.
3634 (ipcp_propagate_stage): Likewise.
3635 (estimate_local_effects): Likewise.
3636 (good_cloning_opportunity_p): Likewise.
3637 (devirtualization_time_bonus): Likewise.
3638 (propagate_constants_accross_call): Likewise.
3639 (initialize_node_lattices): Likewise.
3640 (ipcp_cloning_candidate_p): Likewise.
3641 (determine_versionability): Likewise.
3642 (print_all_lattices): Likewise.
3643 (print_lattice): Likewise.
3644 (ipcp_discover_new_direct_edges): Likewise.
3645 * ipa-devirt.c (ipa_devirt): Likewise.
3646 (likely_target_p): Likewise.
3647 (update_type_inheritance_graph): Likewise.
3648 (possible_polymorphic_call_target_p): Likewise.
3649 (dump_possible_polymorphic_call_targets): Likewise.
3650 (devirt_variable_node_removal_hook): Likewise.
3651 (record_binfo): Likewise.
3652 (maybe_record_node): Likewise.
3653 (build_type_inheritance_graph): Likewise.
3654 * ipa-inline-analysis.c (inline_write_summary): Likewise.
3655 (inline_generate_summary): Likewise.
3656 (inline_analyze_function): Likewise.
3657 (do_estimate_growth): Likewise.
3658 (simple_edge_hints): Likewise.
3659 (estimate_node_size_and_time): Likewise.
3660 (estimate_edge_devirt_benefit): Likewise.
3661 (compute_inline_parameters): Likewise.
3662 (estimate_function_body_sizes): Likewise.
3663 (compute_bb_predicates): Likewise.
3664 (initialize_inline_failed): Likewise.
3665 (dump_inline_summary): Likewise.
3666 (dump_inline_edge_summary): Likewise.
3667 * ipa-inline-transform.c (inline_transform): Likewise.
3668 (preserve_function_body_p): Likewise.
3669 (save_inline_function_body): Likewise.
3670 (inline_call): Likewise.
3671 (clone_inlined_nodes): Likewise.
3672 (can_remove_node_now_p): Likewise.
3673 (can_remove_node_now_p_1): Likewise.
3674 * ipa-inline.c (early_inliner): Likewise.
3675 (early_inline_small_functions): Likewise.
3676 (inline_always_inline_functions): Likewise.
3677 (ipa_inline): Likewise.
3678 (flatten_function): Likewise.
3679 (inline_small_functions): Likewise.
3680 (speculation_useful_p): Likewise.
3681 (recursive_inlining): Likewise.
3682 (update_caller_keys): Likewise.
3683 (reset_edge_caches): Likewise.
3684 (update_edge_key): Likewise.
3685 (edge_badness): Likewise.
3686 (relative_time_benefit): Likewise.
3687 (want_inline_self_recursive_call_p): Likewise.
3688 (want_inline_small_function_p): Likewise.
3689 (want_early_inline_function_p): Likewise.
3690 (num_calls): Likewise.
3691 (can_early_inline_edge_p): Likewise.
3692 (can_inline_edge_p): Likewise.
3693 (report_inline_failed_reason): Likewise.
3694 * ipa-profile.c (ipa_profile): Likewise.
3695 (ipa_propagate_frequency): Likewise.
3696 (ipa_propagate_frequency_1): Likewise.
3697 (ipa_profile_generate_summary): Likewise.
3698 * ipa-prop.c (ipcp_transform_function): Likewise.
3699 (read_replacements_section): Likewise.
3700 (ipa_prop_read_section): Likewise.
3701 (ipa_modify_call_arguments): Likewise.
3702 (ipa_print_node_params): Likewise.
3703 (propagate_controlled_uses): Likewise.
3704 (update_indirect_edges_after_inlining): Likewise.
3705 (remove_described_reference): Likewise.
3706 (ipa_make_edge_direct_to_target): Likewise.
3707 (ipa_analyze_node): Likewise.
3708 (ipa_analyze_params_uses): Likewise.
3709 (ipa_compute_jump_functions): Likewise.
3710 (ipa_get_callee_param_type): Likewise.
3711 (ipa_print_node_jump_functions): Likewise.
3712 (ipa_initialize_node_params): Likewise.
3713 (ipa_populate_param_decls): Likewise.
3714 (ipa_func_spec_opts_forbid_analysis_p): Likewise.
3715 (write_agg_replacement_chain): Likewise.
3716 (ipa_write_node_info): Likewise.
3717 (ipa_edge_duplication_hook): Likewise.
3718 (try_decrement_rdesc_refcount): Likewise.
3719 * ipa-pure-const.c (propagate_nothrow): Likewise.
3720 (propagate_pure_const): Likewise.
3721 (pure_const_read_summary): Likewise.
3722 (pure_const_write_summary): Likewise.
3723 (analyze_function): Likewise.
3724 * ipa-ref-inline.h (ipa_ref_referred_ref_list): Likewise.
3725 (ipa_ref_referring_ref_list): Likewise.
3726 * ipa-ref.c (ipa_clear_stmts_in_references): Likewise.
3727 (ipa_remove_stmt_references): Likewise.
3728 (ipa_find_reference): Likewise.
3729 (ipa_dump_referring): Likewise.
3730 (ipa_dump_references): Likewise.
3731 (ipa_record_reference): Likewise.
3732 * ipa-reference.c (ipa_reference_read_optimization_summary): Likewise.
3733 (ipa_reference_write_optimization_summary): Likewise.
3734 (write_node_summary_p): Likewise.
3735 (propagate): Likewise.
3736 (read_write_all_from_decl): Likewise.
3737 (generate_summary): Likewise.
3738 (analyze_function): Likewise.
3739 (propagate_bits): Likewise.
3740 (ipa_reference_get_not_written_global): Likewise.
3741 (ipa_reference_get_not_read_global): Likewise.
3742 * ipa-split.c (execute_split_functions): Likewise.
3743 (split_function): Likewise.
3744 * ipa-utils.c (ipa_merge_profiles): Likewise.
3745 (dump_cgraph_node_set): Likewise.
3746 (ipa_reverse_postorder): Likewise.
3747 (ipa_edge_within_scc): Likewise.
3748 (ipa_get_nodes_in_cycle): Likewise.
3749 (ipa_free_postorder_info): Likewise.
3750 (ipa_reduced_postorder): Likewise.
3751 (searchc): Likewise.
3752 (recursive_call_p): Likewise.
3753 * ipa.c (ipa_cdtor_merge): Likewise.
3754 (record_cdtor_fn): Likewise.
3755 (function_and_variable_visibility): Likewise.
3756 (varpool_externally_visible_p): Likewise.
3757 (cgraph_externally_visible_p): Likewise.
3758 (comdat_can_be_unshared_p): Likewise.
3759 (comdat_can_be_unshared_p_1): Likewise.
3760 (address_taken_from_non_vtable_p): Likewise.
3761 (ipa_discover_readonly_nonaddressable_vars): Likewise.
3762 (symtab_remove_unreachable_nodes): Likewise.
3763 (walk_polymorphic_call_targets): Likewise.
3764 (process_references): Likewise.
3765 (enqueue_node): Likewise.
3766 (has_addr_references_p): Likewise.
3767 (cgraph_non_local_node_p_1): Likewise.
3768 * is-a.h (varpool_analyze_node): Likewise.
3769 * lto-cgraph.c (input_symtab): Likewise.
3770 (merge_profile_summaries): Likewise.
3771 (input_cgraph_1): Likewise.
3772 (input_edge): Likewise.
3773 (input_varpool_node): Likewise.
3774 (input_node): Likewise.
3775 (input_overwrite_node): Likewise.
3776 (compute_ltrans_boundary): Likewise.
3777 (output_refs): Likewise.
3778 (lto_output_varpool_node): Likewise.
3779 (lto_output_node): Likewise.
3780 (reachable_from_other_partition_p): Likewise.
3781 (referenced_from_other_partition_p): Likewise.
3782 (lto_output_edge): Likewise.
3783 (output_node_opt_summary): Likewise.
3784 (add_node_to): Likewise.
3785 (reachable_from_this_partition_p): Likewise.
3786 (lto_set_symtab_encoder_in_partition): Likewise.
3787 (lto_symtab_encoder_in_partition_p): Likewise.
3788 (lto_set_symtab_encoder_encode_initializer): Likewise.
3789 (lto_symtab_encoder_encode_initializer_p): Likewise.
3790 (lto_set_symtab_encoder_encode_body): Likewise.
3791 (lto_symtab_encoder_encode_body_p): Likewise.
3792 * lto-section-in.c (lto_free_function_in_decl_state_for_node):
3794 * lto-streamer-in.c (lto_read_body): Likewise.
3795 (fixup_call_stmt_edges): Likewise.
3796 (fixup_call_stmt_edges_1): Likewise.
3797 * lto-streamer-out.c (produce_symtab): Likewise.
3798 (output_symbol_p): Likewise.
3799 (write_symbol): Likewise.
3800 (lto_output): Likewise.
3801 (copy_function): Likewise.
3802 (output_function): Likewise.
3803 * passes.c (function_called_by_processed_nodes_p): Likewise.
3804 (ipa_write_optimization_summaries): Likewise.
3805 (ipa_write_summaries): Likewise.
3806 (do_per_function_toporder): Likewise.
3807 (do_per_function): Likewise.
3808 (dump_passes): Likewise.
3809 * symtab.c (symtab_semantically_equivalent_p): Likewise.
3810 (symtab_nonoverwritable_alias): Likewise.
3811 (symtab_nonoverwritable_alias_1): Likewise.
3812 (symtab_for_node_and_aliases): Likewise.
3813 (symtab_resolve_alias): Likewise.
3814 (fixup_same_cpp_alias_visibility): Likewise.
3815 (symtab_alias_ultimate_target): Likewise.
3816 (symtab_used_from_object_file_p): Likewise.
3817 (verify_symtab_base): Likewise.
3818 (dump_symtab_base): Likewise.
3819 (symtab_node_name): Likewise.
3820 (symtab_node_asm_name): Likewise.
3821 (symtab_dissolve_same_comdat_group_list): Likewise.
3822 (symtab_add_to_same_comdat_group): Likewise.
3823 (symtab_unregister_node): Likewise.
3824 (symtab_insert_node_to_hashtable): Likewise.
3825 (symtab_register_node): Likewise.
3826 (unlink_from_assembler_name_hash): Likewise.
3827 (insert_to_assembler_name_hash): Likewise.
3828 (eq_assembler_name): Likewise.
3829 (hash_node_by_assembler_name): Likewise.
3830 (eq_node): Likewise.
3831 (hash_node): Likewise.
3832 * toplev.c (wrapup_global_declaration_2): Likewise.
3833 * trans-mem.c (ipa_tm_execute): Likewise.
3834 (ipa_tm_transform_clone): Likewise.
3835 (ipa_tm_transform_transaction): Likewise.
3836 (ipa_tm_transform_calls_redirect): Likewise.
3837 (ipa_tm_insert_gettmclone_call): Likewise.
3838 (ipa_tm_insert_irr_call): Likewise.
3839 (ipa_tm_create_version): Likewise.
3840 (ipa_tm_create_version_alias): Likewise.
3841 (ipa_tm_mark_forced_by_abi_node): Likewise.
3842 (ipa_tm_mark_force_output_node): Likewise.
3843 (ipa_tm_diagnose_tm_safe): Likewise.
3844 (ipa_tm_mayenterirr_function): Likewise.
3845 (ipa_tm_scan_irr_function): Likewise.
3846 (ipa_tm_note_irrevocable): Likewise.
3847 (ipa_tm_scan_calls_clone): Likewise.
3848 (get_cg_data): Likewise.
3849 * tree-eh.c (tree_could_trap_p): Likewise.
3850 * tree-emutls.c (ipa_lower_emutls): Likewise.
3851 (create_emultls_var): Likewise.
3852 (lower_emutls_function_body): Likewise.
3853 (gen_emutls_addr): Likewise.
3854 (emutls_decl): Likewise.
3855 (new_emutls_decl): Likewise.
3856 * tree-inline.c (tree_function_versioning): Likewise.
3857 (optimize_inline_calls): Likewise.
3858 (expand_call_inline): Likewise.
3859 (estimate_num_insns): Likewise.
3860 (copy_bb): Likewise.
3861 (delete_unreachable_blocks_update_callgraph): Likewise.
3862 * tree-nested.c (gimplify_all_functions): Likewise.
3863 (create_nesting_tree): Likewise.
3864 (check_for_nested_with_variably_modified): Likewise.
3865 * tree-pretty-print.c (dump_function_header): Likewise.
3866 * tree-profile.c (tree_profiling): Likewise.
3867 * tree-sra.c (ipa_sra_preliminary_function_checks): Likewise.
3868 (modify_function): Likewise.
3869 (convert_callers): Likewise.
3870 (convert_callers_for_node): Likewise.
3871 * tree-ssa-structalias.c (ipa_pta_execute): Likewise.
3872 (associate_varinfo_to_alias): Likewise.
3873 (create_variable_info_for): Likewise.
3874 (get_constraint_for_ssa_var): Likewise.
3875 * tree-vectorizer.c (increase_alignment): Likewise.
3876 * tree.c (find_decls_types_in_var): Likewise.
3877 (find_decls_types_in_node): Likewise.
3878 (free_lang_data_in_decl): Likewise.
3879 * value-prof.c (gimple_ic_transform): Likewise.
3880 (gimple_ic): Likewise.
3881 (check_ic_target): Likewise.
3882 (init_node_map): Likewise.
3883 * varasm.c (decl_binds_to_current_def_p): Likewise.
3884 (default_binds_local_p_1): Likewise.
3885 (dump_tm_clone_pairs): Likewise.
3886 (assemble_alias): Likewise.
3887 (find_decl): Likewise.
3888 (mark_decl_referenced): Likewise.
3889 * varpool.c (varpool_for_node_and_aliases): Likewise.
3890 (varpool_extra_name_alias): Likewise.
3891 (varpool_create_variable_alias): Likewise.
3892 (add_new_static_var): Likewise.
3893 (varpool_finalize_named_section_flags): Likewise.
3894 (varpool_remove_unreferenced_decls): Likewise.
3895 (enqueue_node): Likewise.
3896 (varpool_assemble_decl): Likewise.
3897 (assemble_aliases): Likewise.
3898 (varpool_analyze_node): Likewise.
3899 (cgraph_variable_initializer_availability): Likewise.
3900 (varpool_add_new_variable): Likewise.
3901 (ctor_for_folding): Likewise.
3902 (dump_varpool_node): Likewise.
3903 (varpool_remove_initializer): Likewise.
3904 (varpool_remove_node): Likewise.
3905 (varpool_node_for_decl): Likewise.
3906 (varpool_create_empty_node): Likewise.
3907 * config/i386/i386.c (ix86_generate_version_dispatcher_body): Likewise.
3908 (ix86_get_function_versions_dispatcher): Likewise.
3910 2013-10-29 David Malcolm <dmalcolm@redhat.com>
3912 * cgraph.h (symtab_node_base): Convert to a class;
3913 add GTY((desc ("%h.type"), tag ("SYMTAB_SYMBOL"))), and take
3914 chain_next/prev from symtab_node_def.
3915 (cgraph_node): Inherit from symtab_node; add GTY option
3916 tag ("SYMTAB_FUNCTION").
3917 (varpool_node): Inherit from symtab_node; add GTY option
3918 tag ("SYMTAB_VARIABLE").
3919 (symtab_node_def): Remove.
3920 (is_a_helper <cgraph_node>::test (symtab_node_def *)): Convert to...
3921 (is_a_helper <cgraph_node>::test (symtab_node_base *)): ...this.
3922 (is_a_helper <varpool_node>::test (symtab_node_def *)): Convert to...
3923 (is_a_helper <varpool_node>::test (symtab_node_base *)): ...this.
3925 * ipa-ref.h (symtab_node_def): Drop.
3926 (symtab_node): Change underlying type from symtab_node_def to
3928 (const_symtab_node): Likwise.
3930 * is-a.h: Update examples in comment.
3932 * symtab.c (symtab_hash): Change symtab_node_def to symtab_node_base.
3933 (assembler_name_hash): Likewise.
3935 2013-10-29 Martin Liska <marxin.liska@gmail.com>
3937 * doc/tree-ssa.texi (gimple_phi_result): Document.
3938 (gimple_phi_num_args, gimple_phi_arg): Likewise.
3939 (gimple_phi_arg_edge, gimple_phi_arg_def): Likewise.
3940 (PHI_RESULT, PHI_NUM_ARGS): Remove.
3941 (PHI_ARG_ELT, PHI_ARG_EDGE, PHI_ARG_DEF): Likewise.
3943 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
3945 * expr.h: Revert change and include tree-core.h.
3946 * rtl.h: Revert change and don't include tree-core.h.
3948 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
3950 * config/darwin.c: Include gimple.h.
3951 * config/i386/winnt.c: Likewise.
3953 2013-10-29 Marc Glisse <marc.glisse@inria.fr>
3955 PR tree-optimization/19831
3956 * tree-ssa-alias.c (stmt_kills_ref_p_1): Handle BUILT_IN_FREE.
3958 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
3960 * tree-outof-ssa.h: Remove include files.
3961 * tree-outof-ssa.c: Add required include files from tree-outof-ssa.h.
3963 * tree-ssa-coalesce.c: Likewise.
3964 * cfgexpand.c: Likewise.
3965 * tree-ssa-ter.c: Likewise.
3966 * ipa-prop.h: Remove gimple.h and tree-core.h from include list.
3967 * lto-streamer.h: Likewise.
3968 * cgraphbuild.c: Add gimple.h to include list.
3969 * data-streamer-in.c: Likewise.
3970 * ipa-cp.c: Likewise.
3971 * tree-streamer.c: Likewise.
3972 * lto-compress.c: Likewise.
3973 * ipa-reference.c: Likewise.
3974 * data-streamer-out.c: Likewise.
3975 * lto-cgraph.c: Likewise.
3976 * cgraphclones.c: Likewise.
3977 * ipa-utils.c: Likewise.
3978 * data-streamer.c: Likewise.
3979 * ipa-split.c: Likewise.
3980 * lto-section-in.c: Likewise.
3981 * tree-streamer-out.c: Likewise.
3982 * ipa-prop.c: Likewise.
3983 * tree-streamer-in.c: Likewise.
3984 * symtab.c: Likewise.
3985 * opts-global.c: Likewise.
3986 * lto-opts.c: Likewise.
3987 * lto-section-out.c: Likewise.
3988 * lto-streamer.c: Likewise.
3989 * rtl.h: Add tree-core.h to include list.
3990 * expr.h: Remove tree-core.h from include list.
3991 * gimple.h: Likewise.
3992 * ipa-utils.h: Likewise.
3993 * streamer-hooks.h: Likewise.
3994 * streamer-hooks.c: Include input.h.
3996 2013-10-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3998 * config/arm/arm.c (cortexa7_extra_costs): New table.
3999 (arm_cortex_a7_tune): New.
4000 * config/arm/arm-cores.def: Use cortex_a7 tuning for cortex-a7.
4002 2013-10-29 Eric Botcazou <ebotcazou@adacore.com>
4004 * expr.c (expand_expr_real_1) <MEM_EXPR>: Eliminate small redundancy.
4006 2013-10-29 David Malcolm <dmalcolm@redhat.com>
4008 * doc/gty.texi ("Inheritance and GTY"): Make it clear that
4009 to use autogenerated markers for a class-hierarchy, every class
4010 must have a GTY marker.
4011 * gengtype.h (struct type): Add linked list of subclasses to
4012 the "s" member of the union.
4013 (add_subclass): New decl.
4014 * gengtype-state.c (read_state_struct_type): Set up subclass
4016 * gengtype.c (get_ultimate_base_class): New.
4017 (add_subclass): New.
4018 (new_structure): Set up subclass linked list.
4019 (set_gc_used_type): Propagate usage information to subclasses.
4020 (output_mangled_typename): Use get_ultimate_base_class.
4021 (walk_subclasses): Use the subclass linked list, avoiding an
4022 O(N^2) when writing out all types.
4023 (walk_type): Issue an error if the base class is missing a tag,
4024 rather than generating bogus C code. Add a gcc_unreachable
4025 default case, in case people omit tags from concrete subclasses,
4026 or get the values wrong.
4027 (write_func_for_structure): Issue an error for subclasses for
4028 which the base doesn't have a "desc", since otherwise the
4029 autogenerated routines for the base would silently fail to visit
4030 any subclass fields.
4031 (write_root): Use get_ultimate_base_class, tweaking constness of
4032 tp to match that function's signature.
4034 2013-10-29 David Malcolm <dmalcolm@redhat.com>
4036 * doc/gty.texi (GTY Options): Add note about inheritance to
4037 description of desc and tag.
4038 (Inheritance and GTY): New.
4040 2013-10-29 David Malcolm <dmalcolm@redhat.com>
4042 * gengtype-parse.c (opts_have): Drop "static" so that
4043 we can use this from gengtype.c.
4044 * gengtype.c (set_gc_used_type): Mark any base class as used;
4045 update field traversal to visit inherited fields.
4046 (output_mangled_typename): Convert references to classes within
4047 an inheritance hierarchy to reference the ultimate base class,
4048 since only it will have gt_ functions.
4049 (get_string_option): New.
4050 (walk_subclasses): New.
4051 (walk_type): Treat GTY structs that have a "desc" as being the
4052 root of an inheritance hierarchy. Generate a switch on it
4053 within the marking function which walks all subclasses, adding
4054 cases for them via walk_subclasses. For subclasses, visit all
4055 fields of the type (including inherited ones).
4056 (write_func_for_structure): Don't write fns for subclasses, only
4057 for the ultimate base class within an inheritance hierarchy.
4058 Subclasses-marking will be handled by the base class marking functions.
4059 (write_types): Likewise.
4060 (write_local_func_for_structure): Likewise.
4061 (USED_BY_TYPED_GC_P): Emit allocators for subclasses that have
4062 a "tag" option (and are thus concrete subclasses).
4063 (write_root): Use the marker function for the ultimate base class.
4064 * gengtype.h (FOR_ALL_INHERITED_FIELDS): New.
4065 (opts_have): Add declaration.
4067 2013-10-28 Vladimir Makarov <vmakarov@redhat.com>
4069 * lra-spills.c (lra_final_code_change): Remove useless move insns
4070 originated from moves of pseudos.
4072 2013-10-28 Jeff Law <law@redhat.com>
4074 * lower-subreg.c (resolve_simple_move): Fix comment typo.
4076 2013-10-28 Trevor Saunders <tsaunders@mozilla.com>
4078 * df-scan.c (df_collection_rec): Adjust.
4079 (copy_defs): New constant.
4080 (copy_uses): Likewise.
4081 (copy_eq_uses): Likewise.
4082 (copy_mw): Likewise.
4083 (copy_all): Likewise.
4084 (df_insn_rescan): Adjust.
4085 (df_notes_rescan): Likewise.
4086 (df_swap_refs): Likewise.
4087 (df_sort_and_compress_refs): Likewise.
4088 (df_sort_and_compress_mws): Likewise.
4089 (df_install_refs): Likewise.
4090 (df_install_mws): Likewise.
4091 (df_refs_add_to_chains): Add flags parameter controlling which vectors
4093 (df_bb_refs_record): Adjust.
4094 (df_record_entry_block_defs): Likewise.
4095 (df_record_exit_block_defs): Likewise.
4096 (df_refs_verify): Likewise.
4097 (df_mws_verify): Likewise.
4098 (df_insn_refs_verify): Likewise.
4099 (df_bb_verify): Likewise.
4100 * ipa-pure-const.c (finish_state): Remove.
4101 (propagate): Adjust.
4102 * tree-data-ref.c tree-ssa-alias.c tree-ssa-loop-ivcanon.c
4103 tree-ssa-threadedge.c tree-vect-loop-manip.c tree-vect-slp.c
4104 var-tracking.c: Adjust.
4105 * vec.c (stack_vecs): Remove.
4106 (register_stack_vec): Likewise.
4107 (stack_vec_register_index): Likewise.
4108 (unregister_stack_vec): Likewise.
4109 * vec.h (struct va_stack): Remove.
4110 (struct vec<T, A, vl_ptr>): Specialize as
4111 struct vec<T, va_heap, vl_ptr> instead since va_heap is the only
4112 allocation strategy compatable with the vl_ptr layout.
4113 (struct vec<T, va_gc, vl_ptr>): Remove because it now gets an empty
4114 specialization anyway.
4115 (class stack_vec): New class.
4116 (vec_stack_alloc): Remove.
4117 (vec<T, va_heap, vl_ptr>::using_auto_storage): New method.
4119 2013-10-28 Alexander Ivchenko <alexander.ivchenko@intel.com>
4120 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4121 Sergey Lega <sergey.s.lega@intel.com>
4122 Anna Tikhonova <anna.tikhonova@intel.com>
4123 Ilya Tocar <ilya.tocar@intel.com>
4124 Andrey Turetskiy <andrey.turetskiy@intel.com>
4125 Ilya Verbin <ilya.verbin@intel.com>
4126 Kirill Yukhin <kirill.yukhin@intel.com>
4127 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4129 * config/i386/i386.md (prefetch): Allow TARGET_AVX512PF.
4130 (*prefetch_avx512pf_<mode>): New.
4131 * config/i386/sse.md (avx512f_vmcmp<mode>3): Ditto.
4132 (avx512f_maskcmp<mode>3): Ditto.
4133 (vashrv16si3): Ditto.
4135 2013-10-28 Alexander Ivchenko <alexander.ivchenko@intel.com>
4136 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4137 Sergey Lega <sergey.s.lega@intel.com>
4138 Anna Tikhonova <anna.tikhonova@intel.com>
4139 Ilya Tocar <ilya.tocar@intel.com>
4140 Andrey Turetskiy <andrey.turetskiy@intel.com>
4141 Ilya Verbin <ilya.verbin@intel.com>
4142 Kirill Yukhin <kirill.yukhin@intel.com>
4143 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4145 * config/i386/i386.md (any_truncate): New.
4146 (trunsuffix): Ditto.
4147 * config/i386/predicates.md (const_8_to_9_operand): New.
4148 (const_10_to_11_operand): Ditto.
4149 (const_12_to_13_operand): Ditto.
4150 (const_14_to_15_operand): Ditto.
4151 (const_16_to_19_operand): Ditto.
4152 (const_20_to_23_operand): Ditto.
4153 (const_24_to_27_operand): Ditto.
4154 (const_28_to_31_operand): Ditto.
4155 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_FIX_NOTRUNC.
4156 (cvtusi2<ssescalarmodesuffix>32): New.
4157 (cvtusi2<ssescalarmodesuffix>64): Ditto.
4158 (ufloatv16siv16sf2): Ditto.
4159 (avx512f_fix_notruncv16sfv16si): Ditto.
4160 (avx512f_ufix_notruncv16sfv16si): Ditto.
4161 (avx512f_vcvtss2usi): Ditto.
4162 (avx512f_vcvtss2usiq): Ditto.
4163 (avx512f_vcvttss2usi): Ditto.
4164 (avx512f_vcvttss2usiq): Ditto.
4165 (avx512f_vcvtsd2usi): Ditto.
4166 (avx512f_vcvtsd2usiq): Ditto.
4167 (avx512f_vcvttsd2usi): Ditto.
4168 (avx512f_vcvttsd2usiq): Ditto.
4169 (ufloatv8siv8df): Ditto.
4170 (avx512f_cvtdq2pd512_2): Ditto.
4171 (avx512f_cvtpd2dq512): Ditto.
4172 (avx512f_ufix_notruncv8dfv8si): Ditto.
4173 (avx512f_cvtpd2ps512): Ditto.
4174 (vec_unpacks_lo_v16sf): Ditto.
4175 (vec_unpacks_hi_v16sf): Ditto.
4176 (vec_unpacks_float_hi_v16si): Ditto.
4177 (vec_unpacks_float_lo_v16si): Ditto.
4178 (avx512f_unpckhps512): Ditto.
4179 (avx512f_unpcklps512): Ditto.
4180 (avx512f_movshdup512): Ditto.
4181 (avx512f_movsldup512): Ditto.
4182 (vec_extract_lo_v32hi): Ditto.
4183 (vec_extract_hi_v32hi): Ditto.
4184 (vec_extract_lo_v64qi): Ditto.
4185 (vec_extract_hi_v64qi): Ditto.
4186 (avx512f_unpckhpd512): Ditto.
4187 (avx512f_movddup512): Ditto.
4188 (avx512f_unpcklpd512): Ditto.
4189 (*avx512f_unpcklpd512): Ditto.
4190 (avx512f_shufps512_1): Ditto.
4191 (avx512f_shufpd512_1): Ditto.
4192 (avx512f_interleave_highv8di): Ditto.
4193 (avx512f_interleave_lowv8di): Ditto.
4194 (PMOV_DST_MODE): Ditto.
4195 (pmov_src_mode): Ditto.
4196 (pmov_src_lower): Ditto.
4198 (*avx512f_<code><pmov_src_lower><mode>2): Ditto.
4199 (*avx512f_<code>v8div16qi2): Ditto.
4200 (*avx512f_<code>v8div16qi2_store): Ditto.
4201 (vec_widen_umult_even_v16si): Ditto.
4202 (*vec_widen_umult_even_v16si): Ditto.
4203 (vec_widen_smult_even_v16si): Ditto.
4204 (*vec_widen_smult_even_v16si): Ditto.
4205 (avx512f_interleave_highv16si): Ditto.
4206 (avx512f_interleave_lowv16si): Ditto.
4207 (avx512f_<code>v16qiv16si2): Ditto.
4208 (avx512f_<code>v16hiv16si2): Ditto.
4209 (avx512f_<code>v8qiv8di2): Ditto.
4210 (avx512f_<code>v8hiv8di2): Ditto.
4211 (avx512f_<code>v8siv8di2): Ditto.
4212 (avx512cd_maskb_vec_dupv8di): Ditto.
4213 (avx512cd_maskw_vec_dupv16si): Ditto.
4214 (avx512f_vcvtph2ps512): Ditto.
4215 (avx512f_vcvtps2ph512): Ditto.
4216 (VEC_EXTRACT_MODE): Extened with wider modes.
4217 (VEC_PERM_AVX2): Ditto.
4218 (VEC_PERM_CONST): Ditto.
4220 2013-10-28 Joern Rennecke <joern.rennecke@embecosm.com>
4222 * config/arc/arc.c (arc_ccfsm_post_advance):
4223 Add comment about TYPE_RETURN.
4225 2013-10-28 Bin Cheng <bin.cheng@arm.com>
4227 * tree-ssa-loop-ivopts.c (strip_offset_1): Change parameter type.
4228 Count DECL_FIELD_BIT_OFFSET for COMPONENT_REF.
4229 (strip_offset): Convert offset to unsigned number.
4231 2013-10-27 Tom de Vries <tom@codesourcery.com>
4233 * cfgexpand.c (gimple_expand_cfg): Remove test for parm_birth_insn.
4234 Don't commit insertions after NOTE_INSN_FUNCTION_BEG.
4236 2013-10-27 Oleg Endo <olegendo@gcc.gnu.org>
4238 * config/sh/sh.c (MSW, LSW): Move and rename macros to...
4239 * config/sh/sh.h (SH_REG_MSW_OFFSET, SH_REG_LSW_OFFSET): ... here.
4240 (TARGET_BIG_ENDIAN): New macro.
4241 * config/sh/sh.md: Use it instead of !TARGET_LITTLE_ENDIAN.
4242 Use SH_REG_MSW_OFFSET and SH_REG_LSW_OFFSET.
4243 * config/sh/sh.c: Likewise.
4244 * config/sh/sh.h: Likewise.
4246 2013-10-27 Hans-Peter Nilsson <hp@axis.com>
4248 * config/cris/cris.c (cris_emit_trap_for_misalignment): Replace the
4249 removed PRED_MUDFLAP with PRED_NORETURN. Correct file-path in comment.
4251 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
4253 * config/sh/sh.md (movmemsi): Remove empty constraints and predicates.
4255 (cmpstr_t, cmpstrsi): Fix formatting.
4257 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
4260 * config/sh/predicates.md (general_movdst_operand): Allow reg+reg
4261 addressing, do not use general_operand for memory operands.
4263 2013-10-26 Vladimir Makarov <vmakarov@redhat.com>
4266 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
4267 * lra-spills.c (lra_final_code_change): Remove useless move insns.
4269 2013-10-26 Jeff Law <law@redhat.com>
4271 * predict.c (PRED_MUDFLAP): Remove.
4272 * targhooks.c (build_va_arg_indirect_ref): Remove mudflap support.
4274 * Makefile.in (C_COMMON_OBJS): Remove tree-mudflap.
4275 (OBJS): Remove tree-nomudflap.o
4276 (GTFILES): Remove tree-mudflap.c
4277 * builtins.c (expand_builtin_alloc): Remove mudflap support.
4278 * gcc.c (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
4279 (mfwrap_spec, mflib_spec): Likewise.
4280 (cpp_unique_options, cc1_options, static_specs): Likewise.
4281 * gimplify (gimplify_vla_decl, build_va_arg_indirect_ref): Likewise.
4282 * passes.def: Likewise.
4283 * toplev.c (compile_file, process_options): Likewise.
4284 * tree-inline.c (copy_tree_r): Likewise.
4285 * tree-pass.,h (make_pass_mudflap_1, make_pass_mudflap_2): Likewise.
4286 * varasm.c (make_decl_rtl, make_decl_rtl_for_debug): Likewise.
4287 (build_constant_desc, output_constant_def_contents): Likewise.
4288 (categorize_decl_for_section): Likewise.
4289 * tree-mudflap.c: Removed.
4290 * tree-mudflap.h: Removed.
4291 * tree-nomudflap.c: Removed.
4292 * bfin/uclinux.h (MFWRAP_SPEC): Remove.
4293 * moxie/uclinux.h (MFWRAP_SPEC): Likewise.
4294 * rs6000/aix.h (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
4295 * config/sol2.h (MFLIB_SPEC): Likewise.
4296 * doc/install.texi: Remove mudflap references.
4297 * doc/passes.texi: Similarly.
4298 * doc/sourcebuild.texi: Similarly.
4299 * doc/invoke.texi: Remove mudlfap related options.
4301 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
4303 PR rtl-optimization/58759
4304 * lra-constraints.c (lra_constraints): Remove wrong condition to
4305 remove insn setting up an equivalent pseudo.
4307 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
4309 * config/rs6000/rs6000-protos.h
4310 (rs6000_secondary_memory_needed_mode): New prototype.
4311 * config/rs6000/rs6000.c: Include ira.h.
4312 (TARGET_LRA_P): Redefine.
4313 (rs6000_legitimate_offset_address_p): Call
4314 legitimate_constant_pool_address_p in strict mode for LRA.
4315 (rs6000_legitimate_address_p): Ditto.
4316 (legitimate_lo_sum_address_p): Add code for LRA. Use lra_in_progress.
4317 (rs6000_emit_move): Add LRA version of code to generate load/store
4319 (rs6000_secondary_memory_needed_mode): New.
4320 (rs6000_alloc_sdmode_stack_slot): Do nothing for LRA.
4321 (rs6000_secondary_reload_class): Return NO_REGS for LRA for
4322 constants, memory, and FP registers.
4323 (rs6000_lra_p): New.
4324 * config/rs6000/rs6000.h (SECONDARY_MEMORY_NEEDED_MODE): New macro.
4325 * config/rs6000/rs6000.opt (mlra): New option.
4326 * lra-spills.c (lra_final_code_change): Remove useless move insns.
4328 2013-10-25 Yufeng Zhang <yufeng.zhang@arm.com>
4330 * tree-ssa-math-opts.c (convert_plusminus_to_widen): Call
4331 has_single_use () and not do the conversion if has_single_use ()
4332 returns false for the multiplication result.
4334 2013-10-25 David Malcolm <dmalcolm@redhat.com>
4336 * tree.h (EXCEPTIONAL_CLASS_P): Rename parameter from "CODE"
4337 to "NODE", since this works on a "tree", not an
4339 (CONSTANT_CLASS_P): Likewise.
4342 (INDIRECT_REF_P): Likewise.
4343 (REFERENCE_CLASS_P): Likewise.
4344 (COMPARISON_CLASS_P): Likewise.
4345 (UNARY_CLASS_P): Likewise.
4346 (BINARY_CLASS_P): Likewise.
4347 (STATEMENT_CLASS_P): Likewise.
4348 (VL_EXP_CLASS_P): Likewise.
4349 (EXPRESSION_CLASS_P): Likewise.
4350 (IS_TYPE_OR_DECL_P): Likewise.
4352 2013-10-25 Marc Glisse <marc.glisse@inria.fr>
4354 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Look for an
4355 ADDR_EXPR in the defining statement.
4357 2013-10-25 Richard Biener <rguenther@suse.de>
4359 PR tree-optimization/58626
4360 * tree-loop-distribution.c (enum rdg_dep_type): Remove
4361 anti_dd, output_dd and input_dd.
4362 (struct rdg_edge): Remove level and relation members.
4363 (RDGE_LEVEL, RDGE_RELATION): Remove.
4364 (dot_rdg_1): Adjust.
4365 (create_rdg_edge_for_ddr): Remove.
4366 (create_rdg_edges_for_scalar): Adjust.
4367 (create_edge_for_control_dependence): Likewise.
4368 (create_rdg_edges): Split into ...
4369 (create_rdg_flow_edges): ... this
4370 (create_rdg_cd_edges): ... and this.
4372 (build_rdg): Likewise, do not compute data dependences or
4374 (pg_add_dependence_edges): New function.
4376 (distribute_loop): First apply all non-dependence based
4377 partition mergings. Then compute dependences between partitions
4378 and merge and order partitions according to them.
4380 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
4382 PR rtl-optimization/58831
4383 * alias.c (init_alias_analysis): At the beginning of each iteration,
4384 set the reg_seen[N] bit if static_reg_base_value[N] is non-null.
4386 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
4388 * recog.c (search_ofs): New static variable moved from...
4389 (peep2_find_free_register): ...here.
4390 (peephole2_optimize): Initialize it.
4392 2013-10-25 Tobias Burnus <burnus@net-b.de>
4394 * doc/invoke.texi (fopenmp): Change supported OpenMP version to 4.0.
4396 2013-10-25 Uros Bizjak <ubizjak@gmail.com>
4398 * config/i386/i386.h (TARGET_MPX): New define.
4399 (TARGET_MPX_P): Ditto.
4401 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
4403 * config/i386/constraints.md (B): New.
4406 * config/i386/i386-c.c (ix86_target_macros_internal): Add __MPX__.
4407 * config/i386/i386-modes.def (BND32): New.
4409 * config/i386/i386-protos.h (ix86_bnd_prefixed_insn_p): New.
4410 * config/i386/i386.c (isa_opts): Add mmpx.
4411 (regclass_map): Add bound registers.
4412 (dbx_register_map): Likewise.
4413 (dbx64_register_map): Likewise.
4414 (svr4_dbx_register_map): Likewise.
4416 (ix86_option_override_internal): Support MPX ISA.
4417 (ix86_conditional_register_usage): Support bound registers.
4418 (print_reg): Likewise.
4419 (ix86_code_end): Add MPX bnd prefix.
4420 (output_set_got): Likewise.
4421 (ix86_output_call_insn): Likewise.
4422 (ix86_print_operand): Add '!' (MPX bnd) print prefix support.
4423 (ix86_print_operand_punct_valid_p): Likewise.
4424 (ix86_print_operand_address): Support UNSPEC_BNDMK_ADDR and
4426 (ix86_class_likely_spilled_p): Add bound regs support.
4427 (ix86_hard_regno_mode_ok): Likewise.
4428 (x86_order_regs_for_local_alloc): Likewise.
4429 (ix86_bnd_prefixed_insn_p): New.
4430 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Fix to new value.
4431 (FIXED_REGISTERS): Add bound registers.
4432 (CALL_USED_REGISTERS): Likewise.
4433 (REG_ALLOC_ORDER): Likewise.
4434 (HARD_REGNO_NREGS): Likewise.
4436 (VALID_BND_REG_MODE): New.
4437 (FIRST_BND_REG): New.
4438 (LAST_BND_REG): New.
4439 (reg_class): Add BND_REGS.
4440 (REG_CLASS_NAMES): Likewise.
4441 (REG_CLASS_CONTENTS): Likewise.
4443 (ANY_BND_REG_P): New.
4445 (HI_REGISTER_NAMES): Add bound registers.
4446 * config/i386/i386.md (UNSPEC_BNDMK): New.
4447 (UNSPEC_BNDMK_ADDR): New.
4448 (UNSPEC_BNDSTX): New.
4449 (UNSPEC_BNDLDX): New.
4450 (UNSPEC_BNDLDX_ADDR): New.
4451 (UNSPEC_BNDCL): New.
4452 (UNSPEC_BNDCU): New.
4453 (UNSPEC_BNDCN): New.
4454 (UNSPEC_MPX_FENCE): New.
4457 (type): Add mpxmov, mpxmk, mpxchk, mpxld, mpxst.
4458 (length_immediate): Likewise.
4459 (prefix_0f): Likewise.
4461 (prefix_rep): Check for bnd prefix.
4462 (length_nobnd): New.
4463 (length): Use length_nobnd if specified.
4468 (*jcc_1): Add bnd prefix and rename length attr to length_nobnd.
4471 (simple_return_internal): Likewise.
4472 (simple_return_pop_internal): Likewise.
4473 (*indirect_jump): Add MPX bnd prefix.
4474 (*tablejump_1): Likewise.
4475 (simple_return_internal_long): Likewise.
4476 (simple_return_indirect_internal): Likewise.
4480 (*mov<mode>_internal_mpx): New.
4481 (<mode>_<bndcheck>): New.
4482 (*<mode>_<bndcheck>): New.
4487 * config/i386/predicates.md (lea_address_operand): Rename to...
4488 (address_no_seg_operand): ... this.
4489 (address_mpx_no_base_operand): New.
4490 (address_mpx_no_index_operand): New.
4491 (bnd_mem_operator): New.
4492 * config/i386/i386.opt (mmpx): New.
4493 * doc/invoke.texi: Add documentation for the flags -mmpx, -mno-mpx.
4494 * doc/rtl.texi Add documentation for BND32mode and BND64mode.
4496 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
4498 * mode-classes.def (MODE_POINTER_BOUNDS): New.
4499 * tree.def (POINTER_BOUNDS_TYPE): New.
4500 * genmodes.c (complete_mode): Support MODE_POINTER_BOUNDS.
4501 (POINTER_BOUNDS_MODE): New.
4502 (make_pointer_bounds_mode): New.
4503 * machmode.h (POINTER_BOUNDS_MODE_P): New.
4504 * stor-layout.c (int_mode_for_mode): Support MODE_POINTER_BOUNDS.
4505 (layout_type): Support POINTER_BOUNDS_TYPE.
4506 * tree-pretty-print.c (dump_generic_node): Support POINTER_BOUNDS_TYPE.
4507 * tree.c (build_int_cst_wide): Support POINTER_BOUNDS_TYPE.
4508 (type_contains_placeholder_1): Likewise.
4509 * tree.h (POINTER_BOUNDS_TYPE_P): New.
4510 * varasm.c (output_constant): Support POINTER_BOUNDS_TYPE.
4511 * doc/rtl.texi (MODE_POINTER_BOUNDS): New.
4513 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
4515 * expr.c (expand_expr_real_1): Use mode of memory reference rather than
4516 mode of address computation when calling memory_address_addr_space.
4518 2013-08-24 Richard Henderson <rth@twiddle.net>
4521 * optabs.c (maybe_emit_atomic_exchange): Use create_input_operand
4522 instead of create_convert_operand_to.
4523 (maybe_emit_sync_lock_test_and_set): Likewise.
4524 (expand_atomic_compare_and_swap): Likewise.
4525 (maybe_emit_compare_and_swap_exchange_loop): Don't convert_modes.
4527 2013-08-24 Sriraman Tallam <tmsriram@google.com>
4529 * cgraph.c (cgraph_fnver_htab): Move GTY((...)) to be before htab_t.
4530 Change param_is to use the struct and not the pointer to the struct.
4532 2013-10-24 Andrew MacLeod <amacleod@redhat.com>
4534 * builtins.c (dummy_object, gimplify_va_arg_expr): Move to gimplify.c.
4535 * gimplify.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
4536 Move to targhooks.c.
4537 (dummy_object, gimplify_va_arg_expr): Relocate from builtins.c.
4538 * targhooks.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
4539 Relocate from gimplify.c.
4540 * targhooks.h: Add 2 prototypes.
4541 * tree.h: Delete 2 prototypes.
4543 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
4545 * tree-ssa-loop-ivopts.c (multiplier_allowed_in_address_p ): Check both
4546 [reg+mult*reg] and [mult*reg] to determine if multiplier is allowed.
4548 2013-10-24 Cong Hou <congh@google.com>
4550 * convert.c (convert_to_real): Guard those unsafe math function
4551 convertions with flag_unsafe_math_optimizations. Handle sqrt()
4554 2013-10-24 Markus Trippelsdorf <markus@trippelsdorf.de>
4557 * cgraph.c (cgraph_create_edge_1): Add indirect_unknown_callee
4559 (cgraph_create_edge): Use the new argument.
4560 (cgraph_create_indirect_edge): Likewise.
4562 2013-10-24 Joern Rennecke <joern.rennecke@embecosm.com>
4564 * config/arc/arc.c (arc_ccfsm_post_advance): Also handle
4566 (arc_ifcvt) <case 1 and 2>: Check that arc_ccfsm_post_advance
4567 changes statep->state.
4569 2013-10-24 Tobias Burnus <burnus@net-b.de>
4572 * tree-cfg.c (replace_loop_annotate): New function.
4573 (execute_build_cfg): Call it.
4574 * gimplify.c (gimple_boolify, gimplify_expr): Handle ANNOTATE_EXPR.
4575 * internal-fn.c (expand_ANNOTATE): New function.
4576 * internal-fn.def (ANNOTATE): Define as new internal function.
4577 * tree-core.h (tree_node_kind): Add annot_expr_ivdep_kind.
4578 * tree-pretty-print.c (dump_generic_node): Handle ANNOTATE_EXPR.
4579 * tree.def (ANNOTATE_EXPR): New DEFTREECODE.
4580 * doc/extend.texi (Pragmas): Document #pragma ivdep.
4581 * doc/generic.texi (Expressions): Document ANNOTATE_EXPR.
4583 2013-10-17 Ian Bolton <ian.bolton@arm.com>
4584 Marcus Shawcroft <marcus.shawcroft@arm.com>
4586 * config/aarch64/aarch64.c (aarch64_preferred_reload_class):
4587 Special case reload SP+C into none GENERAL_REGS.
4589 2013-10-24 Michael Matz <matz@suse.de>
4591 * gengtype.c (is_file_equal): Check that files will be same length.
4593 2013-10-25 Christian Bruel <christian.bruel@st.com>
4595 * config.gcc (sh-*): Add sh-mem.o to extra_obj.
4596 * config/sh/t-sh (sh-mem.o): New rule.
4597 * config/sh/sh-mem.cc (expand_block_move): Moved here.
4598 (sh_expand_cmpstr): New function.
4599 * config/sh/sh.c (force_into, expand_block_move): Move to sh-mem.c.
4600 * config/sh/sh-protos.h (sh_expand_cmpstr): Declare.
4601 * config/sh/sh.md (cmpstrsi, cmpstr_t): New patterns.
4602 (rotlhi3_8): Rename.
4604 2013-10-24 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4606 * configure.ac (ZW_PROG_COMPILER_DEPENDENCIES): Use CXX instead of CC.
4607 * Makefile.in (CXXDEPMODE): Assign and change users.
4608 (CCDEPMODE): Delete.
4609 * configure: Regenerate.
4611 2013-10-23 David Malcolm <dmalcolm@redhat.com>
4613 * gengtype-parse.c (require_without_advance): New.
4614 (type): For GTY-marked types that are not GTY((user)), parse any
4615 base classes, requiring them to be single-inheritance, and not
4616 be templates. For non-GTY-marked types and GTY((user)),
4617 continue to skip over any C++ inheritance specification.
4618 * gengtype-state.c (state_writer::write_state_struct_type):
4619 Write base class of type (if any).
4620 (read_state_struct_type): Read base class of type (if any).
4621 * gengtype.c (new_structure): Add a "base_class" parameter.
4622 (create_optional_field_): Update for new parameter to new_structure.
4623 (adjust_field_rtx_def): Likewise.
4624 (adjust_field_tree_exp): Likewise.
4625 * gengtype.h (struct type): Add "base_class" field to the s
4627 (new_structure): Add "base" parameter.
4629 2013-10-23 Sriraman Tallam <tmsriram@google.com>
4632 * config/i386/i386.c (ix86_option_override_internal):
4633 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...)
4634 (ix86_valid_target_attribute_tree):
4635 Change TARGET_64BIT to TARGET_64BIT_P (opts->...)
4636 Change TARGET_SSE to TARGET_SSE_P (opts->...)
4638 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
4640 * tree-ssa-loop.h: Remove include files.
4641 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
4642 * cfgloopmanip.c: Move required includes from tree-ssa-loop.h.
4643 * graphite-clast-to-gimple.c: Likewise.
4644 * graphite-scop-detection.c: Likewise.
4645 * graphite-sese-to-poly.c: Likewise.
4646 * ipa-inline-analysis.c: Likewise.
4647 * ipa-pure-const.c: Likewise.
4648 * loop-init.c: Likewise.
4649 * passes.c: Likewise.
4650 * predict.c: Likewise.
4651 * tree-cfg.c: Likewise.
4652 * tree-cfgcleanup.c: Likewise.
4653 * tree-chrec.c: Likewise.
4654 * tree-data-ref.c: Likewise.
4655 * tree-loop-distribution.c: Likewise.
4656 * tree-parloops.c: Likewise.
4657 * tree-predcom.c: Likewise.
4658 * tree-scalar-evolution.c: Likewise.
4659 * tree-ssa-address.c: Likewise.
4660 * tree-ssa.c: Likewise.
4661 * tree-ssa-dce.c: Likewise.
4662 * tree-ssa-loop.c: Likewise.
4663 * tree-ssa-loop-im.c: Likewise.
4664 * tree-ssa-loop-ivcanon.c: Likewise.
4665 * tree-ssa-loop-ivopts.c: Likewise.
4666 * tree-ssa-loop-manip.c: Likewise.
4667 * tree-ssa-loop-niter.c: Likewise.
4668 * tree-ssa-loop-prefetch.c: Likewise.
4669 * tree-ssa-loop-unswitch.c: Likewise.
4670 * tree-ssa-reassoc.c: Likewise.
4671 * tree-vect-data-refs.c: Likewise.
4672 * tree-vect-loop.c: Likewise.
4673 * tree-vect-loop-manip.c: Likewise.
4674 * tree-vectorizer.c: Likewise.
4675 * tree-vect-stmts.c: Likewise.
4676 * tree-vrp.c: Likewise.
4678 2013-10-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4680 * config/rs6000/altivec.md (mulv8hi3): Adjust for little endian.
4682 2013-10-23 Jakub Jelinek <jakub@redhat.com>
4684 PR tree-optimization/58775
4685 PR tree-optimization/58791
4686 * tree-ssa-reassoc.c (reassoc_stmt_dominates_stmt_p): New function.
4687 (insert_stmt_after): Rewritten, don't move the stmt, but really
4689 (get_stmt_uid_with_default): Remove.
4690 (build_and_add_sum): Use insert_stmt_after and
4691 reassoc_stmt_dominates_stmt_p. Fix up uid if bb contains only labels.
4692 (update_range_test): Set uid on stmts added by
4693 force_gimple_operand_gsi. Don't immediately modify statements
4694 in inter-bb optimization, just update oe->op values.
4695 (optimize_range_tests): Return bool whether any changed have been made.
4696 (update_ops): New function.
4697 (struct inter_bb_range_test_entry): New type.
4698 (maybe_optimize_range_tests): Perform statement changes here.
4699 (not_dominated_by, appears_later_in_bb, get_def_stmt,
4700 ensure_ops_are_available): Remove.
4701 (find_insert_point): Rewritten.
4702 (rewrite_expr_tree): Remove MOVED argument, add CHANGED argument,
4703 return LHS of the (new resp. old) stmt. Don't call
4704 ensure_ops_are_available, don't reuse SSA_NAMEs, recurse first
4705 instead of last, move new stmt at the right place.
4706 (linearize_expr, repropagate_negates): Don't reuse SSA_NAMEs.
4707 (negate_value): Likewise. Set uids.
4708 (break_up_subtract_bb): Initialize uids.
4709 (reassociate_bb): Adjust rewrite_expr_tree caller.
4710 (do_reassoc): Don't call renumber_gimple_stmt_uids.
4712 2013-10-23 David Edelsohn <dje.gcc@gmail.com>
4715 * config/rs6000/rs6000.md (mulsi3_internal1 and splitter): Add
4716 TARGET_32BIT final condition.
4717 (mulsi3_internal2 and splitter): Same.
4719 2013-10-23 Jeff Law <law@redhat.com>
4721 * tree-ssa-threadedge.c (thread_across_edge): Do not allow threading
4722 through joiner blocks with abnormal outgoing edges.
4724 * tree-ssa-threadupdate.c (thread_block_1): Renamed from thread_block.
4725 Add parameter JOINERS, to allow/disallow threading through joiner
4727 (thread_block): New. Call thread_block_1.
4728 (mark_threaded_blocks): Remove code to filter out certain cases
4729 of threading through joiner blocks.
4730 (thread_through_all_blocks): Document how we can have a dangling
4731 edge AUX field and clear it.
4733 2013-10-23 Ian Lance Taylor <iant@google.com>
4735 * doc/invoke.texi (Option Summary): Remove -fno-default-inline.
4736 (C++ Dialect Options): Likewise.
4737 (Optimize Options): Likewise.
4739 2013-10-23 Tom de Vries <tom@codesourcery.com>
4741 PR tree-optimization/58805
4742 * tree-ssa-tail-merge.c (stmt_local_def): Add gimple_vdef check.
4744 2013-10-23 Jakub Jelinek <jakub@redhat.com>
4746 * tree-vect-patterns.c (vect_recog_divmod_pattern): Optimize
4747 sequence based on get_range_info returned range.
4749 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
4751 * tree-ssa.h: Remove all #include's
4752 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
4753 * alias.c: Move required includes from tree-ssa.h.
4755 * builtins.c: Likewise.
4756 * calls.c: Likewise.
4757 * cfgexpand.c: Likewise.
4758 * cfghooks.c: Likewise.
4759 * cfgloop.c: Likewise.
4760 * cfgloopmanip.c: Likewise.
4761 * cgraph.c: Likewise.
4762 * cgraphbuild.c: Likewise.
4763 * cgraphclones.c: Likewise.
4764 * cgraphunit.c: Likewise.
4766 * except.c: Likewise.
4768 * final.c: Likewise.
4769 * fold-const.c: Likewise.
4770 * ggc-page.c: Likewise.
4771 * gimple-builder.c: Likewise.
4772 * gimple-fold.c: Likewise.
4773 * gimple-iterator.c: Likewise.
4774 * gimple-low.c: Likewise.
4775 * gimple-pretty-print.c: Likewise.
4776 * gimple-ssa-strength-reduction.c: Likewise.
4777 * gimple-streamer-in.c: Likewise.
4778 * gimple-streamer-out.c: Likewise.
4779 * gimplify.c: Likewise.
4780 * graphite-blocking.c: Likewise.
4781 * graphite-clast-to-gimple.c: Likewise.
4782 * graphite-dependences.c: Likewise.
4783 * graphite-interchange.c: Likewise.
4784 * graphite-optimize-isl.c: Likewise.
4785 * graphite-poly.c: Likewise.
4786 * graphite-scop-detection.c: Likewise.
4787 * graphite-sese-to-poly.c: Likewise.
4788 * graphite.c: Likewise.
4789 * ipa-cp.c: Likewise.
4790 * ipa-inline-analysis.c: Likewise.
4791 * ipa-inline-transform.c: Likewise.
4792 * ipa-inline.c: Likewise.
4793 * ipa-prop.c: Likewise.
4794 * ipa-pure-const.c: Likewise.
4795 * ipa-reference.c: Likewise.
4796 * ipa-split.c: Likewise.
4797 * ipa-utils.c: Likewise.
4798 * loop-init.c: Likewise.
4799 * lto-cgraph.c: Likewise.
4800 * lto-section-in.c: Likewise.
4801 * lto-section-out.c: Likewise.
4802 * lto-streamer-in.c: Likewise.
4803 * lto-streamer-out.c: Likewise.
4804 * lto-streamer.c: Likewise.
4805 * omp-low.c: Likewise.
4806 * passes.c: Likewise.
4807 * predict.c: Likewise.
4808 * print-tree.c: Likewise.
4809 * profile.c: Likewise.
4811 * targhooks.c: Likewise.
4812 * tracer.c: Likewise.
4813 * trans-mem.c: Likewise.
4814 * tree-call-cdce.c: Likewise.
4815 * tree-cfg.c: Likewise.
4816 * tree-cfgcleanup.c: Likewise.
4817 * tree-chrec.c: Likewise.
4818 * tree-complex.c: Likewise.
4819 * tree-data-ref.c: Likewise.
4820 * tree-dfa.c: Likewise.
4821 * tree-eh.c: Likewise.
4822 * tree-emutls.c: Likewise.
4823 * tree-if-conv.c: Likewise.
4824 * tree-inline.c: Likewise.
4825 * tree-into-ssa.c: Likewise.
4826 * tree-loop-distribution.c: Likewise.
4827 * tree-mudflap.c: Likewise.
4828 * tree-nested.c: Likewise.
4829 * tree-nrv.c: Likewise.
4830 * tree-object-size.c: Likewise.
4831 * tree-outof-ssa.c: Likewise.
4832 * tree-parloops.c: Likewise.
4833 * tree-phinodes.c: Likewise.
4834 * tree-predcom.c: Likewise.
4835 * tree-pretty-print.c: Likewise.
4836 * tree-profile.c: Likewise.
4837 * tree-scalar-evolution.c: Likewise.
4838 * tree-sra.c: Likewise.
4839 * tree-ssa-address.c: Likewise.
4840 * tree-ssa-alias.c: Likewise.
4841 * tree-ssa-ccp.c: Likewise.
4842 * tree-ssa-coalesce.c: Likewise.
4843 * tree-ssa-copy.c: Likewise.
4844 * tree-ssa-copyrename.c: Likewise.
4845 * tree-ssa-dce.c: Likewise.
4846 * tree-ssa-dom.c: Likewise.
4847 * tree-ssa-dse.c: Likewise.
4848 * tree-ssa-forwprop.c: Likewise.
4849 * tree-ssa-ifcombine.c: Likewise.
4850 * tree-ssa-live.c: Likewise.
4851 * tree-ssa-loop-ch.c: Likewise.
4852 * tree-ssa-loop-im.c: Likewise.
4853 * tree-ssa-loop-ivcanon.c: Likewise.
4854 * tree-ssa-loop-ivopts.c: Likewise.
4855 * tree-ssa-loop-manip.c: Likewise.
4856 * tree-ssa-loop-niter.c: Likewise.
4857 * tree-ssa-loop-prefetch.c: Likewise.
4858 * tree-ssa-loop-unswitch.c: Likewise.
4859 * tree-ssa-loop.c: Likewise.
4860 * tree-ssa-math-opts.c: Likewise.
4861 * tree-ssa-operands.c: Likewise.
4862 * tree-ssa-phiopt.c: Likewise.
4863 * tree-ssa-phiprop.c: Likewise.
4864 * tree-ssa-pre.c: Likewise.
4865 * tree-ssa-propagate.c: Likewise.
4866 * tree-ssa-reassoc.c: Likewise.
4867 * tree-ssa-sccvn.c: Likewise.
4868 * tree-ssa-sink.c: Likewise.
4869 * tree-ssa-strlen.c: Likewise.
4870 * tree-ssa-structalias.c: Likewise.
4871 * tree-ssa-tail-merge.c: Likewise.
4872 * tree-ssa-ter.c: Likewise.
4873 * tree-ssa-threadedge.c: Likewise.
4874 * tree-ssa-threadupdate.c: Likewise.
4875 * tree-ssa-uncprop.c: Likewise.
4876 * tree-ssa-uninit.c: Likewise.
4877 * tree-ssa.c: Likewise.
4878 * tree-ssanames.c: Likewise.
4879 * tree-stdarg.c: Likewise.
4880 * tree-streamer-in.c: Likewise.
4881 * tree-switch-conversion.c: Likewise.
4882 * tree-tailcall.c: Likewise.
4883 * tree-vect-data-refs.c: Likewise.
4884 * tree-vect-generic.c: Likewise.
4885 * tree-vect-loop-manip.c: Likewise.
4886 * tree-vect-loop.c: Likewise.
4887 * tree-vect-patterns.c: Likewise.
4888 * tree-vect-slp.c: Likewise.
4889 * tree-vect-stmts.c: Likewise.
4890 * tree-vectorizer.c: Likewise.
4891 * tree-vrp.c: Likewise.
4894 * value-prof.c: Likewise.
4895 * var-tracking.c: Likewise.
4896 * varpool.c: Likewise.
4897 * vtable-verify.c: Likewise.
4899 2013-10-23 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4901 * config/tilegx/tilegx.c: Include "tree.h".
4903 2013-10-23 Jakub Jelinek <jakub@redhat.com>
4905 * gimple-pretty-print.c (dump_ssaname_info): Always print "# " before
4906 the info, not after it.
4907 (gump_gimple_phi): Add COMMENT argument, if true, print "# " after
4908 dump_ssaname_info call.
4909 (pp_gimple_stmt_1): Adjust caller.
4910 (dump_phi_nodes): Likewise. Don't print "# " here.
4912 2013-10-22 Jan Hubicka <jh@suse.cz>
4914 * i386.h (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES_EPILOGUES): New
4916 * x86-tune.def (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES): Define it.
4917 * i386.c (expand_small_movmem_or_setmem): New function.
4918 (expand_set_or_movmem_prologue_epilogue_by_misaligned_moves): New
4920 (alg_usable_p): Add support for value ranges; cleanup.
4921 (ix86_expand_set_or_movmem): Add support for misaligned moves.
4923 2013-10-22 Sterling Augustine <saugustine@google.com>
4925 * doc/invoke.texi: Document -ggnu-pubnames.
4926 * common.opt: Add new option -ggnu-pubnames and modify -gpubnames
4928 * dwarf2out.c: Include gdb/gdb-index.h.
4929 (DEBUG_PUBNAMES_SECTION, DEBUG_PUBTYPES_SECTION): Handle
4930 debug_generate_pub_sections.
4931 (is_java, output_pubtables, output_pubname): New functions.
4932 (include_pubname_in_output): Handle debug_generate_pub_sections at
4934 (size_of_pubnames): Use new local space_for_flags based on
4935 debug_generate_pub_sections.
4936 (output_pubnames): Unify pubnames and pubtypes output logic.
4937 Genericize comments. Call output_pubname.
4938 (dwarf2out_finish): Move logic to output_pubnames and call it.
4940 2013-10-22 Uros Bizjak <ubizjak@gmail.com>
4943 * config/i386/i386.c (put_condition_code) <case GTU, case LEU>:
4944 Remove CCCmode handling.
4945 <case LTU>: Return 'c' suffix for CCCmode.
4946 <case GEU>: Return 'nc' suffix for CCCmode.
4947 (ix86_cc_mode) <case GTU, case LEU>: Do not generate overflow checks.
4948 * config/i386/i386.md (*sub<mode>3_cconly_overflow): Remove.
4949 (*sub<mode>3_cc_overflow): Ditto.
4950 (*subsi3_zext_cc_overflow): Ditto.
4952 2013-10-22 Steve Ellcey <sellcey@mips.com>
4954 * config/mips/mips.c (mips_rtx_costs): Fix cost estimate for nor
4955 (AND (NOT OP1) (NOT OP2)).
4957 2013-10-22 Bill Schmidt <wschmidt@vnet.ibm.com>
4959 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
4960 meaning of merge-high and merge-low masks for little endian; avoid
4961 use of vector-pack masks for little endian for mismatched modes.
4963 2013-10-22 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4965 * config/tilepro/tilepro.c: Include "tree.h".
4967 2013-10-22 Andreas Schwab <schwab@suse.de>
4969 * config/m68k/m68k.c (notice_update_cc): Handle register conflict
4972 2013-10-22 Paolo Carlini <paolo.carlini@oracle.com>
4974 * doc/contrib.texi ([Fran@,{c}ois Dumont], [Tim Shen],
4975 [Ed Smith-Rowland]): New entries.
4976 ([Stephen M. Webb]): Update.
4978 2013-10-22 Andrew MacLeod <amacleod@redhat.com>
4980 * tree-ssa-ter.h: Remove duplicate copy of file contents.
4982 2013-10-21 Marek Polacek <polacek@redhat.com>
4985 * fold-const.c (fold_range_test): Return 0 if the type is not
4988 2013-10-21 Richard Sandiford <rdsandiford@googlemail.com>
4990 * system.h: Move hwint.h include further down.
4991 * hwint.h (sext_hwi, zext_hwi): Define unconditionally. Add
4992 gcc_checking_asserts.
4993 * hwint.c (sext_hwi, zext_hwi): Delete ENABLE_CHECKING versions.
4995 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
4997 Fix volatile issues in optimize_bit_field_compare.
4998 * fold-const.c (optimize_bit_field_compare): Bail out if
4999 lvolatilep or rvolatilep.
5001 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
5003 Fix DECL_BIT_FIELD depencency on flag_strict_volatile_bitfields
5004 and get_inner_reference returning different pmode for non-volatile
5005 bit-field members dependent on flag_strict_volatile_bitfields.
5006 * stor-layout.c (layout_decl): Remove special handling of
5007 flag_strict_volatile_bitfields.
5008 * expr.c (get_inner_reference): Don't use DECL_BIT_FIELD
5009 if flag_strict_volatile_bitfields > 0 and TREE_THIS_VOLATILE.
5011 2013-10-21 Paulo Matos <pmatos@broadcom.com>
5013 * ipa-inline.c (edge_badness): Cap edge->count at max_count for badness
5016 2013-10-21 Jeff Law <law@redhat.com>
5018 * tree-ssa-threadedge.c (thread_through_normal_block): New
5019 argument VISITED. Remove VISISTED as a local variable. When we
5020 have a threadable jump, verify the destination of the jump has not
5022 (thread_across_edge): Allocate VISITED bitmap once at function
5023 scope and use it throughout. Make sure to set appropriate bits in
5024 VISITED for E (start of jump thread path).
5025 * tree-ssa-threadupdate.c (mark_threaded_blocks): Reject threading
5026 through a joiner if any edge on the path has a recorded jump thread.
5028 2013-10-21 Ian Lance Taylor <iant@google.com>
5030 * doc/invoke.texi (Optimize Options): For -fno-toplevel-reorder,
5031 don't imply that attributes can solve all problems.
5032 (Directory Options): Fix typo.
5034 2013-10-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5036 * config/arm/arm.c (cortexa9_extra_costs): Update mult costs for
5037 extend and extend_add.
5039 2013-10-21 Richard Biener <rguenther@suse.de>
5041 PR tree-optimization/58794
5042 * fold-const.c (operand_equal_p): Compare FIELD_DECL operand
5043 of COMPONENT_REFs with OEP_CONSTANT_ADDRESS_OF left in place.
5045 2013-10-21 Richard Biener <rguenther@suse.de>
5048 * fold-const.c (fold_binary_loc): Fold ((T) (X /[ex] C)) * C
5049 to (T) X for sign-changing conversions (or no conversion).
5051 2013-10-20 Uros Bizjak <ubizjak@gmail.com>
5053 * config/i386/i386.md (kxnor<mode>): Add FLAGS_REG clobber.
5055 2013-10-20 Jan Hubicka <jh@suse.cz>
5057 * config/i386/i386-tune.def: Add comment; organize into categories
5059 2013-10-21 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
5061 * config/i386/i386.c (expand_set_or_movmem_via_loop): Add issetmem
5062 argument. Update function comment.
5063 (expand_set_or_movmem_via_rep): New function combining
5064 expand_movmem_via_rep_mov and expand_setmem_via_rep_stos.
5065 (expand_movmem_via_rep_mov): Remove.
5066 expand_setmem_via_rep_stos): Remove.
5067 (expand_movmem_epilogue): Update calls correspondingly.
5068 (expand_setmem_epilogue_via_loop): Likewise.
5070 (expand_setmem_epilogue): Add VEC_VALUE argument, refactor.
5071 (expand_set_or_movmem_prologue): New function combining
5072 expand_movmem_prologue and expand_setmem_prologue.
5073 (expand_movmem_prologue): Remove.
5074 (expand_setmem_prologue): Remove.
5075 (expand_set_or_movmem_constant_prologue): New function combining
5076 expand_constant_movmem_prologue and expand_constant_setmem_prologue.
5077 (expand_constant_movmem_prologue): Remove.
5078 (expand_constant_setmem_prologue): Remove.
5079 (promote_duplicated_reg): Allow vector-const0 value.
5080 (ix86_expand_set_or_movmem): New function combining ix86_expand_movmem
5081 and ix86_expand_setmem.
5082 (ix86_expand_movmem): Call ix86_expand_set_or_movmem.
5083 (ix86_expand_setmem): Call ix86_expand_set_or_movmem.
5085 2013-10-21 Diego Novillo <dnovillo@google.com>
5087 * asan.c: Include tree.h
5088 * bb-reorder.c: Likewise.
5089 * cfgcleanup.c: Likewise.
5090 * cfgloopmanip.c: Likewise.
5091 * data-streamer-in.c: Likewise.
5092 * data-streamer-out.c: Likewise.
5093 * data-streamer.c: Likewise.
5094 * dwarf2cfi.c: Likewise.
5095 * graphite-blocking.c: Likewise.
5096 * graphite-clast-to-gimple.c: Likewise.
5097 * graphite-dependences.c: Likewise.
5098 * graphite-interchange.c: Likewise.
5099 * graphite-optimize-isl.c: Likewise.
5100 * graphite-poly.c: Likewise.
5101 * graphite-scop-detection.c: Likewise.
5102 * graphite-sese-to-poly.c: Likewise.
5103 * graphite.c: Likewise.
5104 * ipa-devirt.c: Likewise.
5105 * ipa-profile.c: Likewise.
5108 * loop-init.c: Likewise.
5109 * loop-unroll.c: Likewise.
5110 * lower-subreg.c: Likewise.
5111 * lto/lto-object.c: Likewise.
5112 * recog.c: Likewise.
5113 * reginfo.c: Likewise.
5114 * tree-loop-distribution.c: Likewise.
5115 * tree-parloops.c: Likewise.
5116 * tree-ssa-strlen.c: Likewise.
5117 * tree-streamer.c: Likewise.
5118 * value-prof.c: Likewise.
5119 * target-globals.c: Likewise.
5120 * expr.h: Include tree-core.h instead of tree.h.
5121 * gimple.h: Likewise.
5122 * ipa-prop.h: Likewise.
5123 * ipa-utils.h: Likewise.
5124 * lto-streamer.h: Likewise.
5125 * streamer-hooks.h: Likewise.
5126 * ipa-reference.h: Include cgraph.h instead of tree.h.
5127 * cgraph.h: Include basic-block.h instead of tree.h.
5128 * tree-streamer.h: Do not include tree.h.
5129 * genattrtab.c (write_header): Generate inclusion of tree.h.
5130 * genautomata.c (main): Likewise.
5131 * genemit.c: Likewise.
5132 * genopinit.c: Likewise.
5133 * genoutput.c (output_prologue): Likewise.
5134 * genpeep.c: Likewise.
5136 2013-10-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5138 * config/rs6000/altivec.md (vec_unpacku_hi_v16qi): Adjust for
5140 (vec_unpacku_hi_v8hi): Likewise.
5141 (vec_unpacku_lo_v16qi): Likewise.
5142 (vec_unpacku_lo_v8hi): Likewise.
5144 2013-10-20 Jan Hubicka <jh@suse.cz>
5146 * config/i386/x86-tune.def (X86_TUNE_SLOW_IMUL_IMM32_MEM,
5147 X86_TUNE_SLOW_IMUL_IMM8): Keep enabled only for K8 and AMDFAM10.
5148 (X86_TUNE_USE_VECTOR_FP_CONVERTS): Disable for generic.
5150 2013-10-20 Richard Sandiford <rdsandiford@googlemail.com>
5152 * config/mips/mips.h (ISA_HAS_WSBH): Define.
5153 * config/mips/mips.md (UNSPEC_WSBH, UNSPEC_DSBH, UNSPEC_DSHD): New
5155 (bswaphi2, bswapsi2, bswapdi2, wsbh, dsbh, dshd): New patterns.
5157 2013-10-19 John David Anglin <danglin@gcc.gnu.org>
5160 * system.h: Undef m_slot.
5162 2013-10-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5164 * config/rs6000/rs6000.c (vspltis_constant): Make sure we check
5165 all elements for both endian flavors.
5167 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
5170 * config/i386/i386.c (ix86_function_value_regno): Add DX_REG,
5171 ST1_REG and XMM1_REG for 32bit and 64bit targets. Also add DI_REG
5172 and SI_REG for 64bit SYSV ABI targets.
5174 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
5176 * mode-switching.c (create_pre_exit): Rename maybe_builtin_apply
5177 to multi_reg_return. Clarify that we are skipping USEs of multiple
5178 return registers. Use bool type where appropriate.
5180 2013-10-18 Jan Hubicka <jh@suse.cz>
5182 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Disable accumulation
5184 * x86-tune.def (X86_TUNE_USE_LEAVE): Update comment.
5185 (X86_TUNE_PUSH_MEMORY): Likewise.
5186 (X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL,
5187 X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL): New.
5188 (X86_TUNE_ACCUMULATE_OUTGOING_ARGS, X86_TUNE_ALWAYS_FANCY_MATH_387):
5190 * i386.c (x86_accumulate_outgoing_args, x86_arch_always_fancy_math_387,
5191 x86_avx256_split_unaligned_load, x86_avx256_split_unaligned_store):
5193 (ix86_option_override_internal): Update to use tune features instead
5196 2013-10-18 Cong Hou <congh@google.com>
5198 PR tree-optimization/58508
5199 * tree-vect-loop-manip.c (vect_loop_versioning): Hoist loop invariant
5200 statement that contains data refs with zero-step.
5202 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
5204 * tree-ssa.h: Don't include gimple-low.h, tree-ssa-address.h,
5205 sbitmap.h, tree-ssa-threadedge.h, tree-ssa-dom.h and tree-cfgcleanup.h.
5206 * gimple-low.c (gimple_check_call_arg,
5207 gimple_check_call_matching_types): Move to cgraph.c.
5208 * gimple-low.h: Remove prototype.
5209 * cgraph.c: (gimple_check_call_arg, gimple_check_call_matching_types):
5210 Relocate from gimple-low.c.
5211 * cgraph.h: Add prototype. Don't include basic-block.h.
5212 * gimplify.c: Add gimple-low to include list.
5213 * omp-low.c: Add gimple-low and tree-cfgcleanup.h to include list.
5214 * tree-eh.c: Add gimple-low to include list.
5215 * tree-nested.c: Likewise.
5216 * cfgexpand.c: Add tree-ssa-address.h to include list.
5218 * gimple-fold.c: Likewise.
5219 * gimple-ssa-strength-reduction.c: Likewise.
5220 * trans-mem.c: Likewise.
5221 * tree-mudflap.c: Likewise.
5222 * tree-ssa-loop-ivopts.c: Likewise.
5223 * tree-ssa-dom.c: Include tree-ssa-threadedge.h and tree-ssa-dom.h.
5224 (degenerate_phi_result): Move to tree-phinodes.c.
5225 * tree-ssa-loop-ch.c: Include tree-ssa-threadedge.h.
5226 * tree-ssa-threadedge.c: Likewise.
5227 * tree-vrp.c: Likewise.
5228 * tree-phinodes.c (degenerate_phi_result): Relocate here.
5229 * tree-ssa-dom.h (degenerate_phi_result): Remove Prototype.
5230 * tree-phinodes.h (degenerate_phi_result): Add prototype.
5231 * tree-ssa-copy.c: Include tree-ssa-dom.h.
5232 * tree-ssa-forwprop.c: Likewise.
5233 * tree-cfgcleanup.c (execute_cleanup_cfg_post_optimizing,
5234 pass_data_cleanup_cfg_post_optimizing,
5235 make_pass_cleanup_cfg_post_optimizing): Relocate from tree-optimize.c.
5236 * tree-optimize.c: Delete File.
5237 * graphite.c: Include tree-cfgcleanup.h.
5238 * passes.c: Likewise.
5239 * tree-cfg.c: Likewise.
5240 * tree-profile.c: Likewise.
5241 * tree-ssa-dse.c: Likewise.
5242 * tree-ssa-loop-ivcanon.c: Likewise.
5243 * tree-switch-conversion.c: Don't include tree-ssa-operands.h.
5244 * tree-outof-ssa.c: Include sbitmap.h.
5245 * tree-ssa-live.c: Likewise.
5246 * tree-ssa-propagate.c: Likewise.
5247 * tree-ssa-structalias.c: Likewise.
5248 * tree-stdarg.c: Likewise.
5249 * Makefile.in (OBJS): Delete tree-optimize.o.
5250 * basic-block.h (gcov_type, gcov_type_unsigned): Move to coretypes.h.
5251 * coretypes.h (gcov_type, gcov_type_unsigned): Relocate here.
5252 * varasm.c: Include basic-block.h.
5253 * cfgloop.h: Include function.h instead of basic-block.h
5254 (bb_loop_depth): Move to cfgloop.c.
5255 * cfgloop.c (bb_loop_depth): Relocate from cfgloop.h.
5257 2013-10-18 Teresa Johnson <tejohnson@google.com>
5259 * predict.c (probably_never_executed): Compare frequency-based
5260 count to number of training runs.
5261 * params.def (UNLIKELY_BB_COUNT_FRACTION): New parameter.
5263 2013-10-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5265 * config/arm/arm.c (cortexa9_extra_costs): New table.
5266 (arm_cortex_a9_tune): Use cortexa9_extra_costs.
5268 2013-10-18 Jeff Law <law@redhat.com>
5270 * tree-ssa-threadupdate.c: Do not include "tm.h" or "tm_p.h".
5272 * tree-ssa-threadupdate.c: Include "dbgcnt.h".
5273 (register_jump_thread): Add "registered_jump_thread" debug
5275 * dbgcnt.def (registered_jump_thread): New debug counter.
5277 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
5279 * config/rs6000/rs6000.c: Include cgraph.h.
5281 2013-10-18 Teresa Johnson <tejohnson@google.com>
5283 * tree-ssa-tail-merge.c (replace_block_by): Update edge
5284 weights during merging.
5286 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
5288 * tree-cfg.h: Rename from tree-flow.h. Remove #includes.
5289 * tree-ssa.h: Relocate required #includes from tree-cfg.h.
5290 * tree-ssa-operands.h: Remove prototype.
5291 * tree-ssa-operands.c (virtual_operand_p): Move to gimple.c.
5292 * gimple.c (virtual_operand_p): Relocate from gimple.c.
5293 * gimple.h: Add prototype.
5294 * gimple-ssa.h: Include tree-ssa-operands.h.
5295 * tree-dump.c: Add tree-cfg.h to include list.
5296 * tree-ssa-alias.c: Add ipa-reference.h to include list.
5297 * config/alpha/alpha.c: Include gimple-ssa.h instead of tree-flow.h.
5298 * config/i386/i386.c: Don't include tree-flow.h.
5299 * config/rs6000/rs6000.c: Likewise.
5301 2013-10-18 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5303 * config/frv/frv.c (frv_init_cumulative_args): Fix wrong cast.
5305 2013-10-18 Richard Biener <rguenther@suse.de>
5307 * stor-layout.c (layout_type): Do not change TYPE_PRECISION
5308 or TYPE_UNSIGNED of integral types.
5309 (set_min_and_max_values_for_integral_type): Leave TYPE_MIN/MAX_VALUE
5310 NULL_TREE for zero-precision integral types.
5312 2013-10-18 James Greenhalgh <james.greenhalgh@arm.com>
5314 * config/aarch64/arm_neon.h
5315 (vcvt<ds>_n_<fsu><32,64>_<fsu><32,64>): Correct argument types.
5317 2013-10-17 Sriraman Tallam <tmsriram@google.com>
5320 * opth-gen.awk: Define target_flags_explicit.
5322 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
5324 * config/rs6000/rs6000.c (enum rs6000_reload_reg_type): Add new
5325 fields to the reg_addr array that describes the valid addressing
5326 mode for any register, general purpose registers, floating point
5327 registers, and Altivec registers.
5328 (FIRST_RELOAD_REG_CLASS): Likewise.
5329 (LAST_RELOAD_REG_CLASS): Likewise.
5330 (struct reload_reg_map_type): Likewise.
5331 (reload_reg_map_type): Likewise.
5332 (RELOAD_REG_VALID): Likewise.
5333 (RELOAD_REG_MULTIPLE): Likewise.
5334 (RELOAD_REG_INDEXED): Likewise.
5335 (RELOAD_REG_OFFSET): Likewise.
5336 (RELOAD_REG_PRE_INCDEC): Likewise.
5337 (RELOAD_REG_PRE_MODIFY): Likewise.
5338 (reg_addr): Likewise.
5339 (mode_supports_pre_incdec_p): New helper functions to say whether
5340 a given mode supports PRE_INC, PRE_DEC, and PRE_MODIFY.
5341 (mode_supports_pre_modify_p): Likewise.
5342 (rs6000_debug_vector_unit): Rearrange the -mdebug=reg output to
5343 print the valid address mode bits for each mode.
5344 (rs6000_debug_print_mode): Likewise.
5345 (rs6000_debug_reg_global): Likewise.
5346 (rs6000_setup_reg_addr_masks): New function to set up the address
5347 mask bits for each type.
5348 (rs6000_init_hard_regno_mode_ok): Use memset to clear arrays.
5349 Call rs6000_setup_reg_addr_masks to set up the address mask bits.
5350 (rs6000_legitimate_address_p): Use mode_supports_pre_incdec_p and
5351 mode_supports_pre_modify_p to determine if PRE_INC, PRE_DEC, and
5352 PRE_MODIFY are supported.
5353 (rs6000_output_move_128bit): Change to use {src,dest}_vmx_p for altivec
5354 registers, instead of {src,dest}_av_p.
5355 (rs6000_print_options_internal): Tweak the debug output slightly.
5357 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
5359 * config/i386/sse.md (*vec_widen_smult_even_v8si): Remove
5362 2013-10-17 Andrew MacLeod <amacleod@redhat.com>
5364 * tree-flow.h (struct omp_region): Move to omp-low.c.
5365 Remove omp_ prototypes and variables.
5366 * gimple.h (omp_reduction_init): Move prototype to omp-low.h.
5367 (copy_var_decl): Relocate prototype from tree-flow.h.
5368 * gimple.c (copy_var_decl): Relocate from omp-low.c.
5369 * tree.h: Move prototype to omp-low.h.
5370 * omp-low.h: New File. Relocate prototypes here.
5371 * omp-low.c (struct omp_region): Make local here.
5372 (root_omp_region): Make static.
5373 (copy_var_decl) Move to gimple.c.
5374 (new_omp_region): Make static.
5375 (make_gimple_omp_edges): New. Refactored from tree-cfg.c make_edges.
5376 * tree-cfg.c: Include omp-low.h.
5377 (make_edges): Factor out OMP specific bits to make_gimple_omp_edges.
5378 * gimplify.c: Include omp-low.h.
5379 * tree-parloops.c: Likewise.
5381 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
5383 * config/i386/i386.c (ix86_fixup_binary_operands): When both source
5384 operands are in memory, prefer to force non-matched operand 1 to
5387 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
5390 * config/rs6000/rs6000.c (rs6000_legitimate_address_p): Only
5391 restrict TImode addresses to single indirect registers if both
5392 -mquad-memory and -mvsx-timode are used.
5393 (rs6000_output_move_128bit): Use quad_load_store_p to determine if
5394 we should emit load/store quad. Remove using %y for quad memory
5397 * config/rs6000/rs6000.md (mov<mode>_ppc64, TI/PTImode): Add
5398 constraints to allow load/store quad on machines where TImode is
5399 not allowed in VSX registers. Use 'n' instead of 'F' constraint
5400 for TImode to load integer constants.
5402 2013-10-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5404 * config/aarch64/aarch64.c (aarch64_print_operand): Handle 'c'.
5406 2013-10-17 Marcus Shawcroft <marcus.shawcroft@arm.com>
5408 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Adjust
5409 handling of STACK_REG.
5411 2013-10-17 Richard Biener <rguenther@suse.de>
5413 PR tree-optimization/58143
5414 * tree-ssa-loop-im.c (arith_code_with_undefined_signed_overflow):
5416 (rewrite_to_defined_overflow): Likewise.
5417 (move_computations_dom_walker::before_dom): Rewrite stmts
5418 with undefined signed overflow that are not always executed
5419 into unsigned arithmetic.
5421 2013-10-16 Michael Meissner <meissner@linux.vnet.ibm.com>
5424 * config/rs6000/rs6000.opt (rs6000_isa_flags_explicit): Move the
5425 explicit isa flag to be an options variable, instead of using
5426 global_options_set. Remove define from rs6000.h.
5427 * config/rs6000/rs6000.h (rs6000_isa_flags_explicit): Likewise.
5429 * config/rs6000/rs6000.c (rs6000_option_override_internal):
5430 Initialize rs6000_isa_flags_explicit.
5431 (rs6000_function_specific_save): Add gcc_options* parameter, so
5432 that the powerpc builds after the 2013-10-15 changes.
5433 (rs6000_function_specific_restore): Likewise.
5435 2013-10-16 DJ Delorie <dj@redhat.com>
5437 * config/rl78/rl78.c (rl78_alloc_address_registers_macax): Verify
5438 op is a REG before checking REGNO.
5439 (rl78_alloc_physical_registers): Verify pattern is a SET before
5442 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5444 * config/rs6000/vector.md (vec_unpacks_hi_v4sf): Correct for
5446 (vec_unpacks_lo_v4sf): Likewise.
5447 (vec_unpacks_float_hi_v4si): Likewise.
5448 (vec_unpacks_float_lo_v4si): Likewise.
5449 (vec_unpacku_float_hi_v4si): Likewise.
5450 (vec_unpacku_float_lo_v4si): Likewise.
5452 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5454 * config/rs6000/vsx.md (vsx_concat_<mode>): Adjust output for LE.
5455 (vsx_concat_v2sf): Likewise.
5457 2013-10-16 James Greenhalgh <james.greenhalgh@arm.com>
5459 * config/aarch64/aarch64.md
5460 (*mov<mode>_aarch64): Fix output template for DUP (element) Scalar.
5462 2013-10-16 Andrew MacLeod <amacleod@redhat.com>
5464 PR tree-optimization/58697
5465 * cfgloop.c (get_estimated_loop_iterations_int): Rename from
5466 estimated_loop_iterations_int.
5467 (max_stmt_executions_int): Call get_max_loop_iterations_int.
5468 (get_max_loop_iterations_int): New. HWINT version of
5469 get_max_loop_iterations.
5470 * cfgloop.h: Add prototypes.
5471 * loop-iv.c (find_simple_exit): call get_estimated_loop_iterations_int.
5472 * loop-unroll.c (decide_peel_once_rolling): Call
5473 get_estimated_loop_iterations_int.
5474 * tree-ssa-loop-niter.c (estimated_loop_iterations_int): Add back.
5475 * tree-ssa-loop-niter.h: Tweak prototypes.
5477 2013-10-16 David Malcolm <dmalcolm@redhat.com>
5479 * gengtype-parse.c (struct_field_seq): Ignore access-control
5480 keywords ("public:" etc).
5482 2013-10-16 Marcus Shawcroft <marcus.shawcroft@arm.com>
5484 * config/aarch64/aarch64.c (aarch64_regno_regclass): Classify
5485 FRAME_POINTER_REGNUM and ARG_POINTER_REGNUM as POINTER_REGS.
5487 2013-10-16 Yvan Roux <yvan.roux@linaro.org>
5489 * config/arm/arm.opt (mlra): New option.
5490 * config/arm/arm.c (arm_lra_p): New function.
5491 (TARGET_LRA_P): Define.
5493 2013-10-16 Paulo Matos <pmatos@broadcom.com>
5495 * tree-core.h (tree_code_name): Remove.
5496 * tree.h (get_tree_code_name): New prototype.
5497 * tree.c (tree_code_name): Make static.
5498 (get_tree_code_name): New function.
5499 (dump_tree_statistics, tree_check_failed, tree_not_check_failed,
5500 tree_class_check_failed, tree_range_check_failed,
5501 tree_not_class_check_failed, omp_clause_check_failed,
5502 tree_contains_struct_check_failed, tree_operand_check_failed): Use new
5503 wrapper get_tree_code_name instead of calling tree_code_name directly.
5504 * tree-vrp.c (dump_asserts_for): Likewise.
5505 * tree-dump.c (dequeue_and_dump): Likewise.
5506 * tree-pretty-print.c (do_niy, dump_generic_node): Likewise.
5507 * tree-pretty-print.h (pp_unsupported_tree): Likewise.
5508 * lto-streamer-out.c (lto_write_tree, DFS_write_tree): Likewise.
5509 * tree-ssa-dom.c (print_expr_hash_elt): Likewise.
5510 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs,
5511 dump_ternary_rhs, dump_gimple_assign, dump_gimple_cond,
5512 dump_gimple_omp_for): Likewise.
5513 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Likewise.
5514 * tree-ssa-pre.c (print_pre_expr): Likewise.
5515 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Likewise.
5516 * print-tree.c (print_node_brief, print_node): Likewise.
5517 * gimple.c (gimple_check_failed): Likewise.
5518 * lto-streamer.c (lto_tag_name, print_lto_report): Likewise.
5519 * config/frv/frv.c (frv_init_cumulative_args): Likewise.
5520 * config/mep/mep.c (mep_validate_vliw): Likewise.
5521 * config/iq2000/iq2000.c (init_cumulative_args): Likewise.
5522 * config/rs6000/rs6000.c (init_cumulative_args): Likewise.
5524 2013-10-16 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
5526 * config/i386/i386.c (ix86_option_override_internal): Enable FMA4
5529 2013-10-16 Hans-Peter Nilsson <hp@axis.com>
5531 * config/cris/t-elfmulti (MULTILIB_OPTIONS, MULTILIB_DIRNAMES)
5532 (MULTILIB_MATCHES): Add multilib for -march=v8.
5534 2013-10-15 Sriraman Tallam <tmsriram@google.com>
5537 * optc-save-gen.awk: Add extra parameter to the save and restore
5539 * opth-gen.awk: Generate new TARGET_* macros to accept a parameter.
5540 * tree.c (build_optimization_node): New parameter. Add extra parameter
5541 to call to cl_optimization_save.
5542 (build_target_option_node): New parameter. Add extra parameter
5543 to call to cl_target_option_save.
5544 * tree.h (build_optimization_node): New parameter.
5545 (build_target_option_node): New parameter.
5546 * c-family/c-common.c (handle_optimize_attribute): Fix calls to
5547 build_optimization_node and build_target_option_node.
5548 * c-family/c-pragma.c (handle_pragma_optimize): Ditto.
5549 (handle_pragma_push_options): Ditto.
5550 * toplev.c (process_options): Ditto.
5551 * opts.c (init_options_struct): Check for opts_set non-null.
5552 * target.def (target_option.save): New parameter.
5553 (target_option.restore): New parameter.
5554 * tm.texi: Generate.
5555 * config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
5556 (ix86_pragma_target_parse): Ditto.
5557 * config/i386/i386-protos.h (ix86_valid_target_attribute_tree): New
5559 * config/rs6000/rs6000.c (rs6000_option_override_internal): Fix calls
5560 to build_optimization_node and build_target_option_node.
5561 (rs6000_valid_attribute_p): Ditto.
5562 (rs6000_pragma_target_parse): Ditto.
5563 * config/i386/i386.opt (x_ix86_target_flags_explicit): New TargetSave
5565 * config/i386/i386.h:
5566 TARGET_64BIT_P: New Macro
5567 TARGET_MMX_P: New Macro.
5568 TARGET_3DNOW_P: New Macro.
5569 TARGET_3DNOW_A_P: New Macro.
5570 TARGET_SSE_P: New Macro.
5571 TARGET_SSE2_P: New Macro.
5572 TARGET_SSE3_P: New Macro.
5573 TARGET_SSSE3_P: New Macro.
5574 TARGET_SSE4_1_P: New Macro.
5575 TARGET_SSE4_2_P: New Macro.
5576 TARGET_AVX_P: New Macro.
5577 TARGET_AVX2_P: New Macro.
5578 TARGET_AVX512F_P: New Macro.
5579 TARGET_AVX512PF_P: New Macro.
5580 TARGET_AVX512ER_P: New Macro.
5581 TARGET_AVX512CD_P: New Macro.
5582 TARGET_FMA_P: New Macro.
5583 TARGET_SSE4A_P: New Macro.
5584 TARGET_FMA4_P: New Macro.
5585 TARGET_XOP_P: New Macro.
5586 TARGET_LWP_P: New Macro.
5587 TARGET_ABM_P: New Macro.
5588 TARGET_BMI_P: New Macro.
5589 TARGET_BMI2_P: New Macro.
5590 TARGET_LZCNT_P: New Macro.
5591 TARGET_TBM_P: New Macro.
5592 TARGET_POPCNT_P: New Macro.
5593 TARGET_SAHF_P: New Macro.
5594 TARGET_MOVBE_P: New Macro.
5595 TARGET_CRC32_P: New Macro.
5596 TARGET_AES_P: New Macro.
5597 TARGET_PCLMUL_P: New Macro.
5598 TARGET_CMPXCHG16B_P: New Macro.
5599 TARGET_FSGSBASE_P: New Macro.
5600 TARGET_RDRND_P: New Macro.
5601 TARGET_F16C_P: New Macro.
5602 TARGET_RTM_P: New Macro.
5603 TARGET_HLE_P: New Macro.
5604 TARGET_RDSEED_P: New Macro.
5605 TARGET_PRFCHW_P: New Macro.
5606 TARGET_ADX_P: New Macro.
5607 TARGET_FXSR_P: New Macro.
5608 TARGET_XSAVE_P: New Macro.
5609 TARGET_XSAVEOPT_P: New Macro.
5610 TARGET_LP64_P: New Macro.
5611 TARGET_X32_P: New Macro.
5612 TARGET_FPMATH_DEFAULT_P: New Macro.
5613 TARGET_FLOAT_RETURNS_IN_80387_P: New Macro.
5614 * config/i386/i386.c (ix86_option_override_internal): New parameters.
5616 Change ix86_tune_string to access opts->x_ix86_tune_string.
5617 Change ix86_isa_flags to access opts->x_ix86_isa_flags.
5618 Change ix86_arch_string to access opts->x_ix86_arch_string.
5619 Change ix86_stringop_alg to access opts->x_ix86_stringop_alg.
5620 Change ix86_pmode to access opts->x_ix86_pmode.
5621 Change ix86_abi to access opts->x_ix86_abi.
5622 Change ix86_cmodel to access opts->x_ix86_cmodel.
5623 Change ix86_asm_dialect to access opts->x_ix86_asm_dialect.
5624 Change ix86_isa_flags_explicit to access
5625 opts->x_ix86_isa_flags_explicit.
5626 Change ix86_dump_tunes to access opts->x_ix86_dump_tunes.
5627 Change ix86_regparm to access opts->x_ix86_regparm.
5628 Change ix86_branch_cost to access opts->x_ix86_branch_cost.
5629 Change ix86_preferred_stack_boundary_arg to access
5630 opts->x_ix86_preferred_stack_boundary_arg.
5631 Change ix86_force_align_arg_pointer to access
5632 opts->x_ix86_force_align_arg_pointer.
5633 Change ix86_incoming_stack_boundar_arg to access
5634 opts->x_ix86_incoming_stack_boundar_arg.
5635 Change ix86_fpmath to access opts->x_ix86_fpmath.
5636 Change ix86_veclibabi_type to access opts->x_ix86_veclibabi_type.
5637 Change ix86_recip_name to access opts->x_ix86_recip_name.
5638 Change ix86_stack_protector_guard to access
5639 opts->x_ix86_stack_protector_guard.
5640 Change ix86_tune_memcpy_strategy to access
5641 opts->x_ix86_tune_memcpy_strategy.
5642 Change ix86_tune_memset_strategy to access
5643 opts->x_ix86_tune_memset_strategy.
5644 Change global_options to access opts.
5645 Change global_options_set to access opts_set.
5646 Change TARGET_64BIT to TARGET_64BIT_P (opts->...).
5647 Change TARGET_MMX to TARGET_MMX_P (opts->...).
5648 Change TARGET_3DNOW to TARGET_3DNOW_P (opts->...).
5649 Change TARGET_3DNOW_A to TARGET_3DNOW_A_P (opts->...).
5650 Change TARGET_SSE to TARGET_SSE_P (opts->...).
5651 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...).
5652 Change TARGET_SSE3 to TARGET_SSE3_P (opts->...).
5653 Change TARGET_SSSE3 to TARGET_SSSE3_P (opts->...).
5654 Change TARGET_SSE4_1 to TARGET_SSE4_1_P (opts->...).
5655 Change TARGET_SSE4_2 to TARGET_SSE4_2_P (opts->...).
5656 Change TARGET_AVX to TARGET_AVX_P (opts->...).
5657 Change TARGET_AVX2 to TARGET_AVX2_P (opts->...).
5658 Change TARGET_AVX512F to TARGET_AVX512F_P (opts->...).
5659 Change TARGET_AVX512PF to TARGET_AVX512PF_P (opts->...).
5660 Change TARGET_AVX512ER to TARGET_AVX512ER_P (opts->...).
5661 Change TARGET_AVX512CD to TARGET_AVX512CD_P (opts->...).
5662 Change TARGET_FMA to TARGET_FMA_P (opts->...).
5663 Change TARGET_SSE4A to TARGET_SSE4A_P (opts->...).
5664 Change TARGET_FMA4 to TARGET_FMA4_P (opts->...).
5665 Change TARGET_XOP to TARGET_XOP_P (opts->...).
5666 Change TARGET_LWP to TARGET_LWP_P (opts->...).
5667 Change TARGET_ABM to TARGET_ABM_P (opts->...).
5668 Change TARGET_BMI to TARGET_BMI_P (opts->...).
5669 Change TARGET_BMI2 to TARGET_BMI2_P (opts->...).
5670 Change TARGET_LZCNT to TARGET_LZCNT_P (opts->...).
5671 Change TARGET_TBM to TARGET_TBM_P (opts->...).
5672 Change TARGET_POPCNT to TARGET_POPCNT_P (opts->...).
5673 Change TARGET_SAHF to TARGET_SAHF_P (opts->...).
5674 Change TARGET_MOVBE to TARGET_MOVBE_P (opts->...).
5675 Change TARGET_CRC32 to TARGET_CRC32_P (opts->...).
5676 Change TARGET_AES to TARGET_AES_P (opts->...).
5677 Change TARGET_PCLMUL to TARGET_PCLMUL_P (opts->...).
5678 Change TARGET_CMPXCHG16B to TARGET_CMPXCHG16B_P (opts->...).
5679 Change TARGET_FSGSBASE to TARGET_FSGSBASE_P (opts->...).
5680 Change TARGET_RDRND to TARGET_RDRND_P (opts->...).
5681 Change TARGET_F16C to TARGET_F16C_P (opts->...).
5682 Change TARGET_RTM to TARGET_RTM_P (opts->...).
5683 Change TARGET_HLE to TARGET_HLE_P (opts->...).
5684 Change TARGET_RDSEED to TARGET_RDSEED_P (opts->...).
5685 Change TARGET_PRFCHW to TARGET_PRFCHW_P (opts->...).
5686 Change TARGET_ADX to TARGET_ADX_P (opts->...).
5687 Change TARGET_FXSR to TARGET_FXSR_P (opts->...).
5688 Change TARGET_XSAVE to TARGET_XSAVE_P (opts->...).
5689 Change TARGET_XSAVEOPT to TARGET_XSAVEOPT_P (opts->...).
5690 Change TARGET_LP64 to TARGET_LP64_P (opts->...).
5691 Change TARGET_X32 to TARGET_X32_P (opts->...).
5692 Change TARGET_FPMATH_DEFAULT to TARGET_FPMATH_DEFAULT_P (opts->...).
5693 Change TARGET_FLOAT_RETURNS_IN_80387 to
5694 TARGET_FLOAT_RETURNS_IN_80387_P (opts->...).
5695 (ix86_function_specific_save): New parameter. Use opts-> fields
5696 to replace global fields.
5697 (ix86_function_specific_restore): Ditto.
5698 (ix86_valid_target_attribute_inner_p): New parameters.
5700 Fix call to ix86_handle_option and set_option.
5701 (ix86_valid_target_attribute_tree): New parameters.
5702 Change global_options to access opts.
5703 Change global_options_set to access opts_set.
5704 Fix call to ix86_valid_target_attribute_inner_p.
5705 Change ix86_tune_string to access opts->x_ix86_tune_string.
5706 Change ix86_arch_string to access opts->x_ix86_arch_string.
5707 Change ix86_fpmath to access opts->x_ix86_fpmath
5708 Fix call to ix86_option_override_internal.
5709 Fix call to ix86_add_new_builtins.
5710 Fix calls to build_optimization_node and build_target_option_node.
5711 (ix86_valid_target_attribute_p): Remove access to global_options.
5712 Use new gcc_options structure func_options.
5713 Fix call to ix86_valid_target_attribute_tree.
5714 Fix call to build_optimization_node.
5715 (get_builtin_code_for_version): Fix call to
5716 ix86_valid_target_attribute_tree.
5718 2013-10-15 David Malcolm <dmalcolm@redhat.com>
5720 * Makefile.in (PICFLAG): New.
5721 (enable_host_shared): New.
5722 (INTERNAL_CFLAGS): Use PICFLAG.
5723 (LIBIBERTY): Use pic build of libiberty.a if configured with
5724 --enable-host-shared.
5725 * configure.ac: Add --enable-host-shared, setting up new
5727 * configure: Regenerate.
5728 * doc/install.texi (--enable-shared): Add note contrasting it with ...
5729 (--enable-host-shared): New option.
5731 2013-10-15 Richard Biener <rguenther@suse.de>
5733 * tree-tailcall.c (find_tail_calls): Don't use tail-call recursion
5734 for built-in functions.
5736 2013-10-15 Zhenqiang Chen <zhenqiang.chen@arm.com>
5738 * tree-ssa-reassoc.c: Include rtl.h and tm_p.h.
5739 (optimize_range_tests_1): New function,
5740 extracted from optimize_range_tests.
5741 (optimize_range_tests_xor): Similarly.
5742 (optimize_range_tests_diff): New function.
5743 (optimize_range_tests): Use optimize_range_tests_1.
5745 2013-10-15 Cong Hou <congh@google.com>
5747 * tree-vect-loop.c (vect_is_simple_reduction_1): Relax the
5748 requirement of the reduction pattern so that one operand of the
5749 reduction operation can come from outside of the loop.
5751 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
5753 * config/arm/neon-schedgen.ml: Remove.
5754 * config/arm/cortex-a9-neon.md: Remove comment regarding
5757 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
5759 * config/arm/types: Remove old neon types.
5761 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
5763 * config/arm/cortex-a7.md
5764 (cortex_a7_neon_type): New.
5765 (cortex_a7_neon_mul): Update for new types.
5766 (cortex_a7_neon_mla): Likewise.
5767 (cortex_a7_neon): Likewise.
5769 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
5771 * config/arm/cortex-a15-neon.md
5772 (cortex_a15_neon_type): New,
5774 (cortex_a15_neon_int_1): Remove.
5775 (cortex_a15_neon_int_2): Likewise.
5776 (cortex_a15_neon_int_3): Likewise.
5777 (cortex_a15_neon_int_4): Likewise.
5778 (cortex_a15_neon_int_5): Likewise.
5779 (cortex_a15_neon_vqneg_vqabs): Likewise.
5780 (cortex_a15_neon_vmov): Likewise.
5781 (cortex_a15_neon_vaba): Likewise.
5782 (cortex_a15_neon_vaba_qqq): Likewise.
5783 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
5784 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
5785 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
5787 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
5788 (cortex_a15_neon_mla_qqq_8_16): Likewise.
5789 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
5790 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
5791 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
5792 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
5793 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
5794 (cortex_a15_neon_shift_1): Likewise.
5795 (cortex_a15_neon_shift_2): Likewise.
5796 (cortex_a15_neon_shift_3): Likewise.
5797 (cortex_a15_neon_vshl_ddd): Likewise.
5798 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
5799 (cortex_a15_neon_vsra_vrsra): Likewise.
5800 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
5801 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
5802 (cortex_a15_neon_bp_3cycle): Likewise.
5803 (cortex_a15_neon_ldm_2): Likewise.
5804 (cortex_a15_neon_stm_2): Likewise.
5805 (cortex_a15_neon_mcr): Likewise.
5806 (cortex_a15_neon_mrc): Likewise.
5807 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
5808 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
5809 (cortex_a15_neon_fp_vmul_ddd): Likewise.
5810 (cortex_a15_neon_fp_vmul_qqd): Likewise.
5811 (cortex_a15_neon_fp_vmla_ddd): Likewise.
5812 (cortex_a15_neon_fp_vmla_qqq): Likewise.
5813 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
5814 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
5815 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
5816 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
5817 (cortex_a15_neon_bp_simple): Likewise.
5818 (cortex_a15_neon_bp_2cycle): Likewise.
5819 (cortex_a15_neon_bp_3cycle): Likewise.
5820 (cortex_a15_neon_vld1_1_2_regs): Likewise.
5821 (cortex_a15_neon_vld1_3_4_regs): Likewise.
5822 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
5823 (cortex_a15_neon_vld2_4_regs): Likewise.
5824 (cortex_a15_neon_vld3_vld4): Likewise.
5825 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
5826 (cortex_a15_neon_vst1_3_4_regs): Likewise.
5827 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
5828 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
5829 (cortex_a15_neon_vst3_vst4): Rename to...
5830 (cortex_a15_neon_vst4): This, update for new attributes.
5831 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
5832 (cortex_a15_neon_vld3_vld4_lane): Likewise.
5833 (cortex_a15_neon_vst1_vst2_lane): Likewise.
5834 (cortex_a15_neon_vst3_vst4_lane): Likewise.
5835 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
5836 (cortex_a15_neon_ldm_2): Likewise.
5837 (cortex_a15_neon_stm_2): Likewise.
5838 (cortex_a15_neon_mcr): Likewise.
5839 (cortex_a15_neon_mcr_2_mcrr): Likewise.
5840 (cortex_a15_neon_mrc): Likewise.
5841 (cortex_a15_neon_mrrc): Likewise.
5843 (cortex_a15_neon_abd): New.
5844 (cortex_a15_neon_abd_q): Likewise.
5845 (cortex_a15_neon_aba): Likewise.
5846 (cortex_a15_neon_aba_q): Likewise.
5847 (cortex_a15_neon_acc): Likewise.
5848 (cortex_a15_neon_acc_q): Likewise.
5849 (cortex_a15_neon_arith_basic): Likewise.
5850 (cortex_a15_neon_arith_complex): Likewise.
5851 (cortex_a15_neon_multiply): Likewise.
5852 (cortex_a15_neon_multiply_q): Likewise.
5853 (cortex_a15_neon_mla): Likewise.
5854 (cortex_a15_neon_mla_q): Likewise.
5855 (cortex_a15_neon_sat_mla_long): Likewise.
5856 (cortex_a15_neon_shift_acc): Likewise.
5857 (cortex_a15_neon_shift_imm_basic): Likewise.
5858 (cortex_a15_neon_shift_imm_complex): Likewise.
5859 (cortex_a15_neon_shift_reg_basic): Likewise.
5860 (cortex_a15_neon_shift_reg_basic_q): Likewise.
5861 (cortex_a15_neon_shift_reg_complex): Likewise.
5862 (cortex_a15_neon_shift_reg_complex_q): Likewise.
5863 (cortex_a15_neon_fp_negabs): Likewise
5864 (cortex_a15_neon_fp_arith): Likewise
5865 (cortex_a15_neon_fp_arith_q): Likewise
5866 (cortex_a15_neon_fp_cvt_int): Likewise
5867 (cortex_a15_neon_fp_cvt_int_q): Likewise
5868 (cortex_a15_neon_fp_cvt_16): Likewise
5869 (cortex_a15_neon_fp_mul): Likewise
5870 (cortex_a15_neon_fp_mul_q): Likewise
5871 (cortex_a15_neon_fp_mla): Likewise
5872 (cortex_a15_neon_fp_mla_q): Likewise
5873 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
5874 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
5875 (cortex_a15_neon_bitops): Likewise.
5876 (cortex_a15_neon_bitops_q): Likewise.
5877 (cortex_a15_neon_from_gp): Likewise.
5878 (cortex_a15_neon_from_gp_q): Likewise.
5879 (cortex_a15_neon_tbl3_tbl4): Likewise.
5880 (cortex_a15_neon_zip_q): Likewise.
5881 (cortex_a15_neon_to_gp): Likewise.
5882 (cortex_a15_neon_load_a): Likewise.
5883 (cortex_a15_neon_load_b): Likewise.
5884 (cortex_a15_neon_load_c): Likewise.
5885 (cortex_a15_neon_load_d): Likewise.
5886 (cortex_a15_neon_load_e): Likewise.
5887 (cortex_a15_neon_load_f): Likewise.
5888 (cortex_a15_neon_store_a): Likewise.
5889 (cortex_a15_neon_store_b): Likewise.
5890 (cortex_a15_neon_store_c): Likewise.
5891 (cortex_a15_neon_store_d): Likewise.
5892 (cortex_a15_neon_store_e): Likewise.
5893 (cortex_a15_neon_store_f): Likewise.
5894 (cortex_a15_neon_store_g): Likewise.
5895 (cortex_a15_neon_store_h): Likewise.
5896 (cortex_a15_vfp_to_from_gp): Likewise.
5898 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
5900 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
5902 (cortex_a9_neon_vshl_ddd): Remove.
5903 (cortex_a9_neon_vst3_vst4): Likewise.
5904 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
5906 (cortex_a9_neon_bit_ops_q): New.
5908 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
5909 (cortex_a9_neon_int_2): Likewise.
5910 (cortex_a9_neon_int_3): Likewise.
5911 (cortex_a9_neon_int_4): Likewise.
5912 (cortex_a9_neon_int_5): Likewise.
5913 (cortex_a9_neon_vqneg_vqabs): Likewise.
5914 (cortex_a9_neon_vmov): Likewise.
5915 (cortex_a9_neon_vaba): Likewise.
5916 (cortex_a9_neon_vaba_qqq): Likewise.
5917 (cortex_a9_neon_shift_1): Likewise.
5918 (cortex_a9_neon_shift_2): Likewise.
5919 (cortex_a9_neon_shift_3): Likewise.
5920 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
5921 (cortex_a9_neon_vsra_vrsra): Likewise.
5922 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
5923 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
5924 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
5926 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
5927 (cortex_a9_neon_mla_qqq_8_16): Likewise.
5928 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
5930 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
5931 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
5932 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
5933 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
5934 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
5935 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
5936 (cortex_a9_neon_fp_vsum): Likewise.
5937 (cortex_a9_neon_fp_vmul_ddd): Likewise.
5938 (cortex_a9_neon_fp_vmul_qqd): Likewise.
5939 (cortex_a9_neon_fp_vmla_ddd): Likewise.
5940 (cortex_a9_neon_fp_vmla_qqq): Likewise.
5941 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
5942 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
5943 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
5944 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
5945 (cortex_a9_neon_bp_simple): Likewise.
5946 (cortex_a9_neon_bp_2cycle): Likewise.
5947 (cortex_a9_neon_bp_3cycle): Likewise.
5948 (cortex_a9_neon_ldr): Likewise.
5949 (cortex_a9_neon_str): Likewise.
5950 (cortex_a9_neon_vld1_1_2_regs): Likewise.
5951 (cortex_a9_neon_vld1_3_4_regs): Likewise.
5952 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
5953 (cortex_a9_neon_vld2_4_regs): Likewise.
5954 (cortex_a9_neon_vld3_vld4): Likewise.
5955 (cortex_a9_neon_vld1_vld2_lane): Likewise.
5956 (cortex_a9_neon_vld3_vld4_lane): Likewise.
5957 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
5958 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
5959 (cortex_a9_neon_vst1_3_4_regs): Likewise.
5960 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
5961 (cortex_a9_neon_vst1_vst2_lane): Likewise.
5962 (cortex_a9_neon_vst3_vst4_lane): Likewise.
5963 (cortex_a9_neon_mcr): Likewise.
5964 (cortex_a9_neon_mcr_2_mcrr): Likewise.
5965 (cortex_a9_neon_mrc): Likewise.
5966 (cortex_a9_neon_mrrc): Likewise.
5968 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
5970 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
5972 (cortex_a8_neon_vshl_ddd): Remove.
5973 (cortex_a8_neon_vst3_vst4): Likewise.
5974 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
5976 (cortex_a8_neon_bit_ops_q): New.
5978 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
5979 (cortex_a8_neon_int_2): Likewise..
5980 (cortex_a8_neon_int_3): Likewise.
5981 (cortex_a8_neon_int_5): Likewise.
5982 (cortex_a8_neon_vqneg_vqabs): Likewise.
5983 (cortex_a8_neon_int_4): Likewise.
5984 (cortex_a8_neon_vaba): Likewise.
5985 (cortex_a8_neon_vaba_qqq): Likewise.
5986 (cortex_a8_neon_shift_1): Likewise.
5987 (cortex_a8_neon_shift_2): Likewise.
5988 (cortex_a8_neon_shift_3): Likewise.
5989 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
5990 (cortex_a8_neon_vsra_vrsra): Likewise.
5991 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
5992 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
5993 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
5995 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
5996 (cortex_a8_neon_mla_qqq_8_16): Likewise.
5997 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
5999 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
6000 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
6001 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
6002 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
6003 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
6004 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
6005 (cortex_a8_neon_fp_vsum): Likewise.
6006 (cortex_a8_neon_fp_vmul_ddd): Likewise.
6007 (cortex_a8_neon_fp_vmul_qqd): Likewise.
6008 (cortex_a8_neon_fp_vmla_ddd): Likewise.
6009 (cortex_a8_neon_fp_vmla_qqq): Likewise.
6010 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
6011 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
6012 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
6013 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
6014 (cortex_a8_neon_bp_simple): Likewise.
6015 (cortex_a8_neon_bp_2cycle): Likewise.
6016 (cortex_a8_neon_bp_3cycle): Likewise.
6017 (cortex_a8_neon_ldr): Likewise.
6018 (cortex_a8_neon_str): Likewise.
6019 (cortex_a8_neon_vld1_1_2_regs): Likewise.
6020 (cortex_a8_neon_vld1_3_4_regs): Likewise.
6021 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
6022 (cortex_a8_neon_vld2_4_regs): Likewise.
6023 (cortex_a8_neon_vld3_vld4): Likewise.
6024 (cortex_a8_neon_vld1_vld2_lane): Likewise.
6025 (cortex_a8_neon_vld3_vld4_lane): Likewise.
6026 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
6027 (cortex_a8_neon_vst1_3_4_regs): Likewise.
6028 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
6029 (cortex_a8_neon_vst1_vst2_lane): Likewise.
6030 (cortex_a8_neon_vst3_vst4_lane): Likewise.
6031 (cortex_a8_neon_mcr): Likewise.
6032 (cortex_a8_neon_mcr_2_mcrr): Likewise.
6033 (cortex_a8_neon_mrc): Likewise.
6034 (cortex_a8_neon_mrrc): Likewise.
6036 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
6038 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
6040 * config/aarch64/aarch64-simd.md (neon_type): Remove.
6041 (aarch64_simd_dup<mode>): Add "type" attribute.
6042 (aarch64_dup_lane<mode>): Likewise.
6043 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
6044 (*aarch64_simd_mov<mode>): Likewise.
6045 (aarch64_simd_mov_from_<mode>low): Likewise.
6046 (aarch64_simd_mov_from_<mode>high): Likewise.
6047 (orn<mode>3): Likewise.
6048 (bic<mode>3): Likewise.
6049 (add<mode>3): Likewise.
6050 (sub<mode>3): Likewise.
6051 (mul<mode>3): Likewise.
6052 (*aarch64_mul3_elt<mode>): Likewise.
6053 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
6054 (*aarch64_mul3_elt_to_128df): Likewise.
6055 (*aarch64_mul3_elt_to_64v2df): Likewise.
6056 (neg<mode>2): Likewise.
6057 (abs<mode>2): Likewise.
6058 (abd<mode>_3): Likewise.
6059 (aba<mode>_3): Likewise.
6060 (fabd<mode>_3): Likewise.
6061 (*fabd_scalar<mode>3): Likewise.
6062 (and<mode>3): Likewise.
6063 (ior<mode>3): Likewise.
6064 (xor<mode>3): Likewise.
6065 (one_cmpl<mode>2): Likewise.
6066 (aarch64_simd_vec_set<mode>): Likewise.
6067 (aarch64_simd_lshr<mode>): Likewise.
6068 (aarch64_simd_ashr<mode>): Likewise.
6069 (aarch64_simd_imm_shl<mode>): Likewise.
6070 (aarch64_simd_reg_sshl<mode): Likewise.
6071 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
6072 (aarch64_simd_reg_shl<mode>_signed): Likewise.
6073 (aarch64_simd_vec_setv2di): Likewise.
6074 (aarch64_simd_vec_set<mode>): Likewise.
6075 (aarch64_mla<mode>): Likewise.
6076 (*aarch64_mla_elt<mode>): Likewise.
6077 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
6078 (aarch64_mls<mode>): Likewise.
6079 (*aarch64_mls_elt<mode>): Likewise.
6080 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
6081 (<su><maxmin><mode>3): Likewise.
6082 (move_lo_quad_<mode>): Likewise.
6083 (aarch64_simd_move_hi_quad_<mode>): Likewise.
6084 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
6085 (vec_pack_trunc_<mode>): Likewise.
6086 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
6087 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
6088 (*aarch64_<su>mlal_lo<mode>): Likewise.
6089 (*aarch64_<su>mlal_hi<mode>): Likewise.
6090 (*aarch64_<su>mlsl_lo<mode>): Likewise.
6091 (*aarch64_<su>mlsl_hi<mode>): Likewise.
6092 (*aarch64_<su>mlal<mode>): Likewise.
6093 (*aarch64_<su>mlsl<mode>): Likewise.
6094 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
6095 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
6096 (add<mode>3): Likewise.
6097 (sub<mode>3): Likewise.
6098 (mul<mode>3): Likewise.
6099 (div<mode>3): Likewise.
6100 (neg<mode>2): Likewise.
6101 (abs<mode>2): Likewise.
6102 (fma<mode>4): Likewise.
6103 (*aarch64_fma4_elt<mode>): Likewise.
6104 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
6105 (*aarch64_fma4_elt_to_128df): Likewise.
6106 (*aarch64_fma4_elt_to_64v2df): Likewise.
6107 (fnma<mode>4): Likewise.
6108 (*aarch64_fnma4_elt<mode>): Likewise.
6109 (*aarch64_fnma4_elt_<vswap_width_name><mode>
6110 (*aarch64_fnma4_elt_to_128df): Likewise.
6111 (*aarch64_fnma4_elt_to_64v2df): Likewise.
6112 (<frint_pattern><mode>2): Likewise.
6113 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
6114 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
6115 (vec_unpacks_lo_v4sf): Likewise.
6116 (aarch64_float_extend_lo_v2df): Likewise.
6117 (vec_unpacks_hi_v4sf): Likewise.
6118 (aarch64_float_truncate_lo_v2sf): Likewise.
6119 (aarch64_float_truncate_hi_v4sf): Likewise.
6120 (aarch64_vmls<mode>): Likewise.
6121 (<su><maxmin><mode>3): Likewise.
6122 (<maxmin_uns><mode>3): Likewise.
6123 (reduc_<sur>plus_<mode>): Likewise.
6124 (reduc_<sur>plus_v2di): Likewise.
6125 (reduc_<sur>plus_v2si): Likewise.
6126 (reduc_<sur>plus_<mode>): Likewise.
6127 (aarch64_addpv4sf): Likewise.
6128 (clz<mode>2): Likewise.
6129 (reduc_<maxmin_uns>_<mode>): Likewise.
6130 (reduc_<maxmin_uns>_v2di): Likewise.
6131 (reduc_<maxmin_uns>_v2si): Likewise.
6132 (reduc_<maxmin_uns>_<mode>): Likewise.
6133 (reduc_<maxmin_uns>_v4sf): Likewise.
6134 (aarch64_simd_bsl<mode>_internal): Likewise.
6135 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
6136 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
6137 (aarch64_get_lane<mode>): Likewise.
6138 (*aarch64_combinez<mode>): Likewise.
6139 (aarch64_combine<mode>): Likewise.
6140 (aarch64_simd_combine<mode>): Likewise.
6141 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
6142 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
6143 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
6144 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
6145 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
6146 (aarch64_<sur>h<addsub><mode>): Likewise.
6147 (aarch64_<sur><addsub>hn<mode>): Likewise.
6148 (aarch64_<sur><addsub>hn2<mode>): Likewise.
6149 (aarch64_pmul<mode>): Likewise.
6150 (aarch64_<su_optab><optab><mode>): Likewise.
6151 (aarch64_<sur>qadd<mode>): Likewise.
6152 (aarch64_sqmovun<mode>): Likewise.
6153 (aarch64_<sur>qmovn<mode>): Likewise.
6154 (aarch64_s<optab><mode>): Likewise.
6155 (aarch64_sq<r>dmulh<mode>): Likewise.
6156 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
6157 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
6158 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
6159 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
6160 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
6161 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
6162 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
6163 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
6164 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
6165 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
6166 (aarch64_sqdmull<mode>): Likewise.
6167 (aarch64_sqdmull_lane<mode>_internal): Likewise.
6168 (aarch64_sqdmull_n<mode>): Likewise.
6169 (aarch64_sqdmull2<mode>_internal): Likewise.
6170 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
6171 (aarch64_sqdmull2_n<mode>_internal): Likewise.
6172 (aarch64_<sur>shl<mode>): Likewise.
6173 (aarch64_<sur>q<r>shl<mode>
6174 (aarch64_<sur>shll_n<mode>): Likewise.
6175 (aarch64_<sur>shll2_n<mode>): Likewise.
6176 (aarch64_<sur>shr_n<mode>): Likewise.
6177 (aarch64_<sur>sra_n<mode>): Likewise.
6178 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
6179 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
6180 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
6181 (aarch64_cm<optab><mode>): Likewise.
6182 (aarch64_cm<optab>di): Likewise.
6183 (aarch64_cm<optab><mode>): Likewise.
6184 (aarch64_cm<optab>di): Likewise.
6185 (aarch64_cmtst<mode>): Likewise.
6186 (aarch64_cmtstdi): Likewise.
6187 (aarch64_cm<optab><mode>): Likewise.
6188 (*aarch64_fac<optab><mode>): Likewise.
6189 (aarch64_addp<mode>): Likewise.
6190 (aarch64_addpdi): Likewise.
6191 (sqrt<mode>2): Likewise.
6192 (vec_load_lanesoi<mode>): Likewise.
6193 (vec_store_lanesoi<mode>): Likewise.
6194 (vec_load_lanesci<mode>): Likewise.
6195 (vec_store_lanesci<mode>): Likewise.
6196 (vec_load_lanesxi<mode>): Likewise.
6197 (vec_store_lanesxi<mode>): Likewise.
6198 (*aarch64_mov<mode>): Likewise.
6199 (aarch64_ld2<mode>_dreg): Likewise.
6200 (aarch64_ld2<mode>_dreg): Likewise.
6201 (aarch64_ld3<mode>_dreg): Likewise.
6202 (aarch64_ld3<mode>_dreg): Likewise.
6203 (aarch64_ld4<mode>_dreg): Likewise.
6204 (aarch64_ld4<mode>_dreg): Likewise.
6205 (aarch64_tbl1<mode>): Likewise.
6206 (aarch64_tbl2v16qi): Likewise.
6207 (aarch64_combinev16qi): Likewise.
6208 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
6209 (aarch64_st2<mode>_dreg): Likewise.
6210 (aarch64_st2<mode>_dreg): Likewise.
6211 (aarch64_st3<mode>_dreg): Likewise.
6212 (aarch64_st3<mode>_dreg): Likewise.
6213 (aarch64_st4<mode>_dreg): Likewise.
6214 (aarch64_st4<mode>_dreg): Likewise.
6215 (*aarch64_simd_ld1r<mode>): Likewise.
6216 (aarch64_frecpe<mode>): Likewise.
6217 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
6218 (aarch64_frecps<mode>): Likewise.
6220 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
6222 * config/arm/iterators.md (V_elem_ch): New.
6224 (VQH_type): Likewise.
6225 * config/arm/arm.md (is_neon_type): New.
6226 (conds): Use is_neon_type.
6227 (anddi3_insn): Update type attribute.
6228 (xordi3_insn): Likewise.
6229 (one_cmpldi2): Likewise.
6230 * config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
6231 * config/arm/neon.md (neon_mov): Update type attribute.
6232 (*movmisalign<mode>_neon_store): Likewise.
6233 (*movmisalign<mode>_neon_load): Likewise.
6234 (vec_set<mode>_internal): Likewise.
6235 (vec_set<mode>_internal): Likewise.
6236 (vec_setv2di_internal): Likewise.
6237 (vec_extract<mode>): Likewise.
6238 (vec_extract<mode>): Likewise.
6239 (vec_extractv2di): Likewise.
6240 (*add<mode>3_neon): Likewise.
6241 (adddi3_neon): Likewise.
6242 (*sub<mode>3_neon): Likewise.
6243 (subdi3_neon): Likewise.
6244 (fma<VCVTF:mode>4): Likewise.
6245 (fma<VCVTF:mode>4_intrinsic): Likewise.
6246 (*fmsub<VCVTF:mode>4): Likewise.
6247 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
6248 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
6249 (ior<mode>3): Likewise.
6250 (and<mode>3): Likewise.
6251 (orn<mode>3_neon): Likewise.
6252 (orndi3_neon): Likewise.
6253 (bic<mode>3_neon): Likewise.
6254 (bicdi3_neon): Likewise.
6255 (xor<mode>3): Likewise.
6256 (one_cmpl<mode>2): Likewise.
6257 (abs<mode>2): Likewise.
6258 (neg<mode>2): Likewise.
6259 (negdi2_neon): Likewise.
6260 (*umin<mode>3_neon): Likewise.
6261 (*umax<mode>3_neon): Likewise.
6262 (*smin<mode>3_neon): Likewise.
6263 (*smax<mode>3_neon): Likewise.
6264 (vashl<mode>3): Likewise.
6265 (vashr<mode>3_imm): Likewise.
6266 (vlshr<mode>3_imm): Likewise.
6267 (ashl<mode>3_signed): Likewise.
6268 (ashl<mode>3_unsigned): Likewise.
6269 (neon_load_count): Likewise.
6270 (ashldi3_neon_noclobber): Likewise.
6271 (ashldi3_neon): Likewise.
6272 (signed_shift_di3_neon): Likewise.
6273 (unsigned_shift_di3_neon): Likewise.
6274 (ashrdi3_neon_imm_noclobber): Likewise.
6275 (lshrdi3_neon_imm_noclobber): Likewise.
6276 (<shift>di3_neon): Likewise.
6277 (widen_ssum<mode>3): Likewise.
6278 (widen_usum<mode>3): Likewise.
6279 (quad_halves_<code>v4si): Likewise.
6280 (quad_halves_<code>v4sf): Likewise.
6281 (quad_halves_<code>v8hi): Likewise.
6282 (quad_halves_<code>v16qi): Likewise.
6283 (reduc_splus_v2di): Likewise.
6284 (neon_vpadd_internal<mode>): Likewise.
6285 (neon_vpsmin<mode>): Likewise.
6286 (neon_vpsmax<mode>): Likewise.
6287 (neon_vpumin<mode>): Likewise.
6288 (neon_vpumax<mode>): Likewise.
6289 (*ss_add<mode>_neon): Likewise.
6290 (*us_add<mode>_neon): Likewise.
6291 (*ss_sub<mode>_neon): Likewise.
6292 (*us_sub<mode>_neon): Likewise.
6293 (neon_vadd<mode>_unspec): Likewise.
6294 (neon_vaddl<mode>): Likewise.
6295 (neon_vaddw<mode>): Likewise.
6296 (neon_vhadd<mode>): Likewise.
6297 (neon_vqadd<mode>): Likewise.
6298 (neon_vaddhn<mode>): Likewise.
6299 (neon_vmul<mode>): Likewise.
6300 (neon_vfms<VCVTF:mode>): Likewise.
6301 (neon_vmlal<mode>): Likewise.
6302 (neon_vmls<mode>): Likewise.
6303 (neon_vmlsl<mode>): Likewise.
6304 (neon_vqdmulh<mode>): Likewise.
6305 (neon_vqdmlal<mode>): Likewise.
6306 (neon_vqdmlsl<mode>): Likewise.
6307 (neon_vmull<mode>): Likewise.
6308 (neon_vqdmull<mode>): Likewise.
6309 (neon_vsub<mode>_unspec): Likewise.
6310 (neon_vsubl<mode>): Likewise.
6311 (neon_vsubw<mode>): Likewise.
6312 (neon_vqsub<mode>): Likewise.
6313 (neon_vhsub<mode>): Likewise.
6314 (neon_vsubhn<mode>): Likewise.
6315 (neon_vceq<mode>): Likewise.
6316 (neon_vcge<mode>): Likewise.
6317 (neon_vcgeu<mode>): Likewise.
6318 (neon_vcgt<mode>): Likewise.
6319 (neon_vcgtu<mode>): Likewise.
6320 (neon_vcle<mode>): Likewise.
6321 (neon_vclt<mode>): Likewise.
6322 (neon_vcage<mode>): Likewise.
6323 (neon_vcagt<mode>): Likewise.
6324 (neon_vtst<mode>): Likewise.
6325 (neon_vabd<mode>): Likewise.
6326 (neon_vabdl<mode>): Likewise.
6327 (neon_vaba<mode>): Likewise.
6328 (neon_vabal<mode>): Likewise.
6329 (neon_vmax<mode>): Likewise.
6330 (neon_vmin<mode>): Likewise.
6331 (neon_vpaddl<mode>): Likewise.
6332 (neon_vpadal<mode>): Likewise.
6333 (neon_vpmax<mode>): Likewise.
6334 (neon_vpmin<mode>): Likewise.
6335 (neon_vrecps<mode>): Likewise.
6336 (neon_vrsqrts<mode>): Likewise.
6337 (neon_vqabs<mode>): Likewise.
6338 (neon_vqneg<mode>): Likewise.
6339 (neon_vcls<mode>): Likewise.
6340 (clz<mode>2): Likewise.
6341 (popcount<mode>2): Likewise.
6342 (neon_vrecpe<mode>): Likewise.
6343 (neon_vrsqrte<mode>): Likewise.
6344 (neon_vget_lane<mode>_sext_internal): Likewise.
6345 (neon_vget_lane<mode>_zext_internal): Likewise.
6346 (neon_vdup_n<mode>): Likewise.
6347 (neon_vdup_n<mode>): Likewise.
6348 (neon_vdup_nv2di): Likewise.
6349 (neon_vdup_lane<mode>_interal): Likewise.
6350 (*neon_vswp<mode>): Likewise.
6351 (neon_vcombine<mode>): Likewise.
6352 (float<mode><V_cvtto>2): Likewise.
6353 (floatuns<mode><V_cvtto>2): Likewise.
6354 (fix_trunc<mode><V_cvtto>2): Likewise.
6355 (fixuns_trunc<mode><V_cvtto>2
6356 (neon_vcvt<mode>): Likewise.
6357 (neon_vcvt<mode>): Likewise.
6358 (neon_vcvtv4sfv4hf): Likewise.
6359 (neon_vcvtv4hfv4sf): Likewise.
6360 (neon_vcvt_n<mode>): Likewise.
6361 (neon_vcvt_n<mode>): Likewise.
6362 (neon_vmovn<mode>): Likewise.
6363 (neon_vqmovn<mode>): Likewise.
6364 (neon_vqmovun<mode>): Likewise.
6365 (neon_vmovl<mode>): Likewise.
6366 (neon_vmul_lane<mode>): Likewise.
6367 (neon_vmul_lane<mode>): Likewise.
6368 (neon_vmull_lane<mode>): Likewise.
6369 (neon_vqdmull_lane<mode>): Likewise.
6370 (neon_vqdmulh_lane<mode>): Likewise.
6371 (neon_vqdmulh_lane<mode>): Likewise.
6372 (neon_vmla_lane<mode>): Likewise.
6373 (neon_vmla_lane<mode>): Likewise.
6374 (neon_vmlal_lane<mode>): Likewise.
6375 (neon_vqdmlal_lane<mode>): Likewise.
6376 (neon_vmls_lane<mode>): Likewise.
6377 (neon_vmls_lane<mode>): Likewise.
6378 (neon_vmlsl_lane<mode>): Likewise.
6379 (neon_vqdmlsl_lane<mode>): Likewise.
6380 (neon_vext<mode>): Likewise.
6381 (neon_vrev64<mode>): Likewise.
6382 (neon_vrev32<mode>): Likewise.
6383 (neon_vrev16<mode>): Likewise.
6384 (neon_vbsl<mode>_internal): Likewise.
6385 (neon_vshl<mode>): Likewise.
6386 (neon_vqshl<mode>): Likewise.
6387 (neon_vshr_n<mode>): Likewise.
6388 (neon_vshrn_n<mode>): Likewise.
6389 (neon_vqshrn_n<mode>): Likewise.
6390 (neon_vqshrun_n<mode>): Likewise.
6391 (neon_vshl_n<mode>): Likewise.
6392 (neon_vqshl_n<mode>): Likewise.
6393 (neon_vqshlu_n<mode>): Likewise.
6394 (neon_vshll_n<mode>): Likewise.
6395 (neon_vsra_n<mode>): Likewise.
6396 (neon_vsri_n<mode>): Likewise.
6397 (neon_vsli_n<mode>): Likewise.
6398 (neon_vtbl1v8qi): Likewise.
6399 (neon_vtbl2v8qi): Likewise.
6400 (neon_vtbl3v8qi): Likewise.
6401 (neon_vtbl4v8qi): Likewise.
6402 (neon_vtbl1v16qi): Likewise.
6403 (neon_vtbl2v16qi): Likewise.
6404 (neon_vcombinev16qi): Likewise.
6405 (neon_vtbx1v8qi): Likewise.
6406 (neon_vtbx2v8qi): Likewise.
6407 (neon_vtbx3v8qi): Likewise.
6408 (neon_vtbx4v8qi): Likewise.
6409 (*neon_vtrn<mode>_insn): Likewise.
6410 (*neon_vzip<mode>_insn): Likewise.
6411 (*neon_vuzp<mode>_insn): Likewise.
6412 (neon_vld1<mode>): Likewise.
6413 (neon_vld1_lane<mode>): Likewise.
6414 (neon_vld1_lane<mode>): Likewise.
6415 (neon_vld1_dup<mode>): Likewise.
6416 (neon_vld1_dup<mode>): Likewise.
6417 (neon_vld1_dupv2di): Likewise.
6418 (neon_vst1<mode>): Likewise.
6419 (neon_vst1_lane<mode>): Likewise.
6420 (neon_vst1_lane<mode>): Likewise.
6421 (neon_vld2<mode>): Likewise.
6422 (neon_vld2<mode>): Likewise.
6423 (neon_vld2_lane<mode>): Likewise.
6424 (neon_vld2_lane<mode>): Likewise.
6425 (neon_vld2_dup<mode>): Likewise.
6426 (neon_vst2<mode>): Likewise.
6427 (neon_vst2<mode>): Likewise.
6428 (neon_vst2_lane<mode>): Likewise.
6429 (neon_vst2_lane<mode>): Likewise.
6430 (neon_vld3<mode>): Likewise.
6431 (neon_vld3qa<mode>): Likewise.
6432 (neon_vld3qb<mode>): Likewise.
6433 (neon_vld3_lane<mode>): Likewise.
6434 (neon_vld3_lane<mode>): Likewise.
6435 (neon_vld3_dup<mode>): Likewise.
6436 (neon_vst3<mode>): Likewise.
6437 (neon_vst3qa<mode>): Likewise.
6438 (neon_vst3qb<mode>): Likewise.
6439 (neon_vst3_lane<mode>): Likewise.
6440 (neon_vst3_lane<mode>): Likewise.
6441 (neon_vld4<mode>): Likewise.
6442 (neon_vld4qa<mode>): Likewise.
6443 (neon_vld4qb<mode>): Likewise.
6444 (neon_vld4_lane<mode>): Likewise.
6445 (neon_vld4_lane<mode>): Likewise.
6446 (neon_vld4_dup<mode>): Likewise.
6447 (neon_vst4<mode>): Likewise.
6448 (neon_vst4qa<mode>): Likewise.
6449 (neon_vst4qb<mode>): Likewise.
6450 (neon_vst4_lane<mode>): Likewise.
6451 (neon_vst4_lane<mode>): Likewise.
6452 (neon_vec_unpack<US>_lo_<mode>): Likewise.
6453 (neon_vec_unpack<US>_hi_<mode>): Likewise.
6454 (neon_vec_<US>mult_lo_<mode>): Likewise.
6455 (neon_vec_<US>mult_hi_<mode>): Likewise.
6456 (neon_vec_<US>shiftl_<mode>): Likewise.
6457 (neon_unpack<US>_<mode>): Likewise.
6458 (neon_vec_<US>mult_<mode>): Likewise.
6459 (vec_pack_trunc_<mode>): Likewise.
6460 (neon_vec_pack_trunc_<mode>): Likewise.
6461 (neon_vabd<mode>_2): Likewise.
6462 (neon_vabd<mode>_3): Likewise.
6464 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
6466 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
6467 (load_pair): Update type attribute.
6468 (store_pair): Update type attribute.
6469 * config/aarch64/iterators.md (q): New.
6471 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
6473 * config/arm/types.md: Add new types for Neon insns.
6475 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
6476 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6477 Sergey Lega <sergey.s.lega@intel.com>
6478 Anna Tikhonova <anna.tikhonova@intel.com>
6479 Ilya Tocar <ilya.tocar@intel.com>
6480 Andrey Turetskiy <andrey.turetskiy@intel.com>
6481 Ilya Verbin <ilya.verbin@intel.com>
6482 Kirill Yukhin <kirill.yukhin@intel.com>
6483 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6485 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
6486 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
6487 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
6489 (srcp14<mode>): Ditto.
6490 (rsqrt14<mode>): Ditto.
6491 (rsqrt14<mode>): Ditto.
6492 (avx512f_vmscalef<mode>): Ditto.
6493 (avx512f_scalef<mode>): Ditto.
6494 (avx512f_getexp<mode>): Ditto.
6495 (avx512f_sgetexp<mode>): Ditto.
6496 (avx512f_fixupimm<mode>): Ditto.
6497 (avx512f_sfixupimm<mode>): Ditto.
6498 (avx512f_rndscale<mode>): Ditto.
6499 (*avx512er_exp2<mode>): Ditto.
6500 (*avx512er_rcp28<mode>): Ditto.
6501 (avx512er_rsqrt28<mode>): Ditto.
6502 (avx512f_getmant<mode>): Ditto.
6503 (avx512f_getmant<mode>): Ditto.
6504 (avx512f_rndscale<mode>): Fix formatting.
6506 2013-10-15 Martin Jambor <mjambor@suse.cz>
6508 * ipa-utils.h (ipa_edge_within_scc): Declare.
6509 * ipa-cp.c (edge_within_scc): Moved...
6510 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
6512 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
6513 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6514 Sergey Lega <sergey.s.lega@intel.com>
6515 Anna Tikhonova <anna.tikhonova@intel.com>
6516 Ilya Tocar <ilya.tocar@intel.com>
6517 Andrey Turetskiy <andrey.turetskiy@intel.com>
6518 Ilya Verbin <ilya.verbin@intel.com>
6519 Kirill Yukhin <kirill.yukhin@intel.com>
6520 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6522 * config/i386/predicates.md (const_8_to_15_operand): New.
6523 (const_16_to_31_operand): Ditto.
6524 * config/i386/sse.md (V8FI): New.
6526 (reduc_splus_v8df): Ditto.
6527 (reduc_splus_v16sf): Ditto.
6528 (avx512f_vextract<shuffletype>32x4_1): Ditto.
6529 (vec_extract_hi_<mode>): Ditto.
6530 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
6531 (vec_set_lo_<mode>): Ditto.
6532 (vec_set_hi_<mode>): Ditto.
6533 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
6534 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
6535 (avx512f_pshufd_1): Ditto.
6536 (avx512f_broadcast<mode>): Ditto.
6537 (avx512f_broadcast<mode>): Ditto.
6538 (define_split): Split vec_extract_lo into move.
6539 (ssequartermode): Ditto.
6540 (ssedoublemode): Extened with wider modes.
6541 (vec_extract_lo_<mode>): Ditto.
6543 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
6544 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6545 Sergey Lega <sergey.s.lega@intel.com>
6546 Anna Tikhonova <anna.tikhonova@intel.com>
6547 Ilya Tocar <ilya.tocar@intel.com>
6548 Andrey Turetskiy <andrey.turetskiy@intel.com>
6549 Ilya Verbin <ilya.verbin@intel.com>
6550 Kirill Yukhin <kirill.yukhin@intel.com>
6551 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6553 * config/i386/predicates.md (register_or_constm1_operand): New.
6554 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
6555 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
6556 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
6557 UNSPEC_SCATTER_PREFETCH
6559 (avx512f_ucmp<mode>3): Ditto.
6560 (avx512f_vternlog<mode>): Ditto.
6561 (avx512f_align<mode>): Ditto.
6562 (<shift_insn><mode>3): Ditto.
6563 (avx512f_<rotate>v<mode>): Ditto.
6564 (avx512f_<rotate><mode>): Ditto.
6565 (avx512f_eq<mode>3): Ditto.
6566 (avx512f_eq<mode>3_1): Ditto.
6567 (avx512f_gt<mode>3): Ditto.
6568 (avx512f_testm<mode>3): Ditto.
6569 (avx512f_testnm<mode>3): Ditto.
6570 (avx512pf_gatherpf<mode>): Ditto.
6571 (*avx512pf_gatherpf<mode>_mask): Ditto.
6572 (*avx512pf_gatherpf<mode>): Ditto.
6573 (avx512pf_scatterpf<mode>): Ditto.
6574 (*avx512pf_scatterpf<mode>_mask): Ditto.
6575 (*avx512pf_scatterpf<mode>): Ditto.
6576 (avx512f_vec_dup_gpr<mode>): Ditto.
6577 (clz<mode>2): Ditto.
6578 (conflict<mode>): Ditto.
6579 (REDUC_SMINMAX_MODE): Extened with wider modes.
6580 (reduc_<code>_<mode>): Ditto.
6581 (vlshr<mode>3): Ditto.
6582 (vashl<mode>3): Ditto.
6584 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
6585 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6586 Sergey Lega <sergey.s.lega@intel.com>
6587 Anna Tikhonova <anna.tikhonova@intel.com>
6588 Ilya Tocar <ilya.tocar@intel.com>
6589 Andrey Turetskiy <andrey.turetskiy@intel.com>
6590 Ilya Verbin <ilya.verbin@intel.com>
6591 Kirill Yukhin <kirill.yukhin@intel.com>
6592 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6594 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
6597 (avx512fmaskmode): Ditto.
6598 (bcstscalarsuff): Ditto.
6599 (avx512f_blendm<mode>): Ditto.
6600 (cmp_imm_predicate): Ditto.
6601 (avx512f_cmp<mode>3): Ditto.
6602 (avx512f_vec_dup<mode>): Ditto.
6603 (avx512f_vec_dup_mem<mode>): Ditto.
6604 (avx512f_vpermi2var<mode>3): Ditto.
6605 (avx512f_vpermt2var<mode>3): Ditto.
6606 (vec_init<mode>): Ditto.
6607 (avx512f_gathersi<mode>): Ditto.
6608 (*avx512f_gathersi<mode>): Ditto.
6609 (*avx512f_gathersi<mode>_2): Ditto.
6610 (avx512f_gatherdi<mode>): Ditto.
6611 (*avx512f_gatherdi<mode>): Ditto.
6612 (*avx512f_gatherdi<mode>_2): Ditto.
6613 (avx512f_scattersi<mode>): Ditto.
6614 (*avx512f_scattersi<mode>): Ditto.
6615 (avx512f_scatterdi<mode>): Ditto.
6616 (*avx512f_scatterdi<mode>): Ditto.
6617 (sseintprefix): Extened with wider modes.
6618 (VEC_GATHER_IDXSI): Ditto.
6619 (VEC_GATHER_IDXDI): Ditto.
6620 (VEC_GATHER_SRCDI): Ditto.
6622 2013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
6623 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6625 * config/arm/t-aprofile: New file.
6626 * config.gcc: Handle --with-multilib-list option.
6628 2013-10-15 Bernd Schmidt <bernds@codesourcery.com>
6630 * reload1.c (reloads_unique_chain_p): Ensure that r1 is
6633 2013-10-15 Richard Biener <rguenther@suse.de>
6635 * tree-loop-distribution.c (build_empty_rdg): Inline into
6637 (rdg_flag_vertex): Inline into single user.
6638 (rdg_flag_vertex_and_dependent): Likewise.
6639 (build_rdg_partition_for_vertex): Remove processed bitmap.
6640 (rdg_build_partitions): Simplify.
6642 2013-10-15 Richard Biener <rguenther@suse.de>
6644 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
6645 Restructure forwarding through conversions and copies to
6646 avoid performing copy-propagation the wrong way. Adjust
6647 recursion invocations.
6648 (forward_propagate_addr_expr): Add argument stating if we
6649 are recursing from a single-use.
6650 (ssa_forward_propagate_and_combine): Adjust.
6652 2013-10-14 David Malcolm <dmalcolm@redhat.com>
6654 * dumpfile.h (gcc::dump_manager): New class, to hold state
6655 relating to dumpfile management.
6656 (get_dump_file_name): Remove in favor of method of dump_manager.
6657 (dump_initialized_p): Likewise.
6658 (dump_start): Likewise.
6659 (dump_finish): Likewise.
6660 (dump_switch_p): Likewise.
6661 (dump_register): Likewise.
6662 (get_dump_file_info): Likewise.
6663 * context.c (gcc::context::context): Construct the dump_manager
6665 * context.h (gcc::context::get_dumps): New.
6666 (gcc::context::m_dumps): New.
6667 * coverage.c (coverage_init): Port to dump_manager API.
6668 * dumpfile.c (extra_dump_files): Convert to field of gcc::dump_manager.
6669 (extra_dump_files_in_use): Likewise.
6670 (extra_dump_files_alloced): Likewise.
6671 (gcc::dump_manager::dump_manager): New.
6672 (dump_register): Convert to...
6673 (gcc::dump_manager::dump_register): ...method, replacing
6674 function-static next_dump with m_next_dump field.
6675 (get_dump_file_info): Convert to...
6676 (gcc::dump_manager::get_dump_file_info): ...method.
6677 (get_dump_file_name): Convert to...
6678 (gcc::dump_manager::get_dump_file_name): ...method.
6679 (dump_start): Convert to...
6680 (gcc::dump_manager::dump_start): ...method.
6681 (dump_finish): Convert to...
6682 (gcc::dump_manager::dump_finish): ...method.
6683 (dump_begin): Replace body with...
6684 (gcc::dump_manager::dump_begin): ...new method.
6685 (dump_phase_enabled_p): Convert to...
6686 (gcc::dump_manager::dump_phase_enabled_p): ...method.
6687 (dump_phase_enabled_p): Convert to...
6688 (gcc::dump_manager::dump_phase_enabled_p): ...method.
6689 (dump_initialized_p): Convert to...
6690 (gcc::dump_manager::dump_initialized_p): ...method.
6691 (dump_flag_name): Replace body with...
6692 (gcc::dump_manager::dump_flag_name): ...new method.
6693 (dump_enable_all): Convert to...
6694 (gcc::dump_manager::dump_enable_all): ...new method.
6695 (opt_info_enable_passes): Convert to...
6696 (gcc::dump_manager::opt_info_enable_passes): ...new method.
6697 (dump_switch_p_1): Convert to...
6698 (gcc::dump_manager::dump_switch_p_1): ...new method.
6699 (dump_switch_p): Convert to...
6700 (gcc::dump_manager::dump_switch_p): ...new method.
6701 (opt_info_switch_p): Port to dump_manager API.
6702 (enable_rtl_dump_file): Likewise.
6703 * opts-global.c (handle_common_deferred_options): Port to new
6705 * passes.c (pass_manager::finish_optimization_passes): Likewise.
6706 (pass_manager::register_one_dump_file): Likewise.
6707 (pass_manager::register_pass): Likewise.
6708 (pass_init_dump_file): Likewise.
6709 (pass_fini_dump_file): Likewise.
6710 * statistics.c (statistics_early_init): Likewise.
6712 2013-10-14 Richard Biener <rguenther@suse.de>
6714 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
6715 iterative_hash_canonical_type, gimple_canonical_type_hash,
6716 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
6717 gimple_register_canonical_type, print_gimple_types_stats,
6718 free_gimple_type_tables): Move to lto/lto.c
6719 (gt-gimple.h): Do not include.
6720 * gimple.h (gimple_register_canonical_type,
6721 print_gimple_types_stats, free_gimple_type_tables): Remove.
6722 * Makefile.in (GTFILES): Remove gimple.c.
6724 2013-10-14 Travis Snoozy <quandary@remstate.com>
6727 * config/msp430/msp430.c (msp430_option_override): Correct thinko
6728 scanning for msp430x targets.
6730 2013-10-14 Eric Botcazou <ebotcazou@adacore.com>
6733 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
6734 (registers_ok_for_ldd_peep): Move around.
6735 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
6736 * config/sparc/sparc.md (widening peepholes): Use it.
6738 2013-10-14 Richard Biener <rguenther@suse.de>
6742 * gimple.c (iterative_hash_canonical_type): Make sure to
6743 record the hash into the correct hashtable slot.
6745 2013-10-13 Eric Botcazou <ebotcazou@adacore.com>
6747 PR rtl-optimization/58662
6748 * combine.c (try_combine): Take into account death nodes on I2 when
6749 splitting a PARALLEL of two independent SETs. Fix dump message.
6751 2013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
6754 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
6755 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
6756 * config/sh/t-sh (sh_treg_combine.o): New entry.
6757 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
6758 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
6759 (register_sh_passes): New function. Register sh_treg_combine pass.
6760 (sh_option_override): Invoke it.
6761 (sh_canonicalize_comparison): Handle op0_preserve_value.
6762 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
6763 opportunities. Canonicalize branch condition.
6764 (nott): Allow only if pseudos can be created for non-SH2A.
6766 2013-10-12 H.J. Lu <hongjiu.lu@intel.com>
6769 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
6770 (ix86_expand_movmem): Replace copy_addr_to_reg with
6771 ix86_copy_addr_to_reg.
6772 (ix86_expand_setmem): Likewise.
6774 2013-10-12 Alexander Monakov <amonakov@ispras.ru>
6776 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
6777 provided by ix86_fp_compare_mode instead of CCFPUmode.
6779 2013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
6781 * config/aarch64/arm_neon.h
6782 (vtbx<1,3>_<psu>8): Fix register constriants.
6784 2013-10-11 Jeff Law <law@redhat.com>
6786 PR tree-optimization/58640
6787 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
6788 threading paths that cross over two loop entry points.
6790 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6792 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
6793 handle vector float as well.
6794 (*vsx_le_perm_load_v4si): Likewise.
6795 (*vsx_le_perm_store_v2di): Likewise.
6796 (*vsx_le_perm_store_v4si): Likewise.
6798 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6800 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
6801 directly to circumvent subtract from splat{31} workaround.
6802 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
6804 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
6805 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
6806 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
6807 (altivec_vperm_<mode>): Convert to define_insn_and_split to
6808 separate big and little endian logic.
6809 (*altivec_vperm_<mode>_internal): New define_insn.
6810 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
6811 separate big and little endian logic.
6812 (*altivec_vperm_<mode>_uns_internal): New define_insn.
6813 (vec_permv16qi): Add little endian logic.
6815 2013-10-11 Marc Glisse <marc.glisse@inria.fr>
6817 * doc/extend.texi (returns_nonnull): Remove arguments.
6819 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
6820 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6821 Sergey Lega <sergey.s.lega@intel.com>
6822 Anna Tikhonova <anna.tikhonova@intel.com>
6823 Ilya Tocar <ilya.tocar@intel.com>
6824 Andrey Turetskiy <andrey.turetskiy@intel.com>
6825 Ilya Verbin <ilya.verbin@intel.com>
6826 Kirill Yukhin <kirill.yukhin@intel.com>
6827 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6829 * config/i386/sse.md (VI48F_256_512): New.
6830 (avx2_permvar<mode>): Change to ...
6831 (<avx2_avx512f>_permvar<mode>): This.
6833 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
6834 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6835 Sergey Lega <sergey.s.lega@intel.com>
6836 Anna Tikhonova <anna.tikhonova@intel.com>
6837 Ilya Tocar <ilya.tocar@intel.com>
6838 Andrey Turetskiy <andrey.turetskiy@intel.com>
6839 Ilya Verbin <ilya.verbin@intel.com>
6840 Kirill Yukhin <kirill.yukhin@intel.com>
6841 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6843 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
6844 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
6845 * config/i386/sse.md (VI4_AVX): New.
6846 (sf2simodelower): Ditto.
6847 (sse2_cvtps2dq): Change to ...
6848 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
6850 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
6851 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6852 Sergey Lega <sergey.s.lega@intel.com>
6853 Anna Tikhonova <anna.tikhonova@intel.com>
6854 Ilya Tocar <ilya.tocar@intel.com>
6855 Andrey Turetskiy <andrey.turetskiy@intel.com>
6856 Ilya Verbin <ilya.verbin@intel.com>
6857 Kirill Yukhin <kirill.yukhin@intel.com>
6858 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6860 * config/i386/sse.md (V_512): New.
6862 (vcond<V_512:mode><VF_512:mode>): Ditto.
6863 (vcond<V_512:mode><VI_512:mode>): Ditto.
6864 (vcondu<V_512:mode><VI_512:mode>): Ditto.
6866 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
6867 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6868 Sergey Lega <sergey.s.lega@intel.com>
6869 Anna Tikhonova <anna.tikhonova@intel.com>
6870 Ilya Tocar <ilya.tocar@intel.com>
6871 Andrey Turetskiy <andrey.turetskiy@intel.com>
6872 Ilya Verbin <ilya.verbin@intel.com>
6873 Kirill Yukhin <kirill.yukhin@intel.com>
6874 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6876 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
6877 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
6879 (fma<mode>4): Removed condition.
6880 (fms<mode>4): Ditto.
6881 (fnma<mode>4): Ditto.
6882 (fnms<mode>4): Ditto.
6883 (fma4i_fmadd_<mode>): Ditto.
6884 (*fma_fmadd_<mode>): Ditto.
6885 (*fma_fmsub_<mode>): Ditto.
6886 (*fma_fnmadd_<mode>): Ditto.
6887 (*fma_fnmsub_<mode>): Ditto.
6888 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
6889 (*fma_fmaddsub_<mode>): Ditto.
6890 (*fma_fmsubadd_<mode>): Ditto.
6891 (*fmai_fmadd_<mode>): Ditto.
6892 (*fmai_fmsub_<mode>): Ditto.
6893 (*fmai_fnmadd_<mode>): Ditto.
6894 (*fmai_fnmsub_<mode>): Ditto.
6896 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
6897 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6898 Sergey Lega <sergey.s.lega@intel.com>
6899 Anna Tikhonova <anna.tikhonova@intel.com>
6900 Ilya Tocar <ilya.tocar@intel.com>
6901 Andrey Turetskiy <andrey.turetskiy@intel.com>
6902 Ilya Verbin <ilya.verbin@intel.com>
6903 Kirill Yukhin <kirill.yukhin@intel.com>
6904 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6906 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
6907 (VI124_256): Changed to ...
6908 (VI124_256_48_512): This.
6909 (ssepackmode): Extended with wider modes.
6910 (<code><mode>3): Changed iterator.
6911 (*avx2_<code><mode>3): Ditto.
6912 (vec_pack_trunc_<mode>): Ditto.
6914 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
6915 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6916 Sergey Lega <sergey.s.lega@intel.com>
6917 Anna Tikhonova <anna.tikhonova@intel.com>
6918 Ilya Tocar <ilya.tocar@intel.com>
6919 Andrey Turetskiy <andrey.turetskiy@intel.com>
6920 Ilya Verbin <ilya.verbin@intel.com>
6921 Kirill Yukhin <kirill.yukhin@intel.com>
6922 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6924 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
6925 (VI8F_256_512): Ditto.
6926 (abs<mode>2): Changed iterator.
6927 (avx2_perm<mode>): Changed to ...
6928 (<avx2_avx512f>_perm<mode>): This.
6929 (avx2_perm<mode>_1): Changed to ...
6930 (<avx2_avx512f>_perm<mode>_1): This.
6932 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
6933 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6934 Sergey Lega <sergey.s.lega@intel.com>
6935 Anna Tikhonova <anna.tikhonova@intel.com>
6936 Ilya Tocar <ilya.tocar@intel.com>
6937 Andrey Turetskiy <andrey.turetskiy@intel.com>
6938 Ilya Verbin <ilya.verbin@intel.com>
6939 Kirill Yukhin <kirill.yukhin@intel.com>
6940 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6942 * config/i386/sse.md (VI48_AVX512F): New.
6943 (VI48_AVX2): Changed to ...
6944 (VI48_AVX2_48_AVX512F): This.
6945 (avx2_ashrv<mode>): Changed to ...
6946 (<avx2_avx512f>_ashrv<mode>): This.
6947 (avx2_<shift_insn>v<mode>): Changed to ...
6948 (<avx2_avx512f>_<shift_insn>v<mode>): This.
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/sse.md (VI4_AVX512F): New.
6961 (VI8_AVX2_AVX512F): Ditto.
6962 (mul<mode>3): Extended with wider modes.
6963 (*<sse4_1_avx2>_mul<mode>3): Ditto.
6964 (mul<mode>3): Ditto.
6965 (vec_widen_<s>mult_odd_<mode>): Ditto.
6967 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
6968 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6969 Sergey Lega <sergey.s.lega@intel.com>
6970 Anna Tikhonova <anna.tikhonova@intel.com>
6971 Ilya Tocar <ilya.tocar@intel.com>
6972 Andrey Turetskiy <andrey.turetskiy@intel.com>
6973 Ilya Verbin <ilya.verbin@intel.com>
6974 Kirill Yukhin <kirill.yukhin@intel.com>
6975 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6977 * config/i386/sse.md (VI2_AVX512F): New.
6978 (VI124_AVX512F): Ditto.
6979 (sseunpackmode): Extended with wider modes.
6980 (sseunpackfltmode): Ditto.
6981 (vec_unpacks_float_hi_<mode>): Ditto.
6982 (vec_unpacks_float_lo_<mode>): Ditto.
6983 (vec_unpacku_float_hi_<mode>): Ditto.
6984 (vec_unpacku_float_lo_<mode>): Ditto.
6985 (vec_unpacks_lo_<mode>): Ditto.
6986 (vec_unpacks_hi_<mode>): Ditto.
6987 (vec_unpacku_lo_<mode>): Ditto.
6988 (vec_unpacku_hi_<mode>): Ditto.
6990 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
6991 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6992 Sergey Lega <sergey.s.lega@intel.com>
6993 Anna Tikhonova <anna.tikhonova@intel.com>
6994 Ilya Tocar <ilya.tocar@intel.com>
6995 Andrey Turetskiy <andrey.turetskiy@intel.com>
6996 Ilya Verbin <ilya.verbin@intel.com>
6997 Kirill Yukhin <kirill.yukhin@intel.com>
6998 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
7000 * config/i386/i386.md (multdiv): New.
7001 (multdiv_mnemonic): Ditto.
7002 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
7003 (<sse>_vm<multdiv_mnemonic><mode>3): This.
7004 (<sse>_vmdiv<mode>3): Removed.
7006 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
7007 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
7008 Sergey Lega <sergey.s.lega@intel.com>
7009 Anna Tikhonova <anna.tikhonova@intel.com>
7010 Ilya Tocar <ilya.tocar@intel.com>
7011 Andrey Turetskiy <andrey.turetskiy@intel.com>
7012 Ilya Verbin <ilya.verbin@intel.com>
7013 Kirill Yukhin <kirill.yukhin@intel.com>
7014 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
7016 * config/i386/sse.md (V): Extended with wider modes.
7018 (ssehalfvecmode): Ditto.
7020 (ssepackfltmode): Ditto.
7021 (avx_vec_concat<mode>): Ditto.
7022 (V_256_512): New iterator.
7023 (VF2_512_256): Ditto.
7024 (si2dfmode): New attribute.
7025 (si2dfmodelower): Ditto.
7027 (concat_tg_mode): Ditto.
7028 (floatv4siv4df2): Changed to ...
7029 (float<si2dfmodelower><mode>2): This.
7030 (avx_cvtps2pd256): Changed to ...
7031 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
7032 (vec_pack_trunc_v4df): Changed to ...
7033 (vec_pack_trunc_<mode>): This.
7034 (avx_vpermil<mode>): Changed to ...
7035 (<sse2_avx_avx512f>_vpermil<mode>): This.
7036 (<fixsuffix>fix_truncv8dfv8si2): New.
7037 (vec_pack_sfix_trunc_v8df): Ditto.
7038 (avx512f_rndscale<mode>): Ditto.
7039 (avx512f_roundpd512): Ditto.
7040 (vec_pack_ufix_trunc_<mode>): Updated iterator.
7042 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
7043 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
7044 Sergey Lega <sergey.s.lega@intel.com>
7045 Anna Tikhonova <anna.tikhonova@intel.com>
7046 Ilya Tocar <ilya.tocar@intel.com>
7047 Andrey Turetskiy <andrey.turetskiy@intel.com>
7048 Ilya Verbin <ilya.verbin@intel.com>
7049 Kirill Yukhin <kirill.yukhin@intel.com>
7050 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
7052 * config/i386/i386.md (any_fix): New iterator.
7053 (fixsuffix): New attribute.
7054 * config/i386/sse.md (VF1): Extened with wider modes.
7058 (sseintvecmodelower): Ditto.
7059 (ssescalarmode): Ditto.
7060 (ssescalarnum): Ditto.
7062 (ssexmmmode): Ditto.
7063 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
7064 (<sse>_rcp<mode>2): Change iterator.
7065 (rsqrt<mode>2): Ditto.
7066 (<sse>_rsqrt<mode>2): Ditto.
7067 (avx2_vec_dup<mode>): Ditto.
7068 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
7069 (round<mode>2_sfix): Ditto.
7070 (avx2_pbroadcast<mode>): Ditto.
7071 (*andnot<mode>3): Handle XI mode.
7072 (*<code><mode>3): Ditto.
7073 (AVXTOSSEMODE): Removed.
7074 (avx_vpermil<mode>): Changed to ...
7075 (<sse2_avx_avx512f>_vpermil<mode>): This.
7077 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
7078 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
7079 Sergey Lega <sergey.s.lega@intel.com>
7080 Anna Tikhonova <anna.tikhonova@intel.com>
7081 Ilya Tocar <ilya.tocar@intel.com>
7082 Andrey Turetskiy <andrey.turetskiy@intel.com>
7083 Ilya Verbin <ilya.verbin@intel.com>
7084 Kirill Yukhin <kirill.yukhin@intel.com>
7085 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
7087 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
7088 (<sse>_comi): Ditto.
7089 (<sse>_ucomi): Ditto.
7090 (sse_cvtss2siq_2): Ditto.
7091 (sse2_cvtsd2si): Ditto.
7092 (sse2_cvtsd2siq): Ditto.
7093 (sse2_cvttsd2si): Ditto.
7094 (sse2_cvttsd2siq): Ditto.
7095 (<shift_insn><mode>3): Ditto.
7096 (sse2_cvtsi2sdq): Update constraint and prefix.
7097 (sse_cvtsi2ss): Update prefix.
7098 (sse_cvtsi2ssq): Ditto.
7100 2013-10-11 Jakub Jelinek <jakub@redhat.com>
7102 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
7103 ignore internal calls.
7105 2013-10-11 Richard Biener <rguenther@suse.de>
7107 * tree-pretty-print.c (dump_generic_node): Allow to dump both (D)
7108 and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with (OVF) if
7109 TREE_OVERFLOW is set.
7111 2013-10-11 Thomas Schwinge <thomas@codesourcery.com>
7113 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
7115 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
7116 * gimple.h: Likewise.
7118 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
7120 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
7123 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
7125 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
7126 * configure: Regenerate.
7128 2013-10-11 Jakub Jelinek <jakub@redhat.com>
7130 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
7131 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
7132 formatting fixes, use pp_colon instead of pp_character (..., ':'),
7133 similarly pp_right_paren.
7134 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
7135 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
7136 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
7137 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
7139 (omp_declare_simd_clauses_equal,
7140 omp_remove_redundant_declare_simd_attrs): New functions.
7141 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
7142 (walk_tree_1): Handle new OpenMP 4.0 clauses.
7143 * tree.h (OMP_LOOP_CHECK): Define.
7144 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
7145 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
7146 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
7147 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
7148 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
7149 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
7150 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
7151 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
7152 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
7153 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
7154 OMP_CLAUSE_SIMDLEN_EXPR): Define.
7155 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
7156 (omp_remove_redundant_declare_simd_attrs): New prototype.
7157 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
7158 GIMPLE_OMP_TEAMS): New codes.
7159 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
7160 * omp-low.c (struct omp_context): Add cancel_label and cancellable
7162 (target_nesting_level): New variable.
7163 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
7164 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
7165 for collapse > 1 static schedule unless ordered.
7166 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
7167 (determine_parallel_type): Adjust get_ws_args_for caller.
7168 (install_var_field): Handle mask & 4 for double indirection.
7169 (scan_sharing_clauses): Ignore shared clause on teams construct.
7170 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
7171 (create_omp_child_function): If inside target or declare target
7172 constructs, set "omp declare target" attribute on the child function.
7173 (find_combined_for): New function.
7174 (scan_omp_parallel): Handle combined loops.
7175 (scan_omp_target, scan_omp_teams): New functions.
7176 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
7177 restrictions and set ctx->cancellable for cancellable constructs.
7178 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
7179 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
7180 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
7181 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
7182 (omp_clause_aligned_alignment): New function.
7183 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR on decls.
7184 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
7185 on teams constructs. Handle user defined reductions and new
7187 (lower_reduction_clauses): Don't set placeholder to address of ref
7188 if it has already the right type.
7189 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
7190 (expand_parallel_call): Use the new non-_start suffixed builtins,
7191 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
7192 and GOMP_parallel_end after the call.
7193 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
7194 (expand_omp_for_init_counts): Handle combined loops.
7195 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
7197 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
7198 end of cancellable loops.
7199 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
7200 Likewise. Handle collapse > 1 loops.
7201 (expand_omp_simd): Handle combined loops.
7202 (expand_omp_for): Add inner_stmt argument, adjust callers of
7203 expand_omp_for* functions, use expand_omp_for_static*chunk even
7204 for collapse > 1 unless ordered.
7205 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
7206 of cancellable sections.
7207 (expand_omp_single): Remove need_barrier variable, just rely on
7208 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
7209 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
7210 (expand_omp_atomic_load, expand_omp_atomic_store,
7211 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
7212 (expand_omp_target): New function.
7213 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
7214 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
7215 (build_omp_regions_1): Immediately close region for
7216 GF_OMP_TARGET_KIND_UPDATE.
7217 (maybe_add_implicit_barrier_cancel): New function.
7218 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
7220 (lower_omp_single): Likewise. Add clobber after the barrier.
7221 (lower_omp_taskgroup): New function.
7222 (lower_omp_for): Handle combined loops. Adjust
7223 lower_rec_input_clauses caller. Handle cancellation.
7224 (lower_depend_clauses): New function.
7225 (lower_omp_taskreg): Lower depend clauses. Adjust
7226 lower_rec_input_clauses caller. Add clobber after the call. Handle
7228 (lower_omp_target, lower_omp_teams): New functions.
7229 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
7230 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
7231 and GOMP_cancellation_point calls.
7232 (lower_omp): Fold stmts inside of target region.
7233 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
7234 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
7235 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
7236 (BT_FN_VOID_OMPFN_PTR_UINT,
7237 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
7238 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
7239 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
7240 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
7241 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
7242 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
7243 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
7244 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
7245 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
7246 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
7247 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
7248 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
7249 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
7250 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
7251 BUILT_IN_GOMP_PARALLEL_END.
7252 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
7253 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
7254 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
7255 GF_OMP_FOR_KIND_DISTRIBUTE.
7256 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
7257 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
7258 (dump_gimple_omp_return): Print lhs if it has any.
7259 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
7260 gimple_omp_atomic_seq_cst_p.
7261 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
7262 and GIMPLE_OMP_TEAMS.
7263 * langhooks.c (lhd_omp_mappable_type): New function.
7264 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
7265 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
7267 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
7268 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
7269 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and ORT_TARGET.
7270 (struct gimplify_omp_ctx): Add combined_loop field.
7271 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
7272 on stmts inside of target region.
7273 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
7274 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
7275 ORT_TARGET and ORT_TARGET_DATA.
7276 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
7278 (omp_notice_threadprivate_variable): Complain about threadprivate
7279 variables in target region.
7280 (omp_notice_variable): Complain about vars with non-mappable type
7281 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
7282 (omp_check_private): Ignore ORT_TARGET* regions.
7283 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
7284 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
7285 (find_combined_omp_for): New function.
7286 (gimplify_omp_for): Handle gimplification of combined loops.
7287 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
7289 (gimplify_omp_target_update): New function.
7290 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
7291 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
7292 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
7293 (gimplify_body): If fndecl has "omp declare target" attribute, add
7294 implicit ORT_TARGET context around it.
7295 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
7296 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
7297 * tree-nested.c (convert_nonlocal_reference_stmt,
7298 convert_local_reference_stmt, convert_gimple_call): Handle
7299 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
7300 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
7301 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
7302 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
7303 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
7304 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
7305 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
7306 BUILT_IN_GOMP_LOOP_END_CANCEL,
7307 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
7308 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
7309 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
7310 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
7311 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
7312 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
7313 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
7314 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
7315 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
7316 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
7317 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
7318 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
7319 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
7320 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
7321 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
7322 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
7323 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
7324 gimple_build_omp_teams): New functions.
7325 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
7326 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
7327 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
7328 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
7329 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
7330 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
7331 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
7332 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
7333 GF_OMP_ATOMIC_SEQ_CST): New.
7334 (gimple_build_omp_taskgroup, gimple_build_omp_target,
7335 gimple_build_omp_teams): New prototypes.
7336 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
7337 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
7338 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
7339 GIMPLE_OMP_SINGLE as end of range.
7340 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
7341 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
7342 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
7343 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
7344 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
7345 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
7346 gimple_omp_target_kind, gimple_omp_target_set_kind,
7347 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
7348 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
7349 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
7350 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
7351 gimple_omp_teams_set_clauses): New inlines.
7352 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
7353 and GIMPLE_OMP_TASKGROUP.
7354 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
7355 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
7356 enum omp_clause_proc_bind_kind): New.
7357 (union omp_clause_subcode): Add depend_kind, map_kind and
7358 proc_bind_kind fields.
7359 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
7360 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
7361 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
7362 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
7363 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
7365 2013-10-10 Teresa Johnson <tejohnson@google.com>
7367 * predict.c (tree_estimate_probability): Add new parameter
7368 for estimate_bb_frequencies.
7369 (estimate_bb_frequencies): Add new parameter to force estimation.
7370 (rebuild_frequencies): When max frequency in function is small,
7371 recompute counts from frequencies.
7372 * predict.h (estimate_bb_frequencies): New parameter.
7374 2013-10-10 David Malcolm <dmalcolm@redhat.com>
7376 * ipa-inline.c (ipa_inline): Fix leak of "order" when
7377 optimizations are disabled.
7379 2013-10-10 David Malcolm <dmalcolm@redhat.com>
7381 * coverage.c (coverage_finish): Fix leak of da_file_name.
7383 2013-10-10 Jan Hubicka <jh@suse.cz>
7385 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
7386 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
7389 2013-10-10 Jakub Jelinek <jakub@redhat.com>
7392 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
7393 if any labels are in FALLTHRU_BB, use a special label emitted
7394 immediately after the asm goto insn rather than label_rtx
7396 (expand_asm_stmt): Adjust caller.
7397 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
7398 edge if the last insn in predecessor is a jump with single successor,
7399 but it isn't simplejump_p.
7401 2013-10-10 Richard Biener <rguenther@suse.de>
7403 PR tree-optimization/58656
7404 * tree-ssa-pre.c (phi_translate): Do not cache failed translations.
7406 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
7408 * gimplify.c: Include expr.h and tm_p.h for targets with special
7409 va-arg padding requirements.
7411 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
7413 * tree-flow.h: Move some prototypes to gimple.h.
7414 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
7415 * gimple.h: Relocate some prototypes from tree-flow.h
7416 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
7418 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
7419 (build_va_arg_indirect_ref): Relocate and make static.
7420 (std_gimplify_va_arg_expr): Relocate here.
7421 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
7422 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
7424 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7426 * doc/md.texi: Document the mnemonic attribute.
7428 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7431 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
7432 (set_attr_alternative x ...) when searching for user defined
7435 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
7437 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
7439 2013-10-09 Easwaran Raman <eraman@google.com>
7441 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
7442 * cfgexpand.c (defer_stack_allocation): ...use here
7443 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
7445 2013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
7447 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
7448 (operand_equal_for_value_replacement): New function, extracted from
7449 value_replacement and enhanced to catch more cases.
7450 (value_replacement): Use operand_equal_for_value_replacement.
7452 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
7454 * loop-doloop.c (doloop_modify, doloop_optimize): Use
7455 get_max_loop_iterations.
7457 2013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7459 * config/arm/aarch-common.c (arm_early_load_addr_dep):
7460 Place comment above function.
7462 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
7464 * tree-flow.h: Remove all remaining prototypes, enums and structs that
7465 are not related to tree-cfg.c.
7466 * tree-ssa-address.h: New file. Relocate prototypes.
7467 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
7468 (addr_for_mem_ref): New. Combine call to get_address_description and
7469 return addr_for_mem_ref.
7470 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
7471 * tree-ssa-live.h: Adjust prototypes.
7472 * passes.c: Include tree-ssa-live.h.
7473 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
7474 * graphite.c (graphite_transform_loops): Make static.
7475 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
7476 make_pass_graphite, pass_data_graphite_transforms,
7477 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
7478 * ipa-pure-const.c (warn_function_noreturn): Make static.
7479 (execute_warn_function_noreturn, gate_warn_function_noreturn,
7480 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
7481 Relocate from tree-cfg.c
7482 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
7484 (execute_warn_function_noreturn, gate_warn_function_noreturn,
7485 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
7486 Move to ipa-pure-const.c.
7487 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
7488 Relocate from tree-optimize.c.
7489 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
7490 make_pass_fixup_cfg): Move to tree-cfg.c.
7491 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
7492 Relocate some prototypes.
7493 * tree-data-ref.h (tree_check_data_deps) Add prototype.
7494 * tree-dump.c (dump_function_to_file): Remove prototype.
7495 Add tree-flow.h to the include file.
7496 * tree-dump.h: Remove prototype.
7497 * tree-parloops.h: New File. Add prototypes.
7498 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
7499 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
7500 from tree-ssa-loop.c.
7501 * tree-predcom.c (run_tree_predictive_commoning,
7502 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
7503 Relocate here from tree-ssa-loop.c.
7504 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
7505 ssa_name_values.release ().
7506 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
7507 (ssa_name_values): Relocate from tree-flow.h.
7508 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
7509 * tree-ssa-loop.c (run_tree_predictive_commoning,
7510 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
7511 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
7512 make_pass_graphite, pass_data_graphite_transforms,
7513 make_pass_graphite_transforms, gate_tree_parallelize_loops,
7514 tree_parallelize_loops, pass_data_parallelize_loops,
7515 make_pass_parallelize_loops): Move to other files.
7516 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
7518 * tree.h: Remove prototypes from tree-address.c.
7520 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
7522 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
7523 (struct int_tree_map): Move to tree-hasher.h
7524 (SCALE, LABEL, PERCENT): Move to gimple.h
7525 * tree-flow-inline.h: Delete. Move functions to other files.
7526 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
7527 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
7528 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
7529 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
7530 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
7531 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
7532 * tree-hasher.h: Don't include tree-flow.h.
7533 (struct int_tree_map): Relocate from tree-flow.h.
7534 * tree-sra.c (contains_view_convert_expr_p): Relocate from
7535 tree-flow-inline.h and make static.
7536 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
7538 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
7539 tree-flow-inline.h and make static.
7540 * tree.h (is_global_var, may_be_aliased): Relocate from
7542 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
7543 * value-prof.c: No longer include tree-flow-inline.h.
7544 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
7546 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
7548 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
7549 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
7550 (enum move_pos): Move to tree-ssa-loop-im.h
7551 * cfgloop.h: Move some prototypes.
7552 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
7553 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
7554 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
7555 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
7556 (loop_containing_stmt): Relocate from tree-flow-inline.h.
7557 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
7558 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
7559 (enum move_pos): Relocate here.
7560 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
7562 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
7563 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
7564 make_pass_lim): Relocate here from tree-ssa-loop.c.
7565 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
7567 (loop_edge_to_cancel, unloop_loops): Make static.
7568 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
7569 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
7570 (tree_complete_unroll, gate_tree_complete_unroll,
7571 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
7572 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
7573 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
7574 * tree-ssa-loop-ivopts.c: Remove local prototypes.
7575 (stmt_invariant_in_loop_p): Remove unused function.
7576 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
7577 * tree-ssa-loop-manip.h: New file. Add prototypes.
7578 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
7579 (gcov_type_to_double_int): Move to cfgloop.h.
7580 (double_int_cmp, bound_index,
7581 estimate_numbers_of_iterations_loop): Make static.
7582 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
7583 (max_loop_iterations): Factor out get_max_loop_iterations.
7584 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
7586 * tree-ssa-loop-niter.h: New file. Add prototypes.
7587 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
7588 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
7589 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
7590 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
7591 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
7592 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
7593 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
7594 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
7595 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
7596 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
7597 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
7598 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
7599 pass_data_complete_unroll, make_pass_complete_unroll,
7600 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
7601 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
7602 tree-ssa-loop-ivcanon.c.
7603 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
7604 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
7605 tree-ssa-loop-prefetch.c.
7606 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
7608 (get_lsm_tmp_name): Relocate and add suffix parameter.
7609 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
7610 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
7611 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
7612 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
7613 (get_estimated_loop_iterations): Factor out accessor from
7614 estimated_loop_iterations in tree-ssa-loop-niter.c.
7615 (get_max_loop_iterations): Factor out accessor from
7616 _max_loop_iterations in tree-ssa-niter.c.
7617 * loop-unroll.c (decide_unroll_constant_iterations,
7618 decide_unroll_runtime_iterations, decide_peel_simple,
7619 decide_unroll_stupid): Use new get_* accessors.
7621 2013-10-09 Marc Glisse <marc.glisse@inria.fr>
7623 PR tree-optimization/20318
7624 * doc/extend.texi (returns_nonnull): New function attribute.
7625 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
7627 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
7628 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
7630 2013-10-09 Eric Botcazou <ebotcazou@adacore.com>
7633 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
7634 false if both components are bitfields.
7636 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
7638 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
7639 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
7640 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
7641 * config/aarch64/aarch64.h
7642 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
7643 * config/aarch64/aarch64-simd-builtins.def
7644 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
7646 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
7648 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
7650 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
7652 * config/aarch64/arm_neon.h (vdiv_f64): Added.
7654 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
7656 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
7657 (vneg_f64): New intrinsic.
7658 (vneg_s8): Asm replaced with C.
7659 (vneg_s16): Likewise.
7660 (vneg_s32): Likewise.
7661 (vneg_s64): New intrinsic.
7662 (vnegq_f32): Asm replaced with C.
7663 (vnegq_f64): Likewise.
7664 (vnegq_s8): Likewise.
7665 (vnegq_s16): Likewise.
7666 (vnegq_s32): Likewise.
7667 (vnegq_s64): Likewise.
7669 2013-10-09 Renlin Li <Renlin.Li@arm.com>
7671 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
7673 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7675 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
7676 packed stack special handling.
7677 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
7678 back to fixed stack slots for FPRs saved due to stdarg.
7680 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7682 * config/s390/s390.c (s390_frame_info): Restructure function.
7684 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7686 * config/s390/s390.c (struct s390_frame_layout): New field
7688 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
7689 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
7690 regs_ever_clobbered to char*.
7691 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
7692 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
7693 (s390_register_info): Enable FPR save slots. Move/Copy some
7694 functionality into ...
7695 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
7696 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
7698 (s390_frame_info): Do gpr slot allocation here now. stdarg does
7699 not imply a stack frame.
7700 (s390_init_frame_layout): Remove variable clobbered_regs.
7701 (s390_update_register_info): Remove function.
7702 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
7703 cannot be used for register renaming.
7704 (s390_hard_regno_scratch_ok): New function.
7705 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
7706 (s390_initial_elimination_offset): Change offset calculation of
7707 the return address pointer.
7708 (save_gprs): Deal with only r6 being saved from the call-saved regs.
7709 (restore_gprs): Set frame related flag.
7710 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
7711 (s390_emit_prologue): Call s390_register_info instead of
7712 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
7713 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
7714 (s390_optimize_prologue): Call s390_optimize_register_info.
7715 Try to remove also FPR slot save/restore INSNs. Remove frame
7716 related flags from restore INSNs.
7718 2013-10-08 DJ Delorie <dj@redhat.com>
7720 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
7723 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
7724 avoid conflict with the MI use of %c.
7725 * config/rl78/rl78-real.md: change %c to %C throughout.
7726 * config/rl78/rl78-virt.md: Likewise.
7728 2013-10-08 Jan Hubicka <jh@suse.cz>
7730 * config/i386/i386.c (ix86_option_override_internal): Switch
7731 to SSE math for -ffast-math when target ISA supports SSE2.
7733 2013-10-08 Andrew MacLeod <amacleod@redhat.com>
7735 * tree-flow.h: Remove some prototypes.
7736 * tree.h: Remove some protypes, add a couple.
7737 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
7738 using_eh_for_cleanups_p): Add interface routines for front ends.
7739 * tree-eh.h: New file. Add protoptyes.
7740 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
7741 (add_stmt_to_eh_lp_fn): Make static.
7742 (lower_try_finally): Use new using_eh_for_cleanups_p.
7743 * emit-rtl.c: Include tree-eh.h.
7744 * gimple.h: Include tree-eh.h.
7746 2013-10-08 Marc Glisse <marc.glisse@inria.fr>
7748 PR tree-optimization/58480
7749 * tree-vrp.c (infer_nonnull_range): New function.
7750 (infer_value_range): Call infer_nonnull_range.
7752 2013-10-08 Dehao Chen <dehao@google.com>
7754 PR tree-optimization/58619
7755 * tree-inline.c (copy_phis_for_bb): Combine location data
7758 2013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
7761 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
7762 RTX_FRAME_RELATED_P on INSN.
7764 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7766 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
7767 (altivec_expand_vec_perm_const): Call it.
7769 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7771 * config/rs6000/vector.md (mov<mode>): Emit permuted move
7772 sequences for LE VSX loads and stores at expand time.
7773 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
7775 * config/rs6000/rs6000.c (rs6000_const_vec): New.
7776 (rs6000_gen_le_vsx_permute): New.
7777 (rs6000_gen_le_vsx_load): New.
7778 (rs6000_gen_le_vsx_store): New.
7779 (rs6000_gen_le_vsx_move): New.
7780 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
7781 (*vsx_le_perm_load_v4si): New.
7782 (*vsx_le_perm_load_v8hi): New.
7783 (*vsx_le_perm_load_v16qi): New.
7784 (*vsx_le_perm_store_v2di): New.
7785 (*vsx_le_perm_store_v4si): New.
7786 (*vsx_le_perm_store_v8hi): New.
7787 (*vsx_le_perm_store_v16qi): New.
7788 (*vsx_xxpermdi2_le_<mode>): New.
7789 (*vsx_xxpermdi4_le_<mode>): New.
7790 (*vsx_xxpermdi8_le_V8HI): New.
7791 (*vsx_xxpermdi16_le_V16QI): New.
7792 (*vsx_lxvd2x2_le_<mode>): New.
7793 (*vsx_lxvd2x4_le_<mode>): New.
7794 (*vsx_lxvd2x8_le_V8HI): New.
7795 (*vsx_lxvd2x16_le_V16QI): New.
7796 (*vsx_stxvd2x2_le_<mode>): New.
7797 (*vsx_stxvd2x4_le_<mode>): New.
7798 (*vsx_stxvd2x8_le_V8HI): New.
7799 (*vsx_stxvd2x16_le_V16QI): New.
7801 2013-10-07 Renlin Li <Renlin.Li@arm.com>
7803 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
7805 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7807 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
7808 loop to work also for 31bit ABI.
7809 Save the stack pointer for frame_size > 0.
7811 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7813 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
7814 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
7815 constraint letters from expanders.
7816 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
7817 retry count to general_operand.
7818 ("tabort"): Give operand 0 a mode.
7819 ("tabort_1"): Add mode and constraint letter for operand 0.
7820 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
7822 2013-10-04 Jeff Law <law@redhat.com>
7824 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
7826 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
7828 (thread_across_edge): Here. Call it.
7830 2013-10-04 Cary Coutant <ccoutant@google.com>
7832 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
7833 discarding a location list expression (or a piece of one).
7835 2013-10-03 Jan Hubicka <jh@suse.cz>
7837 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
7838 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
7839 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
7841 2013-10-03 Jan Hubicka <jh@suse.cz>
7843 * config/i386/i386.c (ix86_option_override_internal): Do not enable
7844 accumulate-outgoing-args when producing unwind info.
7846 2013-10-03 Wei Mi <wmi@google.com>
7848 * lra-constraints.c (insert_move_for_subreg): New function
7849 extracted from simplify_operand_subreg.
7850 (simplify_operand_subreg): Add reload for paradoxical subreg.
7852 2013-10-03 Rong Xu <xur@google.com>
7854 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
7855 the candidate of builtin_expect such that we should fix the
7856 size/time estimation.
7857 (estimate_function_body_sizes): Do the acutally size/time fix-up
7860 2013-10-03 Rong Xu <xur@google.com>
7862 * predict.c (tree_predict_by_opcode): Get the probability
7863 for builtin_expect from param builtin_expect_probability.
7864 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
7865 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
7866 * doc/invoke.texi: Add documentation for builtin-expect-probability.
7868 2013-10-03 Marc Glisse <marc.glisse@inria.fr>
7871 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
7872 * calls.c (alloca_call_p): Use get_callee_fndecl.
7873 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
7874 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
7876 (vrp_visit_stmt): Remove duplicated code.
7878 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
7880 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
7881 ceildf2, btruncdf2, instead of vsx_* name.
7883 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
7884 iterators to only do V2DF and V4SF here. Move the DF code to
7885 rs6000.md where it is combined with SF mode. Replace <VSv> with
7886 just 'v' since only vector operations are handled with these insns
7887 after moving the DF support to rs6000.md.
7888 (vsx_sub<mode>3): Likewise.
7889 (vsx_mul<mode>3): Likewise.
7890 (vsx_div<mode>3): Likewise.
7891 (vsx_fre<mode>2): Likewise.
7892 (vsx_neg<mode>2): Likewise.
7893 (vsx_abs<mode>2): Likewise.
7894 (vsx_nabs<mode>2): Likewise.
7895 (vsx_smax<mode>3): Likewise.
7896 (vsx_smin<mode>3): Likewise.
7897 (vsx_sqrt<mode>2): Likewise.
7898 (vsx_rsqrte<mode>2): Likewise.
7899 (vsx_fms<mode>4): Likewise.
7900 (vsx_nfma<mode>4): Likewise.
7901 (vsx_copysign<mode>3): Likewise.
7902 (vsx_btrunc<mode>2): Likewise.
7903 (vsx_floor<mode>2): Likewise.
7904 (vsx_ceil<mode>2): Likewise.
7905 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
7906 (vsx_sminsf3): Likewise.
7907 (vsx_fmadf4): Likewise.
7908 (vsx_fmsdf4): Likewise.
7909 (vsx_nfmadf4): Likewise.
7910 (vsx_nfmsdf4): Likewise.
7911 (vsx_cmpdf_internal1): Likewise.
7913 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
7914 simpler to select whether a target has SPE or traditional floating
7915 point support in iterators.
7916 (TARGET_DF_SPE): Likewise.
7917 (TARGET_SF_FPR): Likewise.
7918 (TARGET_DF_FPR): Likewise.
7919 (TARGET_SF_INSN): Macros to say whether floating point support
7920 exists for a given operation for expanders.
7921 (TARGET_DF_INSN): Likewise.
7923 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
7924 combining of SF/DF mode operations, using both traditional and VSX
7932 (abs<mode>2): Combine SF/DF modes using traditional floating point
7933 instructions. Add support for using the upper DF registers with
7934 VSX support, and SF registers with power8-vector support. Update
7935 expanders for operations supported by both the SPE and traditional
7936 floating point units.
7937 (abs<mode>2_fpr): Likewise.
7938 (nabs<mode>2): Likewise.
7939 (nabs<mode>2_fpr): Likewise.
7940 (neg<mode>2): Likewise.
7941 (neg<mode>2_fpr): Likewise.
7942 (add<mode>3): Likewise.
7943 (add<mode>3_fpr): Likewise.
7944 (sub<mode>3): Likewise.
7945 (sub<mode>3_fpr): Likewise.
7946 (mul<mode>3): Likewise.
7947 (mul<mode>3_fpr): Likewise.
7948 (div<mode>3): Likewise.
7949 (div<mode>3_fpr): Likewise.
7950 (sqrt<mode>3): Likewise.
7951 (sqrt<mode>3_fpr): Likewise.
7952 (fre<Fs>): Likewise.
7953 (rsqrt<mode>2): Likewise.
7954 (cmp<mode>_fpr): Likewise.
7955 (smax<mode>3): Likewise.
7956 (smin<mode>3): Likewise.
7957 (smax<mode>3_vsx): Likewise.
7958 (smin<mode>3_vsx): Likewise.
7959 (negsf2): Delete SF operations that are merged with DF.
7966 (fmasf4_fpr): Likewise.
7967 (fmssf4_fpr): Likewise.
7968 (nfmasf4_fpr): Likewise.
7969 (nfmssf4_fpr): Likewise.
7970 (sqrtsf2): Likewise.
7971 (rsqrtsf_internal1): Likewise.
7972 (smaxsf3): Likewise.
7973 (sminsf3): Likewise.
7974 (cmpsf_internal1): Likewise.
7975 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
7976 (negdf2): Delete DF operations that are merged with SF.
7978 (nabsdf2): Likewise.
7984 (rsqrtdf_internal1): Likewise.
7985 (fmadf4_fpr): Likewise.
7986 (fmsdf4_fpr): Likewise.
7987 (nfmadf4_fpr): Likewise.
7988 (nfmsdf4_fpr): Likewise.
7989 (sqrtdf2): Likewise.
7990 (smaxdf3): Likewise.
7991 (smindf3): Likewise.
7992 (cmpdf_internal1): Likewise.
7993 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
7994 (btrunc<mode>2): Delete separate expander, and combine with the
7995 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
7996 (btrunc<mode>2_fpr): Likewise.
7997 (ceil<mode>2): Likewise.
7998 (ceil<mode>2_fpr): Likewise.
7999 (floor<mode>2): Likewise.
8000 (floor<mode>2_fpr): Likewise.
8001 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
8002 Add support for using the upper registers with VSX and
8003 power8-vector. Move insns to be closer to the define_expands. On
8004 VSX systems, prefer the traditional form of FMA over the VSX
8005 version, since the traditional form allows the target not to
8006 overlap with the inputs.
8007 (fms<mode>4_fpr): Likewise.
8008 (nfma<mode>4_fpr): Likewise.
8009 (nfms<mode>4_fpr): Likewise.
8011 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8012 Richard Earnshaw <richard.earnshaw@arm.com>
8014 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
8015 (struct mult_cost_table): Likewise.
8016 (struct mem_cost_table): Likewise.
8017 (struct fp_cost_table): Likewise.
8018 (struct vector_cost_table): Likewise.
8019 (cpu_cost_table): Likewise.
8020 * config/arm/arm.opt (mold-rts-costs): New option.
8021 (mnew-generic-costs): Likewise.
8022 * config/arm/arm.c (generic_extra_costs): New table.
8023 (cortexa15_extra_costs): Likewise.
8024 (arm_slowmul_tune): Use NULL as new costs.
8025 (arm_fastmul_tune): Likewise.
8026 (arm_strongarm_tune): Likewise.
8027 (arm_xscale_tune): Likewise.
8028 (arm_9e_tune): Likewise.
8029 (arm_v6t2_tune): Likewise.
8030 (arm_cortex_a5_tune): Likewise.
8031 (arm_cortex_a9_tune): Likewise.
8032 (arm_v6m_tune): Likewise.
8033 (arm_fa726te_tune): Likewise.
8034 (arm_cortex_a15_tune): Use cortex15_extra_costs.
8035 (arm_cortex_tune): Use generict_extra_costs.
8036 (shifter_op_p): New function.
8037 (arm_unspec_cost): Likewise.
8038 (LIBCALL_COST): Define.
8039 (arm_new_rtx_costs): New function.
8040 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
8041 table is available. Use old costs otherwise unless mnew-generic-costs
8043 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
8044 (cpu_cost_table): Declare.
8046 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
8049 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
8050 (*subs_mul_imm_<mode>)
8051 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
8052 (*sub_<shift>_<mode>)
8053 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
8054 Remove k constraint.
8056 2013-10-03 Ian Bolton <ian.bolton@arm.com>
8058 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
8060 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
8062 2013-10-02 Teresa Johnson <tejohnson@google.com>
8064 * predict.c (probably_never_executed): New function.
8065 (probably_never_executed_bb_p): Invoke probably_never_executed.
8066 (probably_never_executed_edge_p): Ditto.
8067 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
8068 Treat profile insanities conservatively.
8070 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
8072 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
8074 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
8076 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
8077 use it. Use smaller increase for scratch. Don't increase reject
8078 for early clobber scratch.
8079 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
8080 setting eliminated regs except setting fp from hfp.
8081 (lra_eliminate): Check lra_insn_recog_data on NULL.
8083 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
8086 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
8087 setting -mvsx-timode by default until the underlying problem is fixed.
8088 (RS6000_CPU, power7 defaults): Likewise.
8090 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
8092 * config/x-linux (host-linux.o): Remove header dependencies.
8093 Use $(COMPILE) and $(POSTCOMPILE).
8094 * config/t-linux-android (linux-android.o): Ditto.
8096 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
8098 * Makefile.in (expmed.o-warn): Remove.
8100 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
8102 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
8103 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
8105 2013-10-02 Teresa Johnson <tejohnson@google.com>
8107 * dojump.c (do_jump_1): Divide probability between
8108 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
8110 2013-10-02 Tom Tromey <tromey@redhat.com>
8112 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
8114 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
8116 * tree-flow.h: Remove some prototypes.
8117 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
8118 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
8119 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
8120 mark_virtual_phi_result_for_renaming): Relocate here.
8121 * tree-into-ssa.h: Add prototypes.
8122 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
8123 single_pred_before_succ_order.
8124 (blocks_in_phiopt_order): Rename and move to cfganal.c.
8125 (nonfreeing_call_p) Move to gimple.c.
8126 * cfganal.c (single_pred_before_succ_order): Move and renamed from
8128 * basic-block.h (single_pred_before_succ_order): Add prototype.
8129 * gimple.c (nonfreeing_call_p): Relocate here.
8130 * gimple.h: Add prototype.
8131 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
8132 * tree-ssa-dom.h: New file. Relocate prototypes here.
8133 * tree-ssa.h: Include tree-ssa-dom.h.
8135 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
8137 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
8138 Use $(COMPILE) and $(POSTCOMPILE).
8140 * config/alpha/x-alpha (driver-alpha.o): Ditto.
8142 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
8144 * tree-flow.h: Remove some prototypes.
8145 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
8146 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
8147 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
8148 propagate_tree_value*): Move from here to...
8149 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
8150 propagate_tree_value*): Relocate here.
8151 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
8152 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
8153 * gimple-fold.c: Remove gimple-fold.h from include list.
8154 * tree-vrp.c: Remove gimple-fold.h from include list.
8155 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
8156 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
8157 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
8158 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
8159 * sese.c: Add tree-ssa-propagate.h to include list.
8161 2013-10-02 Richard Biener <rguenther@suse.de>
8163 * tree-loop-distribution.c: Include tree-vectorizer.h for
8165 (enum partition_kind): Remove PKIND_REDUCTION.
8166 (struct partition_s): Remove has_writes member, add reduction_p member.
8167 (partition_alloc): Adjust.
8168 (partition_builtin_p): Likewise.
8169 (partition_has_writes): Remove.
8170 (partition_reduction_p): New function.
8171 (partition_merge_into): Likewise.
8172 (generate_code_for_partition): Commonize builtin partition
8174 (rdg_cannot_recompute_vertex_p): Remove.
8175 (already_processed_vertex_p): Likewise.
8176 (rdg_flag_vertex): Do not set has_writes.
8177 (classify_partition): Adjust.
8178 (rdg_build_partitions): Do not set has_writes, treat all
8179 partitions as useful.
8180 (distribute_loop): Record number of library calls generated. Adjust.
8181 (tree_loop_distribution): Report number of loops and library
8182 calls generated as opt-info.
8184 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
8186 * tree-flow.h: Include new .h files. Move prototypes.
8187 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
8188 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
8189 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
8190 * tree-pretty-print.h: Add prototypes from tree-flow.h.
8191 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
8192 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
8193 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
8194 * tree.h (get_ref_base_and_extent): Move prototype out.
8195 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
8197 * gimple-low.h: New File. Add prototypes from tree-flow.h.
8198 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
8199 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
8200 * tree-scalar-evolution.c: Include tree.h.
8201 * sese.c: Include tree.h.
8202 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
8203 * dwarf2out.c: Include tree-dfa.h.
8204 * tree-chrec.c: Include tree.h.
8205 * tree-data-ref.c: Include tree.h.
8207 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
8209 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
8212 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8214 * config/t-sol2 (sol2-c.o): Remove header dependencies.
8215 Use $(COMPILE) and $(POSTCOMPILE).
8216 (sol2-cxx.o): Likewise.
8217 (sol2-stubs.o): Likewise.
8219 * config/x-solaris (host-solaris.o): Likewise.
8221 * config/sparc/t-sparc (sparc.o): Remove.
8222 (sparc-c.o): Remove header dependencies.
8223 Use $(COMPILE) and $(POSTCOMPILE).
8224 * config/sparc/x-sparc: Likewise.
8226 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
8228 * config/arc/arc-opts.h: Add 2013 to Copyright years.
8229 * config/arc/arc700.md: Likewise.
8230 * config/arc/arc-modes.def: Likewise.
8231 * config/arc/arc-simd.h: Likewise.
8232 * config/arc/t-arc-uClibc: Likewise.
8233 * config/arc/t-arc-newlib: Likewise.
8235 2013-10-02 Renlin Li <renlin.li@arm.com>
8237 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
8239 (aarch64_expand_epilogue): Likewise.
8241 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8242 Yufeng Zhang <yufeng.zhang@arm.com>
8244 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
8246 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
8247 'base_in' represent a conversion and legal_cast_p_1 holds; set
8248 'base_in' with the returned value from get_unwidened.
8250 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8252 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
8253 plus_constant is not used.
8255 2013-10-01 Wei Mi <wmi@google.com>
8257 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
8258 * config/i386/i386.md: Add define_peephole2 to
8259 break partial reg stall for cvtss2sd/cvtsd2ss.
8261 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
8263 * config/arc/arc.c (pass_arc_ifcvt::clone):
8264 Update for ctxt_ -> m_ctxt change.
8266 2013-10-01 Jeff Law <law@redhat.com>
8268 * tree-ssa-threadupdate.c (struct redirection_data): Delete
8269 outgoing_edge and intermediate_edge fields. Instead store the path.
8270 (redirection_data::hash): Hash on the last edge's destination index.
8271 (redirection_data::equal): Check the entire thread path.
8272 (lookup_redirectio_data): Corresponding changes.
8273 (create_edge_and_update_destination_phis): Likewise.
8274 (thread_single_edge): Likewise.
8276 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
8277 Diego Novillo <dnovillo@google.com>
8279 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
8280 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
8281 (vld32wh_insn, vld32wl_insn): Delete commented-out old
8282 versions of these patterns.
8284 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
8285 (__builtin_arc_aligned): Likewise.
8287 * config/arc/arc.md: Expand adc_0 comment stating the intended
8288 purpose and why it isn't ready.
8289 Replace commented out call_value_via_label_mixed with a
8290 plain comment about bl_s.
8292 * config/arc/arc.c (stdio.h): Don't include directly.
8293 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
8294 Remove if (1) condition.
8295 (arc_encode_section_info): Fix comment.
8297 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
8299 * config/arc/arc.c (arc_conditional_register_usage):
8300 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
8301 Also set reg_alloc_order for DMA config regs.
8303 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
8304 Jeremy Bennett <jeremy.bennett@embecosm.com>
8306 * doc/install.texi (--with-cpu): Mention ARC.
8307 (arc-*-elf32): New paragraph.
8308 (arc-linux-uclibc): Likewise.
8309 * doc/md.texi (Machine Constraints): Add ARC part.
8310 * doc/invoke.texi: (menu): Add ARC Options.
8311 (Machine Dependent Options) <ARC Options>: Add synopsis.
8312 (node ARC Options): Add.
8313 * doc/extend.texi (long_call / short_call attribute): Add ARC.
8314 (ARC Built-in Functions): New section defining
8315 generic ARC built-in functions.
8316 (ARC SIMD Built-in Functions): New section defining SIMD specific
8318 (Declaring Attributes of Functions): Extended
8319 description of short_call and long_call attributes for ARC and
8320 added index entries.
8322 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
8323 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
8324 Joern Rennecke <joern.rennecke@embecosm.com>
8325 Muhammad Khurram Riaz <khurram.riaz@arc.com>
8326 Brendan Kehoe <brendan@zen.org>
8327 Michael Eager <eager@eagercon.com>
8328 Simon Cook <simon.cook@embecosm.com>
8329 Jeremy Bennett <jeremy.bennett@embecosm.com>
8331 * config/arc, common/config/arc: New directories.
8333 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
8334 Brendan Kehoe <brendan@zen.org>
8335 Simon Cook <simon.cook@embecosm.com>
8337 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
8339 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
8341 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
8342 * tree-ssa-coalesce.h: New. Move prototype to here.
8343 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
8344 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
8345 (gimple_can_coalesce_p): Move to...
8346 * gimple.c (gimple_can_coalesce_p): Here.
8348 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
8350 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
8351 (dump_decl_set): Move to gimple.c.
8352 * gimple.h: Don't include tree-ssa-operands.h.
8353 (dump_decl_set): Add prototype.
8354 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
8355 Move to gimple-ssa.h.
8356 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
8357 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
8358 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
8359 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
8360 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
8362 (dump_decl_set): Relocate here.
8363 * gimple-ssa.h: New file.
8364 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
8365 Relocate from gimple.h.
8366 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
8367 * tree-ssa-operands.c (swap_ssa_operands): Rename from
8368 swap_tree_operands and remove non-ssa path.
8369 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
8370 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
8372 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
8373 vect_is_simple_reduction_1): Use swap_ssa_operands.
8374 * tree-flow.h: Move various prototypes to tree-phinodes.h.
8375 (enum need_phi_state): Move to tree-into-ssa.c.
8376 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
8377 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
8378 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
8379 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
8380 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
8381 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
8382 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
8383 num_imm_uses): Move to ssa-iterators.h.
8384 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
8385 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
8387 (op_iter_done, op_iter_next_def, op_iter_next_tree,
8388 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
8389 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
8390 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
8391 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
8392 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
8393 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
8394 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
8395 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
8396 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
8397 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
8398 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
8399 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
8400 (set_phi_nodes): Move to tree-phinodes.h.
8401 * tree-ssa-operands.h (enum ssa_op_iter_type,
8402 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
8403 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
8404 (dump_decl_set): Remove prototype.
8405 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
8406 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
8407 (set_phi_nodes): Relocate from tree-flow-inline.h.
8408 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
8410 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
8411 include list. Temporarily add gimple.h to include list.
8412 * ssa-iterators.h: New file.
8413 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
8414 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
8415 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
8416 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
8417 Relocate from tree-ssa-operands.h.
8418 (delink_imm_use, link_imm_use_to_list, link_imm_use,
8419 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
8420 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
8421 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
8422 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
8423 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
8424 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
8425 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
8426 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
8427 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
8428 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
8429 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
8430 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
8431 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
8432 Relocate from tree-flow-inline.h.
8433 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
8435 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
8438 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
8439 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
8440 Insert '\t' to output template.
8441 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
8442 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
8443 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
8444 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
8446 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
8448 * doc/install.texi (Host/target specific installation notes for GCC):
8449 Put @anchor before @heading.
8450 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
8451 Use @email for email addresses.
8453 2013-10-01 Jeff Law <law@redhat.com>
8455 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
8456 a vec. Only delete the path if we create one without successfully
8457 registering a jump thread.
8458 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
8460 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
8461 (paths): New vector of jump threading paths.
8462 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
8463 (THREAD_PATH): New accessor macro for the entire thread path.
8464 (lookup_redirection_data): Get intermediate and final outgoing edge
8465 from the thread path.
8466 (create_edge_and_update_destination_phis): Copy the threading path.
8467 (ssa_fix_duplicate_block_edges): Get edges and block types from the
8468 jump threading path.
8469 (ssa_redirect_edges): Get edges and block types from the jump threading
8470 path. Free the path vector.
8471 (thread_block): Get edges from the jump threading path. Look at the
8472 entire path to see if we thread to a loop exit. If we cancel a jump
8473 thread request, then free the path vector.
8474 (thread_single_edge): Get edges and block types from the jump threading
8475 path. Free the path vector.
8476 (thread_through_loop_header): Get edges and block types from the jump
8477 threading path. Free the path vector.
8478 (mark_threaded_blocks): Iterate over the vector of paths and store
8479 the path on the appropriate edge. Get edges and block types from the
8480 jump threading path.
8481 (mark_threaded_blocks): Get edges and block types from the jump
8482 threading path. Free the path vector.
8483 (thread_through_all_blocks): Use the vector of paths rather than
8484 a vector of 3-edge sets.
8485 (register_jump_thread): Accept pointer to a path vector rather
8486 than the path vector itself. Store the path vector for later use.
8489 2013-10-01 Jakub Jelinek <jakub@redhat.com>
8490 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8493 * config/s390/s390.c (s390_split_branches): Modify check for table
8495 (s390_chunkify_start): Rearrange table jump insn check in order to
8496 deal with compare and branch insns correctly.
8498 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
8502 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
8503 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
8504 define_insn_and_split.
8505 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
8506 (shiftsi3_compare): New pattern.
8508 * config/arm/unspecs.md (UNSPEC_RRX): New.
8510 2013-10-01 Alan Modra <amodra@gmail.com>
8512 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
8513 casing inout operands.
8515 2013-10-01 Richard Biener <rguenther@suse.de>
8517 PR tree-optimization/58553
8518 * tree-loop-distribution.c (struct partition_s): Add niter member.
8519 (classify_partition): Populate niter member for the partition
8520 and properly identify whether the relevant store happens before
8521 or after the loop exit.
8522 (generate_memset_builtin): Use niter member from the partition.
8523 (generate_memcpy_builtin): Likewise.
8525 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
8527 * vec.h (vec_prefix, vec): Prefix member names with "m_".
8528 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
8530 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
8532 * basic-block.h (edge_list): Prefix member names with "m_".
8533 * context.h (context): Likewise.
8534 * domwalk.h (dom_walker): Likewise.
8535 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
8536 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
8537 * hash-table.h (hash_table): Likewise.
8538 * machmode.h (bit_field_mode_iterator): Likewise.
8539 * pass_manager.h (pass_list): Likewise.
8540 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
8541 * tree-pass.h (pass_data): Likewise.
8542 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
8543 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
8544 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
8545 * asan.c (pass_data_asan): Update accordingly.
8546 * cfganal.c (control_dependences::find_control_dependence): Likewise.
8547 (control_dependences::control_dependences): Likewise.
8548 (control_dependences::~control_dependences): Likewise.
8549 (control_dependences::~control_dependences): Likewise.
8550 (control_dependences::get_edges_dependent_on): Likewise.
8551 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
8552 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
8553 * context.c (gcc::context::context): Likewise.
8554 * cprop.c (pass_rtl_cprop::clone): Likewise.
8555 * domwalk.c (dom_walker::walk): Likewise.
8556 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
8557 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
8558 * mode-switching.c (pass_mode_switching::clone): Likewise.
8559 * passes.c (opt_pass::opt_pass): Likewise.
8560 (pass_manager::pass_manager): Likewise.
8561 * predict.c (pass_strip_predict_hints::clone): Likewise.
8562 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
8563 (pass_split_all_insns::clone): Likewise.
8564 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
8566 (bit_field_mode_iterator::next_mode): Likewise.
8567 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
8568 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
8569 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
8570 * tree-complex.c (pass_lower_complex::clone): Likewise.
8571 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
8572 * tree-object-size.c (pass_object_sizes::clone): Likewise.
8573 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
8574 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
8575 (pass_fold_builtins::clone): Likewise.
8576 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
8577 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
8578 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
8579 (pass_cd_dce::clone): Likewise.
8580 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
8581 (pass_phi_only_cprop::clone): Likewise.
8582 * tree-ssa-dse.c (pass_dse::clone): Likewise.
8583 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
8584 * tree-ssa-loop.c (pass_lim::clone): Likewise.
8585 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
8586 * tree-ssa-pre.c (pass_fre::clone): Likewise.
8587 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
8588 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
8589 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
8590 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
8591 * tree-vrp.c (pass_vrp::clone): Likewise.
8592 * tsan.c (pass_tsan::clone): Likewise.
8594 2013-09-30 Jakub Jelinek <jakub@redhat.com>
8597 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
8598 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
8601 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
8602 optimization, punt if sign_bit_p looked through any zero extension.
8604 2013-09-30 Teresa Johnson <tejohnson@google.com>
8606 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
8607 Update redirected out edge count in joiner case.
8608 (ssa_redirect_edges): Common the joiner and non-joiner cases
8609 so that joiner case gets profile updates.
8611 2013-09-30 Richard Biener <rguenther@suse.de>
8613 PR tree-optimization/58554
8614 * tree-loop-distribution.c (classify_partition): Require
8615 unconditionally executed stores for memcpy and memset recognition.
8616 (tree_loop_distribution): Calculate dominance info.
8618 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
8620 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
8621 (NO_PROFILE_COUNTERS): Likewise.
8622 (PROFILE_HOOK): Likewise.
8623 (FUNCTION_PROFILER): Likewise.
8624 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
8626 2013-09-30 Iain Sandoe <iain@codesourcery.com>
8628 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
8629 calls and defines in TARGET_MACHO conditional.
8630 (load_macho_picbase_di): Likewise.
8631 (reload_macho_picbase): Likewise.
8632 (reload_macho_picbase_si): Likewise.
8633 (reload_macho_picbase_di): Likewise.
8634 (nonlocal_goto_receiver): Likewise.
8636 2013-09-30 Nick Clifton <nickc@redhat.com>
8638 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
8639 that use the MSP430X ISA.
8640 (msp430_option_override): Scan -mmcu command line option for any
8641 MCU name that supports the MSP430X ISA.
8642 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
8643 -mmcu options which enable the MSP430X ISA.
8645 2013-09-30 Richard Biener <rguenther@suse.de>
8648 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
8649 before looking for setjmp-like calls.
8651 2013-09-29 Iain Sandoe <iain@codesourcery.com>
8654 * config/darwin-protos.h (machopic_get_function_picbase): New.
8655 * config/darwin.c (machopic_get_function_picbase): New.
8656 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
8657 label for a new func. (load_macho_picbase_di): Likewise.
8658 (reload_macho_picbase): New expand.
8659 (reload_macho_picbase_si): New insn.
8660 (reload_macho_picbase_di): New insn.
8661 (nonlocal_goto_receiver): New define and split.
8662 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
8663 (unspecv enum): Add UNSPECV_NLGR.
8665 2013-09-29 Iain Sandoe <iain@codesourcery.com>
8667 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
8668 that altivec registers are correctly sized on Darwin.
8670 2013-09-29 Iain Sandoe <iain@codesourcery.com>
8672 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
8673 darwin-driver.o): Use COMPILE and POSTCOMPILE.
8674 * config/x-darwin (host-darwin.o): Likewise.
8675 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
8676 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
8677 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
8679 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
8681 * doc/invoke.texi: Fix usage of @tie{} command.
8683 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
8685 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
8687 2013-09-28 Jan Hubicka <jh@suse.cz>
8689 * config/i386/x86-tune.def: Add documentation for each of the options;
8692 2013-09-28 Jan Hubicka <jh@suse.cz>
8694 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
8696 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
8697 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
8698 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
8699 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
8700 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
8702 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
8704 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
8705 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
8706 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
8707 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
8708 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
8709 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
8710 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
8711 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
8712 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
8713 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
8714 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
8715 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
8716 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
8717 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
8718 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
8719 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
8720 gimple.h, godump.c, graphite-clast-to-gimple.c,
8721 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
8722 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
8723 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
8724 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
8725 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
8726 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
8727 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
8728 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
8729 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
8730 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
8731 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
8732 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
8733 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
8734 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
8735 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
8736 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
8737 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
8738 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
8739 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
8740 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
8741 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
8742 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
8743 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
8744 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
8745 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
8746 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
8747 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
8748 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
8749 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
8750 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
8751 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
8752 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
8753 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
8754 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
8755 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
8756 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
8757 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
8758 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
8759 whitespace before "(".
8761 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
8763 * expr.h (extract_bit_field): Remove packedp parameter.
8764 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
8765 from forward declaration.
8766 (store_split_bit_field): Remove packedp arg from calls to
8767 extract_fixed_bit_field.
8768 (extract_bit_field_1): Remove packedp parameter and packedp
8769 argument from recursive calls and calls to extract_fixed_bit_field.
8770 (extract_bit_field): Remove packedp parameter and corresponding
8771 arg to extract_bit_field_1.
8772 (extract_fixed_bit_field): Remove packedp parameter. Remove code
8774 (extract_split_bit_field): Remove packedp arg from call to
8775 extract_fixed_bit_field.
8776 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
8777 (copy_blkmode_from_reg): Likewise.
8778 (copy_blkmode_to_reg): Likewise.
8779 (read_complex_part): Likewise.
8780 (store_field): Likewise.
8781 (expand_expr_real_1): Likewise.
8782 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
8783 to extract_bit_field.
8784 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
8785 call to extract_bit_field.
8786 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
8787 call to extract_bit_field.
8788 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
8789 and special packedp behavior from -fstrict-volatile-bitfields
8792 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8794 * lra-eliminations.c (init_elim_table): Guard value_p.
8796 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
8798 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
8799 DFmode, DImode, and SFmode in the upper VSX registers based on the
8800 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
8801 if -mpower8-vector. Combine -mvsx-timode handling with the rest
8802 of the VSX register handling.
8804 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
8806 (zero_extendsidi2_lfiwzx): Add support for loading into the
8807 Altivec registers with -mpower8-vector. Use wu/wv constraints to
8808 only do VSX memory options on Altivec registers.
8809 (extendsidi2_lfiwax): Likewise.
8810 (extendsfdf2_fpr): Likewise.
8811 (mov<mode>_hardfloat, SF/SD modes): Likewise.
8812 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
8813 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
8814 (movdi_internal64): Likewise.
8816 2013-09-27 Xinliang David Li <davidxl@google.com>
8818 * opts.c (finish_options): Adjust parameters
8819 according to vect cost model.
8820 (common_handle_option): Set dynamic vect cost
8822 targhooks.c (default_add_stmt_cost): Compute stmt cost
8824 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
8825 Use helper function.
8826 * tree-vectorizer.h (unlimited_cost_model): New function.
8827 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
8828 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
8830 (vect_enhance_data_refs_alignment): Ditto.
8831 * flag-types.h: New enum.
8832 * common/config/i386/i386-common.c (ix86_option_init_struct):
8833 No need to initialize vect_cost_model flag.
8834 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
8837 2013-09-27 Diego Novillo <dnovillo@google.com>
8839 * gimple.h (enum ssa_mode): Remove.
8841 2013-09-27 Paulo Matos <pmatos@broadcom.com>
8843 * cfgloop.h (number_of_loops): Fix typo in check for null.
8845 2013-09-27 Jakub Jelinek <jakub@redhat.com>
8848 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
8849 are children of outermost saved_cfun's loop, and set it up to
8850 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
8851 if loop != loop0 and SESE region contains bbs that belong to loop0.
8853 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
8855 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
8856 (binary_scale_code_p, get_base_term, get_index_term): New functions.
8857 (set_address_segment, set_address_base, set_address_index)
8858 (set_address_disp): Accept the argument unconditionally.
8859 (baseness): Remove must_be_base_p and must_be_index_p checks.
8860 (decompose_normal_address): Classify as much as possible in the
8863 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
8865 * cse.c (count_reg_usage): Handle INT_LIST.
8866 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
8867 * reginfo.c (reg_scan_mark_refs): Likewise.
8868 * reload1.c (eliminate_regs_1): Likewise.
8870 2013-09-27 Iain Sandoe <iain@codesourcery.com>
8873 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
8876 2013-09-27 Richard Biener <rguenther@suse.de>
8878 PR tree-optimization/58459
8879 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
8880 restriction not propagating into loops.
8882 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
8884 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
8885 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
8886 * doc/tree-ssa.texi (Walking use-def chains): Delete.
8888 2013-09-26 Richard Biener <rguenther@suse.de>
8890 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
8892 2013-09-26 Richard Biener <rguenther@suse.de>
8894 * alias.h (component_uses_parent_alias_set): Rename to ...
8895 (component_uses_parent_alias_set_from): ... this.
8896 * alias.c (component_uses_parent_alias_set): Rename to ...
8897 (component_uses_parent_alias_set_from): ... this and return
8899 (reference_alias_ptr_type_1): Use the result from
8900 component_uses_parent_alias_set_from instead of stripping
8901 components one at a time.
8902 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
8904 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
8906 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
8907 Move prototypes to...
8908 * tree-ssa-ter.h: New File. Move prototypes here.
8909 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
8910 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
8911 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
8912 from is_replaceable_p.
8913 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
8914 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
8915 (process_replaceable): Use ter_is_replaceable_p.
8916 (find_replaceable_in_bb): Use ter_is_replaceable_p.
8917 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
8918 newly refactored ssa_is_replaceable_p.
8919 * cfgexpand.c: Include tree-outof-ssa.h.
8920 * ssaexpand.h: Delete.
8922 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
8924 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
8925 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
8927 (create_gimple_tmp): Delete.
8928 (get_expr_type, build_assign, build_type_cast): Move to...
8929 * gimple-builder.c: New File.
8930 (get_expr_type): Relocate from gimple.c.
8931 (build_assign, build_type_cast): Change to only create ssanames.
8932 * gimple.h: Move prototypes to...
8933 * gimple-builder.h: New File. Here.
8934 * tree-ssa.h: And here.
8935 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
8936 count_uses_and_derefs): Relocate from gimple.c.
8937 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
8938 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
8939 * tree-ssa-math-opts (execute_cse_reciprocals): Use
8940 gimple_replace_ssa_lhs.
8941 * asan.c: Include gimple-builder.h.
8942 * Makefile.in: Add gimple-builder.o.
8944 2013-09-26 Richard Biener <rguenther@suse.de>
8946 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
8947 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
8948 (loe_visit_block): Use gcc_checking_assert.
8949 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
8950 gimple_assign_ssa_name_copy_p.
8951 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
8954 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
8956 * config/rs6000/t-rs6000 (rs6000.o): Remove.
8957 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
8959 2013-09-26 Richard Biener <rguenther@suse.de>
8961 PR tree-optimization/58539
8962 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
8963 the fact that debug statements are not taking part in loop-closed
8966 2013-09-26 Nick Clifton <nickc@redhat.com>
8968 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
8969 time warning message.
8970 (msp430_print_operand_raw): Delete unused letter parameter.
8971 (TARGET_PRINT_OPERAND_ADDRESS): Define.
8972 (msp430_print_operand_address): New function.
8973 (msp430_print_operand): Move address printing code from here to
8975 * config/msp430/msp430.md (movsipsi2): Add comment in generated
8977 (zero_extendpsisi2): Likewise.
8978 (extendpsisi2): New pattern.
8979 (andneghi3): New pattern.
8981 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
8983 * config/aarch64/aarch64.opt (mlra): New option.
8984 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
8985 (TARGET_LRA_P): Define.
8987 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
8989 * expr.c (expand_assignment): Remove obsolete comment.
8991 2013-09-25 Jeff Law <law@redhat.com>
8993 * tree-flow.h (thread_through_all_blocks): Prototype moved into
8994 tree-ssa-threadupdate.h.
8995 (register_jump_thread): Similarly.
8996 * tree-ssa-threadupdate.h: New header file.
8997 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
8998 * tree-vrp.c: Likewise.
8999 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
9000 (thread_around_empty_blocks): Change type of path vector argument to
9001 an edge,type pair from just an edge. Initialize both elements when
9002 appending to a jump threading path. Tweak references to elements
9004 (thread_across_edge): Similarly. Release memory for the elements
9006 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
9007 (dump_jump_thread_path): New function broken out from
9008 register_jump_thread.
9009 (register_jump_thread): Use dump_jump_thread_path. Change type of
9010 path vector entries. Search the path for NULL edges and dump
9011 the path if one is found. Tweak the conversion of path to 3-edge
9012 form to use the block copy type information embedded in the path.
9014 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
9016 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
9018 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
9019 Vladimir Makarov <vmakarov@redhat.com>
9021 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
9022 from the least significant bit.
9023 (strip_address_mutations): Add bitfield operations handling.
9024 (must_be_index_p): Add shifting and rotate operations handling.
9025 (set_address_base): Use must_be_base_p predicate.
9026 (set_address_index): Use must_be_index_p predicate.
9028 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
9029 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
9030 Sergey Lega <sergey.s.lega@intel.com>
9031 Anna Tikhonova <anna.tikhonova@intel.com>
9032 Ilya Tocar <ilya.tocar@intel.com>
9033 Andrey Turetskiy <andrey.turetskiy@intel.com>
9034 Ilya Verbin <ilya.verbin@intel.com>
9035 Kirill Yukhin <kirill.yukhin@intel.com>
9036 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
9038 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
9040 (ix86_expand_vector_move_misalign): Support new unaligned load and
9041 stores and use new names.
9042 (CODE_FOR_sse2_storedqu): Rename to ...
9043 (CODE_FOR_sse2_storedquv16qi): ... this.
9044 (CODE_FOR_sse2_loaddqu): Rename to ...
9045 (CODE_FOR_sse2_loaddquv16qi): ... this.
9046 (CODE_FOR_avx_loaddqu256): Rename to ...
9047 (CODE_FOR_avx_loaddquv32qi): ... this.
9048 (CODE_FOR_avx_storedqu256): Rename to ...
9049 (CODE_FOR_avx_storedquv32qi): ... this.
9050 * config/i386/i386.md (fpint_logic): New.
9051 * config/i386/sse.md (VMOVE): Extend for AVX512.
9055 (VI_UNALIGNED_LOADSTORE): Ditto.
9056 (sse2_avx_avx512f): Ditto.
9057 (sse2_avx2): Extend for AVX512.
9058 (sse4_1_avx2): Ditto.
9059 (avx2_avx512f): New.
9060 (sse): Extend for AVX512.
9063 (avxsizesuffix): Ditto.
9064 (sseintvecmode): Ditto.
9066 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
9067 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
9068 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
9069 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
9070 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
9071 (<sse2_avx_avx512f>_storedqu<mode): ... this.
9072 (<sse>_movnt<mode>): Replace constraint "x" with "v".
9073 (STORENT_MODE): Extend for AVX512.
9074 (*absneg<mode>2): Replace constraint "x" with "v".
9075 (*mul<mode>3): Ditto.
9076 (*ieee_smin<mode>3): Ditto.
9077 (*ieee_smax<mode>3): Ditto.
9078 (avx_cmp<mode>3): Replace VF with VF_128_256.
9079 (*<sse>_maskcmp<mode>3_comm): Ditto.
9080 (<sse>_maskcmp<mode>3): Ditto.
9081 (<sse>_andnot<mode>3): Extend for AVX512.
9082 (<code><mode>3, anylogic): Replace VF with VF_128_256.
9083 (<code><mode>3, fpint_logic): New.
9084 (*<code><mode>3): Extend for AVX512.
9085 (avx512flogicsuff): New.
9086 (avx512f_<logic><mode>): Ditto.
9087 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
9089 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
9090 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
9091 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
9092 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
9093 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
9094 (xop_vpermil2<mode>3): Ditto.
9095 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
9096 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
9097 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
9098 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
9100 2013-09-25 Tom Tromey <tromey@redhat.com>
9102 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
9103 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
9104 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
9105 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
9106 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
9107 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
9108 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
9109 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
9110 (GRAPHITE_HTAB_H): Remove.
9112 2013-09-25 Tom Tromey <tromey@redhat.com>
9114 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
9116 2013-09-25 Tom Tromey <tromey@redhat.com>
9118 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
9120 2013-09-25 Tom Tromey <tromey@redhat.com>
9122 * config/i386/t-i386 (i386.o): Remove.
9123 (i386-c.o): Use COMPILE and POSTCOMPILE.
9125 2013-09-25 Tom Tromey <tromey@redhat.com>
9127 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
9129 2013-09-25 Tom Tromey <tromey@redhat.com>
9131 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
9132 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
9133 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
9134 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
9135 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
9136 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
9137 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
9138 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
9139 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
9140 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
9141 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
9142 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
9143 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
9144 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
9145 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
9146 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
9147 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
9148 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
9149 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
9150 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
9151 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
9152 (tree-ssa-pre.o, tree-ssa-sccvn.o)
9153 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
9154 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
9155 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
9156 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
9157 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
9158 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
9159 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
9160 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
9161 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
9162 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
9163 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
9164 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
9165 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
9166 (graphite-clast-to-gimple.o, graphite-dependences.o)
9167 (graphite-interchange.o, graphite-poly.o)
9168 (graphite-scop-detection.o, graphite-sese-to-poly.o)
9169 (graphite-optimize-isl.o, tree-vect-loop.o)
9170 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
9171 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
9172 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
9173 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
9174 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
9175 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
9176 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
9177 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
9178 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
9179 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
9180 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
9181 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
9182 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
9183 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
9184 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
9185 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
9186 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
9187 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
9188 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
9189 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
9190 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
9191 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
9192 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
9193 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
9194 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
9195 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
9196 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
9197 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
9198 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
9199 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
9200 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
9201 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
9202 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
9203 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
9204 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
9205 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
9206 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
9207 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
9208 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
9209 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
9210 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
9211 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
9212 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
9213 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
9214 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
9215 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
9216 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
9217 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
9218 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
9219 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
9220 (gcov-dump.o): Remove.
9221 (default-c.o): Use COMPILE and POSTCOMPILE.
9222 (CFLAGS-gcc.o): New variable.
9223 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
9225 2013-09-25 Tom Tromey <tromey@redhat.com>
9227 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
9228 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
9229 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
9230 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
9231 (c-family/c-pragma.o, c-family/c-pretty-print.o)
9232 (c-family/c-semantics.o, c-family/c-ada-spec.o)
9233 (c-family/array-notation-common.o, c-family/stub-objc.o)
9234 (c-family/c-ubsan.o): Remove.
9236 2013-09-25 Tom Tromey <tromey@redhat.com>
9238 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
9240 2013-09-25 Tom Tromey <tromey@redhat.com>
9242 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
9243 to add -DENABLE_SHARED_LIBGCC.
9244 (gcc.o): Don't use subshell.
9246 2013-09-25 Tom Tromey <tromey@redhat.com>
9248 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
9249 * configure.ac: Don't invoke AM_PROG_CC_C_O.
9250 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
9251 * configure, config.in: Rebuild.
9253 2013-09-25 Tom Tromey <tromey@redhat.com>
9255 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
9256 (COMPILE, POSTCOMPILE): New variables.
9257 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
9258 (DEPFILES): New variable.
9259 Include ".Po" files.
9260 * configure.ac: Add checks for dependency checking.
9261 * configure, aclocal.m4: Regenerate.
9263 2013-09-25 Tom Tromey <tromey@redhat.com>
9265 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
9266 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
9268 2013-09-25 Tom Tromey <tromey@redhat.com>
9270 * Makefile.in (generated_files): Add options.h,
9271 target-hooks-def.h, insn-opinit.h,
9272 common/common-target-hooks-def.h, pass-instances.def,
9273 c-family/c-target-hooks-def.h.
9275 2013-09-25 Jeff Law <law@redhat.com>
9277 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
9278 than foo[foo.length () - 1] to access last member in a vec.
9279 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
9281 2013-09-25 Richard Biener <rguenther@suse.de>
9284 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
9286 2013-09-25 Jan Hubicka <jh@suse.cz>
9288 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
9291 2013-09-25 Marek Polacek <polacek@redhat.com>
9294 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
9295 when determining the type name.
9297 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
9299 * config/sh/sh.md: Fix formatting.
9301 2013-09-24 Xinliang David Li <davidxl@google.com>
9303 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
9304 max peel iterations parameter.
9305 * param.def: New parameter.
9306 * doc/invoke.texi: Document New parameter.
9308 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
9310 * gimple-pretty-print.c: Various whitespace tweaks.
9311 * tree-core.h: Likewise.
9312 * tree-pretty-print.c: Likewise.
9313 * tree-ssa-alias.c: Likewise.
9314 * tree-ssa-copy.c: Likewise.
9315 * tree-ssanames.c: Likewise.
9316 * tree-ssanames.h: Likewise.
9317 * tree-vrp.c: Likewise.
9319 2013-09-24 Alan Modra <amodra@gmail.com>
9323 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
9324 for output operands that disallow regs. Don't use EXPAND_WRITE on
9327 2013-09-24 Richard Biener <rguenther@suse.de>
9330 * tree.c (reference_alias_ptr_type): Move ...
9331 * alias.c (reference_alias_ptr_type): ... here and implement
9332 in terms of the new reference_alias_ptr_type_1.
9333 (ref_all_alias_ptr_type_p): New helper.
9334 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
9335 use ref_all_alias_ptr_type_p.
9336 (get_deref_alias_set): Add flag_strict_aliasing check here.
9337 (reference_alias_ptr_type_1): New function, split out from ...
9338 (get_alias_set): ... here.
9339 (alias_ptr_types_compatible_p): New function.
9340 * alias.h (reference_alias_ptr_type): Declare.
9341 (alias_ptr_types_compatible_p): Likewise.
9342 * tree.h (reference_alias_ptr_type): Remove.
9343 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
9344 to compare MEM_REF alias types.
9346 2013-09-24 Richard Biener <rguenther@suse.de>
9348 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
9350 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
9352 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
9353 reload helper function arrays into a single array reg_addr.
9354 (reload_fpr_gpr): Likewise.
9355 (reload_gpr_vsx): Likewise.
9356 (reload_vsx_gpr): Likewise.
9357 (struct rs6000_reg_addr): Likewise.
9358 (reg_addr): Likewise.
9359 (rs6000_debug_reg_global): Change rs6000_vector_reload,
9360 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
9361 (rs6000_init_hard_regno_mode_ok): Likewise.
9362 (rs6000_secondary_reload_direct_move): Likewise.
9363 (rs6000_secondary_reload): Likewise.
9365 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
9366 constraints: wu, ww, and wy. Repurpose wv constraint added during
9367 power8 changes. Put wg constraint in alphabetical order.
9369 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
9370 for future work to add ISA 2.07 VSX single precision support.
9371 (-mvsx-scalar-double): Change default from -1 to 1, update
9372 documentation comment.
9373 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
9374 (-mupper-regs-df): New debug switch to control whether DF values
9375 can go in the traditional Altivec registers.
9376 (-mupper-regs-sf): New debug switch to control whether SF values
9377 can go in the traditional Altivec registers.
9379 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
9381 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
9382 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
9383 Add new constraints, wu/ww/wy. Repurpose wv constraint.
9384 (rs6000_debug_legitimate_address_p): Print if we are running
9385 before, during, or after reload.
9386 (rs6000_secondary_reload): Add a comment.
9387 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
9389 * config/rs6000/constraints.md (wa constraint): Sort w<x>
9390 constraints. Update documentation string.
9391 (wd constraint): Likewise.
9392 (wf constraint): Likewise.
9393 (wg constraint): Likewise.
9394 (wn constraint): Likewise.
9395 (ws constraint): Likewise.
9396 (wt constraint): Likewise.
9397 (wx constraint): Likewise.
9398 (wz constraint): Likewise.
9399 (wu constraint): New constraint for ISA 2.07 SFmode scalar
9401 (ww constraint): Likewise.
9402 (wy constraint): Likewise.
9403 (wv constraint): Repurpose ISA 2.07 constraint that we did not use
9404 in the previous submissions.
9405 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
9407 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
9409 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
9410 (REG_BR_PROB): Say that the probability is stored in an int_list.
9411 * reg-notes.def: Update commentary to mention INT_LIST.
9412 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
9413 (INT_LIST): New rtx.
9414 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
9415 * rtlanal.c (int_reg_note_p): New function.
9416 (alloc_reg_note): Assert that the note does not have an int argument.
9417 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
9418 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
9419 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
9420 rather than an INSN_LIST. Handle INT_LIST.
9421 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
9422 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
9423 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
9424 Manipulate them as ints rather than rtxes.
9425 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
9426 * regmove.c (copy_src_to_dest): Likewise.
9427 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
9429 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
9430 into the cases that need it.
9431 * config/arm/arm.c (arm_unwind_emit): Likewise.
9433 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
9434 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
9435 * loop-doloop.c (add_test, doloop_modify): Likewise.
9436 * loop-unswitch.c (compare_and_jump_seq): Likewise.
9437 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
9438 * predict.c (combine_predictions_for_insn): Likewise.
9439 * print-rtl.c (print_rtx): Handle INT_LIST.
9440 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
9441 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
9442 * config/arm/arm.c (emit_unlikely_jump): Likewise.
9443 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
9444 (ix86_split_fp_branch, predict_jump): Likewise.
9445 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
9446 * config/sh/sh.c (expand_cbranchsi4): Likewise.
9447 * config/spu/spu.c (ea_load_store_inline): Likewise.
9449 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
9450 value of a REG_BR_PROB note.
9451 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
9452 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
9453 * emit-rtl.c (try_split): Likewise.
9454 * predict.c (br_prob_note_reliable_p): Likewise.
9455 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
9456 * reorg.c (mostly_true_jump): Likewise.
9457 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
9458 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
9459 * config/i386/i386.c (ix86_print_operand): Likewise.
9460 * config/ia64/ia64.c (ia64_print_operand): Likewise.
9461 * config/mmix/mmix.c (mmix_print_operand): Likewise.
9462 * config/rs6000/rs6000.c (output_cbranch): Likewise.
9463 * config/s390/s390.c (s390_expand_tbegin): Likewise.
9464 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
9465 * config/sparc/sparc.c (output_cbranch): Likewise.
9466 * config/spu/spu.c (get_branch_target): Likewise.
9467 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
9468 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
9470 2013-09-23 Jan Hubicka <jh@suse.cz>
9472 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
9474 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
9475 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
9477 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
9478 (gimple_fold_call): Dump inconsistent devirtualizations; add
9479 sanity check for type based devirtualizations.
9480 * ipa-prop.c: Include ipa-utils.h
9481 (ipa_intraprocedural_devirtualization): Add sanity check.
9482 (try_make_edge_direct_virtual_call): Likewise.
9484 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
9486 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
9487 assignment statements.
9489 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
9491 * gimple-pretty-print.c (dump_ssaname_info): New function.
9492 (dump_gimple_phi): Call it.
9493 (pp_gimple_stmt_1): Likewise.
9494 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
9495 (range_info_def): Declare.
9496 * tree-pretty-print.c (pp_double_int): New function.
9497 (dump_generic_node): Call it.
9498 * tree-pretty-print.h (pp_double_int): Declare.
9499 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
9500 * tree-ssanames.h (range_info_def): New structure.
9501 (value_range_type): Move definition here.
9502 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
9504 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
9506 (set_range_info): New function.
9507 (get_range_info): Likewise.
9508 (duplicate_ssa_name_range_info): Likewise.
9509 (duplicate_ssa_name_fn): Check pointer type and call
9510 duplicate_ssa_name_range_info.
9511 * tree-ssa-copy.c (fini_copy_prop): Likewise.
9512 * tree-vrp.c (value_range_type): Remove definition, now in
9514 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
9515 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
9516 (SSA_NAME_RANGE_INFO): New macro.
9518 2013-09-23 Richard Biener <rguenther@suse.de>
9520 PR tree-optimization/58464
9521 * tree-ssa-pre.c (phi_trans_lookup): Remove.
9522 (phi_trans_add): Change to add conditionally on being not present.
9523 (phi_translate_1): Remove recursion detection here.
9524 (phi_translate): Pre-seed the cache with NULL to catch
9525 recursion here in a more generic way.
9526 (bitmap_find_leader): Adjust comment.
9527 (get_representative_for): Dump value-numbers.
9528 (create_expression_by_pieces): Likewise.
9529 (insert_into_preds_of_block): Likewise.
9531 2013-09-23 Christian Bruel <christian.bruel@st.com>
9534 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
9535 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
9537 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
9540 2013-09-20 Renlin Li <renlin.li@arm.com>
9542 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
9544 (aarch64_expand_epilogue): Likewise.
9545 (aarch64_legitimize_reload_address): Likewise.
9547 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
9549 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
9550 NULL_TREE before pushing them onto the vector. Likewise for labels.
9552 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
9554 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
9555 comparison operators when -fno-trapping-math is in effect.
9556 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
9557 comparison operators in TFmode and assert that unsupported operators
9559 (ia64_print_operand): Likewise.
9561 2013-09-21 Jan Hubicka <jh@suse.cz>
9563 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
9564 (sse_typeless_stores): Enable for core
9565 (sse_load0_by_pxor): Likewise.
9566 (four_jump_limit): Disable for core.
9567 (pad_returns): Likewise.
9568 (avoid_vector_decode): Likewise.
9569 (fuse_cmp_and_branch): Enable for cores.
9570 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
9572 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
9575 * config/pa/pa.c (pa_option_override): Disable auto increment and
9576 decrement instructions until reload is completed.
9578 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
9579 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
9580 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
9582 2013-09-20 DJ Delorie <dj@redhat.com>
9583 Nick Clifton <nickc@redhat.com>
9585 * config/rl78/rl78.c: Various whitespace and comment tweaks.
9586 (need_to_save): Save bank 0 on interrupts.
9587 (characterize_address): Strip far address wrappers.
9588 (rl78_as_legitimate_address): Likewise.
9589 (transcode_memory_rtx): Likewise.
9590 (rl78_peep_movhi_p): Disable this peephole after devirt.
9591 (rl78_propogate_register_origins): Forget all origins when a
9593 * config/rl78/rl78-virt.md: Various whitespace tweaks.
9594 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
9596 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
9597 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
9598 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
9600 * config/rl78/constraints.md: Various whitespace and paren tweaks.
9602 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
9604 * config/pa/pa.md: In "scc" insn patterns, change output template to
9605 handle const0_rtx in reg_or_0_operand operands.
9607 2013-09-20 Martin Husemann <martin@NetBSD.org>
9610 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
9611 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
9613 2013-09-20 Richard Biener <rguenther@suse.de>
9616 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
9617 remembering SSA name version and block index.
9618 (new_scev_info_str): Adjust.
9619 (hash_scev_info): Likewise. Also hash the block index.
9620 (eq_scev_info): Adjust.
9621 (find_var_scev_info): Likewise.
9622 (struct instantiate_cache_entry): Remove.
9623 (struct instantiate_cache_type): Use a htab to map name, block
9625 (instantiate_cache_type::~instantiate_cache_type): Adjust.
9626 (get_instantiated_value_entry): Likewise.
9627 (hash_idx_scev_info, eq_idx_scev_info): New functions.
9628 (instantiate_scev_name): Adjust.
9630 2013-09-20 Jeff Law <law@redhat.com>
9632 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
9634 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
9636 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
9637 Call aarch64_simd_expand_args to update op[argc].
9639 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
9641 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
9644 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
9646 * gengtype.c (file_rules): Added rule for *.cc files.
9647 (get_output_file_with_visibility): Give fatal message when no
9650 2013-09-20 Renlin Li <renlin.li@arm.com>
9652 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
9653 (aarch64_expand_epilogue): Likewise.
9654 (aarch64_legitimize_reload_address): Likewise.
9656 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
9659 * expr.c (expand_assignment): Remove misalignp code path.
9661 2013-09-20 Marek Polacek <polacek@redhat.com>
9664 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
9665 TYPE_PRECISION. Add asserts.
9667 2013-09-20 Richard Biener <rguenther@suse.de>
9669 PR tree-optimization/58453
9670 * tree-loop-distribution.c (distribute_loop): Apply the cost
9671 model for -ftree-loop-distribute-patterns, too.
9673 2013-09-20 Richard Biener <rguenther@suse.de>
9676 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
9677 make type comparison less strict.
9679 2013-09-20 Alan Modra <amodra@gmail.com>
9681 * configure: Regenerate.
9682 * aclocal.m4: Regenerate.
9684 2013-09-20 Marek Polacek <polacek@redhat.com>
9687 * doc/extend.texi: Document that attribute used is meant to be used
9688 on variables with static storage duration.
9690 2013-09-19 Jakub Jelinek <jakub@redhat.com>
9692 PR tree-optimization/58472
9693 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
9694 simd_lane_access set inv_p = false.
9695 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
9696 the simduid magic VAR_DECL.
9698 2013-09-19 Jan Hubicka <jh@suse.cz>
9700 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
9702 2013-09-17 Jeff Law <law@redhat.com>
9704 * tree-ssa-dom.c (record_temporary_equivalences): New function
9705 split out of dom_opt_dom_walker::after_dom_children.
9706 (dom_opt_dom_walker::thread_across_edge): Move common code
9707 in here from dom_opt_dom_walker::after_dom_children.
9708 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
9710 2013-09-19 Jan Hubicka <jh@suse.cz>
9712 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
9713 (TARGET_GENERIC): Use PROCESOR_GENERIC
9714 (enum processor_type): Unify generic32 and 64.
9715 * i386.md (cpu): Likewise.
9716 * x86-tune.def (use_leave): Enable for generic32.
9717 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
9718 * athlon.md: Change generic64 to generic in all occurences.
9719 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
9720 (ix86_target_macros_internal): Likewise.
9721 * driver-i386.c (host_detect_local_cpu): Likewise.
9722 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
9724 (generic_memcpy, generic_memset, generic_cost): This one.
9725 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
9726 (m_GENERIC32, m_GENERIC64): Remove.
9727 (m_GENERIC): Turn into one flag.
9728 (processor_target): Unify generic tunnings.
9729 (ix86_option_override_internal): Replace generic32/64 by generic.
9730 (ix86_issue_rate): Likewise.
9731 (ix86_adjust_cost): Likewise.
9733 2013-09-19 Jan Hubicka <jh@suse.cz>
9735 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
9736 of speculative flag.
9738 2013-09-19 Jakub Jelinek <jakub@redhat.com>
9740 * omp-low.c (expand_omp_sections): Always pass len - 1 to
9741 GOMP_sections_start, even if !exit_reachable.
9743 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
9745 * lra-constraints.c (need_for_all_save_p): Use macro
9746 HARD_REGNO_CALL_PART_CLOBBERED.
9747 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
9748 set up pseudo conflict hard regs.
9750 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
9753 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
9756 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
9758 PR rtl-optimization/58438
9759 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
9760 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
9761 from previous subpasses.
9763 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
9765 * arm.c (arm_get_frame_offsets): Validate architecture supports
9766 LDRD/STRD before accepting the tuning preference.
9767 (arm_expand_prologue): Likewise.
9768 (arm_expand_epilogue): Likewise.
9770 2013-09-18 Richard Biener <rguenther@suse.de>
9772 PR tree-optimization/58417
9773 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
9774 have chrecs with symbols defined in the loop as operands.
9775 (chrec_fold_multiply): Likewise.
9776 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
9777 parameters before folding binary operations.
9778 (struct instantiate_cache_entry_hasher): Remove.
9779 (struct instantiate_cache_type): Use a pointer-map.
9780 (instantiate_cache_type::instantiate_cache_type): New function.
9781 (instantiate_cache_type::get): Likewise.
9782 (instantiate_cache_type::set): Likewise.
9783 (instantiate_cache_type::~instantiate_cache_type): Adjust.
9784 (get_instantiated_value_entry): Likewise.
9785 (global_cache): New global.
9786 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
9787 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
9788 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
9789 (instantiate_scev_name): Adjust.
9790 (instantiate_scev): Construct global instead of local cache.
9791 (resolve_mixers): Likewise.
9793 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
9794 Paolo Carlini <paolo.carlini@oracle.com>
9797 * doc/implement-cxx.texi: Fix references to the C++ standards.
9799 2013-09-18 Jakub Jelinek <jakub@redhat.com>
9801 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
9802 * tree-vect-data-refs.c (vect_analyze_data_refs): For
9803 simd_lane_access drs, update also DR_ALIGNED_TO.
9805 2013-09-18 Marek Polacek <polacek@redhat.com>
9808 * doc/extend.texi: Document no_sanitize_undefined attribute.
9809 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
9810 no_sanitize_undefined attribute.
9812 2013-09-18 Nick Clifton <nickc@redhat.com>
9814 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
9815 (ASM_DECLARE_FUNCTION_NAME): Define.
9817 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
9819 * compare-elim.c (find_comparison_dom_walker): New class
9820 (find_comparisons_in_bb): Rename to
9821 find_comparison_dom_walker::before_dom_children
9822 (find_comparisons): Adjust
9823 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
9825 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
9826 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
9827 (init_walk_dominator_tree): Remove declaration.
9828 (fini_walk_dominator_tree): Remove declaration.
9829 * fwprop.c (single_def_use_dom_walker): New class
9830 (single_def_use_enter_block): Convert to
9831 single_def_use_dom_walker::before_dom_children.
9832 (single_def_use_leave_block): Convert to
9833 single_def_use_dom_walker::after_dom_children.
9834 (build_single_def_use_links): Adjust.
9835 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
9837 (find_candidates_in_block): Convert to
9838 find_candidates_dom_walker::before_dom_children.
9839 (execute_strength_reduction): Adjust.
9840 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
9841 (sese_dom_walker): New class.
9842 (sese_dom_walker::sese_dom_walker): New constructor.
9843 (sese_dom_walker::~sese_dom_walker): New destructor.
9844 (build_sese_conditions_before): Convert to
9845 sese_dom_walker::before_dom_children.
9846 (build_sese_conditions_after): Convert to
9847 sese_dom_walker::after_dom_children.
9848 (build_poly_scop): Adjust
9849 * tree-into-ssa.c (rewrite_dom_walker): New class
9850 (rewrite_enter_block): Convert to
9851 rewrite_dom_walker::before_dom_children.
9852 (rewrite_leave_block): Convert to
9853 rewrite_dom_walker::after_dom_children.
9854 (rewrite_update_dom_walker): New class.
9855 (rewrite_update_enter_block): Convert to
9856 rewrite_update_dom_walker::before_dom_children.
9857 (rewrite_update_leave_block): Convert to
9858 rewrite_update_dom_walker::after_dom_children.
9859 (rewrite_blocks, rewrite_into_ssa): Adjust.
9860 (mark_def_dom_walker): New class.
9861 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
9862 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
9863 (mark_def_sites_blocks): Convert to
9864 mark_def_dom_walker::before_dom_children.
9865 (mark_def_site_blocks): Remove.
9866 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
9867 (tree_ssa_dominator_optimize): Adjust.
9868 (dom_thread_across_edge): Convert to method
9869 dom_opt_dom_walker::thread_across_edge.
9870 (dom_opt_enter_block): Convert to member function
9871 dom_opt_dom_walker::before_dom_children.
9872 (dom_opt_leave_block): Convert to member function
9873 dom_opt_dom_walker::after_dom_children.
9874 * tree-ssa-dse.c (dse_dom_walker): New class.
9875 (dse_enter_block): Convert to member function
9876 dse_dom_walker::before_dom_children.
9877 (tree_ssa_dse): Adjust.
9878 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
9879 (determine_invariantness_stmt): Convert to method
9880 invariantness_dom_walker::before_dom_children.
9881 (determine_invariantness): Remove
9882 (move_computations_dom_walker): New class.
9883 (move_computations_stmt): Convert to method
9884 move_computations_dom_walker::before_dom_children.
9885 (move_computations, tree_ssa_lim): Adjust.
9886 * tree-ssa-phiopt.c (nontrapping_dom_walker): New class.
9887 (nt_init_block): Convert to method
9888 notrappping_dom_walker::before_dom_children.
9889 (nt_fini_block): Convert to method
9890 method nontrapping_dom_walker::after_dom_children.
9891 (get_non_trapping): Adjust.
9892 * tree-ssa-pre.c (eliminate_dom_walker): New class.
9893 (eliminate_bb): Convert to method
9894 eliminate_dom_walker::before_dom_children.
9895 (eliminate_leave_block): Convert to method
9896 eliminate_dom_walker::after_dom_children.
9897 (eliminate): Adjust.
9898 * tree-ssa-strlen.c (strlen_dom_walker): New class.
9899 (strlen_enter_block): Convert to method
9900 strlen_dom_walker::before_dom_children.
9901 (strlen_leave_block): Convert to method
9902 method strlen_dom_walker::after_dom_children.
9903 (tree_ssa_strlen): Adjust.
9904 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
9905 (tree_ssa_uncprop): Adjust.
9906 (uncprop_leave_block): Convert to method
9907 uncprop_dom_walker::after_dom_children.
9908 (uncprop_leave_block): Convert to method
9909 uncprop_dom_walker::before_dom_children.
9911 2013-09-18 Bin Cheng <bin.cheng@arm.com>
9913 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
9914 branch in same basic block. Check both src and dest of the move insn.
9916 2013-09-17 Nick Clifton <nickc@redhat.com>
9918 * config/rl78/rl78-real.md (bf): New pattern.
9920 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
9921 (rl78_print_operand): Do not put a # before a %B.
9922 * config/rl78/rl78.opt: Tweak doc strings.
9924 2013-09-17 DJ Delorie <dj@redhat.com>
9926 * config/rl78/constraints.md (Wcv): Allow up to $r31.
9927 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
9928 (rl78_option_override): Likewise, if -mallregs.
9929 (is_virtual_register): Likewise.
9930 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
9931 (REGNO_OK_FOR_BASE_P): Likewise.
9932 * config/rl78/rl78.opt (-mallregs): New.
9934 2013-09-17 Nick Clifton <nickc@redhat.com>
9936 * config/rl78/rl78.c (need_to_save): Change return type to bool.
9937 For interrupt functions: save all call clobbered registers if the
9938 interrupt handler is not a leaf function.
9939 (rl78_expand_prologue): Always recompute the frame information.
9940 For interrupt functions: only select bank 0 if one of the bank 0
9941 registers is going to be psuhed.
9943 2013-09-17 DJ Delorie <dj@redhat.com>
9945 * config/rl78/constraints.md: For each W* constraint, rename to C*
9946 and create a W* constraint that checks for an optional ES: prefix
9948 * config/rl78/rl78.md (UNS_ES_ADDR): New.
9949 (es_addr): New. Used to wrap far addresses.
9950 * config/rl78/rl78-protos.h (rl78_es_addr): New.
9951 (rl78_es_base): New.
9952 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
9953 wrapped far addresses.
9954 (rl78_print_operand_1): Unwrap far addresses before processing.
9955 (rl78_lo16): Wrap far addresses in unspecs.
9956 (rl78_es_addr): New.
9957 (rl78_es_base): New.
9958 (insn_ok_now): Check for not-yet-wrapped far addresses.
9959 (transcode_memory_rtx): Properly re-wrap far addresses.
9961 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
9963 * config/sparc/t-rtems: Add leon3 multilibs.
9965 2013-09-17 Cong Hou <congh@google.com>
9967 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
9968 when checking the dot production pattern. The type of rhs operand
9969 of multiply is now checked correctly.
9971 2013-09-17 Jeff Law <law@redhat.com>
9973 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
9974 edge implied equivalences into successor phis.
9975 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
9976 here from tree-ssa-threadedge.c.
9977 (mark_threaded_blocks): When threading through a joiner, if both
9978 successors of the joiner's clone reach the same block, verify the
9979 PHI arguments are equal. If not, cancel the jump threading request.
9980 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
9981 tree-ssa-threadupdate.c
9982 (thread_across_edge): Don't check PHI argument equality when
9983 threading through joiner block here.
9985 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
9987 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
9988 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
9989 (warn_uninit, warn_uninitialized_vars,
9990 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
9991 Move to tree-ssa-uninit.c.
9992 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
9993 (has_undefined_value_p): New. Pass dependant parts of
9994 ssa_undefined_value_p.
9995 (uninit_undefined_value_p): Use has_undefined_value_p.
9996 (warn_uninit, warn_uninitialized_vars,
9997 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
9998 Move from tree-ssa.c.
9999 * tree-ssa.h: Adjust prototypes.
10001 2013-09-17 Jan Hubicka <jh@suse.cz>
10003 PR middle-end/58332
10004 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
10005 * ipa-inline.c (can_inline_edge_p): Do not downgrade
10006 FUNCTION_NOT_OPTIMIZED.
10007 * ipa-inline-analysis.c (compute_inline_parameters): Function
10008 not optimized is not inlinable unless it is alwaysinline.
10009 (inline_analyze_function): Force calls in not optimized
10010 function not inlinable.
10012 2013-09-17 Jan Hubicka <jh@suse.cz>
10014 PR middle-end/58329
10015 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
10017 * ipa.c (function_and_variable_visibility): Likewise.
10018 * ipa-profile.c (ipa_profile): Likewise.
10020 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
10023 * cgraph.c (cgraph_function_body_availability): Check for ifunc
10024 attribute, and don't inline the resolver in this case.
10026 2013-09-17 Teresa Johnson <tejohnson@google.com>
10028 * coverage.c (get_coverage_counts): Add missing newline.
10030 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10032 PR tree-optimization/58088
10033 * fold-const.c (mask_with_trailing_zeros): New function.
10034 (fold_binary_loc): Make sure we don't recurse infinitely
10035 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
10036 Use mask_with_trailing_zeros where appropriate.
10038 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
10040 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
10041 of 'prev' var to get better distance estimation.
10043 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
10045 * tree-inline.h (struct copy_body_data): Add transform_parameter.
10046 * tree-inline.c (is_parameter_of): New predicate.
10047 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
10048 a parameter has been remapped.
10049 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
10050 (optimize_inline_calls): Initialize transform_parameter.
10051 (copy_gimple_seq_and_replace_locals): Likewise.
10052 (tree_function_versioning): Likewise.
10053 (maybe_inline_call_in_expr): Likewise.
10055 2013-09-17 Nick Clifton <nickc@redhat.com>
10057 * config/msp430/msp430-protos.h: Add prototypes for new functions.
10058 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
10059 interrupt handlers.
10060 (is_attr_func): New function.
10061 (msp430_is_interrupt_func): New function.
10062 (is_naked_func): New function.
10063 (is_reentrant_func): New function.
10064 (is_critical_func): New function.
10065 (msp430_start_function): Add annotations for function attributes.
10066 (msp430_attr): New function.
10067 (msp430_attribute_table): New.
10068 (msp430_function_section): New function.
10069 (TARGET_ASM_FUNCTION_SECTION): Define.
10070 (msp430_builtin): New enum.
10071 (msp430_init_builtins): New function.
10072 (msp430_builtin_devl): New function.
10073 (msp430_expand_builtin): New function.
10074 (TARGET_INIT_BUILTINS): Define.
10075 (TARGET_EXPAND_BUILTINS): Define.
10076 (TARGET_BUILTIN_DECL): Define.
10077 (msp430_expand_prologue): Add support for naked, interrupt,
10078 critical and reentrant functions.
10079 (msp430_expand_epilogue): Likewise.
10080 (msp430_print_operand): Handle 'O' character.
10081 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
10083 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
10084 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
10085 (pushm): Use a 'n' rather than an 'i' constraint.
10086 (msp_return): Add generation of the interrupt return instruction.
10087 (disable_interrupts): New pattern.
10088 (enable_interrupts): New pattern.
10089 (push_intr_state): New pattern.
10090 (pop_intr_state): New pattern.
10091 (bic_SR): New pattern.
10092 (bis_SR): New pattern.
10093 * doc/extend.texi: Document MSP430 function attributes and builtin
10096 2013-09-17 Richard Biener <rguenther@suse.de>
10098 PR tree-optimization/58432
10099 * tree-loop-distribution.c (tree_loop_distribution): Also
10100 scan PHIs for outside loop uses and seed a partition from them.
10102 2013-09-17 Bin Cheng <bin.cheng@arm.com>
10104 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
10105 (restructure_reference): Call backtrace_base_for_ref.
10107 2013-09-17 Alan Modra <amodra@gmail.com>
10110 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
10113 2013-09-16 DJ Delorie <dj@redhat.com>
10115 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
10116 vregs location for RL78/G10.
10117 (rl78_expand_prologue): Avoid SEL on G10.
10118 (rl78_expand_epilogue): Likewise.
10119 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
10120 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
10121 __RL78_G10__ when appropriate.
10122 (ASM_SPEC): Pass -mg10 along to the assembler.
10123 * config/rl78/rl78.md (sel_rb): Disable for G10.
10124 * config/rl78/rl78.opt: Add -mg10 option.
10125 * config/rl78/t-rl78: Add -mg10 multilib.
10127 2013-09-16 Xinliang David Li <davidxl@google.com>
10129 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
10130 * omp-low.c (omp_max_vf): Ditto.
10131 (expand_omp_simd): Ditto.
10132 * tree-vectorizer.c (vectorize_loops): Ditto.
10133 (gate_vect_slp): Ditto.
10134 (gate_increase_alignment): Ditto.
10135 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
10136 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
10137 (gate_tree_vectorize): Name change.
10138 (tree_vectorize): Ditto.
10139 (pass_vectorize::gate): Call new function.
10140 (pass_vectorize::execute): Ditto.
10141 * opts.c: O3 default setting change.
10142 (finish_options): Check new flag.
10143 * doc/invoke.texi: Document new flags.
10144 * common.opt: New flags.
10146 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
10148 * doc/tm.texi.in (Cond Exec Macros): Remove node.
10149 (Condition Code): Don't reference it.
10150 * doc/tm.texi: Regenerate.
10152 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
10154 PR middle-end/58418
10155 * lra-constraints.c (undo_optional_reloads): Consider all optional
10156 reload even if it did not get a hard reg.
10158 2013-09-16 Teresa Johnson <tejohnson@google.com>
10160 * dumpfile.c (dump_loc): Remove newline emission.
10161 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
10162 emission to dump_printf_loc calls where missing.
10163 (vect_mark_for_runtime_alias_test): Ditto.
10164 (vect_analyze_data_ref_dependence): Ditto.
10165 (vect_analyze_data_ref_dependences): Ditto.
10166 (vect_slp_analyze_data_ref_dependence): Ditto.
10167 (vect_slp_analyze_data_ref_dependences): Ditto.
10168 (vect_compute_data_ref_alignment): Ditto.
10169 (vect_update_misalignment_for_peel): Ditto.
10170 (vect_verify_datarefs_alignment): Ditto.
10171 (vector_alignment_reachable_p): Ditto.
10172 (vect_get_data_access_cost): Ditto.
10173 (vect_enhance_data_refs_alignment): Ditto.
10174 (vect_find_same_alignment_drs): Ditto.
10175 (vect_analyze_data_refs_alignment): Ditto.
10176 (vect_analyze_group_access): Ditto.
10177 (vect_analyze_data_ref_access): Ditto.
10178 (vect_analyze_data_ref_accesses): Ditto.
10179 (vect_prune_runtime_alias_test_list): Ditto.
10180 (vect_analyze_data_refs): Ditto.
10181 (vect_create_addr_base_for_vector_ref): Ditto.
10182 (vect_create_data_ref_ptr): Ditto.
10183 (vect_grouped_store_supported): Ditto.
10184 (vect_grouped_load_supported): Ditto.
10185 * value-prof.c (check_counter): Ditto.
10186 (check_ic_target): Ditto.
10187 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
10188 (vect_recog_widen_mult_pattern): Ditto.
10189 (vect_recog_widen_sum_pattern): Ditto.
10190 (vect_recog_over_widening_pattern): Ditto.
10191 (vect_recog_widen_shift_pattern): Ditto.
10192 (vect_recog_rotate_pattern): Ditto.
10193 (vect_recog_vector_vector_shift_pattern): Ditto.
10194 (vect_recog_divmod_pattern): Ditto.
10195 (vect_recog_mixed_size_cond_pattern): Ditto.
10196 (vect_recog_bool_pattern): Ditto.
10197 (vect_pattern_recog_1): Ditto.
10198 (vect_pattern_recog): Ditto.
10199 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
10200 (vect_is_simple_iv_evolution): Ditto.
10201 (vect_analyze_scalar_cycles_1): Ditto.
10202 (vect_get_loop_niters): Ditto.
10203 (vect_analyze_loop_1): Ditto.
10204 (vect_analyze_loop_form): Ditto.
10205 (vect_analyze_loop_operations): Ditto.
10206 (vect_analyze_loop_2): Ditto.
10207 (vect_analyze_loop): Ditto.
10208 (report_vect_op): Ditto.
10209 (vect_is_slp_reduction): Ditto.
10210 (vect_is_simple_reduction_1): Ditto.
10211 (vect_get_known_peeling_cost): Ditto.
10212 (vect_estimate_min_profitable_iters): Ditto.
10213 (vect_model_reduction_cost): Ditto.
10214 (vect_model_induction_cost): Ditto.
10215 (get_initial_def_for_induction): Ditto.
10216 (vect_create_epilog_for_reduction): Ditto.
10217 (vectorizable_reduction): Ditto.
10218 (vectorizable_induction): Ditto.
10219 (vectorizable_live_operation): Ditto.
10220 (vect_loop_kill_debug_uses): Ditto.
10221 (vect_transform_loop): Ditto.
10222 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
10223 (vect_stmt_relevant_p): Ditto.
10224 (process_use): Ditto.
10225 (vect_mark_stmts_to_be_vectorized): Ditto.
10226 (vect_model_simple_cost): Ditto.
10227 (vect_model_promotion_demotion_cost): Ditto.
10228 (vect_model_store_cost): Ditto.
10229 (vect_get_store_cost): Ditto.
10230 (vect_model_load_cost): Ditto.
10231 (vect_get_load_cost): Ditto.
10232 (vect_init_vector_1): Ditto.
10233 (vect_get_vec_def_for_operand): Ditto.
10234 (vect_finish_stmt_generation): Ditto.
10235 (vectorizable_call): Ditto.
10236 (vectorizable_conversion): Ditto.
10237 (vectorizable_assignment): Ditto.
10238 (vectorizable_shift): Ditto.
10239 (vectorizable_operation): Ditto.
10240 (vectorizable_store): Ditto.
10241 (vectorizable_load): Ditto.
10242 (vectorizable_condition): Ditto.
10243 (vect_analyze_stmt): Ditto.
10244 (vect_transform_stmt): Ditto.
10245 (vect_is_simple_use): Ditto.
10246 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
10247 (vect_can_advance_ivs_p): Ditto.
10248 (vect_update_ivs_after_vectorizer): Ditto.
10249 (vect_do_peeling_for_loop_bound): Ditto.
10250 (vect_gen_niters_for_prolog_loop): Ditto.
10251 (vect_update_inits_of_drs): Ditto.
10252 (vect_create_cond_for_alias_checks): Ditto.
10253 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
10254 (vect_build_slp_tree_1): Ditto.
10255 (vect_supported_load_permutation_p): Ditto.
10256 (vect_analyze_slp_instance): Ditto.
10257 (vect_analyze_slp): Ditto.
10258 (vect_make_slp_decision): Ditto.
10259 (vect_detect_hybrid_slp): Ditto.
10260 (vect_bb_vectorization_profitable_p): Ditto.
10261 (vect_slp_analyze_bb_1): Ditto.
10262 (vect_update_slp_costs_according_to_vf): Ditto.
10263 (vect_get_mask_element): Ditto.
10264 (vect_transform_slp_perm_load): Ditto.
10265 (vect_schedule_slp_instance): Ditto.
10266 (vect_schedule_slp): Ditto.
10267 (vect_slp_transform_bb): Ditto.
10268 * profile.c (read_profile_edge_counts): Ditto.
10269 (compute_branch_probabilities): Ditto.
10270 * coverage.c (get_coverage_counts): Ditto.
10272 2013-09-16 Diego Novillo <dnovillo@google.com>
10274 * tree-core.h: Add missing comment lines from refactoring of tree.h.
10276 2013-09-16 Jan Hubicka <jh@suse.cz>
10278 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
10279 abstract functions; for static functions check the presence of body.
10281 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
10283 * config/aarch64/aarch64-simd-builtins.def (fma): New.
10284 * config/aarch64/aarch64-simd.md
10285 (aarch64_mla_elt<mode>): New.
10286 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
10287 (aarch64_mls_elt<mode>): Likewise.
10288 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
10289 (aarch64_fma4_elt<mode>): Likewise.
10290 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
10291 (aarch64_fma4_elt_to_128v2df): Likewise.
10292 (aarch64_fma4_elt_to_64df): Likewise.
10293 (fnma<mode>4): Likewise.
10294 (aarch64_fnma4_elt<mode>): Likewise.
10295 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
10296 (aarch64_fnma4_elt_to_128v2df): Likewise.
10297 (aarch64_fnma4_elt_to_64df): Likewise.
10298 * config/aarch64/iterators.md (VDQSF): New.
10299 * config/aarch64/arm_neon.h
10300 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
10301 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
10303 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
10305 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
10306 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
10307 (aarch64_mul3_elt_to_128df): Likewise.
10308 (aarch64_mul3_elt_to_64v2df): Likewise.
10309 * config/aarch64/iterators.md (VEL): Also handle DFmode.
10311 (VMUL_CHANGE_NLANES) Likewise.
10314 * config/aarch64/arm_neon.h
10315 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
10317 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
10319 * config/aarch64/arm_neon.h
10320 (vcvtx_high_f32_f64): Fix parameters.
10322 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
10323 Uros Bizjak <ubizjak@gmail.com>
10325 * config/alpha.c: Include tree-ssanames.h.
10327 2013-09-16 Richard Biener <rguenther@suse.de>
10329 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
10330 (dot_rdg_1): Handle control_dd.
10331 (create_edge_for_control_dependence): New function.
10332 (create_rdg_edges): Add control dependences if asked for.
10333 (build_rdg): Likewise.
10334 (generate_loops_for_partition): If there are not necessary
10335 control stmts remove all their dependencies.
10336 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
10337 (distribute_loop): Pass on control dependences.
10338 (tree_loop_distribution): Compute control dependences and remove
10339 restriction on number of loop nodes.
10341 2013-09-16 Jakub Jelinek <jakub@redhat.com>
10343 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
10344 for internal calls.
10346 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
10348 * cse.c (try_const_anchors): Punt on CC modes.
10350 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
10352 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
10354 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
10357 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
10358 calls to word_mode.
10360 2013-09-14 Iain Sandoe <iain@codesourcery.com>
10363 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
10365 (darwin_objc1_section): Likewise.
10366 (darwin_file_end): Emit Image Info section when required.
10368 2013-09-14 Jan Hubicka <jh@suse.cz>
10370 * tree-into-ssa.c (gate_into_ssa): New.
10371 (pass_data_build_ssa): Use it.
10372 * cgraph.h (expand_thunk): Update prototype.
10373 * cgraphunit.c (analyze_function): Expand thunks early.
10374 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
10375 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
10376 set lowered flag; do not add new function.
10377 (assemble_thunks_and_aliases): Update.
10378 * tree-ssa.c (gate_init_datastructures): New gate.
10379 (pass_data_init_datastructures): Use it.
10381 2013-09-14 Iain Sandoe <iain@codesourcery.com>
10384 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
10385 xmm register set described in the psABI.
10387 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
10389 * dwarf2out.c (should_emit_struct_debug): Add check
10390 for type_decl variable is not NULL.
10392 2013-09-13 Jacek Caban <jacek@codeweavers.com>
10394 * config.gcc: Use new winnt-c.c target hooks
10395 * config/t-winnt: New file
10396 * config/winnt-c.c: New file
10397 * doc/tm.texi.in: Document new hook
10398 * doc/tm.texi: Regenerated
10400 2013-09-13 Jan Hubicka <jh@suse.cz>
10402 PR middle-end/58094
10403 * ipa-inline.c (check_callers): New function.
10404 (check_caller_edge): Remove.
10405 (want_inline_function_to_all_callers_p): Also permit alises that are
10407 (inline_to_all_callers): Terminate the walk when devirtualization
10408 introduce new calls.
10410 2013-09-13 Jan Hubicka <jh@suse.cz>
10412 * ipa-inline-analysis.c (struct growth_data): Add node.
10413 (do_estimate_growth_1): Fix detection of recursion.
10415 2013-09-13 Jakub Jelinek <jakub@redhat.com>
10417 PR tree-optimization/58392
10418 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
10419 to avoid shadowing of outer loop variable. If
10420 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
10421 replace_by_duplicate_decl simduid of loops that have it set and
10422 set dest_cfun->has_simduid_loops and/or
10423 dest_cfun->has_force_vect_loops.
10424 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
10425 instead of maybe_lookup_decl.
10426 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
10427 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
10428 call. Copy over force_vect and copy and remap simduid. Set
10429 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
10430 (copy_cfg_body): Remove blocks_to_copy argument. Use
10431 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
10432 caller. Don't set cfun->has_simduid_loops and/or
10433 cfun->has_force_vect_loops here.
10434 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
10436 (expand_call_inline, tree_function_versioning): Adjust copy_body
10439 2013-09-13 Martin Jambor <mjambor@suse.cz>
10442 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
10443 the assert if the edge was a speculative one.
10445 2013-09-13 Richard Biener <rguenther@suse.de>
10447 * tree-data-ref.h (known_dependences_p): Move here ...
10448 * tree-loop-distribution.c (known_dependences_p): ... from here.
10449 (dump_rdg_component, debug_rdg_component): Remove.
10450 (dump_rdg): Adjust.
10451 (generate_loops_for_partition): Use gimple_uid instead of
10452 relying on matching stmt visit order.
10453 (rdg_build_partitions): Take starting stmt vector.
10454 (ldist_gen): Merge into ...
10455 (distribute_loop): ... this function. Do not compute starting
10457 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
10459 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10461 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
10462 Set type attribute correctly. Set predicable_short_it attribute.
10463 (cmpsi_shiftsi): Remove %? from output template.
10465 2013-09-13 Richard Biener <rguenther@suse.de>
10467 * tree-loop-distribution.c (struct rdg_component,
10468 rdg_defs_used_in_other_loops_p, free_rdg_components,
10469 rdg_build_components): Remove.
10470 (stmts_from_loop): Do not record virtual PHIs.
10471 (generate_loops_for_partition): Skip virtual PHIs.
10472 (build_rdg_partition_for_component): Rename to ...
10473 (build_rdg_partition_for_vertex): ... this and adjust.
10474 (rdg_build_partitions): Take a vector of starting vertices
10475 instead of components. Remove unnecessary leftover handling.
10476 (ldist_gen): Do not build components or record other stores.
10477 (distribute_loop): Do not distribute loops containing stmts
10480 2013-09-13 Christian Bruel <christian.bruel@st.com>
10483 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
10485 2013-09-13 Kai Tietz <ktietz@redhat.com>
10487 * config.gcc: Separate cases for mingw and cygwin targets,
10488 and add 64-bit cygwin target case.
10490 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
10491 dll-export inline-functions.
10492 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
10494 2013-09-13 Jeff Law <law@redhat.com>
10496 PR middle-end/58387
10498 2013-09-06 Jeff Law <law@redhat.com>
10500 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
10501 edge implied equivalences into successor phis.
10503 2013-09-12 DJ Delorie <dj@redhat.com>
10505 * config/rl78/rl78-virt.md: Change from | to \; for asm line
10508 2013-09-12 Brooks Moses <bmoses@google.com>
10511 * Makefile.in: Do not install driver binaries in $(target)/bin.
10513 2013-09-12 DJ Delorie <dj@redhat.com>
10515 * config/rl78/rl78.opt (mrelax): New.
10516 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
10517 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
10519 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
10521 (rl78_expand_epilogue): Likewise.
10523 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
10525 PR middle-end/58335
10526 * lra-eliminations.c (remove_reg_equal_offset_note): New.
10527 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
10528 pointer elimination with using remove_reg_equal_offset_note.
10530 2013-09-12 DJ Delorie <dj@redhat.com>
10532 * config/msp430/: New port.
10533 * config.gcc (msp430): Added.
10534 * doc/invoke.texi: Document MSP430 options.
10535 * doc/install.texi: Document msp430-elf
10536 * doc/md.texi: Document msp430-elf
10537 * doc/contrib.texi: Document msp430-elf
10539 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
10542 2013-09-12 Martin Jambor <mjambor@suse.cz>
10545 * ipa-prop.c (remove_described_reference): Give up if the edge in the
10546 reference descriptor is NULL.
10547 (ipa_edge_removal_hook): If owning a reference descriptor, set its
10550 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
10552 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
10553 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
10554 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
10555 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
10556 make_temp_ssa_name): move to tree-ssanames.h
10557 * tree-ssa-alias.h: Move prototype.
10558 * tree-ssa.h: Include tree-ssanames.h.
10559 * tree-ssanames.c (FREE_SSANAMES): Move to here.
10560 * tree-ssanames.h: New. Move items from tree-flow*.h
10561 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
10563 2013-09-12 Richard Biener <rguenther@suse.de>
10565 PR tree-optimization/58404
10566 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
10567 propagate non-invariant addresses into dereferences wrapped
10568 in component references.
10570 2013-09-12 Richard Biener <rguenther@suse.de>
10572 PR tree-optimization/58402
10573 * passes.def: Move pass_late_warn_uninitialized later.
10575 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
10577 * tree-ssa.h: New. Move content from tree-flow.h and
10578 tree-flow-inline.h.
10579 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
10580 Move prototypes belonging to tree-ssa.c.
10581 * tree-flow-inline.h (redirect_edge_var_map_def,
10582 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
10584 * gimple.h: Adjust prototypes.
10585 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
10587 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
10588 * tree.h: Move prototype to tree-ssa.h.
10589 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
10590 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
10591 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
10592 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
10593 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
10594 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
10595 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
10596 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
10597 graphite-blocking.c, graphite-clast-to-gimple.c,
10598 graphite-dependences.c, graphite-interchange.c,
10599 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
10600 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
10601 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
10602 ipa-reference.c, ipa-split.c, ipa-utils.c,
10603 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
10604 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
10605 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
10606 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
10607 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
10608 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
10609 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
10610 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
10611 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
10612 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
10613 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
10614 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
10615 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
10616 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
10617 value-prof.c, var-tracking.c,
10618 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
10620 2013-09-12 Richard Biener <rguenther@suse.de>
10622 PR tree-optimization/58396
10623 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
10624 (build_rdg): Take a loop-nest parameter, fix memleaks.
10625 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
10627 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
10629 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
10632 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
10634 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
10635 The FMA instruction names should have a 'v' prefix.
10637 2013-09-12 Richard Biener <rguenther@suse.de>
10639 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
10640 (dot_rdg): Use popen instead of system in optional code.
10641 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
10642 (already_processed_vertex_p): Adjust.
10643 (has_anti_or_output_dependence, predecessor_has_mem_write,
10644 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
10645 rdg_flag_uses): Remove.
10646 (rdg_flag_vertex): Simplify.
10647 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
10649 (build_rdg_partition_for_component): Process the first vertex
10650 of a component only.
10651 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
10653 2013-09-12 Alan Modra <amodra@gmail.com>
10655 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
10657 2013-09-11 DJ Delorie <dj@redhat.com>
10658 Nick Clifton <nickc@redhat.com>
10660 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
10661 (rl78_stack_based_mem): New.
10662 * config/rl78/constraints.md (Iv08): New.
10671 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
10673 (movsi): Change from expand to insn-and-split.
10674 (ashrsi3): Clobber AX.
10678 * config/rl78/rl78.md (CC_REG): Fix.
10679 (addsi3): Allow memory and immediate operands.
10680 (addsi3_internal): Split into...
10681 (addsi3_internal_virt): ...new, and ...
10682 (addsi3_internal_real): ...new.
10684 (subsi3_internal_virt): New.
10685 (subsi3_internal_real): New.
10686 (mulsi3): Add memory operand.
10687 (mulsi3_rl78): Likewise.
10688 (mulsi3_g13): Likewise.
10689 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
10690 (cbranchqi4_real): Add more constraint options.
10691 (cbranchhi4_real): Expand pattern.
10692 (cbranchhi4_real_signed): New.
10693 (cbranchhi4_real_inverted): New.
10694 (cbranchsi4_real_lt): New.
10695 (cbranchsi4_real_ge): New.
10696 (cbranchsi4_real_signed): New.
10697 (cbranchsi4_real): New.
10699 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
10701 (lshrsi3_virt): Likewise.
10702 (ashlsi3_virt): Likewise.
10703 (cbranchqi4_virt_signed): New.
10704 (cbranchhi4_virt_signed): New.
10705 (cbranchsi4_virt): New.
10706 * config/rl78/rl78.c: Whitespace fixes throughout.
10707 (move_elim_pass): New.
10708 (pass_data_rl78_move_elim): New.
10709 (pass_rl78_move_elim): New.
10710 (make_pass_rl78_move_elim): New.
10711 (rl78_devirt_info): Run devirt earlier.
10712 (rl78_move_elim_info): New.
10713 (rl78_asm_file_start): Register it.
10714 (rl78_split_movsi): New.
10715 (rl78_as_legitimate_address): Allow virtual base registers when
10717 (rl78_addr_space_convert): Remove spurious debug stuff.
10718 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
10719 (rl78_print_operand): More cases for not printing '#'.
10720 (rl78_expand_compare): Remove most of the logic.
10721 (content_memory): New.
10722 (clear_content_memory): New.
10723 (get_content_index): New.
10724 (get_content_name): New.
10725 (display_content_memory): New.
10726 (update_content): New.
10727 (record_content): New.
10728 (already_contains): New.
10729 (insn_ok_now): Re-recog insns with virtual registers.
10730 (add_postponed_content_update): New.
10731 (process_postponed_content_update): New.
10732 (gen_and_emit_move): New.
10733 (transcode_memory_rtx): Record new location content.
10734 Use gen_and_emit_move.
10735 (force_into_acc): New.
10736 (move_to_acc): Use gen_and_emit_move.
10737 (move_from_acc): Likewise.
10738 (move_acc_to_reg): Likewise.
10739 (move_to_x): Likewise.
10740 (move_to_hl): Likewise.
10741 (move_to_de): Likewise.
10742 (rl78_alloc_physical_registers_op1): Record location content.
10743 (has_constraint): New.
10744 (rl78_alloc_physical_registers_op2): Record location content.
10745 Optimize use of HL.
10746 (rl78_alloc_physical_registers_ro1): Likewise.
10747 (rl78_alloc_physical_registers_cmp): Likewise.
10748 (rl78_alloc_physical_registers_umul): Likewise.
10749 (rl78_alloc_address_registers_macax): New.
10750 (rl78_alloc_physical_registers): Initialize and set location
10751 content memory as needed.
10752 (rl78_reorg): Make sure split2 is called.
10753 (rl78_rtx_costs): New.
10755 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
10757 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
10758 for (not (neg ...)) and (neg (not ...)) cases.
10760 2013-09-11 Richard Biener <rguenther@suse.de>
10762 PR middle-end/58377
10763 * passes.def: Split critical edges before late uninit warning passes.
10764 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
10766 2013-09-11 Jakub Jelinek <jakub@redhat.com>
10768 PR tree-optimization/58385
10769 * fold-const.c (build_range_check): If both low and high are NULL,
10770 use omit_one_operand_loc to preserve exp side-effects.
10772 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10774 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
10776 2013-09-11 Richard Biener <rguenther@suse.de>
10778 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
10779 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
10780 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
10781 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
10782 stmts_from_loop, known_dependences_p, build_empty_rdg,
10783 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
10784 * tree-loop-distribution.c: ... here.
10785 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
10786 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
10787 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
10788 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
10789 * tree-loop-distribution.c: ... here.
10790 * tree-loop-distribution.c: Include gimple-pretty-print.h.
10791 (struct partition_s): Add loops member.
10792 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
10793 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
10794 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
10796 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
10797 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
10798 Sergey Lega <sergey.s.lega@intel.com>
10799 Anna Tikhonova <anna.tikhonova@intel.com>
10800 Ilya Tocar <ilya.tocar@intel.com>
10801 Andrey Turetskiy <andrey.turetskiy@intel.com>
10802 Ilya Verbin <ilya.verbin@intel.com>
10803 Kirill Yukhin <kirill.yukhin@intel.com>
10804 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
10806 * config/i386/constraints.md (k): New.
10808 * config/i386/i386.c (const regclass_map): Add new mask registers.
10809 (dbx_register_map): Ditto.
10810 (dbx64_register_map): Ditto.
10811 (svr4_dbx_register_map): Ditto.
10812 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
10814 (ix86_preferred_reload_class): Disable constants for mask registers.
10815 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
10816 (ix86_hard_regno_mode_ok): Support new mask registers.
10817 (x86_order_regs_for_local_alloc): Ditto.
10818 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
10819 (FIXED_REGISTERS): Add new mask registers.
10820 (CALL_USED_REGISTERS): Ditto.
10821 (REG_ALLOC_ORDER): Ditto.
10822 (VALID_MASK_REG_MODE): New.
10823 (FIRST_MASK_REG): Ditto.
10824 (LAST_MASK_REG): Ditto.
10825 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
10826 (MAYBE_MASK_CLASS_P): New.
10827 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
10828 (REG_CLASS_CONTENTS): Ditto.
10829 (MASK_REGNO_P): New.
10830 (ANY_MASK_REG_P): Ditto.
10831 (HI_REGISTER_NAMES): Add new mask registers.
10832 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
10833 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
10835 (attribute "type"): Add mskmov, msklog.
10836 (attribute "length_immediate"): Support them.
10837 (attribute "memory"): Ditto.
10838 (attribute "prefix_0f"): Ditto.
10839 (*movhi_internal): Support new mask registers.
10840 (*movqi_internal): Ditto.
10841 (define_split): Split out clobber pattern is a logic
10842 insn on mask registers.
10843 (*k<logic><mode>): New.
10844 (*andhi_1): Extend to support mask regs.
10845 (*andqi_1): Extend to support mask regs.
10846 (kandn<mode>): New.
10847 (define_split): Split and-not to and and not if operands
10849 (*<code><mode>_1): Separate HI mode to new pattern...
10850 (*<code>hi_1): This.
10851 (*<code>qi_1): Extend to support mask regs.
10852 (kxnor<mode>): New.
10853 (kortestzhi): Ditto.
10854 (kortestchi): Ditto.
10856 (*one_cmpl<mode>2_1): Remove HImode and handle it...
10857 (*one_cmplhi2_1): ...Here, now with mask registers support.
10858 (*one_cmplqi2_1): Support new mask registers.
10859 (HI/QImode arithmetics splitter): Don't split if mask registers
10861 (HI/QImode not splitter): Ditto.
10862 * config/i386/predicated.md (mask_reg_operand): New.
10863 (general_reg_operand): Ditto.
10865 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
10867 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
10868 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
10870 2013-09-10 Jeff Law <law@redhat.com>
10872 PR tree-optimization/58380
10873 * tree-ssa-threadupdate.c (thread_block): Recognize another case
10874 of threading through a buried loop header.
10876 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
10877 return value for single successor case.
10879 2013-09-10 Jan Hubicka <jh@suse.cz>
10881 * ipa-devirt.c (ipa_devirt): Enable with LTO.
10883 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
10886 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
10887 support conditional execution.
10888 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
10890 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
10892 * lra.c (lra): Clear lra_optional_reload_pseudos before every
10894 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
10895 Check destination too to check move insn.
10896 (undo_optional_reloads): Add check that the original peudo did not
10897 changed its allocation and the optional reload was inherited on last
10898 inheritance pass. Break loop after deciding to keep optional reload.
10899 (lra_undo_inheritance): Add check that inherited pseudo still in
10902 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
10904 * config/aarch64/aarch64.md (generic_sched): New.
10905 * config/aarch64/aarch64-generic.md (load): Make conditional
10906 on generic_sched attribute.
10907 (nonload): Likewise.
10909 2013-09-10 Jan Hubicka <jh@suse.cz>
10911 * lto-cgraph.c: Include ipa-utils.h.
10912 (compute_ltrans_boundary): Also add possible targets into the boundary.
10914 2013-09-10 Jan Hubicka <jh@suse.cz>
10916 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
10917 VAR_DECL of vtable rather than full expression.
10919 2013-09-10 Jan Hubicka <jh@suse.cz>
10920 Paolo Carlini <paolo.carlini@oracle.com>
10922 * cgraphunit.c (analyze_functions): Save input_location, set it
10923 to UNKNOWN_LOCATION and restore it at the end.
10925 2013-09-10 Martin Jambor <mjambor@suse.cz>
10927 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
10928 represented by a thunk.
10930 2013-09-10 Jeff Law <law@redhat.com>
10932 PR tree-optimization/58343
10933 * tree-ssa-threadupdate.c (thread_block): Identify and disable
10934 jump threading requests through loop headers buried in the middle
10935 of a jump threading path.
10937 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
10938 in return value/type.
10940 2013-09-10 Jakub Jelinek <jakub@redhat.com>
10942 PR rtl-optimization/58365
10943 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
10944 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
10947 2013-09-10 Richard Biener <rguenther@suse.de>
10949 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
10950 * tree-data-ref.c (create_rdg_vertices): Collect all data
10951 references, signal failure to the caller, use data-ref API.
10952 (build_rdg): Compute data references only once. Maintain lifetime
10953 of data references and data dependences from within RDG.
10954 (free_rdg): Free dependence relations.
10955 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
10956 inventing extra dependences.
10957 (distribute_loop): Update for RDG API changes.
10959 2013-09-10 Kai Tietz <ktietz@redhat.com>
10961 * doc/invoke.texi (fms-extensions): Document changed
10962 behavior for ms-abi targets.
10963 * config/i386/i386.c (ix86_option_override_internal):
10964 Set default value of option -fms-extension for ms-abi targets.
10966 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
10968 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
10970 2013-09-10 Alan Modra <amodra@gmail.com>
10973 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
10975 2013-09-10 Alan Modra <amodra@gmail.com>
10977 * config/rs6000/predicates.md (add_cint_operand): New.
10978 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
10979 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
10980 using add_cint_operand.
10981 (largetoc_high_plus_aix): Likewise.
10983 2013-09-09 Jakub Jelinek <jakub@redhat.com>
10985 PR tree-optimization/58364
10986 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
10987 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
10988 the current range can't be an unconditional true or false.
10990 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
10992 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
10994 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
10996 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
10998 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
11001 * doc/invoke.texi (-Wdelete-incomplete): Document it.
11003 2013-09-09 Ian Bolton <ian.bolton@arm.com>
11005 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
11006 NO_REGS for immediate that can't be moved directly into FP_REGS.
11008 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11010 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
11011 comparison with negated operand.
11012 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
11015 2013-09-09 Richard Biener <rguenther@suse.de>
11017 PR middle-end/58326
11018 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
11019 of a subloop record all its block as affecting loop-closed SSA form.
11021 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
11023 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
11024 of an rtx/bitpos pair.
11025 (store_fixed_bit_field): Update accordingly.
11027 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
11029 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
11031 * builtins.c (expand_errno_check): Likewise.
11032 * dwarf2cfi.c (init_return_column_size): Likewise.
11033 * except.c (sjlj_mark_call_sites): Likewise.
11034 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
11035 * lra-constraints.c (emit_inc): Likewise.
11036 * ree.c (combine_set_extension): Likewise.
11037 * regmove.c (fixup_match_2): Likewise.
11038 * reload1.c (inc_for_reload): Likewise.
11040 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
11042 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
11043 (force_to_mode, simplify_shift_const_1, simplify_comparison):
11044 Use gen_int_mode with the mode of the associated simplify_* call.
11045 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
11046 * expmed.c (expand_shift_1): Likewise.
11047 * function.c (instantiate_virtual_regs_in_insn): Likewise.
11048 * loop-iv.c (iv_number_of_iterations): Likewise.
11049 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
11050 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
11052 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
11054 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
11055 of the associated expand_* call.
11056 (asan_emit_stack_protection): Likewise.
11057 * builtins.c (round_trampoline_addr): Likewise.
11058 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
11059 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
11060 (emit_store_flag): Likewise.
11061 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
11062 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
11064 * function.c (instantiate_virtual_regs_in_insn): Likewise.
11065 * ifcvt.c (noce_try_store_flag_constants): Likewise.
11066 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
11067 * modulo-sched.c (generate_prolog_epilog): Likewise.
11068 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
11069 (expand_ctz, expand_ffs, expand_unop): Likewise.
11071 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
11073 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
11074 of the associated gen_rtx_* call.
11075 * caller-save.c (init_caller_save): Likewise.
11076 * combine.c (find_split_point, make_extraction): Likewise.
11077 (make_compound_operation): Likewise.
11078 * dwarf2out.c (mem_loc_descriptor): Likewise.
11079 * explow.c (plus_constant, probe_stack_range): Likewise.
11080 * expmed.c (expand_mult_const): Likewise.
11081 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
11082 * reload1.c (init_reload): Likewise.
11083 * valtrack.c (cleanup_auto_inc_dec): Likewise.
11084 * var-tracking.c (adjust_mems): Likewise.
11085 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
11086 rather than gen_rtx_fmt_ee.
11088 2013-09-09 Jan Hubicka <jh@suse.cz>
11090 PR middle-end/58294
11091 * value-prof.c (gimple_ic): Copy also abnormal edges.
11093 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
11095 * asan.c (asan_shadow_cst): Use gen_int_mode.
11097 2013-09-08 Jan Hubicka <jh@suse.cz>
11099 * ipa-profile.c: Add toplevel comment.
11100 (ipa_propagate_frequency_1): Be more conservative when profile is read.
11101 (contains_hot_call_p): New function.
11102 (ipa_propagate_frequency): Set frequencies based on counts when
11104 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
11105 profile; do not tamper with profile after inlining if it is read.
11107 2013-09-08 Jan Hubicka <jh@suse.cz>
11109 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
11112 2013-09-08 Jan Hubicka <jh@suse.cz>
11114 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
11115 summary generation.
11117 2013-09-08 Jeff Law <law@redhat.com>
11120 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
11123 2013-09-08 Andi Kleen <ak@linux.intel.com>
11125 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
11127 2013-09-08 Jan Hubicka <jh@suse.cz>
11129 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
11131 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
11133 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
11134 for non-debug insns.
11135 * lra.c (new_insn_reg): Take the containing insn as a parameter.
11136 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
11137 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
11140 2013-09-08 Jan Hubicka <jh@suse.cz>
11142 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
11143 targets and devirtualize to BUILT_IN_UNREACHABLE.
11144 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
11145 * ipa.c (walk_polymorphic_call_targets): New function.
11146 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
11147 functions; use the new timevar.
11148 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
11149 was removed from the program.
11150 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
11151 not consider it in the walk when its vtable is dead.
11152 (possible_polymorphic_call_targets_1): Pass anonymous flag to
11154 (devirt_variable_node_removal_hook): New function.
11155 (possible_polymorphic_call_targets): Also register
11156 devirt_variable_node_removal_hook.
11157 (ipa_devirt): Do not do non-speculative devirtualization.
11158 (gate_ipa_devirt): One execute if devirtualizing speculatively.
11160 2013-09-08 Jan Hubicka <jh@suse.cz>
11162 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
11163 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
11164 varpool_remove_variable_insertion_hook): Declare.
11165 * varpool.c (varpool_node_hook_list): New structure.
11166 (first_varpool_node_removal_hook,
11167 first_varpool_variable_insertion_hook): New variables.
11168 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
11169 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
11170 varpool_remove_variable_insertion_hook,
11171 varpool_call_variable_insertion_hooks): New functions.
11172 (varpool_remove_node): Use it.
11174 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
11177 * diagnostic.c (diagnostic_build_prefix): When s.file is
11178 "<built-in>" don't output line and column numbers.
11180 2013-09-06 Jan Hubicka <jh@suse.cz>
11182 * cgraphunit.c (expand_thunk): Get body before touching arguments.
11183 * lto-streamer-out.c: Stream thunks, too.
11184 * lto-streamer-in.c (input_function): Pop cfun here
11185 (lto_read_body): Instead of here.
11187 2013-09-06 Caroline Tice <cmtice@google.com>
11189 * doc/install.texi: Add documentation for the --enable-vtable-verify
11190 and the --disable-libvtv configure options.
11192 2013-09-06 Jeff Law <law@redhat.com>
11194 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
11195 edge implied equivalences into successor phis.
11197 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
11199 * resource.c (mark_referenced_resources): Handle COND_EXEC.
11201 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
11203 * resource.c (mark_target_live_regs): Compute resources taking
11204 into account if a call is predicated or not.
11206 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
11208 * toplev.c (output_stack_usage): Be prepared for suffixes created by
11209 the compiler in the function names.
11211 2013-09-06 Jan Hubicka <jh@suse.cz>
11213 PR middle-end/58094
11214 * ipa-inline.c (has_caller_p): New function.
11215 (want_inline_function_to_all_callers_p): Use it.
11216 (sum_callers, inline_to_all_callers): Break out from ...
11217 (ipa_inline): ... here.
11219 2013-09-06 Jan Hubicka <jh@suse.cz>
11221 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
11222 only when AVX is enabled.
11224 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
11226 * config/aarch64/aarch64.md
11227 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
11228 is fpsimd_<load/store>2.
11229 (load_pair<mode>): Likewise.
11230 (store_pair<mode>): Likewise.
11232 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
11234 * config/arm/types.md (type): Add "mrs" type.
11235 * config/aarch64/aarch64.md
11236 (aarch64_load_tp_hard): Make type "mrs".
11237 * config/arm/arm.md
11238 (load_tp_hard): Make type "mrs".
11239 * config/arm/cortex-a15.md: Update with new attributes.
11240 * config/arm/cortex-a5.md: Update with new attributes.
11241 * config/arm/cortex-a53.md: Update with new attributes.
11242 * config/arm/cortex-a7.md: Update with new attributes.
11243 * config/arm/cortex-a8.md: Update with new attributes.
11244 * config/arm/cortex-a9.md: Update with new attributes.
11245 * config/arm/cortex-m4.md: Update with new attributes.
11246 * config/arm/cortex-r4.md: Update with new attributes.
11247 * config/arm/fa526.md: Update with new attributes.
11248 * config/arm/fa606te.md: Update with new attributes.
11249 * config/arm/fa626te.md: Update with new attributes.
11250 * config/arm/fa726te.md: Update with new attributes.
11252 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
11254 * config/aarch64/aarch64.md
11255 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
11256 (*movtf_aarch64): Likewise.
11257 * config/arm/arm.md
11258 (thumb1_movdi_insn): Use "multiple" for type where more than one
11259 instruction is used for a move.
11260 (*arm32_movhf): Likewise.
11261 (*thumb_movdf_insn): Likewise.
11263 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
11265 * config/arm/types.md (type): Rename fcpys to fmov.
11266 * config/arm/vfp.md
11267 (*arm_movsi_vfp): Rename type fcpys as fmov.
11268 (*thumb2_movsi_vfp): Likewise
11269 (*movhf_vfp_neon): Likewise
11270 (*movhf_vfp): Likewise
11271 (*movsf_vfp): Likewise
11272 (*thumb2_movsf_vfp): Likewise
11273 (*movsfcc_vfp): Likewise
11274 (*thumb2_movsfcc_vfp): Likewise
11275 * config/aarch64/aarch64-simd.md
11276 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
11277 * config/aarch64/aarch64.md
11278 (*movsi_aarch64): Replace type mov_reg with fmovs.
11279 (*movdi_aarch64): Likewise
11280 (*movsf_aarch64): Likewise
11281 (*movdf_aarch64): Likewise
11283 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
11284 * config/arm/iwmmxt.md
11285 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
11286 * config/arm/arm1020e.md: Update with new attributes.
11287 * config/arm/cortex-a15-neon.md: Update with new attributes.
11288 * config/arm/cortex-a5.md: Update with new attributes.
11289 * config/arm/cortex-a53.md: Update with new attributes.
11290 * config/arm/cortex-a7.md: Update with new attributes.
11291 * config/arm/cortex-a8-neon.md: Update with new attributes.
11292 * config/arm/cortex-a9.md: Update with new attributes.
11293 * config/arm/cortex-m4-fpu.md: Update with new attributes.
11294 * config/arm/cortex-r4f.md: Update with new attributes.
11295 * config/arm/marvell-pj4.md: Update with new attributes.
11296 * config/arm/vfp11.md: Update with new attributes.
11298 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
11300 * config/aarch64/aarch64.md
11301 (*madd<mode>): Fix type attribute.
11302 (*maddsi_uxtw): Likewise.
11303 (*msub<mode>): Likewise.
11304 (*msubsi_uxtw): Likewise.
11305 (<su_optab>maddsidi4): Likewise.
11306 (<su_optab>msubsidi4): Likewise.
11308 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
11310 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
11311 * config/arm/arm.md (core_cycles): Remove fdiv.
11312 * config/arm/vfp.md:
11313 (*sqrtsf2_vfp): Update for attribute changes.
11314 (*sqrtdf2_vfp): Likewise.
11315 * config/aarch64/aarch64.md:
11316 (sqrt<mode>2): Update for attribute changes.
11317 * config/arm/arm1020e.md: Update with new attributes.
11318 * config/arm/cortex-a15-neon.md: Update with new attributes.
11319 * config/arm/cortex-a5.md: Update with new attributes.
11320 * config/arm/cortex-a53.md: Update with new attributes.
11321 * config/arm/cortex-a7.md: Update with new attributes.
11322 * config/arm/cortex-a8-neon.md: Update with new attributes.
11323 * config/arm/cortex-a9.md: Update with new attributes.
11324 * config/arm/cortex-m4-fpu.md: Update with new attributes.
11325 * config/arm/cortex-r4f.md: Update with new attributes.
11326 * config/arm/marvell-pj4.md: Update with new attributes.
11327 * config/arm/vfp11.md: Update with new attributes.
11329 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
11331 * config/arm/types.md
11332 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
11333 * config/aarch64/aarch64.md
11334 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
11336 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
11337 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
11338 (float<GPI:mode><GPF:mode>2): Likewise.
11339 * config/arm/vfp.md
11340 (*truncsisf2_vfp): Update with new attributes.
11341 (*truncsidf2_vfp): Likewise.
11342 (fixuns_truncsfsi2): Likewise.
11343 (fixuns_truncdfsi2): Likewise.
11344 (*floatsisf2_vfp): Likewise.
11345 (*floatsidf2_vfp): Likewise.
11346 (floatunssisf2): Likewise.
11347 (floatunssidf2): Likewise.
11348 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
11349 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
11350 * config/arm/arm1020e.md: Update with new attributes.
11351 * config/arm/cortex-a15-neon.md: Update with new attributes.
11352 * config/arm/cortex-a5.md: Update with new attributes.
11353 * config/arm/cortex-a53.md: Update with new attributes.
11354 * config/arm/cortex-a7.md: Update with new attributes.
11355 * config/arm/cortex-a8-neon.md: Update with new attributes.
11356 * config/arm/cortex-a9.md: Update with new attributes.
11357 * config/arm/cortex-m4-fpu.md: Update with new attributes.
11358 * config/arm/cortex-r4f.md: Update with new attributes.
11359 * config/arm/marvell-pj4.md: Update with new attributes.
11360 * config/arm/vfp11.md: Update with new attributes.
11362 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
11364 * config/aarch64/arm_neon.h
11365 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
11366 (vqtbx<1,2,3,4><q>_s8): Likewise.
11368 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
11370 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
11371 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
11372 (add<mode>3): Add type attribute.
11373 (add<mode>3): Likewise.
11374 (usadd<mode>3): Likewise.
11375 (ssadd<mode>3): Likewise.
11376 (sub<mode>3): Likewise.
11377 (sub<mode>3): Likewise.
11378 (ussub<mode>3): Likewise.
11379 (sssub<mode>3): Likewise.
11380 (ssmulsa3): Likewise.
11381 (usmulusa3): Likewise.
11382 (arm_usatsihi): Likewise.
11383 * config/arm/vfp.md
11384 (*movdi_vfp): Add types for all instructions.
11385 (*movdi_vfp_cortexa8): Likewise.
11386 (*movhf_vfp_neon): Likewise.
11387 (*movhf_vfp): Likewise.
11388 (*movdf_vfp): Likewise.
11389 (*thumb2_movdf_vfp): Likewise.
11390 (*thumb2_movdfcc_vfp): Likewise.
11391 * config/arm/arm.md: Add type attribute to all insn patterns.
11392 (*thumb1_adddi3): Add type attribute.
11393 (*arm_adddi3): Likewise.
11394 (*adddi_sesidi_di): Likewise.
11395 (*adddi_zesidi_di): Likewise.
11396 (*thumb1_addsi3): Likewise.
11397 (addsi3_compare0): Likewise.
11398 (*addsi3_compare0_scratch): Likewise.
11399 (*compare_negsi_si): Likewise.
11400 (cmpsi2_addneg): Likewise.
11401 (*addsi3_carryin_<optab>): Likewise.
11402 (*addsi3_carryin_alt2_<optab>): Likewise.
11403 (*addsi3_carryin_clobercc_<optab>): Likewise.
11404 (*subsi3_carryin): Likewise.
11405 (*subsi3_carryin_const): Likewise.
11406 (*subsi3_carryin_compare): Likewise.
11407 (*subsi3_carryin_compare_const): Likewise.
11408 (*arm_subdi3): Likewise.
11409 (*thumb_subdi3): Likewise.
11410 (*subdi_di_zesidi): Likewise.
11411 (*subdi_di_sesidi): Likewise.
11412 (*subdi_zesidi_di): Likewise.
11413 (*subdi_sesidi_di): Likewise.
11414 (*subdi_zesidi_ze): Likewise.
11415 (thumb1_subsi3_insn): Likewise.
11416 (*arm_subsi3_insn): Likewise.
11417 (*anddi3_insn): Likewise.
11418 (*anddi_zesidi_di): Likewise.
11419 (*anddi_sesdi_di): Likewise.
11420 (*ne_zeroextracts): Likewise.
11421 (*ne_zeroextracts): Likewise.
11422 (*ite_ne_zeroextr): Likewise.
11423 (*ite_ne_zeroextr): Likewise.
11424 (*anddi_notdi_di): Likewise.
11425 (*anddi_notzesidi): Likewise.
11426 (*anddi_notsesidi): Likewise.
11427 (andsi_notsi_si): Likewise.
11428 (thumb1_bicsi3): Likewise.
11429 (*iordi3_insn): Likewise.
11430 (*iordi_zesidi_di): Likewise.
11431 (*iordi_sesidi_di): Likewise.
11432 (*thumb1_iorsi3_insn): Likewise.
11433 (*xordi3_insn): Likewise.
11434 (*xordi_zesidi_di): Likewise.
11435 (*xordi_sesidi_di): Likewise.
11436 (*arm_xorsi3): Likewise.
11437 (*andsi_iorsi3_no): Likewise.
11438 (*smax_0): Likewise.
11439 (*smax_m1): Likewise.
11440 (*arm_smax_insn): Likewise.
11441 (*smin_0): Likewise.
11442 (*arm_smin_insn): Likewise.
11443 (*arm_umaxsi3): Likewise.
11444 (*arm_uminsi3): Likewise.
11445 (*minmax_arithsi): Likewise.
11446 (*minmax_arithsi_): Likewise.
11447 (*satsi_<SAT:code>): Likewise.
11448 (arm_ashldi3_1bit): Likewise.
11449 (arm_ashrdi3_1bit): Likewise.
11450 (arm_lshrdi3_1bit): Likewise.
11451 (*arm_negdi2): Likewise.
11452 (*thumb1_negdi2): Likewise.
11453 (*arm_negsi2): Likewise.
11454 (*thumb1_negsi2): Likewise.
11455 (*negdi_extendsid): Likewise.
11456 (*negdi_zero_extend): Likewise.
11457 (*arm_abssi2): Likewise.
11458 (*thumb1_abssi2): Likewise.
11459 (*arm_neg_abssi2): Likewise.
11460 (*thumb1_neg_abss): Likewise.
11461 (one_cmpldi2): Likewise.
11462 (extend<mode>di2): Likewise.
11463 (*compareqi_eq0): Likewise.
11464 (*arm_extendhisi2addsi): Likewise.
11465 (*arm_movdi): Likewise.
11466 (*thumb1_movdi_insn): Likewise.
11467 (*arm_movt): Likewise.
11468 (*thumb1_movsi_insn): Likewise.
11469 (pic_add_dot_plus_four): Likewise.
11470 (pic_add_dot_plus_eight): Likewise.
11471 (tls_load_dot_plus_eight): Likewise.
11472 (*thumb1_movhi_insn): Likewise.
11473 (*thumb1_movsf_insn): Likewise.
11474 (*movdf_soft_insn): Likewise.
11475 (*thumb_movdf_insn): Likewise.
11476 (cbranchsi4_insn): Likewise.
11477 (cbranchsi4_scratch): Likewise.
11478 (*negated_cbranchsi4): Likewise.
11479 (*tbit_cbranch): Likewise.
11480 (*tlobits_cbranch): Likewise.
11481 (*tstsi3_cbranch): Likewise.
11482 (*cbranchne_decr1): Likewise.
11483 (*addsi3_cbranch): Likewise.
11484 (*addsi3_cbranch_scratch): Likewise.
11485 (*arm_cmpdi_insn): Likewise.
11486 (*arm_cmpdi_unsig): Likewise.
11487 (*arm_cmpdi_zero): Likewise.
11488 (*thumb_cmpdi_zero): Likewise.
11489 (*deleted_compare): Likewise.
11490 (*mov_scc): Likewise.
11491 (*mov_negscc): Likewise.
11492 (*mov_notscc): Likewise.
11493 (*cstoresi_eq0_thumb1_insn): Likewise.
11494 (cstoresi_nltu_thumb1): Likewise.
11495 (cstoresi_ltu_thu): Likewise.
11496 (thumb1_addsi3_addgeu): Likewise.
11497 (*arm_jump): Likewise.
11498 (*thumb_jump): Likewise.
11499 (*check_arch2): Likewise.
11500 (arm_casesi_internal): Likewise.
11501 (thumb1_casesi_dispatch): Likewise.
11502 (*arm_indirect_jump): Likewise.
11503 (*thumb1_indirect_jump): Likewise.
11505 (*and_scc): Likewise.
11506 (*ior_scc): Likewise.
11507 (*compare_scc): Likewise.
11508 (*cond_move): Likewise.
11509 (*cond_arith): Likewise.
11510 (*cond_sub): Likewise.
11511 (*cmp_ite0): Likewise.
11512 (*cmp_ite1): Likewise.
11513 (*cmp_and): Likewise.
11514 (*cmp_ior): Likewise.
11515 (*ior_scc_scc): Likewise.
11516 (*ior_scc_scc_cmp): Likewise.
11517 (*and_scc_scc): Likewise.
11518 (*and_scc_scc_cmp): Likewise.
11519 (*and_scc_scc_nod): Likewise.
11520 (*negscc): Likewise.
11521 (movcond_addsi): Likewise.
11522 (movcond): Likewise.
11523 (*ifcompare_plus_move): Likewise.
11524 (*if_plus_move): Likewise.
11525 (*ifcompare_move_plus): Likewise.
11526 (*if_move_plus): Likewise.
11527 (*ifcompare_arith_arith): Likewise.
11528 (*if_arith_arith): Likewise.
11529 (*ifcompare_arith_move): Likewise.
11530 (*if_arith_move): Likewise.
11531 (*ifcompare_move_arith): Likewise.
11532 (*if_move_arith): Likewise.
11533 (*ifcompare_move_not): Likewise.
11534 (*if_move_not): Likewise.
11535 (*ifcompare_not_move): Likewise.
11536 (*if_not_move): Likewise.
11537 (*ifcompare_shift_move): Likewise.
11538 (*if_shift_move): Likewise.
11539 (*ifcompare_move_shift): Likewise.
11540 (*if_move_shift): Likewise.
11541 (*ifcompare_shift_shift): Likewise.
11542 (*ifcompare_not_arith): Likewise.
11543 (*ifcompare_arith_not): Likewise.
11544 (*if_arith_not): Likewise.
11545 (*ifcompare_neg_move): Likewise.
11546 (*if_neg_move): Likewise.
11547 (*ifcompare_move_neg): Likewise.
11548 (*if_move_neg): Likewise.
11549 (prologue_thumb1_interwork): Likewise.
11550 (*cond_move_not): Likewise.
11551 (*sign_extract_onebit): Likewise.
11552 (*not_signextract_onebit): Likewise.
11553 (stack_tie): Likewise.
11554 (align_4): Likewise.
11555 (align_8): Likewise.
11556 (consttable_end): Likewise.
11557 (consttable_1): Likewise.
11558 (consttable_2): Likewise.
11559 (consttable_4): Likewise.
11560 (consttable_8): Likewise.
11561 (consttable_16): Likewise.
11562 (*thumb1_tablejump): Likewise.
11563 (prefetch): Likewise.
11564 (force_register_use): Likewise.
11565 (thumb_eh_return): Likewise.
11566 (load_tp_hard): Likewise.
11567 (load_tp_soft): Likewise.
11568 (tlscall): Likewise.
11569 (*arm_movtas_ze): Likewise.
11570 (*arm_rev): Likewise.
11571 (*arm_revsh): Likewise.
11572 (*arm_rev16): Likewise.
11573 * config/arm/thumb2.md
11574 (*thumb2_smaxsi3): Likewise.
11575 (*thumb2_sminsi3): Likewise.
11576 (*thumb32_umaxsi3): Likewise.
11577 (*thumb2_uminsi3): Likewise.
11578 (*thumb2_negdi2): Likewise.
11579 (*thumb2_abssi2): Likewise.
11580 (*thumb2_neg_abss): Likewise.
11581 (*thumb2_movsi_insn): Likewise.
11582 (tls_load_dot_plus_four): Likewise.
11583 (*thumb2_movhi_insn): Likewise.
11584 (*thumb2_mov_scc): Likewise.
11585 (*thumb2_mov_negs): Likewise.
11586 (*thumb2_mov_negs): Likewise.
11587 (*thumb2_mov_nots): Likewise.
11588 (*thumb2_mov_nots): Likewise.
11589 (*thumb2_movsicc_): Likewise.
11590 (*thumb2_movsfcc_soft_insn): Likewise.
11591 (*thumb2_indirect_jump): Likewise.
11592 (*thumb2_and_scc): Likewise.
11593 (*thumb2_ior_scc): Likewise.
11594 (*thumb2_ior_scc_strict_it): Likewise.
11595 (*thumb2_cond_move): Likewise.
11596 (*thumb2_cond_arith): Likewise.
11597 (*thumb2_cond_ari): Likewise.
11598 (*thumb2_cond_sub): Likewise.
11599 (*thumb2_negscc): Likewise.
11600 (*thumb2_movcond): Likewise.
11601 (thumb2_casesi_internal): Likewise.
11602 (thumb2_casesi_internal_pic): Likewise.
11603 (*thumb2_alusi3_short): Likewise.
11604 (*thumb2_mov<mode>_shortim): Likewise.
11605 (*thumb2_addsi_short): Likewise.
11606 (*thumb2_subsi_short): Likewise.
11607 (thumb2_addsi3_compare0): Likewise.
11608 (*thumb2_cbz): Likewise.
11609 (*thumb2_cbnz): Likewise.
11610 (*thumb2_one_cmplsi2_short): Likewise.
11611 (*thumb2_negsi2_short): Likewise.
11612 (*orsi_notsi_si): Likewise.
11613 * config/arm/arm1020e.md: Update with new attributes.
11614 * config/arm/arm1026ejs.md: Update with new attributes.
11615 * config/arm/arm1136jfs.md: Update with new attributes.
11616 * config/arm/arm926ejs.md: Update with new attributes.
11617 * config/arm/cortex-a15.md: Update with new attributes.
11618 * config/arm/cortex-a5.md: Update with new attributes.
11619 * config/arm/cortex-a53.md: Update with new attributes.
11620 * config/arm/cortex-a7.md: Update with new attributes.
11621 * config/arm/cortex-a8.md: Update with new attributes.
11622 * config/arm/cortex-a9.md: Update with new attributes.
11623 * config/arm/cortex-m4.md: Update with new attributes.
11624 * config/arm/cortex-r4.md: Update with new attributes.
11625 * config/arm/fa526.md: Update with new attributes.
11626 * config/arm/fa606te.md: Update with new attributes.
11627 * config/arm/fa626te.md: Update with new attributes.
11628 * config/arm/fa726te.md: Update with new attributes.
11630 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
11632 * config/aarch64/aarch64-simd.md
11633 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
11634 <vwx> iterator to ensure correct register choice.
11635 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
11636 (aarch64_sqdmull_n<mode>): Likewise.
11637 (aarch64_sqdmull2_n<mode>_internal): Likewise.
11638 * config/aarch64/arm_neon.h
11639 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
11640 (vml<as><q>_n_<su>16): Likewise.
11641 (vml<as>l_high_lane<q>_<su>16): Likewise.
11642 (vml<as>l_high_n_<su>16): Likewise.
11643 (vml<as>l_lane<q>_<su>16): Likewise.
11644 (vml<as>l_n_<su>16): Likewise.
11645 (vmul<q>_lane<q>_<su>16): Likewise.
11646 (vmul<q>_n_<su>16): Likewise.
11647 (vmull_lane<q>_<su>16): Likewise.
11648 (vmull_n_<su>16): Likewise.
11649 (vmull_high_lane<q>_<su>16): Likewise.
11650 (vmull_high_n_<su>16): Likewise.
11651 (vqrdmulh<q>_n_s16): Likewise.
11653 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
11655 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
11656 have the correct lane parameter.
11658 2013-09-06 Richard Biener <rguenther@suse.de>
11660 * cfganal.c (control_dependences::~control_dependences):
11661 Properly free all of the vector.
11663 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
11666 * config/i386/i386.c (ix86_conditional_register_usage):
11667 Proper initialize extended SSE registers.
11669 2013-09-06 Jan Hubicka <jh@suse.cz>
11671 PR tree-optimization/58311
11672 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
11674 2013-09-06 Jan Hubicka <jh@suse.cz>
11676 * Makefile.in (tree-sra.o): Update dependencies.
11677 * tree-sra.c: Include ipa-utils.h
11678 (scan_function): Use recursive_call_p.
11679 (has_caller_p): New function.
11680 (cgraph_for_node_and_aliases): Count also callers of aliases.
11682 2013-09-06 Jan Hubicka <jh@suse.cz>
11684 PR middle-end/58094
11685 * cgraph.h (symtab_semantically_equivalent_p): Declare.
11686 * tree-tailcall.c: Include ipa-utils.h.
11687 (find_tail_calls): Use it.
11688 * ipa-pure-const.c (check_call): Likewise.
11689 * ipa-utils.c (recursive_call_p): New function.
11690 * ipa-utils.h (recursive_call_p): Dclare.
11691 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
11692 (symtab_semantically_equivalent_p): New function.
11693 * Makefile.in (tree-tailcall.o): Update dependencies.
11695 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
11697 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
11698 non-inlinable part.
11700 2013-09-06 Richard Biener <rguenther@suse.de>
11702 * lto-streamer.h (lto_global_var_decls): Remove.
11703 * Makefile.in (OBJS): Remove lto-symtab.o.
11704 (lto-symtab.o): Remove.
11705 (GTFILES): Remove lto-symtab.c
11706 * lto-symtab.c: Move to lto/
11708 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11710 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
11711 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
11712 (UNSPEC_FPINT_RINT): New constant definitions.
11713 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
11714 definition with 2 attributes.
11715 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
11716 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
11719 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11721 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
11722 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
11723 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
11724 attribute to "z196_alone".
11725 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
11728 2013-09-06 Richard Biener <rguenther@suse.de>
11730 * basic-block.h (class control_dependences): New.
11731 * tree-ssa-dce.c (control_dependence_map): Remove.
11733 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
11734 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
11735 find_pdom, find_control_dependence, find_all_control_dependences):
11737 (mark_control_dependent_edges_necessary,
11738 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
11739 tree_dce_done, perform_tree_ssa_dce): Adjust.
11740 * cfganal.c (set_control_dependence_map_bit,
11741 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
11742 find_all_control_dependences): Move from tree-ssa-dce.c and
11743 implement as methods of control_dependences class.
11744 (control_dependences::control_dependences): New.
11745 (control_dependences::~control_dependences): Likewise.
11746 (control_dependences::get_edges_dependent_on): Likewise.
11747 (control_dependences::get_edge): Likewise.
11749 2013-09-04 Jan Hubicka <jh@suse.cz>
11751 * tree.c (types_same_for_odr): Drop overactive check.
11752 * ipa-devirt.c (hash_type_name): Likewise.
11754 2013-09-04 Jan Hubicka <jh@suse.cz>
11756 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
11757 (analyze_functions): ... here.
11759 2013-09-04 Jan Hubicka <jh@suse.cz>
11761 PR middle-end/58201
11762 * cgraphunit.c (analyze_functions): Clear AUX fields
11763 after processing; initialize assembler name has.
11765 2013-09-05 Jeff Law <law@redhat.com>
11767 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
11768 from thread_around_empty_block. Record threading path into PATH.
11769 Recurse if threading through the initial block is successful.
11770 (thread_across_edge): Corresponding changes to slightly simplify.
11772 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
11774 * config/aarch64/aarch64.md
11775 (type): Remove frecpe, frecps, frecpx.
11776 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
11777 fix to be a TARGET_SIMD instruction.
11778 (aarch64_frecps): Remove.
11779 * config/aarch64/aarch64-simd.md
11780 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
11781 (aarch64_frecps<mode>): Handle all float/vector of float modes.
11783 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
11784 Sofiane Naci <sofiane.naci@arm.com>
11786 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
11787 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
11788 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
11789 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
11790 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
11791 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
11792 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
11793 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
11794 into "clz, "rbit". Rename "shift" to "shift_imm".
11795 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
11796 changes. Update for attribute changes all occurrences of arlo_* and
11798 * config/arm/arm-fixed.md: Update for attribute changes
11799 all occurrences of arlo_* types.
11800 * config/arm/thumb2.md: Update for attribute changes all occurrences
11802 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
11803 (cortexa7_older_only): Likewise.
11804 (cortexa7_younger): Likewise.
11805 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
11806 (1020alu_shift_op): Likewise.
11807 (1020alu_shift_reg_op): Likewise.
11808 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
11809 (alu_shift_op): Likewise.
11810 (alu_shift_reg_op): Likewise.
11811 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
11812 (11_alu_shift_op): Likewise.
11813 (11_alu_shift_reg_op): Likewise.
11814 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
11815 (9_alu_shift_reg_op): Likewise.
11816 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
11818 (cortex_a15_alu_shift): Likewise.
11819 (cortex_a15_alu_shift_reg): Likewise.
11820 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
11822 (cortex_a5_alu_shift): Likewise.
11823 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
11825 (cortex_a53_alu_shift): Likewise.
11826 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
11828 (cortex_a7_alu_reg): Likewise.
11829 (cortex_a7_alu_shift): Likewise.
11830 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
11832 (cortex_a8_alu_shift): Likewise.
11833 (cortex_a8_alu_shift_reg): Likewise.
11834 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
11835 (cortex_a9_dp_shift): Likewise.
11836 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
11838 * config/arm/cortex-r4.md
11839 (cortex_r4_alu): Update for attribute changes.
11840 (cortex_r4_mov): Likewise.
11841 (cortex_r4_alu_shift_reg): Likewise.
11842 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
11843 (526_alu_shift_op): Likewise.
11844 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
11845 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
11846 (626te_alu_shift_op): Likewise.
11847 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
11848 (726te_alu_shift_op): Likewise.
11849 (726te_alu_shift_reg_op): Likewise.
11850 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
11851 (mp626_alu_shift_op): Likewise.
11852 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
11853 (pj4_alu_conds): Likewise.
11854 (pj4_shift): Likewise.
11855 (pj4_shift_conds): Likewise.
11856 (pj4_alu_shift): Likewise.
11857 (pj4_alu_shift_conds): Likewise.
11858 * config/aarch64/aarch64.md: Update for attribute change
11859 all occurrences of arlo_* and shift* types.
11861 2013-09-05 Mike Stump <mikestump@comcast.net>
11863 * tree.h: Move documentation for tree_function_decl to tree-core.h
11864 with the declaration.
11866 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
11869 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
11870 looking for widest mode.
11872 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
11874 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
11876 2013-09-05 Richard Biener <rguenther@suse.de>
11878 PR tree-optimization/58137
11879 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
11880 Do not create vectors of pointers.
11881 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
11882 types for the components of the vector initializer.
11883 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
11884 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
11886 2013-09-05 Martin Jambor <mjambor@suse.cz>
11888 * ipa-prop.c (remove_described_reference): Accept missing references,
11889 return false if that hppens, otherwise return true.
11890 (cgraph_node_for_jfunc): New function.
11891 (try_decrement_rdesc_refcount): Likewise.
11892 (try_make_edge_direct_simple_call): Use them.
11893 (ipa_edge_removal_hook): Remove references from rdescs.
11894 (ipa_edge_duplication_hook): Clone rdescs and their references
11895 when the new edge has the same caller as the old one.
11896 * cgraph.c (cgraph_resolve_speculation): Remove speculative
11897 reference before removing any edges.
11899 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
11901 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
11903 * thumb2.md (thumb2_storewb_parisi): New pattern.
11905 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
11907 * config/aarch64/aarch64-option-extensions.def: Add
11908 AARCH64_OPT_EXTENSION of 'crc'.
11909 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
11910 (AARCH64_ISA_CRC): Ditto.
11911 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
11912 description of the CRC extension.
11914 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
11916 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
11917 * config/rs6000/linux.h: Ditto.
11918 * alpha/linux.h: Ditto.
11919 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
11920 no_c99_libc_has_function.
11921 * config/c6x/uclinux-elf.h: Ditto.
11922 * config/lm32/uclinux-elf.h: Ditto.
11923 * config/m68k/uclinux.h: Ditto.
11924 * config/moxie/uclinux.h: Ditto.
11925 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
11926 (crisv32-*-linux*, cris-*-linux*): Ditto.
11927 * config/bfin/bfin.c: Include "tm_p.h".
11929 2013-09-05 Richard Biener <rguenther@suse.de>
11931 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
11932 check for a definition without a basic-block.
11934 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
11935 Sofiane Naci <sofiane.naci@arm.com>
11937 * config/aarch64/aarch64.md
11938 (*movti_aarch64): Rename r_2_f and f_2_r.
11939 (*movsf_aarch64): Likewise.
11940 (*movdf_aarch64): Likewise.
11941 (*movtf_aarch64): Likewise.
11942 (aarch64_movdi_<mode>low): Likewise.
11943 (aarch64_movdi_<mode>high): Likewise.
11944 (aarch64_mov<mode>high_di): Likewise.
11945 (aarch64_mov<mode>low_di): Likewise.
11946 (aarch64_movtilow_tilow): Likewise.
11947 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
11948 values to config/arm/types.md
11949 (attribute "conds"): Update for attribute change.
11950 (anddi3_insn): Likewise.
11951 (iordi3_insn): Likewise.
11952 (xordi3_insn): Likewise.
11953 (one_cmpldi2): Likewise.
11954 * config/arm/types.md (type): Add Neon types.
11955 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
11956 use "type" attribute.
11957 (movmisalign<mode>_neon_store): Likewise.
11958 (movmisalign<mode>_neon_load): Likewise.
11959 (vec_set<mode>_internal): Likewise.
11960 (vec_setv2di_internal): Likewise.
11961 (vec_extract<mode>): Likewise.
11962 (vec_extractv2di): Likewise.
11963 (add<mode>3_neon): Likewise.
11964 (adddi3_neon): Likewise.
11965 (sub<mode>3_neon): Likewise.
11966 (subdi3_neon): Likewise.
11967 (mul<mode>3_neon): Likewise.
11968 (mul<mode>3add<mode>_neon): Likewise.
11969 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
11970 (fma<VCVTF:mode>4)): Likewise.
11971 (fma<VCVTF:mode>4_intrinsic): Likewise.
11972 (fmsub<VCVTF:mode>4)): Likewise.
11973 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
11974 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
11975 (ior<mode>3): Likewise.
11976 (and<mode>3): Likewise.
11977 (anddi3_neon): Likewise.
11978 (orn<mode>3_neon): Likewise.
11979 (orndi3_neon): Likewise.
11980 (bic<mode>3_neon): Likewise.
11981 (bicdi3_neon): Likewise.
11982 (xor<mode>3): Likewise.
11983 (one_cmpl<mode>2): Likewise.
11984 (abs<mode>2): Likewise.
11985 (neg<mode>2): Likewise.
11986 (umin<mode>3_neon): Likewise.
11987 (umax<mode>3_neon): Likewise.
11988 (smin<mode>3_neon): Likewise.
11989 (smax<mode>3_neon): Likewise.
11990 (vashl<mode>3): Likewise.
11991 (vashr<mode>3_imm): Likewise.
11992 (vlshr<mode>3_imm): Likewise.
11993 (ashl<mode>3_signed): Likewise.
11994 (ashl<mode>3_unsigned): Likewise.
11995 (neon_load_count): Likewise.
11996 (ashldi3_neon_noclobber): Likewise.
11997 (signed_shift_di3_neon): Likewise.
11998 (unsigned_shift_di3_neon): Likewise.
11999 (ashrdi3_neon_imm_noclobber): Likewise.
12000 (lshrdi3_neon_imm_noclobber): Likewise.
12001 (widen_ssum<mode>3): Likewise.
12002 (widen_usum<mode>3): Likewise.
12003 (quad_halves_<code>v4si): Likewise.
12004 (quad_halves_<code>v4sf): Likewise.
12005 (quad_halves_<code>v8hi): Likewise.
12006 (quad_halves_<code>v16qi): Likewise.
12007 (reduc_splus_v2di): Likewise.
12008 (neon_vpadd_internal<mode>): Likewise.
12009 (neon_vpsmin<mode>): Likewise.
12010 (neon_vpsmax<mode>): Likewise.
12011 (neon_vpumin<mode>): Likewise.
12012 (neon_vpumax<mode>): Likewise.
12013 (ss_add<mode>_neon): Likewise.
12014 (us_add<mode>_neon): Likewise.
12015 (ss_sub<mode>_neon): Likewise.
12016 (us_sub<mode>_neon): Likewise.
12017 (neon_vadd<mode>_unspec): Likewise.
12018 (neon_vaddl<mode>): Likewise.
12019 (neon_vaddw<mode>): Likewise.
12020 (neon_vhadd<mode>): Likewise.
12021 (neon_vqadd<mode>): Likewise.
12022 (neon_vaddhn<mode>): Likewise.
12023 (neon_vmul<mode>): Likewise.
12024 (neon_vmla<mode>): Likewise.
12025 (neon_vmlal<mode>): Likewise.
12026 (neon_vmls<mode>): Likewise.
12027 (neon_vmlsl<mode>): Likewise.
12028 (neon_vqdmulh<mode>): Likewise.
12029 (neon_vqdmlal<mode>): Likewise.
12030 (neon_vqdmlsl<mode>): Likewise.
12031 (neon_vmull<mode>): Likewise.
12032 (neon_vqdmull<mode>): Likewise.
12033 (neon_vsub<mode>_unspec): Likewise.
12034 (neon_vsubl<mode>): Likewise.
12035 (neon_vsubw<mode>): Likewise.
12036 (neon_vqsub<mode>): Likewise.
12037 (neon_vhsub<mode>): Likewise.
12038 (neon_vsubhn<mode>): Likewise.
12039 (neon_vceq<mode>): Likewise.
12040 (neon_vcge<mode>): Likewise.
12041 (neon_vcgeu<mode>): Likewise.
12042 (neon_vcgt<mode>): Likewise.
12043 (neon_vcgtu<mode>): Likewise.
12044 (neon_vcle<mode>): Likewise.
12045 (neon_vclt<mode>): Likewise.
12046 (neon_vcage<mode>): Likewise.
12047 (neon_vcagt<mode>): Likewise.
12048 (neon_vtst<mode>): Likewise.
12049 (neon_vabd<mode>): Likewise.
12050 (neon_vabdl<mode>): Likewise.
12051 (neon_vaba<mode>): Likewise.
12052 (neon_vabal<mode>): Likewise.
12053 (neon_vmax<mode>): Likewise.
12054 (neon_vmin<mode>): Likewise.
12055 (neon_vpaddl<mode>): Likewise.
12056 (neon_vpadal<mode>): Likewise.
12057 (neon_vpmax<mode>): Likewise.
12058 (neon_vpmin<mode>): Likewise.
12059 (neon_vrecps<mode>): Likewise.
12060 (neon_vrsqrts<mode>): Likewise.
12061 (neon_vqabs<mode>): Likewise.
12062 (neon_vqneg<mode>): Likewise.
12063 (neon_vcls<mode>): Likewise.
12064 (clz<mode>2): Likewise.
12065 (popcount<mode>2): Likewise.
12066 (neon_vrecpe): Likewise.
12067 (neon_vrsqrte): Likewise.
12068 (neon_vget_lane<mode>_sext_internal): Likewise.
12069 (neon_vget_lane<mode>_zext_internal): Likewise.
12070 (neon_vdup_n<mode>): Likewise.
12071 (neon_vdup_nv2di): Likewise.
12072 (neon_vdpu_lane<mode>_internal): Likewise.
12073 (neon_vswp<mode>): Likewise.
12074 (float<mode><V_cvtto>2): Likewise.
12075 (floatuns<mode><V_cvtto>2): Likewise.
12076 (fix_trunc<mode><V_cvtto>)2): Likewise
12077 (fixuns_trunc<mode><V_cvtto)2): Likewise.
12078 (neon_vcvt<mode>): Likewise.
12079 (neon_vcvtv4sfv4hf): Likewise.
12080 (neon_vcvtv4hfv4sf): Likewise.
12081 (neon_vcvt_n<mode>): Likewise.
12082 (neon_vmovn<mode>): Likewise.
12083 (neon_vqmovn<mode>): Likewise.
12084 (neon_vqmovun<mode>): Likewise.
12085 (neon_vmovl<mode>): Likewise.
12086 (neon_vmul_lane<mode>): Likewise.
12087 (neon_vmull_lane<mode>): Likewise.
12088 (neon_vqdmull_lane<mode>): Likewise.
12089 (neon_vqdmulh_lane<mode>): Likewise.
12090 (neon_vmla_lane<mode>): Likewise.
12091 (neon_vmlal_lane<mode>): Likewise.
12092 (neon_vqdmlal_lane<mode>): Likewise.
12093 (neon_vmls_lane<mode>): Likewise.
12094 (neon_vmlsl_lane<mode>): Likewise.
12095 (neon_vqdmlsl_lane<mode>): Likewise.
12096 (neon_vext<mode>): Likewise.
12097 (neon_vrev64<mode>): Likewise.
12098 (neon_vrev32<mode>): Likewise.
12099 (neon_vrev16<mode>): Likewise.
12100 (neon_vbsl<mode>_internal): Likewise.
12101 (neon_vshl<mode>): Likewise.
12102 (neon_vqshl<mode>): Likewise.
12103 (neon_vshr_n<mode>): Likewise.
12104 (neon_vshrn_n<mode>): Likewise.
12105 (neon_vqshrn_n<mode>): Likewise.
12106 (neon_vqshrun_n<mode>): Likewise.
12107 (neon_vshl_n<mode>): Likewise.
12108 (neon_vqshl_n<mode>): Likewise.
12109 (neon_vqshlu_n<mode>): Likewise.
12110 (neon_vshll_n<mode>): Likewise.
12111 (neon_vsra_n<mode>): Likewise.
12112 (neon_vsri_n<mode>): Likewise.
12113 (neon_vsli_n<mode>): Likewise.
12114 (neon_vtbl1v8qi): Likewise.
12115 (neon_vtbl2v8qi): Likewise.
12116 (neon_vtbl3v8qi): Likewise.
12117 (neon_vtbl4v8qi): Likewise.
12118 (neon_vtbx1v8qi): Likewise.
12119 (neon_vtbx2v8qi): Likewise.
12120 (neon_vtbx3v8qi): Likewise.
12121 (neon_vtbx4v8qi): Likewise.
12122 (neon_vtrn<mode>_internal): Likewise.
12123 (neon_vzip<mode>_internal): Likewise.
12124 (neon_vuzp<mode>_internal): Likewise.
12125 (neon_vld1<mode>): Likewise.
12126 (neon_vld1_lane<mode>): Likewise.
12127 (neon_vld1_dup<mode>): Likewise.
12128 (neon_vld1_dupv2di): Likewise.
12129 (neon_vst1<mode>): Likewise.
12130 (neon_vst1_lane<mode>): Likewise.
12131 (neon_vld2<mode>): Likewise.
12132 (neon_vld2_lane<mode>): Likewise.
12133 (neon_vld2_dup<mode>): Likewise.
12134 (neon_vst2<mode>): Likewise.
12135 (neon_vst2_lane<mode>): Likewise.
12136 (neon_vld3<mode>): Likewise.
12137 (neon_vld3qa<mode>): Likewise.
12138 (neon_vld3qb<mode>): Likewise.
12139 (neon_vld3_lane<mode>): Likewise.
12140 (neon_vld3_dup<mode>): Likewise.
12141 (neon_vst3<mode>): Likewise.
12142 (neon_vst3qa<mode>): Likewise.
12143 (neon_vst3qb<mode>): Likewise.
12144 (neon_vst3_lane<mode>): Likewise.
12145 (neon_vld4<mode>): Likewise.
12146 (neon_vld4qa<mode>): Likewise.
12147 (neon_vld4qb<mode>): Likewise.
12148 (neon_vld4_lane<mode>): Likewise.
12149 (neon_vld4_dup<mode>): Likewise.
12150 (neon_vst4<mode>): Likewise.
12151 (neon_vst4qa<mode>): Likewise.
12152 (neon_vst4qb<mode>): Likewise.
12153 (neon_vst4_lane<mode>): Likewise.
12154 (neon_vec_unpack<US>_lo_<mode>): Likewise.
12155 (neon_vec_unpack<US>_hi_<mode>): Likewise.
12156 (neon_vec_<US>mult_lo_<mode>): Likewise.
12157 (neon_vec_<US>mult_hi_<mode>): Likewise.
12158 (neon_vec_<US>shiftl_<mode>): Likewise.
12159 (neon_unpack<US>_<mode>): Likewise.
12160 (neon_vec_<US>mult_<mode>): Likewise.
12161 (vec_pack_trunc_<mode>): Likewise.
12162 (neon_vec_pack_trunk_<mode>): Likewise.
12163 (neon_vabd<mode>_2): Likewise.
12164 (neon_vabd<mode>_3): Likewise.
12165 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
12166 (thumb2_movsi_vfp): Likewise.
12167 (movdi_vfp): Likewise.
12168 (movdi_vfp_cortexa8): Likewise.
12169 (movhf_vfp_neon): Likewise.
12170 (movhf_vfp): Likewiwse.
12171 (movsf_vfp): Likewiwse.
12172 (thumb2_movsf_vfp): Likewiwse.
12173 (movdf_vfp): Likewise.
12174 (thumb2_movdf_vfp): Likewise.
12175 (movsfcc_vfp): Likewise.
12176 (thumb2_movsfcc_vfp): Likewise.
12177 (movdfcc_vfp): Likewise.
12178 (thumb2_movdfcc_vfp): Likewise.
12179 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
12180 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
12181 (v10_v2c): Likewise.
12182 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
12184 (cortex_a15_neon_int_2): Likewise.
12185 (cortex_a15_neon_int_3): Likewise.
12186 (cortex_a15_neon_int_4): Likewise.
12187 (cortex_a15_neon_int_5): Likewise.
12188 (cortex_a15_neon_vqneg_vqabs): Likewise.
12189 (cortex_a15_neon_vmov): Likewise.
12190 (cortex_a15_neon_vaba): Likewise.
12191 (cortex_a15_neon_vaba_qqq): Likewise.
12192 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
12193 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
12194 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
12195 scalar_64_32_long_scalar): Likewise.
12196 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
12197 (cortex_a15_neon_mla_qqq_8_16): Likewise.
12198 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
12199 lotype_qdd_64_32_long): Likewise.
12200 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
12201 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
12202 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
12203 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
12204 (cortex_a15_neon_shift_1): Likewise.
12205 (cortex_a15_neon_shift_2): Likewise.
12206 (cortex_a15_neon_shift_3): Likewise.
12207 (cortex_a15_neon_vshl_ddd): Likewise.
12208 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
12209 (cortex_a15_neon_vsra_vrsra): Likewise.
12210 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
12211 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
12212 (cortex_a15_neon_fp_vmul_ddd): Likewise.
12213 (cortex_a15_neon_fp_vmul_qqd): Likewise.
12214 (cortex_a15_neon_fp_vmla_ddd): Likewise.
12215 (cortex_a15_neon_fp_vmla_qqq): Likewise.
12216 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
12217 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
12218 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
12219 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
12220 (cortex_a15_neon_bp_simple): Likewise.
12221 (cortex_a15_neon_bp_2cycle): Likewise.
12222 (cortex_a15_neon_bp_3cycle): Likewise.
12223 (cortex_a15_neon_vld1_1_2_regs): Likewise.
12224 (cortex_a15_neon_vld1_3_4_regs): Likewise.
12225 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
12226 (cortex_a15_neon_vld2_4_regs): Likewise.
12227 (cortex_a15_neon_vld3_vld4): Likewise.
12228 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
12229 (cortex_a15_neon_vst1_3_4_regs): Likewise.
12230 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
12231 (cortex_a15_neon_vst3_vst4): Likewise.
12232 (cortex_a15_neon_vld1_vld2_lane): Likewise.
12233 (cortex_a15_neon_vld3_vld4_lane" 10
12234 (cortex_a15_neon_vst1_vst2_lane): Likewise.
12235 (cortex_a15_neon_vst3_vst4_lane): Likewise.
12236 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
12237 (cortex_a15_neon_ldm_2): Likewise.0
12238 (cortex_a15_neon_stm_2): Likewise.
12239 (cortex_a15_neon_mcr): Likewise.
12240 (cortex_a15_neon_mcr_2_mcrr): Likewise.
12241 (cortex_a15_neon_mrc): Likewise.
12242 (cortex_a15_neon_mrrc): Likewise.
12243 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
12245 (cortex_a15_alu_shift): Likewise.
12246 (cortex_a15_alu_shift_reg): Likewise.
12247 (cortex_a15_mult32): Likewise.
12248 (cortex_a15_mult64): Likewise.
12249 (cortex_a15_block): Likewise.
12250 (cortex_a15_branch): Likewise.
12251 (cortex_a15_load1): Likewise.
12252 (cortex_a15_load3): Likewise.
12253 (cortex_a15_store1): Likewise.
12254 (cortex_a15_store3): Likewise.
12255 (cortex_a15_call): Likewise.
12256 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
12257 (cortex_a5_f2r): Likewise.
12258 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
12260 (cortex_a53_f2r): Likewise.
12261 * config/arm/cortex-a7.md
12262 (cortex_a7_branch): Update for attribute change.
12263 (cortex_a7_call): Likewise.
12264 (cortex_a7_alu_imm): Likewise.
12265 (cortex_a7_alu_reg): Likewise.
12266 (cortex_a7_alu_shift): Likewise.
12267 (cortex_a7_mul): Likewise.
12268 (cortex_a7_load1): Likewise.
12269 (cortex_a7_store1): Likewise.
12270 (cortex_a7_load2): Likewise.
12271 (cortex_a7_store2): Likewise.
12272 (cortex_a7_load3): Likewise.
12273 (cortex_a7_store3): Likewise.
12274 (cortex_a7_load4): Likewise.
12275 (cortex_a7_store4): Likewise.
12276 (cortex_a7_fpalu): Likewise.
12277 (cortex_a7_fconst): Likewise.
12278 (cortex_a7_fpmuls): Likewise.
12279 (cortex_a7_neon_mul): Likewise.
12280 (cortex_a7_fpmacs): Likewise.
12281 (cortex_a7_neon_mla: Likewise.
12282 (cortex_a7_fpmuld: Likewise.
12283 (cortex_a7_fpmacd: Likewise.
12284 (cortex_a7_fpfmad: Likewise.
12285 (cortex_a7_fdivs: Likewise.
12286 (cortex_a7_fdivd: Likewise.
12287 (cortex_a7_r2f: Likewise.
12288 (cortex_a7_f2r: Likewise.
12289 (cortex_a7_f_flags: Likewise.
12290 (cortex_a7_f_loads: Likewise.
12291 (cortex_a7_f_loadd: Likewise.
12292 (cortex_a7_f_stores: Likewise.
12293 (cortex_a7_f_stored: Likewise.
12294 (cortex_a7_neon): Likewise.
12295 * config/arm/cortex-a8-neon.md
12296 (cortex_a8_neon_mrc): Update for attribute change.
12297 (cortex_a8_neon_mrrc): Likewise.
12298 (cortex_a8_neon_int_1): Likewise.
12299 (cortex_a8_neon_int_2): Likewise.
12300 (cortex_a8_neon_int_3): Likewise.
12301 (cortex_a8_neon_int_4): Likewise.
12302 (cortex_a8_neon_int_5): Likewise.
12303 (cortex_a8_neon_vqneg_vqabs): Likewise.
12304 (cortex_a8_neon_vmov): Likewise.
12305 (cortex_a8_neon_vaba): Likewise.
12306 (cortex_a8_neon_vaba_qqq): Likewise.
12307 (cortex_a8_neon_vsma): Likewise.
12308 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
12309 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
12310 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
12311 long_scalar): Likewise.
12312 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
12313 (cortex_a8_neon_mla_qqq_8_16): Likewise.
12314 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
12315 long_scalar_qdd_64_32_long): Likewise.
12316 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
12317 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
12318 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
12319 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
12320 (cortex_a8_neon_shift_1): Likewise.
12321 (cortex_a8_neon_shift_2): Likewise.
12322 (cortex_a8_neon_shift_3): Likewise.
12323 (cortex_a8_neon_vshl_ddd): Likewise.
12324 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
12325 (cortex_a8_neon_vsra_vrsra): Likewise.
12326 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
12327 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
12328 (cortex_a8_neon_fp_vsum): Likewise.
12329 (cortex_a8_neon_fp_vmul_ddd): Likewise.
12330 (cortex_a8_neon_fp_vmul_qqd): Likewise.
12331 (cortex_a8_neon_fp_vmla_ddd): Likewise.
12332 (cortex_a8_neon_fp_vmla_qqq): Likewise.
12333 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
12334 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
12335 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
12336 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
12337 (cortex_a8_neon_bp_simple): Likewise.
12338 (cortex_a8_neon_bp_2cycle): Likewise.
12339 (cortex_a8_neon_bp_3cycle): Likewise.
12340 (cortex_a8_neon_ldr): Likewise.
12341 (cortex_a8_neon_str): Likewise.
12342 (cortex_a8_neon_vld1_1_2_regs): Likewise.
12343 (cortex_a8_neon_vld1_3_4_regs): Likewise.
12344 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
12345 (cortex_a8_neon_vld2_4_regs): Likewise.
12346 (cortex_a8_neon_vld3_vld4): Likewise.
12347 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
12348 (cortex_a8_neon_vst1_3_4_regs): Likewise.
12349 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
12350 (cortex_a8_neon_vst3_vst4): Likewise.
12351 (cortex_a8_neon_vld1_vld2_lane): Likewise.
12352 (cortex_a8_neon_vld3_vld4_lane): Likewise.
12353 (cortex_a8_neon_vst1_vst2_lane): Likewise.
12354 (cortex_a8_neon_vst3_vst4_lane): Likewise.
12355 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
12356 (cortex_a8_neon_mcr): Likewise.
12357 (cortex_a8_neon_mcr_2_mcrr): Likewise.
12358 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
12359 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
12361 (ca9_neon_mrrc): Likewise.
12362 (cortex_a9_neon_int_1): Likewise.
12363 (cortex_a9_neon_int_2): Likewise.
12364 (cortex_a9_neon_int_3): Likewise.
12365 (cortex_a9_neon_int_4): Likewise.
12366 (cortex_a9_neon_int_5): Likewise.
12367 (cortex_a9_neon_vqneg_vqabs): Likewise.
12368 (cortex_a9_neon_vmov): Likewise.
12369 (cortex_a9_neon_vaba): Likewise.
12370 (cortex_a9_neon_vaba_qqq): Likewise.
12371 (cortex_a9_neon_vsma): Likewise.
12372 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
12373 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
12374 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
12375 long_scalar): Likewise.
12376 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
12377 (cortex_a9_neon_mla_qqq_8_16): Likewise.
12378 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
12379 long_scalar_qdd_64_32_long): Likewise.
12380 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
12381 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
12382 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
12383 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
12384 (cortex_a9_neon_shift_1): Likewise.
12385 (cortex_a9_neon_shift_2): Likewise.
12386 (cortex_a9_neon_shift_3): Likewise.
12387 (cortex_a9_neon_vshl_ddd): Likewise.
12388 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
12389 (cortex_a9_neon_vsra_vrsra): Likewise.
12390 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
12391 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
12392 (cortex_a9_neon_fp_vsum): Likewise.
12393 (cortex_a9_neon_fp_vmul_ddd): Likewise.
12394 (cortex_a9_neon_fp_vmul_qqd): Likewise.
12395 (cortex_a9_neon_fp_vmla_ddd): Likewise.
12396 (cortex_a9_neon_fp_vmla_qqq): Likewise.
12397 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
12398 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
12399 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
12400 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
12401 (cortex_a9_neon_bp_simple): Likewise.
12402 (cortex_a9_neon_bp_2cycle): Likewise.
12403 (cortex_a9_neon_bp_3cycle): Likewise.
12404 (cortex_a9_neon_ldr): Likewise.
12405 (cortex_a9_neon_str): Likewise.
12406 (cortex_a9_neon_vld1_1_2_regs): Likewise.
12407 (cortex_a9_neon_vld1_3_4_regs): Likewise.
12408 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
12409 (cortex_a9_neon_vld2_4_regs): Likewise.
12410 (cortex_a9_neon_vld3_vld4): Likewise.
12411 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
12412 (cortex_a9_neon_vst1_3_4_regs): Likewise.
12413 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
12414 (cortex_a9_neon_vst3_vst4): Likewise.
12415 (cortex_a9_neon_vld1_vld2_lane): Likewise.
12416 (cortex_a9_neon_vld3_vld4_lane): Likewise.
12417 (cortex_a9_neon_vst1_vst2_lane): Likewise.
12418 (cortex_a9_neon_vst3_vst4_lane): Likewise.
12419 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
12420 (cortex_a9_neon_mcr): Likewise.
12421 (cortex_a9_neon_mcr_2_mcrr): Likewise.
12422 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
12423 (cortex_a9_fps): Likewise.
12424 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
12426 (cortex_m4_fmuls): Likewise.
12427 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
12429 (cortex_r4_mrc): Likewise.
12430 * config/arm/iterators.md: Update comment referring to neon_type.
12431 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
12432 (iwmmxt_movsi_insn): Likewise.
12433 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
12435 (pj4_core_to_vfp): Likewise.
12436 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
12438 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
12439 (vfp_fstore): Likewise.
12440 * doc/md.texi: Change references to neon_type to refer to type.
12442 2013-09-04 Dodji Seketeli <dodji@redhat.com>
12444 * tree.h (DECL_BUILT_IN): Fix typo in comment.
12446 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
12448 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
12449 lglobl if not weak.
12451 2013-09-04 Easwaran Raman <eraman@google.com>
12453 PR middle-end/57370
12454 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
12455 (build_and_add_sum): Use it.
12456 (appears_later_in_bb): Simplify code.
12458 2013-09-04 Teresa Johnson <tejohnson@google.com>
12460 * dumpfile.c (dump_finish): Don't close stderr/stdout.
12462 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
12464 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
12466 2013-09-04 Jan Hubicka <jh@suse.cz>
12468 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
12469 * ipa-devirt.c: Include diganostic.h
12470 (odr_type_d): Add types and types_set.
12471 (hash_type_name): Work for types with vtables during LTO.
12472 (odr_hasher::remove): Fix comment; destroy types_set.
12473 (add_type_duplicate): New function,
12474 (get_odr_type): Use it.
12475 (dump_type_inheritance_graph): Dump type duplicates.
12476 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
12478 * tree.c (types_same_for_odr): Give exact answers on types with
12481 2013-09-04 Dodji Seketeli <dodji@redhat.com>
12483 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
12484 explaining their differences.
12486 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
12488 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
12490 2013-09-03 Jeff Law <law@redhat.com>
12492 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
12493 when not threading through a joiner block. Pass joiner/no joiner
12494 state to register_jump_thread.
12495 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
12496 state from argument rather than implying on path length.
12497 Dump the entire jump thread path into debugging dump.
12498 * tree-flow.h (register_jump_thread): Update prototype.
12500 2013-08-29 Xinliang David Li <davidxl@google.com>
12502 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
12503 Remove a trivial gcc_assert.
12505 2013-08-29 Xinliang David Li <davidxl@google.com>
12507 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
12508 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
12509 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
12510 Delay base decl alignment adjustment.
12511 * tree-vectorizer.c (vect_destroy_datarefs): New function.
12512 * tree-vectorizer.h: New data structure.
12513 (set_dr_misalignment): New function.
12514 (dr_misalignment): Ditto.
12515 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
12516 (vectorizable_load): Ditto.
12517 (ensure_base_align): New function.
12518 (vectorize_loops): Add dbg_cnt support.
12519 (execute_vect_slp): Ditto.
12520 * dbgcnt.def: New debug counter.
12521 * Makefile: New dependency.
12523 2013-09-03 Meador Inge <meadori@codesourcery.com>
12527 2013-08-30 Meador Inge <meadori@codesourcery.com>
12529 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
12531 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
12533 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
12534 function descriptor.
12536 2013-09-03 Richard Biener <rguenther@suse.de>
12538 * tree-affine.c (add_elt_to_tree): Fix association issue,
12539 avoid useless converts and make sure to always return a
12540 properly typed result.
12542 2013-09-03 Richard Biener <rguenther@suse.de>
12544 PR middle-end/57656
12545 * fold-const.c (negate_expr_p): Fix division case.
12546 (negate_expr): Likewise.
12548 2013-09-03 Richard Biener <rguenther@suse.de>
12551 * tree-streamer-out.c: Include tm.h.
12552 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
12554 2013-09-03 Jan Hubicka <jh@suse.cz>
12556 * tree-profile.c (tree_profiling): Cleanup CFG when done.
12558 2013-09-03 Alan Modra <amodra@gmail.com>
12560 * config.gcc (powerpc*-*-linux*): Add support for little-endian
12561 multilibs to big-endian target and vice versa.
12562 * config/rs6000/t-linux64: Use := assignment on all vars.
12563 (MULTILIB_EXTRA_OPTS): Remove fPIC.
12564 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
12565 * config/rs6000/t-linux64le: New file.
12566 * config/rs6000/t-linux64bele: New file.
12567 * config/rs6000/t-linux64lebe: New file.
12569 2013-09-02 Jan Hubicka <jh@suse.cz>
12571 * ipa-inline-transform.c (inline_transform): Do not
12572 optimize_inline_calls when not optimizing.
12574 2013-09-02 Jan Hubicka <jh@suse.cz>
12576 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
12577 duplicated nodes for assembler names.
12578 * symtab.c (symtab_unregister_node): Do not attempt to unlink
12579 hard registers from assembler name hash.
12581 2013-09-02 Jan Hubicka <jh@suse.cz>
12583 * ipa-split.c (execute_split_functions): Split externally visible
12584 functions called once.
12586 2013-09-02 Martin Jambor <mjambor@suse.cz>
12589 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
12590 linked list. When finding the correct duplicate, also consider also
12591 the caller in additon to its inlined_to node.
12593 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
12595 * config/aarch64/aarch64-simd-builtins.def
12596 (dup_lane_scalar): Remove.
12597 * config/aarch64/aarch64-simd.md
12598 (aarch64_simd_dup): Add 'w->w' alternative.
12599 (aarch64_dup_lane<mode>): Allow for VALL.
12600 (aarch64_dup_lane_scalar<mode>): Remove.
12601 (aarch64_dup_lane_<vswap_width_name><mode>): New.
12602 (aarch64_get_lane_signed<mode>): Add w->w altenative.
12603 (aarch64_get_lane_unsigned<mode>): Likewise.
12604 (aarch64_get_lane<mode>): Likewise.
12605 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
12606 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
12607 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
12608 (VCON): Change container of V2SF.
12609 (vswap_width_name): Likewise.
12610 * config/aarch64/arm_neon.h
12611 (__aarch64_vdup_lane_any): New.
12612 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
12613 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
12614 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
12616 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
12618 PR middle-end/56382
12619 * expr.c (emit_move_complex): Do not move complex FP values as parts if
12620 the source or the destination is a single hard register.
12622 2013-09-02 Richard Biener <rguenther@suse.de>
12624 PR middle-end/57511
12625 * tree-scalar-evolution.c (instantiate_scev_name): Allow
12628 2013-09-02 Richard Biener <rguenther@suse.de>
12630 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
12631 arithmetic to sizetype.
12633 2013-09-02 Bin Cheng <bin.cheng@arm.com>
12635 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
12636 Find auto-increment use both before and after candidate.
12638 2013-09-02 Marek Polacek <polacek@redhat.com>
12640 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
12642 2013-09-01 Jan Hubicka <jh@suse.cz>
12644 * Makefile.in: Add ipa-profile.o
12645 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
12646 * cgraph.c (struct cgraph_propagate_frequency_data,
12647 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
12648 ipa-profile.c; replace cgraph_ by ipa_ prefix.
12649 * cgraph.h (cgraph_propagate_frequency): Remove.
12650 * ipa-inline-analysis.c: Include ipa-utils.h;
12651 drop duplicated cfgloop.h.
12652 (inline_update_callee_summaries): Update.
12653 * ipa-profile.c: New file.
12654 * ipa-utils.h (ipa_propagate_frequency): Declare.
12655 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
12656 data-streamer.h, value-prof.h.
12657 (symtab_remove_unreachable_nodes): Update profile.
12658 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
12659 account_time_size, cmp_counts, dump_histogram,
12660 ipa_profile_generate_summary, ipa_profile_write_summary,
12661 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
12662 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
12663 Move to ipa-profile.c.
12665 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
12667 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
12669 2013-09-01 Jan Hubicka <jh@suse.cz>
12671 * common.opt (fdevirtualize-speculatively): New function.
12672 * invoke.texi (fdevirtualize-speculatively): Document.
12673 * ipa-devirt.c: Include ipa-inline.h
12674 (likely_target_p): New function.
12675 (ipa_devirt): New function.
12676 (gate_ipa_devirt): New function.
12677 (pass_data_ipa_devirt): New static var.
12678 (pass_ipa_devirt): Likewise.
12679 (make_pass_ipa_devirt): New function.
12680 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
12681 (common_handle_option): Disable devirtualization when
12682 value range profiling is available.
12683 * passes.def (pass_ipa_devirt): Add.
12684 * timever.def (TV_IPA_DEVIRT): New timevar.
12685 * tree-pass.h (make_pass_ipa_devirt):
12687 2013-09-01 Iain Sandoe <iain@codesourcery.com>
12689 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
12690 include sanitize(undefined).
12692 2013-08-31 Diego Novillo <dnovillo@google.com>
12694 * Makefile.in (TREE_CORE_H): Define.
12696 (GTFILES): Add tree-core.h.
12697 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
12699 * tree-core.h: New file.
12700 Move all data structures, enum, typedefs, global
12701 declarations and constants from ...
12702 * tree.h: ... here.
12704 2013-08-31 Jan Hubicka <jh@suse.cz>
12706 * bulitins.c (expand_builtin): Do not early exit for gcov
12707 instrumented functions.
12709 2013-08-31 Marek Polacek <polacek@redhat.com>
12711 * ubsan.c: Include tm_p.h.
12713 2013-08-31 Jan Hubicka <jh@suse.cz>
12715 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
12718 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
12719 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
12720 * tree-cfg.c (verify_gimple_label): ... here.
12721 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
12722 (ipa_merge_profiles): New function.
12723 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
12724 (lto_input_function_body): Likewise.
12725 * ipa-utils.h (ipa_merge_profiles): Declare.
12726 * lto-streamer.h (lto_input_function_body): Update prototype.
12727 (emit_label_in_global_context_p): Remove.
12728 * lto-symtab.c: Include ipa-utils.h
12729 (lto_cgraph_replace_node): Use ipa_merge_profiles.
12731 2013-08-31 Jan Hubicka <jh@suse.cz>
12733 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
12735 2013-08-31 Jan Hubicka <jh@suse.cz>
12737 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
12739 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
12741 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
12742 "cmp" RTX before signed_comparison_operator check to account
12743 for "code" changes.
12745 2013-08-30 Jan Hubicka <jh@suse.cz>
12747 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
12748 (detect_type_change_1): Rename to ...
12749 (detect_type_change): ... this one; early return on non-polymorphic
12751 (detect_type_change_ssa): Add comp_type parameter; update
12752 use of detect_type_change.
12753 (compute_complex_assign_jump_func): Add param_type parameter;
12754 update use of detect_type_change_ssa.
12755 (compute_complex_ancestor_jump_func): Likewise.
12756 (ipa_get_callee_param_type): New function.
12757 (ipa_compute_jump_functions_for_edge): Compute parameter type;
12758 update calls to the jump function computation functions.
12760 2013-08-30 Teresa Johnson <tejohnson@google.com>
12761 Steven Bosscher <steven@gcc.gnu.org>
12763 * cfgrtl.c (fixup_new_cold_bb): New routine.
12764 (commit_edge_insertions): Invoke fixup_partitions.
12765 (find_partition_fixes): New routine.
12766 (fixup_partitions): Ditto.
12767 (verify_hot_cold_block_grouping): Update comments.
12768 (rtl_verify_edges): Invoke find_partition_fixes.
12769 (rtl_verify_bb_pointers): Update comments.
12770 (rtl_verify_bb_layout): Ditto.
12771 * basic-block.h (probably_never_executed_edge_p): Declare.
12772 (fixup_partitions): Ditto.
12773 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
12774 * bb-reorder.c (sanitize_hot_paths): New function.
12775 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
12776 sanitize_hot_paths.
12777 * predict.c (probably_never_executed_edge_p): New routine.
12778 * cfg.c (check_bb_profile): Add partition insanity warnings.
12780 2013-08-30 Meador Inge <meadori@codesourcery.com>
12782 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
12784 2013-08-30 Marek Polacek <polacek@redhat.com>
12786 * Makefile.in (ubsan.o): Add.
12787 (c-family/c-ubsan.o): Add.
12788 (builtins.o): Add ubsan.h dependency.
12789 * ubsan.h: New file.
12790 * ubsan.c: New file.
12791 * common.opt: Add -fsanitize=undefined option.
12792 (flag_sanitize): Add variable.
12793 (fsanitize=): Add option. Add Driver.
12794 (fsanitize=thread): Remove option.
12795 (fsanitize=address): Likewise.
12796 (static-libubsan): New option.
12797 * doc/invoke.texi: Document the new flag and -static-libubsan.
12798 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
12799 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
12800 * builtin-attrs.def (ATTR_COLD): Define.
12801 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
12802 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
12803 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
12804 * flag-types.h (sanitize_code): New enum.
12805 * opts.c (common_handle_option): Parse command line arguments
12806 of -fsanitize=. Add -fsanitize=unreachable option.
12807 * varasm.c (get_variable_section): Adjust.
12808 (assemble_noswitch_variable): Likewise.
12809 (assemble_variable): Likewise.
12810 (output_constant_def_contents): Likewise.
12811 (categorize_decl_for_section): Likewise.
12812 (place_block_symbol): Likewise.
12813 (output_object_block): Likewise.
12814 * builtins.def: Likewise.
12815 * toplev.c (compile_file): Likewise.
12816 (process_options): Likewise.
12817 * cppbuiltin.c: Likewise.
12818 * tsan.c (tsan_pass): Likewise.
12819 (tsan_gate): Likewise.
12820 (tsan_gate_O0): Likewise.
12821 * cfgexpand.c (partition_stack_vars): Likewise.
12822 (expand_stack_vars): Likewise.
12823 (defer_stack_allocation): Likewise.
12824 (expand_used_vars): Likewise.
12825 * cfgcleanup.c (old_insns_match_p): Likewise.
12826 * asan.c (asan_finish_file): Likewise.
12827 (asan_instrument): Likewise.
12828 (gate_asan): Likewise.
12829 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
12830 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
12831 (asan_global_struct): Use pointer_sized_int_node instead
12832 calling build_nonstandard_integer_type.
12833 (initialize_sanitizer_builtins): Likewise.
12834 (asan_finish_file): Likewise.
12835 * gcc.c: Document %{%:function(args):X}.
12836 (static_spec_functions): Add sanitize.
12837 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
12838 store funcval != NULL there.
12839 (do_spec_1): Adjust handle_spec_function caller.
12840 (handle_braces): Allow %:function(args) as condition.
12841 (sanitize_spec_function): New function.
12842 (ADD_STATIC_LIBUBSAN_LIBS): Define.
12843 (LIBUBSAN_SPEC): Likewise.
12844 (LIBUBSAN_EARLY_SPEC): Likewise.
12845 (SANITIZER_SPEC): Handle libubsan.
12846 (SANITIZER_EARLY_SPEC): Likewise.
12847 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
12848 instead of fsanitize=address.
12849 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
12850 instead of fsanitize=address*.
12851 * builtins.c: Include ubsan.h.
12852 (fold_builtin_0): Instrument __builtin_unreachable.
12853 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
12854 instead of flag_asan.
12855 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
12856 (pointer_sized_int_node): Define.
12857 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
12859 2013-08-30 Mike Stump <mikestump@comcast.net>
12861 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
12864 2013-08-29 Jan Hubicka <jh@suse.cz>
12866 * cgraph.c (cgraph_function_body_availability): Handle weakref
12868 * passes.def: Remove pass_fixup_cfg.
12869 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
12870 track when we need to remove functions.
12871 (gate_ipa_inline): Execute inlining always; add comment why.
12872 (pass_data_ipa_inline): Remove TODO_remove_functions.
12873 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
12874 do not produce summaries.
12875 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
12876 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
12877 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
12878 constant pool and vtable.
12880 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
12882 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
12883 New arm_neon.h's internal macros to specify 64-bit constants.
12884 Avoid using stdint.h's macros.
12886 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
12888 * recog.c (verify_changes): Verify that changes[i].old is non-zero
12889 before applying REG_P.
12891 2013-08-30 Jakub Jelinek <jakub@redhat.com>
12893 PR tree-optimization/58277
12894 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
12895 after seeing too many stmts with vdef in between dombb and current
12896 bb, invalidate everything.
12898 2013-08-30 Richard Biener <rguenther@suse.de>
12900 * fold-const.c (fold_single_bit_test): Fix overflow test.
12902 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
12904 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
12905 and which can live in a register, always retrieve the value on entry.
12906 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
12907 passed by invisible reference specially.
12908 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
12909 (vt_add_function_parameter): Correctly deal with a parameter passed by
12910 invisible reference.
12912 2013-08-30 Jan Hubicka <jh@suse.cz>
12914 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
12916 2013-08-30 Richard Biener <rguenther@suse.de>
12918 PR tree-optimization/58228
12919 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
12920 allow invariant loads in nested loop vectorization.
12922 2013-08-30 Richard Biener <rguenther@suse.de>
12924 PR tree-optimization/58223
12925 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
12926 (has_anti_or_output_dependence): ... this and adjust to also
12927 look for output dependences.
12928 (mark_nodes_having_upstream_mem_writes): Adjust.
12929 (rdg_flag_uses): Likewise.
12931 2013-08-30 Richard Biener <rguenther@suse.de>
12933 PR tree-optimization/58010
12934 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
12935 assert that we have a loop-closed PHI.
12937 2013-08-29 Jan Hubicka <jh@suse.cz>
12939 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
12940 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
12941 * lto-section-in.c (lto_free_function_in_decl_state): New function.
12942 (lto_free_function_in_decl_state_for_node): New function.
12944 2013-08-29 Xinliang David Li <davidxl@google.com>
12946 * loop-unroll.c (report_unroll_peel): Minor message change.
12947 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
12948 Emit alignment peeling message with default -fopt-info.
12949 (vect_loop_versioning): Emit loop version info message.
12950 * tree-vectorizer.c (vectorize_loops): Minor message change.
12951 (execute_vect_slp): Ditto.
12953 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
12955 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
12956 of the clone from the DECL_NAME of the original function.
12958 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
12960 * passes.c (register_pass): Add overload.
12961 * tree-pass.h (register_pass): Forward declare it. Add comment.
12963 2013-08-29 Jan Hubicka <jh@suse.cz>
12965 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
12966 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
12967 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
12968 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
12969 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
12970 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
12971 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
12972 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
12974 2013-08-29 Teresa Johnson <tejohnson@google.com>
12976 * dumpfile.c (dump_loc): Output column number.
12977 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
12978 * doc/invoke.texi: Document optall -fopt-info flag.
12979 * profile.c (read_profile_edge_counts): Use new dump framework.
12980 (compute_branch_probabilities): Ditto.
12981 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
12982 when pass not in any opt group.
12983 * pass_manager.h (pass_manager::get_pass_profile): New method.
12984 * value-prof.c (check_counter): Use new dump framework.
12985 (check_ic_target): Ditto.
12986 * coverage.c (get_coverage_counts): Ditto.
12987 (coverage_init): Setup new dump framework.
12989 2013-08-29 Richard Biener <rguenther@suse.de>
12991 PR tree-optimization/58246
12992 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
12993 handle the dominance check inside a basic-block.
12995 2013-08-29 Richard Biener <rguenther@suse.de>
12997 PR middle-end/57287
12998 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
12999 of default defs that appear in abnormal PHI nodes.
13001 2013-08-29 Richard Biener <rguenther@suse.de>
13003 PR tree-optimization/57685
13004 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
13005 single-use operands to avoid exponential complexity.
13007 2013-08-28 Dehao Chen <dehao@google.com>
13009 * ipa-inline.c (edge_badness): Fix integer underflow.
13011 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
13013 * gtm-builtins.def (_ITM_free): Declare leaf.
13015 2013-08-28 Jakub Jelinek <jakub@redhat.com>
13018 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
13019 (*tls_local_dynamic_base_64_largepic): Likewise.
13020 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
13021 Remove predicate from call operand.
13022 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
13023 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
13025 2013-08-28 Jeff Law <law@redhat.com>
13027 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
13028 checks for the number of predecessors and successors allowed.
13029 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
13030 which require copying a joiner block if there is a request which
13031 is a subpath that requires no joiner block copying.
13033 2013-08-28 Jan Hubicka <jh@suse.cz>
13035 * lto-streamer-out.c (DFS_write_tree_body): Drop
13036 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
13037 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
13038 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
13039 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
13040 Do not read DECL_ERROR_ISSUED.
13041 (unpack_ts_decl_with_vis_value_fields): Do not read
13043 (lto_input_ts_binfo_tree_pointers): Do not read
13044 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
13045 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
13046 write DECL_ERROR_ISSUED..
13047 (pack_ts_decl_with_vis_value_fields): Do not write
13049 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
13050 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
13051 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
13052 * tree.h (tree_decl_common): Update comment.
13053 (DECL_ERROR_ISSUED): Remove.
13055 2013-08-28 Jakub Jelinek <jakub@redhat.com>
13057 PR middle-end/58257
13058 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
13060 2013-08-28 Jan Hubicka <jh@suse.cz>
13062 * builtins.def (free): Declare leaf.
13064 2013-08-27 David Malcolm <dmalcolm@redhat.com>
13066 * gdbhooks.py: New.
13067 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
13068 * configure: Regenerate.
13070 2013-08-27 Martin Jambor <mjambor@suse.cz>
13072 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
13073 (ipa_ancestor_jf_data): Likewise.
13074 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
13075 (ipa_get_jf_pass_through_type_preserved): New function.
13076 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
13077 (ipa_get_jf_ancestor_type_preserved): New function.
13078 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
13079 (ipa_get_jf_ancestor_result): Likewise.
13080 (propagate_vals_accross_pass_through): Use
13081 ipa_get_jf_pass_through_result to do all the value mappings.
13082 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
13083 type_preserved flag.
13084 (ipa_set_jf_cst_copy): New function.
13085 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
13086 (ipa_set_jf_arith_pass_through): Likewise.
13087 (ipa_set_ancestor_jf): Likewise.
13088 (compute_complex_assign_jump_func): Set type_preserved instead of
13090 (ipa_compute_jump_functions_for_edge): Likewise.
13091 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
13092 (update_jump_functions_after_inlining): Update type_preserved.
13093 Explicitely create jump functions when combining one with pass_through.
13094 (ipa_write_jump_function): Stream the type_preserved flags.
13095 (ipa_read_jump_function): Likewise.
13097 2013-08-27 Jakub Jelinek <jakub@redhat.com>
13098 Aldy Hernandez <aldyh@redhat.com>
13100 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
13101 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
13102 * function.h (struct function): Add has_force_vect_loops and
13104 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
13105 * gimple.c (gimple_build_omp_critical): Add KIND argument and
13107 * gimple.def: Update CLAUSES comments.
13108 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
13109 (gimple_build_omp_for): Add argument to prototype.
13110 (gimple_omp_for_kind): New.
13111 (gimple_omp_for_set_kind): New.
13112 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
13113 GOVD_DATA_SHARE_CLASS.
13114 (enum omp_region_type): Add ORT_SIMD.
13115 (gimple_add_tmp_var): Handle ORT_SIMD.
13116 (gimplify_var_or_parm_decl): Same.
13117 (is_gimple_stmt): Same.
13118 (omp_firstprivatize_variable): Same.
13119 (omp_add_variable): Only use splay_tree_insert if lookup failed.
13120 (omp_notice_variable): Handle ORT_SIMD.
13121 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
13122 (omp_check_private): Handle ORT_SIMD.
13123 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
13124 OMP_CLAUSE_SAFELEN.
13125 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
13126 Handle OMP_CLAUSE_LASTPRIVATE.
13127 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
13128 OMP_CLAUSE_SAFELEN.
13129 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
13130 (gimplify_expr): Handle OMP_SIMD.
13131 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
13132 (expand_GOMP_SIMD_VF): New.
13133 (expand_GOMP_SIMD_LAST_LANE): New.
13134 * internal-fn.def (GOMP_SIMD_LANE): New.
13135 (GOMP_SIMD_VF): New.
13136 (GOMP_SIMD_LAST_LANE): New.
13137 * omp-low.c: Include target.h.
13138 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
13139 OMP_CLAUSE_SAFELEN.
13140 (check_omp_nesting_restrictions): Same.
13142 (lower_rec_simd_input_clauses): New.
13143 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
13145 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
13146 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
13147 (expand_omp_build_assign): New.
13148 (expand_omp_for_init_counts): New.
13149 (expand_omp_for_init_vars): New.
13150 (extract_omp_for_update_vars): New.
13151 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
13152 and rewrite accordingly.
13153 (expand_omp_simd): New.
13154 (expand_omp_for): Use expand_omp_simd.
13155 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
13156 (lower_omp_for): Do not lower the body.
13157 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
13158 in their own loops.
13159 * tree-flow.h (find_omp_clause): Remove prototype.
13160 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
13161 forcing vectorization of the loop, or if flag_tree_vectorize.
13162 (gate_tree_if_conversion): Similarly.
13163 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
13164 gimple_build_omp_for.
13165 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
13166 * tree-parloops (create_parallel_loop): Pass kind argument to
13167 gimple_build_omp_for.
13168 * tree-pretty-print.c (dump_omp_clause): Add cases for
13169 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
13170 OMP_CLAUSE__SIMDUID_.
13171 (dump_generic_node): Handle OMP_SIMD.
13172 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
13173 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
13174 unroll OMP_SIMD loops here.
13175 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
13176 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
13178 (vect_analyze_data_refs): Handle simd loops.
13179 * tree-vect-loop.c (vectorizable_live_operation): Handle
13181 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
13182 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
13183 (vectorizable_load): Same.
13184 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
13185 (struct simduid_to_vf): New.
13186 (simduid_to_vf::hash): New.
13187 (simduid_to-vf::equal): New.
13188 (struct simd_array_to_simduid): New.
13189 (simd_array_to_simduid::hash): New.
13190 (simd_array_to_simduid::equal): New.
13191 (adjust_simduid_builtins): New.
13192 (struct note_simd_array_uses_struct): New.
13193 (note_simd_array_uses_cb): New.
13194 (note_simd_array_uses): New.
13195 (vectorize_loops): Handle simd hints and adjust simd builtins
13197 * tree-vectorizer.h (struct _stmt_vec_info): Add
13198 simd_lane_access_p field.
13199 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
13200 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
13201 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
13202 (omp_clause_code_name): Same.
13203 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
13204 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
13205 * tree.def (OMP_SIMD): New entry.
13206 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
13207 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
13208 (OMP_CLAUSE_DECL): Adjust range for new clauses.
13209 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
13210 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
13211 (OMP_CLAUSE_LINEAR_STEP): New.
13212 (OMP_CLAUSE_SAFELEN_EXPR): New.
13213 (OMP_CLAUSE__SIMDUID__DECL): New.
13214 (find_omp_clause): New prototype.
13216 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
13218 * config/i386/driver-i386.c (host_detect_local_cpu): Update
13219 Haswell processor detection.
13221 2013-08-27 Christian Widmer <shadow@umbrox.de>
13224 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
13225 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
13226 AVX2 capable processors.
13228 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
13230 * config/aarch64/arm_neon.h: Replace all inline asm implementations
13231 of vget_low_* with implementations in terms of other intrinsics.
13233 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
13235 PR middle-end/57219
13236 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
13237 values to -1, 0 and 1.
13239 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
13241 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
13242 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
13243 (<optab><mode>3_insn): Remove.
13244 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
13245 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
13246 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
13247 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
13248 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
13249 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
13250 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
13251 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
13252 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
13253 (ror<mode>3_insn): Likewise.
13254 * config/aarch64/predicates.md (aarch64_simd_register): New.
13256 2013-08-27 Richard Biener <rguenther@suse.de>
13258 PR tree-optimization/57521
13259 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
13260 one edge is non-critical.
13261 (find_phi_replacement_condition): Make sure to use a non-critical
13262 edge. Cleanup and remove old bug workarounds.
13263 (bb_postdominates_preds): Remove.
13264 (if_convertible_loop_p_1): Do not compute post-dominators.
13265 (combine_blocks): Do not free post-dominators.
13266 (main_tree_if_conversion): Likewise.
13267 (pass_data_if_conversion): Add TODO_verify_ssa.
13269 2013-08-27 DJ Delorie <dj@redhat.com>
13271 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
13273 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
13275 * function.c (assign_parm_find_data_types): Set passed_mode and
13276 nominal_mode to the TYPE_MODE of nominal_type for the built
13277 pointer type in case of the struct-pass-by-reference.
13279 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
13281 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
13282 (UINT16_TYPE): Change default to "unsigned int".
13284 * config/avr/avr.opt (mfract-convert-truncate): New option.
13285 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
13286 is set, round negative fractional integers according to n1169
13287 when converting to integer types.
13289 2013-08-26 Jan Hubicka <jh@suse.cz>
13291 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
13292 methods can not be called indirectly when their address is not taken.
13294 2013-08-26 Jan Hubicka <jh@suse.cz>
13296 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
13299 2013-08-26 Jan Hubicka <jh@suse.cz>
13301 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
13304 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
13306 * reload.c (find_valid_class): Allow classes that do not include
13307 FIRST_PSEUDO_REGISTER - 1.
13309 2013-08-26 Jan Hubicka <jh@suse.cz>
13311 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
13312 fix edge count/frequency when speculation failed; fix type check
13313 for the direct call.
13315 2013-08-26 Jan Hubicka <jh@suse.cz>
13317 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
13319 2013-08-26 Jan Hubicka <jh@suse.cz>
13321 * ipa-inline-transform.c (inline_transform): Be ready for basic block
13322 to be changed by edge redirection.
13324 2013-08-26 Jan Hubicka <jh@suse.cz>
13326 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
13327 formating; add sanity check.
13328 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
13329 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
13330 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
13331 (resolve_noninline_speculation): Update callee keys, too.
13333 2013-08-26 Jan Hubicka <jh@suse.cz>
13335 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
13336 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
13338 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
13340 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
13343 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
13345 * config/i386/i386.c (ix86_debug_options): Remove prototype.
13346 (x86_64_elf_select_section): Ditto.
13347 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
13349 (ix86_pass_by_reference): Ditto.
13350 (output_set_got): Ditto.
13351 (ix86_unary_operator_ok): Ditto.
13352 (ix86_expand_builtin): Ditto.
13354 2013-08-23 Jan Hubicka <jh@suse.cz>
13356 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
13358 2013-08-23 Jan Hubicka <jh@suse.cz>
13360 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
13361 (tree_decl_with_vis): Add FINAL field.
13363 2013-08-23 Jeff Law <law@redhat.com>
13365 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
13366 the debugging dump when the expression is fully redundant.
13368 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
13370 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
13371 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
13372 * pretty-print.c (pp_formatted_text_data): Likewise.
13373 (pp_write_text_to_stream): Likewise.
13374 (pp_write_text_as_dot_label_to_stream): Likewise.
13375 (pp_append_r): Likewise.
13376 (pp_format): Likewise.
13377 (pp_flush): Likewise.
13378 (pp_clear_output_area): Likewise.
13379 (pp_append_text): Likewise.
13380 (pp_formatted_text): Likewise.
13381 (pp_remaining_character_count_for_line): Likewise.
13382 (pp_newline): Likewise.
13383 (pp_character): Likewise.
13384 (output_buffer::~output_buffer): Define.
13385 (pretty_printer::~pretty_printer): Destruct output buffer.
13386 * pretty-print.h (output_buffer::~output_buffer): Declare.
13387 (pretty_printer::~pretty_printer): Declare virtual.
13389 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
13392 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
13393 HOST_WIDE_INT_M1U): New macros.
13394 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
13395 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
13396 unsigned -1 for lshift.
13397 * cse.c (cse_insn): Likewise.
13398 * double-int.c (rshift_double, lshift_double): Likewise.
13399 * builtins.c (fold_builtin_bitop): Likewise.
13400 * combine.c (force_to_mode): Likewise.
13401 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
13402 * simplify-rtx.c (simplify_const_unary_operation,
13403 simplify_const_binary_operation): Likewise.
13404 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
13405 check_va_list_escapes): Likewise.
13406 * rtlanal.c (nonzero_bits1): Likewise.
13407 * expmed.c (expand_smod_pow2): Likewise.
13408 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
13410 2013-08-23 Jan Hubicka <jh@suse.cz>
13412 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
13413 as having address taken.
13415 2013-08-23 Jan Hubicka <jh@suse.cz>
13417 * ipa-utils.h (method_class_type): Declare.
13418 * ipa-devirt.c (method_class_type): Export.
13420 * cgraphunit.c (analyze_functions): Do basic devirtualization;
13421 do not walk base classes of anonymous types.
13423 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
13425 PR rtl-optimization/58220
13426 PR regression/58221
13427 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
13428 handle SEQUENCE insns properly.
13430 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
13432 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
13434 (pp_newline_and_indent): Likewise.
13435 (pp_separate_with): Likewise.
13436 * pretty-print.c (pp_newline_and_flush): Define.
13437 (pp_newline_and_indent): Likewise.
13438 (pp_separate_with): Likewise.
13440 2013-08-23 Jakub Jelinek <jakub@redhat.com>
13443 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
13444 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
13446 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
13448 * config/i386/predicates.md (ext_sse_reg_operand): New.
13449 * config/i386/i386.md (*movti_internal): Use
13450 predicate to determine if EVEX is needed.
13451 (*movsi_internal): Ditto.
13452 (*movdf_internal): Ditto.
13453 (*movsf_internal): Ditto.
13454 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
13456 2013-08-23 Jakub Jelinek <jakub@redhat.com>
13458 PR tree-optimization/58209
13459 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
13460 (find_tail_calls): Give up for pointer result types if m is non-NULL.
13461 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
13462 emit POINTER_PLUS_EXPR.
13463 (create_tailcall_accumulator): For pointer result type accumulate in
13466 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
13468 * configure.ac: Add backslashes missing from the last change.
13469 * configure: Regenerate.
13471 2013-08-22 Jan Hubicka <jh@susue.cz>
13473 * ipa.c (function_and_variable_visibility): First remember function
13474 was global and then make it local.
13476 2013-08-22 Julian Brown <julian@codesourcery.com>
13478 * configure.ac: Add aarch64 to list of arches which use "nop" in
13480 * configure: Regenerate.
13482 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13484 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
13485 gnu_libc_has_function.
13486 * config/s390/tpf.h: Likewise.
13488 2013-08-22 Jan Hubicka <jh@susue.cz>
13490 * timevar.c (validate_phases): Add cast.
13492 2013-08-22 Jan Hubicka <jh@susue.cz>
13494 * timevar.c (validate_phases): Use size_t for memory.
13495 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
13497 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
13499 * pretty-print.h (output_buffer::output_buffer): Declare.
13500 (pretty_printer::pretty_printer): Likewise.
13501 (pp_construct): Remove.
13502 * pretty-print.c (output_buffer::output_buffer): Define.
13503 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
13504 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
13505 (print_gimple_expr): Likewise.
13506 (print_gimple_seq): Likewise.
13507 (gimple_dump_bb): Likewise.
13508 * sched-vis.c (dump_value_slim): Likewise.
13509 (dump_insn_slim): Likewise.
13510 (dump_rtl_slim): Likewise.
13511 (str_pattern_slim): Likewise.
13512 * tree-mudflap.c (mf_varname_tree): Likewise.
13513 * graph.c (print_graph_cfg): Likewise.
13514 (start_graph_dump): Likewise.
13515 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
13517 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
13518 pretty printer initialization.
13519 * coretypes.h (diagnostic_context): Remove superflous type alias
13521 (pretty_printer): Likewise. Declare directly as a class.
13522 (pretty_print_info): Remove declaration as class.
13523 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
13524 and pp_clear_output_area.
13525 (asan_add_global): Likewise.
13527 2013-08-22 Jan Hubicka <jh@suse.cz>
13529 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
13530 * ipa-utils.h (update_type_inheritance_graph): Declare.
13531 (possible_polymorphic_call_target_p): Declare.
13532 (possible_polymorphic_call_target_p): New.
13533 * ipa-devirt.c: Update toplevel comments.
13534 (cached_polymorphic_call_targets): Move up.
13535 (odr_type_d): Move ID down.
13536 (polymorphic_type_binfo_p): Update comment.
13537 (odr_hasher::remove): Likewise;
13538 (get_odr_type): Set anonymous_namespace.
13539 (dump_odr_type): Dump it.
13540 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
13541 (maybe_record_node): Record node in cached_polymorphic_call_targets.
13542 (record_binfo): Add comment.
13543 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
13545 (devirt_node_removal_hook): Do not iCE when cache is freed.
13546 (possible_polymorphic_call_target_p): New predicate.
13547 (update_type_inheritance_graph): New function.
13549 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
13550 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
13551 Sergey Lega <sergey.s.lega@intel.com>
13552 Anna Tikhonova <anna.tikhonova@intel.com>
13553 Ilya Tocar <ilya.tocar@intel.com>
13554 Andrey Turetskiy <andrey.turetskiy@intel.com>
13555 Ilya Verbin <ilya.verbin@intel.com>
13556 Kirill Yukhin <kirill.yukhin@intel.com>
13557 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
13559 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
13560 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
13561 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
13562 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
13563 (OPTION_MASK_ISA_AVX2_UNSET): Update.
13564 (OPTION_MASK_ISA_AVX512F_UNSET): New.
13565 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
13566 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
13567 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
13568 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
13569 OPT_mavx512pf, OPT_mavx512er cases.
13570 * config/i386/constraints.md (v): New constraint.
13571 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
13572 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
13573 (bit_AVX512CD): New.
13574 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
13575 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
13576 * config/i386/i386-c.c (ix86_target_macros_internal):
13577 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
13579 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
13580 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
13581 * config/i386/i386.c (regclass_map, dbx_register_map)
13582 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
13583 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
13584 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
13585 -mavx512pf options.
13586 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
13587 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
13588 -mavx512pf options. Fix formatting.
13589 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
13590 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
13591 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
13592 -mavx512cd, -mavx512pf options.
13593 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
13594 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
13595 (ix86_preferred_output_reload_class): Replace SSE_REGS with
13597 (ix86_hard_regno_mode_ok): Support 512-bit registers.
13598 (ix86_set_reg_reg_cost): Ditto.
13599 (x86_order_regs_for_local_alloc): Ditto.
13600 (MAX_VECT_LEN): Extend to 64-byte.
13601 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
13602 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
13603 (TARGET_AVX512ER, TARGET_AVX512CD): New.
13604 (BIGGEST_ALIGNMENT): Extend to 512-bits.
13605 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
13606 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
13607 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
13608 (SSE_REG_MODE_P): Support new modes.
13609 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
13610 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
13611 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
13612 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
13613 (REG_CLASS_CONTENTS): Add new registers.
13614 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
13615 (EXT_REX_SSE_REGNO_P): New.
13616 (HI_REGISTER_NAMES): Add new registers.
13617 * config/i386/i386.md: Define constants for new registers.
13618 (mode): Add new 512-bit modes.
13619 (prefix): Support evex prefix.
13620 (isa): Support avx512f, noavx512f, fma_avx512f.
13621 (ssemodesuffix): Add new 512-bit modes.
13623 (*movxi_internal_avx512f): Ditto.
13624 (*movdi_internal): Replace constraint "x" with the new constraint "v".
13626 (*movsi_internal): Likewise.
13627 (*movdf_internal): Likewise.
13628 (*movsf_internal): Likewise.
13629 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
13630 (<code><mode>3): Likewise.
13631 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
13633 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
13634 with the new constraint "v".
13635 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
13637 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
13638 with the new constraint "v".
13639 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
13640 (<sse2>_storedqu<avxsizesuffix>): Likewise.
13641 (*<plusminus_insn><mode>3): Likewise.
13642 (<sse>_vm<plusminus_insn><mode>3): Likewise.
13643 (*mul<mode>3): Likewise.
13644 (<sse>_vmmul<mode>3): Likewise.
13645 (<sse>_div<mode>3): Likewise.
13646 (<sse>_vmdiv<mode>3): Likewise.
13647 (<sse>_sqrt<mode>2): Likewise.
13648 (<sse>_vmsqrt<mode>2): Likewise.
13649 (*<code><mode>3_finite): Likewise.
13650 (*<code><mode>3) <smaxmin>: Likewise.
13651 (<sse>_vm<code><mode>3): Likewise.
13652 (*<code><mode>3) <any_logic>: Likewise.
13653 (*fma_fmadd_<mode>): Likewise.
13654 (*fma_fmsub_<mode>): Likewise.
13655 (*fma_fnmadd_<mode>): Likewise.
13656 (*fma_fnmsub_<mode>): Likewise.
13657 (*fma_fmaddsub_<mode>): Likewise.
13658 (*fma_fmsubadd_<mode>): Likewise.
13659 (*fmai_fmadd_<mode>): Likewise.
13660 (*fmai_fmsub_<mode>): Likewise.
13661 (*fmai_fnmadd_<mode>): Likewise.
13662 (*fmai_fnmsub_<mode>): Likewise.
13663 (sse_cvtsi2ss): Likewise.
13664 (sse_cvtsi2ssq): Likewise.
13665 (sse_cvtss2si): Likewise.
13666 (sse_cvtss2si_2): Likewise.
13667 (sse_cvtss2siq): Likewise.
13668 (sse_cvtss2siq_2): Likewise.
13669 (sse_cvttss2si): Likewise.
13670 (sse_cvtss2siq_2): Likewise.
13671 (float<sseintvecmodelower><mode>2): Likewise.
13672 (sse2_cvtsd2si_2): Likewise.
13673 (sse2_cvtsd2siq_2): Likewise.
13674 (*<plusminus_insn><mode>3): Likewise.
13675 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
13676 (*<sse4_1_avx2>_mul<mode>3): Likewise.
13677 (ashr<mode>3): Likewise.
13678 (<shift_insn><mode>3): Likewise.
13679 (avx2_<code><mode>3): Likewise.
13680 (*avx2_<code><mode>3): Likewise.
13681 (*andnot<mode>3): Likewise.
13682 (*<code><mode>3) <any_logic>: Likewise.
13683 (abs<mode>2): Likewise.
13684 (avx2_permvar<mode>): Likewise.
13685 (avx2_perm<mode>_1): Likewise.
13686 (*avx_vpermilp<mode>): Likewise.
13687 (avx_vpermilvar<mode>3): Likewise.
13688 (avx2_ashrv<mode>): Likewise.
13689 (avx2_<shift_insn>v<mode>): Likewise.
13690 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
13692 * doc/rtl.texi: Document XImode.
13694 2013-08-21 Jeff Law <law@redhat.com>
13696 * tree-flow.h (register_jump_thread): Pass vector of edges
13697 instead of each important edge.
13698 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
13699 thread path into a vector and pass that to register_jump_thread.
13700 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
13701 passed in edge vector to the current 3-edge form.
13704 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
13706 * dce.c (fini_dce): Call df_analyze again just in case
13707 delete_unmarked_insns removed anything.
13709 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
13711 * reload.h (struct reg_equivs): Rename to ..
13712 (struct reg_equivs_s): .. this.
13714 2013-08-20 Martin Liska <marxin.liska@gmail.com>
13716 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
13718 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
13720 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
13722 2013-08-21 Jeff Law <law@redhat.com>
13724 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
13725 simplify assignments too. If the RHS collapses to a singleton
13726 range, then return the value for the range.
13728 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
13730 * config/i386/sse.md (V16): Rename to...
13732 (mov<mode>): Update iterator name.
13733 (*mov<mode>_internal): Ditto.
13734 (push<mode>1): Ditto.
13735 (movmisalign<mode>): Ditto.
13737 2013-08-20 Jan Hubicka <jh@suse.cz>
13740 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
13741 entry for direct edges.
13742 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
13744 2013-08-20 David Malcolm <dmalcolm@redhat.com>
13746 Revert my last two changes, r201865 and r201864:
13749 2013-08-20 David Malcolm <dmalcolm@redhat.com>
13751 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
13752 instances can own GC refs.
13754 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
13755 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
13756 (gcc::context::gt_pch_nx): Likewise.
13757 (gcc::context::gt_pch_nx): Likewise.
13758 * ggc.h (gt_ggc_mx <T>): New.
13759 (gt_pch_nx_with_op <T>): New.
13760 (gt_pch_nx <T>): New.
13761 * passes.c (opt_pass::gt_ggc_mx): New.
13762 (opt_pass::gt_pch_nx): New.
13763 (opt_pass::gt_pch_nx_with_op): New.
13764 (pass_manager::gt_ggc_mx): New.
13765 (pass_manager::gt_pch_nx): New.
13766 (pass_manager::gt_pch_nx_with_op): New.
13767 (pass_manager::operator new): Use
13768 ggc_internal_cleared_alloc_stat rather than xcalloc.
13769 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
13770 (pass_manager::gt_ggc_mx): New.
13771 (pass_manager::gt_pch_nx): New.
13772 (pass_manager::gt_pch_nx_with_op): New.
13773 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
13774 (opt_pass::operator new): New.
13775 (opt_pass::gt_ggc_mx): New.
13776 (opt_pass::gt_pch_nx): New.
13777 (opt_pass::gt_pch_nx_with_op): New.
13780 2013-08-20 David Malcolm <dmalcolm@redhat.com>
13782 * Makefile.in (GTFILES): Add context.h.
13783 * context.c (gcc::context::operator new): New.
13784 (gcc::context::gt_ggc_mx): New.
13785 (gcc::context::gt_pch_nx): New.
13786 (gcc::context::gt_pch_nx): New.
13787 * context.h (gcc::context): Add GTY((user)) marking.
13788 (gcc::context::operator new): New.
13789 (gcc::context::gt_ggc_mx): New.
13790 (gcc::context::gt_pch_nx): New.
13791 (gcc::context::gt_pch_nx): New.
13792 (g): Add GTY marking.
13793 (gt_ggc_mx (gcc::context *)): New.
13794 (gt_pch_nx (gcc::context *)): New.
13795 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
13796 void *cookie)): New.
13797 * gengtype.c (open_base_files) <ifiles>: Add context.h.
13799 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
13801 * dce.c (fini_dce): Call df_analyze again just in case
13802 delete_unmarked_insns removed anything.
13804 2013-08-20 Teresa Johnson <tejohnson@google.com>
13806 PR rtl-optimizations/57451
13807 * final.c (reemit_insn_block_notes): Prevent lexical blocks
13808 from crossing split section boundaries.
13810 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
13812 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
13813 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
13814 with MULTLIB_DEFAULTS.
13816 2013-08-20 Nick Clifton <nickc@redhat.com>
13818 * target.def (narrow_volatile_bitfield): Note that the default
13819 value is false, not !TARGET_STRICT_ALIGN.
13820 * doc/tm.texi: Regenerate.
13822 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
13824 Fix LIB_SPEC for systems without libpthread.
13826 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
13827 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
13829 * config/i386/linux-common.h: Likewise.
13830 * config/mips/linux-common.h: Likewise.
13832 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
13834 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
13837 2013-08-20 David Malcolm <dmalcolm@redhat.com>
13839 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
13840 instances can own GC refs.
13842 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
13843 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
13844 (gcc::context::gt_pch_nx): Likewise.
13845 (gcc::context::gt_pch_nx): Likewise.
13846 * ggc.h (gt_ggc_mx <T>): New.
13847 (gt_pch_nx_with_op <T>): New.
13848 (gt_pch_nx <T>): New.
13849 * passes.c (opt_pass::gt_ggc_mx): New.
13850 (opt_pass::gt_pch_nx): New.
13851 (opt_pass::gt_pch_nx_with_op): New.
13852 (pass_manager::gt_ggc_mx): New.
13853 (pass_manager::gt_pch_nx): New.
13854 (pass_manager::gt_pch_nx_with_op): New.
13855 (pass_manager::operator new): Use
13856 ggc_internal_cleared_alloc_stat rather than xcalloc.
13857 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
13858 (pass_manager::gt_ggc_mx): New.
13859 (pass_manager::gt_pch_nx): New.
13860 (pass_manager::gt_pch_nx_with_op): New.
13861 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
13862 (opt_pass::operator new): New.
13863 (opt_pass::gt_ggc_mx): New.
13864 (opt_pass::gt_pch_nx): New.
13865 (opt_pass::gt_pch_nx_with_op): New.
13867 2013-08-20 David Malcolm <dmalcolm@redhat.com>
13869 * Makefile.in (GTFILES): Add context.h.
13870 * context.c (gcc::context::operator new): New.
13871 (gcc::context::gt_ggc_mx): New.
13872 (gcc::context::gt_pch_nx): New.
13873 (gcc::context::gt_pch_nx): New.
13874 * context.h (gcc::context): Add GTY((user)) marking.
13875 (gcc::context::operator new): New.
13876 (gcc::context::gt_ggc_mx): New.
13877 (gcc::context::gt_pch_nx): New.
13878 (gcc::context::gt_pch_nx): New.
13879 (g): Add GTY marking.
13880 (gt_ggc_mx (gcc::context *)): New.
13881 (gt_pch_nx (gcc::context *)): New.
13882 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
13883 void *cookie)): New.
13884 * gengtype.c (open_base_files) <ifiles>: Add context.h.
13886 2013-08-20 Alan Modra <amodra@gmail.com>
13889 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
13890 (rs6000_emit_epilogue): Likewise.
13892 2013-08-19 Dehao Chen <dehao@google.com>
13894 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
13896 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
13897 Jakub Jelinek <jakub@redhat.com>
13899 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
13900 (BUILT_IN_FABSD64): Likewise.
13901 (BUILT_IN_FABSD128): Likewise.
13902 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
13903 (fold_builtin_1): Likewise.
13904 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
13905 destination and source operands.
13906 (*abstd2_fpr): Likewise.
13907 (*nabstd2_fpr): Likewise.
13909 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
13911 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
13914 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
13916 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
13919 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
13921 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
13922 * builtins.c (default_libc_has_function): New.
13923 (gnu_libc_has_function): Ditto.
13924 (no_c99_libc_has_function): Ditto.
13925 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
13926 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
13927 (fold_builtin_sincos): Likewise.
13928 (fold_builtin_cexp): Likewise.
13929 * builtins.def (DEF_C94_BUILTIN): Likewise.
13930 (DEF_C99_BUILTIN): Likewise.
13931 (DEF_C99_C90RES_BUILTIN): Likewise.
13932 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
13933 definitions to using this define.
13934 * config/darwin-protos.h (darwin_libc_has_function): New.
13935 * config/darwin.c (darwin_libc_has_function): Ditto.
13936 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
13937 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
13938 * config/darwin.h: Ditto.
13939 * config/elfos.h: Ditto.
13940 * config/freebsd.h: Ditto.
13941 * config/i386/cygming.h: Ditto.
13942 * config/i386/djgpp.h: Ditto.
13943 * config/i386/i386-interix.h: Ditto.
13944 * config/microblaze/microblaze.h: Ditto.
13945 * config/mmix/mmix.h: Ditto.
13946 * config/gnu-user.h: Ditto.
13947 * config/ia64/hpux.h: Ditto.
13948 * config/pa/pa-hpux.h: Ditto.
13949 * config/pdp11/pdp11.h: Ditto.
13950 * config/picochip/picochip.h: Ditto.
13951 * config/linux.h: Ditto.
13952 * config/netbsd.h: Ditto.
13953 * config/openbsd.h: Ditto.
13954 * config/rs6000/aix43.h: Ditto.
13955 * config/rs6000/aix51.h: Ditto.
13956 * config/rs6000/aix52.h: Ditto.
13957 * config/rs6000/aix53.h: Ditto.
13958 * config/rs6000/aix61.h: Ditto.
13959 * config/rs6000/darwin.h: Ditto.
13960 * config/rs6000/linux.h: Ditto.
13961 * config/rs6000/linux64.h: Ditto.
13962 * config/s390/tpf.h: Ditto.
13963 * config/sol2-10.h: Ditto.
13964 * config/sol2.h: Ditto.
13965 * config/vms/vms.h: Ditto.
13966 * config/vxworks.h: Ditto.
13967 * config/linux-android.c (linux_android_libc_has_function):
13968 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
13969 * config/linux-protos.h (linux_android_libc_has_function):
13971 * config/i386/i386.c (ix86_libc_has_function): New.
13972 * config/i386/i386-protos.h
13973 (ix86_libc_has_function): New declaration.
13974 * config/i386/i386.md
13975 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
13976 ("isinf<mode>2): Likewise.
13977 * convert.c (convert_to_integer): Using new target hook
13978 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
13979 TARGET_C99_FUNCTIONS.
13980 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
13981 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
13982 * coretypes.h (function_class): New enum for different
13983 classes of functions.
13984 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
13985 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
13986 (TARGET_HAS_SINCOS): Likewise.
13987 (TARGET_LIBC_HAS_FUNCTION): New.
13988 * doc/tm.texi: Regenerated.
13989 * targhooks.h (default_libc_has_function): New declaration.
13990 (no_c99_libc_has_function): Ditto.
13991 (gnu_libc_has_function): Ditto.
13992 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
13993 and TARGET_HAS_SINCOS.
13995 2013-08-18 Jan Hubicka <jh@suse.cz>
13997 * Makeifle-in (ipa-devirt.o): New.
13998 (GTFILES): Add ipa-utils.h and ipa-devirt.c
13999 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
14000 (analyze_functions): Look into possible targets of polymorphic call.
14001 * dumpfile.c (dump_files): Add type-inheritance dump.
14002 * dumpfile.h (TDI_inheritance): New.
14003 * ipa-devirt.c: New file.
14004 * ipa-utils.h (odr_type_d): Forward declare.
14005 (odr_type): New type.
14006 (build_type_inheritance_graph): Declare.
14007 (possible_polymorphic_call_targets): Declare and introduce inline
14008 variant when only edge is pased.
14009 (dump_possible_polymorphic_call_targets): Likewise.
14010 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
14011 * tree.c (type_in_anonymous_namespace_p): Break out from ...
14012 (types_same_for_odr): ... here.
14013 * tree.h (type_in_anonymous_namespace_p): Declare.
14015 2013-08-18 Jakub Jelinek <jakub@redhat.com>
14017 PR tree-optimization/58006
14018 * tree-parloops.c (take_address_of): Don't ICE if get_name
14020 (eliminate_local_variables_stmt): Remove clobber stmts.
14022 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
14024 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
14025 error message is issued for an alias to undefined symbol.
14027 2013-08-18 Jan Hubicka <jh@suse.cz>
14029 * cgraph.c (cgraph_create_indirect_edge): Discover
14030 polymorphic calls and record basic info into indirect_info.
14031 * gimple-fold.c (gimple_fold_call): When doing BINFO based
14032 devirtualization, ignore objc function calls.
14033 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
14034 call with no parm index info.
14035 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
14036 * tree.c (virtual_method_call_p): New function.
14037 * tree.h (virtual_method_call_p): Declare.
14039 2013-08-16 Jan Hubicka <jh@suse.cz>
14041 PR middle-end/58179
14042 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
14044 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
14046 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
14049 2013-08-16 David Malcolm <dmalcolm@redhat.com>
14051 * gengtype.c (type_for_name): Add special-case support for
14052 locating types within the "gcc::" namespace.
14053 (open_base_files): Emit a "using namespace gcc" directive.
14055 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
14058 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
14059 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
14061 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
14062 array instead of each individual operand as a separate argument.
14063 (emit_fusion_gpr_load): Likewise.
14064 (expand_fusion_gpr_load): Add new function declaration.
14066 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
14067 signature to have the operands passed as an array, instead of as
14068 separate arguments. Allow ZERO_EXTEND to be in the memory
14069 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
14070 depend on the register live/dead flags when peepholes are run.
14071 (expand_fusion_gpr_load): New function to be called from the
14072 peephole2 pass, to change the register that addis sets to be the
14074 (emit_fusion_gpr_load): Change the calling signature to have the
14075 operands passed as an array, instead of as separate arguments.
14076 Allow ZERO_EXTEND to be in the memory address, and also
14077 SIGN_EXTEND if -mpower8-fusion-sign.
14079 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
14080 unspec enumeration.
14081 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
14082 adjust the register addis loads up in the peephole2 pass. Do not
14083 depend on the register live/dead state when the peephole pass is done.
14085 2013-08-16 David Malcolm <dmalcolm@redhat.com>
14087 * gengtype.c (create_user_defined_type): Ensure that the kind
14088 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
14089 declaration is seen before the GTY((user)) marking.
14091 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
14094 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
14096 2013-08-16 Jan Hubicka <jh@suse.cz>
14098 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
14099 arugment expected_type.
14100 (gimple_fold_call): Use it.
14101 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
14102 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
14103 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
14104 (try_make_edge_direct_virtual_call): Likewise.
14105 * tree.c (obj_type_ref_class): New.
14106 * tree.h (obj_type_ref_class): Use it.
14108 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
14110 * sched-vis.c (rtl_slim_pp_initialized): Remove.
14111 (rtl_slim_pp): Likewise.
14112 (init_rtl_slim_pretty_print): Likewise.
14113 (dump_value_slim): Don't call it. Use local pretty printer.
14114 (dump_insn_slim): Likewise.
14115 (dump_rtl_slim): Likewise.
14116 (str_pattern_slim): Likewise.
14117 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
14120 2013-08-16 Jakub Jelinek <jakub@redhat.com>
14122 PR tree-optimization/58164
14123 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
14124 walk gimple_goto_dest of GIMPLE_GOTO.
14126 PR tree-optimization/58165
14127 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
14128 bi_call must be the last stmt in a bb, don't split_block, instead
14129 use fallthru edge from it and give up if there is none.
14130 Release conds vector when returning early.
14132 2013-08-14 Xinliang David Li <davidxl@google.com>
14134 * config/i386/i386.c (ix86_option_override_internal):
14135 Remove unused variable and field.
14137 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14140 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
14141 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
14142 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
14143 and Linux, correct BLKmode alignment when 128-bit alignment is
14144 required and compatibility flag is not set.
14145 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
14146 for zero-size arguments when compatibility flag is not set.
14148 2013-08-14 Jakub Jelinek <jakub@redhat.com>
14150 PR tree-optimization/58145
14151 * tree-sra.c (build_ref_for_offset): If prev_base has
14152 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
14154 2013-08-14 Xinliang David Li <davidxl@google.com>
14156 * config/i386/i386.c (ix86_option_override_internal):
14157 Fix uninitialized variable error.
14159 2013-08-14 Xinliang David Li <davidxl@google.com>
14161 * config/i386/i386.opt: Define two new options.
14162 * config/i386/x86-tune.def: Add arch selector field in macros.
14163 * config/i386/i386.h: Adjust macro definition.
14164 * config/i386/i386.c (ix86_option_override_internal):
14166 (parse_mtune_ctrl_str): New function.
14167 (set_ix86_tune_features): New function.
14168 (ix86_function_specific_restore): Call the new helper function.
14170 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
14172 PR rtl-optimization/57662
14173 * sel-sched.c (code_motion_process_successors): When the current insn
14174 is removed after the recursive traversal, break from the loop.
14175 Add comments and debug printouts.
14177 2013-08-14 Jakub Jelinek <jakub@redhat.com>
14178 Alexandre Oliva <aoliva@redhat.com>
14181 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
14182 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
14183 there also UNSPEC_PLTOFF.
14185 2013-08-14 Marek Polacek <polacek@redhat.com>
14187 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
14188 NUM_CONDITIONS bit positions.
14190 2013-08-13 Cary Coutant <ccoutant@google.com>
14192 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
14193 (attr_checksum): Hash vector contents instead of pointer.
14194 (attr_checksum_ordered): Likewise.
14196 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
14198 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
14199 when Pmode != word_mode. Add length_address attribute.
14200 (sse3_monitor_<mode>): Merge from sse3_monitor and
14201 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
14202 Pmode != word_mode. Update insn length attribute.
14203 * config/i386/i386.c (ix86_option_override_internal): Update
14204 ix86_gen_monitor selection for merged sse3_monitor insn.
14206 2013-08-13 Julian Brown <julian@codesourcery.com>
14208 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
14209 perform invalid legitimization on greater-than-word-size modes for
14210 TARGET_E500_DOUBLE.
14212 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
14214 * ira.c (setup_class_translate_array): Use aclass instead of cl
14215 for classes not fully covered by allocno classes.
14217 2013-08-13 Jakub Jelinek <jakub@redhat.com>
14219 PR tree-optimization/57661
14220 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
14221 * tree-inline.c (tree_function_versioning): Initialize it.
14222 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
14223 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
14224 that is not being copied.
14227 * asan.c (instrument_strlen_call): Fix typo in comment.
14228 Use char * type even for the lhs of POINTER_PLUS_EXPR.
14230 2013-08-13 Steve Ellcey <sellcey@mips.com>
14232 * config/mips/mips.md (prefetch): Use lw instead of ld on
14233 loongson in 32bit mode.
14235 2013-08-13 Nick Clifton <nickc@redhat.com>
14237 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
14239 2013-08-13 Jan Hubicka <jh@suse.cz>
14241 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
14242 introduced edge; fix typo in sanity check.
14243 (cgraph_resolve_speculation): Export; improve diagnostic.
14244 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
14245 speculation at type mismatch.
14246 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
14247 (cgraph_resolve_speculation): Declare.
14248 (symtab_can_be_discarded): New function.
14249 * value-prof.c (gimple_ic_transform): Remove actual transform code.
14250 * ipa-inline-transform.c (speculation_removed): New global var.
14251 (clone_inlined_nodes): See if speculation can be removed.
14252 (inline_call): If speculations was removed, we growths may not match.
14253 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
14254 (speculation_useful_p): New function.
14255 (resolve_noninline_speculation): New function.
14256 (inline_small_functions): Resolve useless speculations.
14257 * ipa-inline.h (speculation_useful_p): Declare
14258 * ipa.c (can_replace_by_local_alias): Simplify.
14259 (ipa_profile): Produce speculative calls in non-lto, too;
14260 add simple cost model; produce local aliases.
14262 2013-08-13 David Malcolm <dmalcolm@redhat.com>
14264 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
14267 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
14269 * config/i386/i386.c (ix86_function_versions): Use error + inform.
14271 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
14273 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
14274 iterator instead of X87MODEF.
14276 2013-08-12 Perez Read <netfirewall@gmail.com>
14279 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
14280 operand 0 for intel asm alternative.
14281 (*movabs<mode>_2): Ditto for operand 1.
14283 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
14285 * config/aarch64/arm_none.h
14286 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
14288 2013-08-12 Nick Clifton <nickc@redhat.com>
14290 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
14292 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
14294 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
14295 expand for QI/HImode operand to produce more effictive code for
14296 unsigned char(short) --> float(double) conversion.
14298 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
14300 * doc/invoke.texi: Mention that -ftls-model does not force the final
14303 2013-08-12 Marek Polacek <polacek@redhat.com>
14304 Marc Glisse <marc.glisse@inria.fr>
14306 PR tree-optimization/57980
14307 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
14308 when creating -1 constant.
14310 2013-08-10 Jan Hubicka <jh@suse.cz>
14312 Workaround binutils PR14342.
14313 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
14314 (gimple_init_edge_profiler): Likewise.
14315 (gimple_gen_ic_func_profiler): Likewise.
14317 2013-08-09 Jan Hubicka <jh@suse.cz>
14319 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
14321 2013-08-09 Xinliang David Li <davidxl@google.com>
14323 * config/i386/stringop.def: New file.
14324 * config/i386/stringop.opt: New file.
14325 * config/i386/i386-opts.h: Include stringopt.def.
14326 * config/i386/i386.opt: Include stringopt.opt.
14327 * config/i386/i386.c (ix86_option_override_internal):
14328 Override default size based stringop inline strategies with options.
14329 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
14332 2013-08-09 Jan Hubicka <jh@suse.cz>
14334 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
14336 2013-08-09 Jan Hubicka <jh@suse.cz>
14338 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
14340 (dump_cgraph_node): Dump profile-id.
14341 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
14342 and common_target_probability.
14343 * lto-cgraph.c (lto_output_edge): Stream common targets.
14344 (lto_output_node): Stream profile ids.
14345 (input_node): Stream profile ids.
14346 (input_edge): Stream common targets.
14347 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
14348 * ipa.c: Include value-prof.h
14349 (ipa_profile_generate_summary): Turn indirect call statement histograms
14350 into common targets.
14351 (ipa_profile): Turn common targets into speculative edges.
14353 2013-08-09 Jan Hubicka <jh@suse.cz>
14355 * cgraph.h (cgraph_node): Add profile_id.
14356 * value-prof.c (cgraph_node_map): Turn into pointer_map.
14357 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
14358 (del_node_map): Update.
14359 (find_func_by_funcdef_no): Replace by ...
14360 (find_func_by_profile_id): ... this one.
14361 (gimple_ic_transform): Do not remove useful histograms when
14362 speculation is not done; dump info when indirect call removal
14364 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
14365 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
14366 (__gcov_indirect_call_profiler_v2): .. this one.
14367 * profile.h (init_node_map): Update.
14368 * coverage.c (coverage_compute_profile_id): New function.
14369 * coverage.h (coverage_compute_profile_id): Declare.
14370 * tree-profile.c (init_ic_make_global_vars): Make
14371 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
14372 (gimple_init_edge_profiler): Update prototype of
14373 __gcov_indirect_call_profiler.
14374 (gimple_gen_ic_func_profiler): Simplify.
14375 (tree_profiling): Use init_node_map
14377 2013-08-09 Jan Hubicka <jh@suse.cz>
14379 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
14380 non-speculative refs.
14381 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
14382 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
14383 (cgraph_set_call_stmt): Likewise.
14384 (cgraph_create_edge_1): Fix release checking compilatoin;
14385 clear lto_stmt_uid.
14386 (cgraph_free_edge): Free indirect info.
14387 (cgraph_turn_edge_to_speculative): New function.
14388 (cgraph_speculative_call_info): New function.
14389 (cgraph_make_edge_direct): Return direct edge; handle speculation.
14390 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
14391 (dump_cgraph_node): Dump speculation.
14392 (verify_edge_count_and_frequency): Accept speculative edges.
14393 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
14394 (verify_cgraph_node): Handle speculation.
14395 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
14396 (cgraph_set_call_stmt): Update prototype.
14397 (cgraph_make_edge_direct): Update prototype.
14398 (cgraph_speculative_call_info): Declare.
14399 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
14400 to change; update call of ipa_find_references.
14401 * ipa-ref.c (ipa_record_reference): Fix return value; clear
14402 lto_stmt_uid and speculative flags.
14403 (ipa_dump_references): Dump speculation.
14404 (ipa_clone_references): Clone speculative flag.
14405 (ipa_clone_referring): Likewise.
14406 (ipa_clone_ref): New function.
14407 (ipa_find_reference): Look into lto_stmt_uids
14408 (ipa_clear_stmts_in_references): Do not clear speculative calls.
14409 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
14410 (ipa_find_reference): Update declaration.
14411 (ipa_clone_ref): Declare.
14412 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
14413 stream speculative flag.
14414 (lto_output_ref): Stream statements uids and speculation.
14415 (input_ref): Likewise.
14416 (input_edge): Stream speuclation.
14417 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
14418 (cgraph_set_call_stmt_including_clones): Handle speculation.
14419 * ipa-inline.c (heap_edge_removal_hook): New function.
14420 (inline_small_functions): Register it.
14421 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
14422 also initialize refs.
14423 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
14425 (try_make_edge_direct_simple_call): Likewise.
14426 (try_make_edge_direct_simple_call): Likewise.
14427 (update_indirect_edges_after_inlining): Likewise.
14428 (remove_described_reference): Look proper lto_stmt_uid.
14429 (propagate_controlled_uses): Likewise.
14430 (propagate_controlled_uses): Liekwise.
14431 * tree-inline.c (copy_bb): Copy speculative edges.
14432 (redirect_all_calls): New function.
14433 (copy_cfg_body): Do redirection after loop info is updated.
14434 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
14436 2013-08-09 Jan Hubicka <jh@suse.cz>
14438 * lto-streamer-out.c (output_function): Renumber PHIs.
14439 * lto-streamer-in.c (input_function): Likewise.
14441 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
14443 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
14444 (get_lane_unsigned): Likewise.
14445 (dup_lane_scalar): Likewise.
14446 (get_lane): enable for VALL.
14447 * config/aarch64/aarch64-simd.md
14448 (aarch64_dup_lane_scalar<mode>): Remove.
14449 (aarch64_get_lane_signed<mode>): Likewise.
14450 (aarch64_get_lane_unsigned<mode>): Likewise.
14451 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
14452 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
14453 (aarch64_get_lane<mode>): Enable for all vector modes.
14454 (aarch64_get_lanedi): Remove misleading constraints.
14455 * config/aarch64/arm_neon.h
14456 (__aarch64_vget_lane_any): Define.
14457 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
14458 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
14459 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
14460 * config/aarch64/iterators.md (VDQQH): New.
14461 (VDQQHS): Likewise.
14462 (vwcore): Likewise.
14464 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
14466 * configure.ac: Add GAS check for LEON instructions on SPARC.
14467 * configure: Regenerate.
14468 * config.in: Likewise.
14469 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
14471 * config/sparc/sparc.opt (LEON, LEON3): New masks.
14472 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
14474 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
14475 (AS_LEON_FLAG): New macro.
14476 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
14477 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
14478 Deal with LEON and LEON3 for the memory model.
14479 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
14480 (atomic_compare_and_swap<mode>_1): Likewise.
14481 (*atomic_compare_and_swap<mode>_1): Likewise.
14483 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
14485 * config/arm/neon.md (vcond): Fix floating-point vector
14486 comparisons against 0.
14488 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
14490 * lra-constraints.c (emit_spill_move): Remove assert.
14491 (process_alt_operands): Add more debugging
14492 output. Increase reject for spilling into memory. Decrease
14493 reject for reloading scratch.
14494 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
14496 2013-08-08 Steve Ellcey <sellcey@mips.com>
14498 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
14499 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
14500 micromips incompatible. Add nan2008.
14501 (MULTILIB_DIRNAMES): Add nan2008.
14502 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
14503 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
14504 and micromips incompatible. Add nan2008.
14505 (MULTILIB_DIRNAMES): Add nan2008.
14506 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
14508 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
14510 PR rtl-optimization/58079
14511 * combine.c (combine_simplify_rtx): Avoid using SUBST if
14512 simplify_comparison has widened a comparison with an integer.
14514 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14516 * config/arm/neon.md (movmisalign<mode>): Disable when we
14517 don't allow unaligned accesses.
14518 (*movmisalign<mode>_neon_store): Likewise.
14519 (*movmisalign<mode>_neon_load): Likewise.
14520 (*movmisalign<mode>_neon_store): Likewise.
14521 (*movmisalign<mode>_neon_load): Likewise.
14523 2013-08-08 Jan Hubicka <jh@suse.cz>
14525 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
14526 (make_pass_rebuild_cgraph_edges): Also clear references.
14527 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
14528 * ipa-inline-transform.c (inline_transform): Remove all references
14530 * cgraphunit.c (expand_function): Remove all references after
14532 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
14533 (ipa_find_reference): Rewrite to iterator.
14534 (remove_stmt_references): Likewise.
14535 (ipa_clear_stmts_in_references): New function.
14536 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
14537 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
14539 * ipa-split.c (split_function): Remove references in split function.
14541 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
14544 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
14545 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
14547 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
14550 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
14551 suggested mode for the assignment isn't compatible with the
14552 registers required.
14554 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
14557 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
14559 2013-08-07 Xinliang David Li <davidxl@google.com>
14561 * config/i386/i386.opt: New option -mtune-ctrl=.
14562 * config/i386/x86-tune.def: New file.
14563 * config/i386/i386.h: include x86-tune.def.
14564 * config/i386/i386.c (ix86_option_override_internal):
14565 Parsing -mtune-ctrl= option and set tune features.
14567 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
14570 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
14571 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
14574 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
14576 * rtl.h (update_alignments): Declare.
14577 * final.c (grow_label_align): New function extracted from...
14578 (shorten_branches): ...here. Call it.
14579 (update_alignments): New function.
14580 * reorg.c (sibling_labels): New variable.
14581 (get_label_before): Add SIBLING parameter. If it is non-zero, push
14582 the new label along with it onto the sibling_labels vector.
14583 (fill_simple_delay_slots): Adjust call to get_label_before.
14584 (fill_slots_from_thread): Likewise.
14585 (relax_delay_slots): Likewise.
14586 (make_return_insns): Likewise.
14587 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
14589 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
14591 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
14592 document its semantics.
14593 (diagnostic_report_diagnostic): Adjust accordingly.
14595 2013-08-07 David Malcolm <dmalcolm@redhat.com>
14597 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
14598 (sparc_option_override): ...and port to new C++ pass API.
14599 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
14601 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
14603 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
14605 2013-08-06 Caroline Tice <cmtice@google.com>
14607 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
14608 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
14609 * tree-pass.h: Add pass_vtable_verify.
14610 * varasm.c (assemble_variable): Add code to properly set the comdat
14611 section and name for the .vtable_map_vars section.
14612 (assemble_vtyv_preinit_initializer): New function.
14613 (default_sectin_type_flags): Make sure .vtable_map_vars section has
14615 * output.h: Add function decl for assemble_vtv_preinit_initializer.
14616 * vtable-verify.c: New file.
14617 * vtable-verify.h: New file.
14618 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
14619 initialiation levels.
14620 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
14621 * passes.def: Insert pass_vtable_verify.
14622 * aclocal.m4: Reorder includes.
14623 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
14624 -fvtv-counts options.
14625 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
14626 as appropriate, if -fvtable-verify=... is used.
14627 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
14628 -fvtable-verify=... is used.
14629 * Makefile.in (OBJS): Add vtable-verify.o to list.
14630 (vtable-verify.o): Add new build rule.
14631 (GTFILES): Add vtable-verify.c to list.
14632 * common.opt (fvtable-verify=): New flag.
14633 (vtv_priority): Values for fvtable-verify= flag.
14634 (fvtv-counts): New flag.
14635 (fvtv-debug): New flag.
14636 * tree.h (save_vtable_map_decl): New extern function decl.
14638 2013-08-07 David Malcolm <dmalcolm@redhat.com>
14640 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
14641 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
14642 (pass_data_rl78_devirt): ...new pass_data instance and...
14643 (make_pass_rl78_devirt): ...new function.
14644 (rl78_asm_file_start): Port pass registration to new C++ API.
14646 2013-08-07 David Malcolm <dmalcolm@redhat.com>
14648 * coretypes.h (rtl_opt_pass): Add.
14649 (gcc::context): Add.
14650 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
14651 (epiphany_init): Port to new C++ pass API.
14652 (epiphany_optimize_mode_switching): Likewise.
14653 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
14654 (pass_manager::get_pass_mode_switching): New.
14655 (pass_manager::get_pass_peephole2): New.
14656 * mode-switching.c (pass_mode_switching): Add clone method.
14657 * recog.c (pass_peephole2): Add clone method.
14658 (pass_split_all_insns): Add clone method.
14660 2013-08-06 David Malcolm <dmalcolm@redhat.com>
14662 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
14663 (mips_option_override): ...here, porting to new C++ API for passes.
14665 2013-08-06 Jan Hubicka <jh@suse.cz>
14667 * cgraph.c (cgraph_get_body): New function based on lto.c
14669 * cgraph.h (cgraph_get_body): Declare.
14670 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
14672 * cgraphunit.c (expand_function): Get body prior expanding.
14673 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
14674 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
14676 * passes.c (do_per_function_toporder): Get body.
14677 * tree-inline.c (expand_call_inline): Get body prior inlining it.
14678 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
14680 2013-08-06 Martin Jambor <mjambor@suse.cz>
14683 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
14684 is not re-finalized. Rename second parameter to no_collect.
14686 2013-08-06 Martin Jambor <mjambor@suse.cz>
14688 PR middle-end/58041
14689 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
14690 MEM_REF has proper alignment information.
14692 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
14695 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
14697 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
14698 rtx (*) (rtx, ...) with insn_gen_fn.
14699 * genoutput.c (output_insn_data): Cast gen_? function pointers to
14700 insn_gen_fn::stored_funcptr. Add initializer braces.
14702 2013-08-05 David Malcolm <dmalcolm@redhat.com>
14704 Rewrite how instances of passes are cloned to remove assumptions
14705 about their sizes (thus allowing pass subclasses to have
14706 additional data fields, albeit non-GC-managed ones at this point).
14708 * passes.c (make_pass_instance): Now that passes have clone
14709 methods, rewrite this function to eliminate XNEW and memcpy
14710 calls that used hardcoded sizes. Since this function no longer
14711 creates pass instances, rename it to...
14712 (add_pass_instance): ...this. Document the old way that passes were
14713 numbered and flagged, and rework this function to continue using it.
14714 (next_pass_1): Add an initial_pass argument for use by
14716 (position_pass): When adding multiple instances of a pass, use
14717 the pass's clone method, rather than relying on the XNEW/memcpy
14718 within the former make_pass_instance (now add_pass_instance).
14719 (pass_manager::pass_manager): When invoking next_pass_1, also supply
14720 the initial instance of the current pass within the pass manager.
14722 2013-08-05 David Malcolm <dmalcolm@redhat.com>
14724 This is the automated part of the conversion of passes from C
14725 structs to C++ classes.
14727 Patch autogenerated by refactor_passes.py from
14728 https://github.com/davidmalcolm/gcc-refactoring-scripts
14729 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
14731 * asan.c (pass_asan): Convert from a global struct to a subclass of
14732 gimple_opt_pass along with...
14733 (pass_data_asan): ...new pass_data instance and...
14734 (make_pass_asan): ...new function.
14735 (pass_asan_O0): Convert from a global struct to a subclass of
14736 gimple_opt_pass along with...
14737 (pass_data_asan_O0): ...new pass_data instance and...
14738 (make_pass_asan_O0): ...new function.
14739 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
14740 subclass of rtl_opt_pass along with...
14741 (pass_data_inc_dec): ...new pass_data instance and...
14742 (make_pass_inc_dec): ...new function.
14743 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
14744 a subclass of rtl_opt_pass along with...
14745 (pass_data_reorder_blocks): ...new pass_data instance and...
14746 (make_pass_reorder_blocks): ...new function.
14747 (pass_duplicate_computed_gotos): Convert from a global struct to a
14748 subclass of rtl_opt_pass along with...
14749 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
14750 (make_pass_duplicate_computed_gotos): ...new function.
14751 (pass_partition_blocks): Convert from a global struct to a subclass of
14752 rtl_opt_pass along with...
14753 (pass_data_partition_blocks): ...new pass_data instance and...
14754 (make_pass_partition_blocks): ...new function.
14755 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
14756 struct to a subclass of rtl_opt_pass along with...
14757 (pass_data_branch_target_load_optimize1): ...new pass_data instance
14759 (make_pass_branch_target_load_optimize1): ...new function.
14760 (pass_branch_target_load_optimize2): Convert from a global struct to a
14761 subclass of rtl_opt_pass along with...
14762 (pass_data_branch_target_load_optimize2): ...new pass_data instance
14764 (make_pass_branch_target_load_optimize2): ...new function.
14765 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
14766 of rtl_opt_pass along with...
14767 (pass_data_jump): ...new pass_data instance and...
14768 (make_pass_jump): ...new function.
14769 (pass_jump2): Convert from a global struct to a subclass of
14770 rtl_opt_pass along with...
14771 (pass_data_jump2): ...new pass_data instance and...
14772 (make_pass_jump2): ...new function.
14773 * cfgexpand.c (pass_expand): Convert from a global struct to a
14774 subclass of rtl_opt_pass along with...
14775 (pass_data_expand): ...new pass_data instance and...
14776 (make_pass_expand): ...new function.
14777 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
14778 of rtl_opt_pass along with...
14779 (pass_data_free_cfg): ...new pass_data instance and...
14780 (make_pass_free_cfg): ...new function.
14781 (pass_into_cfg_layout_mode): Convert from a global struct to a
14782 subclass of rtl_opt_pass along with...
14783 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
14784 (make_pass_into_cfg_layout_mode): ...new function.
14785 (pass_outof_cfg_layout_mode): Convert from a global struct to a
14786 subclass of rtl_opt_pass along with...
14787 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
14788 (make_pass_outof_cfg_layout_mode): ...new function.
14789 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
14790 struct to a subclass of gimple_opt_pass along with...
14791 (pass_data_build_cgraph_edges): ...new pass_data instance and...
14792 (make_pass_build_cgraph_edges): ...new function.
14793 (pass_rebuild_cgraph_edges): Convert from a global struct to a
14794 subclass of gimple_opt_pass along with...
14795 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
14796 (make_pass_rebuild_cgraph_edges): ...new function.
14797 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
14798 subclass of gimple_opt_pass along with...
14799 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
14801 (make_pass_remove_cgraph_callee_edges): ...new function.
14802 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
14803 struct to a subclass of rtl_opt_pass along with...
14804 (pass_data_stack_adjustments): ...new pass_data instance and...
14805 (make_pass_stack_adjustments): ...new function.
14806 * combine.c (pass_combine): Convert from a global struct to a subclass
14807 of rtl_opt_pass along with...
14808 (pass_data_combine): ...new pass_data instance and...
14809 (make_pass_combine): ...new function.
14810 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
14811 global struct to a subclass of rtl_opt_pass along with...
14812 (pass_data_compare_elim_after_reload): ...new pass_data instance
14814 (make_pass_compare_elim_after_reload): ...new function.
14815 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
14816 of rtl_opt_pass along with...
14817 (pass_data_rtl_cprop): ...new pass_data instance and...
14818 (make_pass_rtl_cprop): ...new function.
14819 * cse.c (pass_cse): Convert from a global struct to a subclass of
14820 rtl_opt_pass along with...
14821 (pass_data_cse): ...new pass_data instance and...
14822 (make_pass_cse): ...new function.
14823 (pass_cse2): Convert from a global struct to a subclass of
14824 rtl_opt_pass along with...
14825 (pass_data_cse2): ...new pass_data instance and...
14826 (make_pass_cse2): ...new function.
14827 (pass_cse_after_global_opts): Convert from a global struct to a
14828 subclass of rtl_opt_pass along with...
14829 (pass_data_cse_after_global_opts): ...new pass_data instance and...
14830 (make_pass_cse_after_global_opts): ...new function.
14831 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
14832 of rtl_opt_pass along with...
14833 (pass_data_ud_rtl_dce): ...new pass_data instance and...
14834 (make_pass_ud_rtl_dce): ...new function.
14835 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
14836 rtl_opt_pass along with...
14837 (pass_data_fast_rtl_dce): ...new pass_data instance and...
14838 (make_pass_fast_rtl_dce): ...new function.
14839 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
14840 a subclass of rtl_opt_pass along with...
14841 (pass_data_df_initialize_opt): ...new pass_data instance and...
14842 (make_pass_df_initialize_opt): ...new function.
14843 (pass_df_initialize_no_opt): Convert from a global struct to a
14844 subclass of rtl_opt_pass along with...
14845 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
14846 (make_pass_df_initialize_no_opt): ...new function.
14847 (pass_df_finish): Convert from a global struct to a subclass of
14848 rtl_opt_pass along with...
14849 (pass_data_df_finish): ...new pass_data instance and...
14850 (make_pass_df_finish): ...new function.
14851 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
14852 rtl_opt_pass along with...
14853 (pass_data_rtl_dse1): ...new pass_data instance and...
14854 (make_pass_rtl_dse1): ...new function.
14855 (pass_rtl_dse2): Convert from a global struct to a subclass of
14856 rtl_opt_pass along with...
14857 (pass_data_rtl_dse2): ...new pass_data instance and...
14858 (make_pass_rtl_dse2): ...new function.
14859 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
14860 subclass of rtl_opt_pass along with...
14861 (pass_data_dwarf2_frame): ...new pass_data instance and...
14862 (make_pass_dwarf2_frame): ...new function.
14863 * except.c (pass_set_nothrow_function_flags): Convert from a global
14864 struct to a subclass of rtl_opt_pass along with...
14865 (pass_data_set_nothrow_function_flags): ...new pass_data instance
14867 (make_pass_set_nothrow_function_flags): ...new function.
14868 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
14869 subclass of rtl_opt_pass along with...
14870 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
14872 (make_pass_convert_to_eh_region_ranges): ...new function.
14873 * final.c (pass_compute_alignments): Convert from a global struct to a
14874 subclass of rtl_opt_pass along with...
14875 (pass_data_compute_alignments): ...new pass_data instance and...
14876 (make_pass_compute_alignments): ...new function.
14877 (pass_final): Convert from a global struct to a subclass of
14878 rtl_opt_pass along with...
14879 (pass_data_final): ...new pass_data instance and...
14880 (make_pass_final): ...new function.
14881 (pass_shorten_branches): Convert from a global struct to a subclass of
14882 rtl_opt_pass along with...
14883 (pass_data_shorten_branches): ...new pass_data instance and...
14884 (make_pass_shorten_branches): ...new function.
14885 (pass_clean_state): Convert from a global struct to a subclass of
14886 rtl_opt_pass along with...
14887 (pass_data_clean_state): ...new pass_data instance and...
14888 (make_pass_clean_state): ...new function.
14889 * function.c (pass_instantiate_virtual_regs): Convert from a global
14890 struct to a subclass of rtl_opt_pass along with...
14891 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
14892 (make_pass_instantiate_virtual_regs): ...new function.
14893 (pass_leaf_regs): Convert from a global struct to a subclass of
14894 rtl_opt_pass along with...
14895 (pass_data_leaf_regs): ...new pass_data instance and...
14896 (make_pass_leaf_regs): ...new function.
14897 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
14898 subclass of rtl_opt_pass along with...
14899 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
14901 (make_pass_thread_prologue_and_epilogue): ...new function.
14902 (pass_match_asm_constraints): Convert from a global struct to a
14903 subclass of rtl_opt_pass along with...
14904 (pass_data_match_asm_constraints): ...new pass_data instance and...
14905 (make_pass_match_asm_constraints): ...new function.
14906 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
14907 subclass of rtl_opt_pass along with...
14908 (pass_data_rtl_fwprop): ...new pass_data instance and...
14909 (make_pass_rtl_fwprop): ...new function.
14910 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
14911 rtl_opt_pass along with...
14912 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
14913 (make_pass_rtl_fwprop_addr): ...new function.
14914 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
14915 rtl_opt_pass along with...
14916 (pass_data_rtl_pre): ...new pass_data instance and...
14917 (make_pass_rtl_pre): ...new function.
14918 (pass_rtl_hoist): Convert from a global struct to a subclass of
14919 rtl_opt_pass along with...
14920 (pass_data_rtl_hoist): ...new pass_data instance and...
14921 (make_pass_rtl_hoist): ...new function.
14922 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
14923 subclass of gimple_opt_pass along with...
14924 (pass_data_lower_cf): ...new pass_data instance and...
14925 (make_pass_lower_cf): ...new function.
14926 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
14927 from a global struct to a subclass of gimple_opt_pass along with...
14928 (pass_data_strength_reduction): ...new pass_data instance and...
14929 (make_pass_strength_reduction): ...new function.
14930 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
14931 of rtl_opt_pass along with...
14932 (pass_data_rtl_ifcvt): ...new pass_data instance and...
14933 (make_pass_rtl_ifcvt): ...new function.
14934 (pass_if_after_combine): Convert from a global struct to a subclass of
14935 rtl_opt_pass along with...
14936 (pass_data_if_after_combine): ...new pass_data instance and...
14937 (make_pass_if_after_combine): ...new function.
14938 (pass_if_after_reload): Convert from a global struct to a subclass of
14939 rtl_opt_pass along with...
14940 (pass_data_if_after_reload): ...new pass_data instance and...
14941 (make_pass_if_after_reload): ...new function.
14942 * init-regs.c (pass_initialize_regs): Convert from a global struct to
14943 a subclass of rtl_opt_pass along with...
14944 (pass_data_initialize_regs): ...new pass_data instance and...
14945 (make_pass_initialize_regs): ...new function.
14946 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
14947 of ipa_opt_pass_d along with...
14948 (pass_data_ipa_cp): ...new pass_data instance and...
14949 (make_pass_ipa_cp): ...new function.
14950 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
14951 global struct to a subclass of gimple_opt_pass along with...
14952 (pass_data_inline_parameters): ...new pass_data instance and...
14953 (make_pass_inline_parameters): ...new function.
14954 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
14955 subclass of gimple_opt_pass along with...
14956 (pass_data_early_inline): ...new pass_data instance and...
14957 (make_pass_early_inline): ...new function.
14958 (pass_ipa_inline): Convert from a global struct to a subclass of
14959 ipa_opt_pass_d along with...
14960 (pass_data_ipa_inline): ...new pass_data instance and...
14961 (make_pass_ipa_inline): ...new function.
14962 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
14963 struct to a subclass of gimple_opt_pass along with...
14964 (pass_data_local_pure_const): ...new pass_data instance and...
14965 (make_pass_local_pure_const): ...new function.
14966 (pass_ipa_pure_const): Convert from a global struct to a subclass of
14967 ipa_opt_pass_d along with...
14968 (pass_data_ipa_pure_const): ...new pass_data instance and...
14969 (make_pass_ipa_pure_const): ...new function.
14970 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
14971 to a subclass of ipa_opt_pass_d along with...
14972 (pass_data_ipa_reference): ...new pass_data instance and...
14973 (make_pass_ipa_reference): ...new function.
14974 * ipa-split.c (pass_split_functions): Convert from a global struct to
14975 a subclass of gimple_opt_pass along with...
14976 (pass_data_split_functions): ...new pass_data instance and...
14977 (make_pass_split_functions): ...new function.
14978 (pass_feedback_split_functions): Convert from a global struct to a
14979 subclass of gimple_opt_pass along with...
14980 (pass_data_feedback_split_functions): ...new pass_data instance and...
14981 (make_pass_feedback_split_functions): ...new function.
14982 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
14983 global struct to a subclass of simple_ipa_opt_pass along with...
14984 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
14986 (make_pass_ipa_function_and_variable_visibility): ...new function.
14987 (pass_ipa_free_inline_summary): Convert from a global struct to a
14988 subclass of simple_ipa_opt_pass along with...
14989 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
14990 (make_pass_ipa_free_inline_summary): ...new function.
14991 (pass_ipa_whole_program_visibility): Convert from a global struct to a
14992 subclass of ipa_opt_pass_d along with...
14993 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
14995 (make_pass_ipa_whole_program_visibility): ...new function.
14996 (pass_ipa_profile): Convert from a global struct to a subclass of
14997 ipa_opt_pass_d along with...
14998 (pass_data_ipa_profile): ...new pass_data instance and...
14999 (make_pass_ipa_profile): ...new function.
15000 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
15001 ipa_opt_pass_d along with...
15002 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
15003 (make_pass_ipa_cdtor_merge): ...new function.
15004 * ira.c (pass_ira): Convert from a global struct to a subclass of
15005 rtl_opt_pass along with...
15006 (pass_data_ira): ...new pass_data instance and...
15007 (make_pass_ira): ...new function.
15008 (pass_reload): Convert from a global struct to a subclass of
15009 rtl_opt_pass along with...
15010 (pass_data_reload): ...new pass_data instance and...
15011 (make_pass_reload): ...new function.
15012 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
15013 subclass of rtl_opt_pass along with...
15014 (pass_data_cleanup_barriers): ...new pass_data instance and...
15015 (make_pass_cleanup_barriers): ...new function.
15016 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
15017 of rtl_opt_pass along with...
15018 (pass_data_loop2): ...new pass_data instance and...
15019 (make_pass_loop2): ...new function.
15020 (pass_rtl_loop_init): Convert from a global struct to a subclass of
15021 rtl_opt_pass along with...
15022 (pass_data_rtl_loop_init): ...new pass_data instance and...
15023 (make_pass_rtl_loop_init): ...new function.
15024 (pass_rtl_loop_done): Convert from a global struct to a subclass of
15025 rtl_opt_pass along with...
15026 (pass_data_rtl_loop_done): ...new pass_data instance and...
15027 (make_pass_rtl_loop_done): ...new function.
15028 (pass_rtl_move_loop_invariants): Convert from a global struct to a
15029 subclass of rtl_opt_pass along with...
15030 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
15031 (make_pass_rtl_move_loop_invariants): ...new function.
15032 (pass_rtl_unswitch): Convert from a global struct to a subclass of
15033 rtl_opt_pass along with...
15034 (pass_data_rtl_unswitch): ...new pass_data instance and...
15035 (make_pass_rtl_unswitch): ...new function.
15036 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
15037 subclass of rtl_opt_pass along with...
15038 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
15040 (make_pass_rtl_unroll_and_peel_loops): ...new function.
15041 (pass_rtl_doloop): Convert from a global struct to a subclass of
15042 rtl_opt_pass along with...
15043 (pass_data_rtl_doloop): ...new pass_data instance and...
15044 (make_pass_rtl_doloop): ...new function.
15045 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
15046 a subclass of rtl_opt_pass along with...
15047 (pass_data_lower_subreg): ...new pass_data instance and...
15048 (make_pass_lower_subreg): ...new function.
15049 (pass_lower_subreg2): Convert from a global struct to a subclass of
15050 rtl_opt_pass along with...
15051 (pass_data_lower_subreg2): ...new pass_data instance and...
15052 (make_pass_lower_subreg2): ...new function.
15053 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
15054 struct to a subclass of ipa_opt_pass_d along with...
15055 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
15056 (make_pass_ipa_lto_gimple_out): ...new function.
15057 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
15058 of ipa_opt_pass_d along with...
15059 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
15060 (make_pass_ipa_lto_finish_out): ...new function.
15061 * mode-switching.c (pass_mode_switching): Convert from a global struct
15062 to a subclass of rtl_opt_pass along with...
15063 (pass_data_mode_switching): ...new pass_data instance and...
15064 (make_pass_mode_switching): ...new function.
15065 * modulo-sched.c (pass_sms): Convert from a global struct to a
15066 subclass of rtl_opt_pass along with...
15067 (pass_data_sms): ...new pass_data instance and...
15068 (make_pass_sms): ...new function.
15069 * omp-low.c (pass_expand_omp): Convert from a global struct to a
15070 subclass of gimple_opt_pass along with...
15071 (pass_data_expand_omp): ...new pass_data instance and...
15072 (make_pass_expand_omp): ...new function.
15073 (pass_lower_omp): Convert from a global struct to a subclass of
15074 gimple_opt_pass along with...
15075 (pass_data_lower_omp): ...new pass_data instance and...
15076 (make_pass_lower_omp): ...new function.
15077 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
15078 of gimple_opt_pass along with...
15079 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
15080 (make_pass_diagnose_omp_blocks): ...new function.
15081 * passes.c (pass_early_local_passes): Convert from a global struct to
15082 a subclass of simple_ipa_opt_pass along with...
15083 (pass_data_early_local_passes): ...new pass_data instance and...
15084 (make_pass_early_local_passes): ...new function.
15085 (pass_all_early_optimizations): Convert from a global struct to a
15086 subclass of gimple_opt_pass along with...
15087 (pass_data_all_early_optimizations): ...new pass_data instance and...
15088 (make_pass_all_early_optimizations): ...new function.
15089 (pass_all_optimizations): Convert from a global struct to a subclass
15090 of gimple_opt_pass along with...
15091 (pass_data_all_optimizations): ...new pass_data instance and...
15092 (make_pass_all_optimizations): ...new function.
15093 (pass_all_optimizations_g): Convert from a global struct to a subclass
15094 of gimple_opt_pass along with...
15095 (pass_data_all_optimizations_g): ...new pass_data instance and...
15096 (make_pass_all_optimizations_g): ...new function.
15097 (pass_rest_of_compilation): Convert from a global struct to a subclass
15098 of rtl_opt_pass along with...
15099 (pass_data_rest_of_compilation): ...new pass_data instance and...
15100 (make_pass_rest_of_compilation): ...new function.
15101 (pass_postreload): Convert from a global struct to a subclass of
15102 rtl_opt_pass along with...
15103 (pass_data_postreload): ...new pass_data instance and...
15104 (make_pass_postreload): ...new function.
15105 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
15106 subclass of rtl_opt_pass along with...
15107 (pass_data_gcse2): ...new pass_data instance and...
15108 (make_pass_gcse2): ...new function.
15109 * postreload.c (pass_postreload_cse): Convert from a global struct to
15110 a subclass of rtl_opt_pass along with...
15111 (pass_data_postreload_cse): ...new pass_data instance and...
15112 (make_pass_postreload_cse): ...new function.
15113 * predict.c (pass_profile): Convert from a global struct to a subclass
15114 of gimple_opt_pass along with...
15115 (pass_data_profile): ...new pass_data instance and...
15116 (make_pass_profile): ...new function.
15117 (pass_strip_predict_hints): Convert from a global struct to a subclass
15118 of gimple_opt_pass along with...
15119 (pass_data_strip_predict_hints): ...new pass_data instance and...
15120 (make_pass_strip_predict_hints): ...new function.
15121 * recog.c (pass_peephole2): Convert from a global struct to a subclass
15122 of rtl_opt_pass along with...
15123 (pass_data_peephole2): ...new pass_data instance and...
15124 (make_pass_peephole2): ...new function.
15125 (pass_split_all_insns): Convert from a global struct to a subclass of
15126 rtl_opt_pass along with...
15127 (pass_data_split_all_insns): ...new pass_data instance and...
15128 (make_pass_split_all_insns): ...new function.
15129 (pass_split_after_reload): Convert from a global struct to a subclass
15130 of rtl_opt_pass along with...
15131 (pass_data_split_after_reload): ...new pass_data instance and...
15132 (make_pass_split_after_reload): ...new function.
15133 (pass_split_before_regstack): Convert from a global struct to a
15134 subclass of rtl_opt_pass along with...
15135 (pass_data_split_before_regstack): ...new pass_data instance and...
15136 (make_pass_split_before_regstack): ...new function.
15137 (pass_split_before_sched2): Convert from a global struct to a subclass
15138 of rtl_opt_pass along with...
15139 (pass_data_split_before_sched2): ...new pass_data instance and...
15140 (make_pass_split_before_sched2): ...new function.
15141 (pass_split_for_shorten_branches): Convert from a global struct to a
15142 subclass of rtl_opt_pass along with...
15143 (pass_data_split_for_shorten_branches): ...new pass_data instance
15145 (make_pass_split_for_shorten_branches): ...new function.
15146 * ree.c (pass_ree): Convert from a global struct to a subclass of
15147 rtl_opt_pass along with...
15148 (pass_data_ree): ...new pass_data instance and...
15149 (make_pass_ree): ...new function.
15150 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
15151 subclass of rtl_opt_pass along with...
15152 (pass_data_stack_regs): ...new pass_data instance and...
15153 (make_pass_stack_regs): ...new function.
15154 (pass_stack_regs_run): Convert from a global struct to a subclass of
15155 rtl_opt_pass along with...
15156 (pass_data_stack_regs_run): ...new pass_data instance and...
15157 (make_pass_stack_regs_run): ...new function.
15158 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
15159 subclass of rtl_opt_pass along with...
15160 (pass_data_cprop_hardreg): ...new pass_data instance and...
15161 (make_pass_cprop_hardreg): ...new function.
15162 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
15163 subclass of rtl_opt_pass along with...
15164 (pass_data_reginfo_init): ...new pass_data instance and...
15165 (make_pass_reginfo_init): ...new function.
15166 * regmove.c (pass_regmove): Convert from a global struct to a subclass
15167 of rtl_opt_pass along with...
15168 (pass_data_regmove): ...new pass_data instance and...
15169 (make_pass_regmove): ...new function.
15170 * regrename.c (pass_regrename): Convert from a global struct to a
15171 subclass of rtl_opt_pass along with...
15172 (pass_data_regrename): ...new pass_data instance and...
15173 (make_pass_regrename): ...new function.
15174 * reorg.c (pass_delay_slots): Convert from a global struct to a
15175 subclass of rtl_opt_pass along with...
15176 (pass_data_delay_slots): ...new pass_data instance and...
15177 (make_pass_delay_slots): ...new function.
15178 (pass_machine_reorg): Convert from a global struct to a subclass of
15179 rtl_opt_pass along with...
15180 (pass_data_machine_reorg): ...new pass_data instance and...
15181 (make_pass_machine_reorg): ...new function.
15182 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
15183 of rtl_opt_pass along with...
15184 (pass_data_sched): ...new pass_data instance and...
15185 (make_pass_sched): ...new function.
15186 (pass_sched2): Convert from a global struct to a subclass of
15187 rtl_opt_pass along with...
15188 (pass_data_sched2): ...new pass_data instance and...
15189 (make_pass_sched2): ...new function.
15190 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
15191 to a subclass of rtl_opt_pass along with...
15192 (pass_data_stack_ptr_mod): ...new pass_data instance and...
15193 (make_pass_stack_ptr_mod): ...new function.
15194 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
15195 to a subclass of rtl_opt_pass along with...
15196 (pass_data_rtl_store_motion): ...new pass_data instance and...
15197 (make_pass_rtl_store_motion): ...new function.
15198 * tracer.c (pass_tracer): Convert from a global struct to a subclass
15199 of gimple_opt_pass along with...
15200 (pass_data_tracer): ...new pass_data instance and...
15201 (make_pass_tracer): ...new function.
15202 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
15203 to a subclass of gimple_opt_pass along with...
15204 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
15205 (make_pass_diagnose_tm_blocks): ...new function.
15206 (pass_lower_tm): Convert from a global struct to a subclass of
15207 gimple_opt_pass along with...
15208 (pass_data_lower_tm): ...new pass_data instance and...
15209 (make_pass_lower_tm): ...new function.
15210 (pass_tm_init): Convert from a global struct to a subclass of
15211 gimple_opt_pass along with...
15212 (pass_data_tm_init): ...new pass_data instance and...
15213 (make_pass_tm_init): ...new function.
15214 (pass_tm_mark): Convert from a global struct to a subclass of
15215 gimple_opt_pass along with...
15216 (pass_data_tm_mark): ...new pass_data instance and...
15217 (make_pass_tm_mark): ...new function.
15218 (pass_tm_edges): Convert from a global struct to a subclass of
15219 gimple_opt_pass along with...
15220 (pass_data_tm_edges): ...new pass_data instance and...
15221 (make_pass_tm_edges): ...new function.
15222 (pass_tm_memopt): Convert from a global struct to a subclass of
15223 gimple_opt_pass along with...
15224 (pass_data_tm_memopt): ...new pass_data instance and...
15225 (make_pass_tm_memopt): ...new function.
15226 (pass_ipa_tm): Convert from a global struct to a subclass of
15227 simple_ipa_opt_pass along with...
15228 (pass_data_ipa_tm): ...new pass_data instance and...
15229 (make_pass_ipa_tm): ...new function.
15230 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
15231 subclass of gimple_opt_pass along with...
15232 (pass_data_call_cdce): ...new pass_data instance and...
15233 (make_pass_call_cdce): ...new function.
15234 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
15235 subclass of gimple_opt_pass along with...
15236 (pass_data_build_cfg): ...new pass_data instance and...
15237 (make_pass_build_cfg): ...new function.
15238 (pass_split_crit_edges): Convert from a global struct to a subclass of
15239 gimple_opt_pass along with...
15240 (pass_data_split_crit_edges): ...new pass_data instance and...
15241 (make_pass_split_crit_edges): ...new function.
15242 (pass_warn_function_return): Convert from a global struct to a
15243 subclass of gimple_opt_pass along with...
15244 (pass_data_warn_function_return): ...new pass_data instance and...
15245 (make_pass_warn_function_return): ...new function.
15246 (pass_warn_function_noreturn): Convert from a global struct to a
15247 subclass of gimple_opt_pass along with...
15248 (pass_data_warn_function_noreturn): ...new pass_data instance and...
15249 (make_pass_warn_function_noreturn): ...new function.
15250 (pass_warn_unused_result): Convert from a global struct to a subclass
15251 of gimple_opt_pass along with...
15252 (pass_data_warn_unused_result): ...new pass_data instance and...
15253 (make_pass_warn_unused_result): ...new function.
15254 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
15255 a subclass of gimple_opt_pass along with...
15256 (pass_data_merge_phi): ...new pass_data instance and...
15257 (make_pass_merge_phi): ...new function.
15258 * tree-complex.c (pass_lower_complex): Convert from a global struct to
15259 a subclass of gimple_opt_pass along with...
15260 (pass_data_lower_complex): ...new pass_data instance and...
15261 (make_pass_lower_complex): ...new function.
15262 (pass_lower_complex_O0): Convert from a global struct to a subclass of
15263 gimple_opt_pass along with...
15264 (pass_data_lower_complex_O0): ...new pass_data instance and...
15265 (make_pass_lower_complex_O0): ...new function.
15266 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
15267 subclass of gimple_opt_pass along with...
15268 (pass_data_lower_eh): ...new pass_data instance and...
15269 (make_pass_lower_eh): ...new function.
15270 (pass_refactor_eh): Convert from a global struct to a subclass of
15271 gimple_opt_pass along with...
15272 (pass_data_refactor_eh): ...new pass_data instance and...
15273 (make_pass_refactor_eh): ...new function.
15274 (pass_lower_resx): Convert from a global struct to a subclass of
15275 gimple_opt_pass along with...
15276 (pass_data_lower_resx): ...new pass_data instance and...
15277 (make_pass_lower_resx): ...new function.
15278 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
15279 of gimple_opt_pass along with...
15280 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
15281 (make_pass_lower_eh_dispatch): ...new function.
15282 (pass_cleanup_eh): Convert from a global struct to a subclass of
15283 gimple_opt_pass along with...
15284 (pass_data_cleanup_eh): ...new pass_data instance and...
15285 (make_pass_cleanup_eh): ...new function.
15286 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
15287 to a subclass of simple_ipa_opt_pass along with...
15288 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
15289 (make_pass_ipa_lower_emutls): ...new function.
15290 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
15291 a subclass of gimple_opt_pass along with...
15292 (pass_data_if_conversion): ...new pass_data instance and...
15293 (make_pass_if_conversion): ...new function.
15294 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
15295 subclass of gimple_opt_pass along with...
15296 (pass_data_build_ssa): ...new pass_data instance and...
15297 (make_pass_build_ssa): ...new function.
15298 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
15299 global struct to a subclass of gimple_opt_pass along with...
15300 (pass_data_loop_distribution): ...new pass_data instance and...
15301 (make_pass_loop_distribution): ...new function.
15302 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
15303 subclass of gimple_opt_pass along with...
15304 (pass_data_mudflap_1): ...new pass_data instance and...
15305 (make_pass_mudflap_1): ...new function.
15306 (pass_mudflap_2): Convert from a global struct to a subclass of
15307 gimple_opt_pass along with...
15308 (pass_data_mudflap_2): ...new pass_data instance and...
15309 (make_pass_mudflap_2): ...new function.
15310 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
15311 subclass of gimple_opt_pass along with...
15312 (pass_data_mudflap_1): ...new pass_data instance and...
15313 (make_pass_mudflap_1): ...new function.
15314 (pass_mudflap_2): Convert from a global struct to a subclass of
15315 gimple_opt_pass along with...
15316 (pass_data_mudflap_2): ...new pass_data instance and...
15317 (make_pass_mudflap_2): ...new function.
15318 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
15319 gimple_opt_pass along with...
15320 (pass_data_nrv): ...new pass_data instance and...
15321 (make_pass_nrv): ...new function.
15322 (pass_return_slot): Convert from a global struct to a subclass of
15323 gimple_opt_pass along with...
15324 (pass_data_return_slot): ...new pass_data instance and...
15325 (make_pass_return_slot): ...new function.
15326 * tree-object-size.c (pass_object_sizes): Convert from a global struct
15327 to a subclass of gimple_opt_pass along with...
15328 (pass_data_object_sizes): ...new pass_data instance and...
15329 (make_pass_object_sizes): ...new function.
15330 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
15331 global struct to a subclass of gimple_opt_pass along with...
15332 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
15334 (make_pass_cleanup_cfg_post_optimizing): ...new function.
15335 (pass_fixup_cfg): Convert from a global struct to a subclass of
15336 gimple_opt_pass along with...
15337 (pass_data_fixup_cfg): ...new pass_data instance and...
15338 (make_pass_fixup_cfg): ...new function.
15339 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
15340 (make_pass_mudflap_1): ...new function.
15341 (pass_mudflap_2): Replace declaration with that of...
15342 (make_pass_mudflap_2): ...new function.
15343 (pass_asan): Replace declaration with that of...
15344 (make_pass_asan): ...new function.
15345 (pass_asan_O0): Replace declaration with that of...
15346 (make_pass_asan_O0): ...new function.
15347 (pass_tsan): Replace declaration with that of...
15348 (make_pass_tsan): ...new function.
15349 (pass_tsan_O0): Replace declaration with that of...
15350 (make_pass_tsan_O0): ...new function.
15351 (pass_lower_cf): Replace declaration with that of...
15352 (make_pass_lower_cf): ...new function.
15353 (pass_refactor_eh): Replace declaration with that of...
15354 (make_pass_refactor_eh): ...new function.
15355 (pass_lower_eh): Replace declaration with that of...
15356 (make_pass_lower_eh): ...new function.
15357 (pass_lower_eh_dispatch): Replace declaration with that of...
15358 (make_pass_lower_eh_dispatch): ...new function.
15359 (pass_lower_resx): Replace declaration with that of...
15360 (make_pass_lower_resx): ...new function.
15361 (pass_build_cfg): Replace declaration with that of...
15362 (make_pass_build_cfg): ...new function.
15363 (pass_early_tree_profile): Replace declaration with that of...
15364 (make_pass_early_tree_profile): ...new function.
15365 (pass_cleanup_eh): Replace declaration with that of...
15366 (make_pass_cleanup_eh): ...new function.
15367 (pass_sra): Replace declaration with that of...
15368 (make_pass_sra): ...new function.
15369 (pass_sra_early): Replace declaration with that of...
15370 (make_pass_sra_early): ...new function.
15371 (pass_early_ipa_sra): Replace declaration with that of...
15372 (make_pass_early_ipa_sra): ...new function.
15373 (pass_tail_recursion): Replace declaration with that of...
15374 (make_pass_tail_recursion): ...new function.
15375 (pass_tail_calls): Replace declaration with that of...
15376 (make_pass_tail_calls): ...new function.
15377 (pass_tree_loop): Replace declaration with that of...
15378 (make_pass_tree_loop): ...new function.
15379 (pass_tree_loop_init): Replace declaration with that of...
15380 (make_pass_tree_loop_init): ...new function.
15381 (pass_lim): Replace declaration with that of...
15382 (make_pass_lim): ...new function.
15383 (pass_tree_unswitch): Replace declaration with that of...
15384 (make_pass_tree_unswitch): ...new function.
15385 (pass_predcom): Replace declaration with that of...
15386 (make_pass_predcom): ...new function.
15387 (pass_iv_canon): Replace declaration with that of...
15388 (make_pass_iv_canon): ...new function.
15389 (pass_scev_cprop): Replace declaration with that of...
15390 (make_pass_scev_cprop): ...new function.
15391 (pass_empty_loop): Replace declaration with that of...
15392 (make_pass_empty_loop): ...new function.
15393 (pass_record_bounds): Replace declaration with that of...
15394 (make_pass_record_bounds): ...new function.
15395 (pass_graphite): Replace declaration with that of...
15396 (make_pass_graphite): ...new function.
15397 (pass_graphite_transforms): Replace declaration with that of...
15398 (make_pass_graphite_transforms): ...new function.
15399 (pass_if_conversion): Replace declaration with that of...
15400 (make_pass_if_conversion): ...new function.
15401 (pass_loop_distribution): Replace declaration with that of...
15402 (make_pass_loop_distribution): ...new function.
15403 (pass_vectorize): Replace declaration with that of...
15404 (make_pass_vectorize): ...new function.
15405 (pass_slp_vectorize): Replace declaration with that of...
15406 (make_pass_slp_vectorize): ...new function.
15407 (pass_complete_unroll): Replace declaration with that of...
15408 (make_pass_complete_unroll): ...new function.
15409 (pass_complete_unrolli): Replace declaration with that of...
15410 (make_pass_complete_unrolli): ...new function.
15411 (pass_parallelize_loops): Replace declaration with that of...
15412 (make_pass_parallelize_loops): ...new function.
15413 (pass_loop_prefetch): Replace declaration with that of...
15414 (make_pass_loop_prefetch): ...new function.
15415 (pass_iv_optimize): Replace declaration with that of...
15416 (make_pass_iv_optimize): ...new function.
15417 (pass_tree_loop_done): Replace declaration with that of...
15418 (make_pass_tree_loop_done): ...new function.
15419 (pass_ch): Replace declaration with that of...
15420 (make_pass_ch): ...new function.
15421 (pass_ccp): Replace declaration with that of...
15422 (make_pass_ccp): ...new function.
15423 (pass_phi_only_cprop): Replace declaration with that of...
15424 (make_pass_phi_only_cprop): ...new function.
15425 (pass_build_ssa): Replace declaration with that of...
15426 (make_pass_build_ssa): ...new function.
15427 (pass_build_alias): Replace declaration with that of...
15428 (make_pass_build_alias): ...new function.
15429 (pass_build_ealias): Replace declaration with that of...
15430 (make_pass_build_ealias): ...new function.
15431 (pass_dominator): Replace declaration with that of...
15432 (make_pass_dominator): ...new function.
15433 (pass_dce): Replace declaration with that of...
15434 (make_pass_dce): ...new function.
15435 (pass_dce_loop): Replace declaration with that of...
15436 (make_pass_dce_loop): ...new function.
15437 (pass_cd_dce): Replace declaration with that of...
15438 (make_pass_cd_dce): ...new function.
15439 (pass_call_cdce): Replace declaration with that of...
15440 (make_pass_call_cdce): ...new function.
15441 (pass_merge_phi): Replace declaration with that of...
15442 (make_pass_merge_phi): ...new function.
15443 (pass_split_crit_edges): Replace declaration with that of...
15444 (make_pass_split_crit_edges): ...new function.
15445 (pass_pre): Replace declaration with that of...
15446 (make_pass_pre): ...new function.
15447 (pass_profile): Replace declaration with that of...
15448 (make_pass_profile): ...new function.
15449 (pass_strip_predict_hints): Replace declaration with that of...
15450 (make_pass_strip_predict_hints): ...new function.
15451 (pass_lower_complex_O0): Replace declaration with that of...
15452 (make_pass_lower_complex_O0): ...new function.
15453 (pass_lower_complex): Replace declaration with that of...
15454 (make_pass_lower_complex): ...new function.
15455 (pass_lower_vector): Replace declaration with that of...
15456 (make_pass_lower_vector): ...new function.
15457 (pass_lower_vector_ssa): Replace declaration with that of...
15458 (make_pass_lower_vector_ssa): ...new function.
15459 (pass_lower_omp): Replace declaration with that of...
15460 (make_pass_lower_omp): ...new function.
15461 (pass_diagnose_omp_blocks): Replace declaration with that of...
15462 (make_pass_diagnose_omp_blocks): ...new function.
15463 (pass_expand_omp): Replace declaration with that of...
15464 (make_pass_expand_omp): ...new function.
15465 (pass_expand_omp_ssa): Replace declaration with that of...
15466 (make_pass_expand_omp_ssa): ...new function.
15467 (pass_object_sizes): Replace declaration with that of...
15468 (make_pass_object_sizes): ...new function.
15469 (pass_strlen): Replace declaration with that of...
15470 (make_pass_strlen): ...new function.
15471 (pass_fold_builtins): Replace declaration with that of...
15472 (make_pass_fold_builtins): ...new function.
15473 (pass_stdarg): Replace declaration with that of...
15474 (make_pass_stdarg): ...new function.
15475 (pass_early_warn_uninitialized): Replace declaration with that of...
15476 (make_pass_early_warn_uninitialized): ...new function.
15477 (pass_late_warn_uninitialized): Replace declaration with that of...
15478 (make_pass_late_warn_uninitialized): ...new function.
15479 (pass_cse_reciprocals): Replace declaration with that of...
15480 (make_pass_cse_reciprocals): ...new function.
15481 (pass_cse_sincos): Replace declaration with that of...
15482 (make_pass_cse_sincos): ...new function.
15483 (pass_optimize_bswap): Replace declaration with that of...
15484 (make_pass_optimize_bswap): ...new function.
15485 (pass_optimize_widening_mul): Replace declaration with that of...
15486 (make_pass_optimize_widening_mul): ...new function.
15487 (pass_warn_function_return): Replace declaration with that of...
15488 (make_pass_warn_function_return): ...new function.
15489 (pass_warn_function_noreturn): Replace declaration with that of...
15490 (make_pass_warn_function_noreturn): ...new function.
15491 (pass_cselim): Replace declaration with that of...
15492 (make_pass_cselim): ...new function.
15493 (pass_phiopt): Replace declaration with that of...
15494 (make_pass_phiopt): ...new function.
15495 (pass_forwprop): Replace declaration with that of...
15496 (make_pass_forwprop): ...new function.
15497 (pass_phiprop): Replace declaration with that of...
15498 (make_pass_phiprop): ...new function.
15499 (pass_tree_ifcombine): Replace declaration with that of...
15500 (make_pass_tree_ifcombine): ...new function.
15501 (pass_dse): Replace declaration with that of...
15502 (make_pass_dse): ...new function.
15503 (pass_nrv): Replace declaration with that of...
15504 (make_pass_nrv): ...new function.
15505 (pass_rename_ssa_copies): Replace declaration with that of...
15506 (make_pass_rename_ssa_copies): ...new function.
15507 (pass_sink_code): Replace declaration with that of...
15508 (make_pass_sink_code): ...new function.
15509 (pass_fre): Replace declaration with that of...
15510 (make_pass_fre): ...new function.
15511 (pass_check_data_deps): Replace declaration with that of...
15512 (make_pass_check_data_deps): ...new function.
15513 (pass_copy_prop): Replace declaration with that of...
15514 (make_pass_copy_prop): ...new function.
15515 (pass_vrp): Replace declaration with that of...
15516 (make_pass_vrp): ...new function.
15517 (pass_uncprop): Replace declaration with that of...
15518 (make_pass_uncprop): ...new function.
15519 (pass_return_slot): Replace declaration with that of...
15520 (make_pass_return_slot): ...new function.
15521 (pass_reassoc): Replace declaration with that of...
15522 (make_pass_reassoc): ...new function.
15523 (pass_rebuild_cgraph_edges): Replace declaration with that of...
15524 (make_pass_rebuild_cgraph_edges): ...new function.
15525 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
15526 (make_pass_remove_cgraph_callee_edges): ...new function.
15527 (pass_build_cgraph_edges): Replace declaration with that of...
15528 (make_pass_build_cgraph_edges): ...new function.
15529 (pass_local_pure_const): Replace declaration with that of...
15530 (make_pass_local_pure_const): ...new function.
15531 (pass_tracer): Replace declaration with that of...
15532 (make_pass_tracer): ...new function.
15533 (pass_warn_unused_result): Replace declaration with that of...
15534 (make_pass_warn_unused_result): ...new function.
15535 (pass_diagnose_tm_blocks): Replace declaration with that of...
15536 (make_pass_diagnose_tm_blocks): ...new function.
15537 (pass_lower_tm): Replace declaration with that of...
15538 (make_pass_lower_tm): ...new function.
15539 (pass_tm_init): Replace declaration with that of...
15540 (make_pass_tm_init): ...new function.
15541 (pass_tm_mark): Replace declaration with that of...
15542 (make_pass_tm_mark): ...new function.
15543 (pass_tm_memopt): Replace declaration with that of...
15544 (make_pass_tm_memopt): ...new function.
15545 (pass_tm_edges): Replace declaration with that of...
15546 (make_pass_tm_edges): ...new function.
15547 (pass_split_functions): Replace declaration with that of...
15548 (make_pass_split_functions): ...new function.
15549 (pass_feedback_split_functions): Replace declaration with that of...
15550 (make_pass_feedback_split_functions): ...new function.
15551 (pass_strength_reduction): Replace declaration with that of...
15552 (make_pass_strength_reduction): ...new function.
15553 (pass_ipa_lower_emutls): Replace declaration with that of...
15554 (make_pass_ipa_lower_emutls): ...new function.
15555 (pass_ipa_function_and_variable_visibility): Replace declaration with
15557 (make_pass_ipa_function_and_variable_visibility): ...new function.
15558 (pass_ipa_tree_profile): Replace declaration with that of...
15559 (make_pass_ipa_tree_profile): ...new function.
15560 (pass_early_local_passes): Replace declaration with that of...
15561 (make_pass_early_local_passes): ...new function.
15562 (pass_ipa_whole_program_visibility): Replace declaration with that
15564 (make_pass_ipa_whole_program_visibility): ...new function.
15565 (pass_ipa_lto_gimple_out): Replace declaration with that of...
15566 (make_pass_ipa_lto_gimple_out): ...new function.
15567 (pass_ipa_increase_alignment): Replace declaration with that of...
15568 (make_pass_ipa_increase_alignment): ...new function.
15569 (pass_ipa_inline): Replace declaration with that of...
15570 (make_pass_ipa_inline): ...new function.
15571 (pass_ipa_free_lang_data): Replace declaration with that of...
15572 (make_pass_ipa_free_lang_data): ...new function.
15573 (pass_ipa_free_inline_summary): Replace declaration with that of...
15574 (make_pass_ipa_free_inline_summary): ...new function.
15575 (pass_ipa_cp): Replace declaration with that of...
15576 (make_pass_ipa_cp): ...new function.
15577 (pass_ipa_reference): Replace declaration with that of...
15578 (make_pass_ipa_reference): ...new function.
15579 (pass_ipa_pure_const): Replace declaration with that of...
15580 (make_pass_ipa_pure_const): ...new function.
15581 (pass_ipa_pta): Replace declaration with that of...
15582 (make_pass_ipa_pta): ...new function.
15583 (pass_ipa_lto_finish_out): Replace declaration with that of...
15584 (make_pass_ipa_lto_finish_out): ...new function.
15585 (pass_ipa_tm): Replace declaration with that of...
15586 (make_pass_ipa_tm): ...new function.
15587 (pass_ipa_profile): Replace declaration with that of...
15588 (make_pass_ipa_profile): ...new function.
15589 (pass_ipa_cdtor_merge): Replace declaration with that of...
15590 (make_pass_ipa_cdtor_merge): ...new function.
15591 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
15593 (make_pass_cleanup_cfg_post_optimizing): ...new function.
15594 (pass_init_datastructures): Replace declaration with that of...
15595 (make_pass_init_datastructures): ...new function.
15596 (pass_fixup_cfg): Replace declaration with that of...
15597 (make_pass_fixup_cfg): ...new function.
15598 (pass_expand): Replace declaration with that of...
15599 (make_pass_expand): ...new function.
15600 (pass_instantiate_virtual_regs): Replace declaration with that of...
15601 (make_pass_instantiate_virtual_regs): ...new function.
15602 (pass_rtl_fwprop): Replace declaration with that of...
15603 (make_pass_rtl_fwprop): ...new function.
15604 (pass_rtl_fwprop_addr): Replace declaration with that of...
15605 (make_pass_rtl_fwprop_addr): ...new function.
15606 (pass_jump): Replace declaration with that of...
15607 (make_pass_jump): ...new function.
15608 (pass_jump2): Replace declaration with that of...
15609 (make_pass_jump2): ...new function.
15610 (pass_lower_subreg): Replace declaration with that of...
15611 (make_pass_lower_subreg): ...new function.
15612 (pass_cse): Replace declaration with that of...
15613 (make_pass_cse): ...new function.
15614 (pass_fast_rtl_dce): Replace declaration with that of...
15615 (make_pass_fast_rtl_dce): ...new function.
15616 (pass_ud_rtl_dce): Replace declaration with that of...
15617 (make_pass_ud_rtl_dce): ...new function.
15618 (pass_rtl_dce): Replace declaration with that of...
15619 (make_pass_rtl_dce): ...new function.
15620 (pass_rtl_dse1): Replace declaration with that of...
15621 (make_pass_rtl_dse1): ...new function.
15622 (pass_rtl_dse2): Replace declaration with that of...
15623 (make_pass_rtl_dse2): ...new function.
15624 (pass_rtl_dse3): Replace declaration with that of...
15625 (make_pass_rtl_dse3): ...new function.
15626 (pass_rtl_cprop): Replace declaration with that of...
15627 (make_pass_rtl_cprop): ...new function.
15628 (pass_rtl_pre): Replace declaration with that of...
15629 (make_pass_rtl_pre): ...new function.
15630 (pass_rtl_hoist): Replace declaration with that of...
15631 (make_pass_rtl_hoist): ...new function.
15632 (pass_rtl_store_motion): Replace declaration with that of...
15633 (make_pass_rtl_store_motion): ...new function.
15634 (pass_cse_after_global_opts): Replace declaration with that of...
15635 (make_pass_cse_after_global_opts): ...new function.
15636 (pass_rtl_ifcvt): Replace declaration with that of...
15637 (make_pass_rtl_ifcvt): ...new function.
15638 (pass_into_cfg_layout_mode): Replace declaration with that of...
15639 (make_pass_into_cfg_layout_mode): ...new function.
15640 (pass_outof_cfg_layout_mode): Replace declaration with that of...
15641 (make_pass_outof_cfg_layout_mode): ...new function.
15642 (pass_loop2): Replace declaration with that of...
15643 (make_pass_loop2): ...new function.
15644 (pass_rtl_loop_init): Replace declaration with that of...
15645 (make_pass_rtl_loop_init): ...new function.
15646 (pass_rtl_move_loop_invariants): Replace declaration with that of...
15647 (make_pass_rtl_move_loop_invariants): ...new function.
15648 (pass_rtl_unswitch): Replace declaration with that of...
15649 (make_pass_rtl_unswitch): ...new function.
15650 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
15651 (make_pass_rtl_unroll_and_peel_loops): ...new function.
15652 (pass_rtl_doloop): Replace declaration with that of...
15653 (make_pass_rtl_doloop): ...new function.
15654 (pass_rtl_loop_done): Replace declaration with that of...
15655 (make_pass_rtl_loop_done): ...new function.
15656 (pass_web): Replace declaration with that of...
15657 (make_pass_web): ...new function.
15658 (pass_cse2): Replace declaration with that of...
15659 (make_pass_cse2): ...new function.
15660 (pass_df_initialize_opt): Replace declaration with that of...
15661 (make_pass_df_initialize_opt): ...new function.
15662 (pass_df_initialize_no_opt): Replace declaration with that of...
15663 (make_pass_df_initialize_no_opt): ...new function.
15664 (pass_reginfo_init): Replace declaration with that of...
15665 (make_pass_reginfo_init): ...new function.
15666 (pass_inc_dec): Replace declaration with that of...
15667 (make_pass_inc_dec): ...new function.
15668 (pass_stack_ptr_mod): Replace declaration with that of...
15669 (make_pass_stack_ptr_mod): ...new function.
15670 (pass_initialize_regs): Replace declaration with that of...
15671 (make_pass_initialize_regs): ...new function.
15672 (pass_combine): Replace declaration with that of...
15673 (make_pass_combine): ...new function.
15674 (pass_if_after_combine): Replace declaration with that of...
15675 (make_pass_if_after_combine): ...new function.
15676 (pass_ree): Replace declaration with that of...
15677 (make_pass_ree): ...new function.
15678 (pass_partition_blocks): Replace declaration with that of...
15679 (make_pass_partition_blocks): ...new function.
15680 (pass_match_asm_constraints): Replace declaration with that of...
15681 (make_pass_match_asm_constraints): ...new function.
15682 (pass_regmove): Replace declaration with that of...
15683 (make_pass_regmove): ...new function.
15684 (pass_split_all_insns): Replace declaration with that of...
15685 (make_pass_split_all_insns): ...new function.
15686 (pass_fast_rtl_byte_dce): Replace declaration with that of...
15687 (make_pass_fast_rtl_byte_dce): ...new function.
15688 (pass_lower_subreg2): Replace declaration with that of...
15689 (make_pass_lower_subreg2): ...new function.
15690 (pass_mode_switching): Replace declaration with that of...
15691 (make_pass_mode_switching): ...new function.
15692 (pass_sms): Replace declaration with that of...
15693 (make_pass_sms): ...new function.
15694 (pass_sched): Replace declaration with that of...
15695 (make_pass_sched): ...new function.
15696 (pass_ira): Replace declaration with that of...
15697 (make_pass_ira): ...new function.
15698 (pass_reload): Replace declaration with that of...
15699 (make_pass_reload): ...new function.
15700 (pass_clean_state): Replace declaration with that of...
15701 (make_pass_clean_state): ...new function.
15702 (pass_branch_prob): Replace declaration with that of...
15703 (make_pass_branch_prob): ...new function.
15704 (pass_value_profile_transformations): Replace declaration with that
15706 (make_pass_value_profile_transformations): ...new function.
15707 (pass_postreload_cse): Replace declaration with that of...
15708 (make_pass_postreload_cse): ...new function.
15709 (pass_gcse2): Replace declaration with that of...
15710 (make_pass_gcse2): ...new function.
15711 (pass_split_after_reload): Replace declaration with that of...
15712 (make_pass_split_after_reload): ...new function.
15713 (pass_branch_target_load_optimize1): Replace declaration with that
15715 (make_pass_branch_target_load_optimize1): ...new function.
15716 (pass_thread_prologue_and_epilogue): Replace declaration with that
15718 (make_pass_thread_prologue_and_epilogue): ...new function.
15719 (pass_stack_adjustments): Replace declaration with that of...
15720 (make_pass_stack_adjustments): ...new function.
15721 (pass_peephole2): Replace declaration with that of...
15722 (make_pass_peephole2): ...new function.
15723 (pass_if_after_reload): Replace declaration with that of...
15724 (make_pass_if_after_reload): ...new function.
15725 (pass_regrename): Replace declaration with that of...
15726 (make_pass_regrename): ...new function.
15727 (pass_cprop_hardreg): Replace declaration with that of...
15728 (make_pass_cprop_hardreg): ...new function.
15729 (pass_reorder_blocks): Replace declaration with that of...
15730 (make_pass_reorder_blocks): ...new function.
15731 (pass_branch_target_load_optimize2): Replace declaration with that
15733 (make_pass_branch_target_load_optimize2): ...new function.
15734 (pass_leaf_regs): Replace declaration with that of...
15735 (make_pass_leaf_regs): ...new function.
15736 (pass_split_before_sched2): Replace declaration with that of...
15737 (make_pass_split_before_sched2): ...new function.
15738 (pass_compare_elim_after_reload): Replace declaration with that of...
15739 (make_pass_compare_elim_after_reload): ...new function.
15740 (pass_sched2): Replace declaration with that of...
15741 (make_pass_sched2): ...new function.
15742 (pass_stack_regs): Replace declaration with that of...
15743 (make_pass_stack_regs): ...new function.
15744 (pass_stack_regs_run): Replace declaration with that of...
15745 (make_pass_stack_regs_run): ...new function.
15746 (pass_df_finish): Replace declaration with that of...
15747 (make_pass_df_finish): ...new function.
15748 (pass_compute_alignments): Replace declaration with that of...
15749 (make_pass_compute_alignments): ...new function.
15750 (pass_duplicate_computed_gotos): Replace declaration with that of...
15751 (make_pass_duplicate_computed_gotos): ...new function.
15752 (pass_variable_tracking): Replace declaration with that of...
15753 (make_pass_variable_tracking): ...new function.
15754 (pass_free_cfg): Replace declaration with that of...
15755 (make_pass_free_cfg): ...new function.
15756 (pass_machine_reorg): Replace declaration with that of...
15757 (make_pass_machine_reorg): ...new function.
15758 (pass_cleanup_barriers): Replace declaration with that of...
15759 (make_pass_cleanup_barriers): ...new function.
15760 (pass_delay_slots): Replace declaration with that of...
15761 (make_pass_delay_slots): ...new function.
15762 (pass_split_for_shorten_branches): Replace declaration with that of...
15763 (make_pass_split_for_shorten_branches): ...new function.
15764 (pass_split_before_regstack): Replace declaration with that of...
15765 (make_pass_split_before_regstack): ...new function.
15766 (pass_convert_to_eh_region_ranges): Replace declaration with that
15768 (make_pass_convert_to_eh_region_ranges): ...new function.
15769 (pass_shorten_branches): Replace declaration with that of...
15770 (make_pass_shorten_branches): ...new function.
15771 (pass_set_nothrow_function_flags): Replace declaration with that of...
15772 (make_pass_set_nothrow_function_flags): ...new function.
15773 (pass_dwarf2_frame): Replace declaration with that of...
15774 (make_pass_dwarf2_frame): ...new function.
15775 (pass_final): Replace declaration with that of...
15776 (make_pass_final): ...new function.
15777 (pass_rtl_seqabstr): Replace declaration with that of...
15778 (make_pass_rtl_seqabstr): ...new function.
15779 (pass_release_ssa_names): Replace declaration with that of...
15780 (make_pass_release_ssa_names): ...new function.
15781 (pass_early_inline): Replace declaration with that of...
15782 (make_pass_early_inline): ...new function.
15783 (pass_inline_parameters): Replace declaration with that of...
15784 (make_pass_inline_parameters): ...new function.
15785 (pass_update_address_taken): Replace declaration with that of...
15786 (make_pass_update_address_taken): ...new function.
15787 (pass_convert_switch): Replace declaration with that of...
15788 (make_pass_convert_switch): ...new function.
15789 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
15790 to a subclass of simple_ipa_opt_pass along with...
15791 (pass_data_ipa_tree_profile): ...new pass_data instance and...
15792 (make_pass_ipa_tree_profile): ...new function.
15793 * tree-sra.c (pass_sra_early): Convert from a global struct to a
15794 subclass of gimple_opt_pass along with...
15795 (pass_data_sra_early): ...new pass_data instance and...
15796 (make_pass_sra_early): ...new function.
15797 (pass_sra): Convert from a global struct to a subclass of
15798 gimple_opt_pass along with...
15799 (pass_data_sra): ...new pass_data instance and...
15800 (make_pass_sra): ...new function.
15801 (pass_early_ipa_sra): Convert from a global struct to a subclass of
15802 gimple_opt_pass along with...
15803 (pass_data_early_ipa_sra): ...new pass_data instance and...
15804 (make_pass_early_ipa_sra): ...new function.
15805 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
15806 subclass of gimple_opt_pass along with...
15807 (pass_data_ccp): ...new pass_data instance and...
15808 (make_pass_ccp): ...new function.
15809 (pass_fold_builtins): Convert from a global struct to a subclass of
15810 gimple_opt_pass along with...
15811 (pass_data_fold_builtins): ...new pass_data instance and...
15812 (make_pass_fold_builtins): ...new function.
15813 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
15814 subclass of gimple_opt_pass along with...
15815 (pass_data_copy_prop): ...new pass_data instance and...
15816 (make_pass_copy_prop): ...new function.
15817 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
15818 global struct to a subclass of gimple_opt_pass along with...
15819 (pass_data_rename_ssa_copies): ...new pass_data instance and...
15820 (make_pass_rename_ssa_copies): ...new function.
15821 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
15822 subclass of gimple_opt_pass along with...
15823 (pass_data_dce): ...new pass_data instance and...
15824 (make_pass_dce): ...new function.
15825 (pass_dce_loop): Convert from a global struct to a subclass of
15826 gimple_opt_pass along with...
15827 (pass_data_dce_loop): ...new pass_data instance and...
15828 (make_pass_dce_loop): ...new function.
15829 (pass_cd_dce): Convert from a global struct to a subclass of
15830 gimple_opt_pass along with...
15831 (pass_data_cd_dce): ...new pass_data instance and...
15832 (make_pass_cd_dce): ...new function.
15833 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
15834 subclass of gimple_opt_pass along with...
15835 (pass_data_dominator): ...new pass_data instance and...
15836 (make_pass_dominator): ...new function.
15837 (pass_phi_only_cprop): Convert from a global struct to a subclass of
15838 gimple_opt_pass along with...
15839 (pass_data_phi_only_cprop): ...new pass_data instance and...
15840 (make_pass_phi_only_cprop): ...new function.
15841 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
15842 subclass of gimple_opt_pass along with...
15843 (pass_data_dse): ...new pass_data instance and...
15844 (make_pass_dse): ...new function.
15845 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
15846 a subclass of gimple_opt_pass along with...
15847 (pass_data_forwprop): ...new pass_data instance and...
15848 (make_pass_forwprop): ...new function.
15849 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
15850 struct to a subclass of gimple_opt_pass along with...
15851 (pass_data_tree_ifcombine): ...new pass_data instance and...
15852 (make_pass_tree_ifcombine): ...new function.
15853 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
15854 subclass of gimple_opt_pass along with...
15855 (pass_data_ch): ...new pass_data instance and...
15856 (make_pass_ch): ...new function.
15857 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
15858 subclass of gimple_opt_pass along with...
15859 (pass_data_tree_loop): ...new pass_data instance and...
15860 (make_pass_tree_loop): ...new function.
15861 (pass_tree_loop_init): Convert from a global struct to a subclass of
15862 gimple_opt_pass along with...
15863 (pass_data_tree_loop_init): ...new pass_data instance and...
15864 (make_pass_tree_loop_init): ...new function.
15865 (pass_lim): Convert from a global struct to a subclass of
15866 gimple_opt_pass along with...
15867 (pass_data_lim): ...new pass_data instance and...
15868 (make_pass_lim): ...new function.
15869 (pass_tree_unswitch): Convert from a global struct to a subclass of
15870 gimple_opt_pass along with...
15871 (pass_data_tree_unswitch): ...new pass_data instance and...
15872 (make_pass_tree_unswitch): ...new function.
15873 (pass_predcom): Convert from a global struct to a subclass of
15874 gimple_opt_pass along with...
15875 (pass_data_predcom): ...new pass_data instance and...
15876 (make_pass_predcom): ...new function.
15877 (pass_vectorize): Convert from a global struct to a subclass of
15878 gimple_opt_pass along with...
15879 (pass_data_vectorize): ...new pass_data instance and...
15880 (make_pass_vectorize): ...new function.
15881 (pass_graphite): Convert from a global struct to a subclass of
15882 gimple_opt_pass along with...
15883 (pass_data_graphite): ...new pass_data instance and...
15884 (make_pass_graphite): ...new function.
15885 (pass_graphite_transforms): Convert from a global struct to a subclass
15886 of gimple_opt_pass along with...
15887 (pass_data_graphite_transforms): ...new pass_data instance and...
15888 (make_pass_graphite_transforms): ...new function.
15889 (pass_check_data_deps): Convert from a global struct to a subclass of
15890 gimple_opt_pass along with...
15891 (pass_data_check_data_deps): ...new pass_data instance and...
15892 (make_pass_check_data_deps): ...new function.
15893 (pass_iv_canon): Convert from a global struct to a subclass of
15894 gimple_opt_pass along with...
15895 (pass_data_iv_canon): ...new pass_data instance and...
15896 (make_pass_iv_canon): ...new function.
15897 (pass_scev_cprop): Convert from a global struct to a subclass of
15898 gimple_opt_pass along with...
15899 (pass_data_scev_cprop): ...new pass_data instance and...
15900 (make_pass_scev_cprop): ...new function.
15901 (pass_record_bounds): Convert from a global struct to a subclass of
15902 gimple_opt_pass along with...
15903 (pass_data_record_bounds): ...new pass_data instance and...
15904 (make_pass_record_bounds): ...new function.
15905 (pass_complete_unroll): Convert from a global struct to a subclass of
15906 gimple_opt_pass along with...
15907 (pass_data_complete_unroll): ...new pass_data instance and...
15908 (make_pass_complete_unroll): ...new function.
15909 (pass_complete_unrolli): Convert from a global struct to a subclass of
15910 gimple_opt_pass along with...
15911 (pass_data_complete_unrolli): ...new pass_data instance and...
15912 (make_pass_complete_unrolli): ...new function.
15913 (pass_parallelize_loops): Convert from a global struct to a subclass
15914 of gimple_opt_pass along with...
15915 (pass_data_parallelize_loops): ...new pass_data instance and...
15916 (make_pass_parallelize_loops): ...new function.
15917 (pass_loop_prefetch): Convert from a global struct to a subclass of
15918 gimple_opt_pass along with...
15919 (pass_data_loop_prefetch): ...new pass_data instance and...
15920 (make_pass_loop_prefetch): ...new function.
15921 (pass_iv_optimize): Convert from a global struct to a subclass of
15922 gimple_opt_pass along with...
15923 (pass_data_iv_optimize): ...new pass_data instance and...
15924 (make_pass_iv_optimize): ...new function.
15925 (pass_tree_loop_done): Convert from a global struct to a subclass of
15926 gimple_opt_pass along with...
15927 (pass_data_tree_loop_done): ...new pass_data instance and...
15928 (make_pass_tree_loop_done): ...new function.
15929 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
15930 struct to a subclass of gimple_opt_pass along with...
15931 (pass_data_cse_reciprocals): ...new pass_data instance and...
15932 (make_pass_cse_reciprocals): ...new function.
15933 (pass_cse_sincos): Convert from a global struct to a subclass of
15934 gimple_opt_pass along with...
15935 (pass_data_cse_sincos): ...new pass_data instance and...
15936 (make_pass_cse_sincos): ...new function.
15937 (pass_optimize_bswap): Convert from a global struct to a subclass of
15938 gimple_opt_pass along with...
15939 (pass_data_optimize_bswap): ...new pass_data instance and...
15940 (make_pass_optimize_bswap): ...new function.
15941 (pass_optimize_widening_mul): Convert from a global struct to a
15942 subclass of gimple_opt_pass along with...
15943 (pass_data_optimize_widening_mul): ...new pass_data instance and...
15944 (make_pass_optimize_widening_mul): ...new function.
15945 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
15946 subclass of gimple_opt_pass along with...
15947 (pass_data_phiopt): ...new pass_data instance and...
15948 (make_pass_phiopt): ...new function.
15949 (pass_cselim): Convert from a global struct to a subclass of
15950 gimple_opt_pass along with...
15951 (pass_data_cselim): ...new pass_data instance and...
15952 (make_pass_cselim): ...new function.
15953 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
15954 subclass of gimple_opt_pass along with...
15955 (pass_data_phiprop): ...new pass_data instance and...
15956 (make_pass_phiprop): ...new function.
15957 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
15958 subclass of gimple_opt_pass along with...
15959 (pass_data_pre): ...new pass_data instance and...
15960 (make_pass_pre): ...new function.
15961 (pass_fre): Convert from a global struct to a subclass of
15962 gimple_opt_pass along with...
15963 (pass_data_fre): ...new pass_data instance and...
15964 (make_pass_fre): ...new function.
15965 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
15966 subclass of gimple_opt_pass along with...
15967 (pass_data_reassoc): ...new pass_data instance and...
15968 (make_pass_reassoc): ...new function.
15969 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
15970 subclass of gimple_opt_pass along with...
15971 (pass_data_sink_code): ...new pass_data instance and...
15972 (make_pass_sink_code): ...new function.
15973 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
15974 subclass of gimple_opt_pass along with...
15975 (pass_data_strlen): ...new pass_data instance and...
15976 (make_pass_strlen): ...new function.
15977 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
15978 struct to a subclass of gimple_opt_pass along with...
15979 (pass_data_build_alias): ...new pass_data instance and...
15980 (make_pass_build_alias): ...new function.
15981 (pass_build_ealias): Convert from a global struct to a subclass of
15982 gimple_opt_pass along with...
15983 (pass_data_build_ealias): ...new pass_data instance and...
15984 (make_pass_build_ealias): ...new function.
15985 (pass_ipa_pta): Convert from a global struct to a subclass of
15986 simple_ipa_opt_pass along with...
15987 (pass_data_ipa_pta): ...new pass_data instance and...
15988 (make_pass_ipa_pta): ...new function.
15989 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
15990 subclass of gimple_opt_pass along with...
15991 (pass_data_uncprop): ...new pass_data instance and...
15992 (make_pass_uncprop): ...new function.
15993 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
15994 global struct to a subclass of gimple_opt_pass along with...
15995 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
15996 (make_pass_late_warn_uninitialized): ...new function.
15997 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
15998 to a subclass of gimple_opt_pass along with...
15999 (pass_data_init_datastructures): ...new pass_data instance and...
16000 (make_pass_init_datastructures): ...new function.
16001 (pass_early_warn_uninitialized): Convert from a global struct to a
16002 subclass of gimple_opt_pass along with...
16003 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
16004 (make_pass_early_warn_uninitialized): ...new function.
16005 (pass_update_address_taken): Convert from a global struct to a
16006 subclass of gimple_opt_pass along with...
16007 (pass_data_update_address_taken): ...new pass_data instance and...
16008 (make_pass_update_address_taken): ...new function.
16009 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
16010 struct to a subclass of gimple_opt_pass along with...
16011 (pass_data_release_ssa_names): ...new pass_data instance and...
16012 (make_pass_release_ssa_names): ...new function.
16013 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
16014 subclass of gimple_opt_pass along with...
16015 (pass_data_stdarg): ...new pass_data instance and...
16016 (make_pass_stdarg): ...new function.
16017 * tree-switch-conversion.c (pass_convert_switch): Convert from a
16018 global struct to a subclass of gimple_opt_pass along with...
16019 (pass_data_convert_switch): ...new pass_data instance and...
16020 (make_pass_convert_switch): ...new function.
16021 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
16022 to a subclass of gimple_opt_pass along with...
16023 (pass_data_tail_recursion): ...new pass_data instance and...
16024 (make_pass_tail_recursion): ...new function.
16025 (pass_tail_calls): Convert from a global struct to a subclass of
16026 gimple_opt_pass along with...
16027 (pass_data_tail_calls): ...new pass_data instance and...
16028 (make_pass_tail_calls): ...new function.
16029 * tree-vect-generic.c (pass_lower_vector): Convert from a global
16030 struct to a subclass of gimple_opt_pass along with...
16031 (pass_data_lower_vector): ...new pass_data instance and...
16032 (make_pass_lower_vector): ...new function.
16033 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
16034 gimple_opt_pass along with...
16035 (pass_data_lower_vector_ssa): ...new pass_data instance and...
16036 (make_pass_lower_vector_ssa): ...new function.
16037 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
16038 to a subclass of gimple_opt_pass along with...
16039 (pass_data_slp_vectorize): ...new pass_data instance and...
16040 (make_pass_slp_vectorize): ...new function.
16041 (pass_ipa_increase_alignment): Convert from a global struct to a
16042 subclass of simple_ipa_opt_pass along with...
16043 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
16044 (make_pass_ipa_increase_alignment): ...new function.
16045 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
16046 gimple_opt_pass along with...
16047 (pass_data_vrp): ...new pass_data instance and...
16048 (make_pass_vrp): ...new function.
16049 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
16050 subclass of simple_ipa_opt_pass along with...
16051 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
16052 (make_pass_ipa_free_lang_data): ...new function.
16053 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
16054 gimple_opt_pass along with...
16055 (pass_data_tsan): ...new pass_data instance and...
16056 (make_pass_tsan): ...new function.
16057 (pass_tsan_O0): Convert from a global struct to a subclass of
16058 gimple_opt_pass along with...
16059 (pass_data_tsan_O0): ...new pass_data instance and...
16060 (make_pass_tsan_O0): ...new function.
16061 * var-tracking.c (pass_variable_tracking): Convert from a global
16062 struct to a subclass of rtl_opt_pass along with...
16063 (pass_data_variable_tracking): ...new pass_data instance and...
16064 (make_pass_variable_tracking): ...new function.
16065 * web.c (pass_web): Convert from a global struct to a subclass of
16066 rtl_opt_pass along with...
16067 (pass_data_web): ...new pass_data instance and...
16068 (make_pass_web): ...new function.
16069 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
16070 declaration with that of...
16071 (make_pass_mode_switch_use): ...new function.
16072 (pass_resolve_sw_modes): Replace declaration with that of...
16073 (make_pass_resolve_sw_modes): ...new function.
16074 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
16075 from a global struct to a subclass of rtl_opt_pass along with...
16076 (pass_data_mode_switch_use): ...new pass_data instance and...
16077 (make_pass_mode_switch_use): ...new function.
16078 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
16079 from a global struct to a subclass of rtl_opt_pass along with...
16080 (pass_data_resolve_sw_modes): ...new pass_data instance and...
16081 (make_pass_resolve_sw_modes): ...new function.
16082 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
16083 struct to a subclass of rtl_opt_pass along with...
16084 (pass_data_insert_vzeroupper): ...new pass_data instance and...
16085 (make_pass_insert_vzeroupper): ...new function.
16086 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
16087 global struct to a subclass of rtl_opt_pass along with...
16088 (pass_data_work_around_errata): ...new pass_data instance and...
16089 (make_pass_work_around_errata): ...new function.
16090 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
16091 struct to a subclass of rtl_opt_pass along with...
16092 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
16093 (make_pass_mips_machine_reorg2): ...new function.
16095 2013-08-05 David Malcolm <dmalcolm@redhat.com>
16097 * passes.c (pass_manager::operator new): New.
16099 2013-08-05 David Malcolm <dmalcolm@redhat.com>
16101 Handwritten part of conversion of passes to C++ classes.
16103 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
16104 (toplev.o): Add dep on PASS_MANAGER_H.
16105 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
16106 of early local pases to reflect this moving from a global to a
16107 member of gcc::pass_manager.
16108 (cgraph_add_new_function): Likewise.
16109 * lto-cgraph.c (lto_output_node): Update for conversion of
16110 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
16111 * passes.c (opt_pass::clone): New.
16112 (opt_pass::gate): New.
16113 (opt_pass::execute): New.
16114 (opt_pass::opt_pass): New.
16115 (pass_manager::execute_early_local_passes): New.
16116 (pass_manager::execute_pass_mode_switching): new.
16117 (finish_optimization_passes): Convert to...
16118 (pass_manager::finish_optimization_passes): ...this.
16119 (finish_optimization_passes): Update for conversion of passes to
16121 (register_dump_files_1): Use has_gate since we cannot portably
16122 check a vtable entry against NULL.
16123 (dump_one_pass): Likewise.
16124 (ipa_write_summaries_2): Likewise.
16125 (ipa_write_optimization_summaries_1): Likewise.
16126 (ipa_read_summaries_1): Likewise.
16127 (ipa_read_optimization_summaries_1): Likewise.
16128 (execute_ipa_stmt_fixups): Likewise.
16129 (pass_manager::pass_manager): Rewrite pass-creation, invoking
16130 pass-creation functions rather than wiring up globals, and
16131 storing the results in fields of pass_manager generated using
16132 pass-instances.def.
16133 (pass_manager::dump_profile_report): Update for conversion of
16134 passes to C++ classes.
16135 (pass_manager::execute_ipa_summary_passes): Likewise.
16136 (execute_one_ipa_transform_pass): Likewise.
16137 (execute_one_pass): Use has_gate and has_execute since we cannot
16138 portably check a vtable entry against NULL.
16139 * pass_manager.h (pass_manager::finish_optimization_passes): New.
16140 (pass_manager): Use pass-instances.def to add fields for the
16141 various pass instances.
16142 * toplev.c (finalize): Update for move of
16143 finish_optimization_passes to a method of gcc::pass_manager.
16144 * toplev.h (finish_optimization_passes): Move to method of class
16146 * tree-pass.h (struct pass_data): New.
16147 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
16148 (opt_pass::gate): Convert to virtual function.
16149 (opt_pass::~opt_pass): New.
16150 (opt_pass::clone): New.
16151 (opt_pass::execute): Convert to virtual function.
16152 (opt_pass::opt_pass): New.
16153 (opt_pass::ctxt_): new.
16154 (gimple_opt_pass): Convert to subclass of opt_pass.
16155 (gimple_opt_pass::gimple_opt_pass): New.
16156 (rtl_opt_pass): Convert to subclass of opt_pass.
16157 (rtl_opt_pass::rtl_opt_pass): New.
16158 (ipa_opt_pass_d): Convert to subclass of opt_pass.
16159 (ipa_opt_pass_d::ipa_opt_pass_d): New.
16160 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
16161 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
16162 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
16163 invocation of pass_mode_switching to reflect this moving from a
16164 global to a member of gcc::pass_manager.
16165 (ix86_option_override): Rework how pass_insert_vzeroupper is
16166 added to the pass_manager to reflect autogenerated changes.
16167 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
16169 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
16171 PR rtl-optimization/57708
16172 * recog.c (peep2_find_free_register): Validate all regs in a
16175 2013-08-05 Jan Hubicka <jh@suse.cz>
16178 * cgraph.c (verify_cgraph_node): Accept local flags from other
16180 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
16181 (function_and_variable_visibility): Likewise.
16182 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
16184 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
16186 * graph.c (init_graph_slim_pretty_print): Remove.
16187 (print_graph_cfg): Do not call it. Use local pretty printer.
16188 (start_graph_dump): Likewise.
16190 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
16192 * gimple-pretty-print.c (buffer): Remove.
16193 (initialized): Likewise.
16194 (maybe_init_pretty_print): Likewise.
16195 (print_gimple_stmt): Do not call it. Use non-static local
16196 pretty_printer variable.
16197 (print_gimple_expr): Likewise.
16198 (print_gimple_seq): Likewise.
16199 (gimple_dump_bb): Likewise.
16201 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
16203 * asan.c (asan_pp): Remove.
16204 (asan_pp_initialized): Likewise.
16205 (asan_pp_initialize): Likewise.
16206 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
16207 (asan_emit_stack_protection): Tidy. Use local pretty printer.
16208 (asan_add_global): Likewise.
16210 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
16212 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
16213 * diagnostic.h (diagnostic_flush_buffer): Adjust.
16214 * pretty-print.c (pp_formatted_text_data): Likewise.
16215 (pp_indent): Rename from pp_base_indent.
16216 (pp_format): Rename from pp_base_format.
16217 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
16218 (pp_format_verbatim): Rename from pp_base_format_verbatim.
16219 (pp_flush): Rename from pp_base_flush.
16220 (pp_set_line_maximum_length): Rename from
16221 pp_base_set_line_maximum_length.
16222 (pp_clear_output_area): Rename from pp_base_clear_output_area.
16223 (pp_set_prefix): Rename from pp_base_set_prefix.
16224 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
16225 (pp_emit_prefix): Rename from pp_base_emit_prefix.
16226 (pp_append_text): Rename from pp_base_append_text.
16227 (pp_formatted_text): Rename from pp_base_formatted_text.
16228 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
16229 (pp_remaining_character_count_for_line): Rename from
16230 pp_base_remaining_character_count_for_line.
16231 (pp_newline): Rename from pp_base_newline.
16232 (pp_character): Rename from pp_base_character.
16233 (pp_string): Rename from pp_base_string.
16234 (pp_maybe_space): Rename from pp_base_maybe_space.
16235 * asan.c (asan_pp_string): Adjust.
16236 (asan_emit_stack_protection): Likewise.
16237 (asan_add_global): Likewise.
16238 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
16239 * tree-mudflap.c (mf_varname_tree): Likewise.
16240 * tree-pretty-print.c (pp_tree_identifier): Rename from
16241 pp_base_tree_identifier.
16242 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
16243 Declare as function.
16245 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
16247 * pretty-print.h (pp_bar_bar): New.
16248 (pp_ampersand_ampersand): Likewise.
16249 (pp_less_equal): Likewise.
16250 (pp_greater_equal): Likewise.
16251 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
16252 printer functions instead of pp_string or operators and punctuators.
16253 (dump_gimple_call): Likewise.
16254 (dump_gimple_omp_for): Likewise.
16255 (dump_gimple_transaction): Likewise.
16256 (dump_gimple_phi): Likewise.
16257 (pp_gimple_stmt_1): Likewise.
16258 * sched-vis.c (print_insn): Likewise.
16259 * tree-mudflap.c (mf_varname_tree): Likewise.
16260 * tree-pretty-print.c (dump_block_node): Likewise.
16261 (dump_generic_node): Likewise.
16263 2013-08-02 Jan Hubicka <jh@suse.cz>
16265 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
16267 * lto-streamer-out.c (tree_is_indexable): Results decls and
16268 parm decls are not indexable.
16269 (DFS_write_tree_body): Do not follow args and results.
16270 (hash_tree): Likewise.
16271 (output_functions): Rearrange so struct function is needed
16272 only when real body is output; be able to also ouptut abstract
16273 functions; output DECL_ARGUMENTS and DECL_RESULT.
16274 (lto_output): When not in WPA, ale store abstract functions.
16275 (write_symbol): Do not care about RESULT_DECL.
16276 (output_symbol_p): Handle correctly sbtract decls.
16277 * lto-streamer-in.c (input_function): Rearrange so struct
16278 function can be NULL at entry; allow streaming of
16279 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
16280 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
16281 sanity check during LTO.
16282 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
16283 RESULT_DECl and DECL_ARGUMENTS.
16284 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
16287 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
16289 * pretty-print.h (pp_underscore): New.
16291 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
16292 printer functions instead of pp_character.
16293 (dump_binary_rhs): Likewise.
16294 (dump_ternary_rhs): Likewise.
16295 (dump_gimple_call_args): Likewise.
16296 (pp_points_to_solution): Likewise.
16297 (dump_gimple_call): Likewise.
16298 (dump_gimple_switch): Likewise.
16299 (dump_gimple_cond): Likewise.
16300 (dump_gimple_bind): Likewise.
16301 (dump_gimple_try): Likewise.
16302 (dump_gimple_omp_for): Likewise.
16303 (dump_gimple_omp_continue): Likewise.
16304 (dump_gimple_omp_single): Likewise.
16305 (dump_gimple_omp_sections): Likewise.
16306 (dump_gimple_omp_block): Likewise.
16307 (dump_gimple_omp_critical): Likewise.
16308 (dump_gimple_transaction): Likewise.
16309 (dump_gimple_asm): Likewise.
16310 (dump_gimple_phi): Likewise.
16311 (dump_gimple_omp_parallel): Likewise.
16312 (dump_gimple_omp_task): Likewise.
16313 (dump_gimple_omp_atomic_load): Likewise.
16314 (dump_gimple_omp_atomic_store): Likewise.
16315 (dump_gimple_mem_ops): Likewise.
16316 (pp_gimple_stmt_1): Likewise.
16317 (pp_cfg_jump): Likewise.
16318 (dump_implicit_edges): Likewise.
16319 (gimple_dump_bb_for_graph): Likewise.
16320 * graph.c (draw_cfg_node): Likewise.
16321 * langhooks.c (lhd_print_error_function): Likewise.
16322 * sched-vis.c (print_exp): Likewise.
16323 (print_value): Likewise.
16324 (print_pattern): Likewise.
16325 (print_insn): Likewise.
16326 (rtl_dump_bb_for_graph): Likewise.
16327 * tree-pretty-print.c (dump_function_declaration): Likewise.
16328 (dump_array_domain): Likewise.
16329 (dump_omp_clause): Likewise.
16330 (dump_location): Likewise.
16331 (dump_generic_node): Likewise.
16332 (print_struct_decl): Likewise.
16333 * diagnostic.c (diagnostic_show_locus): Use pp_space.
16335 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
16337 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
16338 candidate table when replacing a candidate statement.
16339 (replace_rhs_if_not_dup): Likewise.
16340 (replace_one_candidate): Likewise.
16342 2013-08-02 Jan Hubicka <jh@suse.cz>
16343 Martin Liska <marxin.liska@gmail.com>
16345 * cgraphunit.c (add_new_function): Fix logic when adding from
16347 (assemble_thunk): Rename to ...
16348 (expand_thunk); .. this one; export; get it working with
16349 general functions; make produced gimple valid.
16350 * cgraph.h (expand_thunk): Declare.
16352 2013-08-02 Jan Hubicka <jh@suse.cz>
16354 * ipa-cp.c (gather_context_independent_values): Use
16355 ipa_get_param_move_cost.
16356 (get_replacement_map): Remove PARAM; move parameter folding
16358 (create_specialized_node): Update.
16359 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
16360 assert that we have gimple body; update move_cost.
16361 (count_formal_params): Assert that we have gimple body.
16362 (ipa_dump_param): New function.
16363 (ipa_alloc_node_params): Break out from ...
16364 (ipa_initialize_node_params): ... here.
16365 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
16366 (ipa_write_node_info): Stream move costs.
16367 (ipa_read_node_info): Read move costs.
16368 (ipa_update_after_lto_read): Do not recompute node params.
16369 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
16370 (ipa_get_param): Check we are not in WPA.
16371 (ipa_get_param_move_cost): New.
16372 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
16373 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
16374 parm numbers to be present.
16376 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
16378 PR rtl-optimization/58048
16379 * lra-constraints.c (process_alt_operands): Don't check asm
16380 operand on register.
16382 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
16384 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
16385 the implied StoreLoad barrier for atomic operations if before.
16387 2013-08-02 Jan Hubicka <jh@suse.cz>
16388 Martin Liska <marxin.liska@gmail.com>
16390 * cgraph.c (cgraph_function_body_availability): Do not check
16392 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
16393 symtab_node_availability): Declare.
16394 * ipa.c (can_replace_by_local_alias): New.
16395 (function_and_variable_visibility): Use it.
16396 * symtab.c (symtab_for_node_and_aliases,
16397 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
16399 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
16401 PR rtl-optimization/57963
16402 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
16403 (lra_constraints): Use them.
16405 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
16407 * config/arm/types.md (define_attr "type"): Add "load_acq"
16409 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
16411 (cortex_a53_store1): Likewise.
16413 2013-08-01 Jan Hubicka <jh@suse.cz>
16415 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
16416 partitions are not needed.
16418 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
16420 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
16421 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
16422 MAYBE_NON_Q_CLASS_P where appropriate.
16424 2013-08-01 Jan Hubicka <jh@suse.cz>
16426 * cgraph.h (release_function_body): Declare.
16427 * tree.c (free_lang_data_in_decl): Free, parameters and return values
16428 of unused delcarations.
16430 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16432 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
16433 RTL form when subtracting a constant.
16435 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16437 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
16438 Generate canonical plus rtx with negated immediate instead of minus
16440 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
16442 2013-08-01 Jan Hubicka <jh@suse.cz>
16444 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
16445 (cgraph_release_function_body): Likewise.
16446 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
16447 * cgraph.h (cgrpah_node): Rename abstract_and_needed
16448 to used_as_abstract_origin.
16449 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
16450 symbols used as abstract origins.
16451 * cgraphunit.c (analyze_functions): Update.
16452 * ipa.c (symtab_remove_unreachable_nodes): Recompute
16453 used_as_abstract_origin.
16454 * tree-inline.c (tree_function_versioning): Update
16455 used_as_abstract_origin; be ready for DECL_RESULT and
16456 DECL_ARGUMENTS to be NULL.
16458 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
16459 for abstract functions.
16460 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
16463 2013-08-01 Jan Hubicka <jh@suse.cz>
16465 * profile.c (compute_value_histograms): Fix thinko.
16467 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
16469 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
16470 aarch-common-protos.h to extra_headers.
16471 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
16472 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
16473 * config/aarch64/t-aarch64 (aarch-common.o): Define.
16475 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
16477 * config/aarch64/aarch64.md (define_attr "type"): Delete.
16478 Include "../arm/types.md". Define "type" attribute for all patterns.
16479 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
16482 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
16484 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
16485 to support power8 load fusion.
16486 (fusion_gpr_mem_load): Likewise.
16488 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
16490 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
16491 declarations for power8 load fusion.
16492 (emit_fusion_gpr_load): Likewise.
16494 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
16495 tuning for power8, turn on fusion mode by default. Turn on sign
16496 extending fusion mode if normal fusion mode is on, and we are at
16498 (fusion_gpr_load_p): New function, return true if we can fuse an
16499 addis instruction with a dependent load to a GPR.
16500 (emit_fusion_gpr_load): Emit the instructions for power8 load
16503 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
16504 (VSX load fusion peepholes): New peepholes to fuse together an
16505 addi instruction with a VSX load instruction.
16507 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
16508 peepholes to fuse an addis instruction with a load to a GPR base
16509 register. If we are supporting sign extending fusions, convert
16510 sign extending loads to zero extending loads and add an explicit
16513 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
16515 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
16516 aarch-common-protos.h to extra_headers.
16517 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
16518 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
16519 (arm_early_store_addr_dep): Likewise.
16520 (arm_no_early_alu_shift_dep): Likewise.
16521 (arm_no_early_alu_shift_value_dep): Likewise.
16522 (arm_no_early_mul_dep): Likewise.
16523 (arm_no_early_store_addr_dep): Likewise.
16524 (arm_mac_accumulator_is_mul_result): Likewise.
16525 (arm_mac_accumulator_is_result): Likewise.
16526 * config/arm/aarch-common.c: ... here. New file.
16527 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
16529 (arm_early_store_addr_dep): Likewise.
16530 (arm_no_early_alu_shift_dep): Likewise.
16531 (arm_no_early_alu_shift_value_dep): Likewise.
16532 (arm_no_early_mul_dep): Likewise.
16533 (arm_no_early_store_addr_dep): Likewise.
16534 (arm_mac_accumulator_is_mul_result): Likewise.
16535 (arm_mac_accumulator_is_result): Likewise.
16536 * config/arm/aarch-common-protos.h: ... here. New file.
16537 * config/arm/t-arm (aarch-common.o): Define.
16539 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
16541 * config/arm/arm.md: Include new file "types.md".
16542 (define_attr "type"): Move from here to ...
16543 (define_attr "mul32"): Likewise.
16544 (define_attr "mul64"): Likewise.
16545 * config/arm/types.md: ... here. New file.
16547 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
16549 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
16550 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
16552 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
16554 * gen-pass-instances.awk: Fix offset of substr().
16556 2013-07-31 David Malcolm <dmalcolm@redhat.com>
16558 * Makefile.in (pass-instances.def): New.
16559 (passes.o): Replace dependency on passes.def with one on
16562 * gen-pass-instances.awk: New.
16564 * passes.c (pass_manager::pass_manager): Use pass-instances.def
16565 rather than passes.def, updating local definition of NEXT_PASS
16566 macro to add an extra NUM parameter (currently unused).
16568 2013-07-30 David Malcolm <dmalcolm@redhat.com>
16570 * Makefile.in (PASS_MANAGER_H): New.
16571 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
16572 (passes.o): Likewise.
16573 (statistics.o): Likewise.
16574 (cgraphunit.o): Likewise.
16575 (context.o): Depend on PASS_MANAGER_H.
16577 * pass_manager.h: New.
16579 * cgraphunit.c (cgraph_add_new_function): Update for moves
16580 of globals to fields of pass_manager.
16581 (analyze_function): Likewise.
16582 (expand_function): Likewise.
16583 (ipa_passes): Likewise.
16584 (compile): Likewise.
16586 * context.c (context::context): New.
16587 * context.h (context::context): New.
16588 (context::get_passes): New.
16589 (context::passes_): New.
16591 * lto-cgraph.c (input_node): Update for moves of globals to
16592 fields of pass_manager.
16594 * passes.c (all_passes): Remove, in favor of a field of the
16595 same name within the new class pass_manager.
16596 (all_small_ipa_passes): Likewise.
16597 (all_lowering_passes): Likewise.
16598 (all_regular_ipa_passes): Likewise.
16599 (all_late_ipa_passes): Likewise.
16600 (all_lto_gen_passes): Likewise.
16601 (passes_by_id): Likewise.
16602 (passes_by_id_size): Likewise.
16603 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
16604 the new class pass_manager.
16605 (set_pass_for_id): Convert to...
16606 (pass_manager::set_pass_for_id): ...method.
16607 (get_pass_for_id): Convert to...
16608 (pass_manager::get_pass_for_id): ...method.
16609 (register_one_dump_file): Move body of implementation into...
16610 (pass_manager::register_one_dump_file): ...here.
16611 (register_dump_files_1): Convert to...
16612 (pass_manager::register_dump_files_1): ...method.
16613 (register_dump_files): Convert to...
16614 (pass_manager::register_dump_files): ...method.
16615 (create_pass_tab): Update for moves of globals to fields of
16617 (dump_passes): Move body of implementation into...
16618 (pass_manager::dump_passes): ...here.
16619 (register_pass): Move body of implementation into...
16620 (pass_manager::register_pass): ...here.
16621 (init_optimization_passes): Convert into...
16622 (pass_manager::pass_manager): ...constructor for new
16623 pass_manager class, and initialize the pass_lists array.
16624 (check_profile_consistency): Update for moves of globals to
16625 fields of pass_manager.
16626 (dump_profile_report): Move body of implementation into...
16627 (pass_manager::dump_profile_report): ...here.
16628 (ipa_write_summaries_1): Update for moves of pass lists from
16629 being globals to fields of pass_manager.
16630 (ipa_write_optimization_summaries): Likewise.
16631 (ipa_read_summaries): Likewise.
16632 (ipa_read_optimization_summaries): Likewise.
16633 (execute_all_ipa_stmt_fixups): Likewise.
16635 * statistics.c (statistics_fini): Update for moves of globals to
16636 fields of pass_manager.
16638 * toplev.c (general_init): Replace call to
16639 init_optimization_passes with construction of the pass_manager
16642 * tree-pass.h (all_passes): Remove, in favor of a field of the
16643 same name within the new class pass_manager.
16644 (all_small_ipa_passes): Likewise.
16645 (all_lowering_passes): Likewise.
16646 (all_regular_ipa_passes): Likewise.
16647 (all_lto_gen_passes): Likewise.
16648 (all_late_ipa_passes): Likewise.
16649 (passes_by_id): Likewise.
16650 (passes_by_id_size): Likewise.
16651 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
16652 the new class pass_manager.
16653 (get_pass_for_id): Remove.
16655 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
16657 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
16660 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
16662 * arm.md (mulhi3): New expand pattern.
16664 2013-07-30 Jan Hubicka <jh@suse.cz>
16665 Martin Liska <marxin.liska@gmail.com>
16667 * profile.c (compute_value_histograms): Do not ICE when
16668 there is mismatch only on some counters.
16670 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
16672 PR rtl-optimization/57637
16673 * function.c (move_insn_for_shrink_wrap): Also check the
16674 GEN set of the LIVE problem for the liveness analysis
16675 if it exists, otherwise give up.
16677 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
16679 PR tree-optimization/57993
16680 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
16681 replaced statement in the candidate table.
16682 (phi_add_costs): Return infinite cost when the hidden basis does
16683 not dominate all phis on which the candidate is dependent.
16684 (replace_one_candidate): Record replaced statement in the
16687 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
16689 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
16690 (ashlv2si3): New expander.
16691 (*ashlv2si3_i): New define_insn_and_split.
16692 * predicates.md (float_operation): Allow patterns with three
16693 basic sub-patterns.
16695 PR rtl-optimization/58021
16696 * mode-switching.c (create_pre_exit): Always split off preceding
16697 insns if we are not at the basic block head.
16699 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
16701 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
16702 (UCLIBC_DYNAMIC_LINKER): New macro.
16703 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
16705 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
16706 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
16708 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
16709 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
16710 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
16711 for SF and DF modes. Use ieee_quad_format for TF mode.
16712 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
16713 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
16714 (mips_option_override): Handle `-mnan=legacy'.
16715 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
16716 `-mabs=2008' and `-mnan=2008'.
16717 (OPTION_DEFAULT_SPECS): Add "nan" default.
16718 (ASM_SPEC): Handle `-mnan='.
16719 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
16720 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
16721 comment accordingly.
16722 (neg<mode>2): Likewise.
16723 * config/mips/mips.opt (mabs, mnan): New options.
16724 * doc/install.texi (Configuration): Document `--with-nan=' option.
16725 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
16727 (MIPS Options): Document them.
16728 * config.gcc <mips*-*-*>: Handle `--with-nan='.
16729 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
16730 * configure: Regenerate.
16731 * config.in: Regenerate.
16733 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
16735 * config/i386/i386.md (float post-reload splitters): Do not check
16736 for subregs of SSE registers.
16738 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
16739 H.J. Lu <hongjiu.lu@intel.com>
16743 * config/i386/i386.md (post-reload splitter
16744 to avoid partial SSE reg dependency stalls): New pattern.
16746 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
16748 * config/s390/s390.md ("movcc"): Swap load and store instructions.
16750 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
16752 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
16753 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
16755 2013-07-26 Cary Coutant <ccoutant@google.com>
16757 * dwarf2out.c (die_checksum_ordered): Don't include template
16758 instantiations in signature.
16759 (is_template_parameter): New function.
16760 (is_template_instantiation): New function.
16761 (generate_skeleton_bottom_up): Don't include template instantiations
16763 (generate_skeleton): Likewise.
16764 (break_out_comdat_types): Move recursive call to break out nested
16766 (prune_unused_types_mark_generic_parms_dies): Call
16767 is_template_parameter.
16769 2013-07-26 Ian Bolton <ian.bolton@arm.com>
16771 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
16772 uses vector registers.
16773 * config/aarch64/iterators.md: Add attributes rtn and vas.
16775 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16776 Richard Earnshaw <richard.earnshaw@arm.com>
16778 * combine.c (simplify_comparison): Re-canonicalize operands
16780 * config/arm/arm.md (movcond_addsi): New splitter.
16782 2013-07-25 Sterling Augustine <saugustine@google.com>
16784 * dwarf2out.c (size_of_pubnames): Move code to...
16785 (include_pubname_in_output): ...here. New.
16786 (want_pubnames): Rearrange.
16787 (output_pubnames): Call include_pubname_in_output. Move assertion.
16789 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
16791 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
16793 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
16796 * doc/extend.texi: Remove obsolete builtins. Fix
16797 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
16799 2013-07-25 Jan Hubicka <jh@suse.cz>
16801 * cgraph.c (release_function_body): Break out from ...
16802 (cgraph_release_function_body): ... this one; also release DECL_RESULT
16803 and DECL_ARGUMENTS.
16804 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
16805 old_tree in the map.
16806 (create_specialized_node): Update.
16807 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
16809 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
16810 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
16811 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
16812 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
16815 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16817 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
16818 addsi3_carryin_alt2_<optab>): Correct output template.
16820 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16822 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
16823 Adjust for arm_restrict_it.
16824 Remove trailing whitespace.
16826 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
16828 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
16829 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
16831 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
16833 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
16835 PR rtl-optimization/57960
16836 * lra-constraints.c (process_alt_operands): Use the right mode
16837 when checking strict_low.
16839 2013-07-25 Jan Hubicka <jh@suse.cz>
16841 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
16842 * cgraph.c (cgraph_remove_node): Do not release function body
16843 when in cgraph streaming.
16844 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
16845 in other partitions are not considered reachable; fix handling of
16848 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16850 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
16852 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16857 * config/arm/arm.md ("*sibcall_insn): Replace use of
16858 Ss with US. Adjust output for v5 and v4t.
16859 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
16861 * config/arm/constraints.md ("Ss"): Rename to US.
16863 2013-07-25 Terry Guo <terry.guo@arm.com>
16865 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
16866 shift_add/shift_sub0/shift_sub1 RTXs.
16868 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
16869 Anton Blanchard <anton@au1.ibm.com>
16871 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
16872 (altivec_vpks<VI_char>ss): Likewise.
16873 (altivec_vpks<VI_char>us): Likewise.
16874 (altivec_vpku<VI_char>us): Likewise.
16875 (altivec_vpku<VI_char>um): Likewise.
16877 2013-07-24 David Malcolm <dmalcolm@redhat.com>
16879 Introduce context class.
16881 * Makefile.in (CONTEXT_H): New.
16882 (OBJS): Add context.o.
16883 (toplev.o): Add CONTEXT_H to dependencies.
16886 * toplev.c (general_init): Create the singleton gcc::context instance.
16892 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
16894 PR rtl-optimization/57968
16895 * mode-switching.c (create_pre_exit): Allow instructions that
16896 don't set a return register to need a non-exit mode.
16898 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
16899 Anton Blanchard <anton@au1.ibm.com>
16901 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
16902 operands to vperm for little endian.
16903 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
16904 of lvsl to create the control mask for a vperm for little endian.
16906 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16907 Anton Blanchard <anton@au1.ibm.com>
16909 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
16910 two operands for little-endian.
16912 2013-07-23 Steve Ellcey <sellcey@mips.com>
16914 * config/mips/mips.c (mips_case_values_threshold): New.
16915 (TARGET_CASE_VALUES_THRESHOLD): Define.
16917 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16918 Anton Blanchard <anton@au1.ibm.com>
16920 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
16921 selection of field for vector splat in little endian mode.
16923 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
16925 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
16926 expanders to rs6000.md.
16927 (ior<mode>3): Likewise.
16928 (and<mode>3): Likewise.
16929 (one_cmpl<mode>2): Likewise.
16930 (nor<mode>3): Likewise.
16931 (andc<mode>3): Likewise.
16932 (eqv<mode>3): Likewise.
16933 (nand<mode>3): Likewise.
16934 (orc<mode>3): Likewise.
16936 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
16939 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
16940 to split multi-word logical operations.
16941 (rs6000_split_logical_di): Likewise.
16942 (rs6000_split_logical): Likewise.
16944 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
16945 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
16946 and allow TImode operations in 32-bit.
16947 (vsx_and<mode>3_64bit): Likewise.
16948 (vsx_ior<mode>3_32bit): Likewise.
16949 (vsx_ior<mode>3_64bit): Likewise.
16950 (vsx_xor<mode>3_32bit): Likewise.
16951 (vsx_xor<mode>3_64bit): Likewise.
16952 (vsx_one_cmpl<mode>2_32bit): Likewise.
16953 (vsx_one_cmpl<mode>2_64bit): Likewise.
16954 (vsx_nor<mode>3_32bit): Likewise.
16955 (vsx_nor<mode>3_64bit): Likewise.
16956 (vsx_andc<mode>3_32bit): Likewise.
16957 (vsx_andc<mode>3_64bit): Likewise.
16958 (vsx_eqv<mode>3_32bit): Likewise.
16959 (vsx_eqv<mode>3_64bit): Likewise.
16960 (vsx_nand<mode>3_32bit): Likewise.
16961 (vsx_nand<mode>3_64bit): Likewise.
16962 (vsx_orc<mode>3_32bit): Likewise.
16963 (vsx_orc<mode>3_64bit): Likewise.
16965 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
16966 logical types in GPRs.
16968 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
16969 logical insns to rs6000.md, and allow TImode operations in
16971 (altivec_ior<mode>3): Likewise.
16972 (altivec_xor<mode>3): Likewise.
16973 (altivec_one_cmpl<mode>2): Likewise.
16974 (altivec_nor<mode>3): Likewise.
16975 (altivec_andc<mode>3): Likewise.
16977 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
16978 attributes for moving the 128-bit logical operations into
16980 (BOOL_REGS_OUTPUT): Likewise.
16981 (BOOL_REGS_OP1): Likewise.
16982 (BOOL_REGS_OP2): Likewise.
16983 (BOOL_REGS_UNARY): Likewise.
16984 (BOOL_REGS_AND_CR0): Likewise.
16985 (one_cmpl<mode>2): Add support for DI logical operations on
16986 32-bit, splitting the operations to 32-bit.
16987 (anddi3): Likewise.
16988 (iordi3): Likewise.
16989 (xordi3): Likewise.
16990 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
16991 changes to combine the 32/64-bit code, allow logical operations on
16992 TI mode in 32-bit, and to use similar match_operator patterns like
16993 scalar mode uses. Combine the Altivec and VSX code for logical
16994 operations, and move it here.
16995 (ior<mode>3, 128-bit types): Likewise.
16996 (xor<mode>3, 128-bit types): Likewise.
16997 (one_cmpl<mode>3, 128-bit types): Likewise.
16998 (nor<mode>3, 128-bit types): Likewise.
16999 (andc<mode>3, 128-bit types): Likewise.
17000 (eqv<mode>3, 128-bit types): Likewise.
17001 (nand<mode>3, 128-bit types): Likewise.
17002 (orc<mode>3, 128-bit types): Likewise.
17003 (and<mode>3_internal): Likewise.
17004 (bool<mode>3_internal): Likewise.
17005 (boolc<mode>3_internal1): Likewise.
17006 (boolc<mode>3_internal2): Likewise.
17007 (boolcc<mode>3_internal1): Likewise.
17008 (boolcc<mode>3_internal2): Likewise.
17009 (eqv<mode>3_internal1): Likewise.
17010 (eqv<mode>3_internal2): Likewise.
17011 (one_cmpl1<mode>3_internal): Likewise.
17013 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
17015 * config/microblaze/microblaze.c (microblaze_expand_prologue):
17016 Rename flag_stack_usage to flag_stack_usage_info.
17018 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
17020 * config/microblaze/sync.md: New file.
17021 * config/microblaze/microblaze.md: Include sync.md
17022 * config/microblaze/microblaze.c: Add print_operand 'y'.
17023 * config/microblaze/constraints.md: Add memory_contraint
17024 'Q' which is a single register.
17026 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
17028 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
17030 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
17032 * reload.c (find_reloads): Exit loop once we find this operand
17033 cannot be reloaded somehow for this alternative.
17035 * reload.c (find_reloads): Exit loop once we find a hard register.
17037 * rtlanal.c (computed_jump_p): Exit loop once we find label
17040 * i386.c (ix86_pad_returns): Exit loop after setting replace.
17042 * cfgloopmanip.c (remove_path): Exit loop after setting
17045 * gensupport.c (subst_dup): Avoid loop if code is not
17046 MATCH_DUP nor MATCH_OP_DUP.
17048 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
17050 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
17052 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
17054 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
17055 true for SP_REGNUM if mode == ptr_mode.
17056 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
17057 with value R0_REGNUM + 31.
17059 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
17061 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
17062 pad pointer-typed argument downward.
17064 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
17066 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
17067 and __ILP32__ when the ILP32 model is in use.
17069 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
17071 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
17072 (aarch64_load_symref_appropriately): In the case of
17073 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
17074 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
17075 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
17076 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
17077 if 'mode' doesn't equal to 'ptr_mode'.
17078 (aarch64_output_mi_thunk): Add an assertion on the alignment of
17079 'vcall_offset'; change to call aarch64_emit_move differently depending
17080 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
17081 to calculate the upper bound of 'vcall_offset'.
17082 (aarch64_cannot_force_const_mem): Change to also return true if
17084 (aarch64_legitimize_reload_address): In the case of large
17085 displacements, add new local variable 'xmode' and an assertion
17086 based on it; change to use 'xmode' to generate the new rtx and
17088 (aarch64_asm_trampoline_template): Change to generate the template
17089 differently depending on TARGET_ILP32 or not; change to use
17090 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
17091 (aarch64_trampoline_size): Removed.
17092 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
17093 and replace immediate literals with it. Change to use 'ptr_mode'
17094 instead of 'DImode' and call convert_memory_address if the mode
17095 of 'fnaddr' doesn't equal to 'ptr_mode'.
17096 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
17098 (aarch64_elf_asm_destructor): Likewise.
17099 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
17100 on TARGET_ILP32 instead of aarch64_trampoline_size.
17101 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
17102 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
17103 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
17104 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
17105 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
17106 (storewb_pair<GPI:mode>_<P:mode>): ... this.
17107 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
17108 depending on the value of 'mode'.
17109 (add_losym_<mode>): New.
17110 (ldr_got_small_<mode>): New, based on ldr_got_small.
17111 (ldr_got_small): Remove.
17112 (ldr_got_small_sidi): New.
17113 * config/aarch64/iterators.md (P): New.
17114 (PTR): Change to 'ptr_mode' in the condition.
17116 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
17118 * config.gcc (aarch64*-*-*): Support --with-abi.
17119 (aarch64*-*-elf): Support --with-multilib-list.
17120 (aarch64*-*-linux*): Likewise.
17121 (supported_defaults): Add abi to aarch64*-*-*.
17122 * configure.ac: Mention AArch64 for --with-multilib-list.
17123 * configure: Re-generated.
17124 * config/aarch64/biarchilp32.h: New file.
17125 * config/aarch64/biarchlp64.h: New file.
17126 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
17128 (MULTILIB_DEFAULTS): Ditto.
17129 (DRIVER_SELF_SPECS): Ditto.
17130 (ASM_SPEC): Update to also substitute -mabi.
17131 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
17132 file whose name depends on -mabi= and -mbig-endian.
17133 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
17135 (POINTER_SIZE): New define.
17136 (POINTERS_EXTEND_UNSIGNED): Ditto.
17137 (enum aarch64_abi_type): New enumeration tag.
17138 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
17139 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
17140 (TARGET_ILP32): New define.
17141 * config/aarch64/aarch64.opt (mabi): New.
17142 (aarch64_abi): New.
17143 (ilp32, lp64): New values for -mabi.
17144 * config/aarch64/t-aarch64 (comma): New define.
17145 (MULTILIB_OPTIONS): Ditto.
17146 (MULTILIB_DIRNAMES): Ditto.
17147 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
17148 * doc/invoke.texi: Document -mabi for AArch64.
17150 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
17152 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
17154 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
17155 Anton Blanchard <anton@au1.ibm.com>
17157 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
17158 endianness when selecting field to splat.
17160 2013-07-22 Eric Christopher <echristo@gmail.com>
17162 * dwarf2out.c (die_odr_checksum): New function to use
17163 CHECKSUM_ macros and ULEB128 for DIE tag.
17164 (generate_type_signature): Use.
17166 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
17168 * config.gcc (sparc*-*-*): Accept leon3 processor.
17169 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
17170 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
17171 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
17172 * config/sparc/sparc.opt (enum processor_type): Add leon3.
17173 (mfix-ut699): Adjust comment.
17174 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
17175 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
17176 (CPP_CPU_SPEC): Likewise.
17177 (ASM_CPU_SPEC): Likewise.
17178 * config/sparc/sparc.c (leon3_cost): New constant.
17179 (sparc_option_override): Add leon3 support.
17180 (mem_ref): New function.
17181 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
17182 (sparc_do_work_around_errata): Look into the instruction in the delay
17183 slot and adjust accordingly. Add fix for the data cache nullify issues
17184 of the UT699. Change insertion position for the NOP.
17185 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
17186 (leon3_load): New reservation.
17187 (leon_store): Bump latency to 2.
17188 (grfpu): New automaton.
17189 (grfpu_alu): New unit.
17190 (grfpu_ds): Likewise.
17191 (leon_fp_alu): Adjust.
17192 (leon_fp_mult): Delete.
17193 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
17194 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
17195 * config/sparc/sparc.md (cpu): Add leon3.
17196 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
17197 (swapsi): Likewise.
17198 (atomic_test_and_set): Likewise.
17199 (ldstub): Likewise.
17201 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
17203 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
17204 default for R5900 targets.
17205 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
17206 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
17207 * config/mips/mips.c (mips_option_override): Report an error for
17208 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
17209 for -march=r5900 -mhard-float.
17211 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
17213 * df-problems.c (can_move_insns_across): Exit loop once we
17214 find a non-fixed, non-global register.
17216 * ipa-pure-const.c (propagate_nothrow): Exit loop after
17219 * omega.c (omega_eliminate_red): Break after setting red_found.
17220 (omega_problem_has_red_equations): Similarly after setting found.
17221 (omega_query_variable): Similarly after setting coupled.
17223 2013-07-22 Marek Polacek <polacek@redhat.com>
17225 * gimplify.c: Don't include gimple.h twice.
17227 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17229 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
17230 instead of TARGET_THUMB1.
17231 (Pz): New constraint.
17232 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
17234 (compare_negsi_si): Likewise.
17235 (compare_addsi2_op0): Likewise.
17236 (compare_addsi2_op1): Likewise.
17237 (addsi3_carryin_<optab>): Likewise.
17238 (addsi3_carryin_alt2_<optab>): Likewise.
17239 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
17240 for arm_restrict_it.
17241 (subsi3_carryin): Likewise.
17242 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
17243 (minmax_arithsi): Disable for arm_restrict_it.
17244 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
17245 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
17246 (satsi_<SAT:code>_shift): Likewise.
17247 (arm_shiftsi3): Add alternative for 16-bit encoding.
17248 (arm32_movhf): Disable for arm_restrict_it.
17249 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
17250 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
17252 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
17254 * config/arm/arm.md (attribute "insn"): Delete.
17255 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
17256 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
17257 (not_shiftsi): Update for attribute change.
17258 (not_shiftsi_compare0): Likewise.
17259 (not_shiftsi_compare0_scratch): Likewise.
17260 (arm_one_cmplsi2): Likewise.
17261 (thumb1_one_cmplsi2): Likewise.
17262 (notsi_compare0): Likewise.
17263 (notsi_compare0_scratch): Likewise.
17264 (thumb1_movdi_insn): Likewise.
17265 (arm_movsi_insn): Likewise.
17266 (movhi_insn_arch4): Likewise.
17267 (movhi_bytes): Likewise.
17268 (arm_movqi_insn): Likewise.
17269 (thumb1_movqi_insn): Likewise.
17270 (arm32_movhf): Likewise.
17271 (thumb1_movhf): Likewise.
17272 (arm_movsf_soft_insn): Likewise.
17273 (thumb1_movsf_insn): Likewise.
17274 (thumb_movdf_insn): Likewise.
17275 (movsicc_insn): Likewise.
17276 (movsfcc_soft_insn): Likewise.
17277 (and_scc): Likewise.
17278 (cond_move): Likewise.
17279 (if_move_not): Likewise.
17280 (if_not_move): Likewise.
17281 (if_shift_move): Likewise.
17282 (if_move_shift): Likewise.
17283 (if_shift_shift): Likewise.
17284 (if_not_arith): Likewise.
17285 (if_arith_not): Likewise.
17286 (cond_move_not): Likewise.
17287 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
17288 (neon_mov<mode>): Likewise.
17289 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
17290 (thumb2_movsi_vfp): Likewise.
17291 (movsf_vfp): Likewise.
17292 (thumb2_movsf_vfp): Likewise.
17293 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
17295 (cortexa7_older_only): Likewise.
17296 (cortexa7_younger): Likewise.
17297 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
17298 (1020alu_shift_op): Likewise.
17299 (1020alu_shift_reg_op): Likewise.
17300 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
17301 (alu_shift_op): Likewise.
17302 (alu_shift_reg_op): Likewise.
17303 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
17304 (11_alu_shift_op): Likewise.
17305 (11_alu_shift_reg_op): Likewise.
17306 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
17307 (9_alu_shift_reg_op): Likewise.
17308 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
17310 (cortex_a15_alu_shift): Likewise.
17311 (cortex_a15_alu_shift_reg): Likewise.
17312 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
17313 (cortex_a5_alu_shift): Likewise.
17314 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
17316 (cortex_a53_alu_shift): Likewise.
17317 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
17319 (cortex_a7_alu_reg): Likewise.
17320 (cortex_a7_alu_shift): Likewise.
17321 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
17322 (cortex_a8_alu_shift): Likewise.
17323 (cortex_a8_alu_shift_reg): Likewise.
17324 (cortex_a8_mov): Likewise.
17325 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
17326 (cortex_a9_dp_shift): Likewise.
17327 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
17328 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
17329 (cortex_r4_mov): Likewise.
17330 (cortex_r4_alu_shift): Likewise.
17331 (cortex_r4_alu_shift_reg): Likewise.
17332 * config/arm/fa526.md (526_alu_op): Update for attribute change.
17333 (526_alu_shift_op): Likewise.
17334 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
17335 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
17336 (626te_alu_shift_op): Likewise.
17337 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
17338 (726te_alu_op): Likewise.
17339 (726te_alu_shift_op): Likewise.
17340 (726te_alu_shift_reg_op): Likewise.
17341 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
17342 (mp626_alu_shift_op): Likewise.
17343 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
17344 (pj4_alu_e1_conds): Likewise.
17345 (pj4_alu): Likewise.
17346 (pj4_alu_conds): Likewise.
17347 (pj4_shift): Likewise.
17348 (pj4_shift_conds): Likewise.
17349 (pj4_alu_shift): Likewise.
17350 (pj4_alu_shift_conds): Likewise.
17352 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17354 * config/arm/predicates.md (shiftable_operator_strict_it):
17356 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
17357 Disable cond_exec version for arm_restrict_it.
17358 (thumb2_smaxsi3): Convert to generate cond_exec.
17359 (thumb2_sminsi3): Likewise.
17360 (thumb32_umaxsi3): Likewise.
17361 (thumb2_uminsi3): Likewise.
17362 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
17363 (thumb2_neg_abssi2): Likewise.
17364 (thumb2_mov_scc): Add alternative for 16-bit encoding.
17365 (thumb2_movsicc_insn): Adjust alternatives.
17366 (thumb2_mov_negscc): Disable for arm_restrict_it.
17367 (thumb2_mov_negscc_strict_it): New pattern.
17368 (thumb2_mov_notscc_strict_it): New pattern.
17369 (thumb2_mov_notscc): Disable for arm_restrict_it.
17370 (thumb2_ior_scc): Likewise.
17371 (thumb2_ior_scc_strict_it): New pattern.
17372 (thumb2_cond_move): Adjust for arm_restrict_it.
17373 (thumb2_cond_arith): Disable for arm_restrict_it.
17374 (thumb2_cond_arith_strict_it): New pattern.
17375 (thumb2_cond_sub): Adjust for arm_restrict_it.
17376 (thumb2_movcond): Likewise.
17377 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
17378 (thumb2_zero_extendhisi2_v6): Likewise.
17379 (thumb2_zero_extendqisi2_v6): Likewise.
17380 (orsi_notsi_si): Likewise.
17381 (orsi_not_shiftsi_si): Likewise.
17383 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
17385 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
17386 instruction sequence is 1 byte shorter.
17388 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
17390 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
17391 it is not needed after split.
17393 2013-07-20 Iain Sandoe <iain@codesourcery.com>
17396 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
17397 second label for nonlocal goto receivers. Don't output pic base labels
17398 unless we're producing PIC; mark that action unreachable().
17399 (ix86_save_reg): If the function contains a nonlocal label, save the
17401 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
17402 * config/darwin.c (emitted_pic_label_num): New GTY.
17403 (update_pic_label_number_if_needed): New.
17404 (machopic_output_function_base_name): Adjust for nonlocal receiver
17406 (machopic_should_output_picbase_label): New.
17407 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
17408 (nonlocal_goto_receiver): New insn and split.
17410 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
17412 * config/aarch64/aarch64-builtins.c
17413 (aarch64_fold_builtin): Fold abs in all modes.
17414 * config/aarch64/aarch64-simd-builtins.def
17415 (abs): Enable for all modes.
17416 * config/aarch64/arm_neon.h
17417 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
17418 (vabs_f64): Add missing intrinsic.
17420 2013-07-19 Ian Bolton <ian.bolton@arm.com>
17422 * config/aarch64/arm_neon.h (vabs_s64): New function
17424 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
17427 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
17428 * config/avr/avr.md (adjust_len): Add `round'.
17429 * config/avr/avr-protos.h (avr_out_round): New prototype.
17430 (avr_out_plus): Add `out_label' argument.
17431 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
17432 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
17433 Handle the case where `insn' is just a pattern.
17434 (avr_out_bitop): Handle the case where `insn' is just a pattern.
17435 (avr_out_round): New function.
17436 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
17438 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
17440 * config/microblaze/microblaze.c (microblaze_expand_prologue):
17441 Add check for flag_stack_usage to handle -fstack-usage support
17443 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
17445 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
17446 interaction for new Power8 flags and VSX.
17448 2013-07-18 Sriraman Tallam <tmsriram@google.com>
17450 PR middle-end/57698
17451 * tree-inline.c (expand_call_inline): Emit errors during
17452 early_inlining only if optimization is not turned on.
17454 2013-07-18 David Malcolm <dmalcolm@redhat.com>
17458 * passes.c (init_optimization_passes): Move the construction of
17459 the pass hierarchy into a new passes.def file.
17461 * Makefile.in (passes.o): Add dependency on passes.def.
17463 2013-07-18 David Malcolm <dmalcolm@redhat.com>
17465 * passes.c (init_optimization_passes): Introduce macros for
17466 constructing the tree of passes (INSERT_PASSES_AFTER,
17467 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
17468 TERMINATE_PASS_LIST).
17470 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
17471 Wei Mi <wmi@google.com>
17473 PR rtl-optimization/57878
17474 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
17476 (reload_pseudo_compare_func): Check nregs first for reload
17479 2013-07-18 David Malcolm <dmalcolm@redhat.com>
17481 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
17483 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
17485 * read-rtl.c (validate_const_int): Once an invalid character is
17486 seen, quit the loop.
17488 * gengtype.c (write_roots): Similarly once we find the "deletable"
17489 or "if_marked" option.
17491 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
17493 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
17494 "xtab" and "sat". Move value "clz" from here to ...
17495 (attriubte "type"): ... here.
17496 (satsi_<SAT:code>): Delete "insn" attribute.
17497 (satsi_<SAT:code>_shift): Likewise.
17498 (arm_zero_extendqisi2addsi): Likewise.
17499 (arm_extendqisi2addsi): Likewise.
17500 (clzsi2): Update for attribute changes.
17501 (rbitsi2): Likewise.
17502 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
17504 (arm_usatsihi): Likewise.
17505 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
17507 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
17509 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
17510 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
17511 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
17512 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
17513 in alphabetical order.
17514 (attribute "core_cycles"): Update for attribute changes.
17515 (arm_addsi3): Likewise.
17516 (addsi3_compare0): Likewise.
17517 (addsi3_compare0_scratch): Likewise.
17518 (addsi3_compare_op1): Likewise.
17519 (addsi3_compare_op2): Likewise.
17520 (compare_addsi2_op0): Likewise.
17521 (compare_addsi2_op1): Likewise.
17522 (addsi3_carryin_shift_<optab>): Likewise.
17523 (subsi3_carryin_shift): Likewise.
17524 (rsbsi3_carryin_shift): Likewise.
17525 (arm_subsi3_insn): Likewise.
17526 (subsi3_compare0): Likewise.
17527 (subsi3_compare): Likewise.
17528 (arm_andsi3_insn): Likewise.
17529 (thumb1_andsi3_insn): Likewise.
17530 (andsi3_compare0): Likewise.
17531 (andsi3_compare0_scratch): Likewise.
17532 (zeroextractsi_compare0_scratch
17533 (andsi_not_shiftsi_si): Likewise.
17534 (iorsi3_insn): Likewise.
17535 (iorsi3_compare0): Likewise.
17536 (iorsi3_compare0_scratch): Likewise.
17537 (arm_xorsi3): Likewise.
17538 (thumb1_xorsi3_insn): Likewise.
17539 (xorsi3_compare0): Likewise.
17540 (xorsi3_compare0_scratch): Likewise.
17541 (satsi_<SAT:code>_shift): Likewise.
17543 (arm_shiftsi3): Likewise.
17544 (shiftsi3_compare0): Likewise.
17545 (not_shiftsi): Likewise.
17546 (not_shiftsi_compare0): Likewise.
17547 (not_shiftsi_compare0_scratch): Likewise.
17548 (arm_one_cmplsi2): Likewise.
17549 (thumb_one_complsi2): Likewise.
17550 (notsi_compare0): Likewise.
17551 (notsi_compare0_scratch): Likewise.
17552 (thumb1_zero_extendhisi2): Likewise.
17553 (arm_zero_extendhisi2): Likewise.
17554 (arm_zero_extendhisi2_v6): Likewise.
17555 (arm_zero_extendhisi2addsi): Likewise.
17556 (thumb1_zero_extendqisi2): Likewise.
17557 (thumb1_zero_extendqisi2_v6): Likewise.
17558 (arm_zero_extendqisi2): Likewise.
17559 (arm_zero_extendqisi2_v6): Likewise.
17560 (arm_zero_extendqisi2addsi): Likewise.
17561 (thumb1_extendhisi2): Likewise.
17562 (arm_extendhisi2): Likewise.
17563 (arm_extendhisi2_v6): Likewise.
17564 (arm_extendqisi): Likewise.
17565 (arm_extendqisi_v6): Likewise.
17566 (arm_extendqisi2addsi): Likewise.
17567 (thumb1_extendqisi2): Likewise.
17568 (thumb1_movdi_insn): Likewise.
17569 (arm_movsi_insn): Likewise.
17570 (movsi_compare0): Likewise.
17571 (movhi_insn_arch4): Likewise.
17572 (movhi_bytes): Likewise.
17573 (arm_movqi_insn): Likewise.
17574 (thumb1_movqi_insn): Likewise.
17575 (arm32_movhf): Likewise.
17576 (thumb1_movhf): Likewise.
17577 (arm_movsf_soft_insn): Likewise.
17578 (thumb1_movsf_insn): Likewise.
17579 (movdf_soft_insn): Likewise.
17580 (thumb_movdf_insn): Likewise.
17581 (arm_cmpsi_insn): Likewise.
17582 (cmpsi_shiftsi): Likewise.
17583 (cmpsi_shiftsi_swp): Likewise.
17584 (arm_cmpsi_negshiftsi_si): Likewise.
17585 (movsicc_insn): Likewise.
17586 (movsfcc_soft_insn): Likewise.
17587 (arith_shiftsi): Likewise.
17588 (arith_shiftsi_compare0
17589 (arith_shiftsi_compare0_scratch
17590 (sub_shiftsi): Likewise.
17591 (sub_shiftsi_compare0
17592 (sub_shiftsi_compare0_scratch
17593 (and_scc): Likewise.
17594 (cond_move): Likewise.
17595 (if_plus_move): Likewise.
17596 (if_move_plus): Likewise.
17597 (if_move_not): Likewise.
17598 (if_not_move): Likewise.
17599 (if_shift_move): Likewise.
17600 (if_move_shift): Likewise.
17601 (if_shift_shift): Likewise.
17602 (if_not_arith): Likewise.
17603 (if_arith_not): Likewise.
17604 (cond_move_not): Likewise.
17605 (thumb1_ashlsi3): Set type attribute.
17606 (thumb1_ashrsi3): Likewise.
17607 (thumb1_lshrsi3): Likewise.
17608 (thumb1_rotrsi3): Likewise.
17609 (shiftsi3_compare0_scratch): Likewise.
17610 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
17611 (neon_mov<mode>): Likewise.
17612 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
17614 (thumb2_movsi_insn): Likewise.
17615 (thumb2_cmpsi_neg_shiftsi): Likewise.
17616 (thumb2_extendqisi_v6): Likewise.
17617 (thumb2_zero_extendhisi2_v6): Likewise.
17618 (thumb2_zero_extendqisi2_v6): Likewise.
17619 (thumb2_shiftsi3_short): Likewise.
17620 (thumb2_addsi3_compare0_scratch): Likewise.
17621 (orsi_not_shiftsi_si): Likewise.
17622 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
17623 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
17625 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
17626 (1020alu_shift_op): Likewise.
17627 (1020alu_shift_reg_op): Likewise.
17628 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
17629 (alu_shift_op): Likewise.
17630 (alu_shift_reg_op): Likewise.
17631 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
17632 (11_alu_shift_op): Likewise.
17633 (11_alu_shift_reg_op): Likewise.
17634 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
17635 (9_alu_shift_reg_op): Likewise.
17636 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
17638 (cortex_a15_alu_shift): Likewise.
17639 (cortex_a15_alu_shift_reg): Likewise.
17640 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
17642 (cortex_a5_alu_shift): Likewise.
17643 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
17645 (cortex_a53_alu_shift): Likewise.
17646 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
17648 (cortex_a7_alu_reg): Likewise.
17649 (cortex_a7_alu_shift): Likewise.
17650 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
17652 (cortex_a8_alu_shift): Likewise.
17653 (cortex_a8_alu_shift_reg): Likewise.
17654 (cortex_a8_mov): Likewise.
17655 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
17656 (cortex_a9_dp_shift): Likewise.
17657 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
17659 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
17661 (cortex_r4_mov): Likewise.
17662 (cortex_r4_alu_shift): Likewise.
17663 (cortex_r4_alu_shift_reg): Likewise.
17664 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
17665 (526_alu_shift_op): Likewise.
17666 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
17667 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
17668 (626te_alu_shift_op): Likewise.
17669 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
17670 (726te_alu_op): Likewise.
17671 (726te_alu_shift_op): Likewise.
17672 (726te_alu_shift_reg_op): Likewise.
17673 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
17674 (mp626_alu_shift_op): Likewise.
17675 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
17676 (pj4_alu_e1_conds): Likewise.
17677 (pj4_alu): Likewise.
17678 (pj4_alu_conds): Likewise.
17679 (pj4_shift): Likewise.
17680 (pj4_shift_conds): Likewise.
17681 (pj4_alu_shift): Likewise.
17682 (pj4_alu_shift_conds): Likewise.
17683 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
17685 (cortexa7_older_only): Likewise.
17686 (cortexa7_younger): Likewise.
17688 2013-07-18 David Malcolm <dmalcolm@redhat.com>
17690 * ipa-pure-const.c (generate_summary): Rename to...
17691 (pure_const_generate_summary): ... this.
17693 2013-07-17 Iain Sandoe <iain@codesourcery.com>
17695 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
17697 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
17700 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
17703 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17705 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
17706 enabled without -march=zEC12.
17707 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
17710 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
17712 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
17713 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
17714 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
17715 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
17716 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
17717 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
17718 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
17720 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
17721 and ISA_HAS_NMADD3_NMSUB3.
17722 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
17723 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
17724 (nmsub4<mode>, nmsub3<mode>): Likewise.
17725 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
17727 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
17729 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
17730 TARGET_MIPS5400 checking.
17732 2013-07-16 Jakub Jelinek <jakub@redhat.com>
17733 Peter Bergner <bergner@vnet.ibm.com>
17735 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
17736 registers in the comment.
17737 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
17738 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
17739 rather than FIRST_PSEUDO_REGISTERS.
17741 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
17743 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
17744 enable extra ISA flags with TARGET_HTM.
17746 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
17748 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
17751 2013-07-15 Cong Hou <congh@google.com>
17753 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
17754 in compare function for sorting.
17756 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
17758 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
17759 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
17760 * config/rs6000/rs6000.opt: Add -mhtm option.
17761 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
17762 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
17763 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
17764 __HTM__ if the HTM instructions are available.
17765 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
17766 htm_spr_reg_operand): New define_predicates.
17767 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
17768 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
17770 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
17771 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
17772 HTM builtin functions.
17773 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
17774 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
17775 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
17776 (rs6000_builtin_mask_calculate): Likewise.
17777 (rs6000_option_override_internal): Likewise.
17778 (bdesc_htm): Add new HTM builtin support.
17779 (htm_spr_num): New function.
17780 (htm_spr_regno): Likewise.
17781 (rs6000_htm_spr_icode): Likewise.
17782 (htm_expand_builtin): Likewise.
17783 (htm_init_builtins): Likewise.
17784 (rs6000_expand_builtin): Add support for HTM builtin functions.
17785 (rs6000_init_builtins): Likewise.
17786 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
17788 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
17789 (TARGET_HTM, MASK_HTM): Define macros.
17790 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
17791 (FIXED_REGISTERS): Likewise.
17792 (CALL_USED_REGISTERS): Likewise.
17793 (CALL_REALLY_USED_REGISTERS): Likewise.
17794 (REG_ALLOC_ORDER): Likewise.
17795 (enum reg_class): Likewise.
17796 (REG_CLASS_NAMES): Likewise.
17797 (REG_CLASS_CONTENTS): Likewise.
17798 (REGISTER_NAMES): Likewise.
17799 (ADDITIONAL_REGISTER_NAMES): Likewise.
17800 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
17801 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
17802 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
17803 * config/rs6000/htm.md: New file.
17804 * config/rs6000/htmintrin.h: New file.
17805 * config/rs6000/htmxlintrin.h: New file.
17807 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
17809 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
17810 Define SYMBOL_TINY_GOT, update comment.
17811 * config/aarch64/aarch64.c
17812 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
17813 (aarch64_expand_mov_immediate): Likewise.
17814 (aarch64_print_operand): Likewise.
17815 (aarch64_classify_symbol): Likewise.
17816 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
17817 (ldr_got_tiny): Define.
17819 2013-07-13 Tobias Grosser <tobias@grosser.es>
17821 PR tree-optimization/54094
17822 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
17823 scheduling dimension for the parallelism check from the polyhedral
17824 information in the AST.
17825 * graphite-dependences.c (carries_deps): Do not assume the schedule is
17828 2013-07-13 Jason Merrill <jason@redhat.com>
17830 * print-tree.c (debug_vec_tree): Use debug_raw.
17831 (debug_raw (vec<tree, va_gc> &)): New.
17832 (debug_raw (vec<tree, va_gc> *)): New.
17833 * tree.h: Declare them.
17835 2013-07-13 Bin Cheng <bin.cheng@arm.com>
17837 * ifcvt.c (ifcvt_after_combine): New static variable.
17838 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
17840 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
17841 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
17842 rest_of_handle_if_after_reload): Pass new argument for if_convert.
17844 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
17846 * config/mips/mips.c (mips_expand_call): Remove empty statement.
17848 2013-07-12 Michael Matz <matz@suse.de>
17850 PR middle-end/55771
17851 * convert.c (convert_to_real): Reject non-float inner types.
17853 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
17855 * config/aarch64/aarch64-protos.h
17856 (aarch64_simd_immediate_valid_for_move): Remove.
17857 * config/aarch64/aarch64.c (simd_immediate_info): New member.
17858 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
17860 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
17862 2013-07-11 Steve Ellcey <sellcey@mips.com>
17864 * config/mips/mips.c (mips_conditional_register_usage): Do not
17865 use t[0-7] registers in MIPS16 mode when optimizing for size.
17867 2013-07-11 Sriraman Tallam <tmsriram@google.com>
17869 * config/i386/i386.c (dispatch_function_versions): Fix array
17870 indexing of function_version_info to match actual_versions.
17872 2013-07-11 Teresa Johnson <tejohnson@google.com>
17874 * vec.h (struct va_gc): Move release out-of-line.
17875 (va_gc::release): Call ggc_free on released vec.
17877 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
17879 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
17880 Require GOT register as additional operand in UNSPEC.
17881 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
17882 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
17883 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
17884 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
17885 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
17886 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
17887 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
17889 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
17892 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
17893 name seen by assembler/linker rather if available.
17895 2013-07-11 Andreas Schwab <schwab@suse.de>
17897 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
17899 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
17901 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
17903 2013-07-10 Joseph Myers <joseph@codesourcery.com>
17905 * doc/tm.texi.in: Move hook documentation to ....
17906 * target.def: ... here.
17908 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
17909 text on @hook line.
17910 * doc/tm.texi: Regenerate.
17912 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
17915 * doc/invoke.texi: Document -Wconditionally-supported.
17917 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
17920 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
17923 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
17926 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
17927 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
17928 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
17929 Remove duplicate devices.
17930 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
17931 * config/avr/t-multilib: Regenerate.
17932 * config/avr/avr-tables.opt: Regenerate.
17933 * doc/avr-mmcu.texi: Regenerate.
17935 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
17938 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
17940 2013-07-10 Graham Stott <graham.stott@btinternet.com>
17942 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
17943 the cost of MULT when optimizing for size.
17945 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
17947 * config/cr16/cr16-protos.h: Don't include target.h.
17949 2013-07-09 Joseph Myers <joseph@codesourcery.com>
17951 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
17952 adjust register size for TDmode and TFmode for VSX registers.
17954 2013-07-08 Kai Tietz <ktietz@redhat.com>
17957 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
17958 hook_bool_const_tree_true.
17960 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17962 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
17963 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
17964 * config/s390/s390.md: Define FPR*_REGNUM constants.
17965 Fix FPR2_REGNUM constant (18 -> 17).
17966 ("*trunc<BFP:mode><DFP_ALL:mode>2")
17967 ("*trunc<DFP_ALL:mode><BFP:mode>2")
17968 ("trunc<BFP:mode><DFP_ALL:mode>2")
17969 ("trunc<DFP_ALL:mode><BFP:mode>2")
17970 ("*extend<BFP:mode><DFP_ALL:mode>2")
17971 ("*extend<DFP_ALL:mode><BFP:mode>2")
17972 ("extend<BFP:mode><DFP_ALL:mode>2")
17973 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
17976 2013-07-08 Graham Stott <graham.stott@btinternet.com>
17978 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
17980 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17982 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
17983 and cfun_fpr_bit_p to cfun_fpr_save_p.
17984 (s390_frame_area, s390_register_info, s390_frame_info)
17985 (s390_emit_prologue, s390_emit_epilogue)
17986 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
17988 * config/s390/s390.h: Define *_REGNUM macros for floating point
17991 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
17993 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
17995 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
17997 PR rtl-optimization/57786
17998 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
17999 and break out of the loop when it is set to false.
18001 2013-07-08 Jakub Jelinek <jakub@redhat.com>
18004 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
18005 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
18006 (const_int 63)) 0)).
18007 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
18008 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
18009 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
18011 PR rtl-optimization/57829
18012 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
18013 mask bits outside of mode are just sign-extension from mode to HWI.
18015 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
18017 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
18018 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
18019 adjust_address instead of change_address to keep info about alignment.
18020 (emit_strmov): Remove.
18021 (emit_memmov): New function.
18022 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
18023 (expand_movmem_epilogue): Likewise and return updated rtx for
18025 (expand_constant_movmem_prologue): Likewise and return updated rtx for
18026 destination and source.
18027 (decide_alignment): Refactor, handle vector_loop.
18028 (ix86_expand_movmem): Likewise.
18029 (ix86_expand_setmem): Likewise.
18030 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
18032 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
18034 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
18035 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
18037 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
18039 * config/i386/sse.md (sse_movlhps): Change alternative 3
18040 of operand 2 to "m".
18042 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
18045 * config/i386/sse.md (iptr): New mode attribute.
18046 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
18047 (<sse>_vm<plusminus_insn><mode>3): Ditto.
18048 (<sse>_vmmul<mode>3): Ditto.
18049 (<sse>_vmdiv<mode>3): Ditto.
18050 (sse_vmrcpv4sf2): Ditto.
18051 (<sse>_vmsqrt<mode>2): Ditto.
18052 (sse_vmrsqrtv4sf2): Ditto.
18053 (<sse>_vm<code><mode>3): Ditto.
18054 (avx_vmcmp<mode>3): Ditto.
18055 (<sse>_vmmaskcmp<mode>3): Ditto.
18056 (<sse>_comi): Ditto.
18057 (<sse>_ucomi): Ditto.
18058 (*xop_vmfrcz_<mode>): Ditto.
18059 (*fmai_fmadd_<mode>): Ditto.
18060 (*fmai_fmsub_<mode>): Ditto.
18061 (*fmai_fnmadd_<mode>): Ditto.
18062 (*fmai_fnmsub_<mode>): Ditto.
18063 (*fma4i_vmfmadd_<mode>): Ditto.
18064 (*fma4i_vmfmsub_<mode>): Ditto.
18065 (*fma4i_vmfnmadd_<mode>): Ditto.
18066 (*fma4i_vmfnmsub_<mode>): Ditto.
18067 (*xop_vmfrcz_<mode>): Ditto.
18068 (sse_cvtps2pi): Ditto.
18069 (sse_cvttps2pi): Ditto.
18070 (sse_cvtss2si): Ditto.
18071 (sse_cvtss2si_2): Ditto.
18072 (sse_cvtss2siq_2): Ditto.
18073 (sse_cvttss2si): Ditto.
18074 (sse_cvttss2siq): Ditto.
18075 (sse_cvtsd2si): Ditto.
18076 (sse_cvtsd2si_2): Ditto.
18077 (sse_cvtsd2siq_2): Ditto.
18078 (sse_cvttsd2si): Ditto.
18079 (sse_cvttsd2siq): Ditto.
18080 (sse_cvtsd2ss): Ditto.
18081 (sse_cvtss2sd): Ditto.
18082 (avx2_pbroadcast<mode>): Ditto.
18083 (avx2_pbroadcast<mode>_1): Ditto.
18084 (*avx_vperm_broadcast_v4sf): Ditto.
18086 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
18087 (sse_movlhps): Ditto.
18088 (sse_storehps): Ditto.
18089 (sse_loadhps): Ditto.
18090 (sse_storelps): Ditto.
18091 (sse_loadlps): Ditto.
18092 (*vec_concatv4sf): Ditto.
18093 (*vec_interleave_highv2df): Ditto.
18094 (*vec_interleave_lowv2df): Ditto.
18095 (*vec_extractv2df_1_sse): Ditto.
18096 (*vec_extractv2df_0_sse): Ditto.
18097 (sse2_storelpd): Ditto.
18098 (sse2_loadlpd): Ditto.
18099 (sse2_movsd): Ditto.
18100 (*vec_concatv4si): Ditto.
18101 (vec_concatv2di): Ditto.
18103 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
18104 for Intel asm dialect.
18105 (mmx_punpcklwd): Ditto.
18106 (mmx_punpckldq): Ditto.
18108 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
18109 for intel assembler dialect.
18111 2013-07-06 Jakub Jelinek <jakub@redhat.com>
18114 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
18115 for BUILT_IN_C{LZ,LRSB}*.
18116 * tree.h (CASE_INT_FN): Add FN##IMAX case.
18117 * tree-vrp.c (extract_range_basic): Handle
18118 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
18119 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
18120 fall thru to code calling set_value*.
18121 * builtins.c (expand_builtin): Remove *IMAX cases.
18122 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
18123 if width is bigger than 2*HWI.
18125 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
18127 PR rtl-optimization/55342
18128 * lra-int.h (lra_subreg_reload_pseudos): New.
18129 * lra.c: Add undoing optional reloads to the block diagram.
18130 (lra_subreg_reload_pseudos): New.
18131 (lra_optional_reload_pseudos): Change comments.
18132 (lra): Init and clear lra_subreg_reload_pseudos. Clear
18133 lra_optional_reload_pseudos after undo transformations.
18134 * lra-assigns.c (pseudo_prefix_title): New.
18135 (lra_setup_reg_renumber): Use it.
18136 (spill_for): Ditto. Check subreg reload pseudos too.
18137 (assign_by_spills): Consider subreg reload pseudos too.
18138 * lra-constraints.c (simplify_operand_subreg): Use
18139 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
18140 (curr_insn_transform): Recognize and do optional reloads.
18141 (undo_optional_reloads): New.
18142 (lra_undo_inheritance): Call undo_optional_reloads.
18144 2013-07-05 Thomas Quinot <quinot@adacore.com>
18146 * tree-complex.c (expand_complex_operations_1): Fix typo.
18148 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
18150 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
18151 (tune_params): New member 'const vec_costs'.
18152 * config/aarch64/aarch64.c (generic_vector_cost): New.
18153 (generic_tunings): New member 'generic_vector_cost'.
18154 (aarch64_builtin_vectorization_cost): New.
18155 (aarch64_add_stmt_cost): New.
18156 (TARGET_VECTORIZE_ADD_STMT_COST): New.
18157 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
18159 2013-07-03 Jakub Jelinek <jakub@redhat.com>
18162 * config/i386/predicates.md (vsib_address_operand): Disallow
18163 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
18165 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
18167 PR middle-end/55030
18168 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
18169 expand_builtin_setjmp_receiver.
18170 (expand_label): Adjust, call expand_builtin_setjmp_receiver
18171 with NULL for the label parameter.
18172 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
18173 the frame-pointer. Adjust comments.
18174 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
18175 only if LABEL is non-NULL.
18177 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
18179 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
18180 (ARM_ABI_AAPCS64): Ditto.
18182 (ARM_DEFAULT_ABI): Ditto.
18184 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
18186 * config/aarch64/aarch64-builtins.c
18187 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
18188 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
18190 * config/aarch64/aarch64-simd.md
18191 (aarch64_ld1<VALL:mode>): New.
18192 (aarch64_st1<VALL:mode>): Likewise.
18193 * config/aarch64/arm_neon.h
18194 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
18196 2013-07-02 Sriraman Tallam <tmsriram@google.com>
18198 * config/i386/i386.c (gate_insert_vzeroupper): Check if
18200 (ix86_option_override_internal):Turn on all -mavx target flags by
18201 default as they are dependent on AVX anyway.
18203 2013-07-02 Cary Coutant <ccoutant@google.com>
18205 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
18206 deterministic hash.
18207 (loc_checksum_ordered): Likewise.
18208 (hash_loc_operands): Remove inline keyword.
18210 2013-07-02 Jakub Jelinek <jakub@redhat.com>
18212 PR tree-optimization/57741
18213 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
18214 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
18215 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
18216 Allow REAL_CST step_exprs if flag_associative_math.
18217 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
18219 2013-07-02 Ian Bolton <ian.bolton@arm.com>
18221 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
18223 2013-07-02 Ian Bolton <ian.bolton@arm.com>
18225 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
18227 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18229 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
18231 (iorsi3_insn): Likewise.
18232 (arm_xorsi3): Likewise.
18234 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
18236 * arm.md (attribute "wtype"): Delete. Move attribute values from here
18238 (attribute "type"): ... here, and prefix with "wmmx_".
18239 (attribute "core_cycles"): Update for attribute changes.
18240 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
18241 (tbcstv4hi): Likewise.
18242 (tbcstv2si): Likewise.
18243 (iwmmxt_iordi3): Likewise.
18244 (iwmmxt_xordi3): Likewise.
18245 (iwmmxt_anddi3): Likewise.
18246 (iwmmxt_nanddi3): Likewise.
18247 (iwmmxt_arm_movdi): Likewise.
18248 (iwmmxt_movsi_insn): Likewise.
18249 (mov<mode>_internal): Likewise.
18250 (and<mode>3_iwmmxt): Likewise.
18251 (ior<mode>3_iwmmxt): Likewise.
18252 (xor<mode>3_iwmmxt): Likewise.
18253 (add<mode>3_iwmmxt): Likewise.
18254 (ssaddv8qi3): Likewise.
18255 (ssaddv4hi3): Likewise.
18256 (ssaddv2si3): Likewise.
18257 (usaddv8qi3): Likewise.
18258 (usaddv4hi3): Likewise.
18259 (usaddv2si3): Likewise.
18260 (sub<mode>3_iwmmxt): Likewise.
18261 (sssubv8qi3): Likewise.
18262 (sssubv4hi3): Likewise.
18263 (sssubv2si3): Likewise.
18264 (ussubv8qi3): Likewise.
18265 (ussubv4hi3): Likewise.
18266 (ussubv2si3): Likewise.
18267 (mulv4hi3_iwmmxt): Likewise.
18268 (smulv4hi3_highpart): Likewise.
18269 (umulv4hi3_highpart): Likewise.
18270 (iwmmxt_wmacs): Likewise.
18271 (iwmmxt_wmacsz): Likewise.
18272 (iwmmxt_wmacu): Likewise.
18273 (iwmmxt_wmacuz): Likewise.
18274 (iwmmxt_clrdi): Likewise.
18275 (iwmmxt_clrv8qi): Likewise.
18276 (iwmmxt_clr4hi): Likewise.
18277 (iwmmxt_clr2si): Likewise.
18278 (iwmmxt_uavgrndv8qi3): Likewise.
18279 (iwmmxt_uavgrndv4hi3): Likewise.
18280 (iwmmxt_uavgv8qi3): Likewise.
18281 (iwmmxt_uavgv4hi3): Likewise.
18282 (iwmmxt_tinsrb): Likewise.
18283 (iwmmxt_tinsrh): Likewise.
18284 (iwmmxt_tinsrw): Likewise.
18285 (iwmmxt_textrmub): Likewise.
18286 (iwmmxt_textrmsb): Likewise.
18287 (iwmmxt_textrmuh): Likewise.
18288 (iwmmxt_textrmsh): Likewise.
18289 (iwmmxt_textrmw): Likewise.
18290 (iwmxxt_wshufh): Likewise.
18291 (eqv8qi3): Likewise.
18292 (eqv4hi3): Likewise.
18293 (eqv2si3): Likewise.
18294 (gtuv8qi3): Likewise.
18295 (gtuv4hi3): Likewise.
18296 (gtuv2si3): Likewise.
18297 (gtv8qi3): Likewise.
18298 (gtv4hi3): Likewise.
18299 (gtv2si3): Likewise.
18300 (smax<mode>3_iwmmxt): Likewise.
18301 (umax<mode>3_iwmmxt): Likewise.
18302 (smin<mode>3_iwmmxt): Likewise.
18303 (umin<mode>3_iwmmxt): Likewise.
18304 (iwmmxt_wpackhss): Likewise.
18305 (iwmmxt_wpackwss): Likewise.
18306 (iwmmxt_wpackdss): Likewise.
18307 (iwmmxt_wpackhus): Likewise.
18308 (iwmmxt_wpackwus): Likewise.
18309 (iwmmxt_wpackdus): Likewise.
18310 (iwmmxt_wunpckihb): Likewise.
18311 (iwmmxt_wunpckihh): Likewise.
18312 (iwmmxt_wunpckihw): Likewise.
18313 (iwmmxt_wunpckilb): Likewise.
18314 (iwmmxt_wunpckilh): Likewise.
18315 (iwmmxt_wunpckilw): Likewise.
18316 (iwmmxt_wunpckehub): Likewise.
18317 (iwmmxt_wunpckehuh): Likewise.
18318 (iwmmxt_wunpckehuw): Likewise.
18319 (iwmmxt_wunpckehsb): Likewise.
18320 (iwmmxt_wunpckehsh): Likewise.
18321 (iwmmxt_wunpckehsw): Likewise.
18322 (iwmmxt_wunpckelub): Likewise.
18323 (iwmmxt_wunpckeluh): Likewise.
18324 (iwmmxt_wunpckeluw): Likewise.
18325 (iwmmxt_wunpckelsb): Likewise.
18326 (iwmmxt_wunpckelsh): Likewise.
18327 (iwmmxt_wunpckelsw): Likewise.
18328 (ror<mode>3): Likewise.
18329 (ashr<mode>3_iwmmxt): Likewise.
18330 (lshr<mode>3_iwmmxt): Likewise.
18331 (ashl<mode>3_iwmmxt): Likewise.
18332 (ror<mode>3_di): Likewise.
18333 (ashr<mode>3_di): Likewise.
18334 (lshr<mode>3_di): Likewise.
18335 (ashl<mode>3_di): Likewise.
18336 (iwmmxt_wmadds): Likewise.
18337 (iwmmxt_wmaddu): Likewise.
18338 (iwmmxt_tmia): Likewise.
18339 (iwmmxt_tmiaph): Likewise.
18340 (iwmmxt_tmiabb): Likewise.
18341 (iwmmxt_tmiatb): Likewise.
18342 (iwmmxt_tmiabt): Likewise.
18343 (iwmmxt_tmiatt): Likewise.
18344 (iwmmxt_tmovmskb): Likewise.
18345 (iwmmxt_tmovmskh): Likewise.
18346 (iwmmxt_tmovmskw): Likewise.
18347 (iwmmxt_waccb): Likewise.
18348 (iwmmxt_wacch): Likewise.
18349 (iwmmxt_waccw): Likewise.
18350 (iwmmxt_waligni): Likewise.
18351 (iwmmxt_walignr): Likewise.
18352 (iwmmxt_walignr0): Likewise.
18353 (iwmmxt_walignr1): Likewise.
18354 (iwmmxt_walignr2): Likewise.
18355 (iwmmxt_walignr3): Likewise.
18356 (iwmmxt_wsadb): Likewise.
18357 (iwmmxt_wsadh): Likewise.
18358 (iwmmxt_wsadbz): Likewise.
18359 (iwmmxt_wsadhz): Likewise.
18360 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
18361 (iwmmxt_wabsdiffb): Likewise.
18362 (iwmmxt_wabsdiffh): Likewise.
18363 (iwmmxt_wabsdiffw): Likewise.
18364 (iwmmxt_waddsubhx): Likewise
18365 (iwmmxt_wsubaddhx): Likewise.
18366 (addc<mode>3): Likewise.
18367 (iwmmxt_avg4): Likewise.
18368 (iwmmxt_avg4r): Likewise.
18369 (iwmmxt_wmaddsx): Likewise.
18370 (iwmmxt_wmaddux): Likewise.
18371 (iwmmxt_wmaddsn): Likewise.
18372 (iwmmxt_wmaddun): Likewise.
18373 (iwmmxt_wmulwsm): Likewise.
18374 (iwmmxt_wmulwum): Likewise.
18375 (iwmmxt_wmulsmr): Likewise.
18376 (iwmmxt_wmulumr): Likewise.
18377 (iwmmxt_wmulwsmr): Likewise.
18378 (iwmmxt_wmulwumr): Likewise.
18379 (iwmmxt_wmulwl): Likewise.
18380 (iwmmxt_wqmulm): Likewise.
18381 (iwmmxt_wqmulwm): Likewise.
18382 (iwmmxt_wqmulmr): Likewise.
18383 (iwmmxt_wqmulwmr): Likewise.
18384 (iwmmxt_waddbhusm): Likewise.
18385 (iwmmxt_waddbhusl): Likewise.
18386 (iwmmxt_wqmiabb): Likewise.
18387 (iwmmxt_wqmiabt): Likewise.
18388 (iwmmxt_wqmiatb): Likewise.
18389 (iwmmxt_wqmiatt): Likewise.
18390 (iwmmxt_wqmiabbn): Likewise.
18391 (iwmmxt_wqmiabtn): Likewise.
18392 (iwmmxt_wqmiatbn): Likewise.
18393 (iwmmxt_wqmiattn): Likewise.
18394 (iwmmxt_wmiabb): Likewise.
18395 (iwmmxt_wmiabt): Likewise.
18396 (iwmmxt_wmiatb): Likewise.
18397 (iwmmxt_wmiatt): Likewise.
18398 (iwmmxt_wmiabbn): Likewise.
18399 (iwmmxt_wmiabtn): Likewise.
18400 (iwmmxt_wmiatbn): Likewise.
18401 (iwmmxt_wmiattn): Likewise.
18402 (iwmmxt_wmiawbb): Likewise.
18403 (iwmmxt_wmiawbt): Likewise.
18404 (iwmmxt_wmiawtb): Likewise.
18405 (iwmmxt_wmiawtt): Likewise.
18406 (iwmmxt_wmiawbbn): Likewise.
18407 (iwmmxt_wmiawbtn): Likewise.
18408 (iwmmxt_wmiawtbn): Likewise.
18409 (iwmmxt_wmiawttn): Likewise.
18410 (iwmmxt_wmerge): Likewise.
18411 (iwmmxt_tandc<mode>3): Likewise.
18412 (iwmmxt_torc<mode>3): Likewise.
18413 (iwmmxt_torvsc<mode>3): Likewise.
18414 (iwmmxt_textrc<mode>3): Likewise.
18415 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
18416 (wmmxt_pack): Likewise.
18417 (wmmxt_mult_c1): Likewise.
18418 (wmmxt_mult_c2): Likewise.
18419 (wmmxt_alu_c1): Likewise.
18420 (wmmxt_alu_c2): Likewise.
18421 (wmmxt_alu_c3): Likewise.
18422 (wmmxt_transfer_c1): Likewise.
18423 (wmmxt_transfer_c2): Likewise.
18424 (wmmxt_transfer_c3): Likewise.
18425 (marvell_f_iwmmxt_wstr): Likewise.
18426 (marvell_f_iwmmxt_wldr): Likewise.
18428 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
18430 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
18432 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
18435 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
18436 * lra-constraints.c (need_for_split_p): Check call used hard regs
18437 living through calls.
18439 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
18440 call used regs for call insn.
18442 2013-06-28 Jakub Jelinek <jakub@redhat.com>
18445 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
18446 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
18448 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
18450 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
18452 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
18454 * lra-constraints.c (need_for_split_p): Check call used hard regs
18455 living through calls.
18457 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
18460 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
18461 to tie with any other modes. Eliminate Altivec vector mode tests,
18462 since these are a subset of ALTIVEC or VSX vector modes. Simplify
18463 code, to return 0 if testing MODE2 for a condition, if we've
18464 already tested MODE1 for the same condition.
18466 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
18468 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
18471 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
18473 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
18474 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
18476 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
18478 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
18480 (aarch64_symbolic_constant_p): Remove.
18481 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
18482 static. Fix line length and white space.
18483 (aarch64_symbolic_constant_p): Remove.
18484 * config/aarch64/predicates.md (aarch64_valid_symref):
18485 Use aarch64_classify_symbol_expression.
18487 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18489 * config/arm/constraints.md (Ts): New constraint.
18490 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
18492 (compare_scc): Use "Ts" constraint for operand 0.
18493 (ior_scc_scc): Likewise.
18494 (and_scc_scc): Likewise.
18495 (and_scc_scc_nodom): Likewise.
18496 (ior_scc_scc_cmp): Likewise for operand 7.
18497 (and_scc_scc_cmp): Likewise.
18498 * config/arm/thumb2.md (thumb2_movsi_insn):
18499 Add alternatives for 16-bit encodings.
18500 (thumb2_movhi_insn): Likewise.
18501 (thumb2_movsicc_insn): Likewise.
18502 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
18503 (thumb2_negscc): Use "Ts" constraint.
18504 Move mvn instruction outside cond_exec block.
18505 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
18506 for 16-bit encodings.
18508 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18510 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
18512 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
18513 (mulsi3subsi): Likewise.
18514 (mulsidi3adddi): Likewise.
18515 (mulsidi3_v6): Likewise.
18516 (umulsidi3_v6): Likewise.
18517 (umulsidi3adddi_v6): Likewise.
18518 (smulsi3_highpart_v6): Likewise.
18519 (umulsi3_highpart_v6): Likewise.
18520 (mulhisi3tb): Likewise.
18521 (mulhisi3bt): Likewise.
18522 (mulhisi3tt): Likewise.
18523 (maddhisi4): Likewise.
18524 (maddhisi4tb): Likewise.
18525 (maddhisi4tt): Likewise.
18526 (maddhidi4): Likewise.
18527 (maddhidi4tb): Likewise.
18528 (maddhidi4tt): Likewise.
18529 (zeroextractsi_compare0_scratch): Likewise.
18530 (insv_zero): Likewise.
18531 (insv_t2): Likewise.
18532 (anddi_notzesidi_di): Likewise.
18533 (anddi_notsesidi_di): Likewise.
18534 (andsi_notsi_si): Likewise.
18535 (iordi_zesidi_di): Likewise.
18536 (xordi_zesidi_di): Likewise.
18537 (andsi_iorsi3_notsi): Likewise.
18538 (smax_0): Likewise.
18539 (smax_m1): Likewise.
18540 (smin_0): Likewise.
18541 (not_shiftsi): Likewise.
18542 (unaligned_loadsi): Likewise.
18543 (unaligned_loadhis): Likewise.
18544 (unaligned_loadhiu): Likewise.
18545 (unaligned_storesi): Likewise.
18546 (unaligned_storehi): Likewise.
18547 (extv_reg): Likewise.
18548 (extzv_t2): Likewise.
18549 (divsi3): Likewise.
18550 (udivsi3): Likewise.
18551 (arm_zero_extendhisi2addsi): Likewise.
18552 (arm_zero_extendqisi2addsi): Likewise.
18553 (compareqi_eq0): Likewise.
18554 (arm_extendhisi2_v6): Likewise.
18555 (arm_extendqisi2addsi): Likewise.
18556 (arm_movt): Likewise.
18557 (thumb2_ldrd): Likewise.
18558 (thumb2_ldrd_base): Likewise.
18559 (thumb2_ldrd_base_neg): Likewise.
18560 (thumb2_strd): Likewise.
18561 (thumb2_strd_base): Likewise.
18562 (thumb2_strd_base_neg): Likewise.
18563 (arm_negsi2): Add alternative for 16-bit encoding.
18564 (arm_one_cmplsi2): Likewise.
18566 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18568 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
18569 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
18570 (movdfcc): Likewise.
18571 * config/arm/vfp.md (*thumb2_movsf_vfp):
18572 Disable predication for arm_restrict_it.
18573 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
18574 (*thumb2_movdfcc_vfp): Likewise.
18575 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
18576 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
18577 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
18578 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
18579 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
18580 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
18581 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
18582 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
18583 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
18584 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
18585 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
18586 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
18587 Disable predication for arm_restrict_it.
18589 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
18591 * config/i386/bmiintrin.h (_bextr_u32): New.
18592 (_bextr_u64): Ditto.
18594 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
18596 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
18597 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
18598 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
18599 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
18600 * config/mips/n32-elf.h: ...this new file.
18602 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
18605 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
18606 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
18608 2013-06-27 Catherine Moore <clm@codesourcery.com>
18610 * config/mips/mips-tables.opt: Regenerate.
18611 * config/mips/mips-cpus.def: Add m14ke and m14kec.
18612 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
18613 * doc/invoke.texi: Add -m14kc.
18615 2013-06-27 Jakub Jelinek <jakub@redhat.com>
18618 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
18619 constraints of operand 1 and 2.
18622 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
18623 to match RTL canonicalization. Swap predicates and
18624 constraints of operand 1 and 2.
18626 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
18628 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
18629 Process OP_INOUT regs for splitting too.
18631 2013-06-27 Jakub Jelinek <jakub@redhat.com>
18633 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
18634 decl before the loop, initialize to NULL.
18635 (vectorizable_load): Initialize ptr_incr to NULL.
18637 2013-06-27 Martin Jambor <mjambor@suse.cz>
18640 * ipa-ref.h (ipa_maybe_record_reference): Declare.
18641 * ipa-ref.c (ipa_maybe_record_reference): New function.
18642 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
18643 * ipa-cp.c (create_specialized_node): Record potential references from
18645 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
18647 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
18649 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
18650 parameter 'mode' of type 'enum machine_mode mode'; change to pass
18651 'mode' to force_reg.
18652 (aarch64_add_offset): Update calls to aarch64_force_temporary.
18653 (aarch64_expand_mov_immediate): Likewise.
18655 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
18657 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
18658 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
18660 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18662 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
18663 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
18664 (struct machine_function): Add tbegin_p.
18665 (s390_canonicalize_comparison): Fold CC mode compares to
18666 conditional jump if possible.
18667 (s390_emit_jump): Return the emitted jump.
18668 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
18669 Handle CCRAWmode compares.
18670 (s390_option_override): Default to -mhtm if available.
18671 (s390_reg_clobbered_rtx): Handle floating point regs as well.
18672 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
18673 FPRs instead of df_regs_ever_live_p.
18674 (s390_optimize_nonescaping_tx): New function.
18675 (s390_init_frame_layout): Extend clobbered_regs array to cover
18677 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
18678 (s390_expand_tbegin): New function.
18679 (enum s390_builtin): New enum definition.
18680 (code_for_builtin): New array definition.
18681 (s390_init_builtins): New function.
18682 (s390_expand_builtin): New function.
18683 (TARGET_INIT_BUILTINS): Define.
18684 (TARGET_EXPAND_BUILTIN): Define.
18685 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
18686 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
18687 (s390_alc_comparison): Likewise.
18688 * config/s390/s390-modes.def: Add CCRAWmode.
18689 * config/s390/s390.h (processor_flags): Add PF_TX.
18690 (TARGET_CPU_HTM): Define macro.
18691 (TARGET_HTM): Define macro.
18692 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
18693 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
18694 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
18695 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
18696 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
18697 (TBEGIN_MASK, TBEGINC_MASK): New constants.
18698 ("*cc_to_int"): Move up.
18699 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
18700 constants other than 0.
18701 ("*ccraw_to_int"): New insn and splitter definition.
18702 ("tbegin", "tbegin_nofloat", "tbegin_retry")
18703 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
18704 ("tx_assist"): New expander.
18705 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
18706 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
18707 * config/s390/s390.opt: Add -mhtm option.
18708 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
18709 * config/s390/htmxlintrin.h: New file.
18710 * config/s390/htmintrin.h: New file.
18711 * config/s390/s390intrin.h: New file.
18712 * doc/extend.texi: Document htm builtins.
18713 * config.gcc: Add the new header files to extra_headers.
18715 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
18717 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
18718 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
18720 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
18721 Pat Haugen <pthaugen@us.ibm.com>
18722 Peter Bergner <bergner@vnet.ibm.com>
18724 * config/rs6000/power8.md: New.
18725 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
18726 setting for power8 entry.
18727 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
18728 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
18729 test for Power4/Power5 only.
18730 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
18732 (force_new_group): Adjust comment.
18733 * config/rs6000/rs6000.md: Include power8.md.
18735 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
18737 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
18738 * config/arm/arm-protos.h (arm_max_conditional_execute): New
18740 (tune_params): Update comment.
18741 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
18742 (arm_max_conditional_execute): New function.
18743 (thumb2_final_prescan_insn): Use max_insn_skipped and
18744 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
18746 2013-06-25 Jakub Jelinek <jakub@redhat.com>
18748 PR tree-optimization/57705
18749 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
18750 SSA_NAME step, provided that it is not defined inside the loop.
18751 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
18752 (get_initial_def_for_induction): Handle SSA_NAME IV step.
18754 2013-06-25 Martin Jambor <mjambor@suse.cz>
18756 PR middle-end/57670
18757 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
18758 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
18760 (ipa_note_param_call): Initialize member_ptr flag.
18761 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
18762 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
18763 (ipa_write_indirect_edge_info): Stream member_ptr flag.
18764 (ipa_read_indirect_edge_info): Likewise.
18766 2013-06-25 Richard Biener <rguenther@suse.de>
18768 PR middle-end/56977
18769 * passes.c (init_optimization_passes): Move pass_fold_builtins
18770 and pass_dce earlier with -Og.
18772 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
18774 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
18775 <BIT_FIELD_REF>: Remove trailing TAB.
18776 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
18779 2013-06-24 Martin Jambor <mjambor@suse.cz>
18781 PR tree-optimization/57358
18782 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
18783 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
18784 (ipa_analyze_params_uses): Generate pessimistic info when true.
18786 2013-06-24 Martin Jambor <mjambor@suse.cz>
18788 PR tree-optimization/57539
18789 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
18790 global.inlined_to of the new node to it. All callers changed.
18791 * ipa-inline-transform.c (clone_inlined_nodes): New variable
18792 inlining_into, pass it to cgraph_clone_node.
18793 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
18794 ipa_free_edge_args_substructures.
18795 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
18796 rdesc linked list. Do not assert rdesc edges have inlined caller.
18797 Assert we have found an rdesc in the rdesc list.
18799 2013-06-24 Richard Biener <rguenther@suse.de>
18801 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
18802 (pointer_set_lookup): Declare.
18803 (class pointer_map): New template class implementing a
18804 generic pointer to T map.
18805 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
18806 pointer_map<T>::contains, pointer_map<T>::insert,
18807 pointer_map<T>::traverse): New functions.
18808 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
18809 (pointer_set_lookup): New function.
18810 (pointer_set_contains): Use pointer_set_lookup.
18811 (pointer_set_insert): Likewise.
18812 (insert_aux): Remove.
18813 (struct pointer_map_t): Embed a pointer_set_t.
18814 (pointer_map_create): Adjust.
18815 (pointer_map_destroy): Likewise.
18816 (pointer_map_contains): Likewise.
18817 (pointer_map_insert): Likewise.
18818 (pointer_map_traverse): Likewise.
18819 * tree-streamer.h (struct streamer_tree_cache_d): Use a
18820 pointer_map<unsigned> instead of a pointer_map_t.
18821 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
18822 (streamer_tree_cache_lookup): Likewise.
18823 (streamer_tree_cache_create): Likewise.
18824 (streamer_tree_cache_delete): Likewise.
18825 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
18826 pointer_map<unsigned> instead of a pointer_map_t.
18827 (lto_init_tree_ref_encoder): Adjust.
18828 (lto_destroy_tree_ref_encoder): Likewise.
18829 * lto-section-out.c (lto_output_decl_index): Likewise.
18830 (lto_record_function_out_decl_state): Likewise.
18831 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
18833 2013-06-24 Richard Biener <rguenther@suse.de>
18835 PR tree-optimization/57488
18836 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
18838 2013-06-24 Alan Modra <amodra@gmail.com>
18840 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
18841 (gen_easy_altivec_constant): Likewise.
18842 * config/rs6000/predicates.md (easy_vector_constant_add_self,
18843 easy_vector_constant_msb): Likewise.
18845 2013-06-23 Jakub Jelinek <jakub@redhat.com>
18848 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
18849 add missing return true.
18851 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
18854 * config/sh/predicates.md (general_extend_operand): Invoke
18855 general_movsrc_operand for memory operands.
18856 (general_movsrc_operand): Allow reg+reg addressing, do not use
18857 general_operand for memory operands.
18859 2013-06-23 Sriraman Tallam <tmsriram@google.com>
18861 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
18862 when current target options does not apply.
18863 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
18864 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
18865 * config/i386/bmiintrin.h: Pass appropriate target
18866 attributes to header.
18867 * config/i386/mmintrin.h: Ditto.
18868 * config/i386/nmmintrin.h: Ditto.
18869 * config/i386/avx2intrin.h: Ditto.
18870 * config/i386/fxsrintrin.h: Ditto.
18871 * config/i386/tbmintrin.h: Ditto.
18872 * config/i386/xsaveintrin.h: Ditto.
18873 * config/i386/f16cintrin.h: Ditto.
18874 * config/i386/xtestintrin.h: Ditto.
18875 * config/i386/xsaveoptintrin.h: Ditto.
18876 * config/i386/bmi2intrin.h: Ditto.
18877 * config/i386/lzcntintrin.h: Ditto.
18878 * config/i386/smmintrin.h: Ditto.
18879 * config/i386/wmmintrin.h: Ditto.
18880 * config/i386/x86intrin.h: Remove all header include guards.
18881 * config/i386/prfchwintrin.h: Ditto.
18882 * config/i386/pmmintrin.h: Ditto.
18883 * config/i386/tmmintrin.h: Ditto.
18884 * config/i386/xmmintrin.h: Ditto.
18885 * config/i386/popcntintrin.h: Ditto.
18886 * config/i386/rdseedintrin.h: Ditto.
18887 * config/i386/ammintrin.h: Ditto.
18888 * config/i386/emmintrin.h: Ditto.
18889 * config/i386/immintrin.h: Remove all header include guards.
18890 * config/i386/fma4intrin.h: Ditto.
18891 * config/i386/lwpintrin.h: Ditto.
18892 * config/i386/xopintrin.h: Ditto.
18893 * config/i386/ia32intrin.h: Ditto.
18894 * config/i386/avxintrin.h: Ditto.
18895 * config/i386/rtmintrin.h: Ditto.
18896 * config/i386/fmaintrin.h: Ditto.
18897 * config/i386/mm3dnow.h: Ditto.
18899 2013-06-22 Sriraman Tallam <tmsriram@google.com>
18901 * common/config/i386/i386-common.c: Handle LZCNT.
18903 2013-06-22 Andi Kleen <ak@linux.intel.com>
18905 * doc/extend.texi: Use __atomic_store_n instead of
18906 __atomic_store in HLE example.
18908 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
18910 * config/sh/sh.c: Remove <cstdlib> workaround.
18912 2013-06-21 Andi Kleen <ak@linux.intel.com>
18914 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
18916 2013-06-21 Andi Kleen <ak@linux.intel.com>
18918 * doc/extend.texi: Document that __atomic_clear and
18919 __atomic_test_and_set should only be used with bool.
18921 2013-06-20 Jan Hubicka <jh@suse.cz>
18923 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
18924 types_same_for_odr.
18925 * tree.c (decls_same_for_odr): New function.
18926 (same_for_edr): New function.
18927 (types_same_for_odr): New function.
18928 (get_binfo_at_offset): Use it.
18929 * tree.h (types_same_for_odr): Declare.
18931 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
18932 Jason Merrill <jason@redhat.com>
18934 * system.h: Include <cstdlib> as well as <stdlib.h>.
18936 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
18939 * config/i386/i386.c (construct_container): Report error if
18940 long double is used with disabled x87 float returns.
18942 2013-06-20 Jan Hubicka <jh@suse.cz>
18944 * lto-cgraph.c (input_symtab): Do not set cgraph state.
18946 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
18948 PR rtl-optimization/57425
18949 PR rtl-optimization/57569
18950 * alias.c (write_dependence_p): Remove parameters mem_mode and
18951 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
18952 Changed all callers.
18953 (canon_anti_dependence): Get comments and semantics in sync.
18954 Add parameter mem_canonicalized. Changed all callers.
18955 * rtl.h (canon_anti_dependence): Update prototype.
18957 2013-06-20 Richard Biener <rguenther@suse.de>
18959 * data-streamer-in.c (streamer_read_uhwi): Optimize single
18960 byte case, inline streamer_read_uchar and defer section
18963 2013-06-20 Richard Biener <rguenther@suse.de>
18965 PR tree-optimization/57584
18966 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
18967 SSA names into the expanded expression that take part in
18968 abnormal coalescing.
18970 2013-06-19 Sharad Singhai <singhai@google.com>
18972 * gcov.c (print_usage): Handle new option.
18973 (process_args): Ditto.
18974 (get_gcov_intermediate_filename): New function.
18975 (output_intermediate_file): New function.
18976 (output_gcov_file): New function
18977 (generate_results): Handle new option.
18978 (release_function): Relase demangled name.
18979 (read_graph_file): Handle demangled name.
18980 (output_lines): Ditto.
18981 * doc/gcov.texi: Document gcov intermediate format.
18983 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
18986 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
18987 (lra_emit_add): Use the functions. Add comment about Y as an
18990 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
18993 * collect2.c (collect_atexit): New.
18994 (collect_exit): Delete.
18995 (main): Register collect_atexit with atexit.
18996 (collect_wait): Change collect_exit to exit.
18998 * collect2.h (collect_exit): Delete.
18999 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
19001 2013-06-19 Wei Mi <wmi@google.com>
19003 PR rtl-optimization/57518
19004 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
19005 if regno is used in paradoxical subreg.
19006 (update_equiv_regs): Check pdx_subregs[regno] before
19007 set a reg to be equivalent with a mem.
19009 2013-06-19 Matthias Klose <doko@ubuntu.com>
19012 * file-find.h (find_a_file): Add a mode parameter.
19013 * file-find.c (find_a_file): Likewise.
19014 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
19015 with X_OK for the executables.
19016 * collect2.c (main): Call find_a_file with X_OK.
19018 2013-06-19 Steve Ellcey <sellcey@mips.com>
19021 * config/mips/mips.md (casesi_internal_mips16_<mode>):
19022 Use NEXT_INSN instead of next_real_insn.
19024 2013-06-19 Jan Hubicka <jh@suse.cz>
19026 * cgraph.h (const_value_known_p): Replace by ...
19027 (ctor_for_folding): .. this one.
19028 * cgraphunit.c (process_function_and_variable_attributes): Use it.
19029 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
19030 * expr.c (expand_expr_real_1): Likewise.
19031 (string_constant): Likewise.
19032 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
19033 * ipa.c (process_references): Likewise.
19034 (symtab_remove_unreachable_nodes): Likewise.
19035 * ipa-inline-analysis.c (param_change_prob): Likewise.
19036 * gimple-fold.c (canonicalize_constructor_val): Likewise.
19037 (get_base_constructor): Likwise.
19038 * varpool.c (varpool_remove_node): Likewise.
19039 (varpool_remove_initializer): LIkewise.
19040 (dump_varpool_node): LIkwise.
19041 (const_value_known_p): Rewrite to ...
19042 (ctor_for_folding): ... this one.
19044 2013-06-19 Jakub Jelinek <jakub@redhat.com>
19047 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
19048 PERSONALITY in $PATH derived prefixes.
19050 2013-06-19 Jeff Law <law@redhat.com>
19052 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
19055 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
19056 (simplify_bitwise_binary): Use it to simpify certain binary ops on
19059 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
19061 * config/arm/vfp.md: Move VFP instruction classification documentation
19063 * config/arm/arm.md: ... here. Update instruction classification
19066 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
19068 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
19069 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
19070 pattern. Use more efficient sequences on ARMv5 and Thumb2.
19072 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
19075 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
19076 with NEXT_INSN. Use tablejump_p to check for jump table data
19079 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
19082 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
19083 that now in C++ the value is correct per the C++ standards.
19085 2013-06-19 Richard Biener <rguenther@suse.de>
19087 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
19088 for global context.
19090 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19093 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19096 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
19097 with next_active_insn.
19099 2013-06-18 Sriraman Tallam <tmsriram@google.com>
19101 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
19102 functions are inlined during failures to flag an error.
19103 * tree-inline.c (expand_call_inline): Allow the error to be flagged
19104 in early inline pass.
19106 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
19108 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
19111 2013-06-18 Julian Brown <julian@codesourcery.com>
19113 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
19114 Permit virtual register pre-reload if !strict.
19115 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
19117 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
19119 * config/arm/neon.md (movmisalign<mode>): Use
19120 neon_perm_struct_or_reg_operand instead of
19121 neon_struct_or_register_operand.
19122 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
19123 neon_permissive_struct_operand instead of neon_struct_operand.
19124 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
19125 neon_vector_mem_operand.
19126 * config/arm/predicates.md (neon_struct_operand): Adjust call to
19127 neon_vector_mem_operand.
19128 (neon_permissive_struct_operand): New.
19129 (neon_struct_or_register_operand): Rename to...
19130 (neon_perm_struct_or_reg_operand): This. Adjust call to
19131 neon_vector_mem_operand.
19133 2013-06-18 Richard Biener <rguenther@suse.de>
19135 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
19136 * lto-streamer.h: Include pointer-set.h.
19137 (struct lto_decl_slot): Remove.
19138 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
19139 Remove next_index entry.
19140 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
19141 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
19142 (lto_init_tree_ref_encoder): Adjust.
19143 (lto_destroy_tree_ref_encoder): Likewise.
19144 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
19145 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
19146 (lto_output_decl_index): Adjust.
19147 (lto_new_out_decl_state): Likewise.
19148 (lto_record_function_out_decl_state): Likewise.
19149 * lto-streamer-out.c (copy_function): Likewise.
19151 2013-06-18 Richard Biener <rguenther@suse.de>
19153 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
19154 * cgraphunit.c: Include cfgloop.h.
19155 (init_lowered_empty_function): Initialize the loop tree.
19156 (assemble_thunk): Insert new BBs into loops.
19158 2013-06-18 Richard Biener <rguenther@suse.de>
19160 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
19161 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
19162 the map from cache entry to cache index optional.
19163 (streamer_tree_cache_replace_tree): Adjust accordingly.
19164 (streamer_tree_cache_append): Likewise.
19165 (streamer_tree_cache_delete): Likewise.
19166 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
19167 streamer cache map from cache entry to cache index.
19168 * lto-streamer-out.c (create_output_block): Adjust.
19170 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
19172 * config/arm/arm.md (attribute "insn"): Move multiplication and
19173 division attributes to...
19174 (attribute "type"): ... here. Remove mult.
19175 (attribute "mul32"): New attribute.
19176 (attribute "mul64"): Add umaal.
19177 (*arm_mulsi3): Update attributes.
19178 (*arm_mulsi3_v6): Likewise.
19179 (*thumb_mulsi3): Likewise.
19180 (*thumb_mulsi3_v6): Likewise.
19181 (*mulsi3_compare0): Likewise.
19182 (*mulsi3_compare0_v6): Likewise.
19183 (*mulsi_compare0_scratch): Likewise.
19184 (*mulsi_compare0_scratch_v6): Likewise.
19185 (*mulsi3addsi): Likewise.
19186 (*mulsi3addsi_v6): Likewise.
19187 (*mulsi3addsi_compare0): Likewise.
19188 (*mulsi3addsi_compare0_v6): Likewise.
19189 (*mulsi3addsi_compare0_scratch): Likewise.
19190 (*mulsi3addsi_compare0_scratch_v6): Likewise.
19191 (*mulsi3subsi): Likewise.
19192 (*mulsidi3adddi): Likewise.
19193 (*mulsi3addsi_v6): Likewise.
19194 (*mulsidi3adddi_v6): Likewise.
19195 (*mulsidi3_nov6): Likewise.
19196 (*mulsidi3_v6): Likewise.
19197 (*umulsidi3_nov6): Likewise.
19198 (*umulsidi3_v6): Likewise.
19199 (*umulsidi3adddi): Likewise.
19200 (*umulsidi3adddi_v6): Likewise.
19201 (*smulsi3_highpart_nov6): Likewise.
19202 (*smulsi3_highpart_v6): Likewise.
19203 (*umulsi3_highpart_nov6): Likewise.
19204 (*umulsi3_highpart_v6): Likewise.
19205 (mulhisi3): Likewise.
19206 (*mulhisi3tb): Likewise.
19207 (*mulhisi3bt): Likewise.
19208 (*mulhisi3tt): Likewise.
19209 (maddhisi4): Likewise.
19210 (*maddhisi4tb): Likewise.
19211 (*maddhisi4tt): Likewise.
19212 (maddhidi4): Likewise.
19213 (*maddhidi4tb): Likewise.
19214 (*maddhidi4tt): Likewise.
19215 (divsi3): Likewise.
19216 (udivsi3): Likewise.
19217 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
19218 (thumb2_mulsi_short_compare0): Likewise.
19219 (thumb2_mulsi_short_compare0_scratch): Likewise.
19220 * config/arm/arm1020e.md (1020mult1): Update attribute change.
19221 (1020mult2): Likewise.
19222 (1020mult3): Likewise.
19223 (1020mult4): Likewise.
19224 (1020mult5): Likewise.
19225 (1020mult6): Likewise.
19226 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
19228 (cortex_a15_mult64): Likewise.
19229 (cortex_a15_sdiv): Likewise.
19230 (cortex_a15_udiv): Likewise.
19231 * config/arm/arm1026ejs.md (mult1): Update attribute change.
19237 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
19238 (pj4_ir_div): Likewise.
19239 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
19240 (11_mult2): Likewise.
19241 (11_mult3): Likewise.
19242 (11_mult4): Likewise.
19243 (11_mult5): Likewise.
19244 (11_mult6): Likewise.
19245 (11_mult7): Likewise.
19246 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
19247 (cortex_a8_mla): Likewise.
19248 (cortex_a8_mull): Likewise.
19249 (cortex_a8_smulwy): Likewise.
19250 (cortex_a8_smlald): Likewise.
19251 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
19252 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
19253 (cortex_r4_mul_3): Likewise.
19254 (cortex_r4_mla_4): Likewise.
19255 (cortex_r4_mla_3): Likewise.
19256 (cortex_r4_smlald): Likewise.
19257 (cortex_r4_mull): Likewise.
19258 (cortex_r4_sdiv): Likewise.
19259 (cortex_r4_udiv): Likewise.
19260 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
19261 (cortex_a7_idiv): Likewise.
19262 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
19263 (9_mult2): Likewise.
19264 (9_mult3): Likewise.
19265 (9_mult4): Likewise.
19266 (9_mult5): Likewise.
19267 (9_mult6): Likewise.
19268 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
19269 (cortex_a53_sdiv): Likewise.
19270 (cortex_a53_udiv): Likewise.
19271 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
19272 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
19273 (mp626_mult2): Likewise.
19274 (mp626_mult3): Likewise.
19275 (mp626_mult4): Likewise.
19276 * config/arm/fa526.md (526_mult1): Update attribute change.
19277 (526_mult2): Likewise.
19278 * config/arm/arm-generic.md (mult): Update attribute change.
19279 (mult_ldsched_strongarm): Likewise.
19280 (mult_ldsched): Likewise.
19281 (multi_cycle): Likewise.
19282 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
19283 * config/arm/fa606te.md (606te_mult1): Update attribute change.
19284 (606te_mult2): Likewise.
19285 (606te_mult3): Likewise.
19286 (606te_mult4): Likewise.
19287 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
19288 (cortex_a9_mac16): Likewise.
19289 (cortex_a9_multiply): Likewise.
19290 (cortex_a9_mac): Likewise.
19291 (cortex_a9_multiply_long): Likewise.
19292 * config/arm/fa626te.md (626te_mult1): Update attribute change.
19293 (626te_mult2): Likewise.
19294 (626te_mult3): Likewise.
19295 (626te_mult4): Likewise.
19297 2013-06-18 Richard Biener <rguenther@suse.de>
19300 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
19302 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19305 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
19306 with next_active_insn.
19308 2013-06-18 Alan Modra <amodra@gmail.com>
19310 * config/rs6000/rs6000.h (enum data_align): New.
19311 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
19312 (DATA_ABI_ALIGNMENT): Define.
19313 (CONSTANT_ALIGNMENT): Correct comment.
19314 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
19315 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
19317 2013-06-17 David Malcolm <dmalcolm@redhat.com>
19319 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
19320 ATTRIBUTE_UNUSED marking.
19322 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
19324 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
19325 alternative and update.
19326 (aarch64_dup_lanedi): Delete.
19327 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
19328 * config/aarch64/aarch64-simd-builtins.def: Update.
19330 2013-06-17 Richard Biener <rguenther@suse.de>
19332 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
19333 (lto_input_scc): Declare.
19334 (lto_input_tree_1): Likewise.
19335 (struct lto_stats_d): Add num_tree_bodies_output and
19336 num_pickle_refs_output.
19337 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
19338 (lto_read_tree_1): Split out from ...
19339 (lto_read_tree): ... this.
19340 (lto_input_scc): New function.
19341 (lto_input_tree_1): Split out from ...
19342 (lto_input_tree): ... this. Handle LTO_tree_scc.
19343 (lto_data_in_create): Create the streamer cache without hashes.
19344 * lto-streamer-out.c (create_output_block): Create the streamer
19345 cache with hashes when not doing WPA.
19346 (lto_write_tree_1): Split out from ...
19347 (lto_write_tree): ... this.
19348 (get_symbol_initial_value): New function.
19349 (lto_output_tree_1): Split out from ...
19350 (lto_output_tree): ... this. Write trees as series of SCCs
19351 using a DFS walk via DFS_write_tree.
19352 (struct sccs, struct scc_entry): New types.
19353 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
19354 (DFS_write_tree_body): New function.
19355 (DFS_write_tree): Likewise.
19356 (hash_tree): Likewise.
19357 (scc_entry_compare): Likewise.
19358 (hash_scc): Likewise.
19359 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
19360 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
19361 TREE_CHAIN as regular reference.
19362 (streamer_read_integer_cst): Remove.
19363 (streamer_get_pickled_tree): Adjust.
19364 * tree-streamer-out.c (streamer_write_chain): Disable streaming
19365 of DECL_EXTERNALs in BLOCK_VARS for now.
19366 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
19368 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
19369 Add hash value argument and record that if hashes are recorded
19371 (streamer_tree_cache_insert_1): Adjust.
19372 (streamer_tree_cache_insert): Likewise.
19373 (streamer_tree_cache_insert_at): Rename to ...
19374 (streamer_tree_cache_replace_tree): ... this and adjust.
19375 (streamer_tree_cache_append): Adjust.
19376 (record_common_node): Likewise.
19377 (streamer_tree_cache_create): Add argument whether to
19378 record hash values together with trees.
19379 (streamer_tree_cache_delete): Adjust.
19380 * tree-streamer.h (struct streamer_tree_cache_d): Add
19382 (streamer_read_integer_cst): Remove.
19383 (streamer_tree_cache_insert): Adjust.
19384 (streamer_tree_cache_append): Likewise.
19385 (streamer_tree_cache_insert_at): Rename to ...
19386 (streamer_tree_cache_replace_tree): ... this and adjust.
19387 (streamer_tree_cache_create): Add argument whether to record hashes.
19388 (streamer_tree_cache_get): Rename to ...
19389 (streamer_tree_cache_get_tree): ... this.
19390 (streamer_tree_cache_get_hash): New function.
19391 * tree.c (cache_integer_cst): New function.
19392 * tree.h (cache_integer_cst): Declare.
19393 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
19394 * lto-symtab.c (lto_varpool_replace_node): Only release
19395 DECL_INITIAL of non-prevailing decls.
19396 * varpool.c (varpool_remove_initializer): Do not release
19397 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
19399 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
19401 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
19402 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
19403 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
19404 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
19405 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
19406 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
19407 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
19408 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
19409 instead of TARGET_64BIT.
19410 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
19411 Require ISA_HAS_<D>DIV.
19413 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
19415 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
19416 (mips*-*-linux*): Move default with_llsc setting to where other
19418 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
19420 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
19421 Likewise. Remove default with_tune setting. Move default float
19422 setting to its own block. Handle with_llsc in the same block as above.
19424 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
19426 PR rtl-optimization/57425
19427 PR rtl-optimization/57569
19428 * alias.c (write_dependence_p): Add new parameters mem_mode,
19429 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
19430 Changed all callers.
19431 (canon_anti_dependence): New function.
19432 * cse.c (check_dependence): Use canon_anti_dependence.
19433 * cselib.c (cselib_invalidate_mem): Likewise.
19434 * rtl.h (canon_anti_dependence): Declare.
19436 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
19438 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
19439 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
19440 ".set mips3" for 64-bit targets.
19442 2013-06-15 Dehao Chen <dehao@google.com>
19444 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
19445 * gimple-low.c (gimple_check_call_matching_types): Likewise.
19446 (gimple_check_call_args): Likewise.
19447 * value-prof.c (check_ic_target): Likewise.
19448 * ipa-inline.c (early_inliner): Likewise.
19449 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
19450 * cgraph.c (cgraph_create_edge_1): Likewise.
19451 (cgraph_make_edge_direct): Likewise.
19453 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
19456 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
19457 rs6000_output_move_128bit to handle emitting quad memory
19458 operations. Set attribute length to 8 bytes.
19460 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
19462 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
19464 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
19465 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
19466 (aarch64_<su>mlsl<mode>): Likewise.
19468 2013-06-14 Mike Stump <mikestump@comcast.net>
19470 * Makefile.in (TARGET_H): Add insn-codes.h.
19472 2013-06-14 Alan Modra <amodra@gmail.com>
19474 PR middle-end/57134
19475 PR middle-end/57586
19476 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
19477 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
19478 bitfield expansion when EXPAND_MEMORY.
19479 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
19481 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
19483 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
19484 test for clearing quad memory on 32-bit later.
19486 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
19488 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
19489 (fold_negate_expr): Likewise.
19490 (fold_real_zero_addition_p): Handle vectors.
19491 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
19493 2013-06-14 Alan Modra <amodra@gmail.com>
19495 * varasm.c (force_const_mem): Revert 2013-06-07 change.
19497 2013-06-13 Jan Hubicka <jh@suse.cz>
19499 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
19500 Local comdats are not externally visible.
19501 * symtab.c (dump_symtab_base): Dump externally visible.
19502 (verify_symtab_base): Verify back links in the symtab hash.
19504 2013-06-13 Bin Cheng <bin.cheng@arm.com>
19506 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
19507 CONVERT_EXPR as equal nodes.
19509 2013-06-13 Bin Cheng <bin.cheng@arm.com>
19511 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
19513 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
19515 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
19516 Generalize to complex and vector.
19517 * tree.c (build_all_ones_cst): New function.
19518 * tree.h (build_all_ones_cst): Declare it.
19520 2013-06-13 Alan Modra <amodra@gmail.com>
19522 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
19523 * config/rs6000/rs6000.md (signbittf2): New insn.
19524 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
19525 (abstf2_internal, cmptf_internal2): Likewise.
19526 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
19528 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
19529 Pat Haugen <pthaugen@us.ibm.com>
19530 Peter Bergner <bergner@vnet.ibm.com>
19532 * config/rs6000/rs6000.c (emit_load_locked): Add support for
19533 power8 byte, half-word, and quad-word atomic instructions.
19534 (emit_store_conditional): Likewise.
19535 (rs6000_expand_atomic_compare_and_swap): Likewise.
19536 (rs6000_expand_atomic_op): Likewise.
19538 * config/rs6000/sync.md (larx): Add new modes for power8.
19540 (AINT): New mode iterator to include TImode as well as normal
19541 integer modes on power8.
19542 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
19543 that VSX registers are not considered. Use AINT mode iterator
19544 instead of INT1 to allow inclusion of quad word atomic operations
19546 (load_locked<mode>): Likewise.
19547 (store_conditional<mode>): Likewise.
19548 (atomic_compare_and_swap<mode>): Likewise.
19549 (atomic_exchange<mode>): Likewise.
19550 (atomic_nand<mode>): Likewise.
19551 (atomic_fetch_<fetchop_name><mode>): Likewise.
19552 (atomic_nand_fetch<mode>): Likewise.
19553 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
19555 (ATOMIC): On power8, add QImode, HImode modes.
19556 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
19557 modes that promote to SImode.
19558 (load_lockedti): Convert TImode arguments to PTImode, so that we
19559 get a guaranteed even/odd register pair.
19560 (load_lockedpti): Likewise.
19561 (store_conditionalti): Likewise.
19562 (store_conditionalpti): Likewise.
19564 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
19565 atomic load/store instructions.
19568 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
19570 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
19572 (insn_count): New attribute, with most cases extracted from...
19573 (length): ...here. Redefine most cases in terms of insn_count.
19574 (single_insn): Delete.
19575 (can_delay): Use insn_count to check for single instructions.
19576 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
19577 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
19578 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
19579 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
19580 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
19581 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
19582 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
19583 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
19584 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
19585 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
19586 rather than "length".
19587 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
19588 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
19589 Use "insn_count" rather than "length".
19590 * config/mips/mips-dsp.md
19591 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
19592 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
19595 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
19597 PR tree-optimization/57361
19598 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
19600 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
19602 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
19604 (aarch64_simd_combine<mode>): New instruction expansion.
19605 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
19606 function prototype.
19607 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
19608 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
19610 2013-06-12 Jan Hubicka <jh@suse.cz>
19612 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
19613 decl has when in streaming stage.
19614 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
19615 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
19617 2013-06-12 Roland Stigge <stigge@antcom.de>
19620 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
19622 2013-06-12 Jakub Jelinek <jakub@redhat.com>
19624 PR tree-optimization/57537
19625 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
19626 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
19628 2013-06-12 Richard Biener <rguenther@suse.de>
19630 * data-streamer.h (streamer_write_char_stream): CSE
19631 obs->current_pointer.
19632 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
19633 streamer_write_char_stream manually and optimize the resulting loop.
19634 (streamer_write_hwi_stream): Likewise.
19636 2013-06-12 Jan Hubicka <jh@suse.cz>
19638 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
19639 * cgraph.h (varpool_create_empty_node): Declare.
19640 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
19642 * symtab.c (symtab_unregister_node): Be lax about missin entries
19644 (symtab_get_node): Update comment.
19645 * varpool.c (varpool_create_empty_node): Break out from ...
19646 (varpool_node_for_decl): ... here.
19647 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
19649 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
19651 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
19652 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
19653 part. Use straight-line flow at the end.
19654 <COMPONENT_REF>: Remove superfluous else.
19655 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
19657 2013-06-12 Jakub Jelinek <jakub@redhat.com>
19660 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
19661 target hook even for !TREE_PUBLIC decls. If no resolution info
19662 is available, return false for common and external decls.
19664 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
19666 * config/rl78/constraints.md (U): New constraint.
19667 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
19670 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
19673 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
19674 to allow returning address to AT_PLATFORM name.
19676 2013-06-11 Jan Hubicka <jh@suse.cz>
19678 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
19679 * cgraph.h (symtab_node_base): Add weakref flag.
19680 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
19681 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
19682 (output_weakrefs): Use weakref flag.
19683 * fold-const.c (simple_operand_p): Handle WEAK.
19684 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
19685 * ipa.c (varpool_externally_visible_p): Drop weakref.
19686 (function_and_variable_visibility): Update comment; fix weakref
19687 sanity checks; do not clear DECL_WEAK on them.
19688 * lto-cgraph.c (lto_output_node): update.
19689 (lto_output_varpool_node): Update.
19690 (input_overwrite_node): Update.
19691 (input_node): Update.
19692 (input_varpool_node): Update.
19693 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
19694 (lto_symtab_merge_symbols): Add sanity check.
19695 (lto_symtab_prevailing_decl): Do not special case weakrefs.
19696 * passes.c (rest_of_decl_compilation): Set static flag, too.
19697 * symtab.c (dump_symtab_base): Dump weakref.
19698 (verify_symtab_base): Sanity check weakrefs.
19699 (symtab_make_decl_local): Remove duplicated code.
19700 (symtab_alias_ultimate_target): Simplify.
19701 * varpool.c (varpool_create_variable_alias): Set weakref flag.
19703 2013-06-11 Tom de Vries <tom@codesourcery.com>
19705 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
19706 sequence_vect == NULL.
19708 2013-06-11 DJ Delorie <dj@redhat.com>
19710 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
19711 (rl78_unwind_word_mode): New.
19713 2013-06-11 David Malcolm <dmalcolm@redhat.com>
19715 * final.c (debug_prefix_maps): Make static.
19717 2013-06-11 David Malcolm <dmalcolm@redhat.com>
19719 * function.c (initial_trampoline): Remove stray copy.
19721 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
19723 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
19725 2013-06-11 Martin Jambor <mjambor@suse.cz>
19727 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
19728 within bounds at the beginning of the function.
19730 2013-06-11 Alan Modra <amodra@gmail.com>
19732 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
19734 (get_named_section): Don't NULL !DECL_P decl.
19736 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
19738 * doc/invoke.texi (core-avx2): Document.
19740 (atom): Updated with MOVBE.
19742 2013-06-11 Richard Biener <rguenther@suse.de>
19744 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
19746 2013-06-11 Anton Blanchard <anton@samba.org>
19748 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
19749 correct shift value in little-endian mode.
19751 2013-06-11 Jakub Jelinek <jakub@redhat.com>
19754 * varasm.c (get_variable_align): Move #endif to the right place.
19756 2013-06-10 Cary Coutant <ccoutant@google.com>
19758 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
19759 for hash so that hash table traversal order is deterministic.
19761 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
19762 Pat Haugen <pthaugen@us.ibm.com>
19763 Peter Bergner <bergner@vnet.ibm.com>
19765 * config/rs6000/vector.md (GPR move splitter): Do not split moves
19766 of vectors in GPRS if they are direct moves or quad word load or
19769 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
19771 (direct_move_p): Likewise.
19772 (quad_load_store_p): Likewise.
19774 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
19775 classes into bins based on the physical register type.
19776 (reg_class_to_reg_type): Likewise.
19777 (IS_STD_REG_TYPE): Likewise.
19778 (IS_FP_VECT_REG_TYPE): Likewise.
19779 (reload_fpr_gpr): Arrays to determine what insn to use if we can
19780 use direct move instructions.
19781 (reload_gpr_vsx): Likewise.
19782 (reload_vsx_gpr): Likewise.
19783 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
19784 information that is a simplification of register classes. Also
19785 precalculate direct move reload helpers.
19786 (direct_move_p): New function to return true if the operation can
19787 be done as a direct move instruciton.
19788 (quad_load_store_p): New function to return true if the operation
19789 is a quad memory operation.
19790 (rs6000_legitimize_address): If quad memory, only allow register
19791 indirect for TImode addresses.
19792 (rs6000_legitimate_address_p): Likewise.
19793 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
19794 (rs6000_reload_register_type): Likewise.
19795 (register_to_reg_type): Return register type.
19796 (rs6000_secondary_reload_simple_move): New helper function for
19797 secondary reload and secondary memory needed to identify anything
19798 that is a simple move, and does not need reloading.
19799 (rs6000_secondary_reload_direct_move): New helper function for
19800 secondary reload to identify cases that can be done with several
19801 instructions via the direct move instructions.
19802 (rs6000_secondary_reload_move): New helper function for secondary
19803 reload to identify moves between register types that can be done.
19804 (rs6000_secondary_reload): Add support for quad memory operations
19805 and for direct move.
19806 (rs6000_secondary_memory_needed): Likewise.
19807 (rs6000_debug_secondary_memory_needed): Change argument names.
19808 (rs6000_output_move_128bit): New function to return the move to
19809 use for 128-bit moves, including knowing about the various
19810 limitations of quad memory operations.
19812 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
19813 memory operations. call rs6000_output_move_128bit for the actual
19814 instruciton(s) to generate.
19815 (vsx_movti_64bit): Likewise.
19817 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
19818 (UNSPEC_P8V_MTVSRWZ): Likewise.
19819 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
19820 (UNSPEC_P8V_MTVSRD): Likewise.
19821 (UNSPEC_P8V_XXPERMDI): Likewise.
19822 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
19823 (UNSPEC_FUSION_GPR): Likewise.
19824 (FMOVE128_GPR): New iterator for direct move.
19825 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
19826 (f32_sv): Likewise.
19827 (f32_dm): Likewise.
19828 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
19829 loads and direct move instructions.
19830 (zero_extendsidi2_lfiwzx): Likewise.
19831 (extendsidi2_lfiwax): Likewise.
19832 (extendsidi2_nocell): Likewise.
19833 (floatsi<mode>2_lfiwax): Likewise.
19834 (lfiwax): Likewise.
19835 (floatunssi<mode>2_lfiwzx): Likewise.
19836 (lfiwzx): Likewise.
19837 (fix_trunc<mode>_stfiwx): Likewise.
19838 (fixuns_trunc<mode>_stfiwx): Likewise.
19839 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
19840 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
19841 (parity<mode>2_cmpb): Set length/type attr.
19842 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
19843 for 'mr.' to fast_compare.
19844 (bpermd_<mode>): Change type attr to popcnt.
19845 (p8_fmrgow_<mode>): New insns for power8 direct move support.
19846 (p8_mtvsrwz_1): Likewise.
19847 (p8_mtvsrwz_2): Likewise.
19848 (reload_fpr_from_gpr<mode>): Likewise.
19849 (p8_mtvsrd_1): Likewise.
19850 (p8_mtvsrd_2): Likewise.
19851 (p8_xxpermdi_<mode>): Likewise.
19852 (reload_vsx_from_gpr<mode>): Likewise.
19853 (reload_vsx_from_gprsf): Likewise.
19854 (p8_mfvsrd_3_<mode>): LIkewise.
19855 (reload_gpr_from_vsx<mode>): Likewise.
19856 (reload_gpr_from_vsxsf): Likewise.
19857 (p8_mfvsrd_4_disf): Likewise.
19858 (multi-word GPR splits): Do not split direct moves or quad memory
19861 2013-06-10 David Malcolm <dmalcolm@redhat.com>
19863 * tree-into-ssa.c (interesting_blocks): Make static.
19865 2013-06-10 Jakub Jelinek <jakub@redhat.com>
19868 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
19869 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
19870 Use DATA_ABI_ALIGNMENT for that case instead if defined.
19871 (get_variable_align): New function.
19872 (get_variable_section, emit_bss, emit_common,
19873 assemble_variable_contents, place_block_symbol): Use
19874 get_variable_align instead of DECL_ALIGN.
19875 (assemble_noswitch_variable): Add align argument, use it
19876 instead of DECL_ALIGN.
19877 (assemble_variable): Adjust caller. Use get_variable_align
19878 instead of DECL_ALIGN.
19879 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
19881 (DATA_ABI_ALIGNMENT): Define.
19882 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
19883 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
19884 opt is false, only return the psABI mandated alignment increase.
19885 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
19886 (DATA_ABI_ALIGNMENT): ... this.
19887 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
19888 (DATA_ABI_ALIGNMENT): ... this.
19889 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
19890 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
19891 (DATA_ABI_ALIGNMENT): ... this.
19892 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
19893 * doc/tm.texi: Regenerated.
19895 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
19897 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
19898 cmp_code to construct REG_EQUAL note.
19900 2013-06-09 Jakub Jelinek <jakub@redhat.com>
19903 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
19904 that operands[2] doesn't overlap with operands[0].
19906 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
19907 Jan Hubicka <jh@suse.cz>
19909 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
19910 hack to mark symbols as used.
19912 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
19914 PR rtl-optimization/57559
19915 * lra-constraints.c (process_alt_operands): Don't discourage
19916 memory with known offset for offsetable memory constraint.
19917 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
19919 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
19921 * varasm.c (struct oc_local_state): Reorder fields.
19922 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
19923 and adjust accordingly.
19924 (output_constructor): Reorder initialization code and adjust call to
19925 output_constructor_bitfield.
19927 2013-06-07 Jan Hubicka <jh@suse.cz>
19929 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
19931 2013-06-07 David Malcolm <dmalcolm@redhat.com>
19933 * tree-object-size.c (unknown): Make const.
19935 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19937 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
19938 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
19939 for last alternative in the cpu_facility attribute.
19941 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19944 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
19945 (xordi3): Change operand 2 constraint to arm_xordi_operand.
19946 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
19947 * config/arm/constraints.md (Dg): New constraint.
19948 * config/arm/neon.md (xordi3_neon): Remove.
19949 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
19950 * config/arm/predicates.md (arm_xordi_operand): New predicate.
19952 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19954 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
19955 Clean up alternatives.
19957 2013-06-07 Alan Modra <amodra@gmail.com>
19959 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
19962 2013-06-07 Alan Modra <amodra@gmail.com>
19964 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
19966 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19968 * config/arm/constraints.md (Df): New constraint.
19969 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
19970 Correct length attribute for last two alternatives.
19972 2013-06-07 Alan Modra <amodra@gmail.com>
19974 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
19975 override user -mfp-in-toc.
19976 (offsettable_ok_by_alignment): Consider just the current access
19977 rather than the whole object, unless BLKmode. Handle
19978 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
19979 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
19980 for -mcmodel=medium.
19981 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
19982 override user -mfp-in-toc or -msum-in-toc. Default to
19983 -mno-fp-in-toc for -mcmodel=medium.
19985 2013-06-06 DJ Delorie <dj@redhat.com>
19987 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
19988 TARGET_VALID_POINTER_MODE.
19990 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
19991 Pat Haugen <pthaugen@us.ibm.com>
19992 Peter Bergner <bergner@vnet.ibm.com>
19994 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
19995 Document new power8 builtins.
19997 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
19998 condition code register, to allow 128-bit logical operations to be
19999 done in the VSX or GPR registers.
20000 (nor<mode>3): Use the canonical form for nor.
20001 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
20002 vclz*, and vpopcnt* vector instructions.
20003 (nand<mode>3): Likewise.
20004 (orc<mode>3): Likewise.
20005 (clz<mode>2): LIkewise.
20006 (popcount<mode>2): Likewise.
20008 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
20009 that only the GPRs are recognized.
20011 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
20012 support for new power8 builtins.
20014 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
20016 (xscvdpspn): Likewise.
20022 (vpopcnt): Likewise.
20023 (vpopcntb): Likewise.
20024 (vpopcnth): Likewise.
20025 (vpopcntw): Likewise.
20026 (vpopcntd): Likewise.
20028 (vmrgew): Likewise.
20029 (vmrgow): Likewise.
20031 (eqv_v16qi3): Likewise.
20032 (eqv_v8hi3): Likewise.
20033 (eqv_v4si3): Likewise.
20034 (eqv_v2di3): Likewise.
20035 (eqv_v4sf3): Likewise.
20036 (eqv_v2df3): Likewise.
20038 (nand_v16qi3): Likewise.
20039 (nand_v8hi3): Likewise.
20040 (nand_v4si3): Likewise.
20041 (nand_v2di3): Likewise.
20042 (nand_v4sf3): Likewise.
20043 (nand_v2df3): Likewise.
20045 (orc_v16qi3): Likewise.
20046 (orc_v8hi3): Likewise.
20047 (orc_v4si3): Likewise.
20048 (orc_v2di3): Likewise.
20049 (orc_v4sf3): Likewise.
20050 (orc_v2df3): Likewise.
20052 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
20053 allow power8 quad mode in 64-bit.
20054 (rs6000_builtin_vectorized_function): Add support to vectorize
20055 ISA 2.07 count leading zeros, population count builtins.
20056 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
20057 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
20058 (builtin_function_type): Add vgbbd builtin function which takes an
20060 (altivec_expand_vec_perm_const): Add support for new power8 merge
20063 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
20064 that does not include TImdoe for use with 32-bit.
20065 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
20067 (UNSPEC_VSX_CVDPSPN): Likewise.
20068 (vsx_xscvdpspn): Likewise.
20069 (vsx_xscvspdpn): Likewise.
20070 (vsx_xscvdpspn_scalar): Likewise.
20071 (vsx_xscvspdpn_directmove): Likewise.
20072 (vsx_and<mode>3): Split logical operations into 32-bit and
20073 64-bit. Add support to do logical operations on TImode as well as
20074 VSX vector types. Allow logical operations to be done in either
20075 VSX registers or in general purpose registers in 64-bit mode. Add
20076 splitters if GPRs were used. For AND, add clobber of CCmode to
20077 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
20079 (vsx_and<mode>3_32bit): Likewise.
20080 (vsx_and<mode>3_64bit): Likewise.
20081 (vsx_ior<mode>3): Likewise.
20082 (vsx_ior<mode>3_32bit): Likewise.
20083 (vsx_ior<mode>3_64bit): Likewise.
20084 (vsx_xor<mode>3): Likewise.
20085 (vsx_xor<mode>3_32bit): Likewise.
20086 (vsx_xor<mode>3_64bit): Likewise.
20087 (vsx_one_cmpl<mode>2): Likewise.
20088 (vsx_one_cmpl<mode>2_32bit): Likewise.
20089 (vsx_one_cmpl<mode>2_64bit): Likewise.
20090 (vsx_nor<mode>3): Likewise.
20091 (vsx_nor<mode>3_32bit): Likewise.
20092 (vsx_nor<mode>3_64bit): Likewise.
20093 (vsx_andc<mode>3): Likewise.
20094 (vsx_andc<mode>3_32bit): Likewise.
20095 (vsx_andc<mode>3_64bit): Likewise.
20096 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
20097 and xxlorc instructions.
20098 (vsx_eqv<mode>3_64bit): Likewise.
20099 (vsx_nand<mode>3_32bit): Likewise.
20100 (vsx_nand<mode>3_64bit): Likewise.
20101 (vsx_orc<mode>3_32bit): Likewise.
20102 (vsx_orc<mode>3_64bit): Likewise.
20104 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
20106 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
20108 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
20109 (p8_vmrgow): Likewise.
20110 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
20111 GPRs to be split under VSX.
20112 (p8v_clz<mode>2): Add power8 count leading zero support.
20113 (p8v_popcount<mode>2): Add power8 population count support.
20114 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
20117 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
20120 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
20122 (vec_nand): Likewise.
20123 (vec_vclz): Likewise.
20124 (vec_vclzb): Likewise.
20125 (vec_vclzd): Likewise.
20126 (vec_vclzh): Likewise.
20127 (vec_vclzw): Likewise.
20128 (vec_vgbbd): Likewise.
20129 (vec_vmrgew): Likewise.
20130 (vec_vmrgow): Likewise.
20131 (vec_vpopcnt): Likewise.
20132 (vec_vpopcntb): Likewise.
20133 (vec_vpopcntd): Likewise.
20134 (vec_vpopcnth): Likewise.
20135 (vec_vpopcntw): Likewise.
20137 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
20139 PR rtl-optimization/57468
20140 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
20143 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
20145 PR rtl-optimization/57459
20146 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
20147 type when setting live regs.
20149 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
20151 * config/s390/s390.opt (mlra): New option.
20152 * config/s390/s390.c (s390_decompose_address): Check displacement
20153 for all registers for LRA.
20154 (s390_secondary_reload): Don't used secondary reloads for LRA.
20155 (s390_lra_p): New function.
20156 (TARGET_LRA_P): Define.
20157 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
20158 of attribute cpu_facility to zarch for the last alternative.
20159 (*cmpmem_short): Ditto.
20161 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
20163 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
20164 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
20165 (arm_expand_prologue): Likewise.
20167 2013-06-06 Teresa Johnson <tejohnson@google.com>
20170 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
20171 as this is now done by redirect_edge_and_branch_force.
20172 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
20173 barriers, and fix interaction with splitting.
20174 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
20175 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
20176 reflect changes made in the routine.
20177 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
20178 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
20179 since this is called in cfglayout mode, and replace partition fixup
20180 with assert as that is now done by force_nonfallthru_and_redirect.
20181 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
20182 already be marked with region crossing note.
20183 (insert_section_boundary_note): Make non-static, gate on flag
20184 has_bb_partition, rewrite to also check for multiple partitions.
20185 (rest_of_handle_reorder_blocks): Remove call to
20186 insert_section_boundary_note, now done later during free_cfg.
20187 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
20188 * bb-reorder.h (insert_section_boundary_note): Declare.
20189 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
20190 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
20191 invoke insert_section_boundary_note.
20192 (try_redirect_by_replacing_jump): Remove unnecessary
20193 check for region crossing note.
20194 (fixup_partition_crossing): New function.
20195 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
20196 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
20197 in non-cfglayout mode.
20198 (force_nonfallthru_and_redirect): Fixup partition boundaries,
20199 remove old code that tried to do this. Emit barrier correctly
20200 when we are in cfglayout mode.
20201 (last_bb_in_partition): New function.
20202 (rtl_split_edge): Correctly fixup partition boundaries.
20203 (commit_one_edge_insertion): Remove old code that tried to
20204 fixup region crossing edge since this is now handled in
20205 split_block, and set up insertion point correctly since
20206 block may now end in a jump.
20207 (verify_hot_cold_block_grouping): Guard against checking when not in
20208 linearized RTL mode.
20209 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
20211 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
20212 rtl_verify_flow_info, so not called in cfglayout mode.
20213 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
20214 (fixup_reorder_chain): Remove old code that attempted to fixup region
20215 crossing note as this is now handled in force_nonfallthru_and_redirect.
20216 (duplicate_insn_chain): Don't duplicate switch section notes.
20217 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
20219 * basic-block.h (emit_barrier_after_bb): Declare.
20221 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20223 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
20224 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
20225 arm_usatsihi): Adjust alternatives for arm_restrict_it.
20227 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20229 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
20231 * config/arm/ldmstm.md: Regenerate.
20233 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20235 * config/arm/sync.md (atomic_loaddi_1):
20236 Disable predication for arm_restrict_it.
20237 (arm_load_exclusive<mode>): Likewise.
20238 (arm_load_exclusivesi): Likewise.
20239 (arm_load_exclusivedi): Likewise.
20240 (arm_load_acquire_exclusive<mode>): Likewise.
20241 (arm_load_acquire_exclusivesi): Likewise.
20242 (arm_load_acquire_exclusivedi): Likewise.
20243 (arm_store_exclusive<mode>): Likewise.
20244 (arm_store_exclusive<mode>): Likewise.
20245 (arm_store_release_exclusivedi): Likewise.
20246 (arm_store_release_exclusive<mode>): Likewise.
20248 2013-06-06 Richard Biener <rguenther@suse.de>
20250 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
20252 (lto_tag_is_tree_code_p): Adjust.
20253 (lto_tag_is_gimple_code_p): Likewise.
20254 (lto_gimple_code_to_tag): Likewise.
20255 (lto_tag_to_gimple_code): Likewise.
20256 (lto_tree_code_to_tag): Likewise.
20257 (lto_tag_to_tree_code): Likewise.
20258 * data-streamer.h (streamer_write_hwi_in_range): Use
20259 uhwi streaming to stream the normalized range.
20260 (streamer_read_hwi_in_range): Likewise.
20262 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20264 * config/arm/arm.md (enabled_for_depr_it): New attribute.
20265 (predicable_short_it): Likewise.
20266 (predicated): Likewise.
20267 (enabled): Handle above.
20268 (define_cond_exec): Set predicated attribute to yes.
20270 2013-06-05 Mike Stump <mikestump@comcast.net>
20272 * gdbinit.in (__FUNCTION__): Add.
20274 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
20276 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
20277 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
20279 2013-06-05 Jan Hubicka <jh@suse.cz>
20281 * varasm.c (mark_decl_referenced): Revert the removal until targets
20284 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
20286 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
20287 instead of mark_decl_referenced.
20289 2013-06-05 Jan Hubicka <jh@suse.cz>
20291 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
20292 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
20293 and symtab_used_from_object_file_p.
20294 (cgraph_make_node_local_1): Clear forced_by_abi.
20295 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
20296 * cgraph.h (symtab_node_base): Add forced_by_abi.
20297 (decide_is_variable_needed): Remove.
20298 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
20299 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
20300 (decide_is_symbol_needed): ... this one; handle symbols in general;
20301 always analyze virtuals; honnor forced_by_abi.
20302 (cgraph_finalize_function): Update.
20303 (varpool_finalize_decl): Update.
20304 (symbol_defined_and_needed): Remove.
20305 (analyze_functions): Update.
20306 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
20307 output_refs, input_overwrite_node): Handle forced_by_abi.
20308 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
20309 (address_taken_from_non_vtable_p): ... this one.
20310 (comdat_can_be_unshared_p_1): New function.
20311 (cgraph_comdat_can_be_unshared_p): Rename to ...
20312 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
20313 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
20314 (function_and_variable_visibility): Clear forced_by_abi as needed.
20315 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
20316 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
20317 * symtab.c (dump_symtab_base): Dump forced_by_abi.
20318 * varpool.c (decide_is_variable_needed): Remove.
20320 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20322 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
20323 (arm_option_override): Override arm_restrict_it where appropriate.
20324 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
20325 * config/arm/arm.opt (mrestrict-it): New command-line option.
20326 * doc/invoke.texi: Document -mrestrict-it.
20328 2013-06-05 David Malcolm <dmalcolm@redhat.com>
20330 * tsan.c (tsan_atomic_table): Make const.
20332 2013-06-05 Richard Biener <rguenther@suse.de>
20334 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
20335 index associated with the tree we are supposed to replace.
20336 * tree-streamer-out.c (pack_ts_base_value_fields): Output
20337 TREE_ASM_WRITTEN as zero for everything but SSA names.
20339 2013-06-05 David Malcolm <dmalcolm@redhat.com>
20341 * tree-ssa-structalias.c (call_stmt_vars): Make static.
20343 2013-06-04 Jan Hubicka <jh@suse.cz>
20345 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
20346 (input_node, input_varpool_node): Handle correctly external same
20348 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
20349 nodes at ltrans stage.
20351 2013-06-04 Jan Hubicka <jh@suse.cz>
20353 * ipa-inline.c (update_caller_keys): Fix availability test.
20354 (update_callee_keys): Likewise.
20355 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
20356 to follow ELF standard.
20358 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
20360 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
20361 (mips64r5900el-*-elf*): New configurations.
20362 * config/mips/mips-cpus.def (r5900): New processor.
20363 * config/mips/mips-tables.opt: Regenerate.
20364 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
20365 (mips_issue_rate): Handle PROCESSOR_R5900.
20366 (mips_reorg_process_insns): Force reorder mode for the R5900.
20367 * config/mips/mips.h (TARGET_MIPS5900): Define.
20368 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
20370 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
20372 * config/mips/mips.md (processor): Add r5900.
20373 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
20375 2013-06-04 Ian Bolton <ian.bolton@arm.com>
20377 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
20378 into function to generate MOVI instruction.
20379 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
20380 (aarch64_preferred_simd_mode): Turn into wrapper.
20381 (aarch64_output_scalar_simd_mov_immediate): New function.
20382 * config/aarch64/aarch64-protos.h: Add prototype for above.
20384 2013-06-04 Ian Bolton <ian.bolton@arm.com>
20386 * config/aarch64/aarch64.c (simd_immediate_info): Remove
20387 element_char member.
20388 (sizetochar): Return signed char.
20389 (aarch64_simd_valid_immediate): Remove elchar and other
20390 unnecessary variables.
20391 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
20392 Calculate element_char as required.
20393 * config/aarch64/aarch64-protos.h: Update and move prototype
20394 for aarch64_output_simd_mov_immediate.
20395 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
20398 2013-06-04 Ian Bolton <ian.bolton@arm.com>
20400 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
20401 information completed by aarch64_simd_valid_immediate.
20402 (aarch64_legitimate_constant_p): Update arguments.
20403 (aarch64_simd_valid_immediate): Work with struct rather than many
20405 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
20406 (aarch64_simd_make_constant): Update arguments.
20407 (aarch64_output_simd_mov_immediate): Work with struct rather than
20408 many pointers. Output immediate directly rather than as operand.
20409 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
20411 * config/aarch64/constraints.md (Dn): Update arguments.
20413 2013-06-04 Ian Bolton <ian.bolton@arm.com>
20415 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
20417 (aarch64_simd_immediate_valid_for_move): Remove.
20418 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
20419 (aarch64_simd_make_constant): Update call.
20420 (aarch64_output_simd_mov_immediate): Update call.
20421 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
20423 * config/aarch64/constraints.md (Dn): Update call.
20425 2013-06-04 Ian Bolton <ian.bolton@arm.com>
20427 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
20428 return type to bool for prototype.
20429 (aarch64_legitimate_constant_p): Check for true instead of not -1.
20430 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
20431 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
20433 2013-06-04 Catherine Moore <clm@codesourcery.com>
20435 * config/mips/mips.opt (meva): New.
20436 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
20437 (ASM_SPEC): Handle -meva.
20438 * doc/invoke.texi (meva): Document.
20440 2013-06-04 Alan Modra <amodra@gmail.com>
20442 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
20445 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20447 * rtl.def: Add extra fourth optional field to define_cond_exec.
20448 * gensupport.c (process_one_cond_exec): Process attributes from
20450 * doc/md.texi: Document fourth field in define_cond_exec.
20452 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
20454 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
20455 out the processing order as in store_bit_field_1.
20457 2013-06-04 Jan Hubicka <jh@suse.cz>
20459 PR middle-end/57500
20460 * cgraphunit.c (cgraph_process_same_body_aliases): Create
20461 non-VAR_DECL node if it does not exist yet.
20463 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
20465 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
20466 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
20467 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
20468 target_cpu_default setting.
20470 2013-06-03 Teresa Johnson <tejohnson@google.com>
20472 * dumpfile.c (opt_info_switch_p): Change -fopt-info
20473 default to -fopt-info=optimized instead of all.
20474 * doc/invoke.texi: Ditto.
20475 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
20476 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
20477 (execute_vect_slp): Emit BB vectorization success under
20478 MSG_OPTIMIZED_LOCATIONS.
20479 * tree-vect-slp.c (vect_slp_transform_bb): Change
20480 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
20481 * tree-vect-loop.c (vect_transform_loop): Ditto.
20483 2013-06-03 Jason Merrill <jason@redhat.com>
20486 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
20487 Use TARGET_EXPR for C++.
20489 2013-06-03 Jakub Jelinek <jakub@redhat.com>
20491 PR rtl-optimization/57268
20492 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
20493 if DEBUG_INSN_P (insn).
20496 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
20498 PR rtl-optimization/57268
20499 * sched-deps.c (sched_analyze_2): Flush dependence lists if
20500 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
20502 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
20504 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
20505 (ix86_avoid_lea_for_addr): Likewise.
20506 (exact_dependency_1): Likewise.
20507 (ix86_adjust_cost): Likewise.
20508 (swap_top_of_ready_list): Fix formatting and !reload_completed check
20510 (do_reorder_for_imul): Fix typo, formatting and
20511 !reload_completed check removed.
20512 (ix86_sched_reorder): Fix typo and formatting.
20513 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
20516 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
20518 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
20520 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
20522 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
20523 <STRING_CST>: Likewise.
20524 <VECTOR_CST>: Likewise.
20526 2013-06-01 Janus Weil <janus@gcc.gnu.org>
20527 Mikael Morin <mikael@gcc.gnu.org>
20529 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
20530 * config.in: Regenerated.
20531 * configure: Regenerated.
20533 2013-06-01 Jan Hubicka <jh@suse.cz>
20535 PR middle-end/57366
20536 * cgraphunit.c (compile): When weakref is not supported,
20537 set up transparent aliases before final output pass.
20538 * varasm.c (assemble_alias): Do not try to do it here.
20540 2013-06-01 Jan Hubicka <jh@suse.cz>
20542 PR middle-end/57467
20543 * passes.c (for_per_function): Skip unanalyzed functions.
20545 2013-06-01 Jan Hubicka <jh@suse.cz>
20547 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
20548 (lto_symtab_merge_symbols_1): ... this one.
20549 (lto_symtab_merge_cgraph_nodes): Rename to ...
20550 (lto_symtab_merge_symbols): ... this one; simplify.
20551 * cgraph.c (same_body_aliases_done): Rename to ...
20552 (cpp_implicit_aliases_done): ... this one.
20553 (cgraph_create_function_alias): Update.
20554 (cgraph_same_body_alias): Update.
20555 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
20556 (verify_edge_corresponds_to_fndecl): Simplify.
20557 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
20558 (cgraph_node): Remove same_body_alias.
20559 (varpool_node): Remove alias_of and extra_name_alias.
20560 (same_body_aliases_done): Rename to ..
20561 (cpp_implicit_aliases_done): ... this one.
20562 (symtab_alias_ultimate_target): Add default parameter.
20563 (symtab_resolve_alias): New function.
20564 (fixup_same_cpp_alias_visibility): Declare.
20565 (cgraph_function_node): Add default parameter.
20566 (cgraph_node_asm_name): Likewise.
20567 (cgraph_function_or_thunk_node): Add default parameter; do
20568 not ICE when it is NULL.
20569 (varpool_variable_node): Likewise.
20570 * tree-emutls.c (create_emultls_var): Update.
20571 (ipa_lower_emutls): Update.
20572 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
20573 (cgraph_reset_node): Reset alias info.
20574 (cgraph_finalize_function): Update.
20575 (fixup_same_cpp_alias_visibility): Move to symtab.c.
20576 (analyze_function): Simplify.
20577 (cgraph_process_same_body_aliases): Simplify.
20578 (analyze_functions): Fixup same body aliases.
20579 (handle_alias_pairs): Simplify.
20580 (assemble_thunk): Update.
20581 (assemble_thunks_and_aliases): Update.
20582 (output_weakrefs): Rewrite.
20583 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
20584 (lto_output_varpool_node): Likewise.
20585 (compute_ltrans_boundary): Remve assert.
20586 (get_alias_symbol): New functoin.
20587 (input_node): Rewrite alias handling.
20588 (input_varpool_node): Likewise.
20589 * ipa-pure-const.c (propagate_pure_const): Fix formating.
20590 * ipa.c (process_references): Handle weakrefs correctly.
20591 (symtab_remove_unreachable_nodes): Likewise.
20592 * trans-mem.c (get_cg_data): Update.
20593 (ipa_tm_create_version_alias): Update.
20594 (ipa_tm_execute): Update.
20595 * symtab.c (dump_symtab_base): Dump aliases.
20596 (verify_symtab_base): Verify aliases.
20597 (symtab_node_availability): New function.
20598 (symtab_alias_ultimate_target): Simplify.
20599 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
20600 handle all the fixup cases.
20601 (symtab_resolve_alias): New function.
20602 * passes.c (ipa_write_summaries): Handle weakrefs.
20603 * varpool.c (varpool_analyze_node): Simplify.
20604 (assemble_aliases): Update.
20605 (varpool_create_variable_alias): Simplify.
20606 (varpool_extra_name_alias): Simplify.
20607 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
20608 (lto_symtab_merge_symbols): ... this one.
20610 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
20613 PR rtl-optimization/57268
20614 * sched-deps.c (sched_analyze_2): Flush dependence lists if
20615 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
20617 2013-06-01 Tobias Burnus <burnus@net-b.de>
20619 Partially reverted:
20620 2013-05-31 Tobias Burnus <burnus@net-b.de>
20622 PR middle-end/57073
20623 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
20626 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
20628 PR rtl-optimization/57268
20629 * sched-deps.c (sched_analyze_2): Flush dependence lists if
20630 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
20632 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
20634 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
20635 unordered comparison operators when -fno-trapping-math is in effect
20637 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
20638 and implement unordered comparison operators properly on the e500.
20640 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
20642 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
20643 for constant scalar integers.
20644 (simplify_relational_operation_1): Likewise.
20646 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
20648 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
20649 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
20652 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
20653 Igor Zamyatin <igor.zamyatin@intel.com>
20655 Silvermont (SLM) architecture performance tuning.
20656 * config/i386/i386.h (enum ix86_tune_indices): Add
20657 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
20658 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
20660 * config/i386/i386.c (initial_ix86_tune_features)
20661 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
20662 (ix86_lea_outperforms): Handle Silvermont tuning.
20663 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
20665 (ix86_use_lea_for_mov): Likewise.
20666 (ix86_avoid_lea_for_addr): Likewise.
20667 (ix86_lea_for_add_ok): Likewise.
20668 (exact_dependency_1): New function.
20669 (exact_store_load_dependency): Likewise.
20670 (ix86_adjust_cost): Handle Silvermont tuning.
20671 (do_reoder_for_imul): Likewise.
20672 (swap_top_of_ready_list): New function.
20673 (ix86_sched_reorder): Changed to handle Silvermont tuning.
20675 * config/i386/i386.md (peepholes that split memory operand in fp
20678 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
20680 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
20681 Remove un-necessary braces.
20683 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
20685 * config/aarch64/aarch64.c (aarch64_classify_symbol):
20686 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
20688 2013-05-31 Tobias Burnus <burnus@net-b.de>
20690 PR middle-end/57073
20691 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
20693 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20696 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
20697 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
20698 * config/arm/neon.md (iordi3_neon): Remove.
20699 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
20700 * config/arm/predicates.md (imm_for_neon_logic_operand):
20701 Move to earlier in the file.
20702 (neon_logic_op2): Likewise.
20703 (arm_iordi_operand_neon): New predicate.
20705 2013-05-31 Richard Biener <rguenther@suse.de>
20707 PR tree-optimization/57478
20708 PR tree-optimization/57453
20709 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
20712 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
20714 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
20715 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
20717 2013-05-30 Tobias Burnus <burnus@net-b.de>
20718 Thomas Koenig <tkoenig@gcc.gnu.org>
20720 PR middle-end/57073
20721 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
20722 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
20724 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
20726 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
20728 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
20730 * target.def (register_usage_leveling_p): New hook.
20731 * targhooks.c (default_register_usage_leveling_p): New.
20732 * targhooks.h (default_register_usage_leveling_p): New prototype.
20733 * lra-assigns.c (register_usage_leveling_p): Use the hook.
20734 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
20735 * doc/tm.texi: Update.
20736 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
20738 2013-05-30 Ian Bolton <ian.bolton@arm.com>
20740 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
20741 (*insv_reg<mode>): New define_insn.
20743 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
20745 PR rtl-optimization/57439
20746 * postreload.c (move2add_valid_value_p): Check that we have
20747 a zero subreg_regno_offset when accessing the register in
20748 the requested mode.
20750 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
20751 Igor Zamyatin <igor.zamyatin@intel.com>
20753 Silvermont (SLM) architecture pipeline model, tuning and
20755 * config.gcc: Add slm config options and target.
20757 * config/i386/slm.md: New.
20759 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
20761 * config/i386/i386-c.c (ix86_target_macros_internal): New case
20763 (ix86_target_macros_internal): Likewise.
20765 * config/i386/i386.c (slm_cost): New cost.
20766 (m_SLM): New macro flag.
20767 (initial_ix86_tune_features): Set m_SLM.
20768 (x86_accumulate_outgoing_args): Likewise.
20769 (x86_arch_always_fancy_math_387): Likewise.
20770 (processor_target_table): Add slm cost.
20771 (cpu_names): Add slm cpu name.
20772 (x86_option_override_internal): Set SLM ISA.
20773 (ix86_issue_rate): New case PROCESSOR_SLM.
20774 (ia32_multipass_dfa_lookahead): Likewise.
20775 (fold_builtin_cpu): Add slm.
20777 * config/i386/i386.h (TARGET_SLM): New target macro.
20778 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
20779 (processor_type): Add PROCESSOR_SLM.
20781 * config/i386/i386.md (cpu): Add new value "slm".
20782 (slm.md): Include slm.md.
20784 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
20785 Zhenqiang Chen <zhenqiang.chen@linaro.org>
20787 * config/arm/arm-protos.h: Add and update function protos.
20788 * config/arm/arm.c (use_simple_return_p): New added.
20789 (thumb2_expand_return): Check simple_return flag.
20790 * config/arm/arm.md: Add simple_return and conditional simple_return.
20791 * config/arm/iterators.md: Add iterator for return and simple_return.
20793 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
20795 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
20796 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
20797 (arm_emit_vfp_multi_reg_pop): Likewise.
20798 (thumb2_emit_ldrd_pop): Likewise.
20799 (arm_expand_epilogue): Add misc REG_CFA notes.
20800 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
20802 2013-05-29 Lawrence Crowl <crowl@google.com>
20804 * config/arm/t-arm: Update for below.
20806 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
20807 Change type to hash_table. Update dependent calls and types.
20809 * config/i386/t-cygming: Update for below.
20811 * config/i386/t-interix: Update for below.
20813 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
20814 Change type to hash_table. Update dependent calls and types.
20815 (i386_find_on_wrapper_list::wrappers): Likewise.
20817 * config/ia64/t-ia64: Update for below.
20819 * config/ia64/ia64.c (bundle_state_table):
20820 Change type to hash_table. Update dependent calls and types.
20822 * config/mips/mips.c (mips_reorg_process_insns::htab):
20823 Change type to hash_table. Update dependent calls and types.
20825 * config/sol2.c (solaris_comdat_htab):
20826 Change type to hash_table. Update dependent calls and types.
20828 * config/t-sol2: Update for above.
20830 2013-05-29 Teresa Johnson <tejohnson@google.com>
20832 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
20833 functions are not yet marked as defined.
20835 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
20836 Pat Haugen <pthaugen@us.ibm.com>
20837 Peter Bergner <bergner@vnet.ibm.com>
20839 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
20843 (vrotl<mode>3): Likewise.
20844 (vashl<mode>3): Likewise.
20845 (vlshr<mode>3): Likewise.
20846 (vashr<mode>3): Likewise.
20848 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
20849 support for power8 V2DI builtins.
20851 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
20852 power8 V2DI builtins.
20853 (vupkhsw): Likewise.
20854 (vupklsw): Likewise.
20855 (vaddudm): Likewise.
20856 (vminsd): Likewise.
20857 (vmaxsd): Likewise.
20858 (vminud): Likewise.
20859 (vmaxud): Likewise.
20860 (vpkudum): Likewise.
20861 (vpksdss): Likewise.
20862 (vpkudus): Likewise.
20863 (vpksdus): Likewise.
20868 (vsubudm): Likewise.
20869 (vcmpequd): Likewise.
20870 (vcmpgtsd): Likewise.
20871 (vcmpgtud): Likewise.
20872 (vcmpequd_p): Likewise.
20873 (vcmpgtsd_p): Likewise.
20874 (vcmpgtud_p): Likewise.
20875 (vupkhsw): Likewise.
20876 (vupklsw): Likewise.
20877 (vaddudm): Likewise.
20878 (vmaxsd): Likewise.
20879 (vmaxud): Likewise.
20880 (vminsd): Likewise.
20881 (vminud): Likewise.
20882 (vpksdss): Likewise.
20883 (vpksdus): Likewise.
20884 (vpkudum): Likewise.
20885 (vpkudus): Likewise.
20890 (vsubudm): Likewise.
20892 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
20893 support for power8 V2DI instructions.
20895 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
20896 power8 V2DI instructions. Combine pack and unpack insns to use an
20897 iterator for each mode. Check whether a particular mode supports
20898 Altivec instructions instead of just checking TARGET_ALTIVEC.
20899 (UNSPEC_VPKUWUM): Likewise.
20900 (UNSPEC_VPKSHSS): Likewise.
20901 (UNSPEC_VPKSWSS): Likewise.
20902 (UNSPEC_VPKUHUS): Likewise.
20903 (UNSPEC_VPKSHUS): Likewise.
20904 (UNSPEC_VPKUWUS): Likewise.
20905 (UNSPEC_VPKSWUS): Likewise.
20906 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
20907 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
20908 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
20909 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
20910 (UNSPEC_VUPKHSB): Likewise.
20911 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
20912 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
20913 (UNSPEC_VUPKHSH): Likewise.
20914 (UNSPEC_VUPKLSB): Likewise.
20915 (UNSPEC_VUPKLSH): Likewise.
20917 (VI_char): Likewise.
20918 (VI_scalar): Likewise.
20919 (VI_unit): Likewise.
20921 (VP_small): Likewise.
20922 (VP_small_lc): Likewise.
20923 (VU_char): Likewise.
20924 (add<mode>3): Likewise.
20925 (altivec_vaddcuw): Likewise.
20926 (altivec_vaddu<VI_char>s): Likewise.
20927 (altivec_vadds<VI_char>s): Likewise.
20928 (sub<mode>3): Likewise.
20929 (altivec_vsubcuw): Likewise.
20930 (altivec_vsubu<VI_char>s): Likewise.
20931 (altivec_vsubs<VI_char>s): Likewise.
20932 (altivec_vavgs<VI_char>): Likewise.
20933 (altivec_vcmpbfp): Likewise.
20934 (altivec_eq<mode>): Likewise.
20935 (altivec_gt<mode>): Likewise.
20936 (altivec_gtu<mode>): Likewise.
20937 (umax<mode>3): Likewise.
20938 (smax<mode>3): Likewise.
20939 (umin<mode>3): Likewise.
20940 (smin<mode>3): Likewise.
20941 (altivec_vpkuhum): Likewise.
20942 (altivec_vpkuwum): Likewise.
20943 (altivec_vpkshss): Likewise.
20944 (altivec_vpkswss): Likewise.
20945 (altivec_vpkuhus): Likewise.
20946 (altivec_vpkshus): Likewise.
20947 (altivec_vpkuwus): Likewise.
20948 (altivec_vpkswus): Likewise.
20949 (altivec_vpks<VI_char>ss): Likewise.
20950 (altivec_vpks<VI_char>us): Likewise.
20951 (altivec_vpku<VI_char>us): Likewise.
20952 (altivec_vpku<VI_char>um): Likewise.
20953 (altivec_vrl<VI_char>): Likewise.
20954 (altivec_vsl<VI_char>): Likewise.
20955 (altivec_vsr<VI_char>): Likewise.
20956 (altivec_vsra<VI_char>): Likewise.
20957 (altivec_vsldoi_<mode>): Likewise.
20958 (altivec_vupkhsb): Likewise.
20959 (altivec_vupkhs<VU_char>): Likewise.
20960 (altivec_vupkls<VU_char>): Likewise.
20961 (altivec_vupkhsh): Likewise.
20962 (altivec_vupklsb): Likewise.
20963 (altivec_vupklsh): Likewise.
20964 (altivec_vcmpequ<VI_char>_p): Likewise.
20965 (altivec_vcmpgts<VI_char>_p): Likewise.
20966 (altivec_vcmpgtu<VI_char>_p): Likewise.
20967 (abs<mode>2): Likewise.
20968 (vec_unpacks_hi_v16qi): Likewise.
20969 (vec_unpacks_hi_v8hi): Likewise.
20970 (vec_unpacks_lo_v16qi): Likewise.
20971 (vec_unpacks_hi_<VP_small_lc>): Likewise.
20972 (vec_unpacks_lo_v8hi): Likewise.
20973 (vec_unpacks_lo_<VP_small_lc>): Likewise.
20974 (vec_pack_trunc_v8h): Likewise.
20975 (vec_pack_trunc_v4si): Likewise.
20976 (vec_pack_trunc_<mode>): Likewise.
20978 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
20980 (vec_vmaxsd): Likewise.
20981 (vec_vmaxud): Likewise.
20982 (vec_vminsd): Likewise.
20983 (vec_vminud): Likewise.
20984 (vec_vpksdss): Likewise.
20985 (vec_vpksdus): Likewise.
20986 (vec_vpkudum): Likewise.
20987 (vec_vpkudus): Likewise.
20988 (vec_vrld): Likewise.
20989 (vec_vsld): Likewise.
20990 (vec_vsrad): Likewise.
20991 (vec_vsrd): Likewise.
20992 (vec_vsubudm): Likewise.
20993 (vec_vupkhsw): Likewise.
20994 (vec_vupklsw): Likewise.
20996 2013-05-29 Jan Hubicka <jh@suse.cz>
20998 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
20999 flags; reorder rest of fields in more consistent way.
21000 (varpool_node): Remove analyzed, finalized and alias.
21001 (cgraph_ndoe): Likewise.
21002 (symtab_alias_ultimate_target): New function.
21003 (cgraph_function_node): Move offline.
21004 (cgraph_reset_node): Declare.
21005 (cgraph_comdat_can_be_unshared_p): Remove.
21006 (varpool_remove_initializer): Declare.
21007 (varpool_first_defined_variable, varpool_next_defined_variable
21008 cgraph_first_defined_function, cgraph_next_defined_function): Update.
21009 (cgraph_function_with_gimple_body_p): Update.
21010 (varpool_all_refs_explicit_p): Update.
21011 (symtab_alias_target): New function.
21012 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
21013 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
21014 (cgraph_function_or_thunk_node): Simplify using
21015 symtab_alias_ultimate_target.
21016 (varpool_variable_node): Likewise.
21017 * cgraph.c (cgraph_create_function_alias): Update.
21018 (cgraph_add_thunk): Update.
21019 (cgraph_remove_node): Update.
21020 (dump_cgraph_node): Do not dump removed flags.
21021 (cgraph_function_body_availability): Update.
21022 (cgraph_propagate_frequency): Update.
21023 (verify_cgraph_node): Check sanity of local flag.
21024 (cgraph_function_node): Move here from cgraph.h; revamp for
21025 cgraph_function_or_thunk_node.
21026 * lto-symtab.c (lto_varpool_replace_node): Update.
21027 (lto_symtab_resolve_can_prevail_p): Update.
21028 (lto_symtab_merge_cgraph_nodes): Update.
21029 * ipa-cp.c (determine_versionability, initialize_node_lattices,
21030 propagate_constants_accross_call, devirtualization_time_bonus,
21031 ipcp_propagate_stage): Update.
21032 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
21033 * ipa-inline-transform.c (clone_inlined_nodes,
21034 preserve_function_body_p): Update.
21035 * ipa-reference.c (propagate): Update.
21036 (write_node_summary_p): Update.
21037 * toplev.c (wrapup_global_declaration_2): Update.
21038 * cgraphunit.c (cgraph_analyze_function): Rename to ...
21039 (analyze_function) ... this one.
21040 (cgraph_process_new_functions): Update.
21041 (cgraph_reset_node): Export.
21042 (cgraph_finalize_function): Update.
21043 (cgraph_add_new_function): Update.
21044 (process_function_and_variable_attributes): Update.
21045 (varpool_finalize_decl): Update.
21046 (symbol_finalized): Remove.
21047 (symbol_finalized_and_needed): Rename to ...
21048 (symbol_defined_and_needed): ... update.
21049 (cgraph_analyze_functions): Update.
21050 (handle_alias_pairs): Update.
21051 (mark_functions_to_output): Update.
21052 (assemble_thunk): Update.
21053 (output_in_order): Update.
21054 (output_weakrefs): Update.
21055 (finalize_compilation_unit): Update.
21056 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
21057 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
21058 input_node, input_varpool_node): Update.
21059 * dbxout.c (dbxout_expand_expr): Update.
21060 * cgraphclones.c (cgraph_clone_node): Update.
21061 (cgraph_copy_node_for_versioning): Update.
21062 (cgraph_materialize_clone): Update.
21063 (cgraph_materialize_all_clones): Update.
21064 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
21065 propagate_pure_const, propagate_nothrow): Update.
21066 * lto-streamer-out.c (lto_output, write_symbol): Update.
21067 * ipa-utils.c (ipa_reverse_postorder): Update.
21068 * ipa-inline.c (can_inline_edge_p): Update.
21069 (update_caller_keys, ipa_inline): Update.
21070 * dwarf2out.c (reference_to_unused,
21071 premark_types_used_by_global_vars_helper): Update.
21072 * tree-eh.c (tree_could_trap_p): Update.
21073 * ipa-split.c (consider_split, execute_split_functions): Update.
21074 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
21075 has_addr_references_p): Update; move ahead in file for better
21077 (process_references): Simplify.
21078 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
21079 bodies are removed.
21080 (cgraph_comdat_can_be_unshared_p): Make static.
21081 (cgraph_externally_visible_p): Update.
21082 (varpool_externally_visible_p): Update.
21083 (function_and_variable_visibility): Update.
21084 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
21085 ipa_tm_mark_force_output_node): Update.
21086 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
21087 estimate_edge_devirt_benefit, inline_generate_summary,
21088 inline_write_summary): Update.
21089 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
21090 * ipa-prop.c (ipa_compute_jump_functions): Update.
21091 (ipa_print_node_params, ipa_prop_read_section,
21092 ipa_update_after_lto_read, read_replacements_section): Update.
21093 * varasm.c (mark_decl_referenced): Update.
21094 (assemble_alias, dump_tm_clone_pairs): Update.
21095 * tree-inline.c (copy_bb): Update.
21096 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
21098 * symtab.c (dump_symtab_base): Print new flags.
21099 (verify_symtab_base): Verify new flags.
21100 (symtab_alias_ultimate_target): New function.
21101 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
21102 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
21104 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
21106 * i386.c (ix86_get_function_versions_dispatcher,
21107 ix86_generate_version_dispatcher_body): Update.
21108 (fold_builtin_cpu): Use varpool_add_new_variable.
21109 * varpool.c (varpool_remove_initializer): Break out from ...
21110 (varpool_remove_node): ... this one.
21111 (dump_varpool_node, varpool_node_for_asm,
21112 cgraph_variable_initializer_availability, varpool_analyze_node,
21113 varpool_assemble_decl, varpool_remove_unreferenced_decls,
21114 varpool_finalize_named_section_flags, varpool_create_variable_alias):
21117 2013-05-29 Jan Hubicka <jh@suse.cz>
21119 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
21121 2013-05-29 Easwaran Raman <eraman@google.com>
21123 PR tree-optimization/57442
21124 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
21125 when control exits the main loop.
21127 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
21129 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
21131 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
21132 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
21133 * rx/t-rx: Add rx100 under multi library matches option for nofpu
21136 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21138 PR tree-optimization/57441
21139 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
21140 Don't limit size of incr_vec to number of candidates.
21142 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
21144 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
21145 and mips16 directories.
21146 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
21147 (MULTILIB_DIRNAMES): Ditto.
21148 (MULTILIB_EXCEPTIONS): Add new exceptions.
21149 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
21150 (MULTILIB_DIRNAMES): Ditto.
21151 (MULTILIB_EXCEPTIONS): Add new exceptions.
21153 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
21154 Marcus Shawcroft <marcus.shawcroft@arm.com>
21156 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
21157 SYMBOL_TINY_ABSOLUTE.
21158 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
21159 SYMBOL_TINY_ABSOLUTE.
21160 (aarch64_expand_mov_immediate): Likewise.
21161 (aarch64_classify_symbol): Likewise.
21162 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
21163 Permit SYMBOL_TINY_ABSOLUTE.
21164 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
21166 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
21167 Marcus Shawcroft <marcus.shawcroft@arm.com>
21169 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
21170 Refactor if/switch. Replace gcc_assert with if.
21172 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
21174 * config/i386/i386.c (initial_ix86_tune_features): Enable
21175 FP Reassociation for AMD bdver1 and bdver2.
21177 2013-05-29 Martin Jambor <mjambor@suse.cz>
21179 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
21180 and IMAGPART_EXPR do not occur within other handled_components.
21182 2013-05-29 Richard Biener <rguenther@suse.de>
21184 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
21185 access on whether the use is in the BB we currently try to
21187 (vect_bb_vectorization_profitable_p): Pass the BB we currently
21188 vectorize to vect_bb_slp_scalar_cost.
21190 2013-05-29 Richard Biener <rguenther@suse.de>
21192 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
21193 computing scalar cost offsetted by stmts that are kept live
21195 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
21196 for computation of scalar cost.
21198 2013-05-28 Steve Ellcey <sellcey@mips.com>
21200 * config/mips/mips-cpus.def (mips32r2): Change processor type.
21202 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
21204 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
21205 array notation built-in reduction functions.
21206 * doc/passes.texi (Passes): Added documentation about changes done
21208 * doc/invoke.texi (C Dialect Options): Added documentation about
21209 the -fcilkplus flag.
21210 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
21211 (BUILTINS_DEF): Depend on cilkplus.def.
21212 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
21213 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
21214 * cilkplus.def: New file.
21216 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
21218 PR rtl-optimization/57439
21219 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
21221 2013-05-28 Easwaran Raman <eraman@google.com>
21223 PR tree-optimization/57337
21224 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
21225 (find_insert_point): Correctly identify the insertion point
21226 when two statements with the same UID is compared.
21228 2013-05-28 Richard Biener <rguenther@suse.de>
21230 PR tree-optimization/56787
21231 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
21232 from the list of data references.
21233 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
21235 (vect_analyze_loop_operations): Likewise.
21236 (vect_transform_loop): Remove clobbers.
21238 2013-05-28 Martin Jambor <mjambor@suse.cz>
21240 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
21241 and REALPART_EXPRs have scalar type.
21243 2013-05-28 Richard Biener <rguenther@suse.de>
21245 PR tree-optimization/57411
21246 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
21248 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
21249 virtual operand propagation.
21251 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
21253 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
21254 destination register for bmasksi_vis.
21255 (vector_init_bshuffle): Likewise.
21256 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
21258 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
21260 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
21261 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
21262 mode if the instruction isn't available in the original mode.
21263 * config/sparc/sparc.opt (mfix-ut699): New option.
21264 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
21265 (divdf3): Turn into expander.
21266 (divdf3_nofix): New insn.
21267 (divdf3_fix): Likewise.
21268 (divsf3): Disable if -mfix-ut699.
21269 (sqrtdf2): Turn into expander.
21270 (sqrtdf2_nofix): New insn.
21271 (sqrtdf2_fix): Likewise.
21272 (sqrtsf2): Disable if -mfix-ut699.
21274 2013-05-27 Richard Biener <rguenther@suse.de>
21276 PR middle-end/57412
21277 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
21278 block for the new loop.
21280 2013-05-27 Richard Biener <rguenther@suse.de>
21282 PR tree-optimization/57343
21283 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
21284 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
21285 (number_of_iterations_cond): Do not build the folded tree.
21287 2013-05-27 Richard Biener <rguenther@suse.de>
21290 PR middle-end/57381
21291 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
21292 OEP_CONSTANT_ADDRESS_OF retained.
21294 PR tree-optimization/57417
21295 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
21296 for unchanged base.
21297 (set_ssa_val_to): Compare addresses using
21298 get_addr_base_and_unit_offset.
21300 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
21302 PR rtl-optimization/56833
21303 * postreload.c (move2add_record_mode): New function.
21304 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
21305 (move2add_use_add2_insn): Use move2add_record_sym_value.
21306 (move2add_use_add3_insn): Likewise.
21307 (reload_cse_move2add): Use move2add_valid_value_p and
21308 move2add_record_mode. Invalidate call-clobbered and REG_INC
21309 affected regs by setting reg_mode to VOIDmode.
21310 (move2add_note_store): Don't pretend the inside of a SUBREG is
21311 the actual destination. Invalidate single/leading registers by
21312 setting reg_mode to VOIDmode.
21313 Use move2add_record_sym_value, move2add_valid_value_p and
21314 move2add_record_mode.
21316 2013-05-27 Richard Biener <rguenther@suse.de>
21318 PR tree-optimization/57396
21319 * tree-affine.c (double_int_constant_multiple_p): Properly
21320 return false for val == 0 and div != 0.
21322 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
21324 * config/mips/mips.h: Use #elif in preprocessor conditions.
21326 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
21329 * config/mips/constraints.md (kl): New constraint.
21330 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
21331 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
21332 constraint for operand 0. Split after CSE for MIPS16. Emit a move
21333 from LO for MIPS16.
21334 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
21336 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
21339 * config/mips/mips.c (mips_can_inline_p): New function.
21340 (TARGET_CAN_INLINE_P): Define.
21342 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
21344 * sched-int.h (ds_t, dw_t): Make unsigned int.
21345 Fix documentation that describes how all the ds_t bits are used.
21346 Reserve the last bit for delayed-branch scheduling.
21347 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
21348 (BITS_PER_DEP_WEAK): Fix definition and documentation.
21349 (gen_dep_weak_1): Remove prototype.
21350 * sched-deps.c (get_dep_weak_1): Make static.
21351 * target.def (speculate_insn, needs_block_p, gen_spec_check,
21352 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
21353 * doc/tm.texi: Regenerate.
21354 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
21356 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
21359 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
21361 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
21362 Sandra Loosemore <sandra@codesourcery.com>
21364 * config.gcc (powerpc-*): Allow native for with-cpu.
21366 2013-05-24 Jeff Law <law@redhat.com>
21368 PR tree-optimization/57124
21369 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
21370 conversion feeding a condition if the range has an overflow
21371 if -fstrict-overflow. Add warnings for when we do make the
21374 2013-05-24 Dehao Chen <dehao@google.com>
21376 * tree-cfg.c (locus_discrim_map): Fix the typo.
21377 (locus_discrim_hasher): Likewise.
21378 (locus_discrim_hasher::hash): Likewise.
21379 (locus_discrim_hasher::equal): Likewise.
21381 2013-05-24 Martin Jambor <mjambor@suse.cz>
21383 PR tree-optimization/57294
21384 * cgraph.h (ipa_record_stmt_references): Declare.
21385 * cgraphbuild.c (ipa_record_stmt_references): New function.
21386 (build_cgraph_edges): Use ipa_record_stmt_references.
21387 (rebuild_cgraph_edges): Likewise.
21388 (cgraph_rebuild_references): Likewise.
21389 * ipa-prop.c (ipa_modify_call_arguments): Discard references
21390 associated with the old statement and build references from the
21391 newly built statements.
21392 * ipa-ref.c (ipa_remove_stmt_references): New function.
21393 * ipa-ref.h (ipa_remove_stmt_references): Declare.
21395 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
21397 * lra-constraints.c (emit_spill_move): Use smaller mode for
21399 (check_and_process_move): Consider mem-reg moves for secondary
21401 (curr_insn_transform): Don't lose insns emitted before for
21402 secondary memory moves.
21403 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
21404 reg set up in the current insn.
21406 2013-05-24 Dehao Chen <dehao@google.com>
21408 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
21410 (locus_descrim_hasher::equal): Likewise.
21411 (build_gimple_cfg): New discriminator assignment algorithm.
21412 (make_edges): Likewise.
21413 (next_discriminator_for_locus): Likewise.
21414 (same_line_p): Likewise.
21415 (assign_discriminators): Likewise.
21416 (make_cond_expr_edges): Likewise.
21417 (make_gimple_switch_edges): Likewise.
21418 (make_goto_expr_edges): Likewise.
21419 (make_gimple_asm_edges): Likewise.
21421 2013-05-24 Ian Bolton <ian.bolton@arm.com>
21423 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
21424 X format specifier to only display bottom 16 bits.
21425 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
21426 immediate to match for operand 2, since it will be masked.
21428 2013-05-24 Richard Biener <rguenther@suse.de>
21430 PR tree-optimization/57287
21431 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
21432 all SSA names that occur in abnormal PHIs.
21434 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
21437 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
21438 that index is not negative.
21440 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
21442 PR rtl-optimization/55177
21443 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
21444 (simplify_byte_swapping_operation): New.
21445 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
21446 (simplify_relational_operation_1): Deal with BSWAP.
21448 2013-05-23 Richard Henderson <rth@redhat.com>
21451 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
21452 (ix86_reorg): Call it.
21454 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
21457 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
21458 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
21459 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
21461 2013-05-23 Christian Bruel <christian.bruel@st.com>
21464 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
21466 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
21467 Marcus Shawcroft <marcus.shawcroft@arm.com>
21469 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
21470 * config/aarch64/constraints.md (Usa): Remove.
21471 * doc/md.texi (AArch64 Usa): Remove.
21473 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
21474 Marcus Shawcroft <marcus.shawcroft@arm.com>
21476 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
21477 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
21478 * config/aarch64/predicates.md (aarch64_const_address): Remove.
21479 (aarch64_mov_operand): Use aarch64_mov_operand_p.
21481 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
21483 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
21484 instruction (AdvSIMD).
21485 * config/aarch64/aarch64-builtins.c
21486 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
21487 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
21489 2013-05-23 Martin Jambor <mjambor@suse.cz>
21491 PR middle-end/57347
21492 * tree.h (contains_bitfld_component_ref_p): Declare.
21493 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
21494 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
21496 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
21497 not access a bit-field. Assert all final offsets are byte-aligned.
21499 2013-05-23 Richard Biener <rguenther@suse.de>
21501 PR tree-optimization/57380
21502 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
21503 least one invariant or re-used load.
21504 * passes.c (init_optimization_passes): Move pass_phiprop before
21507 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
21509 * config/aarch64/aarch64-simd.md
21510 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
21512 2013-05-23 Richard Biener <rguenther@suse.de>
21514 PR middle-end/57381
21515 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
21516 OEP_CONSTANT_ADDRESS_OF retained.
21518 2013-05-23 Jakub Jelinek <jakub@redhat.com>
21520 PR middle-end/57344
21521 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
21522 don't lower unit. Handle unit not being always BITS_PER_WORD.
21524 2013-05-23 Richard Biener <rguenther@suse.de>
21526 PR rtl-optimization/57341
21527 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
21528 instead of true_dependence.
21530 2013-05-22 David Malcolm <dmalcolm@redhat.com>
21532 * bb-reorder.c (branch_threshold): Make const.
21533 (exec_threshold): Ditto.
21535 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
21536 Pat Haugen <pthaugen@us.ibm.com>
21537 Peter Bergner <bergner@vnet.ibm.com>
21539 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
21540 documentation for the power8 crypto builtins.
21542 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
21544 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
21545 macros for defining power8 builtin functions.
21546 (BU_P8V_AV_2): Likewise.
21547 (BU_P8V_AV_P): Likewise.
21548 (BU_P8V_VSX_1): Likewise.
21549 (BU_P8V_OVERLOAD_1): Likewise.
21550 (BU_P8V_OVERLOAD_2): Likewise.
21551 (BU_CRYPTO_1): Likewise.
21552 (BU_CRYPTO_2): Likewise.
21553 (BU_CRYPTO_3): Likewise.
21554 (BU_CRYPTO_OVERLOAD_1): Likewise.
21555 (BU_CRYPTO_OVERLOAD_2): Likewise.
21556 (XSCVSPDP): Fix typo, point to the correct instruction.
21557 (VCIPHER): Add power8 crypto builtins.
21558 (VCIPHERLAST): Likewise.
21559 (VNCIPHER): Likewise.
21560 (VNCIPHERLAST): Likewise.
21561 (VPMSUMB): Likewise.
21562 (VPMSUMH): Likewise.
21563 (VPMSUMW): Likewise.
21564 (VPERMXOR_V2DI): Likewise.
21565 (VPERMXOR_V4SI: Likewise.
21566 (VPERMXOR_V8HI: Likewise.
21567 (VPERMXOR_V16QI: Likewise.
21568 (VSHASIGMAW): Likewise.
21569 (VSHASIGMAD): Likewise.
21570 (VPMSUM): Likewise.
21571 (VPERMXOR): Likewise.
21572 (VSHASIGMA): Likewise.
21574 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
21575 __CRYPTO__ if the crypto instructions are available.
21576 (altivec_overloaded_builtins): Add support for overloaded power8
21579 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
21580 support for power8 crypto builtins.
21581 (builtin_function_type): Likewise.
21582 (altivec_init_builtins): Add support for builtins that take vector
21583 long long (V2DI) arguments.
21585 * config/rs6000/crypto.md: New file, define power8 crypto
21588 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
21589 Pat Haugen <pthaugen@us.ibm.com>
21590 Peter Bergner <bergner@vnet.ibm.com>
21592 * doc/invoke.texi (Option Summary): Add power8 options.
21593 (RS/6000 and PowerPC Options): Likewise.
21595 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
21596 constraints.md instead of rs6000.h. Reorder w* constraints. Add
21597 wm, wn, wr documentation.
21599 * config/rs6000/constraints.md (wm): New constraint for VSX
21600 registers if direct move instructions are enabled.
21601 (wn): New constraint for no registers.
21602 (wq): New constraint for quad word even GPR registers.
21603 (wr): New constraint if 64-bit instructions are enabled.
21604 (wv): New constraint if power8 vector instructions are enabled.
21605 (wQ): New constraint for quad word memory locations.
21607 * config/rs6000/predicates.md (const_0_to_15_operand): New
21608 constraint for 0..15 for crypto instructions.
21609 (gpc_reg_operand): If VSX allow registers in VSX registers as well
21610 as GPR and floating point registers.
21611 (int_reg_operand): New predicate to match only GPR registers.
21612 (base_reg_operand): New predicate to match base registers.
21613 (quad_int_reg_operand): New predicate to match even GPR registers
21614 for quad memory operations.
21615 (vsx_reg_or_cint_operand): New predicate to allow vector logical
21616 operations in both GPR and VSX registers.
21617 (quad_memory_operand): New predicate for quad memory operations.
21618 (reg_or_indexed_operand): New predicate for direct move support.
21620 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
21621 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
21622 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
21623 (POWERPC_MASKS): Add power8 options.
21624 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
21627 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
21628 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
21630 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
21631 (-mpower8-fusion): New power8 options.
21632 (-mpower8-fusion-sign): Likewise.
21633 (-mpower8-vector): Likewise.
21634 (-mcrypto): Likewise.
21635 (-mdirect-move): Likewise.
21636 (-mquad-memory): Likewise.
21638 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
21639 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
21641 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
21642 (rs6000_debug_vector_unit): Add p8_vector.
21643 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
21644 definitions. Also print fusion state.
21645 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
21646 (rs6000_builtin_mask_calculate): Add power8 builtin support.
21647 (rs6000_option_override_internal): Add support for power8.
21648 (rs6000_common_init_builtins): Add debugging for skipped builtins
21649 if -mdebug=builtin.
21650 (rs6000_adjust_cost): Add power8 support.
21651 (rs6000_issue_rate): Likewise.
21652 (insn_must_be_first_in_group): Likewise.
21653 (insn_must_be_last_in_group): Likewise.
21654 (force_new_group): Likewise.
21655 (rs6000_register_move_cost): Likewise.
21656 (rs6000_opt_masks): Likewise.
21658 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
21659 power8 capable assembler, default to power7 options.
21660 (TARGET_DIRECT_MOVE): Likewise.
21661 (TARGET_CRYPTO): Likewise.
21662 (TARGET_P8_VECTOR): Likewise.
21663 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
21664 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
21665 (VECTOR_MEM_P8_VECTOR_P): Likewise.
21666 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
21667 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
21668 (TARGET_XSCVDPSPN): Likewise.
21669 (TARGET_XSCVSPDPN): Likewsie.
21670 (TARGET_SYNC_HI_QI): Likewise.
21671 (TARGET_SYNC_TI): Likewise.
21672 (MASK_CRYPTO): Likewise.
21673 (MASK_DIRECT_MOVE): Likewise.
21674 (MASK_P8_FUSION): Likewise.
21675 (MASK_P8_VECTOR): Likewise.
21676 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
21677 temporary used by some of the direct move instructions to get two FP
21678 temporary registers does not force creation of a stack frame.
21679 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
21680 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
21681 that any VSX registers are tieable, even if they are also an
21682 Altivec vector mode.
21683 (r6000_reg_class_enum): Add wm, wr, wv constraints.
21684 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
21685 (RS6000_BTM_CRYPTO): Likewise.
21686 (RS6000_BTM_COMMON): Likewise.
21688 * config/rs6000/rs6000.md (cpu attribute): Add power8.
21689 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
21690 (enum rs6000_vector): Add power8 vector support.
21692 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
21696 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
21697 (any_sibcall_could_use_r3): this and handle indirect calls.
21698 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
21700 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21702 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
21704 2013-05-22 Richard Biener <rguenther@suse.de>
21706 PR middle-end/57349
21707 * profile.c (branch_prob): Do not split blocks that are
21708 abnormally receiving from ECF_RETURNS_TWICE functions.
21710 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
21712 * recog.c (offsettable_address_addr_space_p): Fix calculation of
21713 address mode. Move pointer mode initialization to the same place.
21715 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
21717 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
21718 while it has any effect.
21720 2013-05-21 Easwaran Raman <eraman@google.com>
21722 PR tree-optimization/57322
21723 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
21724 UID of the statement added to the BB to be 1.
21726 2013-05-21 Jakub Jelinek <jakub@redhat.com>
21728 PR tree-optimization/57331
21729 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
21730 of conversion from pointer type to integral type with integer.
21732 2013-05-21 Martin Jambor <mjambor@suse.cz>
21735 * ipa-prop.c (ipa_read_node_info): Process param_used and
21736 controlled_uses in the same order as when writing.
21738 2013-05-21 Magnus Granberg <baldrick@free.fr>
21741 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
21743 2013-05-21 Richard Biener <rguenther@suse.de>
21745 PR tree-optimization/57318
21746 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
21747 estimate stmts with side-effects as likely eliminated.
21749 2013-05-21 Richard Biener <rguenther@suse.de>
21751 PR tree-optimization/57330
21752 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
21753 preserve the call stmts fntype.
21755 2013-05-21 Richard Biener <rguenther@suse.de>
21757 PR tree-optimization/57303
21758 * tree-ssa-sink.c (statement_sink_location): Improve killing
21759 stmt detection and properly handle self-assignments.
21761 2013-05-21 Christian Bruel <christian.bruel@st.com>
21763 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
21764 spanning registers. LEAF_REG_REMAP is supported only for contiguous
21765 registers. Set register size out of the PARALLEL loop.
21767 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
21770 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
21771 (*fmasf4, *fmasf4_media): New insns.
21773 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
21775 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
21776 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
21777 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
21778 (mips_idiv_insns): Update the comments to say that the returned
21779 instruction counts are in units of BASE_INSN_LENGTH.
21780 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
21781 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
21782 using 2 rather than 4 as the length of indirect MIPS16 and
21783 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
21784 length of a NOP. Don't divide MIPS16 lengths by 2.
21785 (mips16_split_long_branches): Assume a branch is long if the
21786 length is greater than 4 rather than 8.
21787 * config/mips/mips.md (length): Give MIPS16 lengths directly,
21788 rather than multiplying them by 2. Multiply instruction counts
21789 by BASE_INSN_LENGTH rather than 4.
21790 (*jump_mips16, tls_get_tp_mips16_<mode>)
21791 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
21793 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
21795 * config/mips/mips.md (extended_mips16): Remove branch case.
21796 (length): Remove duplicated extended_mips16 test.
21798 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
21800 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
21802 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
21804 * recog.h (Recog_data): Rename to...
21805 (recog_data_d): ...this.
21806 (recog_data): Update accordingly.
21807 * recog.c (recog_data): Likewise.
21808 * reload.c (save_recog_data): Likewise.
21809 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
21810 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
21812 2013-05-17 Julian Brown <julian@codesourcery.com>
21814 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
21815 found in a REG_EQUAL note, invalidate it.
21817 2013-05-17 Easwaran Raman <eraman@google.com>
21819 * tree-ssa-reassoc.c (find_insert_point): New function.
21820 (insert_stmt_after): Likewise.
21821 (get_def_stmt): Likewise.
21822 (ensure_ops_are_available): Likewise.
21823 (not_dominated_by): Likewise.
21824 (rewrite_expr_tree): Do not move statements beyond what is
21825 necessary. Remove call to swap_ops_for_binary_stmt...
21826 (reassociate_bb): ... and move it here.
21827 (build_and_add_sum): Assign UIDs for new statements.
21828 (linearize_expr): Likewise.
21829 (do_reassoc): Renumber gimple statement UIDs.
21831 2013-05-17 Jan Hubicka <jh@suse.cz>
21833 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
21835 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
21836 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
21838 (output_weakrefs): Update.
21840 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
21841 Martin Jambor <mjambor@suse.cz>
21843 PR middle-end/57276
21844 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
21845 value that corresponds to the given aggval is found in values vector.
21847 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
21849 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
21850 sse, sse2, sse3, ssse3 and sse4a flags to options.
21852 2013-05-17 David Malcolm <dmalcolm@redhat.com>
21854 * gengtype-state.c: (s_expr_writer): New class, to handle
21855 prettifying of output layout of s-expressions.
21856 (state_writer): New class, to write out gtype.state.
21857 (state_written_type_count): Move this variable into member data of
21859 (s_expr_writer::s_expr_writer): New code: constructor for new class
21860 (state_writer::state_writer(): ditto
21861 (s_expr_writer::write_new_line): New function
21862 (s_expr_writer::write_any_indent): ditto
21863 (s_expr_writer::begin_s_expr): ditto
21864 (s_expr_writer::end_s_expr): ditto
21865 (write_state_fileloc): convert to method of state_writer...
21866 (state_writer:: write_state_fileloc): ...and use methods of
21867 s_expr_writer to write indentation into the gtype.state output file
21868 to visually represent the hierarchical structure of the list
21870 (write_state_fields): ditto, renaming to...
21871 (state_writer::write_state_fields)
21872 (write_state_a_string): ditto, renaming to...
21873 (state_writer::write_state_a_string)
21874 (write_state_string_option): ditto, renaming to...
21875 (state_writer::write_state_string_option)
21876 (write_state_type_option): ditto, renaming to...
21877 (state_writer::write_state_type_option)
21878 (write_state_nested_option): ditto, renaming to...
21879 (state_writer::write_state_nested_option)
21880 (write_state_option): ditto, renaming to...
21881 (state_writer::write_state_option)
21882 (write_state_options): ditto, renaming to...
21883 (state_writer::write_state_options)
21884 (write_state_lang_bitmap): ditto, renaming to...
21885 (state_writer::write_state_lang_bitmap)
21886 (write_state_version): ditto, renaming to...
21887 (state_writer::write_state_version)
21888 (write_state_scalar_type): ditto, renaming to...
21889 (state_writer::write_state_scalar_type)
21890 (write_state_string_type): ditto, renaming to...
21891 (state_writer::write_state_string_type)
21892 (write_state_undefined_type): ditto, renaming to...
21893 (state_writer::write_state_undefined_type)
21894 (write_state_struct_union_type): ditto, renaming to...
21895 (state_writer::write_state_struct_union_type)
21896 (write_state_struct_type): ditto, renaming to...
21897 (state_writer::write_state_struct_type)
21898 (write_state_user_struct_type): ditto, renaming to...
21899 (state_writer::write_state_user_struct_type)
21900 (write_state_lang_struct_type): ditto, renaming to...
21901 (state_writer::write_state_lang_struct_type)
21902 (write_state_param_struct_type): ditto, renaming to...
21903 (state_writer::write_state_param_struct_type)
21904 (write_state_pointer_type): ditto, renaming to...
21905 (state_writer::write_state_pointer_type)
21906 (write_state_array_type): ditto, renaming to...
21907 (state_writer::write_state_array_type)
21908 (write_state_gc_used): ditto, renaming to...
21909 (state_writer::write_state_gc_used)
21910 (write_state_common_type_content): ditto, renaming to...
21911 (state_writer::write_state_common_type_content)
21912 (write_state_type): ditto, renaming to...
21913 (state_writer::write_state_type)
21914 (write_state_pair_list): ditto, renaming to...
21915 (state_writer::write_state_pair_list)
21916 (write_state_pair): ditto, renaming to...
21917 (state_writer::write_state_pair)
21918 (write_state_typedefs): ditto, renaming to...
21919 (state_writer::write_state_typedefs)
21920 (write_state_structures): ditto, renaming to...
21921 (state_writer::write_state_structures)
21922 (write_state_param_structs): ditto, renaming to...
21923 (state_writer::write_state_param_structs)
21924 (write_state_variables): ditto, renaming to...
21925 (state_writer::write_state_variables)
21926 (write_state_srcdir): ditto, renaming to...
21927 (state_writer::write_state_srcdir)
21928 (write_state_files_list): ditto, renaming to...
21929 (state_writer::write_state_files_list)
21930 (write_state_languages): ditto, renaming to...
21931 (state_writer::write_state_languages)
21932 (write_state): create a state_writer instance and use it when
21933 writing out the state file
21935 2013-05-17 Mike Stump <mikestump@comcast.net>
21937 PR rtl-optimization/57304
21938 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
21939 accessing DF_REF_REAL_LOC.
21941 2013-05-17 Jakub Jelinek <jakub@redhat.com>
21943 PR rtl-optimization/57281
21944 PR rtl-optimization/57300
21945 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
21946 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
21947 what the other splitter did if the registers are dead.
21949 2013-05-17 Richard Biener <rguenther@suse.de>
21951 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
21954 2013-05-17 Jakub Jelinek <jakub@redhat.com>
21956 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
21959 2013-05-17 Marek Polacek <polacek@redhat.com>
21961 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
21962 length when doing non-zero store of storing '\0' to '\0'.
21964 2013-05-17 Jakub Jelinek <jakub@redhat.com>
21966 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
21967 vect_external_def oprnd1 with loop_vinfo, try to emit
21968 optional cast, negation and and stmts on the loop preheader
21969 edge instead of into the pattern def seq.
21971 PR tree-optimization/57051
21972 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
21973 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
21975 2013-05-16 Nick Clifton <nickc@redhat.com>
21977 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
21978 (rl78_is_naked_func): New function.
21979 (rl78_expand_prologue): Skip prologue generation for naked functions.
21980 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
21981 * doc/extend.texi (naked): Add RL78 to the list of processors
21982 that supports this attribute.
21984 2013-05-16 Jeff Law <law@redhat.com>
21986 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
21988 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
21990 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
21991 cache parameters using detect_caches_amd also for CYRIX,
21992 NSC and TM2 signatures.
21994 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
21995 Dzianis Kahanovich <mahatma@eu.by>
21999 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
22000 VIA/Centaur processors and determine their cache parameters
22001 using detect_caches_amd.
22003 2013-05-16 Teresa Johnson <tejohnson@google.com>
22005 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
22006 (rtl_verify_edges): New function.
22007 (rtl_verify_bb_insns): Ditto.
22008 (rtl_verify_bb_pointers): Ditto.
22009 (rtl_verify_bb_insn_chain): Ditto.
22010 (rtl_verify_fallthru): Ditto.
22011 (rtl_verify_bb_layout): Ditto.
22012 (rtl_verify_flow_info_1): Outline checks into new functions.
22013 (rtl_verify_flow_info): Ditto.
22015 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
22017 * cfghooks.c (copy_bbs): Add update_dominance argument.
22018 * cfghooks.h (copy_bbs): Update prototype.
22019 * tree-cfg.c (gimple_duplicate_sese_region):
22020 Add update_dominance argument.
22021 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
22022 * tree-ssa-loop-ch.c (copy_loop_headers): Update
22023 gimple_duplicate_sese_region call.
22024 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
22025 Update copy_bbs call.
22026 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
22027 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
22029 2013-05-16 Jakub Jelinek <jakub@redhat.com>
22031 * tree-vectorizer.h (NUM_PATTERNS): Increment.
22032 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
22033 vect_recog_rotate_pattern.
22034 (vect_recog_rotate_pattern): New function.
22036 2013-05-16 Jason Merrill <jason@redhat.com>
22038 * Makefile.in (LLINKER): New variable.
22039 (mostlyclean): Remove link mutex.
22040 * configure.ac: Handle --enable-link-mutex.
22041 * lock-and-run.sh: New script.
22043 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
22046 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
22049 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
22051 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
22053 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
22055 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
22056 * config/arm/arm.c (next_consecutive_mem): New function.
22057 (gen_movmem_ldrd_strd): Likewise.
22058 * config/arm/arm.md (movmemqi): Update condition and code.
22059 (unaligned_loaddi, unaligned_storedi): New patterns.
22061 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
22063 * config.gcc: Obsolete *-*-solaris2.9*.
22064 * doc/install.texi (Specific, *-*-solaris2*): Document it.
22066 2013-05-16 Richard Biener <rguenther@suse.de>
22068 * passes.c (init_optimization_passes): Move pass_parallelize_loops
22069 earlier, after GRAPHITE transforms and IV canonicalization.
22071 2013-05-16 Jakub Jelinek <jakub@redhat.com>
22073 * omp-low.c (extract_omp_for_data): For collapsed loops,
22074 if at least one of the loops is known at compile time to
22075 iterate zero times, set count to 0.
22076 (expand_omp_regimplify_p): New function.
22077 (expand_omp_for_generic): For collapsed loops, if at least
22078 one of the loops isn't known to iterate at least once,
22079 add runtime check with setting count to 0.
22080 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
22081 For unsigned types if it isn't known at compile time that
22082 the loop will iterate at least once, add runtime check to bypass
22083 the whole loop if initial condition isn't true.
22085 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
22087 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
22089 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
22091 PR middle-end/57286
22092 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
22093 transformations to avoid an infinite loop.
22095 2013-05-16 Marek Polacek <polacek@redhat.com>
22097 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
22099 2013-05-15 Leif Ekblad <leif@rdos.net>
22101 * config/i386/i386.c (ix86_decompose_address): Use
22102 DEFAULT_TLS_SEG_REG to access TLS segment register.
22103 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
22104 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
22105 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
22107 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
22110 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
22111 sibling calls to functions that would normally be lazily bound,
22112 unless $gp is call-clobbered.
22114 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
22116 * config/i386/i386.c (ix86_option_override_internal): Update
22117 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
22118 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
22119 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
22120 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
22122 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
22124 2013-05-15 Andreas Schwab <schwab@suse.de>
22126 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
22127 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
22130 2013-05-15 Teresa Johnson <tejohnson@google.com>
22132 * loop-unroll.c (report_unroll_peel): Check decision before
22133 emitting unroll/peel message.
22135 2013-05-15 Teresa Johnson <tejohnson@google.com>
22137 * function.h (has_bb_partition): New rtl_data flag.
22138 (bb_reorder_complete): Ditto.
22139 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
22140 instead of flag_reorder_blocks_and_partition.
22141 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
22142 with some enhancements.
22143 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
22144 * bb-reorder.c (connect_traces): Check for has_bb_partition
22145 instead of flag_reorder_blocks_and_partition.
22146 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
22147 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
22148 verify_hot_cold_block_grouping.
22149 (partition_hot_cold_basic_blocks): Set has_bb_partition.
22151 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
22154 * config/arm/predicates.md (call_insn_operand): New predicate.
22155 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
22156 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
22157 if insn is not a tail call.
22158 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
22160 * config/arm/arm.h (enum reg_class): New caller save register class.
22161 (REG_CLASS_NAMES): Likewise.
22162 (REG_CLASS_CONTENTS): Likewise.
22163 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
22166 2013-05-15 Richard Biener <rguenther@suse.de>
22168 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
22169 of MSG_OPTIMIZED_LOCATIONS.
22170 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
22171 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
22173 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
22174 of MSG_OPTIMIZED_LOCATIONS.
22175 (execute_vect_slp): Likewise.
22176 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
22177 (vect_create_cond_for_alias_checks): Likewise.
22178 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
22179 (vect_recog_widen_mult_pattern): Likewise.
22180 (vect_recog_widen_sum_pattern): Likewise.
22181 (vect_recog_over_widening_pattern): Likewise.
22182 (vect_recog_widen_shift_pattern): Likewise.
22183 (vect_recog_vector_vector_shift_pattern): Likewise.
22184 (vect_recog_divmod_pattern): Likewise.
22185 (vect_recog_mixed_size_cond_pattern): Likewise.
22186 (vect_recog_bool_pattern): Likewise.
22187 (vect_pattern_recog_1): Likewise.
22189 2013-05-15 Martin Jambor <mjambor@suse.cz>
22191 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
22192 non-functions to builtin_unreachable.
22193 * ipa-inline-transform.c (inline_call): Do not assert estimates were
22194 correct when new direct edges were discovered.
22196 2013-05-15 Martin Jambor <mjambor@suse.cz>
22198 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
22199 header, print symbol order instead of node uid, print more information
22200 about indirect edge targets.
22201 (ipa_make_edge_direct_to_target): Print symbol order instead of node
22203 (ipa_make_edge_direct_to_target): Likewise.
22204 (remove_described_reference): Likewise.
22205 (propagate_controlled_uses): Likewise.
22206 (ipa_print_node_params): Also print symbol order.
22207 (ipcp_transform_function): Print symbol order instead of node uids.
22208 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
22209 (cgraph_get_create_real_symbol_node): Likewise.
22210 * ipa-cp.c (print_lattice): Likewise.
22211 (print_all_lattices): Likewise.
22212 (determine_versionability): Likewise.
22213 (initialize_node_lattices): Likewise.
22214 (estimate_local_effects): Likewise.
22215 (update_profiling_info): Likewise.
22216 (create_specialized_node): Likewise.
22217 (perhaps_add_new_callers): Likewise.
22218 (decide_about_value): Likewise.
22219 (decide_whether_version_node): Likewise.
22220 (identify_dead_nodes): Likewise.
22221 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
22222 (dump_inline_summary): Likewise.
22223 (estimate_node_size_and_time): Likewise.
22224 (inline_analyze_function): Likewise.
22225 * ipa-inline.c (report_inline_failed_reason): Likewise.
22226 (want_early_inline_function_p): Likewise.
22227 (edge_badness): Likewise.
22228 (update_edge_key): Likewise.
22229 (inline_small_functions): Likewise. Add dumping of order to two other
22231 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
22232 instead of node uids.
22233 (propagate_pure_const): Likewise.
22234 (propagate_pure_const): Likewise.
22235 * ipa-utils.c (dump_cgraph_node_set): Likewise.
22236 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
22237 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
22239 * tree-pretty-print.c (dump_function_header): Likewise.
22240 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
22241 Print symbol order instead of node uids.
22243 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
22245 * config/s390/s390.c (s390_register_move_cost): Don't impose the
22246 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
22248 2013-05-15 Richard Biener <rguenther@suse.de>
22250 PR tree-optimization/57275
22251 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
22252 return value for fail to do runtime alias checks for gather loads.
22254 2013-05-15 Jan Hubicka <jh@suse.cz>
22258 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
22259 weakrefs are not external.
22260 (lto_symtab_merge_decls): Fix thinko when dealing with
22261 non-lto_symtab decls.
22262 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
22263 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
22264 * varpool.c (dump_varpool_node): Dump more flags.
22266 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
22268 * config/i386/i386.c (processor_alias_table): Add instruction
22269 FSGSBASE for AMD bdver3 architecture.
22271 2013-05-14 Jakub Jelinek <jakub@redhat.com>
22273 * tree.c (warn_deprecated_use): Print file:line using locus color.
22274 * diagnostic.c (diagnostic_report_current_module): Print file:line
22275 and file:line:column using locus color.
22277 2013-05-14 Mike Stump <mikestump@comcast.net>
22279 * gdbinit.in: Add __null.
22281 2013-05-14 Mike Stump <mikestump@comcast.net>
22283 * recog.h: Rename struct recog_data to Recog_data.
22284 * recog.c: Likewise.
22285 * reload.c (can_reload_into): Likewise.
22286 * config/picochip/picochip.c: Likewise.
22288 2013-05-14 Mike Stump <mikestump@comcast.net>
22290 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
22292 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
22294 * resource.h (struct resources): Remove unch_memory member.
22295 (CLEAR_RESOURCE): Don't clear unch_memory.
22296 * resource.c (mark_referenced_resources): Don't set it.
22297 (mark_set_resources): Likewise.
22298 (mark_target_live_regs): Don't clear it.
22299 (init_resource_info): Likewise.
22300 * reorg.c (resource_conflicts_p): Don't compare it.
22301 (redundant_insn): Don't set it.
22303 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
22305 * emit-rtl.c (next_label): Remove unused function.
22306 (skip_consecutive_labels, link_cc0_insns): Move to ...
22307 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
22308 only place where these functions are used, and make them static.
22310 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
22312 * fold-const.c (fold_negate_expr): Handle vectors.
22313 (fold_truth_not_expr): Make it static.
22314 (fold_invert_truthvalue): New static function.
22315 (invert_truthvalue_loc): Handle vectors. Do not call
22316 fold_truth_not_expr directly.
22317 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
22318 <TRUTH_NOT_EXPR>: Do not cast to boolean.
22319 (fold_comparison): Handle vector constants.
22320 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
22321 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
22322 * tree.h (fold_truth_not_expr): Remove declaration.
22324 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
22326 * config/aarch64/aarch64-simd.md
22327 (aarch64_vcond_internal<mode>): Rename to...
22328 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
22329 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
22330 float modes. Clarify all iterator modes.
22331 (vcond<mode><mode>): Use new name for vcond expanders.
22332 (vcond<v_cmp_result><mode>): Likewise.
22333 (vcondu<mode><mode>: Likewise.
22334 * config/aarch64/iterators.md (VDQF_COND): New.
22336 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
22339 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
22340 variable for the shift amount. Check that we shift by non-negative
22343 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
22346 * config/arm/arm.h (EPILOGUE_USES): Only return true
22347 for LR_REGNUM after epilogue_completed.
22349 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
22351 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
22352 is error_mark_node.
22354 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
22357 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
22358 and Solaris 11+/x86 with gld.
22359 * configure: Regenerate.
22361 2013-05-14 Jakub Jelinek <jakub@redhat.com>
22363 * expmed.c (expand_shift_1): Canonicalize rotates by
22364 constant bitsize / 2 to bitsize - 1.
22365 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
22366 case ROTATERT>: Likewise.
22369 2013-05-10 Jakub Jelinek <jakub@redhat.com>
22371 * config/i386/i386.md (rotateinv): New code attr.
22372 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
22373 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
22374 roll $31, %eax, etc.
22376 2013-05-14 Richard Biener <rguenther@suse.de>
22378 PR middle-end/57235
22379 * tree-eh.c (sink_clobbers): Give up for successors with
22380 multiple predecessors and no virtual uses.
22382 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
22384 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
22385 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
22387 2013-05-14 Jakub Jelinek <jakub@redhat.com>
22389 PR middle-end/57251
22390 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
22391 the case when both op0 and op1 have VOIDmode.
22393 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
22395 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
22396 in multiply-accumulate mode.
22398 2013-05-13 Guozhi Wei <carrot@google.com>
22400 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
22402 2013-05-13 Kai Tietz <ktietz@redhat.com>
22405 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
22406 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
22407 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
22408 * config/i386/i386.c (ix86_option_override_internal): Likewise.
22409 (ix86_expand_prologue): Likewise.
22410 (ix86_expand_split_stack_prologue): Likewise.
22411 (legitimate_pic_address_disp_p): Likewise.
22412 (legitimize_pic_address): Likewise.
22413 (legitimize_tls_address): Likewise.
22414 (legitimize_pe_coff_symbol): Likewise.
22415 (output_pic_addr_const): Likewise.
22416 (construct_plt_address): Likewise.
22417 (ix86_expand_call): Likewise.
22418 (x86_output_mi_thunk): Likewise.
22419 (x86_function_profiler): Likewise.
22421 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
22423 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
22424 similar switch cases.
22425 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
22426 (aarch64_simd_mov_to_<mode>low): Delete.
22427 (aarch64_simd_mov_to_<mode>high): Delete.
22428 (move_lo_quad_<mode>): Add w<-r alternative.
22429 (aarch64_simd_move_hi_quad_<mode>): Likewise.
22430 (aarch64_simd_mov_from_*): Update type attribute.
22431 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
22434 2013-05-13 Jan Hubicka <jh@suse.cz>
22436 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
22437 * config/i386/i386.c (ix86_compute_frame_layout,
22438 ix86_expand_epilogue, emit_i387_cw_initialization,
22439 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
22440 ix86_local_alignment): Fix use of size/speed predicates.
22442 2013-05-13 Jakub Jelinek <jakub@redhat.com>
22444 PR tree-optimization/45216
22445 PR tree-optimization/57157
22446 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
22447 the (-Y) & (B - 1) variant if OP is |.
22448 * expmed.c (expand_shift_1): For rotations by const0_rtx just
22449 return shifted. Use (-op1) & (prec - 1) as other_amount
22450 instead of prec - op1.
22452 2013-05-13 Martin Jambor <mjambor@suse.cz>
22454 PR middle-end/42371
22455 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
22456 (ipa_constant_data): New type.
22457 (ipa_jump_func): Use ipa_constant_data to hold information about
22458 constant jump functions.
22459 (ipa_get_jf_constant): Adjust to jump function type changes.
22460 (ipa_get_jf_constant_rdesc): New function.
22461 (ipa_param_descriptor): New field controlled_uses.
22462 (ipa_get_controlled_uses): New function.
22463 (ipa_set_controlled_uses): Likewise.
22464 * ipa-ref.h (ipa_find_reference): Declare.
22465 * ipa-prop.c (ipa_cst_ref_desc): New type.
22466 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
22468 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
22469 New parameter cs. Adjust all callers.
22470 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
22471 (remove_described_reference): New function.
22472 (jfunc_rdesc_usable): Likewise.
22473 (try_make_edge_direct_simple_call): Decrement controlled use count,
22474 attempt to remove reference if it hits zero.
22475 (combine_controlled_uses_counters): New function.
22476 (propagate_controlled_uses): Likewise.
22477 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
22478 (ipa_edge_duplication_hook): Duplicate reference descriptions.
22479 (ipa_print_node_params): Print described use counter.
22480 (ipa_write_jump_function): Adjust to jump function type changes.
22481 (ipa_read_jump_function): New parameter CS, pass it to
22482 ipa_set_jf_constant. Adjust caller.
22483 (ipa_write_node_info): Stream controlled use count
22484 (ipa_read_node_info): Likewise.
22485 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
22487 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
22488 count. Remove cloning-added reference if it reaches zero.
22489 * ipa-ref.c (ipa_find_reference): New function.
22491 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
22493 * config/i386/i386.c (processor_target_table): Modified default
22494 alignment values for AMD BD and BT architectures.
22496 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
22498 * tree-vect-generic.c (uniform_vector_p): Move ...
22499 * tree.c (uniform_vector_p): ... here.
22500 * tree.h (uniform_vector_p): Declare it.
22501 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
22504 2013-05-13 Jakub Jelinek <jakub@redhat.com>
22506 PR tree-optimization/57230
22507 * tree-ssa-strlen.c (handle_char_store): Record length for
22508 array store from STRING_CST.
22510 PR tree-optimization/57230
22511 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
22514 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
22516 * config/epiphany/epiphany.c (epiphany_init): Check size of
22517 NUM_MODES_FOR_MODE_SWITCHING.
22518 (epiphany_expand_prologue):
22519 Remove CONFIG_REGNUM initial value handling code.
22520 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
22521 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
22522 (emit_set_fp_mode, epiphany_mode_after): Likewise.
22523 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
22524 Don't return 1 for FP_MODE_NONE.
22525 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
22526 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
22527 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
22528 * config/epiphany/epiphany.md (save_config): New pattern.
22530 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
22532 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
22534 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
22536 * config/i386/i386.md (memory): Handle sseishft1.
22537 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
22538 (*vec_extractv2di_1): Ditto.
22540 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
22542 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
22545 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
22547 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
22548 Add mthumb/march=armv7-a multilib.
22549 Add mthumb/march=armv7-r multilib.
22550 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
22552 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
22554 * config/v850/t-rtems: Add more multilibs.
22556 2013-05-10 Richard Biener <rguenther@suse.de>
22558 PR tree-optimization/57214
22559 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
22560 not propagate from SSA names that occur in abnormal PHI nodes.
22562 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
22564 * stor-layout.c (element_precision): New function.
22565 * machmode.h (element_precision): Declare it.
22566 * tree.c (build_minus_one_cst): New function.
22567 (element_precision): Likewise.
22568 * tree.h (build_minus_one_cst): Declare new function.
22569 (element_precision): Likewise.
22570 * fold-const.c (operand_equal_p): Use element_precision.
22571 (fold_binary_loc): Handle vector types.
22572 * convert.c (convert_to_integer): Use element_precision.
22573 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
22576 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
22578 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
22579 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
22580 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
22581 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
22582 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
22583 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
22584 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
22585 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
22586 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
22587 (Uuw8): New constraints.
22588 (Usb4): Move into alphabetical order.
22589 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
22590 (sd8_operand, ub8_operand, uw8_operand): New predicates.
22591 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
22592 previously unnamed patterns.
22593 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
22594 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
22595 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
22596 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
22597 of set_attr_alternative/if_then_else. Use extended_mips16 instead
22598 of specific lengths.
22600 2013-05-10 Jakub Jelinek <jakub@redhat.com>
22602 * config/i386/i386.md (rotateinv): New code attr.
22603 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
22604 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
22605 roll $31, %eax, etc.
22607 PR tree-optimization/45216
22608 PR tree-optimization/57157
22609 * tree-ssa-forwprop.c (simplify_rotate): New function.
22610 (ssa_forward_propagate_and_combine): Call it.
22612 2013-05-10 Richard Biener <rguenther@suse.de>
22614 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
22615 disable peeling when we version for aliasing.
22616 (vector_alignment_reachable_p): Honor explicit user alignment.
22617 (vect_supportable_dr_alignment): Likewise.
22618 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
22619 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
22620 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
22621 then peeling to arrange for the cost-model check to come first.
22623 2013-05-10 Alan Modra <amodra@gmail.com>
22625 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
22626 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
22627 * configure: Regenerate.
22629 2013-05-10 Alan Modra <amodra@gmail.com>
22632 * varasm.c (default_elf_select_section): Move !DECL_P check..
22633 (get_named_section): ..to here before calling get_section_name.
22635 (default_section_type_flags): Add DECL_P check.
22636 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
22637 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
22639 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
22641 * config/epiphany/epiphany.c (epiphany_expand_prologue):
22642 When using gen_stack_adjust_str with a register offset, add a
22643 REG_FRAME_RELATED_EXPR note.
22645 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
22647 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
22648 (*vec_extractv4si_zext_mem): Ditto.
22649 (*vec_extractv2di): Add 0->x and x->x alternatives.
22650 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
22651 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
22653 2013-05-09 Jason Merrill <jason@redhat.com>
22655 N3639 C++1y VLA support
22656 * gimplify.c (gimplify_vla_decl): Don't touch an existing
22659 * tree.c (build_constructor_va): New.
22660 * tree.h: Declare it.
22662 2013-05-09 Martin Jambor <mjambor@suse.cz>
22665 * gimple-fold.c (canonicalize_constructor_val): Call
22666 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
22668 2013-05-09 Jan Hubicka <jh@suse.cz>
22669 Richard Biener <rguenther@suse.de>
22672 * symtab.c (symtab_make_decl_local): Do not add private names.
22674 2013-05-09 Jan Hubicka <jh@suse.cz>
22677 * symtab.c (insert_to_assembler_name_hash): Handle clones.
22678 (unlink_from_assembler_name_hash): Likewise.
22679 (symtab_prevail_in_asm_name_hash, symtab_register_node,
22680 symtab_unregister_node, symtab_initialize_asm_name_hash,
22681 change_decl_assembler_name): Update.
22683 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
22685 * config/aarch64/aarch64.md: New movtf split.
22686 (*movtf_aarch64): Update.
22687 (aarch64_movdi_tilow): Handle TF modes and rename to
22688 aarch64_movdi_<mode>low.
22689 (aarch64_movdi_tihigh): Handle TF modes and rename to
22690 aarch64_movdi_<mode>high
22691 (aarch64_movtihigh_di): Handle TF modes and rename to
22692 aarch64_mov<mode>high_di
22693 (aarch64_movtilow_di): Handle TF modes and rename to
22694 aarch64_mov<mode>low_di
22695 (aarch64_movtilow_tilow): Remove spurious whitespace.
22696 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
22698 (aarch64_print_operand): Update.
22700 2013-05-09 Alan Modra <amodra@gmail.com>
22702 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
22704 * configure: Regenerate.
22706 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
22708 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
22709 splitter preparation statements.
22710 * config/i386/sse.md (*vec_extract* splitters): Ditto.
22711 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
22714 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
22716 * gimple-ssa-strength-reduction.c (count_candidates): Change
22717 return value to int.
22718 (analyze_candidates_and_replace): Change type of length to int.
22720 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
22722 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
22723 (*vec_extract<mode>): Use VI12_128 mode iterator.
22724 (*vec_extract<mode>_mem): Ditto.
22725 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
22728 2013-05-08 Diego Novillo <dnovillo@google.com>
22733 2012-08-17 Diego Novillo <dnovillo@google.com>
22736 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
22737 * config.in: Regenerate.
22738 * configure: Regenerate.
22739 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
22741 2013-05-08 Jan Hubicka <jh@suse.cz>
22744 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
22745 * cgraph.h (symtab_node_base): Add unique_name.
22746 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
22747 input_overwrite_node, input_varpool_node): Stream unique_name.
22748 * cgraphclones.c (cgraph_create_virtual_clone,
22749 cgraph_function_versioning): Set unique_name.
22750 * ipa.c (function_and_variable_visibility): Set unique_name.
22752 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
22754 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
22755 (alloc_cand_and_find_basis): Restrict conditional candidate
22756 processing to CAND_MULTs.
22758 2013-05-08 Jan Hubicka <jh@suse.cz>
22761 lto-symtab.c (lto_symtab_symbol_p): New function.
22762 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
22763 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
22764 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
22765 Skip static symbols.
22767 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
22769 PR tree-optimization/57200
22770 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
22771 Only call inform if the preceding warning_at returns true.
22773 2013-05-07 Han Shen <shenhan@google.com>
22775 * cfgexpand.c (record_or_union_type_has_array_p): New function.
22776 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
22777 * common.opt (fstack-protector-strong): New option.
22778 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
22779 * doc/invoke.texi (Optimization Options): Document
22780 "-fstack-protector-strong".
22781 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
22783 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
22785 * config/mips/mips.c (mips_machine_reorg2): Return 0.
22787 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
22789 * ira.c (update_equiv_regs): Add insn having equiv memory even if
22790 it is not lhs of the insn.
22791 (setup_reg_equiv): Remove insn having equiv memory which it is not
22793 * lra-constraints.c (process_address): Try to improve generation
22794 code for address base + disp.
22795 (lra_constraints): Make correct the code for checking insn setting
22796 up backward equivalence. Remove insn only if it is in the init
22798 * lra-eliminations.c (update_reg_eliminate): Change return value.
22799 (lra_eliminate): Use the result.
22801 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
22803 * config/i386/sse.md (ssescalarnummask): New mode attribute.
22804 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
22805 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
22806 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
22807 register target operands.
22808 (*vec_extractv8hi_sse2): New pattern.
22809 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
22810 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
22811 (*vec_extract<mode>_mem): New insn and split pattern.
22813 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
22815 * config/arm/arm.c (arm_asan_shadow_offset): New function.
22816 (TARGET_ASAN_SHADOW_OFFSET): Define.
22817 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
22818 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
22820 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
22822 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
22823 (incr_vec_index): Return -1 if increment not found.
22824 (create_add_on_incoming_edge): Assert if increment not found.
22825 (record_increment): Limit number of increments recorded.
22826 (all_phi_incrs_profitable): Return false if an increment not found.
22827 (replace_profitable_candidates): Don't process increments that were
22829 (analyze_candidates_and_replace): Limit size of incr_vec.
22831 2013-05-07 Richard Biener <rguenther@suse.de>
22833 * calls.c (special_function_p): setjmp-like functions are leaf.
22834 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
22835 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
22837 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
22839 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
22840 (aarch64_simd_mov<mode>): New expander.
22841 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
22842 (aarch64_simd_mov_to_<mode>high): Likewise.
22843 (aarch64_simd_mov_from_<mode>low): Likewise.
22844 (aarch64_simd_mov_from_<mode>high): Likewise.
22845 (aarch64_dup_lane<mode>): Update.
22846 (aarch64_dup_lanedi): New instruction pattern.
22847 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
22848 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
22850 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
22852 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
22853 (replace_mult_candidate): Remove unnecessary argument; remove
22854 unnecessary parameter from call to introduce_cast_before_cand.
22855 (replace_unconditional_candidate): Remove unnecessary parameter
22856 from call to replace_mult_candidate.
22857 (replace_conditional_candidate): Likewise.
22858 (insert_initializers): Use make_temp_ssa_name.
22859 (introduce_cast_before_cand): Remove unnecessary argument; use
22860 make_temp_ssa_name.
22861 (replace_one_candidate): Remove unnecessary argument; remove
22862 unnecessary parameter from calls to introduce_cast_before_cand.
22863 (replace_profitable_candidates): Remove unnecessary parameters
22864 from calls to replace_one_candidate.
22866 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
22868 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
22869 phi def as possibly hiding a basis for a CAND_ADD whose operands
22870 have been commuted in the analysis.
22871 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
22873 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22875 * config/aarch64/aarch64.md
22876 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
22877 shift value between 0-4.
22879 2013-05-07 Richard Biener <rguenther@suse.de>
22881 * double-int.h (rshift): New overload.
22882 * double-int.c (rshift): New function.
22883 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
22884 (create_reference_ops_from_ref): Remove.
22885 (vn_reference_insert): Use shared ops for constructing the
22886 reference and copy it.
22888 2013-05-07 Richard Biener <rguenther@suse.de>
22890 PR middle-end/57190
22891 * tree-eh.c (sink_clobbers): Properly propagate
22892 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
22894 2013-05-07 Jakub Jelinek <jakub@redhat.com>
22896 PR tree-optimization/57149
22897 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
22898 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
22899 collect_phi_def_edges, execute_late_warn_uninitialized): Use
22900 uninit_undefined_value_p instead of ssa_undefined_value_p.
22903 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
22904 for modifier == EXPAND_INITIALIZER.
22906 2013-05-07 Anton Blanchard <anton@samba.org>
22908 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
22909 for powerpc64 little endian.
22910 * configure: Regenerate.
22912 2013-05-06 Graham Stott <grahams@btinternet.com>
22914 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
22915 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
22916 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
22917 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
22919 2013-05-06 Graham Stott <grahams@btinternet.com>
22921 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
22922 codes which allow non-lvalues.
22924 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
22926 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
22927 components are all 1s.
22928 (integer_minus_onep): New function.
22929 * tree.h (integer_minus_onep): Declare it.
22930 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
22931 integer_minus_onep instead of integer_all_onesp.
22933 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
22936 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
22937 variations of these patterns.
22939 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
22941 * config/i386/i386.md (isa): Add x64_sse4 member.
22942 (enabled): Handle x64_sse4.
22943 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
22944 instruction for 64bit SSE4_1 targets. Update insn attributes.
22945 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
22946 instruction for SSE4_1 targets. Update insn attributes.
22947 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
22948 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
22949 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
22951 (*vec_extractv4si): Rename from *sse4_1_pextrd.
22952 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
22953 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
22955 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
22958 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
22960 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
22962 * final.c (do_assembler_dialects): Don't handle curly braces and
22963 vertical bar escaped by % as dialect delimiters.
22964 (output_asm_insn): Print curly braces and vertical bar if escaped
22965 by % and ASSEMBLER_DIALECT defined.
22966 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
22967 * doc/tm.texi: Regenerated.
22969 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
22971 * config/mips/mips.c: Include tree-pass.h.
22972 (mips_reorg): Split in pre- and post-dbr_schedule parts.
22973 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
22974 (pass_mips_machine_reorg2): New machine specific pass.
22975 (insert_pass_mips_machine_reorg2): New pass plugin definition.
22976 (mips_option_override): Register the new pass.
22977 * rtl.h (cleanup_barriers): Remove prototype.
22978 (dbr_schedule): Likewise.
22979 * jump.c (cleanup_barriers): Make static.
22980 * reorg.c (dbr_schedule): Likewise.
22982 2013-05-06 Richard Biener <rguenther@suse.de>
22984 PR tree-optimization/57185
22985 * tree-parloops.c (add_field_for_reduction): Handle anonymous
22986 SSA names properly.
22988 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
22991 * config/i386/i386.c (add_parameter_dependencies): Add dependence
22992 between "first_arg" and "insn", not "last" and "insn".
22994 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
22996 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
22997 (find_candidates_in_block): Re-enable slsr_process_phi.
22998 (create_phi_basis): Fix double counting of candidate adjustment.
23000 2013-05-06 Richard Biener <rguenther@suse.de>
23002 PR middle-end/57147
23003 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
23004 the edge is also fallthru, preserve it and just clear the
23006 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
23007 also complex, preserve that and just clear the fallthru flag.
23008 * tree-inline.c (update_ssa_across_abnormal_edges): Also
23009 update virtual operands.
23011 2013-05-06 Alan Modra <amodra@gmail.com>
23013 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
23014 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
23015 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
23016 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
23017 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
23018 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
23020 2013-05-06 Alan Modra <amodra@gmail.com>
23022 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
23023 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
23024 (DEFAULT_ASM_ENDIAN): Define.
23025 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
23026 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
23027 Update -K PIC clause from sysv4.h.
23028 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
23029 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
23031 2013-05-06 Alan Modra <amodra@gmail.com>
23033 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
23034 twice for little-endian.
23035 (ashrdi3_no_power, ashrdi3): Support little-endian.
23037 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
23040 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
23041 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
23043 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
23044 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
23045 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
23048 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
23049 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23051 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
23052 * config/pa/pa.opt: Make mbig-switch a no-op.
23053 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
23054 (CASE_VECTOR_MODE): Always return SImode.
23055 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
23056 for the !TARGET_BIG_SWITCH case.
23057 * config/pa/pa-linux.h: Likewise.
23058 * config/pa/pa-openbsd.h: Likewise.
23059 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
23060 * config/pa/pa.md (short_jump): Remove define_insn.
23061 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
23062 (casesi0): Remove define_insn.
23063 (type): Remove btable_branch.
23064 (pa_combine_type): Likewise.
23065 (in_nullified_branch_delay): Likewise.
23066 (in_call_delay): Likewise.
23067 (define_delay): Likewise.
23068 (define_insn_reservation "Z3"): Likewise.
23069 (define_insn_reservation "Z4"): Likewise.
23070 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
23071 (pa_adjust_insn_length): Remove adjustment for btable branches.
23072 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
23075 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
23077 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
23078 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
23079 Add m->r,x alternatives.
23080 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
23081 splitters using SWI48x mode iterator.
23082 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
23083 TARGET_64BIT. Add m->x alternative.
23084 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
23085 Add o->x alternative. Enable for TARGET_SSE.
23086 (sse_storeq): Remove expander.
23087 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
23088 with memory input operand.
23089 (*vec_extractv2di_1 splitter): New.
23090 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
23091 * config/i386/i386.md (ssevecmodelower): New mode attribute.
23093 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
23095 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
23096 (INT_LOWPART): Delete.
23097 (extract_MB): Adjust.
23098 (extract_ME): Adjust.
23099 (print_operand): Adjust.
23101 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
23103 * config/rs6000/predicates.md (reg_or_add_cint_operand,
23104 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
23105 (reg_or_logical_cint_operand, easy_fp_constant,
23106 logical_const_operand): Delete "CONST_DOUBLE" case.
23107 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
23108 "HOST_BITS_PER_WIDE_INT == 64" test.
23109 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
23110 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
23111 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
23112 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
23114 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
23115 CONST_DOUBLE DImode/VOIDmode case.
23116 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
23117 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
23118 CONST_DOUBLE VOIDmode case.
23119 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
23120 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
23121 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
23122 Delete CONST_DOUBLE case.
23123 (splitters for mov FMOVE64 const_double): Delete
23124 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
23125 "HOST_BITS_PER_WIDE_INT >= 64" test.
23126 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
23128 (mov DI const_double): Delete.
23130 2013-05-04 Jakub Jelinek <jakub@redhat.com>
23132 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
23133 on op shows all bits zero in mode of a lowpart subreg, return zero.
23135 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
23138 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
23139 to save TFmode registers and DImode to save TImode registers for
23140 caller save operations.
23141 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
23142 mark being partially clobbered since they only use the first
23145 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
23146 and TDmode only use the upper 64-bits of each VSX register.
23148 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
23150 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
23151 (find_candidates_in_block): Disable slsr_process_phi.
23153 2013-05-03 Guozhi Wei <carrot@google.com>
23155 * coverage.c (coverage_obj_init): Move the construction of gcov
23157 (build_init_ctor): ... here.
23159 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
23161 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
23162 (slsr_cand_d): Redefine def_phi.
23163 (stride_status, phi_adjust_status, count_phis_status): New enums.
23164 (find_phi_def): New.
23165 (find_basis_for_base_expr): New.
23166 (find_basis_for_candidate): Handle hidden bases.
23167 (alloc_cand_and_find_basis): Handle phi candidates.
23168 (slsr_process_phi): New.
23169 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
23170 (create_mul_imm_cand): Likewise.
23171 (create_add_ssa_cand): Exclude phi base candidates.
23172 (create_add_imm_cand): Likewise.
23173 (slsr_process_cast): Likewise.
23174 (slsr_process_copy): Likewise.
23175 (find_candidates_in_block): Handle phi candidates.
23176 (dump_candidate): Likewise.
23177 (unconditional_cands): Delete.
23178 (unconditional_cands_with_known_stride_p): Delete.
23179 (phi_dependent_cand_p): New.
23180 (cand_increment): Handle phi-dependent candidates.
23181 (replace_dependent): Delete.
23182 (replace_mult_candidate): New.
23183 (replace_unconditional_candidate): New.
23184 (incr_vec_index): Move to avoid forward reference.
23185 (create_add_on_incoming_edge): New.
23186 (create_phi_basis): New.
23187 (replace_dependents): Delete.
23188 (replace_conditional_candidate): New.
23189 (phi_add_costs): New.
23190 (replace_uncond_cands_and_profitable_phis): New.
23191 (record_increment): Handle phi adjustments.
23192 (record_phi_increments): New.
23193 (record_increments): Handle phi adjustments.
23194 (phi_incr_cost): New.
23195 (lowest_cost_path): Handle phis.
23196 (total_savings): Likewise.
23197 (analyze_increments): Likewise.
23198 (ncd_with_phi): New.
23199 (ncd_of_cand_and_phis): New.
23200 (nearest_common_dominator_for_cands): Handle phi increments.
23201 (all_phi_incrs_profitable): New.
23202 (replace_profitable_candidates): Handle phi-dependent candidates.
23203 (analyze_candidates_and_replace): Likewise.
23205 2013-05-03 Teresa Johnson <tejohnson@google.com>
23208 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
23209 do not exceed REG_BR_PROB_BASE.
23211 2013-05-03 Jeff Law <law@redhat.com>
23213 PR tree-optimization/57144
23214 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
23215 operand of the condition will bit into the new type when eliminating
23216 a cast feeding a condition.
23218 2013-05-03 Jakub Jelinek <jakub@redhat.com>
23220 PR rtl-optimization/57130
23221 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
23222 of COMPARE as in_code to the recursive call if needed.
23224 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
23226 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
23227 (enabled): Handle new members.
23228 * config/i386/sse.md (*vec_concatv2si): Merge from
23229 *vec_concatv2si_sse2 and vec_concatv2si_sse.
23230 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
23232 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
23234 PR tree-optimization/57027
23235 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
23236 for fnms opportunity, check we got the prerequisite kind
23237 of tree / gimple before using accessor functions.
23239 2013-05-03 Richard Biener <rguenther@suse.de>
23241 * double-int.h (lshift): New overload without precision
23242 and arith argument.
23243 (operator *=, operator +=, operator -=): Move ...
23244 * double-int.c (operator *=, operator +=, operator -=): ... here
23245 and implement more efficiently.
23246 (mul_double_with_sign): Remove.
23247 (lshift_double): Adjust to take unsinged shift argument, push
23248 dispatching code to callers.
23249 (mul_double_wide_with_sign): Add early out for callers that
23250 are not interested in high parts or overflow.
23251 (lshift): New function.
23252 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
23253 dispatch code here.
23254 (lrotate, rrotate): Use logical shifts.
23255 * expr.c (get_inner_reference): Use lshift.
23256 * fixed-value.c (do_fixed_divide): Likewise.
23257 * tree-dfa.c (get_ref_base_and_extent): Likewise.
23258 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
23259 (indirect_refs_may_alias_p): Likewise.
23260 (stmt_kills_ref_p_1): Likewise.
23262 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
23264 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
23266 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
23268 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
23269 scalar form of FABD instruction.
23271 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
23273 * lra-constraints.c (process_alt_operands): Add checking alt
23274 number to choose the best alternative.
23276 2013-05-02 Richard Biener <rguenther@suse.de>
23278 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
23279 bitmap and its handling.
23280 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
23282 2013-05-02 Richard Biener <rguenther@suse.de>
23284 PR middle-end/57140
23285 * tree-inline.c (copy_loops): Properly handle removed loops.
23286 (copy_cfg_body): Mark destination loops for fixup if source
23287 loops needed fixup.
23289 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
23292 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
23293 generating simple_return for naked functions.
23295 2013-05-02 Martin Jambor <mjambor@suse.cz>
23297 PR middle-end/56988
23298 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
23299 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
23301 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
23302 ipa_agg_replacement_value structures.
23303 (known_aggs_to_agg_replacement_list): Likewise.
23304 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
23305 (read_agg_replacement_chain): Likewise.
23306 (ipcp_transform_function): Also check that by_ref flags match.
23308 2013-05-02 Richard Biener <rguenther@suse.de>
23310 * graphds.h (struct graph): Add obstack member.
23311 * graphds.c (new_graph): Initialize obstack and allocate
23313 (add_edge): Allocate edge from the obstack.
23314 (free_graph): Free the obstack instead of all edges and vertices.
23316 2013-05-02 Teresa Johnson <tejohnson@google.com>
23318 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
23320 * cfg.c (update_bb_profile_for_threading): Ditto.
23321 * tree-inline.c (copy_bb): Ditto.
23322 (copy_edges_for_bb): Ditto.
23323 (initialize_cfun): Ditto.
23324 (copy_cfg_body): Ditto.
23325 (expand_call_inline): Ditto.
23326 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
23327 (estimate_node_size_and_time): Ditto.
23328 (inline_merge_summary): Ditto.
23329 * cgraphclones.c (cgraph_clone_edge): Ditto.
23330 (cgraph_clone_node): Ditto.
23331 * sched-rgn.c (compute_dom_prob_ps): Ditto.
23332 (compute_trg_info): Ditto.
23334 2013-05-02 Ian Bolton <ian.bolton@arm.com>
23336 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
23337 S reg when fp attribute set.
23338 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
23340 2013-05-02 Ian Bolton <ian.bolton@arm.com>
23342 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
23344 (*and_one_cmplsi3_compare0_uxtw): Likewise.
23345 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
23346 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
23348 2013-05-02 Richard Biener <rguenther@suse.de>
23350 * tree-scalar-evolution.c (scev_info_hasher): Remove.
23351 (struct instantiate_cache_entry): New type.
23352 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
23353 (struct instantiate_cache_type): New type.
23354 (set_instantiated_value, get_instantiated_value): Remove.
23355 (get_instantiated_value_entry): New function.
23356 (instantiate_scev_name): Use the new cache and adjust.
23357 (instantiate_scev_poly): Adjust.
23358 (instantiate_scev_binary): Likewise.
23359 (instantiate_array_ref): Likewise.
23360 (instantiate_scev_convert): Likewise.
23361 (instantiate_scev_not): Likewise.
23362 (instantiate_scev_3): Likewise.
23363 (instantiate_scev_2): Likewise.
23364 (instantiate_scev_r): Likewise.
23365 (instantiate_scev): Likewise.
23366 (resolve_mixers): Likewise.
23368 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
23371 * lra-constraints.c (best_small_class_operands_num): Remove.
23372 (process_alt_operands): Remove small_class_operands_num. Take
23373 small classes operands into losers and only if the operand is not
23374 matched. Modify debugging output.
23375 (curr_insn_transform): Remove best_small_class_operands_num.
23378 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
23380 * config/aarch64/aarch64-builtins.c
23381 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
23382 * config/aarch64/aarch64-simd-builtins.def
23383 (reduc_splus_): Add new modes.
23384 (reduc_uplus_): New.
23385 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
23386 (reduc_uplus_v4sf): Likewise.
23387 (reduc_splus_v4sf): Likewise.
23388 (aarch64_addv<mode>): Likewise.
23389 (reduc_uplus_<mode>): Likewise.
23390 (reduc_splus_<mode>): Likewise.
23391 (aarch64_addvv2di): Likewise.
23392 (reduc_uplus_v2di): Likewise.
23393 (reduc_splus_v2di): Likewise.
23394 (aarch64_addvv2si): Likewise.
23395 (reduc_uplus_v2si): Likewise.
23396 (reduc_splus_v2si): Likewise.
23397 (reduc_<sur>plus_<mode>): New.
23398 (reduc_<sur>plus_v2di): Likewise.
23399 (reduc_<sur>plus_v2si): Likewise.
23400 (reduc_<sur>plus_v4sf): Likewise.
23401 (aarch64_addpv4sf): Likewise.
23402 * config/aarch64/arm_neon.h
23403 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
23404 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
23405 add UNSPEC_SADDV, UNSPEC_UADDV.
23407 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
23409 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
23411 * config/aarch64/arm_neon.h
23412 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
23414 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
23416 * config/aarch64/aarch64-builtins
23417 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
23419 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
23421 * config/aarch64/aarch64-simd-builtins.def
23422 (reduc_smax_): New.
23423 (reduc_smin_): Likewise.
23424 (reduc_umax_): Likewise.
23425 (reduc_umin_): Likewise.
23426 (reduc_smax_nan_): Likewise.
23427 (reduc_smin_nan_): Likewise.
23430 (smax): Update for V2SF, V4SF and V2DF modes.
23433 (smin_nan): Likewise.
23434 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
23435 (<su><maxmin><mode>3): ...This, refactor.
23436 (s<maxmin><mode>3): New.
23437 (<maxmin_uns><mode>3): Likewise.
23438 (reduc_<maxmin_uns>_<mode>): Refactor.
23439 (reduc_<maxmin_uns>_v4sf): Likewise.
23440 (reduc_<maxmin_uns>_v2si): Likewise.
23441 (aarch64_<fmaxmin><mode>: Remove.
23442 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
23444 (vmin<q>_f<32,64>): Likewise.
23445 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
23447 (su): Add mappings for smax, smin, umax, umin.
23449 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
23450 (FMAXMIN): Rename as...
23451 (FMAXMIN_UNS): ...This.
23453 (fmaxminv): Likewise.
23454 (fmaxmin): Likewise.
23456 (maxmin_uns_op): Likewise.
23458 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
23460 * config/aarch64/arm_neon.h
23461 (vac<ge, gt><sd>_f<32, 64>): Rename to...
23462 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
23463 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
23465 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
23467 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
23468 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
23470 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
23472 * config/aarch64/aarch64-simd.md
23473 (vcond<mode>_internal): Handle special cases for constant masks.
23474 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
23475 (vcondu<mode><mode>): Likewise.
23476 (vcond<v_cmp_result><mode>): New.
23478 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
23480 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
23481 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
23482 * config/aarch64/aarch64-simd-builtins.def
23483 (cmeq): Update to BUILTIN_VALLDI.
23488 * config/aarch64/arm_neon.h
23489 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
23490 to builtins or C as appropriate.
23492 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
23494 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
23496 (cmhi): Rename to...
23498 * config/aarch64/aarch64-simd.md
23499 (simd_mode): Add SF.
23500 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
23501 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
23502 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
23503 (cstore<mode>_neg): ...This.
23504 * config/aarch64/iterators.md
23506 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
23507 (COMPARISONS): New.
23508 (UCOMPARISONS): Likewise.
23509 (optab): Add missing comparisons.
23515 (VCMP_S): Likewise.
23516 (VCMP_U): Likewise.
23517 (V_cmp_result): Add DF, SF modes.
23518 (v_cmp_result): Likewise.
23520 (vmtype): Likewise.
23521 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
23523 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
23525 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
23526 define_insn to define_insn_and_split.
23527 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
23528 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
23529 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
23530 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
23531 (thumb2_negscc): Likewise.
23533 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
23535 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
23537 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
23539 * config/arm/thumb2.md: Remove trailing whitespaces.
23541 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
23543 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
23544 Use gen_int_mode rather than GEN_INT.
23546 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
23548 * value-prof.c (stream_in_histogram_value): Remove the strayed
23551 2013-04-30 Richard Biener <rguenther@suse.de>
23553 PR middle-end/57122
23554 * cfghooks.c (split_edge): Properly check for the loop latch edge.
23556 2013-04-30 Richard Biener <rguenther@suse.de>
23558 PR middle-end/57107
23559 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
23561 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
23563 PR rtl-optimization/56957
23564 PR rtl-optimization/57105
23565 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
23566 variable. Use just INSN_UID for determining whether an insn
23567 should be only disconnected from the insn stream.
23568 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
23570 2013-04-30 Jakub Jelinek <jakub@redhat.com>
23572 PR tree-optimization/57104
23573 * tsan.c (instrument_expr): Don't instrument accesses to
23574 DECL_HARD_REGISTER VAR_DECLs.
23576 2013-04-30 Richard Biener <rguenther@suse.de>
23578 * function.h (loops_for_fn): New inline function.
23579 (set_loops_for_fn): Likewise.
23580 * cfgloop.h (place_new_loop): Add struct function parameter.
23581 (get_loop): Likewise.
23582 (get_loops): Likewise.
23583 (number_of_loops): Likewise.
23584 (fel_next): Adjust.
23585 (fel_init): Likewise.
23586 * cfg.c (get_loop_copy): Adjust.
23587 * cfgloop.c (flow_loops_dump): Likewise.
23588 (record_loop_exits): Likewise.
23589 (verify_loop_structure): Likewise.
23590 * cfgloopanal.c (mark_irreducible_loops): Likewise.
23591 (estimate_reg_pressure_cost): Likewise.
23592 (mark_loop_exit_edges): Likewise.
23593 * cfgloopmanip.c (place_new_loop): Likewise.
23594 (add_loop): Likewise.
23595 (duplicate_loop): Likewise.
23596 * graph.c (draw_cfg_nodes): Likewise.
23597 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
23598 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
23599 (extract_affine_chrec): Likewise.
23600 (build_scop_iteration_domain): Likewise.
23601 * graphite.c (graphite_initialize): Likewise.
23602 * ira-build.c (create_loop_tree_nodes): Likewise.
23603 (more_one_region_p): Likewise.
23604 (rebuild_regno_allocno_maps): Likewise.
23605 (mark_loops_for_removal): Likewise.
23606 (mark_all_loops_for_removal): Likewise.
23607 (remove_unnecessary_regions): Likewise.
23608 (ira_build): Likewise.
23609 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
23610 * loop-init.c (fix_loop_structure): Likewise.
23611 (gate_rtl_move_loop_invariants): Likewise.
23612 (gate_rtl_unswitch): Likewise.
23613 (gate_rtl_unroll_and_peel_loops): Likewise.
23614 (rtl_doloop): Likewise.
23615 * lto-streamer-in.c (input_cfg): Likewise.
23616 * lto-streamer-out.c (output_cfg): Likewise.
23617 * modulo-sched.c (sms_schedule): Likewise.
23618 * predict.c (tree_estimate_probability): Likewise.
23619 (tree_estimate_probability_driver): Likewise.
23620 (estimate_loops): Likewise.
23621 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
23622 (move_sese_region_to_fn): Likewise.
23623 (debug_loop_num): Likewise.
23624 * tree-chrec.c (chrec_evaluate): Likewise.
23625 (hide_evolution_in_other_loops_than_loop): Likewise.
23626 (chrec_component_in_loop_num): Likewise.
23627 (reset_evolution_in_loop): Likewise.
23628 (evolution_function_is_invariant_rec_p): Likewise.
23629 * tree-if-conv.c (main_tree_if_conversion): Likewise.
23630 * tree-inline.c (copy_loops): Likewise.
23631 (copy_cfg_body): Likewise.
23632 (tree_function_versioning): Likewise.
23633 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
23634 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
23636 (add_to_evolution_1): Likewise.
23637 (scev_const_prop): Likewise.
23638 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
23639 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
23640 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
23641 (tree_ssa_lim_initialize): Likewise.
23642 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
23643 (verify_loop_closed_ssa): Likewise.
23644 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
23645 (tree_ssa_loop_im): Likewise.
23646 (tree_ssa_loop_unswitch): Likewise.
23647 (tree_vectorize): Likewise.
23648 (check_data_deps): Likewise.
23649 (tree_ssa_loop_ivcanon): Likewise.
23650 (tree_ssa_loop_bounds): Likewise.
23651 (tree_complete_unroll): Likewise.
23652 (tree_complete_unroll_inner): Likewise.
23653 (tree_parallelize_loops): Likewise.
23654 (tree_ssa_loop_prefetch): Likewise.
23655 (tree_ssa_loop_ivopts): Likewise.
23656 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
23657 * tree-vectorizer.c (vectorize_loops): Likewise.
23659 2013-04-29 Mike Frysinger <vapier@gentoo.org>
23661 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
23662 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
23663 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
23664 with EABI_LINK_SPEC.
23666 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
23669 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
23672 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
23675 * lra-constraints.c (process_alt_operands): Discourage a bit more
23676 using memory for pseudos. Print cost dump for alternatives.
23677 Modify cost values for conflicts with early clobbers.
23678 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
23680 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
23683 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
23685 2013-04-29 Ian Bolton <ian.bolton@arm.com>
23687 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
23688 from/to S register.
23689 (movdi_aarch64): Support LDR/STR from/to D register.
23691 2013-04-29 Ian Bolton <ian.bolton@arm.com>
23693 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
23694 or higher by default.
23696 2013-04-29 Richard Biener <rguenther@suse.de>
23698 PR middle-end/57075
23699 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
23700 even if not adding abnormal edges for calls that can make
23703 2013-04-29 Richard Biener <rguenther@suse.de>
23705 PR middle-end/57103
23706 * tree-cfg.c (move_stmt_op): Fix condition under which to update
23708 (move_stmt_r): Remove redundant checking.
23710 2013-04-29 Teresa Johnson <tejohnson@google.com>
23713 * basic-block.h (apply_scale): New function.
23714 (apply_probability): Use apply_scale.
23715 * gimple-streamer-in.c (input_bb): Ditto.
23716 * lto-streamer-in.c (input_cfg): Ditto.
23717 * lto-cgraph.c (merge_profile_summaries): Ditto.
23718 * tree-optimize.c (execute_fixup_cfg): Ditto.
23719 * tree-inline.c (copy_bb): Update comment to use apply_scale.
23720 (copy_edges_for_bb): Ditto.
23721 (copy_cfg_body): Ditto.
23723 2013-04-29 Tom de Vries <tom@codesourcery.com>
23725 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
23726 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
23727 (tail_merge_optimize): Handle current_loops == NULL.
23729 2013-04-26 Jeff Law <law@redhat.com>
23731 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
23732 (simplify_cond_using_ranges): Generalize code to simplify
23733 COND_EXPRs where one argument is a constant and the other
23734 is an SSA_NAME created by an integral type conversion.
23736 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23738 * config/arm/arm.md (store_minmaxsi): Use only when
23739 optimize_insn_for_size_p.
23741 2013-04-29 Christian Bruel <christian.bruel@st.com>
23744 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
23746 2013-04-29 Richard Biener <rguenther@suse.de>
23748 PR middle-end/57089
23749 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
23750 loop tree make sure to schedule a fixup for the child as well.
23751 (expand_omp_for_generic): Properly add loops.
23752 (expand_omp_for_static_nochunk): Likewise.
23753 (expand_omp_for_static_chunk): Likewise.
23754 (expand_omp_for): For the degenerate case fixup loops.
23755 (expand_omp_sections): Fix default bb placement in loops.
23756 (expand_omp_atomic_pipeline): Properly add loops.
23758 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23760 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
23762 2013-04-29 Tom de Vries <tom@codesourcery.com>
23764 * tree-ssa-tail-merge.c: Update header comment.
23766 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
23768 * config/aarch64/arm_neon.h
23769 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
23770 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
23771 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
23772 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
23773 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
23774 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
23775 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
23776 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
23778 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
23780 * config/aarch64/aarch64-simd.md
23781 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
23782 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
23783 fix_trunc, fixuns_trunc.
23784 (ftrunc<VDQF:mode>2): New.
23785 * config/aarch64/iterators.md (optab): Add fix, fixuns.
23786 (fix_trunc_optab): New.
23788 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
23790 * config/aarch64/aarch64-builtins.c
23791 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
23792 iceilf, lround, iroundf.
23794 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
23797 * config/i386/i386.h (enum ix86_tune_indices)
23798 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
23799 New, split from X86_TUNE_INTER_UNIT_MOVES.
23800 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
23801 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
23802 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
23803 (TARGET_INTER_UNIT_MOVES): Remove.
23804 * config/i386/i386.c (initial_ix86_tune_features): Update.
23805 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
23806 (ix86_expand_convert_uns_didf_sse): Use
23807 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
23808 (ix86_expand_vector_init_one_nonzero): Ditto.
23809 (ix86_expand_vector_init_interleave): Ditto.
23810 (inline_secondary_memory_needed): Return true for moves from SSE class
23811 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
23812 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
23813 * config/i386/constraints.md (Yi, Ym): Depend on
23814 TARGET_INTER_UNIT_MOVES_TO_VEC.
23815 (Yj, Yn): New constraints.
23816 * config/i386/i386.md (*movdi_internal): Change constraints of
23817 operand 1 from Yi to Yj and from Ym to Yn.
23818 (*movsi_internal): Ditto.
23819 (*movdf_internal): Ditto.
23820 (*movsf_internal): Ditto.
23821 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
23822 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
23823 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
23824 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
23825 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
23826 * config/i386/sse.md (movdi_to_sse): Ditto.
23827 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
23828 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
23829 TARGET_INTER_UNIT_MOVES.
23830 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
23831 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
23832 instead of TARGET_INTER_UNIT_MOVES.
23833 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
23834 operand 1 from Yi to Yj and from Ym to Yn.
23836 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
23838 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
23839 (float_truncate_hi_): Likewise.
23840 (float_extend_lo_): Likewise.
23841 (float_truncate_lo_): Likewise.
23842 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
23843 (aarch64_float_extend_lo_v2df): Likewise.
23844 (vec_unpacks_hi_v4sf): Likewise.
23845 (aarch64_float_truncate_lo_v2sf): Likewise.
23846 (aarch64_float_truncate_hi_v4sf): Likewise.
23847 (vec_pack_trunc_v2df): Likewise.
23848 (vec_pack_trunc_df): Likewise.
23850 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
23852 * config/aarch64/aarch64-builtins.c
23853 (aarch64_fold_builtin): Fold float conversions.
23854 * config/aarch64/aarch64-simd-builtins.def
23855 (floatv2si, floatv4si, floatv2di): New.
23856 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
23857 * config/aarch64/aarch64-simd.md
23858 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
23859 * config/aarch64/iterators.md (FLOATUORS): New.
23860 (optab): Add float, floatuns.
23861 (su_optab): Likewise.
23863 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
23865 * config/aarch64/aarch64-builtins.c
23866 (aarch64_builtin_vectorized_function): Use new names for
23868 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
23869 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
23870 (fcvtzu): Split as...
23871 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
23872 (fcvtas): Split as...
23873 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
23874 (fcvtau): Split as...
23875 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
23876 (fcvtps): Split as...
23877 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
23878 (fcvtpu): Split as...
23879 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
23880 (fcvtms): Split as...
23881 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
23882 (fcvtmu): Split as...
23883 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
23884 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
23885 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
23886 (lfrintnusf, lfrintnudf): Likewise.
23887 * config/aarch64/aarch64-simd.md
23888 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
23890 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
23891 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
23892 (fcvt_pattern): Likewise.
23894 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
23896 * config/aarch64/aarch64-simd.md
23897 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
23898 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
23900 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
23902 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
23903 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
23904 (vrnd<a,m,n,p>_f32): Implement using builtins.
23905 (vrnd<i,x><q>_f<32, 64>): New.
23907 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
23909 * config/aarch64/aarch64-builtins.c
23910 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
23911 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
23912 (frintz): Rename to...
23914 (frintp): Rename to...
23916 (frintm): Rename to...
23918 (frinti): Rename to...
23919 (nearbyint): ...this.
23920 (frintx): Rename to...
23922 (frinta): Rename to...
23924 * config/aarch64/aarch64-simd.md
23925 (aarch64_frint<frint_suffix><mode>): Delete.
23926 (<frint_pattern><mode>2): Convert to insn.
23927 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
23928 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
23929 (frint_pattern): Likewise.
23930 (frint_suffix): Likewise.
23932 2013-04-29 Richard Biener <rguenther@suse.de>
23934 PR tree-optimization/57081
23935 * loop-init.c: Include tree-flow.h.
23936 (loop_optimizer_finalize): Free number of iteration estimates.
23937 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
23939 2013-04-29 Jakub Jelinek <jakub@redhat.com>
23941 PR tree-optimization/57083
23942 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
23943 non-singleton shift count range, zero extend low_bound for uns case.
23945 * config/i386/predicates.md (general_vector_operand): New predicate.
23946 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
23947 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
23948 if they aren't nonimmediate operands. If their original values
23949 satisfy const_vector_equal_evenodd_p, don't shift them.
23950 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
23951 predicates. For the SSE4.1 case force operands[{1,2}] into registers
23952 if not nonimmediate_operand.
23953 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
23954 instead of register_operand.
23955 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
23957 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
23959 * stor-layout.c (finalize_size_functions): Allocate a structure and
23960 reset cfun before dumping the functions.
23962 2013-04-27 Jakub Jelinek <jakub@redhat.com>
23964 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
23967 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
23968 use xop_pmacsdqh if uns_p.
23969 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
23970 the immediate rotate count.
23972 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
23974 * rtl.h (struct rtx_def): Add comment for field jump.
23975 (LRA_SUBREG_P): New macro.
23976 * recog.c (register_operand): Check LRA_SUBREG_P.
23977 * lra.c (lra): Add note at the end of RTL code. Align non-empty
23979 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
23980 (lra_final_code_change): Skip subreg change for operators.
23981 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
23982 if there are no operand changes.
23983 * lra-constraints.c (curr_insn_set): New.
23984 (match_reload): Set LRA_SUBREG_P.
23985 (emit_spill_move): Ditto.
23986 (check_and_process_move): Use curr_insn_set. Process only single
23987 set insns. Don't initialize sec_mem_p and change_p.
23988 (simplify_operand_subreg): Use LRA_SUBREG_P.
23989 (reg_in_class_p): New function.
23990 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
23991 of #ifdef. Add code to remove cycling.
23992 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
23993 non-null disp. Reload inner instead of disp when base and index
23994 are null. Try to put lo_sum into register.
23995 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
23996 (check_and_process_move): Move code for move cost check to
23997 simple_move_p. Remove equiv_substitution.
23998 (simple_move_p): New function.
23999 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
24000 curr_insn_set. Call check_and_process_move only for single set
24001 insns. Use the new function. Move call of check_and_process_move
24002 after operand equiv substitution and address process.
24004 2013-04-26 Jakub Jelinek <jakub@redhat.com>
24007 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
24008 with nonlocal goto receivers or returns twice calls, ignore
24009 unininitialized values from abnormal edges to nl goto receiver
24010 or returns twice call.
24012 2013-04-26 Jakub Jelinek <jakub@redhat.com>
24014 PR tree-optimization/57051
24015 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
24016 and VEC_RSHIFT_EXPR if shift count is a multiple of element
24019 2013-04-26 Richard Biener <rguenther@suse.de>
24021 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
24022 (expand_omp_taskreg): Likewise. Mark loops for fixup.
24023 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
24024 (fixup_loop_arrays_after_move): New function.
24025 (move_sese_region_to_fn): Properly outline the loop tree parts
24026 of the SESE region.
24028 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
24030 * config/i386/i386.md (type, unit): Fix long lines.
24032 2013-04-26 Richard Biener <rguenther@suse.de>
24034 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
24035 (lto-streamer-out.o): Likewise.
24036 * cfgloop.c (init_loops_structure): Export, add struct function
24037 argument and adjust.
24038 (flow_loops_find): Adjust.
24039 * cfgloop.h (enum loop_estimation): Add EST_LAST.
24040 (init_loops_structure): Declare.
24041 * lto-streamer-in.c: Include cfgloop.h.
24042 (input_cfg): Input the loop tree.
24043 * lto-streamer-out.c: Include cfgloop.h.
24044 (output_cfg): Output the loop tree.
24045 (output_struct_function_base): Do not drop PROP_loops.
24047 2013-03-26 Richard Biener <rguenther@suse.de>
24049 * tree-cfg.c (execute_build_cfg): Build the loop tree.
24050 (pass_build_cfg): Provide PROP_loops.
24051 (move_sese_region_to_fn): Remove loops that are outlined into fn
24053 * tree-inline.c: Include cfgloop.h.
24054 (initialize_cfun): Do not drop PROP_loops.
24055 (copy_loops): New function.
24056 (copy_cfg_body): Copy loop structure.
24057 (tree_function_versioning): Initialize destination loop tree.
24058 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
24059 (pass_parallelize_loops): Do IL verification.
24060 * loop-init.c (loop_optimizer_init): Fixup loops if required.
24061 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
24062 the CFG make sure we fixup loops as well.
24063 * tree-ssa-tail-merge.c: Include cfgloop.h.
24064 (replace_block_by): When merging loop latches mark loops for fixup.
24065 * lto-streamer-out.c (output_struct_function_base): Drop
24066 PROP_loops for now.
24067 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
24068 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
24069 * ipa-split.c: Include cfgloop.h.
24070 (split_function): Add the new return block to the loop tree root.
24071 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
24072 whether we have removed the forwarder block.
24073 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
24074 * cfgloop.h (place_new_loop): Declare.
24075 * cfgloopmanip.c (place_new_loop): Export.
24076 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
24077 (tree-switch-conversion.o): Likewise.
24078 (tree-complex.o): Likewise.
24079 (tree-inline.o): Likewise.
24080 (tree-ssa-tailmerge.o): Likewise.
24081 (ipa-split.o): Likewise.
24082 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
24083 (tree-ssa-copy.o): Likewise.
24084 * tree-switch-conversion.c: Include cfgloop.h
24085 (process_switch): If we emit a bit-test cascade, schedule loops
24087 * tree-complex.c: Include cfgloop.h.
24088 (expand_complex_div_wide): Properly add new basic-blocks to loops.
24089 * asan.c: Include cfgloop.h.
24090 (create_cond_insert_point): Properly add new basic-blocks to
24091 loops, schedule loop fixup.
24092 * cfgloop.c (verify_loop_structure): Check that looks are not
24094 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
24096 (expand_omp_for_generic): Likewise.
24097 (expand_omp_sections): Likewise.
24098 (expand_omp_atomic_pipeline): Schedule loops for fixup.
24099 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
24100 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
24101 is initialized, not when loops are present.
24102 * tree-parloops.c (parallelize_loops): Remove checking here.
24103 * passes.c (init_optimization_passes): Schedule a copy-propagation
24104 pass before complete unrolling of inner loops.
24106 2013-04-26 Jakub Jelinek <jakub@redhat.com>
24108 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
24109 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
24110 (colorize_init): Add argument to _WIN32 version.
24111 * toplev.c: Include diagnostic-color.h.
24112 (process_options): Default to -fdiagnostics-color=auto if
24113 GCC_COLORS env var is in the environment.
24114 * common.opt (fdiagnostics-color=): Add Var and Init.
24115 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
24116 env var is in the environment, the default is auto rather than never.
24118 * diagnostic.h (file_name_as_prefix): Add context argument.
24119 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
24120 the string as locus.
24121 * langhooks.c (lhd_print_error_function): Adjust caller.
24123 2013-04-25 Lawrence Crowl <crowl@google.com>
24125 * var-tracking.c (shared_hash_def::htab):
24126 Change type to hash_table. Update dependent calls and types.
24128 2013-04-25 Lawrence Crowl <crowl@google.com>
24130 * Makefile.in: Update as needed below.
24132 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
24133 Move declaration to after the type's method definitons.
24135 * attribs.c (htab_t scoped_attributes::attribute_hash):
24136 Change type to hash_table. Update dependent calls and types.
24138 * bitmap.c (htab_t bitmap_desc_hash):
24139 Change type to hash_table. Update dependent calls and types.
24141 * cselib.c (htab_t cselib_hash_table):
24142 Change type to hash_table. Update dependent calls and types.
24144 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
24145 (hash_string_slot_node): Move implementation into lto-streamer.h
24146 struct string_slot_hasher.
24147 (eq_string_slot_node): Likewise.
24149 * data-streamer-out.c: Update output_block::string_hash_table
24150 dependent calls and types.
24152 * dwarf2cfi.c (htab_t trace_index):
24153 Change type to hash_table. Update dependent calls and types.
24155 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
24156 Change type to hash_table. Update dependent calls and types.
24157 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
24158 (htab_t optimize_external_refs::map): Likewise.
24159 (htab_t output_comp_unit::extern_map): Likewise.
24160 (htab_t output_comdat_type_unit::extern_map): Likewise.
24161 (htab_t output_macinfo::macinfo_htab): Likewise.
24162 (htab_t optimize_location_lists::htab): Likewise.
24163 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
24165 * except.c (htab_t ehspec_hash_type):
24166 Change type to hash_table. Update dependent calls and types.
24167 (assign_filter_values::ttypes): Likewise.
24168 (assign_filter_values::ehspec): Likewise.
24169 (sjlj_assign_call_site_values::ar_hash): Likewise.
24170 (convert_to_eh_region_ranges::ar_hash): Likewise.
24172 * gcse.c (htab_t pre_ldst_table):
24173 Change type to hash_table. Update dependent calls and types.
24175 * ggc-common.c (htab_t saving_htab):
24176 Change type to hash_table. Update dependent calls and types.
24177 (htab_t loc_hash): Likewise.
24178 (htab_t ptr_hash): Likewise.
24179 (call_count): Rename ggc_call_count.
24180 (call_alloc): Rename ggc_call_alloc.
24181 (loc_descriptor): Rename make_loc_descriptor.
24182 (add_statistics): Rename ggc_add_statistics.
24184 * ggc-common.c (saving_htab):
24185 Change type to hash_table. Update dependent calls and types.
24187 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
24188 (push_gimplify_context): Likewise.
24189 (pop_gimplify_context): Likewise.
24190 (struct gimple_temp_hash_elt): Added.
24191 (struct gimplify_hasher): Likewise.
24192 (struct gimplify_ctx.temp_htab):
24193 Change type to hash_table. Update dependent calls and types.
24195 * gimple-fold.c: Include gimplify-ctx.h.
24197 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
24198 Change type to hash_table. Update dependent calls and types.
24199 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
24200 avoid potential global name collision.
24202 * gimplify.c: Include gimplify-ctx.h.
24203 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
24204 (htab_t gimplify_ctx::temp_htab):
24205 Update dependent calls and types for new type hash_table.
24206 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
24207 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
24209 * gimplify-ctx.h: New.
24210 (struct gimple_temp_hash_elt): Move from gimplify.c.
24211 (class gimplify_hasher): New.
24212 (struct gimplify_ctx): Move from gimple.h.
24213 (htab_t gimplify_ctx::temp_htab):
24214 Change type to hash_table. Update dependent calls and types.
24216 * graphite-clast-to-gimple.c: Include graphite-htab.h.
24217 (htab_t ivs_params::newivs_index):
24218 Change type to hash_table. Update dependent calls and types.
24219 (htab_t ivs_params::params_index): Likewise.
24220 (htab_t print_generated_program::params_index): Likewise.
24221 (htab_t gloog::newivs_index): Likewise.
24222 (htab_t gloog::params_index): Likewise.
24224 * graphite.c: Include graphite-htab.h.
24225 4htab_t graphite_transform_loops::bb_pbb_mapping):
24226 Change type to hash_table. Update dependent calls and types.
24228 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
24229 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
24230 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
24232 * graphite-dependences.c: Include graphite-htab.h.
24233 (loop_is_parallel_p): Change hash table type of parameter.
24235 * graphite-htab.h: New.
24236 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
24237 (extern find_pbb_via_hash): Move from graphite-poly.h.
24238 (extern loop_is_parallel_p): Move from graphite-poly.h.
24239 (extern get_loop_body_pbbs): Move from graphite-poly.h.
24241 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
24242 (extern loop_is_parallel_p): Move to graphite-htab.h.
24243 (extern get_loop_body_pbbs): Move to graphite-htab.h.
24245 * haifa-sched.c (htab_t delay_htab):
24246 Change type to hash_table. Update dependent calls and types.
24247 (htab_t delay_htab_i2): Likewise.
24249 * ira-color.c (htab_t allocno_hard_regs_htab):
24250 Change type to hash_table. Update dependent calls and types.
24252 * ira-costs.c (htab_t cost_classes_htab):
24253 Change type to hash_table. Update dependent calls and types.
24255 * loop-invariant.c (htab_t merge_identical_invariants::eq):
24256 Change type to hash_table. Update dependent calls and types.
24258 * loop-iv.c (htab_t bivs):
24259 Change type to hash_table. Update dependent calls and types.
24261 * loop-unroll.c (htab_t opt_info::insns_to_split):
24262 Change type to hash_table. Update dependent calls and types.
24263 (htab_t opt_info::insns_with_var_to_expand): Likewise.
24265 * lto-streamer.h (struct string_slot): Move from data-streamer.h
24266 (struct string_slot_hasher): New.
24267 (htab_t output_block::string_hash_table):
24268 Change type to hash_table. Update dependent calls and types.
24270 * lto-streamer-in.c (freeing_string_slot_hasher): New.
24271 (htab_t file_name_hash_table):
24272 Change type to hash_table. Update dependent calls and types.
24274 * lto-streamer-out.c: Update output_block::string_hash_table dependent
24277 * lto-streamer.c (htab_t tree_htab):
24278 Change type to hash_table. Update dependent calls and types.
24280 * omp-low.c: Include gimplify-ctx.h.
24282 * passes.c (htab_t name_to_pass_map):
24283 Change type to hash_table. Update dependent calls and types.
24284 (pass_traverse): Rename to passes_pass_traverse.
24286 * plugin.c (htab_t event_tab):
24287 Change type to hash_table. Update dependent calls and types.
24289 * postreload-gcse.c (htab_t expr_table):
24290 Change type to hash_table. Update dependent calls and types.
24291 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
24293 * sese.c (debug_rename_map_1): Make extern.
24294 (htab_t copy_bb_and_scalar_dependences::rename_map):
24295 Change type to hash_table. Update dependent calls and types.
24297 * sese.h (extern debug_rename_map): Move to .c file.
24299 * store-motion.c (htab_t store_motion_mems_table):
24300 Change type to hash_table. Update dependent calls and types.
24302 * trans-mem.c (htab_t tm_new_mem_hash):
24303 Change type to hash_table. Update dependent calls and types.
24305 * tree-browser.c (htab_t TB_up_ht):
24306 Change type to hash_table. Update dependent calls and types.
24308 * tree-cfg.c (htab_t discriminator_per_locus):
24309 Change type to hash_table. Update dependent calls and types.
24311 * tree-complex.c: Include tree-hasher.h
24312 (htab_t complex_variable_components):
24313 Change type to hash_table. Update dependent calls and types.
24315 * tree-eh.c (htab_t finally_tree):
24316 Change type to hash_table. Update dependent calls and types.
24318 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
24319 struct int_tree_hasher.
24320 (extern int_tree_map_eq): Likewise.
24321 (uid_decl_map_hash): Removed.
24322 (extern decl_tree_map_eq): Likewise.
24324 * tree-hasher.h: New.
24325 (struct int_tree_hasher): New.
24326 (typedef int_tree_htab_type): New.
24328 * tree-inline.c: Include gimplify-ctx.h.
24330 * tree-mudflap.c: Include gimplify-ctx.h.
24332 * tree-parloops.c: Include tree-hasher.h.
24333 (htab_t eliminate_local_variables_stmt::decl_address):
24334 Change type to hash_table. Update dependent calls and types.
24335 (htab_t separate_decls_in_region::decl_copies): Likewise.
24337 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
24338 Change type to hash_table. Update dependent calls and types.
24340 * tree-sra.c (candidates):
24341 Change type to hash_table. Update dependent calls and types.
24343 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
24345 (int_tree_map_hash): Likewise.
24347 * tree-ssa-dom.c (htab_t avail_exprs):
24348 Change type to hash_table. Update dependent calls and types.
24350 * tree-ssa-live.c (var_map_base_init::tree_to_index):
24351 Change type to hash_table. Update dependent calls and types.
24353 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
24354 Change type to hash_table. Update dependent calls and types.
24356 * tree-ssa-phiopt.c (seen_ssa_names):
24357 Change type to hash_table. Update dependent calls and types.
24359 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
24360 Change type to hash_table. Update dependent calls and types.
24362 * tree-ssa-uncprop.c (equiv):
24363 Change type to hash_table. Update dependent calls and types.
24365 2013-04-25 Jakub Jelinek <jakub@redhat.com>
24367 PR rtl-optimization/57003
24368 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
24369 call note_stores with kill_clobbered_value callback again after
24370 killing regs_invalidated_by_call.
24372 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
24374 * config/aarch64/aarch64-simd.md
24375 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
24376 (aarch64_simd_bsl<mode>): Likewise.
24377 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
24379 2013-04-25 Marek Polacek <polacek@redhat.com>
24381 PR tree-optimization/57066
24382 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
24384 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
24386 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
24388 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
24390 * config/aarch64/aarch64-builtins.c
24391 (aarch64_fold_builtin): New.
24392 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
24393 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
24394 * config/aarch64/aarch64-simd-builtins.def (abs): New.
24395 * config/aarch64/arm_neon.h
24396 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
24398 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
24399 Tejas Belagod <tejas.belagod@arm.com>
24401 * config/aarch64/aarch64-builtins.c
24402 (aarch64_gimple_fold_builtin): New.
24403 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
24404 * config/aarch64/aarch64-simd-builtins.def (addv): New.
24405 * config/aarch64/aarch64-simd.md (addpv4sf): New.
24406 (addvv4sf): Update.
24407 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
24409 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
24411 * config/aarch64/aarch64.md
24412 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
24414 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
24416 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
24417 (*ngcsi_uxtw): New pattern.
24419 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24420 Julian Brown <julian@codesourcery.com>
24422 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
24423 (TB_DREG): Add T_V4HF.
24424 (v4hf_UP): New macro.
24425 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
24426 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
24427 Handle initialisation of V4HF. Adjust initialisation of reinterpret
24429 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
24430 (arm_vector_mode_supported_p): Handle V4HF.
24431 (arm_mangle_map): Handle V4HFmode.
24432 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
24433 * config/arm/arm_neon_builtins.def: Add entries for
24434 vcvtv4hfv4sf, vcvtv4sfv4hf.
24435 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
24436 (neon_vcvtv4hfv4sf): Likewise.
24437 * config/arm/neon-gen.ml: Handle half-precision floating point
24439 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
24440 * config/arm/arm_neon.h: Regenerate.
24441 * config/arm/neon.ml (type elts): Add F16.
24442 (type vectype): Add T_float16x4, T_floatHF.
24443 (type vecmode): Add V4HF.
24444 (type features): Add Requires_FP_bit feature.
24445 (elt_width): Handle F16.
24446 (elt_class): Likewise.
24447 (elt_of_class_width): Likewise.
24448 (mode_of_elt): Refactor.
24449 (type_for_elt): Handle F16, fix error messages.
24450 (vectype_size): Handle T_float16x4.
24451 (vcvt_sh): New function.
24452 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
24453 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
24454 (string_of_mode): Handle V4HF.
24455 * doc/arm-neon-intrinsics.texi: Regenerate.
24457 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
24459 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
24460 format specifier in 'X' case.
24462 2013-04-25 Alan Modra <amodra@gmail.com>
24465 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
24466 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
24467 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
24468 Repeat for many other rotate/shift and mask patterns using subregs.
24469 Name lshiftrt insns.
24470 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
24471 on WORDS_BIG_ENDIAN.
24473 2013-04-25 Alan Modra <amodra@gmail.com>
24475 * config.gcc: Support little-endian powerpc-linux targets.
24476 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
24477 (LINK_OS_LINUX_SPEC): Define.
24478 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
24479 Preserve MASK_LITTLE_ENDIAN.
24480 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
24481 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
24482 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
24483 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
24484 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
24485 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
24486 Correct fp word order for little-endian. Don't shift toc entries
24487 smaller than a word for little-endian.
24488 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
24489 (bswapdi2 splits): Correct low-part subreg for little-endian.
24490 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
24491 low/high where such is correct only for be.
24492 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
24493 little-endian for -mcall-aixdesc.
24495 2013-04-25 Alan Modra <amodra@gmail.com>
24497 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
24498 replace_equiv_address_nv.
24500 2013-04-25 Alan Modra <amodra@gmail.com>
24502 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
24504 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
24507 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
24508 * rtl.h (struct rtx_def): ...
24510 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
24512 PR rtl-optimizations/57046
24513 * lra-constraints (split_reg): Set up lra_risky_transformations_p
24514 for multi-reg splits.
24516 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
24518 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
24520 2013-04-24 Sterling Augustine <saugustine@google.com>
24522 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
24523 (comp_dir_string, debug_str_dwo_section): New.
24524 (DEBUG_STR_DWO_SECTION): Rename to ...
24525 (DEBUG_DWO_STR_SECTION): ... this.
24526 (DEBUG_NORM_STR_SECTION): Delete.
24527 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
24528 (DEBUG_STR_DWO_SECTION_FLAGS): New.
24529 (find_AT_string): Move most logic to ...
24530 (find_AT_string_in_table): ... here. New.
24531 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
24532 add_skeleton_AT_string. Delete logic.
24533 (output_skeleton_debug_sections): Remove call to
24534 add_top_level_skeleton_die_attrs.
24535 (add_comp_dir_attribute): Move logic to comp_dir_string.
24536 (dwarf2out_init): Initialize debug_str_dwo_section.
24537 (output_indirect_string): Call find_string_form.
24538 (output_indirect_strings): Rewrite.
24539 (prune_unused_types): Empty skeleton_debug_str_hash.
24540 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
24541 (dwarf2out_finish): Call output_indirect_strings.
24543 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
24545 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
24547 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
24549 * rtl.h (struct rtx_def): Add comment for field jump.
24550 (LRA_SUBREG_P): New macro.
24551 * recog.c (register_operand): Check LRA_SUBREG_P.
24552 * lra.c (lra): Add note at the end of RTL code. Align non-empty
24554 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
24555 (lra_final_code_change): Skip subreg change for operators.
24556 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
24557 if there are no operand changes.
24558 * lra-constraints.c (curr_insn_set): New.
24559 (match_reload): Set LRA_SUBREG_P.
24560 (emit_spill_move): Ditto.
24561 (check_and_process_move): Use curr_insn_set. Process only single
24562 set insns. Don't initialize sec_mem_p and change_p.
24563 (simplify_operand_subreg): Use LRA_SUBREG_P.
24564 (reg_in_class_p): New function.
24565 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
24566 of #ifdef. Add code to remove cycling.
24567 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
24568 non-null disp. Reload inner instead of disp when base and index
24569 are null. Try to put lo_sum into register.
24570 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
24571 (check_and_process_move): Move code for move cost check to
24572 simple_move_p. Remove equiv_substitution.
24573 (simple_move_p): New function.
24574 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
24575 curr_insn_set. Call check_and_process_move only for single set
24576 insns. Use the new function. Move call of check_and_process_move
24577 after operand equiv substitution and address process.
24579 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
24581 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
24582 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
24583 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
24585 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
24587 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
24589 2013-04-24 Marek Polacek <polacek@redhat.com>
24591 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
24592 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
24593 (select_loops_exit_conditions): Likewise.
24594 (number_of_iterations_for_all_loops): Likewise.
24595 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
24596 (scev_analysis): Likewise.
24598 2013-04-02 Catherine Moore <clm@codesourcery.com>
24599 Chao-ying Fu <fu@mips.com>
24601 * config/mips/micromips.md (jraddiusp): New pattern.
24602 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
24603 instruction if possible.
24605 2013-04-24 Alan Modra <amodra@gmail.com>
24607 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
24609 2013-04-24 Julian Brown <julian@codesourcery.com>
24610 Chung-Lin Tang <cltang@codesourcery.com>
24612 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
24613 dependency behavior in enumeration type DIE generation. Add TODO note
24614 to comments about future DW_FORM_sdata/udata re-work of related code.
24616 2013-04-23 Lawrence Crowl <crowl@google.com>
24618 * Makefile.in: Update as needed below.
24620 * hash-table.h (class hash_table):
24621 Correct many methods with parameter types compare_type to the correct
24622 value_type. (Correct code was unlikely to notice the change.)
24623 (hash_table::elements_with_deleted) New.
24624 (class hashtable::iterator): New.
24625 (hashtable::begin()): New.
24626 (hashtable::end()): New.
24627 (FOR_EACH_HASH_TABLE_ELEMENT): New.
24629 * statistics.c (statistics_hashes):
24630 Change type to hash_table. Update dependent calls and types.
24632 * tree-into-ssa.c (var_infos):
24633 Change type to hash_table. Update dependent calls and types.
24635 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
24636 Change type to hash_table. Update dependent calls and types.
24638 * tree-ssa-loop-im.c (struct mem_ref.refs):
24639 Change type to hash_table. Update dependent calls and types.
24641 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
24642 Change type to hash_table. Update dependent calls and types.
24644 * tree-ssa-sccvn.c (vn_tables_s::nary):
24645 Change type to hash_table. Update dependent calls and types.
24646 (vn_tables_s::phis): Likewise.
24647 (vn_tables_s::references): Likewise.
24649 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
24650 (vn_reference_eq): Update parameter and return types.
24652 * tree-ssa-structalias.c (pointer_equiv_class_table):
24653 Change type to hash_table. Update dependent calls and types.
24654 (location_equiv_class_table): Likewise.
24656 * tree-vect-data-refs.c: Consequential changes for making
24657 peeling a hash_table.
24659 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
24660 (destroy_loop_vec_info): Dependent hash_table update.
24662 * tree-vectorizer.h (peeling_htab):
24663 Change type to hash_table. Update dependent calls and types.
24665 2013-04-23 Shiva Chen <shiva0217@gmail.com>
24667 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
24668 to check the register content is equal or not.
24669 * lra-constraints.c (match_reload): Use lra_assign_reg_val
24670 to assign register content record.
24671 * lra-eliminations.c (update_reg_eliminate): Use
24672 lra_update_reg_val_offset to update register content offset.
24673 * lra-int.h (struct lra_reg): Add offset member.
24674 (lra_reg_val_equal_p): New static inline function.
24675 (lra_update_reg_val_offset): New static inline function.
24676 (lra_assign_reg_val): New static inline function.
24677 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
24678 to assign register content record.
24679 (initialize_lra_reg_info_element): Initial offset to zero.
24681 2013-04-23 Catherine Moore <clm@codesourcery.com>
24683 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
24684 operands. Record compression.
24686 2013-04-23 Xinliang David Li <davidxl@google.com>
24688 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
24690 2013-04-23 Richard Biener <rguenther@suse.de>
24692 PR middle-end/57036
24693 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
24694 parameter, only add abnormal goto edges from the copied body
24695 if the call could perform abnormal gotos.
24696 (copy_cfg_body): Adjust.
24698 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
24700 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
24702 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
24704 * coretypes.h (gimple_stmt_iterator): Add struct to make
24707 2013-04-23 Richard Biener <rguenther@suse.de>
24709 PR tree-optimization/57026
24710 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
24711 from SSA names occuring in abnormal PHI nodes.
24713 2013-04-22 Andi Kleen <ak@linux.intel.com>
24715 * lto/lto.c (print_lto_report_1): Fix LTO report names.
24717 2013-04-22 Andi Kleen <ak@linux.intel.com>
24719 * lto/lto.c (print_lto_report_1): Declare early.
24720 (read_cgraph_and_symbols): Call print_lto_report_1 early.
24722 2013-04-22 Andi Kleen <ak@linux.intel.com>
24724 * common.opt (-flto-report-wpa): Add.
24725 * doc/invoke.texi (-flto-report-wpa): Add.
24726 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
24729 2013-04-22 Xinliang David Li <davidxl@google.com>
24731 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
24732 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
24733 * Makefile.in: New dependency
24735 David Daney <ddaney.cavm@gmail.com>
24737 * configure.ac (gcc_cv_as_micromips_support): Use the
24738 --fatal-warnings option.
24739 * configure: Regenerate.
24741 2013-04-22 Marek Polacek <polacek@redhat.com>
24744 * tsan.c (instrument_expr): Don't instrument expression
24745 in case its size is zero.
24747 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
24751 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
24753 * config/alpha/alpha.c (TARGET_LRA_P): New define.
24755 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
24757 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
24758 (gimple_stmt_iterator): New typedef.
24759 * gimple.h (gimple_stmt_iterator): Rename to...
24760 (gimple_stmt_iterator_d): ... This.
24761 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
24762 trees be valid for GIMPLE and GENERIC.
24763 (TARGET_GIMPLE_FOLD_BUILTIN): New.
24764 * gimple-fold.c (gimple_fold_call): Call target hook
24765 gimple_fold_builtin.
24766 * hooks.c (hook_bool_gsiptr_false): New.
24767 * hooks.h (hook_bool_gsiptr_false): New.
24768 * target.def (fold_stmt): New.
24769 * doc/tm.texi: Regenerate.
24771 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
24774 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
24775 a set sp if no stack realignment.
24777 2013-04-22 Nick Clifton <nickc@redhat.com>
24779 * config.gcc (tilegx-linux): Extend extra_objs rather than
24781 (tilepro-linux): Likewise.
24783 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
24785 * config/aarch64/aarch64-builtins.c
24787 (CF0, CF1, CF2, CF3, CF4, CF10): New.
24788 (VAR<1-12>): Add MAP parameter.
24789 (BUILTIN_*): Likewise.
24790 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
24791 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
24792 (aarch64_ushl_n<mode>): Likewise.
24793 (aarch64_sshr_n<mode>): Likewise.
24794 (aarch64_ushr_n<mode>): Likewise.
24795 (aarch64_<maxmin><mode>): Likewise.
24796 (aarch64_sqrt<mode>): Likewise.
24797 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
24798 (vshr<q>_n_*): Likewise.
24800 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
24802 * config/aarch64/aarch64-builtins.c
24803 (aarch64_simd_builtin_type_mode): Handle SF types.
24805 (BUILTIN_GPF): Define.
24806 (aarch64_init_simd_builtins): Handle SF types.
24807 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
24808 (frecps): Likewise.
24809 (frecpx): Likewise.
24810 * config/aarch64/aarch64-simd.md
24811 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
24812 (aarch64_frecpe<mode>): New.
24813 (aarch64_frecps<mode>): Likewise.
24814 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
24815 (v8type): Add frecp<esx>.
24816 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
24817 (aarch64_frecps<mode>): Likewise.
24818 * config/aarch64/iterators.md (FRECP): New.
24819 (frecp_suffix): Likewise.
24820 * config/aarch64/arm_neon.h
24821 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
24823 2013-04-22 Christian Bruel <christian.bruel@st.com>
24826 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
24827 (REG_CLASS_NAMES): Idem.
24828 (REG_CLASS_CONTENTS): Idem.
24829 (REGCLASS_HAS_FP_REG): Idem.
24830 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
24831 (sh_conditional_register_usage): Idem.
24833 2013-04-21 Jeff Law <law@redhat.com>
24835 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
24836 (ssa_forward_propagate_and_combine): Use it.
24838 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
24840 * lra.c: Update the flow chart diagram.
24842 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
24844 PR rtl-optimization/56847
24845 * lra-constraints.c (process_alt_operands): Discourage alternative
24846 with non-matche doffsettable memory constraint fro memory with
24849 2013-04-19 Richard Biener <rguenther@suse.de>
24851 PR tree-optimization/56982
24852 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
24854 * gimplify.c (gimplify_call_expr): Notice special calls.
24855 (gimplify_modify_expr): Likewise.
24856 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
24857 abnormal control flow receivers.
24858 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
24859 in the same way as cfun->has_nonlocal_labels.
24860 (gimple_purge_dead_abnormal_call_edges): Likewise.
24861 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
24862 receivers start a basic-block.
24864 2013-04-19 Richard Biener <rguenther@suse.de>
24866 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
24868 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
24869 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
24870 (SLP_TREE_LOAD_PERMUTATION): Add.
24871 (vect_transform_slp_perm_load): Adjust prototype.
24872 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
24873 (vect_free_slp_instance): Likewise.
24874 (vect_create_new_slp_node): Likewise.
24875 (vect_supported_slp_permutation_p): Remove.
24876 (vect_slp_rearrange_stmts): Adjust.
24877 (vect_supported_load_permutation_p): Likewise. Inline
24878 vect_supported_slp_permutation_p here.
24879 (vect_analyze_slp_instance): Compute load permutations per
24880 slp node instead of per instance.
24881 (vect_get_slp_defs): Adjust.
24882 (vect_transform_slp_perm_load): Likewise.
24883 (vect_schedule_slp_instance): Remove redundant code.
24884 (vect_schedule_slp): Remove hack for PR56270, add it ...
24885 * tree-vect-stmts.c (vectorizable_load): ... here, do not
24886 CSE loads for SLP. Adjust.
24888 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
24890 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
24891 spelling in two comments.
24893 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
24896 * config/arm/arm.c (load_multiple_sequence): Require SP
24897 as base register for loads if SP is in the register list.
24899 2013-04-19 Martin Jambor <mjambor@suse.cz>
24901 PR tree-optimization/56718
24902 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
24903 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
24904 and made public. Adjusted all callers.
24905 (ipa_intraprocedural_devirtualization): New function.
24906 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
24907 (ipa_intraprocedural_devirtualization): Likewise.
24908 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
24910 2013-04-19 Richard Biener <rguenther@suse.de>
24912 PR tree-optimization/57000
24913 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
24915 2013-04-19 Terry Guo <terry.guo@arm.com>
24917 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
24919 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
24920 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
24921 (cortex_m4_fmacs): Use new reservations.
24922 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
24924 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
24926 PR rtl-optimization/56999
24927 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
24929 (lra_coalesce): Remove split_origin_bitmap and related code.
24930 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
24931 ranges if necessary.
24933 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
24935 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
24937 (ix86_expand_call): Remove clobbered_registers array and use
24938 x86_64_ms_sysv_extra_clobbered_registers instead.
24939 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
24941 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
24943 * config/i386/i386.md (*call_rex64_ms_sysv): Use
24944 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
24945 (*call_value_rex64_ms_sysv): Ditto.
24947 2013-04-18 Cary Coutant <ccoutant@google.com>
24949 * dwarf2out.c (output_pubnames): Check die_perennial_p of
24950 parent instead of die_mark.
24952 2013-04-18 Diego Novillo <dnovillo@google.com>
24954 * gimple.c (create_gimple_tmp): New.
24955 (get_expr_type): New.
24956 (build_assign): New.
24957 (build_type_cast): New.
24958 * gimple.h (enum ssa_mode): Define.
24959 (gimple_seq_set_location): New.
24960 * asan.c (build_check_stmt): Change some gimple_build_* calls
24961 to use build_assign and build_type_cast.
24963 2013-04-18 Richard Biener <rguenther@suse.de>
24965 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
24966 handle negative step. Remove redundant checks.
24967 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
24968 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
24969 for negative step and grouped loads fail to vectorize.
24971 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
24973 * emit-rtl.c (reset_insn_used_flags): New function.
24974 (reset_all_used_flags): Use it.
24975 (verify_insn_sharing): New function.
24976 (verify_rtl_sharing): Fix verification for SEQUENCEs.
24978 2013-04-18 Jakub Jelinek <jakub@redhat.com>
24980 PR tree-optimization/56984
24981 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
24982 and (x >> M) >= N don't register any assertion if N << M is the
24985 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
24987 * lower-subreg.c (resolve_simple_move): If called self-recursive,
24988 do not delete_insn insns that have not yet been emitted, only
24989 unlink them with remove_insn.
24990 * df-scan.c (df_insn_delete): Revert r197492.
24992 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
24994 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
24995 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
24997 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
24999 * config/arm/arm.md (movsicc_insn): Convert define_insn into
25000 define_insn_and_split.
25001 (and_scc,ior_scc,negscc): Likewise.
25002 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
25004 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
25006 * config/arm/arm.c (use_return_insn): Return 0 for targets that
25007 can benefit from using a sequence of LDRD instructions in epilogue
25008 instead of a single LDM instruction.
25010 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
25013 * doc/extend.texi: Fix typo.
25015 2013-04-17 Richard Biener <rguenther@suse.de>
25017 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
25018 (vect_build_slp_tree): ... here.
25019 (vect_build_slp_tree_1): Compute which stmts of the SLP group
25020 match. Remove special-casing of mismatched complex loads.
25021 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
25022 re-try the match with swapped commutative operands.
25023 (vect_supported_load_permutation_p): Remove special-casing of
25024 mismatched complex loads.
25025 (vect_analyze_slp_instance): Adjust.
25027 2013-04-17 Richard Biener <rguenther@suse.de>
25029 PR rtl-optimization/56921
25030 * cfgloop.h (struct loop): Add simple_loop_desc member.
25031 (struct niter_desc): Mark with GTY(()).
25032 (simple_loop_desc): Do not use aux field but simple_loop_desc.
25033 * loop-iv.c (get_simple_loop_desc): Likewise.
25034 (free_simple_loop_desc): Likewise.
25037 2013-04-16 Richard Biener <rguenther@suse.de>
25039 PR rtl-optimization/56921
25040 * loop-init.c (pass_rtl_move_loop_invariants): Add
25041 TODO_do_not_ggc_collect to todo_flags_finish.
25042 (pass_rtl_unswitch): Same.
25043 (pass_rtl_unroll_and_peel_loops): Same.
25044 (pass_rtl_doloop): Same.
25046 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
25048 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
25049 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
25050 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
25052 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
25053 * tree-streamer.c (record_common_node): Adjust reference in comment.
25055 2013-04-17 Terry Guo <terry.guo@arm.com>
25057 * config/arm/cortex-m4.md: Add a new bypass.
25059 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25061 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
25063 (*subs_<optab><mode>_multp2): New pattern.
25064 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
25065 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
25067 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25069 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
25070 (*subs_mul_imm_<mode>): New pattern.
25072 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
25075 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
25076 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
25077 (vsx_movti_32bit): Change j->wa to O->wa.
25079 2013-04-16 Richard Biener <rguenther@suse.de>
25081 PR rtl-optimization/56921
25082 * loop-init.c (pass_rtl_move_loop_invariants): Add
25083 TODO_do_not_ggc_collect to todo_flags_finish.
25084 (pass_rtl_unswitch): Same.
25085 (pass_rtl_unroll_and_peel_loops): Same.
25086 (pass_rtl_doloop): Same.
25088 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
25090 * config/arm/arm.c (emit_multi_reg_push): New declaration
25091 for an existing function.
25092 (arm_emit_strd_push): New function.
25093 (arm_expand_prologue): Used here.
25094 (arm_emit_ldrd_pop): New function.
25095 (arm_expand_epilogue): Used here.
25096 (arm_get_frame_offsets): Update condition.
25097 (arm_emit_multi_reg_pop): Add a special case for load of a single
25098 register with writeback.
25100 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
25102 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
25105 2013-04-16 Richard Biener <rguenther@suse.de>
25107 PR tree-optimization/56756
25108 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
25109 (first_mem_ref_loc): New.
25110 (execute_sm): Place the load temporarily before a previous
25111 access instead of in the latch edge to ensure its SSA dependencies
25112 are defined at points dominating the load.
25114 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
25116 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
25117 correct fix by moving header and footer insn to the footer of
25118 the merged basic block. Clear BB_END of the merged-away block.
25120 PR middle-end/43631
25121 * emit-rtl.c (make_note_raw): New function.
25122 (link_insn_into_chain): New static inline function.
25123 (add_insn): Use it.
25124 (add_insn_before, add_insn_after): Factor insn chain linking code...
25125 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
25126 using link_insn_into_chain.
25127 (note_outside_basic_block_p): New helper function for emit_note_after
25128 and emit_note_before.
25129 (emit_note_after): Use nobb variant of add_insn_after if the note
25130 should not be contained in a basic block.
25131 (emit_note_before): Use nobb variant of add_insn_before if the note
25132 should not be contained in a basic block.
25133 (emit_note_copy): Use make_note_raw.
25134 (emit_note): Likewise.
25135 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
25136 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
25137 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
25138 the moved barrier the tail of the basic block it follows.
25139 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
25141 2013-04-15 Jakub Jelinek <jakub@redhat.com>
25143 PR tree-optimization/56962
25144 * gimple-ssa-strength-reduction.c (record_increment): Only set
25145 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
25146 either rhs1 or rhs2 is equal to c->base_expr.
25148 2013-04-15 Richard Biener <rguenther@suse.de>
25150 PR tree-optimization/56933
25151 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
25153 (GROUP_READ_WRITE_DEPENDENCE): Remove.
25154 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
25155 * tree-vect-data-refs.c (vect_analyze_group_access): Move
25156 dependence check ...
25157 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
25159 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
25160 GROUP_READ_WRITE_DEPENDENCE.
25162 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
25164 * emit-rtl.c (reset_all_used_flags): New function.
25165 (verify_rtl_sharing): Call reset_all_used_flags before and after
25166 performing the checks.
25168 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
25170 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
25171 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
25172 * config/arm/constraints.md (De): New constraint.
25173 * config/arm/neon.md (anddi3_neon): Delete.
25174 (neon_vand<mode>): Expand to standard anddi3 pattern.
25175 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
25176 Move earlier in the file.
25177 (neon_inv_logic_op2): Likewise.
25178 (arm_anddi_operand_neon): New predicate.
25180 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
25182 * configure.ac (gcc_cv_ld_as_needed): Set
25183 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
25184 Use -z ignore, -z record on *-*-solaris2*.
25185 (HAVE_LD_AS_NEEDED): Update comment.
25186 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
25187 * configure: Regenerate.
25188 * config.in: Regenerate.
25189 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
25190 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
25191 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
25192 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
25193 equivalents. Fix markup.
25194 * doc/tm.texi: Regenerate.
25196 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
25198 * config/i386/i386.opt: New option mstack-protector-guard=.
25199 * config/i386/i386-opts.h: Add enum stack_protector_guard.
25200 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
25201 TARGET_SSP_TLS_GUARD.
25202 * config/i386/i386.c (ix86_option_override_internal): Set
25203 ix86_stack_protector_guard.
25204 * config/i386/i386.md (stack_protect_set): Enable for
25205 TARGET_SSP_TLS_GUARD only.
25206 (stack_protect_set_<mode>): Ditto.
25207 (stack_protect_test): Ditto.
25208 (stack_protect_test_<mode>): Ditto.
25209 * doc/invoke.texi (i386 Option): Document.
25211 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
25214 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
25215 (S_MODES): Set H_MODE bit.
25216 (SF_MODES): Set only S_MODE and SF_MODE bits.
25217 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
25218 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
25219 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
25220 <MODE_FLOAT>: Likewise.
25222 2013-04-15 Joey Ye <joey.ye@arm.com>
25224 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
25226 2013-04-15 Joey Ye <joey.ye@arm.com>
25228 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
25230 (thumb_far_jump_used_p): Count instruction size and set
25233 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
25235 * reorg.c (fill_simple_delay_slots): Reindent block of code.
25236 * resource.c (mark_target_live_regs): Reformat conditional block.
25238 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
25240 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
25241 notes, they are emitted only just before final.
25242 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
25244 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
25246 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
25247 * cfgrtl.c (delete_insn): Call it here instead.
25248 * lra-spills.c (lra_final_code_change): Use delete_insn.
25249 * haifa-sched.c (sched_remove_insn): Likewise.
25250 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
25251 returning to the nop pool.
25252 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
25253 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
25255 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
25257 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
25258 * doc/tm.texi: Regenerated.
25260 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
25262 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
25265 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
25267 * df-core.c (df_find_def): Compare register numbers.
25268 (df_find_use): Likewise.
25270 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
25273 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
25274 lra_in_progress for return.
25276 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
25278 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
25279 define_insn into define_insn_and_split and emit movsicc patterns.
25281 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
25283 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
25285 2013-04-12 Richard Biener <rguenther@suse.de>
25287 * tree-pass.h (TODO_do_not_ggc_collect): New.
25288 * passes.c (execute_one_ipa_transform_pass): Honor
25289 TODO_do_not_ggc_collect.
25290 (execute_one_pass): Likewise.
25293 2013-04-10 Richard Biener <rguenther@suse.de>
25295 * passes.c (init_optimization_passes): Remove reload pass.
25296 * ira.c (do_reload): Merge into ...
25298 (rest_of_handle_reload): Remove.
25299 (pass_reload): Likewise.
25300 * config/i386/i386.c (ix86_option_override): Refer to ira instead
25301 of reload for vzeroupper pass placement.
25303 2013-04-12 Jakub Jelinek <jakub@redhat.com>
25305 PR tree-optimization/56918
25306 PR tree-optimization/56920
25307 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
25308 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
25309 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
25310 use wide_mul_with_sign method.
25312 2013-04-12 Richard Biener <rguenther@suse.de>
25314 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
25315 not be considered a gimple constant.
25317 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
25319 * fold-const.c (const_binop): Handle vector shifts by a scalar.
25320 (fold_binary_loc): Call const_binop also for mixed vector-scalar
25323 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
25324 Jakub Jelinek <jakub@redhat.com>
25326 * opts.c: Include diagnostic-color.h.
25327 (common_handle_option): Handle OPT_fdiagnostics_color_.
25328 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
25329 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
25330 (diagnostic-color.o): New.
25331 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
25332 (diagnostic_color_rule): New enum.
25333 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
25334 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
25335 the location string.
25336 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
25337 either NULL, or color kind.
25338 * diagnostic-color.c: New file.
25339 * diagnostic-color.h: New file.
25340 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
25342 * doc/invoke.texi (-fdiagnostics-color): Document.
25343 * pretty-print.h (pp_show_color): Define.
25344 (struct pretty_print_info): Add show_color field.
25345 * diagnostic.c: Include diagnostic-color.h.
25346 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
25347 macros. Colorize error:, warning: etc. strings and also the location
25349 (diagnostic_show_locus): Colorize the caret line.
25350 * pretty-print.c: Include diagnostic-color.h.
25351 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
25352 inside of %< %> quotes or quoted through q format modifier.
25354 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
25356 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
25358 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25360 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
25361 code in CC_NZ mode.
25362 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
25365 2013-04-11 Marek Polacek <polacek@redhat.com>
25367 PR tree-optimization/48184
25368 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
25370 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
25372 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
25373 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
25374 (skip_simple_arithmetic): Tidy up.
25375 * tree.h (skip_simple_constant_arithmetic): Declare.
25377 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25379 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
25381 2013-04-11 Richard Biener <rguenther@suse.de>
25383 * tree-vect-loop.c (get_initial_def_for_induction): Properly
25384 generate vector constants.
25386 2013-04-11 Richard Biener <rguenther@suse.de>
25388 PR tree-optimization/56878
25389 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
25390 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
25392 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
25393 Prefer to align the DR with the most invariant base address.
25395 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
25397 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
25400 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
25402 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
25403 floating-point vector comparisons against 0.
25405 2013-04-11 Jakub Jelinek <jakub@redhat.com>
25407 PR tree-optimization/56899
25408 * fold-const.c (extract_muldiv_1): Apply distributive law
25409 only if TYPE_OVERFLOW_WRAPS (ctype).
25411 2013-04-11 Bin Cheng <bin.cheng@arm.com>
25414 * ira-costs.c (scan_one_insn): Check whether the source rtx of
25415 loading has side effect.
25417 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
25419 * config/sparc/sparc.c: Include tree-pass.h.
25420 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
25421 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
25422 head of file. Change return type. Split off gate function.
25423 (sparc_gate_work_around_errata): New function.
25424 (pass_work_around_errata): New pass definition.
25425 (insert_pass_work_around_errata) New pass insert definition to
25426 insert pass_work_around_errata just after delayed-branch scheduling.
25427 (sparc_option_override): Insert the pass.
25428 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
25430 2013-04-10 David S. Miller <davem@davemloft.net>
25432 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
25433 or -mcpu=hypersparc.
25435 * target.def (cstore_mode): New hook.
25436 * target.h: Include insn-codes.h
25437 * targhooks.c: Likewise.
25438 (default_cstore_mode): New function.
25439 * targhooks.h: Declare it.
25440 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
25441 * doc/tm.texi: Rebuild.
25442 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
25443 target hook, rather than inspecting the insn_data.
25444 * config/sparc/sparc.c (sparc_cstore_mode): New function.
25445 (TARGET_CSTORE_MODE): Redefine.
25446 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
25448 * config/sparc/predicates.md (cstore_result_operand): New special
25450 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
25451 Use it for operand 0.
25452 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
25453 (*snesi_special): Likewise.
25454 (*snesi_zero): Likewise.
25455 (*seqsi_zero): Likewise.
25456 (*sltu_insn): Likewise.
25457 (*sgeu_insn): Likewise.
25458 (*seqdi_special): Make operand 0 and comparison operation be of
25460 (*snedi_special): Likewise.
25461 (*snedi_special_vis3): Likewise.
25462 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
25463 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
25464 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
25465 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
25466 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
25467 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
25468 (*sltu_extend_sp64): Likewise.
25469 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
25470 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
25471 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
25472 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
25473 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
25475 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
25477 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
25478 (aarch64_start_file): Use the new function.
25480 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
25481 Jason Merrill <jason@redhat.com>
25483 * common.opt: Add -gdwarf.
25484 * opts.c (common_handle_option): Handle it.
25485 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
25487 2013-04-10 Richard Biener <rguenther@suse.de>
25489 * passes.c (execute_todo): Do not call ggc_collect conditional here.
25490 (execute_one_ipa_transform_pass): But unconditionally here.
25491 (execute_one_pass): And here.
25492 (init_optimization_passes): Remove reload pass.
25493 * tree-pass.h (TODO_ggc_collect): Remove.
25494 (pass_reload): Likewise.
25495 * ira.c (do_reload): Merge into ...
25497 (rest_of_handle_reload): Remove.
25498 (pass_reload): Likewise.
25499 * config/i386/i386.c (ix86_option_override): Refer to ira instead
25500 of reload for vzeroupper pass placement.
25501 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
25502 and todo_flags_finish of all passes.
25504 2013-04-10 Richard Biener <rguenther@suse.de>
25506 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
25507 first_const_oprnd field, rename first_def_type to first_op_type.
25508 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
25509 (vect_get_and_check_slp_defs): Always use the type of the
25510 operand. Allow mixed vect_external_def, vect_constant_def types.
25511 (vect_get_constant_vectors): Handle mixed vect_external_def,
25512 vect_constant_def types.
25514 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
25516 PR tree-optimization/55524
25517 * tree-ssa-math-opts.c
25518 (convert_mult_to_fma): Don't use an fms construct
25519 when we don't have an fms operation, but fnma, and it looks
25520 likely that we'll be able to use the latter.
25522 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
25524 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
25526 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
25527 inline fail caused by overwritable functions.
25529 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
25531 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
25532 unnecessary bits in the constant power of two case.
25534 2013-04-10 Richard Biener <rguenther@suse.de>
25536 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
25537 broken code swapping operands.
25538 (vect_build_slp_tree): Do not compute load permutations here.
25539 (vect_analyze_slp_instance): Compute load permutations here,
25540 after building the SLP tree.
25542 2013-04-09 Christian Bruel <christian.bruel@st.com>
25544 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
25545 of next/prev_real_insn.
25547 2013-04-09 Jan Hubicka <jh@suse.cz>
25549 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
25550 Drop aliased parameter.
25551 (function_and_variable_visibility): Do not handle alias pairs.
25552 * cgraph.c (varpool_externally_visible_p): Update prototype.
25553 * varpool.c (varpool_add_new_variable): Update.
25555 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
25557 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
25559 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
25561 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
25563 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
25565 2013-04-09 Marek Polacek <polacek@redhat.com>
25567 PR tree-optimization/48762
25568 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
25570 2013-04-09 Richard Biener <rguenther@suse.de>
25572 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
25574 (vect_build_slp_tree): Likewise.
25575 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
25576 calculating the cost of a SLP instance.
25577 (vect_analyze_slp_instance): Use it from here, after building
25580 2013-04-09 Jakub Jelinek <jakub@redhat.com>
25582 PR middle-end/56883
25583 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
25584 expand_omp_for_static_chunk): Use simple_p = true in
25585 force_gimple_operand_gsi calls when assigning to addressable decls.
25587 2013-04-09 Jeff Law <law@redhat.com>
25589 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
25590 when the boolean was created by converting a wider object which
25591 had a boolean range.
25593 2013-04-09 Richard Biener <rguenther@suse.de>
25595 * tree-vectorizer.h (slp_void_p): Remove.
25596 (slp_tree): Typedef before _slp_tree declaration.
25597 (struct _slp_tree): Use a vector of slp_tree as children.
25598 (vect_get_place_in_interleaving_chain): Remove.
25599 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
25601 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
25603 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
25604 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
25605 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
25606 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
25607 Use slp_node instead of slp_void_p and adjust.
25609 2013-04-09 Richard Biener <rguenther@suse.de>
25611 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
25612 work that is not necessary.
25614 2013-04-09 Jakub Jelinek <jakub@redhat.com>
25616 PR tree-optimization/56854
25617 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
25618 forward into clobber stmts if it would change MEM_REF lhs into
25621 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
25623 * tree.c (type_hash_lookup, type_hash_add): Make static.
25624 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
25626 2013-04-09 Richard Biener <rguenther@suse.de>
25628 * tree.h (unsave_expr_now): Remove.
25629 * tree-inline.c (mark_local_for_remap_r): Remove.
25630 (unsave_expr_1): Likewise.
25631 (unsave_r): Likewise.
25632 (unsave_expr_now): Likewise.
25633 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
25634 (propagate_tree_value): Likewise.
25636 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
25638 * doc/rtl.texi (sequence): Rewrite documentation to match the
25639 current use of SEQUENCE rtl objects.
25640 * rtl.def (SEQUENCE): Likewise.
25642 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
25643 Update documentation.
25644 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
25645 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
25647 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
25649 2013-04-08 Teresa Johnson <tejohnson@google.com>
25651 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
25652 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
25654 (estimate_edge_size_and_time): Add comment to suggest using rounding
25656 (estimate_node_size_and_time): Ditto.
25657 (remap_edge_change_prob): Use helper rounding divide methods.
25658 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
25659 (gimple_mod_pow2_value_transform): Ditto.
25660 (gimple_mod_subtract_transform): Ditto.
25661 (gimple_ic_transform): Ditto.
25662 (gimple_stringops_transform): Ditto.
25663 * stmt.c (conditional_probability): Ditto.
25664 (emit_case_dispatch_table): Ditto.
25665 * lto-cgraph.c (merge_profile_summaries): Ditto.
25666 * tree-optimize.c (execute_fixup_cfg): Ditto.
25667 * cfgcleanup.c (try_forward_edges): Ditto.
25668 * cfgloopmanip.c (scale_loop_profile): Ditto.
25670 (duplicate_loop_to_header_edge): Ditto.
25671 (lv_adjust_loop_entry_edge): Ditto.
25672 * tree-vect-loop.c (vect_transform_loop): Ditto.
25673 * profile.c (compute_branch_probabilities): Ditto.
25674 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
25675 * lto-streamer-in.c (input_cfg): Ditto.
25676 * gimple-streamer-in.c (input_bb): Ditto.
25677 * ipa-cp.c (update_profiling_info): Ditto.
25678 (update_specialized_profile): Ditto.
25679 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
25680 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
25682 * sched-rgn.c (compute_dom_prob_ps): Ditto.
25683 (compute_trg_info): Ditto.
25684 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
25685 (purge_dead_edges): Ditto.
25686 * loop-unswitch.c (unswitch_loop): Ditto.
25687 * cgraphclones.c (cgraph_clone_edge): Ditto.
25688 (cgraph_clone_node): Ditto.
25689 * tree-inline.c (copy_bb): Ditto.
25690 (copy_edges_for_bb): Ditto.
25691 (initialize_cfun): Ditto.
25692 (copy_cfg_body): Ditto.
25693 (expand_call_inline): Ditto.
25695 2013-04-08 Kai Tietz <ktietz@redhat.com>
25697 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
25698 TARGET_CYGWIN64 by TARGET_64BIT.
25700 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
25702 * config/epiphany/epiphany.md (GPR_1): New constant.
25703 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
25704 * config/epiphany/epiphany.c (gen_compare_reg):
25705 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
25706 is already in place.
25707 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
25708 Don't require being called during rtl expansion; If y operlaps r0,
25710 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
25711 (epiphany_expand_epilogue): Likewise.
25713 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
25714 Don't use CC_FPmode for ORDERED / UNORDERED.
25715 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
25717 * config/epiphany/constraints.md (CnL): New constraint.
25718 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
25719 * config/epiphany/predicates.md (add_operand): Allow 1024.
25721 * config/epiphany/epiphany.md (logical_op): New code iterator.
25722 (op_mnc): New code attribute.
25723 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
25724 (mov_f+1, mov_f+2): New peephole2 patterns.
25726 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
25727 (cstoresi4): Also allow re-use of zero result when doing a NE
25728 comparison to a non-zero operand.
25729 Use (clobber (scratch)) for first insn if the gpr output is not needed.
25731 * config/epiphany/epiphany.md (<insn_opname>v2si3):
25732 Use gen_addsi3_i / gen_subsi3_i.
25734 2013-04-08 Jakub Jelinek <jakub@redhat.com>
25738 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
25739 contain anything but clobbers, at most one __builtin_stack_restore,
25740 optionally debug stmts and final resx, and if it has at least one
25741 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
25742 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
25743 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
25744 which isn't defaut definition, remove them.
25745 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
25746 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
25747 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
25748 with MEM_REF LHS with SSA_NAME address.
25750 2013-04-08 Jeff Law <law@redhat.com>
25752 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
25754 2013-04-08 Richard Biener <rguenther@suse.de>
25756 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
25758 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
25759 determined vector type.
25760 (vect_analyze_data_refs): Likewise.
25761 (vect_get_new_vect_var): Adjust.
25762 (vect_create_destination_var): Preserve SSA name versions.
25763 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
25764 not dump anything here.
25766 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
25768 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
25769 Add member lr_slot_known.
25770 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
25772 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
25773 Remove code that sets lr_slot_offset according to what a previous
25774 version of epiphany_emit_save_restore used to do.
25775 (epiphany_emit_save_restore): When doing an lr save or restore,
25776 set/verify lr_slot_known and lr_slot_offset.
25778 2013-04-08 Xinyu Qi <xyqi@marvell.com>
25781 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
25784 2013-04-08 Richard Biener <rguenther@suse.de>
25786 * alias.c (find_base_term): Fix thinko in previous change.
25788 2013-04-08 Jakub Jelinek <jakub@redhat.com>
25790 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
25791 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
25792 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
25793 if possible to compute val.
25794 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
25795 For QImode integers don't require anything about precision. Use
25796 const_with_all_bytes_same to find out if the constant doesn't have
25797 repeated bytes in it.
25799 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
25801 * config/s390/s390.c (s390_expand_insv): Only accept insertions
25804 2013-04-08 Marek Polacek <polacek@redhat.com>
25806 PR rtl-optimization/48182
25807 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
25810 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25813 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
25814 nuses, make sure we have a label.
25816 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
25819 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
25820 (rs6000_emit_swdiv_low_precision): Remove.
25821 (rs6000_emit_swdiv): Rewrite to handle between one and four
25822 iterations of Newton-Raphson generally; modify required number of
25823 iterations for some cases.
25824 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
25826 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
25828 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
25829 set-but-unused variable.
25831 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
25832 basic blocks of released function bodies garbage-collectable.
25834 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
25835 (struct rtl_opt_pass): Add TODO_df_finish.
25837 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
25839 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
25841 * config/arm/constraints.md (q): New constraint.
25842 * config/arm/ldrdstrd.md: New file.
25843 * config/arm/arm.md (ldrdstrd.md) New include.
25844 (arm_movdi): Use "q" instead of "r" constraint
25845 for double-word memory access.
25846 (movdf_soft_insn): Likewise.
25847 * config/arm/vfp.md (movdi_vfp): Likewise.
25848 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
25849 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
25850 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
25851 (mem_ok_for_ldrd_strd): Likewise.
25852 (output_move_double): Update assertion.
25854 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
25856 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
25858 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
25860 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
25861 define_insn_and_split.
25862 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
25864 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
25866 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
25867 define_insn_and_split.
25868 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
25869 (shiftsi3_compare): New pattern.
25870 (rrx): New pattern.
25871 * config/arm/unspecs.md (UNSPEC_RRX): New.
25873 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
25875 * config/arm/arm.md (negdi_extendsidi): New pattern.
25876 (negdi_zero_extendsidi): Likewise.
25878 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
25880 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
25881 define_insn_and_split.
25882 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
25883 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
25885 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
25887 * config/arm/arm.md (arm_subdi3): Convert define_insn into
25888 define_insn_and_split.
25889 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
25890 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
25892 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
25894 * config/arm/arm.md (subsi3_carryin): New pattern.
25895 (subsi3_carryin_const): Likewise.
25896 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
25897 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
25899 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
25901 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
25903 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
25905 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
25906 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
25908 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
25910 * config/arm/arm.c (arm_expand_builtin): Change fcode
25911 type to unsigned int.
25913 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
25915 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
25917 2013-04-04 Ian Lance Taylor <iant@google.com>
25919 * doc/standards.texi (Standards): The Go frontend supports the Go 1
25922 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
25924 PR middle-end/56729
25925 * df-scan.c (df_insn_delete): Disable failing assert.
25927 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
25929 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
25930 New function prototype.
25931 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
25932 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
25933 (arm_builtin_vectorized_function): New function.
25935 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
25937 * config/arm/arm_neon_builtins.def: New file.
25938 * config/arm/arm.c (neon_builtin_data): Move contents to
25939 arm_neon_builtins.def.
25940 (enum arm_builtins): Include neon builtin definitions.
25941 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
25942 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
25944 2013-04-04 Marek Polacek <polacek@redhat.com>
25946 PR tree-optimization/48186
25947 * predict.c (maybe_hot_frequency_p): Return false if
25948 HOT_BB_FREQUENCY_FRACTION is 0.
25949 (cgraph_maybe_hot_edge_p): Likewise.
25951 2013-04-04 Richard Biener <rguenther@suse.de>
25953 PR tree-optimization/56826
25954 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
25957 2013-04-04 Richard Biener <rguenther@suse.de>
25959 PR tree-optimization/56213
25960 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
25961 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
25963 2013-04-04 Richard Biener <rguenther@suse.de>
25965 PR tree-optimization/56837
25966 * tree-loop-distribution.c (classify_partition): For non-zero
25967 values require that the value has the same precision as its
25968 mode to be useful as memset value.
25970 2013-04-03 Nick Clifton <nickc@redhat.com>
25972 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
25973 (fmssf4): Use fmsf.s on E3V5 architectures.
25974 (fnmasf4): Use fnmaf.s on E3V5 architectures.
25975 (fnmssf4): Use fnmsf.s on E3V5 architectures.
25977 2013-04-03 Jeff Law <law@redhat.com>
25979 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
25980 (lra-eliminations.o): Likewise.
25982 2013-04-03 Teresa Johnson <tejohnson@google.com>
25984 * gcov-io.c (compute_working_sets): Moved most of body of old
25985 compute_working_sets here from profile.c.
25986 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
25987 (gcov_working_set_t): Moved typedef here from basic-block.h
25988 (compute_working_set): Declare.
25989 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
25990 (get_working_sets): Renamed from compute_working_set,
25991 replace most of body with call to new compute_working_sets.
25992 (get_exec_counts): Replace call to compute_working_sets
25993 to get_working_sets.
25994 * profile.h (get_working_sets): Renamed from compute_working_set.
25995 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
25996 to get_working_sets.
25997 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
25998 * gcov-dump.c (dump_working_sets): New function.
26000 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
26002 * hwint.c (sext_hwi, zext_hwi): New functions.
26003 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
26004 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
26005 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
26006 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
26007 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
26008 (sext_hwi, zext_hwi): New functions.
26010 2013-04-03 Jeff Law <law@redhat.com>
26012 PR tree-optimization/56799
26013 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
26014 back test for widening conversion erroneously dropped in prior change.
26016 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
26019 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
26020 instead of next_real_insn.
26022 2013-04-03 Marek Polacek <polacek@redhat.com>
26025 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
26027 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
26030 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
26032 (thumb1_output_casesi): Likewise.
26033 (thumb2_output_casesi): Likewise.
26035 2013-04-03 Richard Biener <rguenther@suse.de>
26037 PR tree-optimization/56817
26038 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
26040 (tree_unroll_loops_completely_1): ... new function to manually
26041 walk the loop tree, properly defering outer loops of unrolled
26042 loops to later iterations.
26044 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
26046 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
26047 (vectorizable_load): Likewise.
26048 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
26049 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
26051 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
26053 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
26056 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
26058 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
26060 2013-04-03 Bin Cheng <bin.cheng@arm.com>
26062 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
26064 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
26066 PR tree-optimization/56790
26067 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
26070 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
26072 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
26074 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
26075 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
26078 2013-04-03 Jakub Jelinek <jakub@redhat.com>
26081 * tree.h (force_folding_builtin_constant_p): New decl.
26082 * builtins.c (force_folding_builtin_constant_p): New variable.
26083 (fold_builtin_constant_p): Fold immediately also if
26084 force_folding_builtin_constant_p.
26086 2013-04-03 Richard Biener <rguenther@suse.de>
26088 PR tree-optimization/56812
26089 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
26090 DRs of the same interleaving chain are independent.
26092 2013-04-02 Jason Merrill <jason@redhat.com>
26094 * gdbinit.in (pbb): Use debug fn.
26096 2013-04-02 Lawrence Crowl <crowl@google.com>
26098 * sese.h (struct ivtype_map_elt_s): Remove unused.
26099 (extern debug_ivtype_map): Remove unused.
26100 (extern eq_ivtype_map_elts): Remove unused.
26101 * sese.c (debug_ivtype_map): Removed unused.
26102 (debug_ivtype_map_1): Removed unused.
26103 (debug_ivtype_elt): Remove unused.
26104 (eq_ivtype_map_elts): Remove unused.
26106 2013-04-02 Kai Tietz <ktietz@redhat.com>
26109 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
26110 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
26111 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
26113 (legitimize_pe_coff_symbol): Likewise.
26114 (is_imported_p): New helper-function.
26115 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
26116 for Windows x64 targets.
26117 (ix86_expand_prologue): Optimize for pe-coff targets.
26118 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
26119 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
26120 medium/large code-model.
26121 (legitimize_pic_address): Likewise.
26122 (legitimize_tls_address): Likewise.
26123 (ix86_expand_call): Likewise.
26124 (x86_output_mi_thunk): Likewise.
26125 (get_dllimport_decl): Add new beimport argument.
26126 (construct_plt_address): Don't assert for x64 pe-coff targets.
26127 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
26129 (SYMBOL_FLAG_STUBVAR): New macro.
26130 (SYMBOL_REF_STUBVAR_P): Likewise.
26131 * config/i386/winnt.c (stub_list): New structure.
26132 (stub_head): New local variable.
26133 (i386_pe_record_stub): New function.
26134 (i386_pe_file_end): Emit refptr-stubs.
26136 2013-04-02 Jakub Jelinek <jakub@redhat.com>
26138 PR rtl-optimization/56745
26139 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
26140 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
26143 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
26144 and both of them are MEM_REFs, just compare first argument for
26145 equality and attempt to deal even with differing offsets.
26148 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
26149 of gimple_clobber_p to be MEM_REF.
26150 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
26151 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
26152 after gimplification.
26153 * asan.c (get_mem_ref_of_assignment): Don't instrument
26154 gimple_clobber_p stmts.
26155 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
26156 gimple_clobber_p stmt if they have MEM_REF lhs and
26157 are dead because of another gimple_clobber_p stmt.
26158 * tree-ssa-live.c (clear_unused_block_pointer): Treat
26159 gimple_clobber_p stmts like debug stmts.
26160 (remove_unused_locals): Remove clobbers with MEM_REF lhs
26161 that refer to unused VAR_DECLs or uninitialized values.
26162 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
26163 gimple_clobber_p stmts if they refer to removed parameters.
26164 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
26167 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
26169 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
26170 using SWI48 mode attribute.
26172 2013-04-02 Wei Mi <wmi@google.com>
26174 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
26175 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
26176 *<rotate_insn><mode>3_mask in i386.md.
26178 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
26180 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
26182 2013-04-02 Richard Biener <rguenther@suse.de>
26184 PR tree-optimization/56778
26185 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
26186 Runtime alias tests are not supported for gather loads.
26187 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
26188 stmts referenced from SSA operands before updating SSA form.
26190 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
26191 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
26193 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
26194 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
26195 * config/arm/cortex-a53.md: New file.
26196 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
26197 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
26198 * config/arm/arm.c (arm_issue_rate): Likewise.
26199 * config/arm/arm-tune.md: Regenerate
26200 * config/arm/arm-tables.opt: Regenerate.
26201 * config/arm/arm-cores.def: Add cortex-a53.
26203 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
26205 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
26208 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
26210 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
26211 scalar load/store operations using B/H registers.
26212 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
26214 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
26216 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
26218 * config/aarch64/aarch64.c
26219 (aarch64_simd_scalar_immediate_valid_for_move): New.
26220 * config/aarch64/aarch64-protos.h
26221 (aarch64_simd_scalar_immediate_valid_for_move): New.
26222 * config/aarch64/constraints.md (Dh, Dq): New.
26223 * config/aarch64/iterators.md (hq): New.
26225 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
26227 * reorg.c (get_branch_condition): Deal with conditional returns.
26228 (fill_simple_delay_slots): Remove dead code dealing with jumps.
26230 2013-04-01 Wei Mi <wmi@google.com>
26232 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
26233 Truncate operand 2 using %b asm operand modifier.
26234 (*<shift_insn><mode>3_mask): Ditto.
26235 (*<rotate_insn><mode>3_mask): Ditto.
26237 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
26239 PR middle-end/56798
26240 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
26242 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
26244 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
26246 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
26248 2013-03-30 Lawrence Crowl <crowl@google.com>
26250 * dse.c (clear_alias_sets): Remove never set.
26251 (disqualified_clear_alias_sets): Remove never set.
26252 (clear_alias_mode_pool): Remove never set.
26253 (dse_step0): Remove condition that is never true.
26254 (canon_address): Remove condition that is never true.
26255 (dse_step7): Remove condition that is never true.
26256 (rest_of_handle_dse): Remove condition that is never true.
26257 (rest_of_handle_dse::did_global): Remove never read from above.
26258 (dse_step2_spill): Remove never called from above.
26259 (dse_step5_spill): Remove never called from above.
26261 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
26263 * doc/md.texi (Standard Names) <casesi>: Update documentation for
26264 JUMP_TABLE_DATA changes.
26265 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
26266 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
26267 (Insns) <jump_table_data>: New entry.
26268 * doc/tm.texi: Regenerate.
26270 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
26272 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
26273 for table jump at the end of a basic block using tablejump_p.
26274 * targhooks.c (default_invalid_within_doloop): Likewise.
26275 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
26276 target hook implementation that is identical to the default hook.
26277 (rs6000_invalid_within_doloop): Remove.
26279 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
26280 unused variable from tablejump_p call.
26282 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
26283 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
26284 (INSN_DELETED_P): Likewise.
26285 (emit_jump_table_data): New prototype.
26286 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
26287 after 4th as unused.
26288 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
26289 * sched-vis.c (print_insn): Likewise.
26290 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
26291 insn for compatibility with back ends that use next_active_insn to
26292 identify jump table data.
26293 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
26294 (remove_insn): Likewise.
26295 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
26297 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
26298 (emit_jump_table_data): New function.
26300 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
26301 basic block, a JUMP_TABLE_DATA never is.
26302 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
26303 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
26304 off from code handling real insns.
26305 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
26306 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
26307 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
26308 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
26309 is not a NONDEBUG_INSN_P.
26310 * ira-costs.c (scan_one_insn): Likewise.
26311 * jump.c (mark_all_labels): Likewise.
26312 (mark_jump_label_1): Likewise.
26313 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
26314 * lra.c (get_insn_freq): Expect all insns reaching here to be in
26316 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
26317 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
26318 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
26319 JUMP_TABLE_DATA_P insns.
26320 (calculate_elim_costs_all_insns): Likewise.
26321 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
26322 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
26323 (delete_output_reload): Code style fixups.
26324 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
26325 insn flags on this non-insn.
26326 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
26327 as scheduling barriers, for pre-change compatibility.
26328 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
26329 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
26331 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
26332 redundant JUMP_TABLE_DATA_P test.
26333 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
26334 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
26335 (frv_for_each_packet): Likewise.
26336 * config/i386/i386.c (min_insn_size): Likewise.
26337 (ix86_avoid_jump_mispredicts): Likewise.
26338 * config/m32r/m32r.c (m32r_is_insn): Likewise.
26339 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
26340 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
26341 (mips16_insn_length): Robustify.
26342 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
26343 (mips16_split_long_branches): Likewise.
26344 * config/pa/pa.c (pa_combine_instructions): Likewise.
26345 * config/rs6000/rs6000.c (get_next_active_insn): Treat
26346 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
26347 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
26348 as contributing to pool range lengths.
26349 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
26350 Remove redundant JUMP_TABLE_DATA_P test.
26351 (sh_loop_align): Likewise.
26352 (split_branches): Likewise.
26353 (sh_insn_length_adjustment): Likewise.
26354 * config/spu/spu.c (get_branch_target): Likewise.
26356 2013-03-29 Jan Hubicka <jh@suse.cz>
26358 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
26359 gcov streaming; stream hot bb threshold to ltrans.
26360 * predict.c (get_hot_bb_threshold): Break out from ....
26361 (maybe_hot_count_p): ... here.
26362 (set_hot_bb_threshold): New function.
26363 * lto-section-in.c (lto_section_name): Add profile.
26364 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
26365 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
26366 and data-streamer.h
26367 (histogram_entry): New structure.
26368 (histogram, histogram_pool): New global vars.
26369 (histogram_hash): New structure.
26370 (histogram_hash::hash): New method.
26371 (histogram_hash::equal): Likewise.
26372 (account_time_size): New function.
26373 (cmp_counts): New function.
26374 (dump_histogram): New function.
26375 (ipa_profile_generate_summary): New function.
26376 (ipa_profile_write_summary): New function.
26377 (ipa_profile_read_summary): New function.
26378 (ipa_profile): Decide on threshold.
26379 (pass_ipa_profile): Add ipa_profile_write_summary and
26380 ipa_profile_read_summary.
26381 * Makefile.in (ipa.o): Update dependencies.
26382 * lto-streamer.h (LTO_section_ipa_profile): New section.
26384 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
26386 * tree.h (VAR_P): New.
26388 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
26391 * doc/invoke.texi ([-fwhole-program]): Fix typo.
26393 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
26395 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
26396 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
26397 (control_flow_insn_p): Likewise.
26398 * cfgrtl.c (duplicate_insn_chain): Likewise.
26399 * final.c (get_attr_length_1): Likewise.
26400 (shorten_branches): Likewise.
26401 (final_scan_insn): Likewise.
26402 * function.c (instantiate_virtual_regs): Likewise.
26403 * gcse.c (insert_insn_end_basic_block): Likewise.
26404 * ira-costs.c (scan_one_insn): Likewise.
26405 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
26406 * lra.c (check_rtl): Likewise.
26407 * reload1.c (elimination_costs_in_insn): Likewise.
26408 * reorg.c (follow_jumps): Likewise.
26410 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
26411 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
26412 (thumb_far_jump_used_p): Likewise.
26413 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
26414 (workaround_speculation): Likewise.
26415 (add_sched_insns_for_speculation): Likewise.
26416 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
26417 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
26418 (frv_for_each_packet): Likewise.
26419 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
26420 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
26421 (final_emit_insn_group_barriers): Likewise.
26422 * config/m32r/m32r.c (m32r_is_insn): Likewise.
26423 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
26424 (mips16_insn_length): Likewise.
26425 * config/pa/pa.c (pa_reorg): Likewise.
26426 (pa_combine_instructions): Likewise.
26427 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
26428 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
26429 (sh_reorg): Likewise.
26430 (split_branches): Likewise.
26431 * config/spu/spu.c (get_branch_target): Likewise.
26433 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
26436 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
26438 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
26439 Fix declaration name.
26441 2013-03-28 Lawrence Crowl <crowl@google.com>
26443 * graphds.h (struct graph.indicies): Remove unused.
26444 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
26445 (SCOP_ORIGINAL_PDDRS): Remove unused.
26446 * sese.h (extern insert_loop_close_phis): Removed unused.
26447 (extern insert_guard_phis): Removed unused.
26448 (extern ivtype_map_elt_info): Removed unused.
26449 (new_ivtype_map_elt): Removed unused.
26450 * sese.c (ivtype_map_elt_info): Removed unused.
26452 2013-03-28 Lawrence Crowl <crowl@google.com>
26454 * Makefile.in: Add several missing include dependences.
26456 (test-dump.o): New. This object is not added to any executable,
26457 but is present for ad-hoc testing.
26459 (debug (const bitmap_head_def &)): New.
26460 (debug (const bitmap_head_def *)): New.
26462 (extern debug (const bitmap_head_def &)): New.
26463 (extern debug (const bitmap_head_def *)): New.
26465 (debug (edge_def &)): New.
26466 (debug (edge_def *)): New.
26468 (debug (basic_block_def &)): New.
26469 (debug (basic_block_def *)): New.
26471 (dump_node (const_tree, int, FILE *)): Correct source file.
26473 (debug (die_struct &)): New.
26474 (debug (die_struct *)): New.
26476 (extern debug (die_struct &)): New.
26477 (extern debug (die_struct *)): New.
26478 * gimple-pretty-print.c
26479 (debug (gimple_statement_d &)): New.
26480 (debug (gimple_statement_d *)): New.
26481 * gimple-pretty-print.h
26482 (extern debug (gimple_statement_d &)): New.
26483 (extern debug (gimple_statement_d *)): New.
26485 (debug (ira_allocno_copy &)): New.
26486 (debug (ira_allocno_copy *)): New.
26487 (debug (ira_allocno &)): New.
26488 (debug (ira_allocno *)): New.
26490 (extern debug (ira_allocno_copy &)): New.
26491 (extern debug (ira_allocno_copy *)): New.
26492 (extern debug (ira_allocno &)): New.
26493 (extern debug (ira_allocno *)): New.
26495 (debug (live_range &)): New.
26496 (debug (live_range *)): New.
26498 (debug (lra_live_range &)): New.
26499 (debug (lra_live_range *)): New.
26501 (debug (lra_live_range &)): New.
26502 (debug (lra_live_range *)): New.
26504 (debug (omega_pb_d &)): New.
26505 (debug (omega_pb_d *)): New.
26507 (extern debug (omega_pb_d &)): New.
26508 (extern debug (omega_pb_d *)): New.
26510 (debug (const rtx_def &)): New.
26511 (debug (const rtx_def *)): New.
26513 (debug_tree (tree): Move within file.
26514 (debug_raw (const tree_node &)): New.
26515 (debug_raw (const tree_node *)): New.
26516 (dump_tree_via_hooks (const tree_node *, int)): New.
26517 (debug (const tree_node &)): New.
26518 (debug (const tree_node *)): New.
26519 (debug_verbose (const tree_node &)): New.
26520 (debug_verbose (const tree_node *)): New.
26521 (debug_head (const tree_node &)): New.
26522 (debug_head (const tree_node *)): New.
26523 (debug_body (const tree_node &)): New.
26524 (debug_body (const tree_node *)): New.
26525 (debug_vec_tree (tree): Move and reimplement in terms of dump.
26526 (debug (vec<tree, va_gc> &)): New.
26527 (debug (vec<tree, va_gc> *)): New.
26529 (extern debug (const rtx_def &)): New.
26530 (extern debug (const rtx_def *)): New.
26532 (debug_raw (simple_bitmap_def &)): New.
26533 (debug_raw (simple_bitmap_def *)): New.
26534 (debug (simple_bitmap_def &)): New.
26535 (debug (simple_bitmap_def *)): New.
26537 (extern debug (simple_bitmap_def &)): New.
26538 (extern debug (simple_bitmap_def *)): New.
26539 (extern debug_raw (simple_bitmap_def &)): New.
26540 (extern debug_raw (simple_bitmap_def *)): New.
26542 (debug (vinsn_def &)): New.
26543 (debug (vinsn_def *)): New.
26544 (debug_verbose (vinsn_def &)): New.
26545 (debug_verbose (vinsn_def *)): New.
26546 (debug (expr_def &)): New.
26547 (debug (expr_def *)): New.
26548 (debug_verbose (expr_def &)): New.
26549 (debug_verbose (expr_def *)): New.
26550 (debug (vec<rtx> &)): New.
26551 (debug (vec<rtx> *)): New.
26553 (extern debug (vinsn_def &)): New.
26554 (extern debug (vinsn_def *)): New.
26555 (extern debug_verbose (vinsn_def &)): New.
26556 (extern debug_verbose (vinsn_def *)): New.
26557 (extern debug (expr_def &)): New.
26558 (extern debug (expr_def *)): New.
26559 (extern debug_verbose (expr_def &)): New.
26560 (extern debug_verbose (expr_def *)): New.
26561 (extern debug (vec<rtx> &)): New.
26562 (extern debug (vec<rtx> *)): New.
26564 (_list_iter_cond_expr): Make inline instead of static.
26566 (debug (sreal &)): New.
26567 (debug (sreal *)): New.
26569 (extern debug (sreal &)): New.
26570 (extern debug (sreal *)): New.
26572 (extern debug_raw (const tree_node &)): New.
26573 (extern debug_raw (const tree_node *)): New.
26574 (extern debug (const tree_node &)): New.
26575 (extern debug (const tree_node *)): New.
26576 (extern debug_verbose (const tree_node &)): New.
26577 (extern debug_verbose (const tree_node *)): New.
26578 (extern debug_head (const tree_node &)): New.
26579 (extern debug_head (const tree_node *)): New.
26580 (extern debug_body (const tree_node &)): New.
26581 (extern debug_body (const tree_node *)): New.
26582 (extern debug (vec<tree, va_gc> &)): New.
26583 (extern debug (vec<tree, va_gc> *)): New.
26585 (debug (struct loop &)): New.
26586 (debug (struct loop *)): New.
26587 (debug_verbose (struct loop &)): New.
26588 (debug_verbose (struct loop *)): New.
26589 * tree-dump.c: Add header dependence.
26591 (extern debug (struct loop &)): New.
26592 (extern debug (struct loop *)): New.
26593 (extern debug_verbose (struct loop &)): New.
26594 (extern debug_verbose (struct loop *)): New.
26596 (debug (data_reference &)): New.
26597 (debug (data_reference *)): New.
26598 (debug (vec<data_reference_p> &)): New.
26599 (debug (vec<data_reference_p> *)): New.
26600 (debug (vec<ddr_p> &)): New.
26601 (debug (vec<ddr_p> *)): New.
26603 (extern debug (data_reference &)): New.
26604 (extern debug (data_reference *)): New.
26605 (extern debug (vec<data_reference_p> &)): New.
26606 (extern debug (vec<data_reference_p> *)): New.
26607 (extern debug (vec<ddr_p> &)): New.
26608 (extern debug (vec<ddr_p> *)): New.
26610 (debug (pt_solution &)): New.
26611 (debug (pt_solution *)): New.
26613 (extern debug (pt_solution &)): New.
26614 (extern debug (pt_solution *)): New.
26616 (debug (_var_map &)): New.
26617 (debug (_var_map *)): New.
26618 (debug (tree_live_info_d &)): New.
26619 (debug (tree_live_info_d *)): New.
26621 (extern debug (_var_map &)): New.
26622 (extern debug (_var_map *)): New.
26623 (extern debug (tree_live_info_d &)): New.
26624 (extern debug (tree_live_info_d *)): New.
26626 2013-03-28 Jan Hubicka <jh@suse.cz>
26628 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
26630 2013-03-28 Ian Bolton <ian.bolton@arm.com>
26632 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
26633 record only when desired or required.
26635 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
26637 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
26638 *vec_extractv2di_1_rex64. Use x64 isa attribute.
26640 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
26642 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
26643 (*andsi3_compare0_uxtw): New pattern.
26644 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
26645 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
26647 2013-03-28 Jan Hubicka <jh@suse.cz>
26649 * data-streamer-in.c (streamer_read_gcov_count): New function.
26650 * gimple-streamer-out.c: Include value-prof.h.
26651 (output_gimple_stmt): Output histogram.
26652 (output_bb): Use streamer_write_gcov_count.
26653 * value-prof.c: Include data-streamer.h
26654 (dump_histogram_value): Add HIST_TYPE_MAX.
26655 (stream_out_histogram_value): New function.
26656 (stream_in_histogram_value): New function.
26657 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
26658 (stream_out_histogram_value, stream_in_histogram_value): Declare.
26659 * data-streamer-out.c (streamer_write_gcov_count): New function.
26660 (streamer_write_gcov_count_stream): New function.
26661 * lto-cgraph.c (lto_output_edge): Update counter streaming.
26662 (lto_output_node): Likewise.
26663 (input_node, input_edge): Likewise.
26664 * lto-streamer-out.c (output_cfg): Update streaming.
26665 * lto-streamer-in.c (input_cfg): Likewise.
26666 * data-streamer.h (streamer_write_gcov_count,
26667 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
26668 * gimple-streamer-in.c: Include value-prof.h
26669 (input_gimple_stmt): Input histograms.
26670 (input_bb): Update profile streaming.
26672 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
26674 * genmodes.c (emit_max_int): New function.
26675 (emit_insn_modes_h): Added call to emit_max_function.
26676 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
26678 * machmode.def: Fixed comment.
26680 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
26682 * combine.c (try_combine): Removed useless assert.
26683 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
26685 2013-03-28 Marek Polacek <polacek@redhat.com>
26686 Richard Biener <rguenther@suse.de>
26688 PR tree-optimization/56695
26689 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
26690 build signed result of a vector comparison.
26691 * tree-cfg.c (verify_gimple_comparison): Check that a result
26692 of a vector comparison has signed type.
26694 2013-03-28 Richard Biener <rguenther@suse.de>
26696 PR tree-optimization/37021
26697 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
26698 do not restrict gaps between groups.
26699 * tree-vect-stmts.c (vectorizable_load): Properly account for
26700 a gap between groups.
26702 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
26704 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
26705 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
26708 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
26710 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
26711 * doc/extend.texi (Named Address Spaces): Ditto.
26712 (Variable Attributes): Ditto.
26714 2013-03-27 Kai Tietz <ktietz@redhat.com>
26716 * config.build: Add support for cygwin x64 target.
26717 * config.gcc: Likewise.
26718 * config.host: Likewise.
26719 * configure.ac: Likewise
26720 * configure: Regenerated.
26722 2013-03-27 Kai Tietz <ktietz@redhat.com>
26724 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
26725 * config/i386/t-cygwin-w64: New file.
26726 * config/i386/cygwin-w64.h: New file.
26727 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
26728 and add support for x64-cygwin target.
26729 (CPP_SPEC): Likewise.
26730 (CXX_WRAP_SPEC_LIST): Undefine before define.
26731 (LIBGCJ_SONAME): Use 15 as version.
26733 2013-03-27 Richard Biener <rguenther@suse.de>
26735 PR tree-optimization/56716
26736 * tree-ssa-structalias.c (perform_var_substitution): Adjust
26737 dumping for ref nodes.
26739 2013-03-27 Martin Jambor <mjambor@suse.cz>
26741 PR tree-optimization/55334
26742 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
26743 restricted pointers to arrays.
26745 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
26747 * Makefile.in (.SUFFIXES): Add .cc.
26748 (.c.o): Apply same recipe for implicit rule .cc.o.
26750 2013-03-27 Richard Biener <rguenther@suse.de>
26752 PR tree-optimization/37021
26753 * tree-vect-data-refs.c (vect_check_strided_load): Allow
26754 REALPART/IMAGPART_EXPRs around the supported refs.
26755 * tree-ssa-structalias.c (find_func_aliases): Assume that
26756 floating-point values are not used to transfer pointers.
26758 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
26760 * target.def (TARGET_HAS_IFUNC_P): New target hook.
26761 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
26762 * doc/tm.texi: Regenerate.
26763 * targhooks.h (default_has_ifunc_p): New.
26764 * targhooks.c (default_has_ifunc_p): Ditto.
26765 * config/linux-protos.h: New file.
26766 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
26767 hook for linux which disables support of indirect functions in android.
26768 * config/linux-android.c: New file.
26769 * config/t-linux-android.c: Ditto.
26770 * config.gcc: Added new object file linux-android.o.
26771 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
26772 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
26773 * varasm.c (do_assemble_alias): Likewise.
26774 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
26775 doesn't support indirect functions.
26776 * configure: Regenerate.
26778 2013-03-27 Bin Cheng <bin.cheng@arm.com>
26781 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
26782 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
26785 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26787 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
26789 2013-03-27 Terry Guo <terry.guo@arm.com>
26791 * config/arm/arm-cores.def: Added core cortex-r7.
26792 * config/arm/arm-tune.md: Regenerated.
26793 * config/arm/arm-tables.opt: Regenerated.
26794 * doc/invoke.texi: Added entry for core cortex-r7.
26796 2013-03-27 Walter Lee <walt@tilera.com>
26798 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
26799 double-decrement of next_scratch_regno.
26801 2013-03-27 Walter Lee <walt@tilera.com>
26803 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
26805 (insn_v1mulus): Ditto.
26806 (insn_v2muls): Ditto.
26808 2013-03-27 Walter Lee <walt@tilera.com>
26810 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
26811 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
26813 2013-03-27 Walter Lee <walt@tilera.com>
26815 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
26816 (*sibcall_value): Ditto.
26818 2013-03-27 Walter Lee <walt@tilera.com>
26820 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
26821 (insn_mnz_v8qi): ... this ...
26822 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
26824 (insn_v<n>mnz): Replaced by ...
26825 (insn_v1mnz): ... this ...
26826 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
26828 (insn_mz_<mode>): Replaced by ...
26829 (insn_mz_v8qi): ... this ...
26830 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
26832 (insn_v<n>mz): Replaced by ...
26833 (insn_v1mz): ... this ...
26834 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
26837 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
26839 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
26841 2013-03-26 Roland McGrath <mcgrathr@google.com>
26843 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
26844 than fprintf with a non-constant, non-format string.
26846 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
26848 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
26849 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
26850 operand 0 predicate.
26851 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
26852 attribute. Use general_x64nomem_operand as operand 1 predicate.
26853 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
26854 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
26855 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
26856 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
26857 (mov<mode>_insv_1): Remove expander. Merge insn with
26858 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
26859 Use general_x64nomem_operand as operand 1 predicate.
26860 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
26861 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
26862 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
26863 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
26864 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
26865 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
26866 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
26867 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
26868 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
26869 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
26870 (general_x64nomem_operand): Ditto.
26872 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
26874 * config/rtems.opt: Add -pthread option.
26876 2013-03-26 Richard Biener <rguenther@suse.de>
26878 * alias.c (find_base_term): Avoid redundant and not used recursion.
26879 (base_alias_check): Get the initial base term from the caller.
26880 (true_dependence_1): Compute and pass base terms to base_alias_check.
26881 (write_dependence_p): Likewise.
26882 (may_alias_p): Likewise.
26884 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
26886 * config/aarch64/aarch64.c (aarch64_classify_address): Support
26887 PC-relative load in SI modes and above only.
26889 2013-03-26 Xinyu Qi <xyqi@marvell.com>
26891 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
26892 * config/arm/iwmmxt.md (WCGR0): Update.
26893 (WCGR1, WCGR2, WCGR3): Likewise.
26895 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
26897 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
26898 Use x64 and nox64 isa attributes.
26900 2013-03-26 Richard Biener <rguenther@suse.de>
26902 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
26903 alignment computations and rely on get_object_alignment_1
26904 for the !TYPE_P case.
26905 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
26907 2013-03-26 Walter Lee <walt@tilera.com>
26909 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
26910 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
26912 2013-03-25 Jeff Law <law@redhat.com>
26914 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
26915 check for INTEGRAL_TYPE_P that was missing due to checking in
26916 wrong version of prior patch.
26918 2013-03-25 Walter Lee <walt@tilera.com>
26920 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
26921 TILEGX_INSN_SHUFFLEBYTES1.
26922 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
26924 (tilegx_builtins): Ditto.
26925 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
26927 2013-03-25 Walter Lee <walt@tilera.com>
26929 * config/tilegx/tilegx.md (floatsisf2): New pattern.
26930 (floatunssisf2): New pattern.
26931 (floatsidf2): New pattern.
26932 (floatunssidf2): New pattern.
26934 2013-03-25 Walter Lee <walt@tilera.com>
26936 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
26937 tests for constraint J, K, N, P.
26939 2013-03-25 Walter Lee <walt@tilera.com>
26941 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
26942 Use indirect/pcrel encoding.
26943 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
26946 2013-03-25 Steve Ellcey <sellcey@mips.com>
26948 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
26949 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
26950 * config/mips/mips.c (mips_option_override): Set IMADD default.
26951 * config/mips/mips.h (PTF_AVOID_IMADD): New.
26952 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
26953 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
26954 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
26955 * doc/invoke.texi (-mimadd/-mno-imadd): New.
26957 2013-03-25 Jeff Law <law@redhat.com>
26959 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
26960 slightly to avoid creating and folding useless trees. Simplify
26961 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
26963 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
26965 * config/i386/i386.md (*zero_extendsidi2): Merge with
26966 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
26967 * config/i386/predicates.md (x86_64_zext_operand): Rename from
26968 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
26969 targets. Clarify comment.
26971 2013-03-25 Martin Jambor <mjambor@suse.cz>
26973 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
26974 pass-through jump functions differently.
26975 (ipa_read_jump_function): Likewise. Also use setter functions to set
26978 2013-03-25 Martin Jambor <mjambor@suse.cz>
26980 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
26981 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
26983 (ipa_get_indirect_edge_target): New function.
26984 (devirtualization_time_bonus): New parameter known_aggs, pass it to
26985 ipa_get_indirect_edge_target. Update all callers.
26986 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
26987 ipa_get_indirect_edge_target_1 instead of calling
26988 ipa_get_indirect_edge_target.
26989 (create_specialized_node): Pass aggvlas to
26990 ipcp_discover_new_direct_edges.
26992 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
26994 * config/arm/arm.md (f_sels, f_seld): New types.
26995 (*cmov<mode>): New pattern.
26996 * config/arm/predicates.md (arm_vsel_comparison_operator): New
26999 2013-03-25 Kai Tietz <ktietz@redhat.com>
27001 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
27002 POSIX-printf for mingw-hosted builds.
27004 2013-03-25 Richard Biener <rguenther@suse.de>
27006 PR middle-end/56694
27007 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
27008 must-not-throw stmt location.
27010 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
27012 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
27013 Emit load-acquire versions when acq is true.
27014 (arm_emit_store_exclusive): Add rel parameter.
27015 Emit store-release versions when rel is true.
27016 (arm_split_compare_and_swap): Use acquire-release instructions
27018 of barriers when appropriate.
27019 (arm_split_atomic_op): Likewise.
27020 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
27021 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
27022 (VUNSPEC_SLX): Likewise.
27023 (VUNSPEC_LDA): Likewise.
27024 (VUNSPEC_STL): Likewise.
27025 * config/arm/sync.md (atomic_load<mode>): New pattern.
27026 (atomic_store<mode>): Likewise.
27027 (arm_load_acquire_exclusive<mode>): Likewise.
27028 (arm_load_acquire_exclusivesi): Likewise.
27029 (arm_load_acquire_exclusivedi): Likewise.
27030 (arm_store_release_exclusive<mode>): Likewise.
27032 2013-03-25 Catherine Moore <clm@codesourcery.com>
27034 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
27035 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
27036 * config/mip/predicates.md (lwsp_swsp_operand,
27037 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
27038 sb16_operand, db4_operand, db7_operand, ib3_operand,
27039 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
27040 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
27041 andi16_operand): New predicates.
27042 * config/mips/mips.md (compression): New attribute.
27043 (enabled): New attribute.
27044 (length): Consider compression in computing length.
27045 (shift_compression): New code attribute.
27046 (*add<mode>3): New operands. Record compression.
27047 (sub<mode>3): Likewise.
27048 (one_cmpl<mode>2): Likewise.
27049 (*and<mode>3): Likewise.
27050 (*ior<mode>3): Likewise.
27051 (unnamed pattern for xor): Likewise.
27052 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
27053 (*<optab><mode>3): Likewise.
27054 (*mov<mode>_internal: Likewise.
27055 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
27056 (mips_unsigned_immediate_p): New.
27057 (umips_lwsp_swsp_address_p): New.
27058 (m16_based_address_p): New.
27059 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
27060 (mips_unsigned_immediate_p): New prototype.
27061 (lwsp_swsp_address_p): New prototype.
27062 (m16_based_address_p): New prototype.
27063 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
27064 (mips_signed_immediate_p): New function.
27065 (m16_based_address_p): New function.
27066 (lwsp_swsp_address_p): New function.
27067 (mips_print_operand_punctuation): Recognize short delay slot insns
27068 for microMIPS.add<mode>3"
27070 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
27073 * config/arm/iterators.md (v_cmp_result): New mode attribute.
27074 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
27076 2013-03-25 Richard Biener <rguenther@suse.de>
27078 PR tree-optimization/56689
27079 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
27082 2013-03-25 Richard Biener <rguenther@suse.de>
27084 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
27086 (memory_references): Likewise.
27087 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
27088 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
27089 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
27090 (gather_mem_refs_in_loops): Fold into ...
27091 (analyze_memory_references): ... this. Move initialization
27092 to tree_ssa_lim_initialize.
27093 (fill_always_executed_in): Rename to ...
27094 (fill_always_executed_in_1): ... this.
27095 (fill_always_executed_in): Move contains_call computation to
27096 this new function from ...
27097 (tree_ssa_lim_initialize): ... here.
27098 (tree_ssa_lim): Call fill_always_executed_in.
27100 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
27102 * postreload.c (reload_combine): Fix code detecting returns.
27104 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
27106 * function.c (emit_use_return_register_into_block): On cc0 targets,
27107 do not emit the sequence between cc0 setter and user.
27109 2013-03-25 Kai Tietz <ktietz@redhat.com>
27111 * config/i386/predicates.md (local_symbolic_operand): Interpret
27112 dll-imported symbols as none-local.
27114 2013-03-25 Richard Biener <rguenther@suse.de>
27116 * tree-ssa-loop-im.c (struct depend): Remove.
27117 (struct lim_aux_data): Make depends a vec of gimples.
27118 (free_lim_aux_data): Adjust.
27119 (add_dependency): Likewise.
27120 (set_level): Likewise.
27122 2013-03-25 Richard Biener <rguenther@suse.de>
27124 PR middle-end/56434
27125 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
27126 the pointer returned by calls with ECF_MALLOC set.
27128 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
27130 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
27132 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
27134 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
27135 using MMXMODE mode iterator.
27136 (*move<mode>_internal): Merge with *movv2sf_internal and
27137 *movv2sf_internal_rex64 using MMXMODE mode iterator.
27139 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
27141 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
27142 (record_last_mem_set_info): Likewise.
27144 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
27145 of XNEWVEC followed by memset.
27146 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
27148 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
27150 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
27151 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
27152 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
27153 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
27154 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
27155 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
27156 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
27157 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
27158 BARRIER_P instead of GET_CODE.
27160 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
27162 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
27163 inaccuracy in the probing code.
27165 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
27166 (ctrapdi4): Likewise.
27168 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
27170 * calls.c (expand_call): Add missing guard to code handling return
27171 of non-BLKmode structures in MSB.
27172 * function.c (expand_function_end): Likewise.
27174 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
27176 * combine.c (try_combine): Adjust comment. Do not add the set of
27177 insn #0 if the destination indirectly is set or dies in insn #2.
27178 Tidy up code to distribute a new note.
27180 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
27182 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
27183 also for alternatives 16 and 17.
27185 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
27187 * config/i386/sse.md (*mov<mode>_internal): Merge with
27188 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
27189 Emit insn template depending on type attribute. Use
27190 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
27191 movd instead of movq mnemonic for interunit moves. Rewrite mode
27192 attribute calculation. Remove unit attribute calculation.
27193 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
27194 Set prefix_data16 attribute for DImode ssemov types.
27195 Use Ym instead of y for SSE-MMX conversion alternatives.
27196 Reorder operand constraints.
27198 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
27200 * df.h (df_insn_delete): Adjust prototype.
27201 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
27202 and let it decide whether mark the basic block dirty.
27203 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
27204 * df-scan.c (df_insn_info_delete): New helper function, split
27205 off from df_insn_delete.
27206 (df_scan_free_bb_info): Use it.
27207 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
27209 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
27210 that the insn is actually an insn and it has a non-NULL basic block.
27211 Do not mark basic block dirty if only deleting a DEBUG_INSN.
27213 2013-03-22 Richard Biener <rguenther@suse.de>
27215 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
27217 (mem_ref_alloc): Do not allocate them.
27218 (refs_independent_p): Do not query or maintain a cache.
27220 2013-03-22 Richard Biener <rguenther@suse.de>
27222 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
27223 (gather_mem_refs_in_loops): Do not compute it.
27224 (analyze_memory_references): Do not allocate it.
27225 (tree_ssa_lim_finalize): Do not free it.
27226 (for_all_locs_in_loop): Do not query all_refs_in_loop.
27228 2013-03-22 Richard Biener <rguenther@suse.de>
27230 * is-a.h (as_a): Use gcc_checking_assert.
27232 2013-03-22 Ian Bolton <ian.bolton@arm.com>
27234 * config/aarch64/aarch64.c (aarch64_print_operand): New
27235 format specifier for printing a constant in hex.
27236 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
27237 format specifier for printing second operand.
27239 2013-03-22 Richard Biener <rguenther@suse.de>
27241 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
27243 (gather_mem_refs_in_loops): Perform store accumulation here.
27244 (create_vop_ref_mapping_loop): Remove.
27245 (create_vop_ref_mapping): Likewise.
27246 (analyze_memory_references): Initialize refs_stored_in_loop.
27247 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
27248 (record_indep_loop): Remove.
27249 (record_dep_loop): New function.
27250 (ref_indep_loop_p_1): Adjust to only walk over references
27251 in the loop, not its subloops.
27252 (ref_indep_loop_p): Rename to ...
27253 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
27254 maintaining a more fine-grained cache.
27255 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
27256 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
27258 2013-03-22 Richard Biener <rguenther@suse.de>
27260 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
27261 (struct mem_ref): Make accesses_in_loop a vec of a vec of
27262 aggregate mem_ref_loc.
27263 (free_mem_ref_locs): Inline into ...
27264 (memref_free): ... this and adjust.
27265 (mem_ref_alloc): Adjust.
27266 (mem_ref_locs_alloc): Remove.
27267 (record_mem_ref_loc): Adjust.
27268 (get_all_locs_in_loop): Rewrite into ...
27269 (for_all_locs_in_loop): ... this iterator.
27270 (rewrite_mem_ref_loc): New functor.
27271 (rewrite_mem_refs): Use for_all_locs_in_loop.
27272 (sm_set_flag_if_changed): New functor.
27273 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
27274 (ref_always_accessed): New functor.
27275 (ref_always_accessed_p): Use for_all_locs_in_loop.
27277 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
27279 * tree-pass.h (PROP_gimple_lvec): New.
27280 * passes.c (dump_properties): Handle PROP_gimple_lvec.
27281 (init_optimization_passes): Move pass_lower_vector.
27282 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
27284 (pass_lower_vector): Provide PROP_gimple_lvec.
27285 (pass_lower_vector_ssa): Likewise.
27286 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
27288 2013-03-21 Mark Wielaard <mjw@redhat.com>
27290 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
27292 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
27294 * config/i386/i386.md (*movdi_internal): Disparage slightly
27295 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
27296 conversion alternatives.
27298 2013-03-21 Jakub Jelinek <jakub@redhat.com>
27300 PR middle-end/48087
27301 * diagnostic.def (DK_WERROR): New kind.
27302 * diagnostic.h (werrorcount): Define.
27303 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
27304 promoted to DK_ERROR, increment DK_WERROR counter instead of
27306 * toplev.c (toplev_main): Call print_ignored_options even if
27307 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
27308 even if just werrorcount is non-zero.
27311 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
27313 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
27314 (string_cst_pool_decl): New function.
27315 (optimize_one_addr_into_implicit_ptr): New function.
27316 (resolve_addr_in_expr): Optimize DWARF location expression
27317 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
27318 which doesn't live in memory, but has DW_AT_location or
27319 DW_AT_const_value, or refers to a string literal, into
27320 DW_OP_GNU_implicit_pointer.
27321 (optimize_location_into_implicit_ptr): New function.
27322 (resolve_addr): If removing DW_AT_location of a variable because
27323 it was DW_OP_addr of address of the variable, but the variable doesn't
27324 live in memory, try to emit const value attribute for the initializer.
27326 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
27328 * tree.h (VECTOR_TYPE_P): New macro.
27329 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
27330 TYPE_MODE): Use it.
27331 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
27332 VEC_COND_EXPR cannot be lvalues.
27333 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
27335 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
27337 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
27338 Restrict the transformation to equal modes.
27340 2013-03-21 Richard Biener <rguenther@suse.de>
27342 PR tree-optimization/39326
27343 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
27344 (MEM_ANALYZABLE): Adjust.
27345 (record_mem_ref_loc): Move bitmap ops ...
27346 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
27347 unanalyzable refs, do not record locations for it.
27348 (analyze_memory_references): Allocate ref zero as shared
27350 (refs_independent_p): Do not test for unanalyzed mems here.
27351 (ref_indep_loop_p_1): Special-case disambiguation against
27352 the unanalyzed ref.
27353 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
27355 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
27357 * config/arm/arm-protos.h (tune_params): Add
27358 prefer_neon_for_64bits field.
27359 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
27360 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
27361 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
27362 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
27363 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
27364 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
27365 (arm_option_override): Handle -mneon-for-64bits new option.
27366 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
27367 (prefer_neon_for_64bits): Declare new variable.
27368 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
27369 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
27370 (arch_enabled): Handle new arch types. Remove support for onlya8
27372 (one_cmpldi2): Use new arch names.
27373 (zero_extend<mode>di2, extend<mode>di2): Ditto.
27374 * config/arm/arm.opt (mneon-for-64bits): Add option.
27375 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
27376 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
27377 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
27379 * doc/invoke.texi (-mneon-for-64bits): Document.
27381 2013-03-21 Richard Biener <rguenther@suse.de>
27383 PR tree-optimization/39326
27384 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
27385 (sort_bbs_in_loop_postorder_cmp): New function.
27386 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
27388 2013-03-21 Richard Biener <rguenther@suse.de>
27390 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
27391 (vect_insert_into_interleaving_chain): Likewise.
27392 (vect_drs_dependent_in_basic_block): Inline ...
27393 (vect_slp_analyze_data_ref_dependence): ... here. New function,
27395 (vect_analyze_data_ref_dependence): ... here. Simplify.
27396 (vect_check_interleaving): Simplify.
27397 (vect_analyze_data_ref_dependences): Likewise. Split out ...
27398 (vect_slp_analyze_data_ref_dependences): ... this new function.
27399 (dr_group_sort_cmp): New function.
27400 (vect_analyze_data_ref_accesses): Compute data-reference groups
27401 here instead of in vect_analyze_data_ref_dependence. Use
27402 a more efficient algorithm.
27403 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
27404 vect_slp_analyze_data_ref_dependences. Call
27405 vect_analyze_data_ref_accesses earlier.
27406 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
27407 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
27408 (vect_slp_analyze_data_ref_dependences): New prototype.
27410 2013-03-21 Richard Biener <rguenther@suse.de>
27412 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
27413 ref is stored in the loop.
27414 (find_refs_for_sm): Walk only over all stores.
27415 (store_motion_loop): Allocate from lim_bitmap_obstack.
27416 (store_motion): Likewise.
27418 2013-03-21 Richard Biener <rguenther@suse.de>
27420 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
27421 Update virtual SSA form.
27423 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
27425 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
27426 * configure: Regenerate.
27427 * config.in: Regenerate.
27428 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
27429 if !HAVE_LD_EH_FRAME_CIEV3.
27431 2013-03-21 Richard Biener <rguenther@suse.de>
27433 * tree-cfg.c (verify_expr_no_block): New function.
27434 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
27435 nor DECL_VALUE_EXPR have locations with associated blocks.
27436 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
27437 (clear_unused_block_pointer): Remove code dealing with
27438 blocks in DECL_DEBUG_EXPR locations.
27440 2013-03-21 Richard Biener <rguenther@suse.de>
27442 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
27443 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
27444 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
27445 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
27446 instead of DECL_DEBUG_EXPR_IS_FROM.
27447 * gimplify.c (gimplify_modify_expr): Likewise.
27448 * tree-cfg.c (verify_expr_location_1): Likewise.
27449 * tree-complex.c (create_one_component_var): Likewise.
27450 * tree-sra.c (create_access_replacement): Likewise.
27451 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
27452 (clear_unused_block_pointer): Likewise.
27453 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
27454 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
27455 * var-tracking.c (var_debug_decl): Likewise.
27456 (track_expr_p): Likewise.
27457 * tree-inline.c (add_local_variables): Likewise. Set
27458 DECL_HAS_DEBUG_EXPR_P after copying it.
27459 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
27460 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
27462 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
27465 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
27466 * configure: Regenerate.
27467 * config.in: Regenerate.
27468 * config/i386/i386.md (*movdf_internal): Use
27469 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
27470 movd instead of movq mnemonic for interunit moves.
27471 (*movdi_internal): Ditto.
27473 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
27475 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
27476 (abd<mode>_3): New pattern.
27477 (aba<mode>_3): New pattern.
27478 (fabd<mode>_3): New pattern.
27480 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
27482 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
27483 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
27484 occurrence of REGISTER_PREFIX as its empty string.
27486 2013-03-20 Jeff Law <law@redhat.com>
27488 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
27489 addititional equivalences for equality comparisons between an SSA_NAME
27490 and a constant where the SSA_NAME was set from a widening conversion.
27492 2013-03-20 Walter Lee <walt@tilera.com>
27494 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
27496 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
27498 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
27499 depending on type attribute.
27500 (*movti_internal): Ditto.
27501 (*movtf_internal): Ditto.
27502 (*movxf_internal): Ditto.
27503 (*movdf_internal): Ditto.
27504 (*movsf_internal): Ditto.
27506 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
27508 * config/i386/i386.md (*movti_internal): Set prefix attribute to
27509 maybe_vex for sselog1 and ssemov types.
27510 (*movdi_internal): Reorder operand constraints.
27511 (*movsi_internal): Ditto. Set prefix attribute to
27512 maybe_vex for sselog1 and ssemov types.
27513 (*movtf_internal): Set prefix attribute to maybe_vex
27514 for sselog1 and ssemov types.
27515 (*movdf_internal): Ditto. Set prefix_data16 attribute for
27516 DImode ssemov types. Reorder operand constraints.
27517 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
27518 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
27519 attribute for SImode ssemov types. Reorder operand constraints.
27521 2013-03-20 Martin Jambor <mjambor@suse.cz>
27523 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
27524 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
27526 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
27528 * config/rs6000/predicates.md (indexed_address, update_address_mem
27529 update_indexed_address_mem): New predicates.
27530 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
27531 attribute for load/store instructions.
27532 * config/rs6000/dfp.md (movsd_store): Likewise.
27533 (movsd_load): Likewise.
27534 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
27535 (unnamed HI->DI extend define_insn): Likewise.
27536 (unnamed SI->DI extend define_insn): Likewise.
27537 (unnamed QI->SI extend define_insn): Likewise.
27538 (unnamed QI->HI extend define_insn): Likewise.
27539 (unnamed HI->SI extend define_insn): Likewise.
27540 (unnamed HI->SI extend define_insn): Likewise.
27541 (extendsfdf2_fpr): Likewise.
27542 (movsi_internal1): Likewise.
27543 (movsi_internal1_single): Likewise.
27544 (movhi_internal): Likewise.
27545 (movqi_internal): Likewise.
27546 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
27547 attribute for load/store instructions.
27548 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
27550 (mov<mode>_softfloat): Likewise.
27551 (mov<mode>_hardfloat32): Likewise.
27552 (mov<mode>_hardfloat64): Likewise.
27553 (mov<mode>_softfloat64): Likewise.
27554 (movdi_internal32): Likewise.
27555 (movdi_internal64): Likewise.
27556 (probe_stack_<mode>): Likewise.
27558 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
27560 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
27561 floating point, and decimal floating point to reload iterator.
27563 * config/rs6000/constraints.md (wl constraint): New constraints to
27564 return FLOAT_REGS if certain options are used to reduce the number
27565 of separate patterns that exist in the file.
27566 (wx constraint): Likewise.
27567 (wz constraint): Likewise.
27569 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
27570 -mdebug=reg, print wg, wl, wx, and wz constraints.
27571 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
27572 Initialize the reload functions for 64-bit binary/decimal floating
27574 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
27575 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
27576 create the buffer on the stack to overcome not having a 32-bit
27578 (rs6000_emit_move): Likewise.
27579 (rs6000_secondary_memory_needed_rtx): Likewise.
27580 (rs6000_alloc_sdmode_stack_slot): Likewise.
27581 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
27582 via xxlxor, just like DFmode 0.0.
27584 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
27585 define as 1 if we are running on a power7 or newer.
27586 (enum r6000_reg_class_enum): Add new constraints.
27588 * config/rs6000/dfp.md (movsd): Delete, combine with binary
27589 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
27590 with other moves by using conditional constraits (wg). Use LFIWZX
27591 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
27592 (movsd splitter): Likewise.
27593 (movsd_hardfloat): Likewise.
27594 (movsd_softfloat): Likewise.
27596 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
27597 binary and decimal floating point moves.
27598 (fmove_ok): New attributes to combine binary and decimal floating
27599 point moves, and to combine power6x (mfpgpr) moves along normal
27601 (real_value_to_target): Likewise.
27602 (f32_lr): Likewise.
27603 (f32_lm): Likewise.
27604 (f32_li): Likewise.
27605 (f32_sr): Likewise.
27606 (f32_sm): Likewise.
27607 (f32_si): Likewise.
27608 (movsf): Combine binary and decimal floating point moves. Combine
27609 power6x (mfpgpr) moves with other moves by using conditional
27610 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
27611 (mov<mode> for SFmode/SDmode); Likewise.
27612 (SFmode/SDmode splitters): Likewise.
27613 (movsf_hardfloat): Likewise.
27614 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
27615 (movsf_softfloat): Likewise.
27616 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
27618 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
27619 wx and wz constraints.
27621 * config/rs6000/constraints.md (wg constraint): New constraint to
27622 return FLOAT_REGS if -mmfpgpr (power6x) was used.
27624 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
27627 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
27628 -mdebug=reg, print wg, wl, wx, and wz constraints.
27629 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
27630 Initialize the reload functions for 64-bit binary/decimal floating
27632 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
27633 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
27634 create the buffer on the stack to overcome not having a 32-bit
27636 (rs6000_emit_move): Likewise.
27637 (rs6000_secondary_memory_needed_rtx): Likewise.
27638 (rs6000_alloc_sdmode_stack_slot): Likewise.
27639 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
27640 via xxlxor, just like DFmode 0.0.
27642 * config/rs6000/dfp.md (movdd): Delete, combine with binary
27643 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
27644 with other moves by using conditional constraits (wg). Use LFIWZX
27645 and STFIWX for loading SDmode on power7.
27646 (movdd splitters): Likewise.
27647 (movdd_hardfloat32): Likewise.
27648 (movdd_softfloat32): Likewise.
27649 (movdd_hardfloat64_mfpgpr): Likewise.
27650 (movdd_hardfloat64): Likewise.
27651 (movdd_softfloat64): Likewise.
27653 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
27654 64-bit binary and decimal floating point moves.
27655 (FMOVE64X): Likewise.
27656 (movdf): Combine 64-bit binary and decimal floating point moves.
27657 Combine power6x (mfpgpr) moves with other moves by using
27658 conditional constraits (wg).
27659 (mov<mode> for DFmode/DDmode): Likewise.
27660 (DFmode/DDmode splitters): Likewise.
27661 (movdf_hardfloat32): Likewise.
27662 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
27663 (movdf_softfloat32): Likewise.
27664 (movdf_hardfloat64_mfpgpr): Likewise.
27665 (movdf_hardfloat64): Likewise.
27666 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
27667 (movdf_softfloat64): Likewise.
27668 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
27669 (reload_<mode>_load): Move to later in the file so they aren't in
27670 the middle of the floating point move insns.
27671 (reload_<mode>_store): Likewise.
27673 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
27676 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
27677 constraint if -mdebug=reg.
27678 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
27679 Enable using dd reload support if needed.
27681 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
27682 binary and decimal floating point moves in rs6000.md.
27683 (movtd_internal): Likewise.
27685 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
27686 decimal floating point moves.
27688 (movtf_internal): Likewise.
27689 (mov<mode>_internal, TDmode/TFmode): Likewise.
27690 (movtf_softfloat): Likewise.
27691 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
27693 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
27694 movdi_internal64, using wg constraint for move direct operations.
27695 (movdi_internal64): Likewise.
27697 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
27698 MODES_TIEABLE_P for selected modes. Print the numerical value of
27699 the various virtual registers. Use GPR/FPR first/last values,
27700 instead of hard coding the register numbers. Print which modes
27701 have reload functions registered.
27702 (rs6000_option_override_internal): If -mdebug=reg, trace the options
27703 settings before/after setting cpu, target and subtarget settings.
27704 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
27705 and for secondary reload failures in rs6000_secondary_reload_inner.
27706 (rs6000_secondary_reload_fail): Likewise.
27707 (rs6000_secondary_reload_inner): Likewise.
27709 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
27710 macros for first/last GPR and FPR registers.
27711 (LAST_GPR_REGNO): Likewise.
27712 (FIRST_FPR_REGNO): Likewise.
27713 (LAST_FPR_REGNO): Likewise.
27715 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
27716 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
27717 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
27718 (vcond<mode><mode>): Likewise.
27719 (vcondu<mode><mode>): Likewise.
27720 (vector_gtu<mode>): Likewise.
27721 (vector_gte<mode>): Likewise.
27722 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
27723 to prevent the compiler from converting DImode operations to TImode.
27724 (ior<mode>3): Likewise.
27725 (and<mode>3): Likewise.
27726 (one_cmpl<mode>2): Likewise.
27727 (nor<mode>3): Likewise.
27728 (andc<mode>3): Likewise.
27730 * config/rs6000/constraints.md (wt constraint): New constraint
27731 that returns VSX_REGS if TImode is allowed in VSX registers.
27733 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
27734 constant under VSX.
27736 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
27737 similar to TImode, but it is restricted to being in the GPRs.
27739 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
27740 TImode to occupy a single VSX register.
27742 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
27743 -mvsx-timode for power7/power8.
27744 (power7 cpu): Likewise.
27745 (power8 cpu): Likewise.
27747 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
27748 sure that TFmode/TDmode take up two registers if they are ever
27749 allowed in the upper VSX registers.
27750 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
27752 (rs6000_init_hard_regno_mode_ok): Likewise.
27753 (rs6000_debug_reg_global): Add debugging for PTImode and wt
27754 constraint. Print if LRA is turned on.
27755 (rs6000_option_override_internal): Give an error if -mvsx-timode
27756 and VSX is not enabled.
27757 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
27758 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
27759 to reg+offset addressing. Use PTImode when checking offset
27760 addresses for validity.
27761 (reg_offset_addressing_ok_p): Likewise.
27762 (rs6000_legitimate_offset_address_p): Likewise.
27763 (rs6000_legitimize_address): Likewise.
27764 (rs6000_legitimize_reload_address): Likewise.
27765 (rs6000_legitimate_address_p): Likewise.
27766 (rs6000_eliminate_indexed_memrefs): Likewise.
27767 (rs6000_emit_move): Likewise.
27768 (rs6000_secondary_reload): Likewise.
27769 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
27770 reloads to fpr registers to continue to use reg+offset addressing,
27771 but 64-bit reloads to altivec registers need reg+reg addressing.
27772 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
27773 it. Treat LO_SUM like a PLUS operation.
27774 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
27775 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
27776 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
27777 registers to share a register with a smaller sized type, since VSX
27778 puts scalars in the upper 64-bits.
27779 (print_operand): Add support for PTImode.
27780 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
27781 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
27782 registers, but don't have arithmetic support.
27783 (rs6000_memory_move_cost): Add test for VSX.
27784 (rs6000_opt_masks): Add -mvsx-timode.
27786 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
27789 (VSr): Use wt constraint for TImode.
27790 (VSv): Drop TImode support.
27791 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
27792 (vsx_movti_64bit): Likewise.
27793 (vsx_movti_32bit): Likewise.
27794 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
27795 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
27796 one '?' on the appropriate output constraint. Do not allow TImode
27797 logical operations on 32-bit systems.
27798 (vsx_ior<mode>3): Likewise.
27799 (vsx_xor<mode>3): Likewise.
27800 (vsx_one_cmpl<mode>2): Likewise.
27801 (vsx_nor<mode>3): Likewise.
27802 (vsx_andc<mode>3): Likewise.
27803 (vsx_concat_<mode>): Likewise.
27804 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
27806 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
27807 OPTION_MASK_VSX_TIMODE.
27808 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
27809 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
27811 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
27812 (TI2 iterator): New iterator for TImode, PTImode.
27813 (wd mode attribute): Add values for vector types.
27814 (movti_string): Replace TI move operations with operations for TImode
27815 and PTImode. Add support for TImode being allowed in VSX registers.
27816 (mov<mode>_string, TImode/PTImode): Likewise.
27817 (movti_ppc64): Likewise.
27818 (mov<mode>_ppc64, TImode/PTImode): Likewise.
27819 (TI mode splitters): Likewise.
27821 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
27824 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
27826 PR tree-optimization/56355
27827 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
27828 Also handle integers with undefined overflow.
27830 2013-03-20 Catherine Moore <clm@codesourcery.com>
27831 Maciej W. Rozycki <macro@codesourcery.com>
27832 Tom de Vries <tom@codesourcery.com>
27833 Nathan Sidwell <nathan@codesourcery.com>
27834 Iain Sandoe <iain@codesourcery.com>
27835 Nathan Froyd <froydnj@codesourcery.com>
27836 Chao-ying Fu <fu@mips.com>
27838 * doc/extend.texi: (micromips, nomicromips, nocompression):
27839 Document new function attributes.
27840 * doc/invoke.texi (minterlink-compressed, mmicromips,
27841 m14k, m14ke, m14kec): Document new options.
27842 (minterlink-mips16): Update documentation.
27843 * doc/md.texi (ZC, ZD): Document new constraints.
27844 * configure.ac (gcc_cv_as_micromips): Check if linker
27845 supports the .set micromips directive.
27846 * configure: Regenerate.
27847 * config.in: Regenerate.
27848 * config/mips/mips-tables.opt: Regenerate.
27849 * config/mips/micromips.md: New file.
27850 * constraints.md (ZC, ZD): New constraints.
27851 * config/mips/predicates.md (movep_src_register): New predicate.
27852 (movep_src_operand): New predicate.
27853 (non_volatile_mem_operand): New predicate.
27854 * config/mips/mips.md (multimem): New type.
27855 (length): Differentiate between 17-bit and 18-bit branch offsets.
27856 (MOVEP1, MOVEP2): New mode iterator.
27857 (mov_<load>l): Use ZC constraint.
27858 (mov_<load>r): Likewise.
27859 (mov_<store>l): Likewise.
27860 (mov_<store>r): Likewise.
27861 (*branch_equality<mode>_inverted): Add microMIPS support.
27862 (*branch_equality<mode>): Likewise.
27863 (*jump_absolute): Likewise.
27864 (indirect_jump_<mode>): Likewise.
27865 (tablejump_<mode>): Likewise.
27866 (<optab>_internal): Likewise.
27867 (sibcall_internal): Likewise.
27868 (sibcall_value_internal): Likewise.
27869 (prefetch): Use constraint ZD.
27870 * config/mips/mips.opt (minterlink-compressed): New option.
27871 (minterlink-mips16): Now an alias for minterlink-compressed.
27872 (mmicromips): New option.
27873 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
27874 (compare_and_swap_12): Likewise.
27875 (sync_add<mode>): Likewise.
27876 (sync_<optab>_12): Likewise.
27877 (sync_old_<optab>_12): Likewise.
27878 (sync_new_<optab>_12): Likewise.
27879 (sync_nand_12): Likewise.
27880 (sync_old_nand_12): Likewise.
27881 (sync_new_nand_12): Likewise.
27882 (sync_sub<mode>): Likewise.
27883 (sync_old_add<mode>): Likewise.
27884 (sync_old_sub<mode>): Likewise.
27885 (sync_new_add<mode>): Likewise.
27886 (sync_new_sub<mode>): Likewise.
27887 (sync_<optab><mode>): Likewise.
27888 (sync_old_<optab><mode>): Likewise.
27889 (sync_new_<optab><mode>): Likewise.
27890 (sync_nand<mode>): Likewise.
27891 (sync_old_nand<mode>): Likewise.
27892 (sync_new_nand<mode>): Likewise.
27893 (sync_lock_test_and_set<mode>): Likewise.
27894 (test_and_set_12): Likewise.
27895 (atomic_compare_and_swap<mode>): Likewise.
27896 (atomic_exchange<mode>_llsc): Likewise.
27897 (atomic_fetch_add<mode>_llsc): Likewise.
27898 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
27899 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
27900 (umips_save_restore_pattern_p): Likewise.
27901 (umips_load_store_pair_p): Likewise.
27902 (umips_output_load_store_pair): Likewise.
27903 (umips_movep_target_p): Likewise.
27904 (umips_12bit_offset_address_p): Likewise.
27905 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
27906 (mips_base_mips16): Rename this...
27907 (mips_base_compression_flags): ...to this. Update all uses.
27908 (mips_attribute_table): Add micromips, nomicromips and nocompression.
27909 (mips_mips16_decl_p): Delete.
27910 (mips_nomips16_decl_p): Delete.
27911 (mips_get_compress_on_flags): New function.
27912 (mips_get_compress_off_flags): New function.
27913 (mips_get_compress_mode): New function.
27914 (mips_get_compress_on_name): New function.
27915 (mips_get_compress_off_name): New function.
27916 (mips_insert_attributes): Support multiple compression types.
27917 (mips_merge_decl_attributes): Likewise.
27918 (umips_12bit_offset_address_p): New function.
27919 (mips_start_function_definition): Emit .set micromips directive.
27920 (mips_call_may_need_jalx_p): New function.
27921 (mips_function_ok_for_sibcall): Add microMIPS support.
27922 (mips_print_operand_punctuation): Support short delay slots and
27924 (umips_swm_mask, umips_swm_encoding): New.
27925 (umips_build_save_restore): New function.
27926 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
27927 (was_mips16_p): Remove.
27928 (old_compression_mode): New.
27929 (mips_set_compression_mode): New function.
27930 (mips_set_current_function): Add microMIPS support.
27931 (mips_option_override): Likewise.
27932 (umips_save_restore_pattern_p): New function.
27933 (umips_output_save_restore): New function.
27934 (umips_load_store_pair_p_1): New function.
27935 (umips_load_store_pair_p): New function.
27936 (umips_output_load_store_pair_1): New function.
27937 (umips_output_load_store_pair): New function.
27938 (umips_movep_target_p) New function.
27939 (mips_prepare_pch_save): Add microMIPS support.
27940 * config/mips/mips.h (TARGET_COMPRESSION): New.
27941 (TARGET_CPU_CPP_BUILTINS): Update macro
27942 to use new compression flags and to support microMIPS.
27943 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
27944 (MIPS_ARCH_FLOAT_SPEC): Likewise.
27945 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
27946 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
27947 (ASM_SPEC): Support mmicromips and mno-micromips.
27948 (M16STORE_REG_P): New macro.
27949 (MIPS_CALL): Support TARGET_MICROMIPS.
27950 (MICROMIPS_J): New macro.
27951 (mips_base_mips16): Rename this...
27952 (mips_base_compression_flags): ...to this.
27953 (UMIPS_12BIT_OFFSET_P): New macro.
27954 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
27955 (MULTILIB_DIRNAMES): Likewise.
27956 2013-03-20 Richard Biener <rguenther@suse.de>
27958 PR tree-optimization/56661
27959 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
27960 the result does not have to be distinct.
27962 2013-03-20 Richard Biener <rguenther@suse.de>
27964 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
27967 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
27968 Steven Bosscher <steven@gcc.gnu.org>
27970 PR rtl-optimization/56605
27971 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
27973 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
27976 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
27977 that require movd instead of movq.
27979 2013-03-20 Richard Biener <rguenther@suse.de>
27981 * tree-ssa-structalias.c (struct variable_info): Add pointer
27982 to the first field of an aggregate with sub-vars. Make
27983 this and the pointer to the next subfield its ID.
27984 (vi_next): New function.
27985 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
27986 storedanything_id, integer_id): Increment by one.
27987 (new_var_info, get_call_vi, lookup_call_clobber_vi,
27988 get_call_clobber_vi): Adjust.
27989 (solution_set_expand): Simplify and speedup.
27990 (solution_set_add): Inline into ...
27991 (set_union_with_increment): ... this. Adjust accordingly.
27992 (do_sd_constraint): Likewise.
27993 (do_ds_constraint): Likewise.
27994 (do_complex_constraint): Simplify.
27995 (build_pred_graph): Adjust.
27996 (solve_graph): Likewise. Simplify and speedup.
27997 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
27998 get_constraint_for_component_ref, get_constraint_for_1,
27999 first_vi_for_offset, first_or_preceding_vi_for_offset,
28000 create_function_info_for, create_variable_info_for_1,
28001 create_variable_info_for, intra_create_variable_infos): Adjust.
28002 (init_base_vars): Push NULL for ID zero.
28003 (compute_points_to_sets): Adjust.
28005 2013-03-20 Richard Biener <rguenther@suse.de>
28007 * cfgloop.c (verify_loop_structure): Streamline and avoid
28008 ICEing on corrupt loop tree.
28009 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
28012 2013-03-20 Richard Biener <rguenther@suse.de>
28014 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
28015 check whether an SSA update is needed.
28017 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
28019 * config/mips/constraints.md (T): Rename to...
28023 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
28024 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
28026 2013-03-19 Ian Bolton <ian.bolton@arm.com>
28028 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
28029 (*subsi3_carryin_uxtw): Likewise.
28031 2013-03-19 Ian Bolton <ian.bolton@arm.com>
28033 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
28034 (*rorsi3_insn_uxtw): Likewise.
28036 2013-03-19 Ian Bolton <ian.bolton@arm.com>
28038 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
28039 (*extrsi5_insn_uxtw): Likewise.
28041 2013-03-19 Richard Biener <rguenther@suse.de>
28043 PR tree-optimization/56273
28044 * passes.c (init_optimization_passes): Move second VRP after DOM.
28046 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
28048 * config/i386/i386.md (*movti_internal): Merge from
28049 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
28050 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
28051 nox64 isa attributes.
28053 2013-03-18 Richard Biener <rguenther@suse.de>
28055 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
28057 (merge_node_constraints): Likewise.
28058 (build_succ_graph): Likewise.
28059 (valid_graph_edge): Inline into single caller.
28060 (unify_nodes): Likewise. Use bitmap_set_bit return value
28062 (scc_visit): Fix formatting and variable use.
28063 (do_sd_constraint): Use gcc_checking_assert.
28064 (do_ds_constraint): Likewise.
28065 (do_complex_constraint): Likewise.
28066 (condense_visit): Likewise. Cleanup.
28067 (dump_pred_graph): New function.
28068 (perform_var_substitution): Dump the pred-graph before
28069 variable substitution.
28070 (find_equivalent_node): Use gcc_checking_assert.
28071 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
28073 2013-03-18 Richard Biener <rguenther@suse.de>
28075 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
28076 Remove cond_expr_stmt_list argument and do not gimplify the
28078 (vect_loop_versioning): Adjust.
28079 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
28080 Cleanup to use less temporaries.
28081 (vect_create_data_ref_ptr): Cleanup.
28083 2013-03-18 Jakub Jelinek <jakub@redhat.com>
28085 PR tree-optimization/56635
28086 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
28087 require types_compatible_p types.
28089 2013-03-18 Nick Clifton <nickc@redhat.com>
28091 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
28092 spurious backslash.
28094 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
28095 Add missing line to comment describing function.
28097 2013-03-18 Richard Biener <rguenther@suse.de>
28099 PR tree-optimization/56210
28100 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
28101 Handle string / character search functions.
28102 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
28104 2013-03-18 Richard Biener <rguenther@suse.de>
28106 PR middle-end/56483
28107 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
28108 and implement properly.
28109 * gimple.h (gimple_cond_single_var_p): Remove.
28111 2013-03-18 Richard Biener <rguenther@suse.de>
28113 * tree-data-ref.h (find_data_references_in_loop): Declare.
28114 * tree-data-ref.c (get_references_in_stmt): Use a stack
28115 vector pre-allocated in the callers.
28116 (find_data_references_in_stmt): Adjust.
28117 (graphite_find_data_references_in_stmt): Likewise.
28118 (create_rdg_vertices): Likewise.
28119 (find_data_references_in_loop): Export.
28120 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
28121 Compute dependences here...
28122 (vect_analyze_data_refs): ...not here. When we encounter
28123 a non-vectorizable data reference in basic-block vectorization
28124 truncate the data reference vector. Do not bother to
28125 fixup data-dependence information for gather loads.
28126 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
28127 of data references, as reported.
28129 2013-03-18 Richard Biener <rguenther@suse.de>
28131 PR tree-optimization/3713
28132 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
28133 has_constants and expr.
28134 (stmt_has_constants): Properly valueize SSA names when deciding
28135 whether the stmt has constants.
28137 2013-03-18 Richard Biener <rguenther@suse.de>
28139 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
28140 whole function when there is nothing to do.
28141 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
28142 * tree-vectorizer.c (vectorize_loops): Update virtual and
28143 loop-closed SSA once.
28144 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
28146 2013-03-18 Richard Biener <rguenther@suse.de>
28148 PR middle-end/56113
28149 * domwalk.c (bb_postorder): New global static.
28150 (cmp_bb_postorder): New function.
28151 (walk_dominator_tree): Replace scheme imposing an order for
28152 visiting dominator sons by one sorting them at the time they
28153 are pushed on the stack.
28155 2013-03-18 Richard Biener <rguenther@suse.de>
28157 PR tree-optimization/39326
28158 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
28159 (struct mem_ref): Replace mem member with ao_ref typed member.
28160 (MEM_ANALYZABLE): Adjust.
28161 (memref_eq): Likewise.
28162 (mem_ref_alloc): Likewise.
28163 (gather_mem_refs_stmt): Likewise.
28164 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
28165 (execute_sm_if_changed_flag_set): Adjust.
28166 (execute_sm): Likewise.
28167 (ref_always_accessed_p): Likewise.
28168 (refs_independent_p): Likewise.
28169 (can_sm_ref_p): Likewise.
28171 2013-03-18 Jakub Jelinek <jakub@redhat.com>
28174 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
28175 return 1 even for !unsignedp.
28177 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
28179 * config/i386/i386.md (isa): Add x64 and nox64.
28180 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
28181 (*pushtf): Enable *roF alternative for x64 isa only.
28182 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
28183 mode attribute of integer alternatives to DImode for TARGET_64BIT.
28184 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
28185 (*movtf_internal): Merge from *movtf_internal_rex64 and
28186 *movtf_internal_sse. Use x64 and nox64 isa attributes.
28187 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
28188 nox64 isa attributes.
28189 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
28190 nox64 isa attributes.
28191 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
28193 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
28195 * config/alpha/alpha.c (TARGET_LRA_P): New define.
28197 2013-03-17 Jakub Jelinek <jakub@redhat.com>
28200 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
28201 class names. Remove trailing comma after "ALL_REGS".
28203 2013-03-16 Jan Hubicka <jh@suse.cz>
28205 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
28206 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
28207 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
28208 of cgraph_get_create_node.
28209 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
28211 2013-03-16 Jason Merrill <jason@redhat.com>
28214 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
28215 with DW_AT_default_value.
28217 2013-03-16 Jakub Jelinek <jakub@redhat.com>
28219 * BASE-VER: Set to 4.9.0.
28221 2013-03-14 Andi Kleen <ak@linux.intel.com>
28224 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
28225 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
28226 Document _x* TSX intrinsics.
28228 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
28229 David Holsgrove <david.holsgrove@xilinx.com>
28231 * configure.ac: Add MicroBlaze TLS support detection.
28232 * configure: Regenerate.
28233 * config/microblaze/microblaze-protos.h
28234 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
28235 symbol_mentioned_p, label_mentioned_p): Add prototypes.
28236 * config/microblaze/microblaze.c (microblaze_address_type): Add
28237 ADDRESS_TLS and tls_reloc address types.
28238 (microblaze_address_info): Add tls_reloc.
28239 (TARGET_HAVE_TLS): Define.
28240 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
28241 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
28242 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
28243 load_tls_operand, microblaze_call_tls_get_addr,
28244 microblaze_legitimize_tls_address): New functions.
28245 (microblaze_classify_unspec): Handle UNSPEC_TLS.
28246 (get_base_reg): Use microblaze_tls_symbol_p.
28247 (microblaze_classify_address): Handle TLS.
28248 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
28249 label_mentioned_p and microblaze_tls_referenced_p.
28250 (microblaze_legitimize_address): Handle TLS.
28251 (microblaze_address_insns): Handle ADDRESS_TLS.
28252 (pic_address_needs_scratch): Handle TLS.
28253 (print_operand_address): Handle TLS.
28254 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
28255 (microblaze_expand_move): Handle TLS.
28256 (microblaze_legitimate_constant_p): Check
28257 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
28258 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
28259 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
28260 (PIC_OFFSET_TABLE_REGNUM): Set.
28261 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
28262 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
28263 (addsi3, movsi_internal2, movdf_internal): Update constraints
28264 * config/microblaze/predicates.md (arith_plus_operand): Define
28265 (move_operand): Redefine as move_src_operand,
28266 check microblaze_tls_referenced_p.
28268 2013-03-14 Ian Bolton <ian.bolton@arm.com>
28270 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
28271 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
28273 2013-03-14 Ian Bolton <ian.bolton@arm.com>
28275 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
28278 2013-03-14 Jakub Jelinek <jakub@redhat.com>
28280 PR tree-optimization/53265
28281 * common.opt (Waggressive-loop-optimizations): New option.
28282 * tree-ssa-loop-niter.c: Include tree-pass.h.
28283 (do_warn_aggressive_loop_optimizations): New function.
28284 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
28285 if number_of_latch_executions returned constant.
28286 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
28287 early. If number_of_latch_executions returned constant, set
28288 nb_iterations_upper_bound back to it.
28289 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
28291 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
28292 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
28294 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
28295 (MULTILIB_OSDIRNAMES): Set.
28296 * genmultilib: If defaultosdirname doesn't start with :: , set
28297 defaultosdirname2 instead, clear it and emit two . multilib_raw
28298 entries instead of just one.
28300 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
28302 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
28303 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
28304 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
28305 (SUBTARGET_OVERRIDE_OPTIONS): New.
28307 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
28310 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
28311 (musermode): Convert to Var(TARGET_USERMODE).
28312 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
28313 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
28314 * config/sh/sh.c (sh_option_override): Use
28315 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
28316 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
28318 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
28320 (udivsi3_i4_single, divsi3_i4_single): Use
28321 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
28323 2013-03-13 Dave Korn <dave.korn.cygwin@gmail.com>
28325 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
28328 2013-03-13 Richard Biener <rguenther@suse.de>
28330 PR tree-optimization/56608
28331 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
28332 calls when vectorizing basic-blocks.
28334 2013-03-13 Jakub Jelinek <jakub@redhat.com>
28337 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
28340 2013-03-12 Jakub Jelinek <jakub@redhat.com>
28342 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
28344 2013-03-11 Jan Hubicka <jh@suse.cz>
28347 * lto-streamer-out.c (output_symbol_p): Skip references from
28348 constructors of external variables.
28350 2013-03-11 Jan Hubicka <jh@suse.cz>
28352 PR middle-end/56571
28353 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
28355 * emit-rtl.c (verify_rtx_sharing): Likewise.
28356 (copy_insn_1): Likewise.
28357 * rtl.c (copy_rtx): Likewise.
28359 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
28362 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
28363 output_operand_lossage message.
28365 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
28368 * arm.c (shift_op): Validate RTL pattern on the fly.
28369 (arm_print_operand, case 'S'): Don't use shift_operator to validate
28372 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
28375 * config/pa/pa.md (call_value): Check for calls to powf and direct to
28376 new call patterns that clobber %fr12.
28377 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
28378 split and postreload patterns.
28379 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
28380 registers %fr12 and %fr12R as call used.
28382 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
28384 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
28385 (canon_address, record_store, replace_read, check_mem_read_rtx,
28386 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
28387 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
28388 rest_of_handle_dse): Likewise.
28390 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
28392 PR middle-end/56524
28393 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
28395 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
28396 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
28397 (save_optabs_if_changed): Replace with...
28398 (init_tree_optimization_optabs): ...this.
28399 * optabs.c (save_optabs_if_changed): Rename to...
28400 (init_tree_optimization_optabs): ...this. Take the optimization node
28401 as argument. Do nothing if the base optabs are already correct.
28402 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
28403 to recompute optabs.
28404 * function.h (function): Remove optabs field.
28405 * function.c (invoke_set_current_function_hook): Call
28406 init_tree_optimization_optabs. Use the result to initialize
28409 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
28411 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
28412 if GTMA_HAS_NO_INSTRUMENTATION.
28413 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
28414 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
28415 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
28416 * gimple-pretty-print.c (dump_gimple_transaction): Handle
28417 GTMA_HAS_NO_INSTRUMENTATION.
28419 2013-03-08 Jakub Jelinek <jakub@redhat.com>
28421 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
28424 2013-03-08 Marek Polacek <polacek@redhat.com>
28425 Jakub Jelinek <jakub@redhat.com>
28427 PR tree-optimization/56478
28428 * predict.c (is_comparison_with_loop_invariant_p): Change the
28429 type of loop_step to tree.
28430 (predict_loops): Adjust.
28431 (predict_iv_comparison): Perform the computations on double_ints.
28433 2013-03-08 Richard Biener <rguenther@suse.de>
28435 PR tree-optimization/56570
28436 * tree-cfg.c (verify_expr_location_1): Verify locations for
28438 * tree-sra.c (create_access_replacement): Strip locations
28439 from DECL_DEBUG_EXPRs.
28441 2013-03-08 Richard Biener <rguenther@suse.de>
28443 * tree-inline.c (expand_call_inline): Do not associate
28444 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
28445 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
28447 2013-03-08 Richard Biener <rguenther@suse.de>
28449 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
28450 or block changes with -Og. Fix for location / block encoding
28451 changes and PHI arguments with locations.
28453 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
28455 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
28457 (struct output_info): Likewise.
28458 (register_overhead): Remove bad gcc_assert.
28459 (bitmap_find_bit): If there is only a single bitmap element, do not
28460 count a miss as a search.
28461 (print_statistics): Update for counter type changes.
28462 (dump_bitmap_statistics): Likewise. Print headers such that they
28463 are properly lined up with the printed counters.
28465 2013-03-07 Jakub Jelinek <jakub@redhat.com>
28467 PR tree-optimization/56559
28468 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
28469 check that it has only a single use.
28471 2013-03-07 Richard Biener <rguenther@suse.de>
28473 * doc/invoke.texi (fwhole-program): Discourage use in combination
28476 2013-03-06 Jakub Jelinek <jakub@redhat.com>
28478 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
28480 PR tree-optimization/56539
28481 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
28482 instead of GSI_CONTINUE_LINKING as last argument to
28483 force_gimple_operand_gsi. Adjust function comment.
28485 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
28488 PR middle-end/56548
28489 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
28490 promoted mode, convert the result back to the original mode.
28492 2013-03-06 Richard Biener <rguenther@suse.de>
28494 PR middle-end/56294
28495 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
28496 (insert_updated_phi_nodes_compare_uids): New function.
28497 (update_ssa): Sort symbols_to_rename after UID before
28498 traversing it to insert PHI nodes.
28500 2013-03-06 Richard Biener <rguenther@suse.de>
28502 PR middle-end/50494
28503 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
28504 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
28507 2013-02-13 Richard Biener <rguenther@suse.de>
28510 * varasm.c (output_constant_def_1): Get the decl representing
28511 the constant as argument.
28512 (output_constant_def): Wrap output_constant_def_1.
28513 (make_decl_rtl): Use output_constant_def_1 with the decl
28514 representing the constant.
28515 (build_constant_desc): Optionally re-use a decl already
28516 representing the constant.
28517 (tree_output_constant_def): Adjust.
28519 2013-03-06 Joey Ye <joey.ye@arm.com>
28522 * gcc.c (convert_white_space): New function.
28523 (main): Handles white space in function name.
28525 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
28528 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
28529 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
28530 to SH_DIV_CALL_TABLE for TARGET_SH2.
28531 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
28533 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
28534 call-table options.
28536 2013-03-05 Sterling Augustine <saugustine@google.com>
28537 Cary Coutant <ccoutant@google.com>
28540 * dwarf2out.c (resolve_addr): Don't call
28541 remove_loc_list_addr_table_entries a second time for the same
28544 2013-03-05 Jakub Jelinek <jakub@redhat.com>
28547 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
28548 (avoid_complex_debug_insns): New function.
28549 (expand_debug_locations): Call it.
28551 PR rtl-optimization/56484
28552 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
28553 lifetimes of hard registers on small register class machines.
28555 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
28557 * config/microblaze/microblaze-protos.h: Rename
28558 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
28559 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
28561 (microblaze_fast_interrupt_function_p): New function.
28562 (microblaze_is_interrupt_handler): Rename to
28563 microblaze_is_interrupt_variant and add fast_interrupt check.
28564 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
28565 (save_restore_insns): Likewise.
28566 (compute_frame_size): Likewise.
28567 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
28568 (microblaze_globalize_label): Likewise.
28569 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
28570 * config/microblaze/microblaze.md: Use wrapper
28571 microblaze_is_interrupt_variant.
28573 2013-03-05 Kai Tietz <ktietz@redhat.com>
28575 * sdbout.c (sdbout_one_type): Switch to current function's section
28576 supporting cold/hot.
28578 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
28580 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
28583 2013-03-05 Jakub Jelinek <jakub@redhat.com>
28585 PR middle-end/56461
28586 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
28587 if VALGRIND_GET_VBITS is defined, temporarily make object
28588 memory all defined, and restore previous valgrind addressability
28589 and definability afterwards. Free this_object at the end.
28591 PR middle-end/56461
28592 * lra.c (lra): Call lra_clear_live_ranges if live_p,
28593 right before calling lra_create_live_ranges, also call it
28594 when clearing live_p. Only call lra_clear_live_ranges
28595 at the end if live_p.
28597 PR middle-end/56461
28598 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
28600 2013-03-05 Richard Biener <rguenther@suse.de>
28602 PR tree-optimization/56521
28603 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
28606 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
28609 * except.h (remove_unreachable_eh_regions): New prototype.
28610 * except.c (remove_eh_handler_splicer): New function, split out
28611 of remove_eh_handler.
28612 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
28613 warning about running it on many EH regions one at a time.
28614 (remove_unreachable_eh_regions_worker): New function, walk the
28615 EH tree in depth-first order and remove non-marked regions.
28616 (remove_unreachable_eh_regions): New function.
28617 * tree-eh.c (mark_reachable_handlers): New function, split out
28618 from remove_unreachable_handlers.
28619 (remove_unreachable_handlers): Use mark_reachable_handlers and
28620 remove_unreachable_eh_regions.
28621 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
28622 and remove_unreachable_eh_regions.
28624 2013-03-05 Richard Biener <rguenther@suse.de>
28626 PR middle-end/56525
28627 * loop-init.c (fix_loop_structure): Remove loops in two stages,
28628 not freeing them until the end.
28630 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
28632 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
28634 2013-03-05 Richard Biener <rguenther@suse.de>
28636 PR tree-optimization/56270
28637 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
28638 of loads after scheduling an SLP instance.
28640 2013-03-05 Jakub Jelinek <jakub@redhat.com>
28642 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
28644 (check_gcc_parallelize): Run guality.exp as a separate job from
28645 vect.exp with unsorted.exp and $(dg_target_exps) separately from
28646 struct-layout-1.exp with stackalign.exp.
28648 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
28650 PR middle-end/56461
28651 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
28652 load_index sbitmap even if some bit in it isn't set.
28654 PR middle-end/56461
28655 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
28656 (discover_iteration_bound_by_body_walk): Change queues to
28657 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
28658 spelling in comment. Call safe_push on queues[bound_index] directly.
28659 Release queues[queue_index] in every iteration unconditionally.
28660 Release bounds vector.
28662 PR middle-end/56461
28663 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
28664 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
28665 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
28668 2013-03-05 Richard Biener <rguenther@suse.de>
28671 * tree-inline.c (remap_blocks_to_null): New function.
28672 (expand_call_inline): When expanding a call stmt without
28673 an associated block inline remap all callee blocks to NULL.
28675 2013-03-05 Jakub Jelinek <jakub@redhat.com>
28677 PR rtl-optimization/56494
28678 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
28679 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
28680 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
28682 PR middle-end/56461
28683 * sel-sched-ir.c (free_sched_pools): Release
28684 succs_info_pool.stack[succs_info_pool.max_top] vectors too
28685 if succs_info_pool.max_top isn't -1.
28688 * opts.c (opts_obstack, opts_concat): Moved to...
28689 * opts-common.c (opts_obstack, opts_concat): ... here.
28691 2013-03-04 Jakub Jelinek <jakub@redhat.com>
28693 PR middle-end/56461
28694 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
28696 2013-03-04 Martin Jambor <mjambor@suse.cz>
28698 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
28699 all appropriate places.
28701 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
28703 PR tree-optimization/56424
28704 * ipa-split.c (split_function): Do not set the RSO flag if result is
28705 not by reference and its type is a register type.
28707 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
28709 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
28710 (microblaze_legitimate_pic_operand): Likewise
28711 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
28712 new function microblaze_legitimate_pic_operand
28713 * config/microblaze/microblaze-protos.h
28714 (microblaze_legitimate_pic_operand): Declare.
28716 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28718 * config/microblaze/predicates.md (call_insn_simple_operand):
28719 New predicate for supported rtx code types.
28720 * config/microblaze/microblaze.md (call_internal1): Use
28721 call_insn_simple_operand predicate.
28723 2013-03-04 Jakub Jelinek <jakub@redhat.com>
28725 PR middle-end/56461
28726 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
28727 partitions.ordered_remove.
28729 PR middle-end/56461
28730 * tree-vect-stmts.c (vectorizable_conversion): Don't call
28731 vec_oprnds0.create (1) for modifier == NONE.
28733 PR middle-end/56461
28734 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
28735 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
28736 vec_oprnds1 right before pushing anything to it for
28739 PR middle-end/56461
28740 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
28741 set nbbs to 0 instead of having separate code path.
28742 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
28743 instead of false as last argument if returning NULL.
28745 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
28747 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
28748 the attribute is now called "target" instead of "option".
28749 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
28750 * doc/tm.texi.in (Target Attributes): Likewise document the correct
28751 attribute/pragma name for TARGET_OPTION_VALID_P and
28752 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
28753 * doc/tm.texi: Regenerated.
28755 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
28757 * config/microblaze/microblaze.c:
28758 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
28759 * config/microblaze/microblaze.h: Add -mxl-reorder to
28761 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
28762 instructions emitted if TARGET_REORDER.
28763 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
28764 or 0 for -m/-mno case, but initialises as 2 to detect default use case
28767 2013-03-01 Xinliang David Li <davidxl@google.com>
28769 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
28772 2013-03-01 Jakub Jelinek <jakub@redhat.com>
28774 PR middle-end/56461
28775 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
28776 vector even when returning true. Fix up function comment formatting.
28778 PR middle-end/56461
28779 * ira-build.c (ira_loop_nodes_count): New variable.
28780 (create_loop_tree_nodes): Initialize it.
28781 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
28783 PR middle-end/56461
28784 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
28785 method on dr_chain and result_chain.
28786 * tree-vect-stmts.c (vectorizable_store): Only call
28787 result_chain.create if j == 0.
28789 PR middle-end/56461
28790 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
28791 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
28792 before overwriting it.
28794 2013-03-01 Tobias Burnus <burnus@net-b.de>
28796 * doc/extended.texi (C Extensions): Change order in @menu
28798 (Other MIPS Built-in Functions): Move last MIPS entry before
28799 "picoChip Built-in Functions".
28800 (SH Built-in Functions): Move after RX Built-in Functions.
28801 * doc/gcc.texi (Introduction): Change order in @menu
28803 * doc/md.texi (Constraints): Ditto.
28804 * gty.texi (Type Information): Ditto.
28805 (User-provided marking routines for template types): Make
28807 * doc/invoke.texi (AArch64 Options): Move before
28808 "Adapteva Epiphany Options".
28810 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
28811 Jakub Jelinek <jakub@redhat.com>
28814 * asan.c (gate_asan): Lookup no_sanitize_address instead of
28815 no_address_safety_analysis attribute.
28816 * doc/extend.texi (no_address_safety_attribute): Rename to
28817 no_sanitize_address attribute, mention no_address_safety_analysis
28818 attribute as deprecated alias.
28820 2013-02-28 Jakub Jelinek <jakub@redhat.com>
28822 PR middle-end/56461
28823 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
28824 type to vec<vec<tree> > *.
28825 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
28826 to be vec<tree> instead of vec<tree> *, set vec_defs
28827 to vNULL and call vec_defs.create (number_of_vects), adjust other
28829 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
28830 vectorizable_condition): Adjust vect_get_slp_defs callers.
28832 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
28834 * config/aarch64/aarch64.c
28835 (aarch64_float_const_representable): Remove unused variable.
28837 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
28839 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
28841 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
28843 * config/aarch64/aarch64-builtins.c
28844 (aarch64_init_simd_builtins): Make static.
28846 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
28848 * config/aarch64/aarch64.c
28849 (aarch64_simd_make_constant): Make static.
28851 2013-02-28 Martin Jambor <mjambor@suse.cz>
28853 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
28854 with no initialization to the RHS of debug statements.
28856 2013-02-28 Martin Jambor <mjambor@suse.cz>
28858 PR tree-optimization/56294
28859 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
28861 (get_access_replacement): Do not call create_access_replacement.
28862 Assert a replacement exists.
28863 (get_repl_default_def_ssa_name): Create the replacement declaration
28866 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
28868 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
28869 final_end_function.
28871 2013-02-28 Marek Polacek <polacek@redhat.com>
28873 PR rtl-optimization/56466
28874 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
28875 if we're changing a loop.
28876 (peel_loops_completely): Likewise.
28878 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
28881 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
28883 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
28886 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
28887 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
28888 INTX_FTYPE_FX, FX_FTYPE_INTX.
28889 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
28891 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
28893 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
28894 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
28895 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
28896 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
28897 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
28898 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
28899 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
28900 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
28901 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
28902 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
28903 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
28904 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
28905 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
28906 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
28907 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
28908 (avrxmega6): Increase max flash segments from 5 to 6.
28909 * config/avr/t-multilib: Regenerate.
28910 * config/avr/avr-tables.opt: Regenerate.
28911 * doc/avr-mmcu.texi: Regenerate.
28913 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
28915 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
28916 (avr_device_to_arch): Rename to avr_device_to_ld.
28917 (avr_device_to_as): New prototype.
28918 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
28919 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
28920 * config/avr/driver-avr.c (avr_device_to_as): New.
28921 (avr_device_to_arch): Rename to avr_device_to_ld.
28923 2013-02-27 Jakub Jelinek <jakub@redhat.com>
28925 PR middle-end/56461
28926 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
28927 method on dr_chain and result_chain.
28929 PR middle-end/56461
28930 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
28931 pointer_set_destroy on not_executed_last_iteration.
28933 PR middle-end/56461
28934 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
28936 PR middle-end/56461
28937 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
28938 FOR_EACH_DEFINED_FUNCTION when freeing state.
28940 PR middle-end/56461
28941 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
28943 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
28946 PR middle-end/56461
28947 * ipa-cp.c (decide_whether_version_node): Call vec_free on
28948 known_aggs[i].items and release known_aggs vector.
28950 PR middle-end/56461
28951 * ipa-reference.c (propagate): Free node_info even for alias nodes.
28953 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28955 * config/microblaze/microblaze.c (microblaze_emit_compare):
28956 Use xor for EQ/NE comparisions.
28957 * config/microblaze/microblaze.md (cstoresf4): Add constraints
28958 (cbranchsf4): Adjust operator to comparison_operator.
28960 2013-02-27 Jakub Jelinek <jakub@redhat.com>
28962 PR middle-end/56461
28963 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
28965 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
28966 vec_safe_push, always update *slot.
28967 (redirect_edge_var_map_clear): Use vec_free.
28968 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
28969 (free_var_map_entry): Use vec_free.
28970 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
28971 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
28973 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
28975 PR middle-end/45472
28976 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
28977 when the may_trap_p bit of the exprs being merged differs.
28978 Reorder tests for speculativeness in the logical and operator.
28980 2013-02-27 Jakub Jelinek <jakub@redhat.com>
28982 * incpath.c (add_standard_paths): Use reconcat instead of concat
28983 where appropriate and avoid leaking memory.
28985 * opts.h: Include obstack.h.
28986 (opts_concat): New prototype.
28987 (opts_obstack): New declaration.
28988 * opts.c (opts_concat): New function.
28989 (opts_obstack): New variable.
28990 (init_options_struct): Call gcc_init_obstack on opts_obstack.
28991 (finish_options): Use opts_concat instead of concat
28992 and XOBNEWVEC instead of XNEWVEC.
28993 * opts-common.c (generate_canonical_option, decode_cmdline_option,
28994 generate_option): Likewise.
28995 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
28996 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
28999 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
29000 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
29002 2013-02-26 Jakub Jelinek <jakub@redhat.com>
29004 PR middle-end/56461
29005 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
29007 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
29009 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
29010 (arm_block_move_unaligned_straight): Likewise.
29011 (arm_adjust_block_mem): Likewise.
29013 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
29016 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
29017 temp, cond and label.
29018 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
29021 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
29022 * config/c6x/c6x.h (dbx_register_map): Update declaration.
29025 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
29026 of prologue/epilogue functions.
29029 * config/tilegx/tilegx.c (tilegx_expand_prologue):
29030 Remove unused variable cfa_offset.
29031 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
29034 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
29035 type promotion to unsigned.
29038 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
29039 for HOST_WIDE_INT of 32 bit / same size as int.
29040 (arm_block_move_unaligned_straight): Likewise.
29041 (arm_adjust_block_mem): Likewise.
29044 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
29047 2013-02-26 Marek Polacek <polacek@redhat.com>
29049 PR tree-optimization/56426
29050 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
29052 2013-02-26 Richard Biener <rguenther@suse.de>
29055 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
29056 unused variable loops.
29058 2013-02-26 Jakub Jelinek <jakub@redhat.com>
29060 PR tree-optimization/56448
29061 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
29062 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
29063 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
29064 later operands of the references, or even first operand for
29065 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
29067 PR tree-optimization/56443
29068 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
29069 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
29070 to type_for_mode langhook.
29072 2013-02-25 Matt Turner <mattst88@gmail.com>
29074 * doc/invoke.texi: Document r4700.
29076 2013-02-25 Richard Biener <rguenther@suse.de>
29078 PR tree-optimization/56175
29079 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
29081 (simplify_bitwise_binary): ... here. Also guard the conversion
29082 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
29084 2013-02-25 Catherine Moore <clm@codesourcery.com>
29087 2013-02-24 Catherine Moore <clm@codesourcery.com>
29088 Maciej W. Rozycki <macro@codesourcery.com>
29089 Tom de Vries <tom@codesourcery.com>
29090 Nathan Sidwell <nathan@codesourcery.com>
29091 Iain Sandoe <iain@codesourcery.com>
29092 Nathan Froyd <froydnj@codesourcery.com>
29093 Chao-ying Fu <fu@mips.com>
29095 * doc/extend.texi: (micromips, nomicromips, nocompression):
29096 Document new function attributes.
29097 * doc/invoke.texi (minterlink-compressed, mmicromips,
29098 m14k, m14ke, m14kec): Document new options.
29099 (minterlink-mips16): Update documentation.
29100 * doc/md.texi (ZC, ZD): Document new constraints.
29101 * configure.ac (gcc_cv_as_micromips): Check if linker
29102 supports the .set micromips directive.
29103 * configure: Regenerate.
29104 * config.in: Regenerate.
29105 * config/mips/mips-tables.opt: Regenerate.
29106 * config/mips/micromips.md: New file.
29107 * constraints.md (ZC, AD): New constraints.
29108 * config/mips/predicates.md (movep_src_register): New predicate.
29109 (movep_src_operand): New predicate.
29110 (non_volatile_mem_operand): New predicate.
29111 * config/mips/mips.md (multimem): New type.
29112 (length): Differentiate between 17-bit and 18-bit branch offsets.
29113 (MOVEP1, MOVEP2): New mode iterator.
29114 (mov_<load>l): Use ZC constraint.
29115 (mov_<load>r): Likewise.
29116 (mov_<store>l): Likewise.
29117 (mov_<store>r): Likewise.
29118 (*branch_equality<mode>_inverted): Add microMIPS support.
29119 (*branch_equality<mode>): Likewise.
29120 (*jump_absolute): Likewise.
29121 (indirect_jump_<mode>): Likewise.
29122 (tablejump_<mode>): Likewise.
29123 (<optab>_internal): Likewise.
29124 (sibcall_internal): Likewise.
29125 (sibcall_value_internal): Likewise.
29126 (prefetch): Use constraint ZD.
29127 * config/mips/mips.opt (minterlink-compressed): New option.
29128 (minterlink-mips16): Now an alias for minterlink-compressed.
29129 (mmicromips): New option.
29130 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
29131 (compare_and_swap_12): Likewise.
29132 (sync_add<mode>): Likewise.
29133 (sync_<optab>_12): Likewise.
29134 (sync_old_<optab>_12): Likewise.
29135 (sync_new_<optab>_12): Likewise.
29136 (sync_nand_12): Likewise.
29137 (sync_old_nand_12): Likewise.
29138 (sync_new_nand_12): Likewise.
29139 (sync_sub<mode>): Likewise.
29140 (sync_old_add<mode>): Likewise.
29141 (sync_old_sub<mode>): Likewise.
29142 (sync_new_add<mode>): Likewise.
29143 (sync_new_sub<mode>): Likewise.
29144 (sync_<optab><mode>): Likewise.
29145 (sync_old_<optab><mode>): Likewise.
29146 (sync_new_<optab><mode>): Likewise.
29147 (sync_nand<mode>): Likewise.
29148 (sync_old_nand<mode>): Likewise.
29149 (sync_new_nand<mode>): Likewise.
29150 (sync_lock_test_and_set<mode>): Likewise.
29151 (test_and_set_12): Likewise.
29152 (atomic_compare_and_swap<mode>): Likewise.
29153 (atomic_exchange<mode>_llsc): Likewise.
29154 (atomic_fetch_add<mode>_llsc): Likewise.
29155 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
29156 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
29157 (umips_save_restore_pattern_p): Likewise.
29158 (umips_load_store_pair_p): Likewise.
29159 (umips_output_load_store_pair): Likewise.
29160 (umips_movep_target_p): Likewise.
29161 (umips_12bit_offset_address_p): Likewise.
29162 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
29163 (mips_base_mips16): Rename this...
29164 (mips_base_compression_flags): ...to this. Update all uses.
29165 (mips_attribute_table): Add micromips, nomicromips and nocompression.
29166 (mips_mips16_decl_p): Delete.
29167 (mips_nomips16_decl_p): Delete.
29168 (mips_get_compress_on_flags): New function.
29169 (mips_get_compress_off_flags): New function.
29170 (mips_get_compress_mode): New function.
29171 (mips_get_compress_on_name): New function.
29172 (mips_get_compress_off_name): New function.
29173 (mips_insert_attributes): Support multiple compression types.
29174 (mips_merge_decl_attributes): Likewise.
29175 (umips_12bit_offset_address_p): New function.
29176 (mips_start_function_definition): Emit .set micromips directive.
29177 (mips_call_may_need_jalx_p): New function.
29178 (mips_function_ok_for_sibcall): Add microMIPS support.
29179 (mips_print_operand_punctuation): Support short delay slots and
29181 (umips_swm_mask, umips_swm_encoding): New.
29182 (umips_build_save_restore): New function.
29183 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
29184 (was_mips16_p): Remove.
29185 (old_compression_mode): New.
29186 (mips_set_compression_mode): New function.
29187 (mips_set_current_function): Add microMIPS support.
29188 (mips_option_override): Likewise.
29189 (umips_save_restore_pattern_p): New function.
29190 (umips_output_save_restore): New function.
29191 (umips_load_store_pair_p_1): New function.
29192 (umips_load_store_pair_p): New function.
29193 (umips_output_load_store_pair_1): New function.
29194 (umips_output_load_store_pair): New function.
29195 (umips_movep_target_p) New function.
29196 (mips_prepare_pch_save): Add microMIPS support.
29197 * config/mips/mips.h (TARGET_COMPRESSION): New.
29198 (TARGET_CPU_CPP_BUILTINS): Update macro
29199 to use new compression flags and to support microMIPS.
29200 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
29201 (MIPS_ARCH_FLOAT_SPEC): Likewise.
29202 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
29203 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
29204 (ASM_SPEC): Support mmicromips and mno-micromips.
29205 (M16STORE_REG_P): New macro.
29206 (MIPS_CALL): Support TARGET_MICROMIPS.
29207 (MICROMIPS_J): New macro.
29208 (mips_base_mips16): Rename this...
29209 (mips_base_compression_flags): ...to this.
29210 (UMIPS_12BIT_OFFSET_P): New macro.
29211 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
29212 (MULTILIB_DIRNAMES): Likewise.
29214 2013-02-25 Tom de Vries <tom@codesourcery.com>
29216 PR rtl-optimization/56131
29217 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
29218 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
29219 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
29221 2013-02-25 Tobias Burnus <burnus@net-b.de>
29223 * doc/invoke.texi (-fsanitize=): Move from optimization
29224 to debugging options.
29226 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
29228 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
29230 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
29231 Alexander Monakov <amonakov@ispras.ru>
29233 PR middle-end/56077
29234 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
29235 flush pending lists also on non-jumps. Adjust comment.
29237 2013-02-24 Catherine Moore <clm@codesourcery.com>
29238 Maciej W. Rozycki <macro@codesourcery.com>
29239 Tom de Vries <tom@codesourcery.com>
29240 Nathan Sidwell <nathan@codesourcery.com>
29241 Iain Sandoe <iain@codesourcery.com>
29242 Nathan Froyd <froydnj@codesourcery.com>
29243 Chao-ying Fu <fu@mips.com>
29245 * doc/extend.texi: (micromips, nomicromips, nocompression):
29246 Document new function attributes.
29247 * doc/invoke.texi (minterlink-compressed, mmicromips,
29248 m14k, m14ke, m14kec): Document new options.
29249 (minterlink-mips16): Update documentation.
29250 * doc/md.texi (ZC, ZD): Document new constraints.
29251 * configure.ac (gcc_cv_as_micromips): Check if linker
29252 supports the .set micromips directive.
29253 * configure: Regenerate.
29254 * config.in: Regenerate.
29255 * config/mips/mips-tables.opt: Regenerate.
29256 * config/mips/micromips.md: New file.
29257 * constraints.md (ZC, AD): New constraints.
29258 * config/mips/predicates.md (movep_src_register): New predicate.
29259 (movep_src_operand): New predicate.
29260 (non_volatile_mem_operand): New predicate.
29261 * config/mips/mips.md (multimem): New type.
29262 (length): Differentiate between 17-bit and 18-bit branch offsets.
29263 (MOVEP1, MOVEP2): New mode iterator.
29264 (mov_<load>l): Use ZC constraint.
29265 (mov_<load>r): Likewise.
29266 (mov_<store>l): Likewise.
29267 (mov_<store>r): Likewise.
29268 (*branch_equality<mode>_inverted): Add microMIPS support.
29269 (*branch_equality<mode>): Likewise.
29270 (*jump_absolute): Likewise.
29271 (indirect_jump_<mode>): Likewise.
29272 (tablejump_<mode>): Likewise.
29273 (<optab>_internal): Likewise.
29274 (sibcall_internal): Likewise.
29275 (sibcall_value_internal): Likewise.
29276 (prefetch): Use constraint ZD.
29277 * config/mips/mips.opt (minterlink-compressed): New option.
29278 (minterlink-mips16): Now an alias for minterlink-compressed.
29279 (mmicromips): New option.
29280 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
29281 (compare_and_swap_12): Likewise.
29282 (sync_add<mode>): Likewise.
29283 (sync_<optab>_12): Likewise.
29284 (sync_old_<optab>_12): Likewise.
29285 (sync_new_<optab>_12): Likewise.
29286 (sync_nand_12): Likewise.
29287 (sync_old_nand_12): Likewise.
29288 (sync_new_nand_12): Likewise.
29289 (sync_sub<mode>): Likewise.
29290 (sync_old_add<mode>): Likewise.
29291 (sync_old_sub<mode>): Likewise.
29292 (sync_new_add<mode>): Likewise.
29293 (sync_new_sub<mode>): Likewise.
29294 (sync_<optab><mode>): Likewise.
29295 (sync_old_<optab><mode>): Likewise.
29296 (sync_new_<optab><mode>): Likewise.
29297 (sync_nand<mode>): Likewise.
29298 (sync_old_nand<mode>): Likewise.
29299 (sync_new_nand<mode>): Likewise.
29300 (sync_lock_test_and_set<mode>): Likewise.
29301 (test_and_set_12): Likewise.
29302 (atomic_compare_and_swap<mode>): Likewise.
29303 (atomic_exchange<mode>_llsc): Likewise.
29304 (atomic_fetch_add<mode>_llsc): Likewise.
29305 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
29306 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
29307 (umips_save_restore_pattern_p): Likewise.
29308 (umips_load_store_pair_p): Likewise.
29309 (umips_output_load_store_pair): Likewise.
29310 (umips_movep_target_p): Likewise.
29311 (umips_12bit_offset_address_p): Likewise.
29312 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
29313 (mips_base_mips16): Rename this...
29314 (mips_base_compression_flags): ...to this. Update all uses.
29315 (mips_attribute_table): Add micromips, nomicromips and nocompression.
29316 (mips_mips16_decl_p): Delete.
29317 (mips_nomips16_decl_p): Delete.
29318 (mips_get_compress_on_flags): New function.
29319 (mips_get_compress_off_flags): New function.
29320 (mips_get_compress_mode): New function.
29321 (mips_get_compress_on_name): New function.
29322 (mips_get_compress_off_name): New function.
29323 (mips_insert_attributes): Support multiple compression types.
29324 (mips_merge_decl_attributes): Likewise.
29325 (umips_12bit_offset_address_p): New function.
29326 (mips_start_function_definition): Emit .set micromips directive.
29327 (mips_call_may_need_jalx_p): New function.
29328 (mips_function_ok_for_sibcall): Add microMIPS support.
29329 (mips_print_operand_punctuation): Support short delay slots and
29331 (umips_swm_mask, umips_swm_encoding): New.
29332 (umips_build_save_restore): New function.
29333 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
29334 (was_mips16_p): Remove.
29335 (old_compression_mode): New.
29336 (mips_set_compression_mode): New function.
29337 (mips_set_current_function): Add microMIPS support.
29338 (mips_option_override): Likewise.
29339 (umips_save_restore_pattern_p): New function.
29340 (umips_output_save_restore): New function.
29341 (umips_load_store_pair_p_1): New function.
29342 (umips_load_store_pair_p): New function.
29343 (umips_output_load_store_pair_1): New function.
29344 (umips_output_load_store_pair): New function.
29345 (umips_movep_target_p) New function.
29346 (mips_prepare_pch_save): Add microMIPS support.
29347 * config/mips/mips.h (TARGET_COMPRESSION): New.
29348 (TARGET_CPU_CPP_BUILTINS): Update macro
29349 to use new compression flags and to support microMIPS.
29350 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
29351 (MIPS_ARCH_FLOAT_SPEC): Likewise.
29352 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
29353 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
29354 (ASM_SPEC): Support mmicromips and mno-micromips.
29355 (M16STORE_REG_P): New macro.
29356 (MIPS_CALL): Support TARGET_MICROMIPS.
29357 (MICROMIPS_J): New macro.
29358 (mips_base_mips16): Rename this...
29359 (mips_base_compression_flags): ...to this.
29360 (UMIPS_12BIT_OFFSET_P): New macro.
29361 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
29362 (MULTILIB_DIRNAMES): Likewise.
29364 2013-02-24 Jakub Jelinek <jakub@redhat.com>
29367 * target-globals.c (save_target_globals): For init_reg_sets and
29368 target_reinit remporarily set this_fn_optabs to this_target_optabs.
29370 2013-02-22 James Greenhalgh <james.greenhalgh@arm.com>
29372 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
29373 * config/aarch64/t-aarch64
29374 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
29376 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
29378 PR inline-asm/56148
29379 * lra-constraints.c (process_alt_operands): Reload operand
29380 conflicting with earlier clobber only if no more other conflicting
29383 2013-02-22 Jakub Jelinek <jakub@redhat.com>
29386 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
29387 if not linking a shared library.
29389 2013-02-22 Seth LaForge <sethml@google.com>
29391 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
29393 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
29395 * config/arm/arm.md (split for extendsidi): Update condition.
29396 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
29397 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
29398 (qhs_zextenddi_cstr): Likewise.
29400 2013-02-21 Jakub Jelinek <jakub@redhat.com>
29402 PR middle-end/56420
29403 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
29404 avoid signed wrapping.
29405 (expand_mult): Handle properly multiplication by
29406 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
29407 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
29408 in the compiler if coeff is HOST_WIDE_INT_MIN.
29409 (expand_divmod): Don't make ext_op1 static, change it's type to
29410 uhwi. Avoid undefined behavior in -INTVAL (op1).
29412 PR rtl-optimization/50339
29413 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
29415 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
29416 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
29417 into splitting_ashiftrt field.
29418 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
29420 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
29423 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
29425 PR middle-end/56108
29426 * trans-mem.c (execute_tm_mark): Do not expand transactions that
29427 are sure to go irrevocable.
29429 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
29431 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
29432 scalars are valid operands.
29434 2013-02-21 Martin Jambor <mjambor@suse.cz>
29436 PR tree-optimization/56310
29437 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
29438 only matching indices and non-negative final offsets.
29439 (intersect_aggregates_with_edge): Pass src_idx to
29440 agg_replacements_to_vector. Pass src_idx insstead of index to
29441 intersect_with_agg_replacements.
29443 2013-02-21 Martin Jambor <mjambor@suse.cz>
29445 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
29446 instead of hard-wired defaults.
29448 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
29450 * doc/invoke.texi (MIPS Options): Update documentation of the
29451 floating-point multiply-accumulate instruction restrictions.
29453 2013-02-21 Kostya Serebryany <kcc@google.com>
29455 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
29456 asan_shadow_offset on x86_64 linux.
29458 2013-02-21 Richard Biener <rguenther@suse.de>
29460 PR tree-optimization/56415
29462 2013-02-11 Richard Biener <rguenther@suse.de>
29464 PR tree-optimization/56273
29465 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
29468 2013-02-21 Jakub Jelinek <jakub@redhat.com>
29471 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
29474 PR inline-asm/56405
29475 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
29476 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
29478 2013-02-20 Jan Hubicka <jh@suse.cz>
29480 PR tree-optimization/56265
29481 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
29482 when target is referenced for first time.
29484 2013-02-20 Richard Biener <rguenther@suse.de>
29486 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
29487 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
29488 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
29489 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
29490 not return anything.
29491 (rename_ssa_copies): Do not remove unused locals.
29492 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
29493 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
29494 * passes.c (execute_function_todo): Do not schedule unused locals
29495 removal if cleanup_tree_cfg did something.
29496 * tree-ssa-live.c (remove_unused_locals): Dump statistics
29497 about the number of removed locals.
29499 2013-02-20 Richard Biener <rguenther@suse.de>
29501 PR tree-optimization/56398
29502 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
29504 2013-02-20 Martin Jambor <mjambor@suse.cz>
29506 PR tree-optimization/55334
29507 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
29508 restricted pointers to arrays.
29510 2013-02-20 Richard Biener <rguenther@suse.de>
29511 Jakub Jelinek <jakub@redhat.com>
29513 PR tree-optimization/56396
29514 * tree-ssa-ccp.c (n_const_val): New static variable.
29515 (get_value): Return NULL for SSA names we don't have a lattice
29517 (ccp_initialize): Initialize n_const_val.
29518 * tree-ssa-copy.c (n_copy_of): New static variable.
29519 (init_copy_prop): Initialize n_copy_of.
29520 (get_value): Return NULL_TREE for SSA names we don't have a
29523 2013-02-20 Martin Jambor <mjambor@suse.cz>
29525 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
29527 2013-02-20 Richard Biener <rguenther@suse.de>
29529 * genpreds.c (write_lookup_constraint): Do not compare first
29530 letter of the constraint again.
29532 2013-02-20 Richard Biener <rguenther@suse.de>
29534 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
29536 (get_use_iv_cost): Terminate hashtable walk when coming across
29539 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
29541 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
29542 reassociation for avx2 targets.
29544 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
29546 * config/microblaze/microblaze.c: microblaze_has_clz = 0
29547 Add version check for v8.10.a to enable microblaze_has_clz
29548 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
29549 version and TARGET_PATTERN_COMPARE check
29550 * config/microblaze/microblaze.md: New clzsi2 instruction
29552 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
29554 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
29555 function before branching.
29557 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
29559 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
29561 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
29563 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
29565 PR middle-end/55889
29566 * sel-sched.c: Include ira.h.
29567 (implicit_clobber_conflict_p): New function.
29568 (moveup_expr): Use it.
29569 * Makefile.in (sel-sched.o): Depend on ira.h.
29571 2013-02-19 Richard Biener <rguenther@suse.de>
29573 PR tree-optimization/56384
29574 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
29575 (vn_hash_type): Split out from ...
29576 (vn_hash_constant_with_type): ... here.
29577 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
29578 (vn_phi_eq): Compare types from vn_phi_s structure.
29579 (vn_phi_lookup): Populate vn_phi_s type.
29580 (vn_phi_insert): Likewise.
29582 2013-02-19 Jakub Jelinek <jakub@redhat.com>
29584 PR tree-optimization/56350
29585 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
29586 if haven't found reduction or nested cycle operand, rather than
29587 asserting we must find it.
29589 PR tree-optimization/56381
29590 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
29593 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
29594 Jakub Jelinek <jakub@redhat.com>
29597 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
29598 (swap_optab_enable): Same.
29599 (init_all_optabs): Use argument instead of global.
29600 * tree.h (struct tree_optimization_option): New field target_optabs.
29601 * expr.h (init_all_optabs): Add argument to prototype.
29602 (TREE_OPTIMIZATION_OPTABS): New.
29603 (save_optabs_if_changed): Protoize.
29604 * optabs.h: Declare this_fn_optabs.
29605 * optabs.c (save_optabs_if_changed): New.
29606 Declare this_fn_optabs.
29607 (init_optabs): Add argument to init_all_optabs() call.
29608 * function.c (invoke_set_current_function_hook): Handle per
29610 * function.h (struct function): New field optabs.
29611 * config/mips/mips.c (mips_set_mips16_mode): Handle when
29612 optimization_current_node has changed.
29613 * target-globals.h (save_target_globals_default_opts): Protoize.
29614 * target-globals.c (save_target_globals_default_opts): New.
29616 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29619 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
29620 registers %fr12 and %fr12R as call used.
29623 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
29624 and HImode, require all displacements to be an integer multiple of
29626 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
29627 only allow QImode and HImode when reload is in progress and strict is
29628 true. Likewise for symbolic addresses. Use base14_operand to check
29629 displacements in REG+BASE addresses.
29631 2013-02-18 Richard Biener <rguenther@suse.de>
29633 PR tree-optimization/56366
29634 * tree-vect-loop.c (get_initial_def_for_induction): Properly
29635 handle sign-conversion of outer-loop initial induction value.
29637 2013-02-18 Richard Biener <rguenther@suse.de>
29639 PR middle-end/56349
29640 * cfghooks.c (merge_blocks): If we merge a latch into another
29641 block adjust references to it.
29642 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
29643 (verify_loop_structure): Verify that a recorded latch is in fact
29646 2013-02-18 Richard Biener <rguenther@suse.de>
29648 PR tree-optimization/56321
29649 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
29650 order SSA name release and virtual operand unlinking.
29652 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
29654 * config/microblaze/microblaze.md (save_stack_block): Define.
29655 (restore_stack_block): Likewise.
29657 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
29659 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
29660 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
29661 * config/microblaze/microblaze.c (microblaze_option_override):
29662 Bail out early for PIC modes when target does not support PIC.
29664 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
29666 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
29667 Replace with a microblaze version.
29668 (microblaze_trampoline_init): Adapt for microblaze.
29669 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
29672 2013-02-16 Jakub Jelinek <jakub@redhat.com>
29673 Dodji Seketeli <dodji@redhat.com>
29676 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
29677 (instrument_mem_region_access): Do not forget to always put
29678 instrumentation of the of 'base' and 'base + len' in a "if (len !=
29679 0) statement, even for cases where either 'base' or 'base + len'
29680 are not instrumented -- because they have been previously
29681 instrumented. Simplify the logic by putting all the statements
29682 instrument 'base + len' inside a sequence, and then insert that
29683 sequence right before the current insertion point. Then, to
29684 instrument 'base + len', just get an iterator on that statement.
29685 And do not forget to update the pointer to iterator the function
29686 received as argument.
29688 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
29690 PR rtl-optimization/56348
29691 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
29693 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
29695 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
29696 (clean_graph_dump_file): Pass base to start_graph_dump.
29698 2013-02-14 Richard Henderson <rth@redhat.com>
29701 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
29703 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
29705 * collect2-aix.h: Define F_LOADONLY.
29707 2013-02-14 Richard Biener <rguenther@suse.de>
29710 * varasm.c (output_constant_def_1): Get the decl representing
29711 the constant as argument.
29712 (output_constant_def): Wrap output_constant_def_1.
29713 (make_decl_rtl): Use output_constant_def_1 with the decl
29714 representing the constant.
29715 (build_constant_desc): Optionally re-use a decl already
29716 representing the constant.
29717 (tree_output_constant_def): Adjust.
29719 2013-02-14 Dodji Seketeli <dodji@redhat.com>
29722 * asan.c (instrument_builtin_call): Really put the length of the
29723 second source argument into src1_len.
29725 2013-02-13 Jakub Jelinek <jakub@redhat.com>
29727 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
29728 argument. If it is false, don't create edge from then_bb to
29730 (insert_if_then_before_iter): Pass true to it.
29731 (build_check_stmt): Pass false to it.
29732 (transform_statements): Flush hash table only on extended basic
29733 block boundaries, rather than at the beginning of every bb.
29734 Don't flush hash table on nonfreeing_call_p calls.
29735 * tree-flow.h (nonfreeing_call_p): New prototype.
29736 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
29738 2013-02-13 David S. Miller <davem@davemloft.net>
29740 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
29742 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
29745 * ira.c (max_regno_before_ira): Move from ...
29747 (fix_reg_equiv_init): Use max_regno_before_ira instead of
29750 2013-02-13 Jakub Jelinek <jakub@redhat.com>
29752 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
29754 2013-02-13 Richard Biener <rguenther@suse.de>
29757 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
29758 globals in MEM_REFs.
29760 2013-02-13 Richard Biener <rguenther@suse.de>
29762 * loop-init.c (loop_optimizer_init): Clear loop state when
29763 re-initializing preserved loops.
29764 * loop-unswitch.c (unswitch_single_loop): Return whether
29765 we unswitched the loop. Do not verify loop state here.
29766 (unswitch_loops): When we unswitched a loop discover new loops.
29768 2013-02-13 Kostya Serebryany <kcc@google.com>
29770 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
29772 * sanitizer.def: Rename __asan_init to __asan_init_v1.
29774 2013-02-12 Dodji Seketeli <dodji@redhat.com>
29776 Avoid instrumenting duplicated memory access in the same basic block
29777 * Makefile.in (asan.o): Add new dependency on hash-table.h
29778 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
29779 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
29780 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
29781 (free_mem_ref_resources, has_mem_ref_been_instrumented)
29782 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
29783 (get_mem_ref_of_assignment): New functions.
29784 (get_mem_refs_of_builtin_call): Extract from
29785 instrument_builtin_call and tweak a little bit to make it fit with
29787 (instrument_builtin_call): Use the new
29788 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
29789 of is_gimple_builtin_call.
29790 (instrument_derefs, instrument_mem_region_access): Insert the
29791 instrumented memory reference into the hash table.
29792 (maybe_instrument_assignment): Renamed instrument_assignment into
29793 this, and change it to advance the iterator when instrumentation
29794 actually happened and return true in that case. This makes it
29795 homogeneous with maybe_instrument_assignment, and thus give a
29796 chance to callers to be more 'regular'.
29797 (transform_statements): Clear the memory reference hash table
29798 whenever we enter a new BB, when we cross a function call, or when
29799 we are done transforming statements. Use
29800 maybe_instrument_assignment instead of instrumentation. No more
29801 need to special case maybe_instrument_assignment and advance the
29802 iterator after calling it; it's now handled just like
29803 maybe_instrument_call. Update comment.
29805 2013-02-13 Richard Biener <rguenther@suse.de>
29807 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
29808 Fix loop discovery code.
29810 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
29812 PR inline-asm/56148
29813 * lra-constraints.c (process_alt_operands): Match early clobber
29814 operand with itself. Check conflicts with earlyclobber only if
29815 the operand is not reloaded. Prefer to reload conflicting operand
29816 if earlyclobber and matching operands are the same.
29818 2013-02-12 Richard Biener <rguenther@suse.de>
29821 * lto-streamer-out.c (write_symbol): Do not output symbols
29822 for hard register variables.
29824 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
29827 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
29828 (umulsidi3_insn, mulsidi3_insn): New insns.
29830 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
29832 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
29833 (struct tune_params): Add vec_costs field.
29834 * config/arm/arm.c (arm_builtin_vectorization_cost)
29835 (arm_add_stmt_cost): New functions.
29836 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
29837 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
29838 (arm_default_vec_cost): New struct of type cpu_vec_costs.
29839 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
29840 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
29841 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
29842 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
29844 2013-02-12 Richard Biener <rguenther@suse.de>
29847 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
29848 decls again if possible.
29850 2013-02-12 Richard Biener <rguenther@suse.de>
29852 PR middle-end/56288
29853 * tree-ssa.c (verify_ssa_name): Fix check, move
29854 SSA_NAME_IN_FREE_LIST check up.
29856 2013-02-12 Jakub Jelinek <jakub@redhat.com>
29857 Steven Bosscher <steven@gcc.gnu.org>
29859 PR rtl-optimization/56151
29860 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
29861 equal to op0 or op1, and last_insn pattern is CODE operation
29862 with MEM dest and one of the operands matches that MEM.
29864 2013-02-11 Sriraman Tallam <tmsriram@google.com>
29866 * doc/extend.texi: Document Function Multiversioning and "default"
29867 parameter string to target attribute.
29868 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
29869 target attribute parameter is "default".
29870 (ix86_compare_version_priority): Remove checks for target attribute.
29871 (ix86_mangle_function_version_assembler_name): Change error to sorry.
29872 Remove check for target attribute equal to NULL. Add assert.
29873 (ix86_generate_version_dispatcher_body): Change error to sorry.
29875 2013-02-11 Iain Sandoe <iain@codesourcery.com>
29876 Jack Howarth <howarth@bromo.med.uc.edu>
29877 Patrick Marlier <patrick.marlier@gmail.com>
29880 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
29881 define ENDFILE_SPEC as TM_DESTRUCTOR.
29882 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
29884 2013-02-11 Alexander Potapenko <glider@google.com>
29885 Jack Howarth <howarth@bromo.med.uc.edu>
29886 Jakub Jelinek <jakub@redhat.com>
29889 * config/darwin.c (cdtor_record): Rename ctor_record.
29890 (sort_cdtor_records): Rename sort_ctor_records.
29891 (finalize_dtors): New routine to sort destructors by
29892 priority before use in assemble_integer.
29893 (machopic_asm_out_destructor): Use finalize_dtors if needed.
29895 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
29897 PR rtl-optimization/56275
29898 * simplify-rtx.c (avoid_constant_pool_reference): Check that
29899 offset is non-negative and less than cmode size before
29900 calling simplify_subreg.
29902 2013-02-11 Richard Biener <rguenther@suse.de>
29904 PR tree-optimization/56264
29905 * cfgloop.h (fix_loop_structure): Adjust prototype.
29906 * loop-init.c (fix_loop_structure): Return the number of
29907 newly discovered loops.
29908 * tree-cfgcleanup.c (repair_loop_structures): When new loops
29909 are discovered, do a full loop-closed SSA rewrite.
29911 2013-02-11 Richard Biener <rguenther@suse.de>
29913 PR tree-optimization/56273
29914 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
29916 (check_array_ref): Fix missing newline in dumps.
29917 (search_for_addr_array): Likewise.
29919 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
29921 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
29923 2013-02-09 Jakub Jelinek <jakub@redhat.com>
29926 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
29928 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
29930 PR rtl-optimization/56246
29931 * lra-constraints.c (simplify_operand_subreg): Try to reuse
29933 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
29934 constraints are satisfied.
29936 2013-02-08 Jeff Law <law@redhat.com>
29939 * emit-rtl.c (reg_is_parm_p): New function.
29940 * regs.h (reg_is_parm_p): New prototype.
29941 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
29942 callee-clobbered registers.
29944 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
29947 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
29948 If there is no implicit builtin declaration, just return NULL.
29950 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
29952 * config/i386/sse.md (FMAMODEM): New mode iterator.
29953 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
29954 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
29956 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
29958 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
29959 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
29960 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
29962 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
29964 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
29965 (microblaze*-*-elf): Likewise.
29966 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
29968 * config/microblaze/microblaze-c.c: Add builtin defines for
29969 _LITTLE_ENDIAN and _BIG_ENDIAN.
29970 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
29971 add to TARGET_DEFAULT flags.
29972 Expand ASM_SPEC and LINK_SPEC.
29973 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
29974 * config/microblaze/microblaze.md: Update extendsidi2 and
29975 movdi_internal instructions to use low-order / high-order reg
29977 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
29978 options and inversemask / mask of LITTLE_ENDIAN.
29979 * config/microblaze/t-microblaze: Expand multilib options to
29980 include mlittle-endian (le) and update exceptions patterns.
29982 2013-02-08 Jakub Jelinek <jakub@redhat.com>
29984 PR rtl-optimization/56195
29985 * lra-constraints.c (get_reload_reg): Don't reuse regs
29986 if they have smaller mode than requested, if they have
29987 wider mode than requested, try to return a SUBREG.
29989 PR tree-optimization/56250
29990 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
29991 if type is unsigned and code isn't MULT_EXPR.
29993 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
29995 PR tree-optimization/56064
29996 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
29997 bits according to mode.
29998 * fixed-value.h (fixed_from_double_int)
29999 (const_fixed_from_double_int): Adjust comments.
30001 2013-02-08 Richard Biener <rguenther@suse.de>
30004 * lto-streamer.h (struct data_in): Remove current_file, current_line
30005 and current_col members.
30006 * lto-streamer-out.c (lto_output_location): Stream changed bits
30007 en-block for efficiency.
30008 * lto-streamer-in.c (clear_line_info): Remove.
30009 (lto_input_location): Cache current file, line and column
30010 globally via local statics. Read changed bits en-block.
30011 (input_function): Do not call clear_line_info.
30012 (lto_read_body): Likewise.
30013 (lto_input_toplevel_asms): Likewise.
30015 2013-02-08 Michael Matz <matz@suse.de>
30017 PR tree-optimization/52448
30018 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
30019 (nt_call_phase): New static.
30020 (add_or_mark_expr): Only mark accesses with newer phase than any
30022 (nonfreeing_call_p): New.
30023 (nt_init_block): Update nt_call_phase, mark blocks as visited.
30024 (nt_fini_block): Keep blocks marked as visited.
30025 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
30027 2013-02-08 Richard Biener <rguenther@suse.de>
30029 * ira.c (ira): Free broken dominator information.
30031 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
30033 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
30035 2013-02-08 Marek Polacek <polacek@redhat.com>
30037 * cfgloop.c (verify_loop_structure): Add more checking of headers.
30039 2013-02-08 Richard Biener <rguenther@suse.de>
30041 PR middle-end/56181
30042 * cfgloop.h (flow_loops_find): Adjust.
30043 (bb_loop_header_p): Declare.
30044 * cfgloop.c (bb_loop_header_p): New function split out from ...
30045 (flow_loops_find): ... here. Adjust function signature,
30046 support incremental loop structure update.
30047 (verify_loop_structure): Cleanup. Verify a loop is a loop.
30048 * cfgloopmanip.c (fix_loop_structure): Move ...
30049 * loop-init.c (fix_loop_structure): ... here.
30050 (apply_loop_flags): Split out from ...
30051 (loop_optimizer_init): ... here.
30052 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
30053 in incremental mode, only remove dead loops here.
30055 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
30058 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
30059 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
30060 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
30061 (*round<mode>3.libgcc): New insns for fixed-modes.
30062 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
30063 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
30064 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
30065 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
30066 implementations. Define to __builtin_avr_absFX,
30067 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
30068 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
30069 __builtin_avr_countlsFX, respectively.
30070 * config/avr/avr-c.c (target.h): Include it.
30071 (enum avr_builtin_id): New enum.
30072 (avr_resolve_overloaded_builtin): New static function.
30073 (avr_register_target_pragmas): Use it to set
30074 targetm.resolve_overloaded_builtin.
30075 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
30076 tree nodes used by DEF_BUILTIN.
30077 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
30078 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
30079 <AVR_BUILTIN_xxBITS>: Same.
30081 2013-02-08 Richard Biener <rguenther@suse.de>
30083 * cfgloop.c (verify_loop_structure): Properly handle
30084 a loop exiting to another loop header.
30085 * ira-int.h (ira_loops): Remove.
30086 * ira.c (ira_loops): Remove.
30087 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
30088 (do_reload): Use loop_optimizer_finalize.
30089 * ira-build.c (create_loop_tree_nodes): Use get_loops and
30090 number_of_loops to access the loop tree.
30091 (more_one_region_p): Likewise.
30092 (finish_loop_tree_nodes): Likewise.
30093 (rebuild_regno_allocno_maps): Likewise.
30094 (mark_loops_for_removal): Likewise.
30095 (mark_all_loops_for_removal): Likewise.
30096 (remove_unnecessary_regions): Likewise.
30097 (ira_build): Likewise.
30098 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
30100 2013-02-08 Richard Biener <rguenther@suse.de>
30102 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
30103 * ipa-pure-const.c (analyze_function): Avoid calling
30104 mark_irreducible_loops twice.
30105 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
30107 2013-02-07 David S. Miller <davem@davemloft.net>
30109 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
30111 * var-tracking.c (vt_add_function_parameter): Test the presence of
30112 HAVE_window_save properly and do not remap argument registers when
30113 we have a leaf function.
30115 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
30118 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
30120 * config/i386/i386.c (ix86_print_operand): Ditto.
30122 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
30124 * lra-constraints.c (process_alt_operands): Fix recently added comment.
30126 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
30128 PR rtl-optimization/56225
30129 * lra-constraints.c (process_alt_operands): Check that reload hard
30130 reg can hold value for strict_low_part.
30132 2013-02-07 Jakub Jelinek <jakub@redhat.com>
30135 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
30136 dwarf2out_end_function.
30137 (in_first_function_p, maybe_at_text_label_p,
30138 first_loclabel_num_not_at_text_label): New variables.
30139 (dwarf2out_var_location): In the first function find out
30140 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
30141 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
30144 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
30146 PR rtl-optimization/56178
30147 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
30148 SUBREG of a register. Tidy up related block of code.
30149 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
30150 note if the source is a register or a SUBREG of a register.
30152 2013-02-07 Jakub Jelinek <jakub@redhat.com>
30155 * config/rs6000/rs6000.md (ptrm): New mode attr.
30156 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
30157 call_value_indirect_aix<pttrsize>,
30158 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
30161 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
30163 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
30164 if -bnortl. Convert to strcmp and strncmp.
30166 2013-02-07 Alan Modra <amodra@gmail.com>
30169 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
30170 addresses won't wrap when offsetting.
30171 (rs6000_secondary_reload): Provide secondary reloads needed for
30172 wrapping LO_SUM addresses.
30174 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
30176 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
30177 MACH, just __MACH__.
30179 2013-02-06 Richard Biener <rguenther@suse.de>
30181 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
30182 instead of calling fix_loop_structure.
30184 2013-02-06 Jakub Jelinek <jakub@redhat.com>
30186 PR middle-end/56217
30187 * omp-low.c (use_pointer_for_field): Return false if
30188 lower_send_shared_vars doesn't generate any copy-out code.
30190 2013-02-06 Tom de Vries <tom@codesourcery.com>
30192 PR rtl-optimization/56131
30193 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
30194 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
30195 of the label is NULL. Add comment.
30197 2013-02-05 Jakub Jelinek <jakub@redhat.com>
30199 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
30202 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
30203 (STATIC_LIBTSAN_LIBS): Likewise.
30204 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
30205 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
30206 is defined, don't add anything else beyond that.
30207 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
30208 (LINK_COMMAND_SPEC): Use them.
30210 PR tree-optimization/56205
30211 * tree-stdarg.c (check_all_va_list_escapes): Return true if
30212 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
30213 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
30215 2013-02-05 Richard Biener <rguenther@suse.de>
30217 PR tree-optimization/53342
30218 PR tree-optimization/53185
30219 * tree-vectorizer.h (vect_check_strided_load): Remove.
30220 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
30221 not disallow peeling for vectorized strided loads.
30222 (vect_check_strided_load): Make static and simplify.
30223 (vect_analyze_data_refs): Adjust.
30224 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
30225 correctly when vectorizing strided loads.
30227 2013-02-05 Richard Biener <rguenther@suse.de>
30229 * doc/install.texi: Refer to ISL, not PPL.
30231 2013-02-05 Jan Hubicka <jh@suse.cz>
30233 PR tree-optimization/55789
30234 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
30236 2013-02-05 Jan Hubicka <jh@suse.cz>
30238 PR tree-optimization/55789
30239 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
30240 the dead call anyway.
30242 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
30245 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
30247 2013-02-04 Alexander Potapenko <glider@google.com>
30248 Jack Howarth <howarth@bromo.med.uc.edu>
30249 Jakub Jelinek <jakub@redhat.com>
30252 * config/darwin.c (sort_ctor_records): Stabilized qsort
30253 on constructor priority by using original position.
30254 (finalize_ctors): New routine to sort constructors by
30255 priority before use in assemble_integer.
30256 (machopic_asm_out_constructor): Use finalize_ctors if needed.
30258 2013-02-04 Jakub Jelinek <jakub@redhat.com>
30261 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
30262 about visibility on artificial decls.
30263 * config/sol2.c (solaris_assemble_visibility): Likewise.
30265 2013-02-04 Kai Tietz <ktietz@redhat.com>
30268 * config/i386/i386.c (function_value_ms_64): Add additional valtype
30269 argument and improve checking of return-argument types for 16-byte
30271 (ix86_function_value_1): Add additional valtype argument on call
30272 of function_value_64.
30273 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
30274 handling infunction_value_64 function.
30276 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
30278 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
30280 2013-02-04 Richard Biener <rguenther@suse.de>
30282 PR tree-optimization/56188
30283 * tree-ssa-structalias.c (label_visit): Consider case with
30284 initially non-empty points-to set.
30285 (perform_var_substitution): Dump node mapping and clean up.
30287 2013-02-04 Richard Guenther <rguenther@suse.de>
30290 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
30291 node prevail as last resort.
30292 (lto_symtab_merge_decls): Remove guard on LTRANS here.
30293 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
30295 2013-02-04 Richard Biener <rguenther@suse.de>
30297 PR tree-optimization/56113
30298 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
30300 (equiv_class_lookup_or_add): ... this.
30301 (label_visit): Adjust and fix error in previous patch.
30302 (perform_var_substitution): Adjust.
30304 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
30306 * config/sh/divtab.c: Fix formatting and comments throughout the file.
30307 * config/sh/sh4-300.md: Likewise.
30308 * config/sh/sh4a.md: Likewise.
30309 * config/sh/constraints.md: Likewise.
30310 * config/sh/sh.md: Likewise.
30311 * config/sh/netbsd-elf.h: Likewise.
30312 * config/sh/predicates.md: Likewise.
30313 * config/sh/sh-protos.h: Likewise.
30314 * config/sh/ushmedia.h: Likewise.
30315 * config/sh/linux.h: Likewise.
30316 * config/sh/sh.c: Likewise.
30317 * config/sh/superh.h: Likewise.
30318 * config/sh/elf.h: Likewise.
30319 * config/sh/sh4.md: Likewise.
30320 * config/sh/sh.h: Likewise.
30322 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
30324 * config/pa/constraints.md: Adjust unused letters. Change "T"
30325 constraint to match_test floating_point_store_memory_operand().
30326 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
30327 (base14_operand): New.
30328 (floating_point_store_memory_operand): New.
30329 (integer_store_memory_operand): Revise to use base14_operand and
30330 reg_plus_base_memory_operand.
30331 (move_dest_operand): Allow symbolic_memory_operands.
30332 (symbolic_memory_operand): Check for LO_SOM.
30333 (symbolic_operand): Change default case to break.
30334 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
30335 CONST_DOUBLE values to be reloaded by putting them into memory when
30336 the destination is a floating point register.
30337 (movdf): Remove code to handle CONST_DOUBLE.
30339 (reload_indf_r1): New.
30340 (reload_insf_r1): New.
30341 Consistently use "Q" and "T" constraints with integer and floating
30342 point move instructions, respectively.
30343 (movdi): Remove FAIL.
30344 Change predicate for source operand unamed DImode move from
30345 general_operand to move_src_operand.
30346 (umulsidi3): Change predicate for destination operand to
30348 Likewise for similar unamed patterns.
30349 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
30350 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
30351 (hppa_legitimize_address): Simplify mask calculation.
30352 (pa_emit_move_sequence): Revised handling of secondary reloads from
30353 REG+D addresses for floating point loads and stores. Directly handle
30354 loading CONST0_RTX (mode) to a floating point register.
30355 (pa_secondary_reload): Handle reloading DF and SFmode constant values
30356 to floating point registers. Don't restrict secondary reloads to
30357 floating point registers to integer modes. Revise some comments and
30359 (TARGET_LEGITIMATE_ADDRESS_P): Define.
30360 (pa_legitimate_address_p): New.
30361 (pa_legitimize_reload_address): New.
30362 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
30363 (STRICT_REG_OK_FOR_BASE_P): New.
30364 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
30365 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
30367 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
30368 Andrew Dixie <andrewd@gentrack.com>
30370 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
30373 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
30375 * expmed.c (extract_bit_field_1): Pass the full width of the
30376 structure to get_best_reg_extraction_insn.
30378 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
30381 * configure.ac (use_cxa_atexit): Add AIX.
30382 * configure: Regenerate.
30384 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
30386 2013-02-01 Jakub Jelinek <jakub@redhat.com>
30389 * final.c (need_profile_function): New variable.
30390 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
30391 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
30392 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
30393 notes, targetm.asm_out.function_prologue doesn't emit anything,
30394 HAVE_prologue and profiler should be emitted before prologue,
30395 set need_profile_function instead of emitting it.
30396 (final_scan_insn): If need_profile_function, emit
30397 profile_function on the first NOTE_INSN_BASIC_BLOCK or
30398 NOTE_INSN_FUNCTION_BEG note.
30400 2013-02-01 Richard Henderson <rth@redhat.com>
30402 * config/rs6000/rs6000.md (smulditi3): New.
30405 * config/alpha/alpha.md (umulditi3): New.
30407 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
30409 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
30410 (ASM_OUTPUT_ALIGNED_LOCAL): New.
30412 2013-02-01 Richard Biener <rguenther@suse.de>
30414 PR tree-optimization/56113
30415 * tree-ssa-structalias.c (label_visit): Reduce work for
30416 single-predecessor nodes.
30418 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
30420 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
30421 range isn't testing for zero.
30423 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
30425 PR middle-end/56113
30426 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
30428 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
30429 Nick Clifton <nickc@redhat.com>
30431 * config/v850/constraints.md (Q): Define as a memory constraint.
30432 * config/v850/predicates.md (label_ref_operand): New predicate.
30433 (e3v5_shift_operand): New predicate.
30434 (ior_operator): New predicate.
30435 * config/v850/t-v850: Add e3v5 multilib.
30436 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
30437 (v850_gen_movdi): Prototype.
30438 * config/v850/v850.c: Add support for e3v5 architecture.
30439 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
30441 (construct_save_jarl): Add e3v5 long JARL support.
30442 (v850_adjust_insn_length): New function. Adjust length of call
30443 insns when using e3v5 instructions.
30444 (v850_gen_movdi): New function: Generate instructions to move a
30446 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
30447 (CPP_SPEC): Define __v850e3v5__ as appropriate.
30448 (TARGET_USE_FPU): Enable for e3v5.
30449 (CONST_OK_FOR_W): New macro.
30450 (ADJUST_INSN_LENGTH): Define.
30451 * config/v850/v850.md (UNSPEC_LOOP): Define.
30452 (attr cpu): Add v850e3v5.
30453 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
30454 (movdi): New pattern.
30455 (movdi_internal): New pattern.
30456 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
30457 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
30458 (cstoresf4): Likewise.
30459 (cstoredf4): Likewise.
30460 (insv): New pattern.
30461 (rotlso3_a): New pattern.
30462 (rotlsi3_b): New pattern
30463 (rotlsi3_v850e3v5): New pattern.
30464 (doloop_begin): New pattern.
30465 (fix_loop_counter): New pattern.
30466 (doloop_end): New pattern.
30467 (branch_normal): Add e3v5 long branch support.
30468 (branch_invert): Likewise.
30469 (branch_z_normal): Likewise.
30470 (branch_z_invert): Likewise.
30471 (branch_nz_normal): Likewise.
30472 (branch_nz_invert): Likewise.
30473 (call_internal_short): Add e3v5 register-indirect JARL support.
30474 (call_internal_long): Likewise.
30475 (call_value_internal_short): Likewise.
30476 (call_value_internal_long): Likewise.
30477 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
30478 (mloop): New option.
30479 * config.gcc: Add support for configuring v840e3v5 target.
30480 * doc/invoke.texi: Document new v850 specific command line options.
30482 2013-01-31 Paul Koning <ni1d@arrl.net>
30486 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
30487 children if parent is a class.
30488 (prune_unused_types_prune): Don't add DW_AT_declaration.
30490 2013-01-31 Richard Biener <rguenther@suse.de>
30492 PR tree-optimization/56157
30493 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
30494 match up operand with SLP child.
30496 2013-01-31 Jason Merrill <jason@redhat.com>
30499 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
30500 parameters the first time.
30501 (gen_scheduled_generic_parms_dies): Check completeness here.
30503 2013-01-31 Richard Biener <rguenther@suse.de>
30505 PR middle-end/53073
30506 * common.opt (faggressive-loop-optimizations): New flag,
30507 enabled by default.
30508 * doc/invoke.texi (faggressive-loop-optimizations): Document.
30509 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
30510 infer_loop_bounds_from_undefined by it.
30512 2013-01-31 Richard Biener <rguenther@suse.de>
30514 PR tree-optimization/56150
30515 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
30516 visit virtual operands.
30517 (find_uses_to_rename_bb): Likewise.
30519 2013-01-31 Richard Biener <rguenther@suse.de>
30521 PR tree-optimization/56150
30522 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
30523 mixed store non-store stmts.
30525 2013-01-30 Jakub Jelinek <jakub@redhat.com>
30528 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
30529 LIBASAN_EARLY_SPEC is defined.
30530 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
30531 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
30533 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
30536 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
30537 invalid args instead of ICEing on it.
30538 (ix86_valid_target_attribute_tree): Return error_mark_node if
30539 ix86_valid_target_attribute_inner_p failed.
30540 (ix86_valid_target_attribute_p): Return false only if
30541 ix86_valid_target_attribute_tree returned error_mark_node. Allow
30542 target("default") attribute.
30543 (sorted_attr_string): Change argument from const char * to tree,
30544 merge in all target attribute arguments rather than just one.
30545 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
30546 instead of free. Avoid using strcat.
30547 (ix86_mangle_function_version_assembler_name): Mangle
30548 target("default") as if no target attribute is present. Adjust
30549 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
30550 instead of xmalloc and XDELETEVEC instead of free.
30551 (ix86_function_versions): Don't return true if one of the decls
30552 doesn't have target attribute. If they don't and one of the decls
30553 is DECL_FUNCTION_VERSIONED, report an error. Adjust
30554 sorted_attr_string caller. Use XDELETEVEC instead of free.
30555 (ix86_supports_function_versions): Remove.
30556 (make_name): Fix up formatting.
30557 (make_dispatcher_decl): Remove resolver_name and its initialization.
30558 Avoid leaking memory.
30559 (is_function_default_version): Return true if there is
30560 target("default") attribute rather than no target attribute at all.
30561 (make_resolver_func): Avoid leaking memory.
30562 (ix86_generate_version_dispatcher_body): Likewise.
30563 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
30564 * target.def (supports_function_versions): Remove.
30565 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
30566 * doc/tm.texi: Regenerated.
30568 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
30570 PR rtl-optimization/56144
30571 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
30572 for values with side effects.
30574 2013-01-30 Richard Biener <rguenther@suse.de>
30576 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
30577 (sparseset_pop): Likewise.
30578 * cfganal.c (compute_idf): Likewise. Increase work-stack size
30579 to be able to use quick_push in the worker loop.
30581 2013-01-30 Marek Polacek <polacek@redhat.com>
30583 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
30585 2013-01-30 Richard Biener <rguenther@suse.de>
30588 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
30590 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
30592 PR tree-optimization/56064
30593 * fixed-value.c (fixed_from_double_int): New function.
30594 * fixed-value.h (fixed_from_double_int): New prototype.
30595 (const_fixed_from_double_int): New static inline function.
30596 * fold-const.c (native_interpret_fixed): New static function.
30597 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
30598 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
30599 (native_encode_fixed): New static function.
30600 (native_encode_expr) <FIXED_CST>: Use it.
30601 (native_interpret_int): Move double_int worker code to...
30602 * double-int.c (double_int::from_buffer): ...this new static method.
30603 * double-int.h (double_int::from_buffer): Prototype it.
30605 2013-01-30 Richard Biener <rguenther@suse.de>
30607 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
30608 New pointer-map and obstack.
30609 (init_alias_vars): Allocate pointer-map and obstack.
30610 (delete_points_to_sets): Free them.
30611 (find_what_var_points_to): Cache result.
30612 (find_what_p_points_to): Adjust for changed interface of
30613 find_what_var_points_to.
30614 (compute_points_to_sets): Likewise.
30615 (ipa_pta_execute): Likewise.
30617 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
30619 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
30620 * configure: Regenerate.
30621 * config.in: Regenerate.
30622 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
30623 #nobits/#progbits if supported.
30625 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
30628 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
30629 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
30630 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
30632 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
30634 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
30635 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
30637 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
30639 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
30640 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
30642 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
30644 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
30646 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
30647 * config/arm/cortex-a7.md: New bypasses using
30648 arm_mac_accumulator_is_result.
30650 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
30652 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
30653 (cortex_a7_neon_mla): Likewise.
30654 (cortex_a7_fpfmad): New reservation.
30655 (cortex_a7_fpmacs): Use ffmas and update required units.
30656 (cortex_a7_fpmuld): Update required units and latency.
30657 (cortex_a7_fpmacd): Likewise.
30658 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
30659 (cortex_a7_neon). Likewise.
30660 (bypass) Update participating units.
30662 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
30664 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
30665 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
30667 * config/arm/vfp11.md (vfp_farith): Use ffmas.
30668 (vfp_fmul): Use ffmad.
30669 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
30670 (cortex_r4_fmacd): Use ffmad.
30671 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
30672 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
30673 (cortex_a9_fmacd): Use ffmad.
30674 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
30675 (cortex_a8_vfp_macd): Use ffmad.
30676 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
30677 (cortex_a5_fpmacd): Use ffmad.
30678 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
30679 (cortex_a15_vfp_macd): Use ffmad.
30680 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
30682 2013-01-29 Jason Merrill <jason@redhat.com>
30685 * varasm.c (default_assemble_visibility): Don't warn about
30686 visibility on artificial decls.
30688 2013-01-29 Richard Biener <rguenther@suse.de>
30690 PR tree-optimization/56113
30691 * tree-ssa-structalias.c (equiv_class_lookup): Also return
30693 (label_visit): Free duplicate bitmaps and record the leader instead.
30694 (perform_var_substitution): Adjust.
30696 2013-01-29 Richard Biener <rguenther@suse.de>
30698 PR tree-optimization/55270
30699 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
30700 the CFG, schedule loops for fixup.
30702 2013-01-29 Nick Clifton <nickc@redhat.com>
30704 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
30707 2013-01-28 Leif Ekblad <leif@rdos.net>
30709 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
30710 * config/i386/i386.h (TARGET_RDOS): New macro.
30711 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
30712 * config/i386/i386.c (ix86_option_override_internal): For 64bit
30713 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
30714 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
30715 DEFAULT_LARGE_SECTION_THRESHOLD.
30716 * config/i386/i386.md (R14_REG, R15_REG): New constants.
30717 * config/i386/rdos.h: New file.
30718 * config/i386/rdos64.h: New file.
30720 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
30723 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
30726 2013-01-28 Jakub Jelinek <jakub@redhat.com>
30728 PR rtl-optimization/56117
30729 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
30730 call cselib_lookup_from_insn on the MEM before calling
30731 add_insn_mem_dependence.
30733 2013-01-28 Richard Biener <rguenther@suse.de>
30735 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
30736 to a stmt that didn't have one.
30737 (copy_phis_for_bb): Likewise for PHI arguments.
30738 (copy_debug_stmt): Likewise for debug stmts.
30740 2013-01-28 Richard Biener <rguenther@suse.de>
30742 PR tree-optimization/56034
30743 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
30744 (partition_builtin_p): Adjust.
30745 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
30746 it is the last partition.
30747 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
30748 up the vertex for the definition.
30749 (classify_partition): Classify whether a partition is a
30750 PKIND_REDUCTION, thus has uses outside of the loop.
30751 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
30752 Merge all PKIND_REDUCTION partitions into the last partition.
30753 (tree_loop_distribution): Seed partitions from reductions as well.
30755 2013-01-28 Jakub Jelinek <jakub@redhat.com>
30757 PR tree-optimization/56125
30758 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
30759 pow(x,c) into sqrt(x) * powi(x, n/2) or
30760 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
30761 optimizing for size.
30762 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
30763 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
30766 PR tree-optimization/56094
30767 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
30768 to UNKNOWN_LOCATION while gimplifying expr.
30770 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
30773 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
30774 operand 0 in movabs insn template for -masm=intel asm alternative.
30775 (*movabs<mode>_2): Ditto for operand 1.
30777 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
30780 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
30783 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
30785 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
30788 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
30790 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
30791 Undef to avoid warning.
30793 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
30795 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
30796 * configure: Regenerate.
30798 2013-01-25 Jakub Jelinek <jakub@redhat.com>
30800 PR tree-optimization/56098
30801 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
30802 for stmts with volatile ops.
30803 (cond_store_replacement): Don't optimize if assign has volatile ops.
30804 (cond_if_else_store_replacement_1): Don't optimize if either
30805 then_assign or else_assign have volatile ops.
30806 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
30809 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
30811 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
30813 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
30815 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
30816 missing ':' in asm example.
30818 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
30820 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
30821 entries into lane and laneq entries.
30822 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
30823 Remove AdvSIMD scalar modes.
30824 (aarch64_sq<r>dmulh_laneq<mode>): New.
30825 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
30827 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
30828 builtin implementations to relfect changes in RTL in aarch64-simd.md.
30829 * config/aarch64/iterators.md (VCOND): New.
30832 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
30835 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
30836 Add NULL LIBNAME argument to existing definitions.
30837 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
30838 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
30839 * config/avr/avr.c (DEF_BUILTIN): Same.
30840 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
30841 (avr_expand_builtin): Expand to a vanilla call if a libgcc
30842 implementation is available (DECL_ASSEMBLER_NAME is set).
30843 (avr_fold_absfx): New static function.
30844 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
30845 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
30846 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
30847 AVR_BUILTIN_ABSLLK.
30848 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
30849 (abshk, absk, abslk, absllk): Provide as static inline functions.
30851 2013-01-25 Marek Polacek <polacek@redhat.com>
30853 PR tree-optimization/56035
30854 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
30856 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
30858 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
30859 (*movtf_internal_rex64): Add (!o,C) alternative
30860 (*movxf_internal_rex64): Ditto.
30861 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
30863 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
30865 * doc/invoke.texi: fix typo.
30866 * doc/objc.texi: fix typo.
30868 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
30870 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
30871 for the first two alternatives.
30873 2013-01-24 Diego Novillo <dnovillo@google.com>
30875 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
30876 (ggc-zone.o): Remove.
30877 * configure.ac: Remove option --with-gc.
30878 * configure: Re-generate.
30879 * doc/install.texi: Remove documentation for --with-gc.
30880 * gengtype.c (write_enum_defn): Remove. Update all users.
30881 (write_Types_process_field): Remove generation of gt_e_* argument.
30882 (output_type_enum): Remove. Update all users.
30883 (write_enum_defn): Remove. Update all users.
30884 (enum alloc_zone): Remove. Update all users.
30885 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
30886 * ggc-common.c (ggc_splay_alloc): Remove first argument.
30887 Update all callers.
30888 (struct ptr_data): Remove field TYPE. Update all users.
30889 (gt_pch_note_object): Remove argument TYPE. Update all users.
30890 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
30892 * ggc-none.c (ggc_alloc_typed_stat): Remove.
30893 (struct alloc_zone): Remove.
30894 (ggc_internal_alloc_zone_stat): Remove.
30895 (ggc_internal_cleared_alloc_zone_stat): Remove.
30896 * ggc-page.c (ggc_alloc_typed_stat): Remove.
30897 (ggc_pch_count_object): Remove last argument. Update all users.
30898 (ggc_pch_alloc_object): Remove last argument. Update all users.
30899 (struct alloc_zone): Remove.
30900 * ggc-zone.c: Remove.
30901 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
30902 (struct alloc_zone): Remove.
30903 (ggc_alloc_typed_stat): Remove.
30904 (ggc_alloc_typed): Remove.
30905 (ggc_splay_alloc): Remove first argument.
30906 (rtl_zone): Remove. Update all users.
30907 (tree_zone): Remove. Update all users.
30908 (tree_id_zone): Remove. Update all users.
30909 (ggc_internal_zone_alloc_stat): Remove. Update all users.
30910 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
30911 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
30912 * tree-ssanames.c: Remove references to zone allocator in comments.
30914 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
30916 * config/avr/avr.c (avr_out_fract): Make register numbers that
30917 might be outside of source operand signed.
30919 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
30921 * config/i386/constraints.md (Yf): New constraint.
30922 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
30923 of f constraint to conditionaly disable x87 register preferences.
30924 (*movdf_internal): Ditto.
30925 (*movsf_internal): Ditto.
30927 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
30929 PR inline-asm/55934
30930 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
30931 that have operands with impossible constraints.
30932 Add a FIXME for a speed-up opportunity.
30933 * lra-constraints.c (process_alt_operands): Verify that a class
30934 selected from constraints on asms is valid for the operand mode.
30935 (curr_insn_transform): Remove incorrect comment.
30937 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
30939 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
30940 TOC operand is a valid symbol ref in the constant pool.
30942 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
30944 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
30946 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
30949 * config/avr/stdfix.h: New file.
30950 * t-avr (stdfix-gcc.h): New rule to build it.
30951 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
30953 2013-01-23 Kostya Serebryany <kcc@google.com>
30955 * config/darwin.h: remove dependency on
30956 CoreFoundation (asan on Mac OS).
30958 2013-01-23 Jakub Jelinek <jakub@redhat.com>
30961 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
30962 instead of cmpdi_operand for first comparison operand.
30963 Don't assert that comparison operands aren't both constants.
30965 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
30967 * doc/install.texi (Downloading the Source): Update references to
30968 downloading separate components.
30970 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
30972 * doc/extend.texi (__int128): Improve grammar.
30974 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
30977 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
30978 alternative to (o,r).
30979 (*movdi_internal_rex64): Remove (!o,n) alternative.
30980 (DImode immediate->memory splitter): Remove.
30981 (DImode immediate->memory peephole2): Remove.
30982 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
30983 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
30984 alternative to (!o,*r).
30985 (*movtf_internal_sse): New pattern.
30986 (*movxf_internal_rex64): New pattern.
30987 (*movxf_internal): Disable for TARGET_64BIT.
30988 (*movdf_internal_rex64): Remove (!o,F) alternative.
30990 2013-01-22 Jakub Jelinek <jakub@redhat.com>
30992 PR middle-end/56074
30993 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
30994 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
30995 * tree-vect-loop-manip.c (find_loop_location): Also ignore
30996 stmt locations where LOCATION_LOCUS of the stmt location is
30997 UNKNOWN_LOCATION or BUILTINS_LOCATION.
31000 * config/i386/i386.md (UNSPEC_STOS): New.
31001 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
31002 *strsetqi_1): Add UNSPEC_STOS.
31004 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
31007 * doc/invoke.texi: Remove left over -Wsynth example.
31009 2013-01-21 Jakub Jelinek <jakub@redhat.com>
31011 PR tree-optimization/56051
31012 * fold-const.c (fold_binary_loc): Don't fold
31013 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
31014 a narrowing conversion, or widening conversion from signed
31017 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
31019 PR rtl-optimization/56023
31020 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
31021 dependent on debug instruction.
31023 2013-01-21 Martin Jambor <mjambor@suse.cz>
31025 PR middle-end/56022
31026 * function.c (allocate_struct_function): Call
31027 invoke_set_current_function_hook earlier.
31029 2013-01-21 Jakub Jelinek <jakub@redhat.com>
31031 * reload1.c (init_reload): Only initialize reload_obstack
31032 during the first call.
31034 2013-01-21 Marek Polacek <polacek@redhat.com>
31036 * cfgloop.c (verify_loop_structure): Fix up grammar.
31038 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
31040 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
31041 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
31043 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
31046 * config/arm/marvell-pj4.md: Update copyright year.
31047 Fix up use of alu to alu_reg and simple_alu_imm.
31049 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
31051 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
31053 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
31056 * lra-constraints.c (curr_insn_transform): Don't reuse original
31057 insn for secondary memory move when memory mode should be different.
31059 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
31061 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
31062 atomic_storedi_1): New patterns.
31064 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
31066 btver2 pipeline descriptions.
31067 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
31069 * config/i386/i386.md (btver2_decode): New type attributes.
31070 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
31072 * config/i386/btver2.md: New file describing btver2 pipelines.
31074 2013-01-19 Andrew Pinski <apinski@cavium.com>
31076 PR tree-optimization/52631
31077 * tree-ssa-sccvn (visit_use): Before looking up the original
31078 statement, try looking up the simplified expression.
31080 2013-01-19 Anthony Green <green@moxielogic.com>
31082 * config/moxie/moxie.c (moxie_expand_prologue): Set
31083 current_function_static_stack_size.
31085 2013-01-18 Jakub Jelinek <jakub@redhat.com>
31087 PR tree-optimization/56029
31088 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
31089 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
31091 2013-01-18 Sharad Singhai <singhai@google.com>
31093 PR tree-optimization/55995
31094 * dumpfile.c (dump_loc): Print location only if available.
31095 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
31097 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
31100 * lra-constraints.c (curr_insn_transform): Reuse original insn for
31101 secondary memory move.
31102 (inherit_reload_reg): Use rclass instead of cl for
31103 check_secondary_memory_needed_p.
31105 2013-01-18 Jakub Jelinek <jakub@redhat.com>
31107 PR middle-end/56015
31108 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
31109 the case where writing real complex part of target modifies op1.
31111 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
31113 * config/aarch64/aarch64-simd.md
31114 (aarch64_vcond_internal<mode>): Handle unordered cases.
31115 * config/aarch64/iterators.md (v_cmp_result): New.
31117 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
31118 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
31120 * config/arm/marvell-pj4.md: New file.
31121 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
31122 * config/arm/arm.md (generic_sched): Add marvell_pj4.
31123 (generic_vfp): Likewise.
31124 * config/arm/arm-cores.def: Add marvell-pj4.
31125 * config/arm/arm-tune.md: Regenerate.
31126 * config/arm/arm-tables.opt: Regenerate.
31127 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
31128 * doc/invoke.texi: Document marvell-pj4.
31130 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
31132 * config/aarch64/arm_neon.h: Map scalar types to standard types.
31134 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
31139 * var-tracking.c (negative_power_of_two_p): New.
31140 (global_get_addr_cache, local_get_addr_cache): New.
31141 (get_addr_from_global_cache, get_addr_from_local_cache): New.
31142 (vt_canonicalize_addr): Rewrite using the above. Adjust the
31144 (vt_stack_offset_p): Remove.
31145 (vt_canon_true_dep): Always canonicalize loc's address.
31146 (clobber_overlapping_mems): Make sure we have a MEM.
31147 (local_get_addr_clear_given_value): New.
31148 (val_reset): Clear local cached entries.
31149 (compute_bb_dataflow): Create and release the local cache.
31150 Disable duplicate MEMs clobbering.
31151 (emit_notes_in_bb): Clobber MEMs likewise.
31152 (vt_emit_notes): Create and release the local cache.
31153 (vt_initialize, vt_finalize): Create and release the global
31154 cache, respectively.
31155 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
31157 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
31159 PR libmudflap/53359
31160 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
31161 not found in the symtab.
31163 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
31166 PR rtl-optimization/55547
31167 PR rtl-optimization/53827
31170 * alias.c (offset_overlap_p): New, factored out of...
31171 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
31172 the conservative special case for symbolic constants. Don't
31173 adjust zero sizes on alignment.
31175 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
31177 PR rtl-optimization/52573
31178 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
31179 REG_UNUSED for the same register.
31181 2013-01-17 Richard Biener <rguenther@suse.de>
31182 Marek Polacek <polacek@redhat.com>
31184 PR rtl-optimization/55833
31185 * loop-unswitch.c (unswitch_loops): Move loop verification...
31186 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
31187 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
31188 Set it to true when we're removing a loop from hierarchy tree in
31189 an irreducible region.
31190 (fix_bb_placements): Adjust caller.
31191 (fix_loop_placements): Likewise.
31193 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
31195 * config/avr/builtins.def (DEF_BUILTIN): Factor out
31196 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
31197 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
31198 Remove ID. Adjust comments.
31199 * config/avr/avr-c.c (avr_builtin_name): Remove.
31200 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
31201 * config/avr/avr.c (avr_tolower): New static function.
31202 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
31203 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
31204 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
31207 2013-01-17 Jan Hubicka <jh@suse.cz>
31209 PR tree-optimization/55273
31210 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
31212 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
31215 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
31216 store through atomic_store<mode>_1.
31217 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
31219 2013-01-17 Martin Jambor <mjambor@suse.cz>
31221 PR tree-optimizations/55264
31222 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
31223 for virtual methods.
31224 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
31225 virtual methods before inlining is over.
31226 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
31228 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
31231 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
31233 PR rtl-optimization/56005
31234 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
31235 pending reads for prefetch.
31237 2013-01-16 Ian Bolton <ian.bolton@arm.com>
31239 * config/aarch64/aarch64.md
31240 (*cstoresi_neg_uxtw): New pattern.
31241 (*cmovsi_insn_uxtw): New pattern.
31242 (*<optab>si3_uxtw): New pattern.
31243 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
31244 (*<optab>si3_insn_uxtw): New pattern.
31245 (*bswapsi2_uxtw): New pattern.
31247 2013-01-16 Richard Biener <rguenther@suse.de>
31249 * tree-inline.c (tree_function_versioning): Remove set but
31250 never used variable.
31252 2013-01-16 Richard Biener <rguenther@suse.de>
31254 PR tree-optimization/55964
31255 * tree-flow.h (rename_variables_in_loop): Remove.
31256 (rename_variables_in_bb): Likewise.
31257 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
31258 (copy_loop_before): Adjust and delete update-ssa status.
31259 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
31260 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
31261 (rename_variables_in_loop): Remove.
31262 (slpeel_update_phis_for_duplicate_loop): Likewise.
31263 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
31264 use available cfg machinery instead of duplicating it.
31265 Update PHI nodes and perform poor-mans SSA update here.
31266 (slpeel_tree_peel_loop_to_edge): Adjust.
31268 2013-01-16 Richard Biener <rguenther@suse.de>
31270 PR tree-optimization/54767
31271 PR tree-optimization/53465
31272 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
31273 (vrp_visit_phi_node): For PHI arguments coming via backedges
31274 drop all symbolical range information.
31275 (execute_vrp): Compute backedges.
31277 2013-01-16 Richard Biener <rguenther@suse.de>
31279 * doc/install.texi: Update CLooG and ISL requirements to
31282 2013-01-16 Christian Bruel <christian.bruel@st.com>
31285 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
31286 (broken_move): Handle UNSPECV_SP_SWITCH_B.
31287 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
31289 2013-01-16 DJ Delorie <dj@redhat.com>
31291 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
31292 (UNSPECV_SP_SWITCH_E): New.
31293 (sp_switch_1): Change to an unspec.
31294 (sp_switch_2): Change to an unspec. Don't use post-inc when we
31297 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
31299 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
31300 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
31301 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
31302 (expand_mem_thread_fence): Ditto.
31303 (expand_mem_signal_fence): Ditto.
31304 (expand_atomic_load): Ditto.
31305 (expand_atomic_store): Ditto.
31307 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
31309 PR rtl-optimization/55547
31310 PR rtl-optimization/53827
31313 * alias.c (memrefs_conflict_p): Set sizes to negative after
31316 2013-01-15 Jakub Jelinek <jakub@redhat.com>
31319 * function.c (thread_prologue_and_epilogue_insns): Always
31320 add crtl->drap_reg to set_up_by_prologue.set, even if
31321 stack_realign_drap is false.
31323 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
31325 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
31326 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
31327 *call): Fix indention.
31329 2013-01-15 Tom de Vries <tom@codesourcery.com>
31332 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
31335 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
31337 PR rtl-optimization/55153
31338 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
31340 2013-01-15 Martin Jambor <mjambor@suse.cz>
31342 PR tree-optimization/55920
31343 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
31344 accesses as grp_to_be_debug_replaced.
31346 2013-01-15 Jakub Jelinek <jakub@redhat.com>
31348 PR tree-optimization/55920
31349 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
31350 there is non-useless type conversion needed from debug rhs to lhs,
31351 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
31353 2013-01-15 Joseph Myers <joseph@codesourcery.com>
31354 Mikael Pettersson <mikpe@it.uu.se>
31357 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
31359 (ASM_OUTPUT_CASE_LABEL): Remove.
31360 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
31361 * final.c (shorten_branches): Update alignment of labels before
31362 jump tables if CASE_VECTOR_SHORTEN_MODE.
31364 2013-01-15 Richard Biener <rguenther@suse.de>
31367 * system.h: Do not include gmp.h for building host tools.
31369 2013-01-15 Richard Biener <rguenther@suse.de>
31371 PR middle-end/55882
31372 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
31373 account for bitpos when computing alignment.
31375 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
31377 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
31378 (ix86_target_macros_internal): Likewise.
31380 * config/i386/i386.c (m_CORE2I7): Removed.
31381 (m_CORE_HASWELL): New macro.
31382 (m_CORE_ALL): Likewise.
31383 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
31384 (initial_ix86_arch_features): Likewise.
31385 (processor_target_table): Initializations for Core avx2.
31386 (cpu_names): New names "core-avx2".
31387 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
31388 PROCESSOR_CORE_HASWELL.
31389 (ix86_issue_rate): New case.
31390 (ia32_multipass_dfa_lookahead): Likewise.
31391 (ix86_sched_init_global): Likewise.
31393 * config/i386/i386.h (TARGET_HASWELL): New macro.
31394 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
31395 (processor_type): New PROCESSOR_HASWELL.
31397 2013-01-15 Jakub Jelinek <jakub@redhat.com>
31399 PR tree-optimization/55955
31400 * tree-vect-loop.c (vectorizable_reduction): Give up early on
31401 *SHIFT_EXPR and *ROTATE_EXPR codes.
31403 PR tree-optimization/48766
31404 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
31405 -ftrapv disable -fwrapv.
31407 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
31410 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
31411 etc. to 1 and not to __flash.
31412 Use LL suffix for __INT24_MAX__ with -mint8.
31413 Use ULL suffix for __UINT24_MAX__ with -mint8.
31415 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
31417 * config/avr/avr-arch.h
31418 (struct base_arch_s): Use typedef avr_arch_t instead.
31419 (struct arch_info_s): Use typedef avr_arch_info_t instead.
31420 (struct mcu_type_s): Use typedef avr_mcu_t instead.
31421 * config/avr/avr.c: Same.
31422 * config/avr/avr-devices.c: Same.
31423 * config/avr/driver-avr.c: Same.
31424 * config/avr/gen-avr-mmcu-texi.c: Same.
31425 * config/avr/avr-mcus.def: Adjust comment.
31427 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
31429 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
31430 * config/aarch64/iterators.md (VALLDI): New.
31432 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
31433 Andi Kleen <ak@linux.intel.com>
31436 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
31437 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
31440 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
31442 * config/avr/avr-stdint.h: Remove trailing blanks.
31443 * config/avr/avr-log.h: Same.
31444 * config/avr/avr-arch.h: Same.
31445 * config/avr/avr-devices.c: Same.
31446 * config/avr/avr-dimode.md: Same.
31447 * config/avr/predicates.md: Same.
31448 * config/avr/avr-c.c: Same. And fix typo.
31450 * config/avr/avr-protos.h: Same. And:
31451 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
31452 (init_cumulative_args): Rename to avr_init_cumulative_args.
31453 (expand_prologue): Rename to avr_expand_prologue.
31454 (expand_epilogue): Rename to avr_expand_epilogue.
31455 (adjust_insn_length): Rename to avr_adjust_insn_length.
31456 (notice_update_cc): Rename to avr_notice_update_cc.
31457 (final_prescan_insn): Rename to avr_final_prescan_insn.
31458 * config/avr/avr.c: Same.
31459 * config/avr/avr.h: Same.
31460 * config/avr/avr.md: Remove trailing blanks.
31461 (prologue): Use avr_expand_prologue.
31462 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
31464 2013-01-14 Richard Biener <rguenther@suse.de>
31466 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
31467 verify_location, collect_subblocks): New functions.
31468 (verify_gimple_in_cfg): Verify that locations only reference
31469 BLOCKs in the functions BLOCK tree.
31471 2013-01-14 Richard Biener <rguenther@suse.de>
31473 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
31475 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
31477 (insert_out_of_ssa_copy_on_edge): Likewise.
31478 (rewrite_close_phi_out_of_ssa): Likewise.
31479 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
31481 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
31482 propagated constants.
31483 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
31486 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
31488 * config/avr/avr-modes.def: Add GPL copyright notice.
31490 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
31492 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
31493 MEMMODEL_MASK to determine memory model.
31494 (atomic_store<mode>): Ditto from operands[2].
31495 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
31497 2013-01-13 Jakub Jelinek <jakub@redhat.com>
31500 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
31501 (fold_gimple_assign): Don't call unshare_expr here.
31502 (fold_ctor_reference): Call unshare_expr.
31504 2013-01-13 Terry Guo <terry.guo@arm.com>
31506 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
31507 * doc/fragments.texi: Document MULTILIB_REUSE.
31508 * gcc.c (multilib_reuse): New internal spec.
31509 (set_multilib_dir): Also search multilib from multilib_reuse.
31510 * genmultilib (tmpmultilib3): Refactor code.
31511 (tmpmultilib4): Ditto.
31512 (multilib_reuse): New multilib argument.
31514 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
31516 * Makefile.in: Update copyright.
31518 2013-01-12 Tom de Vries <tom@codesourcery.com>
31520 PR middle-end/55890
31521 * calls.c (expand_call): Check if arg_nr is valid.
31523 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
31525 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
31526 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
31527 documentation. Add missing '__' in front of
31528 __builtin_ia32_packssdw256.
31530 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
31533 * config/s390/s390.c (s390_preferred_reload_class): Do not return
31534 NO_REGS for larl operands.
31535 (s390_reload_larl_operand): Use s390_load_address instead of
31538 2013-01-11 Richard Biener <rguenther@suse.de>
31540 * tree-cfg.c (verify_node_sharing_1): Split out from ...
31541 (verify_node_sharing): ... here.
31542 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
31544 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
31546 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
31547 Substitute TREECHECKING.
31548 * configure: Regenerate.
31549 * Makefile.in (TREECHECKING): New.
31551 2013-01-11 Richard Guenther <rguenther@suse.de>
31553 PR tree-optimization/44061
31554 * tree-vrp.c (extract_range_basic): Compute zero as
31555 value-range for __builtin_constant_p of function parameters.
31557 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
31559 Update copyright years.
31561 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
31563 PR rtl-optimization/55672
31564 * lra-eliminations.c (mark_not_eliminable): Permit addition with
31565 const to be eliminable.
31567 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
31569 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
31570 * configure: Regenerate.
31572 2013-01-10 Richard Biener <rguenther@suse.de>
31574 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
31576 2013-01-10 Richard Biener <rguenther@suse.de>
31579 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
31580 locations for virtual PHI arguments.
31581 (rewrite_update_phi_arguments): Likewise.
31583 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
31585 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
31588 2013-01-10 Jakub Jelinek <jakub@redhat.com>
31590 PR tree-optimization/55921
31591 * tree-complex.c (expand_complex_asm): New function.
31592 (expand_complex_operations_1): Call it for GIMPLE_ASM.
31594 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
31597 * config/s390/s390.c (s390_symref_operand_p)
31598 (s390_loadrelative_operand_p): Merge the two functions.
31599 (s390_check_qrst_address, print_operand_address): Add parameters
31600 to s390_loadrelative_operand_p invokation.
31601 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
31602 (s390_reload_larl_operand, s390_secondary_reload): Use
31603 s390_loadrelative_operand_p instead of s390_symref_operand_p.
31604 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
31606 2013-01-09 Mike Stump <mikestump@comcast.net>
31608 * dse.c (record_store): Remove unnecessary assert.
31610 2013-01-09 Jan Hubicka <jh@suse.cz>
31612 PR tree-optimization/55569
31613 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
31614 * cfgloop.h (scale_loop_profile): Likewise.
31616 2013-01-09 Jan Hubicka <jh@suse.cz>
31619 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
31621 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
31623 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
31625 PR middle-end/55114
31626 * expr.h (maybe_emit_group_store): Declare.
31627 * expr.c (maybe_emit_group_store): New function.
31628 * builtins.c (expand_builtin_int_roundingfn): Call it.
31629 (expand_builtin_int_roundingfn_2): Likewise.
31631 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
31633 PR rtl-optimization/55829
31634 * lra-constraints.c (match_reload): Add code for absent output.
31635 (curr_insn_transform): Add code for reloads of matched inputs
31638 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
31640 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
31641 attribute of movddup insn to DF.
31642 (*vec_interleave_lowv2df): Ditto.
31643 (vec_dupv2df): Ditto.
31645 2013-01-09 Jan Hubicka <jh@suse.cz>
31647 PR tree-optimiation/55875
31648 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
31649 EVERY_ITERATION parameter.
31650 (number_of_iterations_exit): Check if exit is executed every iteration.
31651 (idx_infer_loop_bounds): Similarly here.
31652 (n_of_executions_at_most): Simplify
31653 to only test for cases where statement is dominated by the
31654 particular bound; handle correctly the "postdominance" test.
31655 (scev_probably_wraps_p): Use max loop iterations info
31656 as a global bound first.
31658 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
31659 Nick Clifton <nickc@redhat.com>
31661 * config/v850/v850.md (cbranchsf4): New pattern.
31662 (cstoresf4): New pattern.
31663 (cbranchdf4): New pattern.
31664 (cstoredf4): New pattern.
31665 (movsicc): Disallow floating point comparisons.
31666 (cmpsf_le_insn): Fix order of operators.
31667 (cmpsf_lt_insn): Likewise.
31668 (cmpsf_eq_insn): Likewise.
31669 (cmpdf_le_insn): Likewise.
31670 (cmpdf_lt_insn): Likewise.
31671 (cmpdf_eq_insn): Likewise.
31672 (cmpsf_ge_insn): Use LE comparison.
31673 (cmpdf_ge_insn): Likewise.
31674 (cmpsf_gt_insn): Use LT comparison.
31675 (cmpdf_gt_insn): Likewise.
31676 (cmpsf_ne_insn): Delete pattern.
31677 (cmpdf_ne_insn): Delete pattern.
31678 * config/v850/v850.c (v850_gen_float_compare): Use
31679 gen_cmpdf_eq_insn for NE comparison.
31680 (v850_float_z_comparison_operator)
31681 (v850_float_nz_comparison_operator): Move from here ...
31682 * config/v850/predicates.md: ... to here. Move GT and GE
31683 comparisons into v850_float_z_comparison_operator.
31684 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
31686 (v850_float_nz_comparison_operator): Likewise.
31688 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
31690 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
31691 with calls to gen_insvsi/gen_insvdi.
31693 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
31695 * config/i386/i386.c (initial_ix86_tune_features): Set up
31696 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
31698 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
31699 Jakub Jelinek <jakub@redhat.com>
31701 PR tree-optimization/48189
31702 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
31703 If nitercst is 0, don't predict the exit edge.
31705 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
31707 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
31708 in asm_fprintf with reg_names.
31709 (aarch64_print_operand_address): Likewise.
31710 (aarch64_return_addr): Likewise.
31711 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
31713 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
31715 * config/pa/pa.h (VAL_U6_BITS_P): Define.
31716 (INT_U6_BITS): Likewise.
31717 * config/pa/predicates.md (uint6_operand): New predicate.
31718 (shift5_operand, shift6_operand): Likewise.
31719 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
31721 (lshrdi3): Use shift6_operand.
31722 (shrpsi4, shrpdi4): New insn patterns.
31723 (extzv): Delete expander.
31724 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
31725 predicates in unamed zero extract patterns. Tighten common constraint.
31726 (extv): Delete expander.
31727 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
31728 predicates in unamed sign extract patterns. Tighten common constraint.
31729 (insv): Delete expander.
31730 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
31731 predicates in unamed insert patterns. Tighten common constraint.
31732 Change uint32_operand predicate to uint6_operand predicate in unamed
31733 DImode pattern to insert constant values of type 1...1xxxx.
31735 2013-01-04 Jan Hubicka <jh@suse.cz>
31737 PR tree-optimization/55823
31738 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
31741 2013-01-08 Jakub Jelinek <jakub@redhat.com>
31742 Uros Bizjak <ubizjak@gmail.com>
31744 PR rtl-optimization/55845
31745 * df-problems.c (can_move_insns_across): Stop scanning at
31746 volatile_insn_p source instruction or give up if
31747 across_from .. across_to range contains any volatile_insn_p
31750 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
31752 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
31753 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
31755 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
31756 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
31758 2013-01-08 Jakub Jelinek <jakub@redhat.com>
31761 * asan.c (asan_clear_shadow): New function.
31762 (asan_emit_stack_protection): Use it.
31764 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
31766 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
31767 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
31768 with tab instead of space.
31770 2013-01-08 Nick Clifton <nickc@redhat.com>
31772 * config/rl78/rl78.c (rl78_expand_prologue): Always select
31773 register bank 0 at the start of an interrupt handler.
31774 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
31777 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
31779 * config/aarch64/aarch64-simd.md
31780 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
31781 (aarch64_simd_bsl): Likewise.
31782 (aarch64_vcond_internal<mode>): Likewise.
31783 (vcond<mode><mode>): Likewise.
31784 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
31785 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
31787 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
31789 * config/aarch64/aarch64-builtins.c
31790 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
31792 2013-01-08 Martin Jambor <mjambor@suse.cz>
31795 * tree-sra.c (analyze_access_subtree): Return true also after
31796 potentially creating a debug-only replacement.
31798 2013-01-08 Jakub Jelinek <jakub@redhat.com>
31800 PR middle-end/55890
31801 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
31803 PR tree-optimization/54120
31804 * tree-vrp.c (range_fits_type_p): Don't allow
31805 src_precision < precision from signed vr to unsigned_p
31806 if vr->min or vr->max is negative.
31807 (simplify_float_conversion_using_ranges): Test can_float_p
31808 against CODE_FOR_nothing.
31810 2013-01-08 Jakub Jelinek <jakub@redhat.com>
31811 Richard Biener <rguenther@suse.de>
31813 PR middle-end/55851
31814 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
31815 types instead of just INTEGER_TYPE types.
31817 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
31819 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
31822 2013-01-07 Steve Ellcey <sellcey@mips.com>
31825 * config/mips/mips.opt: Change mad to mmad to match documentation.
31827 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
31830 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
31831 .progmemx.data now.
31833 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
31836 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
31837 (avr_addrspace_t): Add .section_name field.
31838 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
31840 (avr_addrspace): Same. Initialize .section_name. Remove last
31841 NULL entry. Put __memx into .progmemx.data.
31842 (progmem_section_prefix): Remove.
31843 (avr_asm_init_sections): No need to initialize progmem_section.
31844 (avr_asm_named_section): Use avr_addrspace[].section_name to get
31845 section name prefix.
31846 (avr_asm_select_section): Ditto. And use get_unnamed_section to
31847 retrieve the progmem section.
31848 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
31849 boundary to run over avr_addrspace[].
31850 (avr_register_target_pragmas): Ditto.
31852 2013-01-06 Jakub Jelinek <jakub@redhat.com>
31854 * varasm.c (output_constant_def_contents): For asan_protect_global
31855 protected strings, adjust DECL_ALIGN if needed, before testing for
31857 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
31858 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
31860 (output_object_block): For asan protected decls, emit asan padding
31861 after their contents.
31862 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
31863 (asan_finish_file): Test it here instead.
31865 2013-01-07 Nick Clifton <nickc@redhat.com>
31866 Matthias Klose <doko@debian.org>
31867 Doug Kwan <dougkwan@google.com>
31868 H.J. Lu <hongjiu.lu@intel.com>
31871 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
31873 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
31875 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
31877 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
31879 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
31881 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
31884 * doc/install.texi (Cross-Compiler-Specific Options): Document
31887 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
31889 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
31890 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
31891 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
31892 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
31893 vqmovun_high_s64): Fix source operand number and update copyright.
31895 2013-01-07 Richard Biener <rguenther@suse.de>
31897 PR middle-end/55890
31898 * gimple.h (gimple_call_builtin_p): New overload.
31899 * gimple.c (validate_call): New function.
31900 (gimple_call_builtin_p): Likewise.
31901 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
31902 Use gimple_call_builtin_p.
31903 (find_func_clobbers): Likewise.
31904 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
31905 (strlen_optimize_stmt): Likewise.
31907 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
31909 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
31910 (vld1q_dup_*): Likewise.
31911 (vld1_*): Likewise.
31912 (vld1q_*): Likewise.
31913 (vld1_lane_*): Likewise.
31914 (vld1q_lane_*): Likewise.
31916 2013-01-07 Richard Biener <rguenther@suse.de>
31918 * lto-streamer.h (LTO_minor_version): Bump to 2.
31920 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
31922 * config/aarch64/aarch64-protos.h
31923 (aarch64_const_double_zero_rtx_p): Rename to...
31924 (aarch64_float_const_zero_rtx_p): ...this.
31925 (aarch64_float_const_representable_p): New.
31926 (aarch64_output_simd_mov_immediate): Likewise.
31927 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
31928 move immediate case.
31929 * config/aarch64/aarch64.c
31930 (aarch64_const_double_zero_rtx_p): Rename to...
31931 (aarch64_float_const_zero_rtx_p): ...this.
31932 (aarch64_print_operand): Allow printing of new constants.
31933 (aarch64_valid_floating_const): New.
31934 (aarch64_legitimate_constant_p): Check for valid floating-point
31936 (aarch64_simd_valid_immediate): Likewise.
31937 (aarch64_vect_float_const_representable_p): New.
31938 (aarch64_float_const_representable_p): Likewise.
31939 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
31940 (aarch64_output_simd_mov_immediate): New.
31941 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
31942 (*movdf_aarch64): Likewise.
31943 * config/aarch64/constraints.md (Ufc): New.
31944 (Y): call aarch64_float_const_zero_rtx.
31945 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
31947 2013-01-07 Richard Biener <rguenther@suse.de>
31949 PR tree-optimization/55888
31950 PR tree-optimization/55862
31951 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
31952 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
31953 not if it is contained therein.
31955 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
31957 * config/avr/t-avr: Typo.
31959 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
31962 * config/avr/t-avr: Don't automatically rebuild
31963 $(srcdir)/config/avr/t-multilib
31964 $(srcdir)/config/avr/avr-tables.opt
31965 $(srcdir)/doc/avr-mmcu.texi
31966 (avr-mcus): New phony target to build them on request.
31967 (s-avr-mlib, s-avr-mmcu-texi): Remove.
31968 * avr/avr-mcus.def: Adjust comments.
31970 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
31972 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
31974 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
31976 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
31978 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
31980 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
31982 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
31984 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
31985 to generate profiling.
31986 * config/rs6000/aix64.h (LIB_SPEC): Same.
31988 2013-01-04 Andrew Pinski <apinski@cavium.com>
31990 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
31992 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
31994 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
31996 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
31998 (ix86_expand_move): Ditto.
31999 (ix86_zero_extend_to_Pmode): Ditto.
32000 (ix86_expand_call): Ditto.
32001 (ix86_expand_special_args_builtin): Ditto.
32002 (ix86_expand_builtin): Ditto.
32004 2013-01-04 Richard Biener <rguenther@suse.de>
32006 PR tree-optimization/55862
32007 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
32008 translating them through PHI nodes.
32010 2013-01-04 Martin Jambor <mjambor@suse.cz>
32012 PR tree-optimization/55755
32013 * tree-sra.c (sra_modify_assign): Do not check that an access has no
32014 children when trying to avoid producing a VIEW_CONVERT_EXPR.
32016 2013-01-04 Marek Polacek <polacek@redhat.com>
32018 PR middle-end/55859
32019 * opts.c (default_options_optimization): Clarify error message.
32021 2013-01-04 Richard Biener <rguenther@suse.de>
32023 PR middle-end/55863
32024 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
32027 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
32030 * config/pa/pa.md (movsi): Revert previous change.
32031 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
32034 2013-01-03 Richard Henderson <rth@redhat.com>
32036 * config/i386/i386.c (ix86_expand_move): Always assign to op1
32037 after eliminating TLS symbols.
32039 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
32042 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
32043 * graphite-poly.c (debug_gmp_value): Likewise.
32045 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
32048 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
32049 selected code model, define __code_mode_small__, __code_model_medium__,
32050 __code_model_large__, __code_model_32__ or __code_model_kernel__.
32051 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
32052 xchg temporary register with %k. Declare temporary register as
32054 [__x86_64__]: For medium and large code models, preserve %rbx register.
32056 2013-01-03 Richard Biener <rguenther@suse.de>
32058 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
32059 (dump_subscript): Adjust.
32060 (finalize_ddr_dependent): Do not dump redundant info.
32061 (analyze_siv_subscript): Adjust.
32062 (subscript_dependence_tester): Likewise.
32063 (compute_affine_dependence): Likewise.
32065 2013-01-03 Richard Biener <rguenther@suse.de>
32068 2013-01-03 Richard Biener <rguenther@suse.de>
32070 PR tree-optimization/55857
32071 * tree-vect-stmts.c (vectorizable_load): Do not setup
32072 re-alignment for invariant loads.
32074 2013-01-02 Richard Biener <rguenther@suse.de>
32076 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
32077 invariant load do not generate a vector load from the scalar location.
32079 2013-01-03 Richard Biener <rguenther@suse.de>
32081 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
32082 for not vectorizing.
32083 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
32084 not build INDIRECT_REFs, call get_name once only.
32085 (vect_create_data_ref_ptr): Likewise. Dump base object kind
32086 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
32088 2013-01-03 Richard Biener <rguenther@suse.de>
32090 PR tree-optimization/55857
32091 * tree-vect-stmts.c (vectorizable_load): Do not setup
32092 re-alignment for invariant loads.
32094 2013-01-03 Richard Biener <rguenther@suse.de>
32097 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
32098 prefer a built-in decl.
32100 2013-01-03 Jakub Jelinek <jakub@redhat.com>
32102 * gcc.c (process_command): Update copyright notice dates.
32103 * gcov.c (print_version): Likewise.
32104 * gcov-dump.c (print_version): Likewise.
32106 PR rtl-optimization/55838
32107 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
32108 iv0.step, iv1.step and step.
32110 2013-01-03 Jakub Jelinek <jakub@redhat.com>
32111 Marc Glisse <marc.glisse@inria.fr>
32113 PR tree-optimization/55832
32114 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
32115 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
32116 integer_{one,zero}_node.
32118 2013-01-03 Jakub Jelinek <jakub@redhat.com>
32121 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
32122 * var-tracking.c (reverse_op): Don't add reverse ops to
32123 VALUEs that have already
32124 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
32126 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
32128 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
32130 2013-01-02 Teresa Johnson <tejohnson@google.com>
32132 * dumpfile.c (dump_loc): Print filename with location.
32133 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
32134 new location_t parameter to emit complete unroll message with
32135 new dump framework.
32136 (canonicalize_loop_induction_variables): Compute loops location
32137 and pass to try_unroll_loop_completely.
32138 * loop-unroll.c (report_unroll_peel): New function.
32139 (peel_loops_completely): Use new dump format with location
32140 for main dumpfile message, and invoke report_unroll_peel on success.
32141 (decide_unrolling_and_peeling): Ditto.
32142 (decide_peel_once_rolling): Remove old dumpfile message subsumed
32143 by report_unroll_peel.
32144 (decide_peel_completely): Ditto.
32145 (decide_unroll_constant_iterations): Ditto.
32146 (decide_unroll_runtime_iterations): Ditto.
32147 (decide_peel_simple): Ditto.
32148 (decide_unroll_stupid): Ditto.
32149 * cfgloop.c (get_loop_location): New function.
32150 * cfgloop.h (get_loop_location): Declare.
32152 2013-01-02 Sriraman Tallam <tmsriram@google.com>
32154 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
32157 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
32159 PR middle-end/55198
32160 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
32161 BLKmode objects when EXPAND_MEMORY is specified.
32163 2013-01-02 Sriraman Tallam <tmsriram@google.com>
32165 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
32167 (fold_builtin_cpu): Do not share cpu model decls across statements.
32169 2013-01-02 Jason Merrill <jason@redhat.com>
32172 * tree.c (build_array_type_1): Revert earlier change.
32174 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
32176 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
32178 * config/aarch64/aarch64-tune.md: Re-generate.
32180 2013-01-02 Richard Biener <rguenther@suse.de>
32182 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
32183 invariant load do not generate a vector load from the scalar location.
32185 2013-01-02 Richard Biener <rguenther@suse.de>
32188 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
32189 * configure: Regenerate.
32191 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
32193 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
32194 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
32195 (expand_builtin_int_roundingfn_2): Keep the original target around
32196 for the fallback case.
32198 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
32200 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
32201 to be clear for sign changes.
32203 2013-01-01 Jan Hubicka <jh@suse.cz>
32205 * ipa-inline-analysis.c: Fix formatting.
32207 2013-01-01 Jakub Jelinek <jakub@redhat.com>
32209 PR tree-optimization/55831
32210 * tree-vect-loop.c (get_initial_def_for_induction): Use
32211 gsi_after_labels instead of gsi_start_bb.
32213 Copyright (C) 2013 Free Software Foundation, Inc.
32215 Copying and distribution of this file, with or without modification,
32216 are permitted in any medium without royalty provided the copyright
32217 notice and this notice are preserved.