1 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
3 * vec.h (vec_prefix, vec): Prefix member names with "m_".
4 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
6 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
8 * basic-block.h (edge_list): Prefix member names with "m_".
9 * context.h (context): Likewise.
10 * domwalk.h (dom_walker): Likewise.
11 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
12 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
13 * hash-table.h (hash_table): Likewise.
14 * machmode.h (bit_field_mode_iterator): Likewise.
15 * pass_manager.h (pass_list): Likewise.
16 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
17 * tree-pass.h (pass_data): Likewise.
18 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
19 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
20 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
21 * asan.c (pass_data_asan): Update accordingly.
22 * cfganal.c (control_dependences::find_control_dependence): Likewise.
23 (control_dependences::control_dependences): Likewise.
24 (control_dependences::~control_dependences): Likewise.
25 (control_dependences::~control_dependences): Likewise.
26 (control_dependences::get_edges_dependent_on): Likewise.
27 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
28 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
29 * context.c (gcc::context::context): Likewise.
30 * cprop.c (pass_rtl_cprop::clone): Likewise.
31 * domwalk.c (dom_walker::walk): Likewise.
32 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
33 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
34 * mode-switching.c (pass_mode_switching::clone): Likewise.
35 * passes.c (opt_pass::opt_pass): Likewise.
36 (pass_manager::pass_manager): Likewise.
37 * predict.c (pass_strip_predict_hints::clone): Likewise.
38 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
39 (pass_split_all_insns::clone): Likewise.
40 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
42 (bit_field_mode_iterator::next_mode): Likewise.
43 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
44 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
45 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
46 * tree-complex.c (pass_lower_complex::clone): Likewise.
47 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
48 * tree-object-size.c (pass_object_sizes::clone): Likewise.
49 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
50 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
51 (pass_fold_builtins::clone): Likewise.
52 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
53 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
54 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
55 (pass_cd_dce::clone): Likewise.
56 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
57 (pass_phi_only_cprop::clone): Likewise.
58 * tree-ssa-dse.c (pass_dse::clone): Likewise.
59 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
60 * tree-ssa-loop.c (pass_lim::clone): Likewise.
61 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
62 * tree-ssa-pre.c (pass_fre::clone): Likewise.
63 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
64 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
65 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
66 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
67 * tree-vrp.c (pass_vrp::clone): Likewise.
68 * tsan.c (pass_tsan::clone): Likewise.
70 2013-09-30 Jakub Jelinek <jakub@redhat.com>
73 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
74 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
77 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
78 optimization, punt if sign_bit_p looked through any zero extension.
80 2013-09-30 Teresa Johnson <tejohnson@google.com>
82 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
83 Update redirected out edge count in joiner case.
84 (ssa_redirect_edges): Common the joiner and non-joiner cases
85 so that joiner case gets profile updates.
87 2013-09-30 Richard Biener <rguenther@suse.de>
89 PR tree-optimization/58554
90 * tree-loop-distribution.c (classify_partition): Require unconditionally
91 executed stores for memcpy and memset recognition.
92 (tree_loop_distribution): Calculate dominance info.
94 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
96 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
97 (NO_PROFILE_COUNTERS): Likewise.
98 (PROFILE_HOOK): Likewise.
99 (FUNCTION_PROFILER): Likewise.
100 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
102 2013-09-30 Iain Sandoe <iain@codesourcery.com>
104 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
105 calls and defines in TARGET_MACHO conditional.
106 (load_macho_picbase_di): Likewise.
107 (reload_macho_picbase): Likewise.
108 (reload_macho_picbase_si): Likewise.
109 (reload_macho_picbase_di): Likewise.
110 (nonlocal_goto_receiver): Likewise.
112 2013-09-30 Nick Clifton <nickc@redhat.com>
114 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
115 that use the MSP430X ISA.
116 (msp430_option_override): Scan -mmcu command line option for any
117 MCU name that supports the MSP430X ISA.
118 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
119 -mmcu options which enable the MSP430X ISA.
121 2013-09-30 Richard Biener <rguenther@suse.de>
124 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
125 before looking for setjmp-like calls.
127 2013-09-29 Iain Sandoe <iain@codesourcery.com>
130 * config/darwin-protos.h (machopic_get_function_picbase): New.
131 * config/darwin.c (machopic_get_function_picbase): New.
132 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
133 label for a new func. (load_macho_picbase_di): Likewise.
134 (reload_macho_picbase): New expand.
135 (reload_macho_picbase_si): New insn.
136 (reload_macho_picbase_di): New insn.
137 (nonlocal_goto_receiver): New define and split.
138 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
139 (unspecv enum): Add UNSPECV_NLGR.
141 2013-09-29 Iain Sandoe <iain@codesourcery.com>
143 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
144 that altivec registers are correctly sized on Darwin.
146 2013-09-29 Iain Sandoe <iain@codesourcery.com>
148 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
149 darwin-driver.o): Use COMPILE and POSTCOMPILE.
150 * config/x-darwin (host-darwin.o): Likewise.
151 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
152 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
153 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
155 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
157 * doc/invoke.texi: Fix usage of @tie{} command.
159 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
161 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
163 2013-09-28 Jan Hubicka <jh@suse.cz>
165 * config/i386/x86-tune.def: Add documentation for each of the options;
168 2013-09-28 Jan Hubicka <jh@suse.cz>
170 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
172 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
173 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
174 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
175 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
176 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
178 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
180 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
181 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
182 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
183 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
184 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
185 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
186 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
187 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
188 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
189 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
190 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
191 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
192 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
193 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
194 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
195 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
196 gimple.h, godump.c, graphite-clast-to-gimple.c,
197 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
198 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
199 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
200 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
201 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
202 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
203 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
204 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
205 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
206 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
207 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
208 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
209 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
210 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
211 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
212 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
213 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
214 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
215 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
216 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
217 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
218 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
219 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
220 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
221 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
222 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
223 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
224 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
225 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
226 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
227 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
228 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
229 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
230 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
231 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
232 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
233 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
234 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
235 whitespace before "(".
237 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
239 * expr.h (extract_bit_field): Remove packedp parameter.
240 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
241 from forward declaration.
242 (store_split_bit_field): Remove packedp arg from calls to
243 extract_fixed_bit_field.
244 (extract_bit_field_1): Remove packedp parameter and packedp
245 argument from recursive calls and calls to extract_fixed_bit_field.
246 (extract_bit_field): Remove packedp parameter and corresponding
247 arg to extract_bit_field_1.
248 (extract_fixed_bit_field): Remove packedp parameter. Remove code
250 (extract_split_bit_field): Remove packedp arg from call to
251 extract_fixed_bit_field.
252 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
253 (copy_blkmode_from_reg): Likewise.
254 (copy_blkmode_to_reg): Likewise.
255 (read_complex_part): Likewise.
256 (store_field): Likewise.
257 (expand_expr_real_1): Likewise.
258 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
259 to extract_bit_field.
260 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
261 call to extract_bit_field.
262 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
263 call to extract_bit_field.
264 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
265 and special packedp behavior from -fstrict-volatile-bitfields
268 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
270 * lra-eliminations.c (init_elim_table): Guard value_p.
272 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
274 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
275 DFmode, DImode, and SFmode in the upper VSX registers based on the
276 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
277 if -mpower8-vector. Combine -mvsx-timode handling with the rest
278 of the VSX register handling.
280 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
282 (zero_extendsidi2_lfiwzx): Add support for loading into the
283 Altivec registers with -mpower8-vector. Use wu/wv constraints to
284 only do VSX memory options on Altivec registers.
285 (extendsidi2_lfiwax): Likewise.
286 (extendsfdf2_fpr): Likewise.
287 (mov<mode>_hardfloat, SF/SD modes): Likewise.
288 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
289 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
290 (movdi_internal64): Likewise.
292 2013-09-27 Xinliang David Li <davidxl@google.com>
294 * opts.c (finish_options): Adjust parameters
295 according to vect cost model.
296 (common_handle_option): Set dynamic vect cost
298 targhooks.c (default_add_stmt_cost): Compute stmt cost
300 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
302 * tree-vectorizer.h (unlimited_cost_model): New function.
303 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
304 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
306 (vect_enhance_data_refs_alignment): Ditto.
307 * flag-types.h: New enum.
308 * common/config/i386/i386-common.c (ix86_option_init_struct):
309 No need to initialize vect_cost_model flag.
310 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
313 2013-09-27 Diego Novillo <dnovillo@google.com>
315 * gimple.h (enum ssa_mode): Remove.
317 2013-09-27 Paulo Matos <pmatos@broadcom.com>
320 * gcc.dg/pr58463.c: New test.
322 2013-09-27 Paulo Matos <pmatos@broadcom.com>
324 * cfgloop.h (number_of_loops): Fix typo in check for null.
326 2013-09-27 Jakub Jelinek <jakub@redhat.com>
329 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
330 are children of outermost saved_cfun's loop, and set it up to
331 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
332 if loop != loop0 and SESE region contains bbs that belong to loop0.
334 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
336 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
337 (binary_scale_code_p, get_base_term, get_index_term): New functions.
338 (set_address_segment, set_address_base, set_address_index)
339 (set_address_disp): Accept the argument unconditionally.
340 (baseness): Remove must_be_base_p and must_be_index_p checks.
341 (decompose_normal_address): Classify as much as possible in the
344 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
346 * cse.c (count_reg_usage): Handle INT_LIST.
347 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
348 * reginfo.c (reg_scan_mark_refs): Likewise.
349 * reload1.c (eliminate_regs_1): Likewise.
351 2013-09-27 Iain Sandoe <iain@codesourcery.com>
354 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
357 2013-09-27 Richard Biener <rguenther@suse.de>
359 PR tree-optimization/58459
360 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
361 restriction not propagating into loops.
363 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
365 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
366 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
367 * doc/tree-ssa.texi (Walking use-def chains): Delete.
369 2013-09-26 Richard Biener <rguenther@suse.de>
371 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
373 2013-09-26 Richard Biener <rguenther@suse.de>
375 * alias.h (component_uses_parent_alias_set): Rename to ...
376 (component_uses_parent_alias_set_from): ... this.
377 * alias.c (component_uses_parent_alias_set): Rename to ...
378 (component_uses_parent_alias_set_from): ... this and return
380 (reference_alias_ptr_type_1): Use the result from
381 component_uses_parent_alias_set_from instead of stripping
382 components one at a time.
383 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
385 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
387 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
388 Move prototypes to...
389 * tree-ssa-ter.h: New File. Move prototypes here.
390 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
391 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
392 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
393 from is_replaceable_p.
394 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
395 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
396 (process_replaceable): Use ter_is_replaceable_p.
397 (find_replaceable_in_bb): Use ter_is_replaceable_p.
398 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
399 newly refactored ssa_is_replaceable_p.
400 * cfgexpand.c: Include tree-outof-ssa.h.
401 * ssaexpand.h: Delete.
403 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
405 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
406 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
408 (create_gimple_tmp): Delete.
409 (get_expr_type, build_assign, build_type_cast): Move to...
410 * gimple-builder.c: New File.
411 (get_expr_type): Relocate from gimple.c.
412 (build_assign, build_type_cast): Change to only create ssanames.
413 * gimple.h: Move prototypes to...
414 * gimple-builder.h: New File. Here.
415 * tree-ssa.h: And here.
416 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
417 count_uses_and_derefs): Relocate from gimple.c.
418 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
419 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
420 * tree-ssa-math-opts (execute_cse_reciprocals): Use
421 gimple_replace_ssa_lhs.
422 * asan.c: Include gimple-builder.h.
423 * Makefile.in: Add gimple-builder.o.
425 2013-09-26 Richard Biener <rguenther@suse.de>
427 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
428 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
429 (loe_visit_block): Use gcc_checking_assert.
430 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
431 gimple_assign_ssa_name_copy_p.
432 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
435 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
437 * config/rs6000/t-rs6000 (rs6000.o): Remove.
438 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
440 2013-09-26 Richard Biener <rguenther@suse.de>
442 PR tree-optimization/58539
443 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
444 the fact that debug statements are not taking part in loop-closed
447 2013-09-26 Nick Clifton <nickc@redhat.com>
449 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
450 time warning message.
451 (msp430_print_operand_raw): Delete unused letter parameter.
452 (TARGET_PRINT_OPERAND_ADDRESS): Define.
453 (msp430_print_operand_address): New function.
454 (msp430_print_operand): Move address printing code from here to
456 * config/msp430/msp430.md (movsipsi2): Add comment in generated
458 (zero_extendpsisi2): Likewise.
459 (extendpsisi2): New pattern.
460 (andneghi3): New pattern.
462 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
464 * config/aarch64/aarch64.opt (mlra): New option.
465 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
466 (TARGET_LRA_P): Define.
468 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
470 * expr.c (expand_assignment): Remove obsolete comment.
472 2013-09-25 Jeff Law <law@redhat.com>
474 * tree-flow.h (thread_through_all_blocks): Prototype moved into
475 tree-ssa-threadupdate.h.
476 (register_jump_thread): Similarly.
477 * tree-ssa-threadupdate.h: New header file.
478 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
479 * tree-vrp.c: Likewise.
480 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
481 (thread_around_empty_blocks): Change type of path vector argument to
482 an edge,type pair from just an edge. Initialize both elements when
483 appending to a jump threading path. Tweak references to elements
485 (thread_across_edge): Similarly. Release memory for the elements
487 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
488 (dump_jump_thread_path): New function broken out from
489 register_jump_thread.
490 (register_jump_thread): Use dump_jump_thread_path. Change type of
491 path vector entries. Search the path for NULL edges and dump
492 the path if one is found. Tweak the conversion of path to 3-edge
493 form to use the block copy type information embedded in the path.
495 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
497 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
499 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
500 Vladimir Makarov <vmakarov@redhat.com>
502 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
503 from the least significant bit.
504 (strip_address_mutations): Add bitfield operations handling.
505 (must_be_index_p): Add shifting and rotate operations handling.
506 (set_address_base): Use must_be_base_p predicate.
507 (set_address_index): Use must_be_index_p predicate.
509 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
510 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
511 Sergey Lega <sergey.s.lega@intel.com>
512 Anna Tikhonova <anna.tikhonova@intel.com>
513 Ilya Tocar <ilya.tocar@intel.com>
514 Andrey Turetskiy <andrey.turetskiy@intel.com>
515 Ilya Verbin <ilya.verbin@intel.com>
516 Kirill Yukhin <kirill.yukhin@intel.com>
517 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
519 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
521 (ix86_expand_vector_move_misalign): Support new unaligned load and
522 stores and use new names.
523 (CODE_FOR_sse2_storedqu): Rename to ...
524 (CODE_FOR_sse2_storedquv16qi): ... this.
525 (CODE_FOR_sse2_loaddqu): Rename to ...
526 (CODE_FOR_sse2_loaddquv16qi): ... this.
527 (CODE_FOR_avx_loaddqu256): Rename to ...
528 (CODE_FOR_avx_loaddquv32qi): ... this.
529 (CODE_FOR_avx_storedqu256): Rename to ...
530 (CODE_FOR_avx_storedquv32qi): ... this.
531 * config/i386/i386.md (fpint_logic): New.
532 * config/i386/sse.md (VMOVE): Extend for AVX512.
536 (VI_UNALIGNED_LOADSTORE): Ditto.
537 (sse2_avx_avx512f): Ditto.
538 (sse2_avx2): Extend for AVX512.
539 (sse4_1_avx2): Ditto.
541 (sse): Extend for AVX512.
544 (avxsizesuffix): Ditto.
545 (sseintvecmode): Ditto.
547 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
548 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
549 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
550 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
551 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
552 (<sse2_avx_avx512f>_storedqu<mode): ... this.
553 (<sse>_movnt<mode>): Replace constraint "x" with "v".
554 (STORENT_MODE): Extend for AVX512.
555 (*absneg<mode>2): Replace constraint "x" with "v".
556 (*mul<mode>3): Ditto.
557 (*ieee_smin<mode>3): Ditto.
558 (*ieee_smax<mode>3): Ditto.
559 (avx_cmp<mode>3): Replace VF with VF_128_256.
560 (*<sse>_maskcmp<mode>3_comm): Ditto.
561 (<sse>_maskcmp<mode>3): Ditto.
562 (<sse>_andnot<mode>3): Extend for AVX512.
563 (<code><mode>3, anylogic): Replace VF with VF_128_256.
564 (<code><mode>3, fpint_logic): New.
565 (*<code><mode>3): Extend for AVX512.
566 (avx512flogicsuff): New.
567 (avx512f_<logic><mode>): Ditto.
568 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
570 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
571 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
572 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
573 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
574 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
575 (xop_vpermil2<mode>3): Ditto.
576 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
577 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
578 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
579 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
581 2013-09-25 Tom Tromey <tromey@redhat.com>
583 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
584 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
585 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
586 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
587 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
588 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
589 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
590 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
591 (GRAPHITE_HTAB_H): Remove.
593 2013-09-25 Tom Tromey <tromey@redhat.com>
595 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
597 2013-09-25 Tom Tromey <tromey@redhat.com>
599 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
601 2013-09-25 Tom Tromey <tromey@redhat.com>
603 * config/i386/t-i386 (i386.o): Remove.
604 (i386-c.o): Use COMPILE and POSTCOMPILE.
606 2013-09-25 Tom Tromey <tromey@redhat.com>
608 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
610 2013-09-25 Tom Tromey <tromey@redhat.com>
612 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
613 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
614 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
615 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
616 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
617 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
618 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
619 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
620 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
621 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
622 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
623 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
624 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
625 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
626 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
627 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
628 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
629 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
630 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
631 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
632 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
633 (tree-ssa-pre.o, tree-ssa-sccvn.o)
634 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
635 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
636 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
637 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
638 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
639 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
640 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
641 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
642 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
643 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
644 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
645 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
646 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
647 (graphite-clast-to-gimple.o, graphite-dependences.o)
648 (graphite-interchange.o, graphite-poly.o)
649 (graphite-scop-detection.o, graphite-sese-to-poly.o)
650 (graphite-optimize-isl.o, tree-vect-loop.o)
651 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
652 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
653 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
654 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
655 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
656 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
657 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
658 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
659 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
660 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
661 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
662 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
663 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
664 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
665 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
666 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
667 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
668 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
669 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
670 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
671 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
672 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
673 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
674 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
675 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
676 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
677 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
678 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
679 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
680 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
681 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
682 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
683 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
684 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
685 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
686 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
687 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
688 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
689 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
690 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
691 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
692 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
693 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
694 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
695 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
696 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
697 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
698 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
699 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
700 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
701 (gcov-dump.o): Remove.
702 (default-c.o): Use COMPILE and POSTCOMPILE.
703 (CFLAGS-gcc.o): New variable.
704 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
706 2013-09-25 Tom Tromey <tromey@redhat.com>
708 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
709 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
710 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
711 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
712 (c-family/c-pragma.o, c-family/c-pretty-print.o)
713 (c-family/c-semantics.o, c-family/c-ada-spec.o)
714 (c-family/array-notation-common.o, c-family/stub-objc.o)
715 (c-family/c-ubsan.o): Remove.
717 2013-09-25 Tom Tromey <tromey@redhat.com>
719 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
721 2013-09-25 Tom Tromey <tromey@redhat.com>
723 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
724 to add -DENABLE_SHARED_LIBGCC.
725 (gcc.o): Don't use subshell.
727 2013-09-25 Tom Tromey <tromey@redhat.com>
729 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
730 * configure.ac: Don't invoke AM_PROG_CC_C_O.
731 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
732 * configure, config.in: Rebuild.
734 2013-09-25 Tom Tromey <tromey@redhat.com>
736 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
737 (COMPILE, POSTCOMPILE): New variables.
738 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
739 (DEPFILES): New variable.
741 * configure.ac: Add checks for dependency checking.
742 * configure, aclocal.m4: Regenerate.
744 2013-09-25 Tom Tromey <tromey@redhat.com>
746 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
747 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
749 2013-09-25 Tom Tromey <tromey@redhat.com>
751 * Makefile.in (generated_files): Add options.h,
752 target-hooks-def.h, insn-opinit.h,
753 common/common-target-hooks-def.h, pass-instances.def,
754 c-family/c-target-hooks-def.h.
756 2013-09-25 Jeff Law <law@redhat.com>
758 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
759 than foo[foo.length () - 1] to access last member in a vec.
760 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
762 2013-09-25 Richard Biener <rguenther@suse.de>
765 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
767 2013-09-25 Jan Hubicka <jh@suse.cz>
769 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
772 2013-09-25 Marek Polacek <polacek@redhat.com>
775 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
776 when determining the type name.
778 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
780 * config/sh/sh.md: Fix formatting.
782 2013-09-24 Xinliang David Li <davidxl@google.com>
784 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
785 max peel iterations parameter.
786 * param.def: New parameter.
787 * doc/invoke.texi: Document New parameter.
789 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
791 * gimple-pretty-print.c: Various whitespace tweaks.
792 * tree-core.h: Likewise.
793 * tree-pretty-print.c: Likewise.
794 * tree-ssa-alias.c: Likewise.
795 * tree-ssa-copy.c: Likewise.
796 * tree-ssanames.c: Likewise.
797 * tree-ssanames.h: Likewise.
798 * tree-vrp.c: Likewise.
800 2013-09-24 Alan Modra <amodra@gmail.com>
804 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
805 for output operands that disallow regs. Don't use EXPAND_WRITE on
808 2013-09-24 Richard Biener <rguenther@suse.de>
811 * tree.c (reference_alias_ptr_type): Move ...
812 * alias.c (reference_alias_ptr_type): ... here and implement
813 in terms of the new reference_alias_ptr_type_1.
814 (ref_all_alias_ptr_type_p): New helper.
815 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
816 use ref_all_alias_ptr_type_p.
817 (get_deref_alias_set): Add flag_strict_aliasing check here.
818 (reference_alias_ptr_type_1): New function, split out from ...
819 (get_alias_set): ... here.
820 (alias_ptr_types_compatible_p): New function.
821 * alias.h (reference_alias_ptr_type): Declare.
822 (alias_ptr_types_compatible_p): Likewise.
823 * tree.h (reference_alias_ptr_type): Remove.
824 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
825 to compare MEM_REF alias types.
827 2013-09-24 Richard Biener <rguenther@suse.de>
829 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
831 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
833 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
834 reload helper function arrays into a single array reg_addr.
835 (reload_fpr_gpr): Likewise.
836 (reload_gpr_vsx): Likewise.
837 (reload_vsx_gpr): Likewise.
838 (struct rs6000_reg_addr): Likewise.
839 (reg_addr): Likewise.
840 (rs6000_debug_reg_global): Change rs6000_vector_reload,
841 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
842 (rs6000_init_hard_regno_mode_ok): Likewise.
843 (rs6000_secondary_reload_direct_move): Likewise.
844 (rs6000_secondary_reload): Likewise.
846 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
847 constraints: wu, ww, and wy. Repurpose wv constraint added during
848 power8 changes. Put wg constraint in alphabetical order.
850 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
851 for future work to add ISA 2.07 VSX single precision support.
852 (-mvsx-scalar-double): Change default from -1 to 1, update
853 documentation comment.
854 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
855 (-mupper-regs-df): New debug switch to control whether DF values
856 can go in the traditional Altivec registers.
857 (-mupper-regs-sf): New debug switch to control whether SF values
858 can go in the traditional Altivec registers.
860 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
862 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
863 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
864 Add new constraints, wu/ww/wy. Repurpose wv constraint.
865 (rs6000_debug_legitimate_address_p): Print if we are running
866 before, during, or after reload.
867 (rs6000_secondary_reload): Add a comment.
868 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
870 * config/rs6000/constraints.md (wa constraint): Sort w<x>
871 constraints. Update documentation string.
872 (wd constraint): Likewise.
873 (wf constraint): Likewise.
874 (wg constraint): Likewise.
875 (wn constraint): Likewise.
876 (ws constraint): Likewise.
877 (wt constraint): Likewise.
878 (wx constraint): Likewise.
879 (wz constraint): Likewise.
880 (wu constraint): New constraint for ISA 2.07 SFmode scalar
882 (ww constraint): Likewise.
883 (wy constraint): Likewise.
884 (wv constraint): Repurpose ISA 2.07 constraint that we not used in
885 the previous submissions.
886 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
888 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
890 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
891 (REG_BR_PROB): Say that the probability is stored in an int_list.
892 * reg-notes.def: Update commentary to mention INT_LIST.
893 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
895 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
896 * rtlanal.c (int_reg_note_p): New function.
897 (alloc_reg_note): Assert that the note does not have an int argument.
898 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
899 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
900 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
901 rather than an INSN_LIST. Handle INT_LIST.
902 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
903 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
904 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
905 Manipulate them as ints rather than rtxes.
906 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
907 * regmove.c (copy_src_to_dest): Likewise.
908 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
910 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
911 into the cases that need it.
912 * config/arm/arm.c (arm_unwind_emit): Likewise.
914 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
915 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
916 * loop-doloop.c (add_test, doloop_modify): Likewise.
917 * loop-unswitch.c (compare_and_jump_seq): Likewise.
918 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
919 * predict.c (combine_predictions_for_insn): Likewise.
920 * print-rtl.c (print_rtx): Handle INT_LIST.
921 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
922 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
923 * config/arm/arm.c (emit_unlikely_jump): Likewise.
924 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
925 (ix86_split_fp_branch, predict_jump): Likewise.
926 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
927 * config/sh/sh.c (expand_cbranchsi4): Likewise.
928 * config/spu/spu.c (ea_load_store_inline): Likewise.
930 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
931 value of a REG_BR_PROB note.
932 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
933 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
934 * emit-rtl.c (try_split): Likewise.
935 * predict.c (br_prob_note_reliable_p): Likewise.
936 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
937 * reorg.c (mostly_true_jump): Likewise.
938 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
939 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
940 * config/i386/i386.c (ix86_print_operand): Likewise.
941 * config/ia64/ia64.c (ia64_print_operand): Likewise.
942 * config/mmix/mmix.c (mmix_print_operand): Likewise.
943 * config/rs6000/rs6000.c (output_cbranch): Likewise.
944 * config/s390/s390.c (s390_expand_tbegin): Likewise.
945 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
946 * config/sparc/sparc.c (output_cbranch): Likewise.
947 * config/spu/spu.c (get_branch_target): Likewise.
948 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
949 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
951 2013-09-23 Jan Hubicka <jh@suse.cz>
953 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
955 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
956 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
958 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
959 (gimple_fold_call): Dump inconsistent devirtualizations; add
960 sanity check for type based devirtualizations.
961 * ipa-prop.c: Include ipa-utils.h
962 (ipa_intraprocedural_devirtualization): Add sanity check.
963 (try_make_edge_direct_virtual_call): Likewise.
965 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
967 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
968 assignment statements.
970 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
972 * gimple-pretty-print.c (dump_ssaname_info): New function.
973 (dump_gimple_phi): Call it.
974 (pp_gimple_stmt_1): Likewise.
975 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
976 (range_info_def): Declare.
977 * tree-pretty-print.c (pp_double_int): New function.
978 (dump_generic_node): Call it.
979 * tree-pretty-print.h (pp_double_int): Declare.
980 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
981 * tree-ssanames.h (range_info_def): New structure.
982 (value_range_type): Move definition here.
983 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
985 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
987 (set_range_info): New function.
988 (get_range_info): Likewise.
989 (duplicate_ssa_name_range_info): Likewise.
990 (duplicate_ssa_name_fn): Check pointer type and call
991 duplicate_ssa_name_range_info.
992 * tree-ssa-copy.c (fini_copy_prop): Likewise.
993 * tree-vrp.c (value_range_type): Remove definition, now in
995 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
996 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
997 (SSA_NAME_RANGE_INFO): New macro.
999 2013-09-23 Richard Biener <rguenther@suse.de>
1001 PR tree-optimization/58464
1002 * tree-ssa-pre.c (phi_trans_lookup): Remove.
1003 (phi_trans_add): Change to add conditionally on being not present.
1004 (phi_translate_1): Remove recursion detection here.
1005 (phi_translate): Pre-seed the cache with NULL to catch
1006 recursion here in a more generic way.
1007 (bitmap_find_leader): Adjust comment.
1008 (get_representative_for): Dump value-numbers.
1009 (create_expression_by_pieces): Likewise.
1010 (insert_into_preds_of_block): Likewise.
1012 2013-09-23 Christian Bruel <christian.bruel@st.com>
1015 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
1016 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
1018 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
1021 2013-09-20 Renlin Li <renlin.li@arm.com>
1023 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
1025 (aarch64_expand_epilogue): Likewise.
1026 (aarch64_legitimize_reload_address): Likewise.
1028 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
1030 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
1031 NULL_TREE before pushing them onto the vector. Likewise for labels.
1033 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
1035 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
1036 comparison operators when -fno-trapping-math is in effect.
1037 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
1038 comparison operators in TFmode and assert that unsupported operators
1040 (ia64_print_operand): Likewise.
1042 2013-09-21 Jan Hubicka <jh@suse.cz>
1044 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
1045 (sse_typeless_stores): Enable for core
1046 (sse_load0_by_pxor): Likewise.
1047 (four_jump_limit): Disable for core.
1048 (pad_returns): Likewise.
1049 (avoid_vector_decode): Likewise.
1050 (fuse_cmp_and_branch): Enable for cores.
1051 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
1053 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
1056 * config/pa/pa.c (pa_option_override): Disable auto increment and
1057 decrement instructions until reload is completed.
1059 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
1060 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
1061 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
1063 2013-09-20 DJ Delorie <dj@redhat.com>
1064 Nick Clifton <nickc@redhat.com>
1066 * config/rl78/rl78.c: Various whitespace and comment tweaks.
1067 (need_to_save): Save bank 0 on interrupts.
1068 (characterize_address): Strip far address wrappers.
1069 (rl78_as_legitimate_address): Likewise.
1070 (transcode_memory_rtx): Likewise.
1071 (rl78_peep_movhi_p): Disable this peephole after devirt.
1072 (rl78_propogate_register_origins): Forget all origins when a
1074 * config/rl78/rl78-virt.md: Various whitespace tweaks.
1075 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
1077 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
1078 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
1079 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
1081 * config/rl78/constraints.md: Various whitespace and paren tweaks.
1083 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
1085 * config/pa/pa.md: In "scc" insn patterns, change output template to
1086 handle const0_rtx in reg_or_0_operand operands.
1088 2013-09-20 Martin Husemann <martin@NetBSD.org>
1091 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
1092 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
1094 2013-09-20 Richard Biener <rguenther@suse.de>
1097 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
1098 remembering SSA name version and block index.
1099 (new_scev_info_str): Adjust.
1100 (hash_scev_info): Likewise. Also hash the block index.
1101 (eq_scev_info): Adjust.
1102 (find_var_scev_info): Likewise.
1103 (struct instantiate_cache_entry): Remove.
1104 (struct instantiate_cache_type): Use a htab to map name, block
1106 (instantiate_cache_type::~instantiate_cache_type): Adjust.
1107 (get_instantiated_value_entry): Likewise.
1108 (hash_idx_scev_info, eq_idx_scev_info): New functions.
1109 (instantiate_scev_name): Adjust.
1111 2013-09-20 Jeff Law <law@redhat.com>
1113 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
1115 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
1117 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
1118 Call aarch64_simd_expand_args to update op[argc].
1120 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
1122 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
1125 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
1127 * gengtype.c (file_rules): Added rule for *.cc files.
1128 (get_output_file_with_visibility): Give fatal message when no
1131 2013-09-20 Renlin Li <renlin.li@arm.com>
1133 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
1134 (aarch64_expand_epilogue): Likewise.
1135 (aarch64_legitimize_reload_address): Likewise.
1137 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
1140 * expr.c (expand_assignment): Remove misalignp code path.
1142 2013-09-20 Marek Polacek <polacek@redhat.com>
1145 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
1146 TYPE_PRECISION. Add asserts.
1148 2013-09-20 Richard Biener <rguenther@suse.de>
1150 PR tree-optimization/58453
1151 * tree-loop-distribution.c (distribute_loop): Apply the cost
1152 model for -ftree-loop-distribute-patterns, too.
1154 2013-09-20 Richard Biener <rguenther@suse.de>
1157 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
1158 make type comparison less strict.
1160 2013-09-20 Alan Modra <amodra@gmail.com>
1162 * configure: Regenerate.
1163 * aclocal.m4: Regenerate.
1165 2013-09-20 Marek Polacek <polacek@redhat.com>
1168 * doc/extend.texi: Document that attribute used is meant to be used
1169 on variables with static storage duration.
1171 2013-09-19 Jakub Jelinek <jakub@redhat.com>
1173 PR tree-optimization/58472
1174 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
1175 simd_lane_access set inv_p = false.
1176 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
1177 the simduid magic VAR_DECL.
1179 2013-09-19 Jan Hubicka <jh@suse.cz>
1181 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
1183 2013-09-17 Jeff Law <law@redhat.com>
1185 * tree-ssa-dom.c (record_temporary_equivalences): New function
1186 split out of dom_opt_dom_walker::after_dom_children.
1187 (dom_opt_dom_walker::thread_across_edge): Move common code
1188 in here from dom_opt_dom_walker::after_dom_children.
1189 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
1191 2013-09-19 Jan Hubicka <jh@suse.cz>
1193 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
1194 (TARGET_GENERIC): Use PROCESOR_GENERIC
1195 (enum processor_type): Unify generic32 and 64.
1196 * i386.md (cpu): Likewise.
1197 * x86-tune.def (use_leave): Enable for generic32.
1198 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
1199 * athlon.md: Change generic64 to generic in all occurences.
1200 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
1201 (ix86_target_macros_internal): Likewise.
1202 * driver-i386.c (host_detect_local_cpu): Likewise.
1203 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
1205 (generic_memcpy, generic_memset, generic_cost): This one.
1206 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
1207 (m_GENERIC32, m_GENERIC64): Remove.
1208 (m_GENERIC): Turn into one flag.
1209 (processor_target): Unify generic tunnings.
1210 (ix86_option_override_internal): Replace generic32/64 by generic.
1211 (ix86_issue_rate): Likewise.
1212 (ix86_adjust_cost): Likewise.
1214 2013-09-19 Jan Hubicka <jh@suse.cz>
1216 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
1217 of speculative flag.
1219 2013-09-19 Jakub Jelinek <jakub@redhat.com>
1221 * omp-low.c (expand_omp_sections): Always pass len - 1 to
1222 GOMP_sections_start, even if !exit_reachable.
1224 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
1226 * lra-constraints.c (need_for_all_save_p): Use macro
1227 HARD_REGNO_CALL_PART_CLOBBERED.
1228 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
1229 set up pseudo conflict hard regs.
1231 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
1234 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
1237 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
1239 PR rtl-optimization/58438
1240 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
1241 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
1242 from previous subpasses.
1244 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
1246 * arm.c (arm_get_frame_offsets): Validate architecture supports
1247 LDRD/STRD before accepting the tuning preference.
1248 (arm_expand_prologue): Likewise.
1249 (arm_expand_epilogue): Likewise.
1251 2013-09-18 Richard Biener <rguenther@suse.de>
1253 PR tree-optimization/58417
1254 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
1255 have chrecs with symbols defined in the loop as operands.
1256 (chrec_fold_multiply): Likewise.
1257 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
1258 parameters before folding binary operations.
1259 (struct instantiate_cache_entry_hasher): Remove.
1260 (struct instantiate_cache_type): Use a pointer-map.
1261 (instantiate_cache_type::instantiate_cache_type): New function.
1262 (instantiate_cache_type::get): Likewise.
1263 (instantiate_cache_type::set): Likewise.
1264 (instantiate_cache_type::~instantiate_cache_type): Adjust.
1265 (get_instantiated_value_entry): Likewise.
1266 (global_cache): New global.
1267 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
1268 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
1269 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
1270 (instantiate_scev_name): Adjust.
1271 (instantiate_scev): Construct global instead of local cache.
1272 (resolve_mixers): Likewise.
1274 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
1275 Paolo Carlini <paolo.carlini@oracle.com>
1278 * doc/implement-cxx.texi: Fix references to the C++ standards.
1280 2013-09-18 Jakub Jelinek <jakub@redhat.com>
1282 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
1283 * tree-vect-data-refs.c (vect_analyze_data_refs): For
1284 simd_lane_access drs, update also DR_ALIGNED_TO.
1286 2013-09-18 Marek Polacek <polacek@redhat.com>
1289 * doc/extend.texi: Document no_sanitize_undefined attribute.
1290 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
1291 no_sanitize_undefined attribute.
1293 2013-09-18 Nick Clifton <nickc@redhat.com>
1295 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
1296 (ASM_DECLARE_FUNCTION_NAME): Define.
1298 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
1300 * compare-elim.c (find_comparison_dom_walker): New class
1301 (find_comparisons_in_bb): Rename to
1302 find_comparison_dom_walker::before_dom_children
1303 (find_comparisons): Adjust
1304 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
1306 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
1307 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
1308 (init_walk_dominator_tree): Remove declaration.
1309 (fini_walk_dominator_tree): Remove declaration.
1310 * fwprop.c (single_def_use_dom_walker): New class
1311 (single_def_use_enter_block): Convert to
1312 single_def_use_dom_walker::before_dom_children.
1313 (single_def_use_leave_block): Convert to
1314 single_def_use_dom_walker::after_dom_children.
1315 (build_single_def_use_links): Adjust.
1316 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
1318 (find_candidates_in_block): Convert to
1319 find_candidates_dom_walker::before_dom_children.
1320 (execute_strength_reduction): Adjust.
1321 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
1322 (sese_dom_walker): New class.
1323 (sese_dom_walker::sese_dom_walker): New constructor.
1324 (sese_dom_walker::~sese_dom_walker): New destructor.
1325 (build_sese_conditions_before): Convert to
1326 sese_dom_walker::before_dom_children.
1327 (build_sese_conditions_after): Convert to
1328 sese_dom_walker::after_dom_children.
1329 (build_poly_scop): Adjust
1330 * tree-into-ssa.c (rewrite_dom_walker): New class
1331 (rewrite_enter_block): Convert to
1332 rewrite_dom_walker::before_dom_children.
1333 (rewrite_leave_block): Convert to
1334 rewrite_dom_walker::after_dom_children.
1335 (rewrite_update_dom_walker): New class.
1336 (rewrite_update_enter_block): Convert to
1337 rewrite_update_dom_walker::before_dom_children.
1338 (rewrite_update_leave_block): Convert to
1339 rewrite_update_dom_walker::after_dom_children.
1340 (rewrite_blocks, rewrite_into_ssa): Adjust.
1341 (mark_def_dom_walker): New class.
1342 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
1343 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
1344 (mark_def_sites_blocks): Convert to
1345 mark_def_dom_walker::before_dom_children.
1346 (mark_def_site_blocks): Remove.
1347 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
1348 (tree_ssa_dominator_optimize): Adjust.
1349 (dom_thread_across_edge): Convert to method
1350 dom_opt_dom_walker::thread_across_edge.
1351 (dom_opt_enter_block): Convert to member function
1352 dom_opt_dom_walker::before_dom_children.
1353 (dom_opt_leave_block): Convert to member function
1354 dom_opt_dom_walker::after_dom_children.
1355 * tree-ssa-dse.c (dse_dom_walker): New class.
1356 (dse_enter_block): Convert to member function
1357 dse_dom_walker::before_dom_children.
1358 (tree_ssa_dse): Adjust.
1359 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
1360 (determine_invariantness_stmt): Convert to method
1361 invariantness_dom_walker::before_dom_children.
1362 (determine_invariantness): Remove
1363 (move_computations_dom_walker): New class.
1364 (move_computations_stmt): Convert to method
1365 move_computations_dom_walker::before_dom_children.
1366 (move_computations, tree_ssa_lim): Adjust.
1367 * tree-ssa-phiopt.c (nontrapping_dom_walker): new class
1368 (nt_init_block): Make method
1369 notrappping_dom_walker::before_dom_children.
1370 (nt_fini_block): Make
1371 method nontrapping_dom_walker::after_dom_children.
1372 (get_non_trapping): Adjust.
1373 * tree-ssa-pre.c (eliminate_dom_walker): New class.
1374 (eliminate_bb): Make method eliminate_dom_walker::before_dom_children.
1375 (eliminate_leave_block): Make method.
1376 eliminate_dom_walker::after_dom_children.
1378 * tree-ssa-strlen.c (strlen_dom_walker): New class.
1379 (strlen_enter_block): Make method
1380 strlen_dom_walker::before_dom_children.
1381 (strlen_leave_block): Make
1382 method strlen_dom_walker::after_dom_children.
1383 (tree_ssa_strlen): Adjust.
1384 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
1385 (tree_ssa_uncprop): Adjust.
1386 (uncprop_leave_block): Make method
1387 uncprop_dom_walker::after_dom_children.
1388 (uncprop_leave_block): Make method
1389 uncprop_dom_walker::before_dom_children.
1391 2013-09-18 Bin Cheng <bin.cheng@arm.com>
1393 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
1394 branch in same basic block. Check both src and dest of the move insn.
1396 2013-09-17 Nick Clifton <nickc@redhat.com>
1398 * config/rl78/rl78-real.md (bf): New pattern.
1400 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
1401 (rl78_print_operand): Do not put a # before a %B.
1402 * config/rl78/rl78.opt: Tweak doc strings.
1404 2013-09-17 DJ Delorie <dj@redhat.com>
1406 * config/rl78/constraints.md (Wcv): Allow up to $r31.
1407 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
1408 (rl78_option_override): Likewise, if -mallregs.
1409 (is_virtual_register): Likewise.
1410 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
1411 (REGNO_OK_FOR_BASE_P): Likewise.
1412 * config/rl78/rl78.opt (-mallregs): New.
1414 2013-09-17 Nick Clifton <nickc@redhat.com>
1416 * config/rl78/rl78.c (need_to_save): Change return type to bool.
1417 For interrupt functions: save all call clobbered registers if the
1418 interrupt handler is not a leaf function.
1419 (rl78_expand_prologue): Always recompute the frame information.
1420 For interrupt functions: only select bank 0 if one of the bank 0
1421 registers is going to be psuhed.
1423 2013-09-17 DJ Delorie <dj@redhat.com>
1425 * config/rl78/constraints.md: For each W* constraint, rename to C*
1426 and create a W* constraint that checks for an optional ES: prefix
1428 * config/rl78/rl78.md (UNS_ES_ADDR): New.
1429 (es_addr): New. Used to wrap far addresses.
1430 * config/rl78/rl78-protos.h (rl78_es_addr): New.
1431 (rl78_es_base): New.
1432 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
1433 wrapped far addresses.
1434 (rl78_print_operand_1): Unwrap far addresses before processing.
1435 (rl78_lo16): Wrap far addresses in unspecs.
1436 (rl78_es_addr): New.
1437 (rl78_es_base): New.
1438 (insn_ok_now): Check for not-yet-wrapped far addresses.
1439 (transcode_memory_rtx): Properly re-wrap far addresses.
1441 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
1443 * config/sparc/t-rtems: Add leon3 multilibs.
1445 2013-09-17 Cong Hou <congh@google.com>
1447 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
1448 when checking the dot production pattern. The type of rhs operand
1449 of multiply is now checked correctly.
1451 2013-09-17 Jeff Law <law@redhat.com>
1453 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
1454 edge implied equivalences into successor phis.
1455 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
1456 here from tree-ssa-threadedge.c.
1457 (mark_threaded_blocks): When threading through a joiner, if both
1458 successors of the joiner's clone reach the same block, verify the
1459 PHI arguments are equal. If not, cancel the jump threading request.
1460 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
1461 tree-ssa-threadupdate.c
1462 (thread_across_edge): Don't check PHI argument equality when
1463 threading through joiner block here.
1465 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
1467 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
1468 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
1469 (warn_uninit, warn_uninitialized_vars,
1470 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
1471 Move to tree-ssa-uninit.c.
1472 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c
1473 (has_undefined_value_p): New. Pass dependant parts of
1474 ssa_undefined_value_p.
1475 (uninit_undefined_value_p): Use has_undefined_value_p.
1476 (warn_uninit, warn_uninitialized_vars,
1477 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
1478 Move from tree-ssa.c.
1479 * tree-ssa.h: Adjust prototypes.
1481 2013-09-17 Jan Hubicka <jh@suse.cz>
1484 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
1485 * ipa-inline.c (can_inline_edge_p): Do not downgrade
1486 FUNCTION_NOT_OPTIMIZED.
1487 * ipa-inline-analysis.c (compute_inline_parameters): Function
1488 not optimized is not inlinable unless it is alwaysinline.
1489 (inline_analyze_function): Force calls in not optimized
1490 function not inlinable.
1492 2013-09-17 Jan Hubicka <jh@suse.cz>
1495 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
1497 * ipa.c (function_and_variable_visibility): Likewise.
1498 * ipa-profile.c (ipa_profile): Likewise.
1500 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
1503 * cgraph.c (cgraph_function_body_availability): Check for ifunc
1504 attribute, and don't inline the resolver in this case.
1506 2013-09-17 Teresa Johnson <tejohnson@google.com>
1508 * coverage.c (get_coverage_counts): Add missing newline.
1510 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1512 PR tree-optimization/58088
1513 * fold-const.c (mask_with_trailing_zeros): New function.
1514 (fold_binary_loc): Make sure we don't recurse infinitely
1515 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
1516 Use mask_with_trailing_zeros where appropriate.
1518 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
1520 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
1521 of 'prev' var to get better distance estimation.
1523 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
1525 * tree-inline.h (struct copy_body_data): Add transform_parameter.
1526 * tree-inline.c (is_parameter_of): New predicate.
1527 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
1528 a parameter has been remapped.
1529 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
1530 (optimize_inline_calls): Initialize transform_parameter.
1531 (copy_gimple_seq_and_replace_locals): Likewise.
1532 (tree_function_versioning): Likewise.
1533 (maybe_inline_call_in_expr): Likewise.
1535 2013-09-17 Nick Clifton <nickc@redhat.com>
1537 * config/msp430/msp430-protos.h: Add prototypes for new functions.
1538 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
1540 (is_attr_func): New function.
1541 (msp430_is_interrupt_func): New function.
1542 (is_naked_func): New function.
1543 (is_reentrant_func): New function.
1544 (is_critical_func): New function.
1545 (msp430_start_function): Add annotations for function attributes.
1546 (msp430_attr): New function.
1547 (msp430_attribute_table): New.
1548 (msp430_function_section): New function.
1549 (TARGET_ASM_FUNCTION_SECTION): Define.
1550 (msp430_builtin): New enum.
1551 (msp430_init_builtins): New function.
1552 (msp430_builtin_devl): New function.
1553 (msp430_expand_builtin): New function.
1554 (TARGET_INIT_BUILTINS): Define.
1555 (TARGET_EXPAND_BUILTINS): Define.
1556 (TARGET_BUILTIN_DECL): Define.
1557 (msp430_expand_prologue): Add support for naked, interrupt,
1558 critical and reentrant functions.
1559 (msp430_expand_epilogue): Likewise.
1560 (msp430_print_operand): Handle 'O' character.
1561 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
1563 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
1564 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
1565 (pushm): Use a 'n' rather than an 'i' constraint.
1566 (msp_return): Add generation of the interrupt return instruction.
1567 (disable_interrupts): New pattern.
1568 (enable_interrupts): New pattern.
1569 (push_intr_state): New pattern.
1570 (pop_intr_state): New pattern.
1571 (bic_SR): New pattern.
1572 (bis_SR): New pattern.
1573 * doc/extend.texi: Document MSP430 function attributes and builtin
1576 2013-09-17 Richard Biener <rguenther@suse.de>
1578 PR tree-optimization/58432
1579 * tree-loop-distribution.c (tree_loop_distribution): Also
1580 scan PHIs for outside loop uses and seed a partition from them.
1582 2013-09-17 Bin Cheng <bin.cheng@arm.com>
1584 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
1585 (restructure_reference): Call backtrace_base_for_ref.
1587 2013-09-17 Alan Modra <amodra@gmail.com>
1590 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
1593 2013-09-16 DJ Delorie <dj@redhat.com>
1595 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
1596 vregs location for RL78/G10.
1597 (rl78_expand_prologue): Avoid SEL on G10.
1598 (rl78_expand_epilogue): Likewise.
1599 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
1600 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
1601 __RL78_G10__ when appropriate.
1602 (ASM_SPEC): Pass -mg10 along to the assembler.
1603 * config/rl78/rl78.md (sel_rb): Disable for G10.
1604 * config/rl78/rl78.opt: Add -mg10 option.
1605 * config/rl78/t-rl78: Add -mg10 multilib.
1607 2013-09-16 Xinliang David Li <davidxl@google.com>
1609 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
1610 * omp-low.c (omp_max_vf): Ditto.
1611 (expand_omp_simd): Ditto.
1612 * tree-vectorizer.c (vectorize_loops): Ditto.
1613 (gate_vect_slp): Ditto.
1614 (gate_increase_alignment): Ditto.
1615 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
1616 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
1617 (gate_tree_vectorize): Name change.
1618 (tree_vectorize): Ditto.
1619 (pass_vectorize::gate): Call new function.
1620 (pass_vectorize::execute): Ditto.
1621 * opts.c: O3 default setting change.
1622 (finish_options): Check new flag.
1623 * doc/invoke.texi: Document new flags.
1624 * common.opt: New flags.
1626 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
1628 * doc/tm.texi.in (Cond Exec Macros): Remove node.
1629 (Condition Code): Don't reference it.
1630 * doc/tm.texi: Regenerate.
1632 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
1635 * lra-constraints.c (undo_optional_reloads): Consider all optional
1636 reload even if it did not get a hard reg.
1638 2013-09-16 Teresa Johnson <tejohnson@google.com>
1640 * dumpfile.c (dump_loc): Remove newline emission.
1641 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
1642 emission to dump_printf_loc calls where missing.
1643 (vect_mark_for_runtime_alias_test): Ditto.
1644 (vect_analyze_data_ref_dependence): Ditto.
1645 (vect_analyze_data_ref_dependences): Ditto.
1646 (vect_slp_analyze_data_ref_dependence): Ditto.
1647 (vect_slp_analyze_data_ref_dependences): Ditto.
1648 (vect_compute_data_ref_alignment): Ditto.
1649 (vect_update_misalignment_for_peel): Ditto.
1650 (vect_verify_datarefs_alignment): Ditto.
1651 (vector_alignment_reachable_p): Ditto.
1652 (vect_get_data_access_cost): Ditto.
1653 (vect_enhance_data_refs_alignment): Ditto.
1654 (vect_find_same_alignment_drs): Ditto.
1655 (vect_analyze_data_refs_alignment): Ditto.
1656 (vect_analyze_group_access): Ditto.
1657 (vect_analyze_data_ref_access): Ditto.
1658 (vect_analyze_data_ref_accesses): Ditto.
1659 (vect_prune_runtime_alias_test_list): Ditto.
1660 (vect_analyze_data_refs): Ditto.
1661 (vect_create_addr_base_for_vector_ref): Ditto.
1662 (vect_create_data_ref_ptr): Ditto.
1663 (vect_grouped_store_supported): Ditto.
1664 (vect_grouped_load_supported): Ditto.
1665 * value-prof.c (check_counter): Ditto.
1666 (check_ic_target): Ditto.
1667 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
1668 (vect_recog_widen_mult_pattern): Ditto.
1669 (vect_recog_widen_sum_pattern): Ditto.
1670 (vect_recog_over_widening_pattern): Ditto.
1671 (vect_recog_widen_shift_pattern): Ditto.
1672 (vect_recog_rotate_pattern): Ditto.
1673 (vect_recog_vector_vector_shift_pattern): Ditto.
1674 (vect_recog_divmod_pattern): Ditto.
1675 (vect_recog_mixed_size_cond_pattern): Ditto.
1676 (vect_recog_bool_pattern): Ditto.
1677 (vect_pattern_recog_1): Ditto.
1678 (vect_pattern_recog): Ditto.
1679 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
1680 (vect_is_simple_iv_evolution): Ditto.
1681 (vect_analyze_scalar_cycles_1): Ditto.
1682 (vect_get_loop_niters): Ditto.
1683 (vect_analyze_loop_1): Ditto.
1684 (vect_analyze_loop_form): Ditto.
1685 (vect_analyze_loop_operations): Ditto.
1686 (vect_analyze_loop_2): Ditto.
1687 (vect_analyze_loop): Ditto.
1688 (report_vect_op): Ditto.
1689 (vect_is_slp_reduction): Ditto.
1690 (vect_is_simple_reduction_1): Ditto.
1691 (vect_get_known_peeling_cost): Ditto.
1692 (vect_estimate_min_profitable_iters): Ditto.
1693 (vect_model_reduction_cost): Ditto.
1694 (vect_model_induction_cost): Ditto.
1695 (get_initial_def_for_induction): Ditto.
1696 (vect_create_epilog_for_reduction): Ditto.
1697 (vectorizable_reduction): Ditto.
1698 (vectorizable_induction): Ditto.
1699 (vectorizable_live_operation): Ditto.
1700 (vect_loop_kill_debug_uses): Ditto.
1701 (vect_transform_loop): Ditto.
1702 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
1703 (vect_stmt_relevant_p): Ditto.
1704 (process_use): Ditto.
1705 (vect_mark_stmts_to_be_vectorized): Ditto.
1706 (vect_model_simple_cost): Ditto.
1707 (vect_model_promotion_demotion_cost): Ditto.
1708 (vect_model_store_cost): Ditto.
1709 (vect_get_store_cost): Ditto.
1710 (vect_model_load_cost): Ditto.
1711 (vect_get_load_cost): Ditto.
1712 (vect_init_vector_1): Ditto.
1713 (vect_get_vec_def_for_operand): Ditto.
1714 (vect_finish_stmt_generation): Ditto.
1715 (vectorizable_call): Ditto.
1716 (vectorizable_conversion): Ditto.
1717 (vectorizable_assignment): Ditto.
1718 (vectorizable_shift): Ditto.
1719 (vectorizable_operation): Ditto.
1720 (vectorizable_store): Ditto.
1721 (vectorizable_load): Ditto.
1722 (vectorizable_condition): Ditto.
1723 (vect_analyze_stmt): Ditto.
1724 (vect_transform_stmt): Ditto.
1725 (vect_is_simple_use): Ditto.
1726 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
1727 (vect_can_advance_ivs_p): Ditto.
1728 (vect_update_ivs_after_vectorizer): Ditto.
1729 (vect_do_peeling_for_loop_bound): Ditto.
1730 (vect_gen_niters_for_prolog_loop): Ditto.
1731 (vect_update_inits_of_drs): Ditto.
1732 (vect_create_cond_for_alias_checks): Ditto.
1733 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
1734 (vect_build_slp_tree_1): Ditto.
1735 (vect_supported_load_permutation_p): Ditto.
1736 (vect_analyze_slp_instance): Ditto.
1737 (vect_analyze_slp): Ditto.
1738 (vect_make_slp_decision): Ditto.
1739 (vect_detect_hybrid_slp): Ditto.
1740 (vect_bb_vectorization_profitable_p): Ditto.
1741 (vect_slp_analyze_bb_1): Ditto.
1742 (vect_update_slp_costs_according_to_vf): Ditto.
1743 (vect_get_mask_element): Ditto.
1744 (vect_transform_slp_perm_load): Ditto.
1745 (vect_schedule_slp_instance): Ditto.
1746 (vect_schedule_slp): Ditto.
1747 (vect_slp_transform_bb): Ditto.
1748 * profile.c (read_profile_edge_counts): Ditto.
1749 (compute_branch_probabilities): Ditto.
1750 * coverage.c (get_coverage_counts): Ditto.
1752 2013-09-16 Diego Novillo <dnovillo@google.com>
1754 * tree-core.h: Add missing comment lines from refactoring of tree.h.
1756 2013-09-16 Jan Hubicka <jh@suse.cz>
1758 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
1759 abstract functions; for static functions check the presence of body.
1761 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
1763 * config/aarch64/aarch64-simd-builtins.def (fma): New.
1764 * config/aarch64/aarch64-simd.md
1765 (aarch64_mla_elt<mode>): New.
1766 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
1767 (aarch64_mls_elt<mode>): Likewise.
1768 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
1769 (aarch64_fma4_elt<mode>): Likewise.
1770 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
1771 (aarch64_fma4_elt_to_128v2df): Likewise.
1772 (aarch64_fma4_elt_to_64df): Likewise.
1773 (fnma<mode>4): Likewise.
1774 (aarch64_fnma4_elt<mode>): Likewise.
1775 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
1776 (aarch64_fnma4_elt_to_128v2df): Likewise.
1777 (aarch64_fnma4_elt_to_64df): Likewise.
1778 * config/aarch64/iterators.md (VDQSF): New.
1779 * config/aarch64/arm_neon.h
1780 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
1781 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
1783 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
1785 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
1786 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
1787 (aarch64_mul3_elt_to_128df): Likewise.
1788 (aarch64_mul3_elt_to_64v2df): Likewise.
1789 * config/aarch64/iterators.md (VEL): Also handle DFmode.
1791 (VMUL_CHANGE_NLANES) Likewise.
1794 * config/aarch64/arm_neon.h
1795 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
1797 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
1799 * config/aarch64/arm_neon.h
1800 (vcvtx_high_f32_f64): Fix parameters.
1802 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1803 Uros Bizjak <ubizjak@gmail.com>
1805 * config/alpha.c: Include tree-ssanames.h.
1807 2013-09-16 Richard Biener <rguenther@suse.de>
1809 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
1810 (dot_rdg_1): Handle control_dd.
1811 (create_edge_for_control_dependence): New function.
1812 (create_rdg_edges): Add control dependences if asked for.
1813 (build_rdg): Likewise.
1814 (generate_loops_for_partition): If there are not necessary
1815 control stmts remove all their dependencies.
1816 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
1817 (distribute_loop): Pass on control dependences.
1818 (tree_loop_distribution): Compute control dependences and remove
1819 restriction on number of loop nodes.
1821 2013-09-16 Jakub Jelinek <jakub@redhat.com>
1823 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
1826 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
1828 * cse.c (try_const_anchors): Punt on CC modes.
1830 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1832 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
1834 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
1837 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
1840 2013-09-14 Iain Sandoe <iain@codesourcery.com>
1843 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
1845 (darwin_objc1_section): Likewise.
1846 (darwin_file_end): Emit Image Info section when required.
1848 2013-09-14 Jan Hubicka <jh@suse.cz>
1850 * tree-into-ssa.c (gate_into_ssa): New.
1851 (pass_data_build_ssa): Use it.
1852 * cgraph.h (expand_thunk): Update prototype.
1853 * cgraphunit.c (analyze_function): Expand thunks early.
1854 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
1855 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
1856 set lowered flag; do not add new function.
1857 (assemble_thunks_and_aliases): Update.
1858 * tree-ssa.c (gate_init_datastructures): New gate.
1859 (pass_data_init_datastructures): Use it.
1861 2013-09-14 Iain Sandoe <iain@codesourcery.com>
1864 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
1865 xmm register set described in the psABI.
1867 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
1869 * dwarf2out.c (should_emit_struct_debug): Add check
1870 for type_decl variable is not NULL.
1872 2013-09-13 Jacek Caban <jacek@codeweavers.com>
1874 * config.gcc: Use new winnt-c.c target hooks
1875 * config/t-winnt: New file
1876 * config/winnt-c.c: New file
1877 * doc/tm.texi.in: Document new hook
1878 * doc/tm.texi: Regenerated
1880 2013-09-13 Jan Hubicka <jh@suse.cz>
1883 * ipa-inline.c (check_callers): New function.
1884 (check_caller_edge): Remove.
1885 (want_inline_function_to_all_callers_p): Also permit alises that are
1887 (inline_to_all_callers): Terminate the walk when devirtualization
1888 introduce new calls.
1890 2013-09-13 Jan Hubicka <jh@suse.cz>
1892 * ipa-inline-analysis.c (struct growth_data): Add node.
1893 (do_estimate_growth_1): Fix detection of recursion.
1895 2013-09-13 Jakub Jelinek <jakub@redhat.com>
1897 PR tree-optimization/58392
1898 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
1899 to avoid shadowing of outer loop variable. If
1900 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
1901 replace_by_duplicate_decl simduid of loops that have it set and
1902 set dest_cfun->has_simduid_loops and/or
1903 dest_cfun->has_force_vect_loops.
1904 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
1905 instead of maybe_lookup_decl.
1906 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
1907 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
1908 call. Copy over force_vect and copy and remap simduid. Set
1909 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
1910 (copy_cfg_body): Remove blocks_to_copy argument. Use
1911 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
1912 caller. Don't set cfun->has_simduid_loops and/or
1913 cfun->has_force_vect_loops here.
1914 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
1916 (expand_call_inline, tree_function_versioning): Adjust copy_body
1919 2013-09-13 Martin Jambor <mjambor@suse.cz>
1922 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
1923 the assert if the edge was a speculative one.
1925 2013-09-13 Richard Biener <rguenther@suse.de>
1927 * tree-data-ref.h (known_dependences_p): Move here ...
1928 * tree-loop-distribution.c (known_dependences_p): ... from here.
1929 (dump_rdg_component, debug_rdg_component): Remove.
1931 (generate_loops_for_partition): Use gimple_uid instead of
1932 relying on matching stmt visit order.
1933 (rdg_build_partitions): Take starting stmt vector.
1934 (ldist_gen): Merge into ...
1935 (distribute_loop): ... this function. Do not compute starting
1937 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
1939 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1941 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
1942 Set type attribute correctly. Set predicable_short_it attribute.
1943 (cmpsi_shiftsi): Remove %? from output template.
1945 2013-09-13 Richard Biener <rguenther@suse.de>
1947 * tree-loop-distribution.c (struct rdg_component,
1948 rdg_defs_used_in_other_loops_p, free_rdg_components,
1949 rdg_build_components): Remove.
1950 (stmts_from_loop): Do not record virtual PHIs.
1951 (generate_loops_for_partition): Skip virtual PHIs.
1952 (build_rdg_partition_for_component): Rename to ...
1953 (build_rdg_partition_for_vertex): ... this and adjust.
1954 (rdg_build_partitions): Take a vector of starting vertices
1955 instead of components. Remove unnecessary leftover handling.
1956 (ldist_gen): Do not build components or record other stores.
1957 (distribute_loop): Do not distribute loops containing stmts
1960 2013-09-13 Christian Bruel <christian.bruel@st.com>
1963 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
1965 2013-09-13 Kai Tietz <ktietz@redhat.com>
1967 * config.gcc: Separate cases for mingw and cygwin targets,
1968 and add 64-bit cygwin target case.
1970 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
1971 dll-export inline-functions.
1972 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
1975 2013-09-13 Jeff Law <law@redhat.com>
1979 2013-09-06 Jeff Law <law@redhat.com>
1981 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
1982 edge implied equivalences into successor phis.
1984 2013-09-12 DJ Delorie <dj@redhat.com>
1986 * config/rl78/rl78-virt.md: Change from | to \; for asm line
1989 2013-09-12 Brooks Moses <bmoses@google.com>
1992 * Makefile.in: Do not install driver binaries in $(target)/bin.
1994 2013-09-12 DJ Delorie <dj@redhat.com>
1996 * config/rl78/rl78.opt (mrelax): New.
1997 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
1998 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
2000 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
2002 (rl78_expand_epilogue): Likewise.
2004 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
2007 * lra-eliminations.c (remove_reg_equal_offset_note): New.
2008 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
2009 pointer elimination with using remove_reg_equal_offset_note.
2011 2013-09-12 DJ Delorie <dj@redhat.com>
2013 * config/msp430/: New port.
2014 * config.gcc (msp430): Added.
2015 * doc/invoke.texi: Document MSP430 options.
2016 * doc/install.texi: Document msp430-elf
2017 * doc/md.texi: Document msp430-elf
2018 * doc/contrib.texi: Document msp430-elf
2020 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
2023 2013-09-12 Martin Jambor <mjambor@suse.cz>
2026 * ipa-prop.c (remove_described_reference): Give up if the edge in the
2027 reference descriptor is NULL.
2028 (ipa_edge_removal_hook): If owning a reference descriptor, set its
2031 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
2033 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
2034 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
2035 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
2036 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
2037 make_temp_ssa_name): move to tree-ssanames.h
2038 * tree-ssa-alias.h: Move prototype.
2039 * tree-ssa.h: Include tree-ssanames.h.
2040 * tree-ssanames.c (FREE_SSANAMES): Move to here.
2041 * tree-ssanames.h: New. Move items from tree-flow*.h
2042 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
2044 2013-09-12 Richard Biener <rguenther@suse.de>
2046 PR tree-optimization/58404
2047 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
2048 propagate non-invariant addresses into dereferences wrapped
2049 in component references.
2051 2013-09-12 Richard Biener <rguenther@suse.de>
2053 PR tree-optimization/58402
2054 * passes.def: Move pass_late_warn_uninitialized later.
2056 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
2058 * tree-ssa.h: New. Move content from tree-flow.h and
2060 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
2061 Move prototypes belonging to tree-ssa.c.
2062 * tree-flow-inline.h (redirect_edge_var_map_def,
2063 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
2065 * gimple.h: Adjust prototypes.
2066 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
2068 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
2069 * tree.h: Move prototype to tree-ssa.h.
2070 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
2071 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
2072 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
2073 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
2074 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
2075 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
2076 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
2077 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
2078 graphite-blocking.c, graphite-clast-to-gimple.c,
2079 graphite-dependences.c, graphite-interchange.c,
2080 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
2081 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
2082 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
2083 ipa-reference.c, ipa-split.c, ipa-utils.c,
2084 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
2085 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
2086 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
2087 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
2088 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
2089 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
2090 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
2091 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
2092 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
2093 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
2094 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
2095 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
2096 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
2097 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
2098 value-prof.c, var-tracking.c,
2099 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
2101 2013-09-12 Richard Biener <rguenther@suse.de>
2103 PR tree-optimization/58396
2104 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
2105 (build_rdg): Take a loop-nest parameter, fix memleaks.
2106 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
2108 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
2110 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
2113 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
2115 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
2116 The FMA instruction names should have a 'v' prefix.
2118 2013-09-12 Richard Biener <rguenther@suse.de>
2120 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
2121 (dot_rdg): Use popen instead of system in optional code.
2122 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
2123 (already_processed_vertex_p): Adjust.
2124 (has_anti_or_output_dependence, predecessor_has_mem_write,
2125 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
2126 rdg_flag_uses): Remove.
2127 (rdg_flag_vertex): Simplify.
2128 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
2130 (build_rdg_partition_for_component): Process the first vertex
2131 of a component only.
2132 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
2134 2013-09-12 Alan Modra <amodra@gmail.com>
2136 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
2138 2013-09-11 DJ Delorie <dj@redhat.com>
2139 Nick Clifton <nickc@redhat.com>
2141 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
2142 (rl78_stack_based_mem): New.
2143 * config/rl78/constraints.md (Iv08): New.
2152 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
2154 (movsi): Change from expand to insn-and-split.
2155 (ashrsi3): Clobber AX.
2159 * config/rl78/rl78.md (CC_REG): Fix.
2160 (addsi3): Allow memory and immediate operands.
2161 (addsi3_internal): Split into...
2162 (addsi3_internal_virt): ...new, and ...
2163 (addsi3_internal_real): ...new.
2165 (subsi3_internal_virt): New.
2166 (subsi3_internal_real): New.
2167 (mulsi3): Add memory operand.
2168 (mulsi3_rl78): Likewise.
2169 (mulsi3_g13): Likewise.
2170 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
2171 (cbranchqi4_real): Add more constraint options.
2172 (cbranchhi4_real): Expand pattern.
2173 (cbranchhi4_real_signed): New.
2174 (cbranchhi4_real_inverted): New.
2175 (cbranchsi4_real_lt): New.
2176 (cbranchsi4_real_ge): New.
2177 (cbranchsi4_real_signed): New.
2178 (cbranchsi4_real): New.
2180 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
2182 (lshrsi3_virt): Likewise.
2183 (ashlsi3_virt): Likewise.
2184 (cbranchqi4_virt_signed): New.
2185 (cbranchhi4_virt_signed): New.
2186 (cbranchsi4_virt): New.
2187 * config/rl78/rl78.c: Whitespace fixes throughout.
2188 (move_elim_pass): New.
2189 (pass_data_rl78_move_elim): New.
2190 (pass_rl78_move_elim): New.
2191 (make_pass_rl78_move_elim): New.
2192 (rl78_devirt_info): Run devirt earlier.
2193 (rl78_move_elim_info): New.
2194 (rl78_asm_file_start): Register it.
2195 (rl78_split_movsi): New.
2196 (rl78_as_legitimate_address): Allow virtual base registers when
2198 (rl78_addr_space_convert): Remove spurious debug stuff.
2199 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
2200 (rl78_print_operand): More cases for not printing '#'.
2201 (rl78_expand_compare): Remove most of the logic.
2202 (content_memory): New.
2203 (clear_content_memory): New.
2204 (get_content_index): New.
2205 (get_content_name): New.
2206 (display_content_memory): New.
2207 (update_content): New.
2208 (record_content): New.
2209 (already_contains): New.
2210 (insn_ok_now): Re-recog insns with virtual registers.
2211 (add_postponed_content_update): New.
2212 (process_postponed_content_update): New.
2213 (gen_and_emit_move): New.
2214 (transcode_memory_rtx): Record new location content.
2215 Use gen_and_emit_move.
2216 (force_into_acc): New.
2217 (move_to_acc): Use gen_and_emit_move.
2218 (move_from_acc): Likewise.
2219 (move_acc_to_reg): Likewise.
2220 (move_to_x): Likewise.
2221 (move_to_hl): Likewise.
2222 (move_to_de): Likewise.
2223 (rl78_alloc_physical_registers_op1): Record location content.
2224 (has_constraint): New.
2225 (rl78_alloc_physical_registers_op2): Record location content.
2227 (rl78_alloc_physical_registers_ro1): Likewise.
2228 (rl78_alloc_physical_registers_cmp): Likewise.
2229 (rl78_alloc_physical_registers_umul): Likewise.
2230 (rl78_alloc_address_registers_macax): New.
2231 (rl78_alloc_physical_registers): Initialize and set location
2232 content memory as needed.
2233 (rl78_reorg): Make sure split2 is called.
2234 (rl78_rtx_costs): New.
2236 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
2238 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
2239 for (not (neg ...)) and (neg (not ...)) cases.
2241 2013-09-11 Richard Biener <rguenther@suse.de>
2244 * passes.def: Split critical edges before late uninit warning passes.
2245 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
2247 2013-09-11 Jakub Jelinek <jakub@redhat.com>
2249 PR tree-optimization/58385
2250 * fold-const.c (build_range_check): If both low and high are NULL,
2251 use omit_one_operand_loc to preserve exp side-effects.
2253 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2255 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
2257 2013-09-11 Richard Biener <rguenther@suse.de>
2259 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
2260 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
2261 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
2262 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
2263 stmts_from_loop, known_dependences_p, build_empty_rdg,
2264 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
2265 * tree-loop-distribution.c: ... here.
2266 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
2267 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
2268 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
2269 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
2270 * tree-loop-distribution.c: ... here.
2271 * tree-loop-distribution.c: Include gimple-pretty-print.h.
2272 (struct partition_s): Add loops member.
2273 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
2274 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
2275 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
2277 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2278 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2279 Sergey Lega <sergey.s.lega@intel.com>
2280 Anna Tikhonova <anna.tikhonova@intel.com>
2281 Ilya Tocar <ilya.tocar@intel.com>
2282 Andrey Turetskiy <andrey.turetskiy@intel.com>
2283 Ilya Verbin <ilya.verbin@intel.com>
2284 Kirill Yukhin <kirill.yukhin@intel.com>
2285 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2287 * config/i386/constraints.md (k): New.
2289 * config/i386/i386.c (const regclass_map): Add new mask registers.
2290 (dbx_register_map): Ditto.
2291 (dbx64_register_map): Ditto.
2292 (svr4_dbx_register_map): Ditto.
2293 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
2295 (ix86_preferred_reload_class): Disable constants for mask registers.
2296 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
2297 (ix86_hard_regno_mode_ok): Support new mask registers.
2298 (x86_order_regs_for_local_alloc): Ditto.
2299 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
2300 (FIXED_REGISTERS): Add new mask registers.
2301 (CALL_USED_REGISTERS): Ditto.
2302 (REG_ALLOC_ORDER): Ditto.
2303 (VALID_MASK_REG_MODE): New.
2304 (FIRST_MASK_REG): Ditto.
2305 (LAST_MASK_REG): Ditto.
2306 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
2307 (MAYBE_MASK_CLASS_P): New.
2308 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
2309 (REG_CLASS_CONTENTS): Ditto.
2310 (MASK_REGNO_P): New.
2311 (ANY_MASK_REG_P): Ditto.
2312 (HI_REGISTER_NAMES): Add new mask registers.
2313 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
2314 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
2316 (attribute "type"): Add mskmov, msklog.
2317 (attribute "length_immediate"): Support them.
2318 (attribute "memory"): Ditto.
2319 (attribute "prefix_0f"): Ditto.
2320 (*movhi_internal): Support new mask registers.
2321 (*movqi_internal): Ditto.
2322 (define_split): Split out clobber pattern is a logic
2323 insn on mask registers.
2324 (*k<logic><mode>): New.
2325 (*andhi_1): Extend to support mask regs.
2326 (*andqi_1): Extend to support mask regs.
2328 (define_split): Split and-not to and and not if operands
2330 (*<code><mode>_1): Separate HI mode to new pattern...
2331 (*<code>hi_1): This.
2332 (*<code>qi_1): Extend to support mask regs.
2334 (kortestzhi): Ditto.
2335 (kortestchi): Ditto.
2337 (*one_cmpl<mode>2_1): Remove HImode and handle it...
2338 (*one_cmplhi2_1): ...Here, now with mask registers support.
2339 (*one_cmplqi2_1): Support new mask registers.
2340 (HI/QImode arithmetics splitter): Don't split if mask registers
2342 (HI/QImode not splitter): Ditto.
2343 * config/i386/predicated.md (mask_reg_operand): New.
2344 (general_reg_operand): Ditto.
2346 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2348 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
2349 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
2351 2013-09-10 Jeff Law <law@redhat.com>
2353 PR tree-optimization/58380
2354 * tree-ssa-threadupdate.c (thread_block): Recognize another case
2355 of threading through a buried loop header.
2357 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
2358 return value for single successor case.
2360 2013-09-10 Jan Hubicka <jh@suse.cz>
2362 * ipa-devirt.c (ipa_devirt): Enable with LTO.
2364 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
2367 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
2368 support conditional execution.
2369 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
2371 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
2373 * lra.c (lra): Clear lra_optional_reload_pseudos before every
2375 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
2376 Check destination too to check move insn.
2377 (undo_optional_reloads): Add check that the original peudo did not
2378 changed its allocation and the optional reload was inherited on last
2379 inheritance pass. Break loop after deciding to keep optional reload.
2380 (lra_undo_inheritance): Add check that inherited pseudo still in
2383 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
2385 * config/aarch64/aarch64.md (generic_sched): New.
2386 * config/aarch64/aarch64-generic.md (load): Make conditional
2387 on generic_sched attribute.
2388 (nonload): Likewise.
2390 2013-09-10 Jan Hubicka <jh@suse.cz>
2392 * lto-cgraph.c: Include ipa-utils.h.
2393 (compute_ltrans_boundary): Also add possible targets into the boundary.
2395 2013-09-10 Jan Hubicka <jh@suse.cz>
2397 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
2398 VAR_DECL of vtable rather than full expression.
2400 2013-09-10 Jan Hubicka <jh@suse.cz>
2401 Paolo Carlini <paolo.carlini@oracle.com>
2403 * cgraphunit.c (analyze_functions): Save input_location, set it
2404 to UNKNOWN_LOCATION and restore it at the end.
2406 2013-09-10 Martin Jambor <mjambor@suse.cz>
2408 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
2409 represented by a thunk.
2411 2013-09-10 Jeff Law <law@redhat.com>
2413 PR tree-optimization/58343
2414 * tree-ssa-threadupdate.c (thread_block): Identify and disable
2415 jump threading requests through loop headers buried in the middle
2416 of a jump threading path.
2418 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
2419 in return value/type.
2421 2013-09-10 Jakub Jelinek <jakub@redhat.com>
2423 PR rtl-optimization/58365
2424 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
2425 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
2428 2013-09-10 Richard Biener <rguenther@suse.de>
2430 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
2431 * tree-data-ref.c (create_rdg_vertices): Collect all data
2432 references, signal failure to the caller, use data-ref API.
2433 (build_rdg): Compute data references only once. Maintain lifetime
2434 of data references and data dependences from within RDG.
2435 (free_rdg): Free dependence relations.
2436 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
2437 inventing extra dependences.
2438 (distribute_loop): Update for RDG API changes.
2440 2013-09-10 Kai Tietz <ktietz@redhat.com>
2442 * doc/invoke.texi (fms-extensions): Document changed
2443 behavior for ms-abi targets.
2444 * config/i386/i386.c (ix86_option_override_internal):
2445 Set default value of option -fms-extension for ms-abi targets.
2447 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
2449 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
2451 2013-09-10 Alan Modra <amodra@gmail.com>
2454 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
2456 2013-09-10 Alan Modra <amodra@gmail.com>
2458 * config/rs6000/predicates.md (add_cint_operand): New.
2459 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
2460 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
2461 using add_cint_operand.
2462 (largetoc_high_plus_aix): Likewise.
2464 2013-09-09 Jakub Jelinek <jakub@redhat.com>
2466 PR tree-optimization/58364
2467 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
2468 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
2469 the current range can't be an unconditional true or false.
2471 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
2473 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
2475 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
2477 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
2479 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
2482 * doc/invoke.texi (-Wdelete-incomplete): Document it.
2484 2013-09-09 Ian Bolton <ian.bolton@arm.com>
2486 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
2487 NO_REGS for immediate that can't be moved directly into FP_REGS.
2489 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2491 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
2492 comparison with negated operand.
2493 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
2496 2013-09-09 Richard Biener <rguenther@suse.de>
2499 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
2500 of a subloop record all its block as affecting loop-closed SSA form.
2502 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
2504 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
2505 of an rtx/bitpos pair.
2506 (store_fixed_bit_field): Update accordingly.
2508 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
2510 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
2512 * builtins.c (expand_errno_check): Likewise.
2513 * dwarf2cfi.c (init_return_column_size): Likewise.
2514 * except.c (sjlj_mark_call_sites): Likewise.
2515 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
2516 * lra-constraints.c (emit_inc): Likewise.
2517 * ree.c (combine_set_extension): Likewise.
2518 * regmove.c (fixup_match_2): Likewise.
2519 * reload1.c (inc_for_reload): Likewise.
2521 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
2523 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
2524 (force_to_mode, simplify_shift_const_1, simplify_comparison):
2525 Use gen_int_mode with the mode of the associated simplify_* call.
2526 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
2527 * expmed.c (expand_shift_1): Likewise.
2528 * function.c (instantiate_virtual_regs_in_insn): Likewise.
2529 * loop-iv.c (iv_number_of_iterations): Likewise.
2530 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
2531 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
2533 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
2535 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
2536 of the associated expand_* call.
2537 (asan_emit_stack_protection): Likewise.
2538 * builtins.c (round_trampoline_addr): Likewise.
2539 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
2540 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
2541 (emit_store_flag): Likewise.
2542 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
2543 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
2545 * function.c (instantiate_virtual_regs_in_insn): Likewise.
2546 * ifcvt.c (noce_try_store_flag_constants): Likewise.
2547 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
2548 * modulo-sched.c (generate_prolog_epilog): Likewise.
2549 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
2550 (expand_ctz, expand_ffs, expand_unop): Likewise.
2552 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
2554 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
2555 of the associated gen_rtx_* call.
2556 * caller-save.c (init_caller_save): Likewise.
2557 * combine.c (find_split_point, make_extraction): Likewise.
2558 (make_compound_operation): Likewise.
2559 * dwarf2out.c (mem_loc_descriptor): Likewise.
2560 * explow.c (plus_constant, probe_stack_range): Likewise.
2561 * expmed.c (expand_mult_const): Likewise.
2562 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
2563 * reload1.c (init_reload): Likewise.
2564 * valtrack.c (cleanup_auto_inc_dec): Likewise.
2565 * var-tracking.c (adjust_mems): Likewise.
2566 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
2567 rather than gen_rtx_fmt_ee.
2569 2013-09-09 Jan Hubicka <jh@suse.cz>
2572 * value-prof.c (gimple_ic): Copy also abnormal edges.
2574 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
2576 * asan.c (asan_shadow_cst): Use gen_int_mode.
2578 2013-09-08 Jan Hubicka <jh@suse.cz>
2580 * ipa-profile.c: Add toplevel comment.
2581 (ipa_propagate_frequency_1): Be more conservative when profile is read.
2582 (contains_hot_call_p): New function.
2583 (ipa_propagate_frequency): Set frequencies based on counts when
2585 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
2586 profile; do not tamper with profile after inlining if it is read.
2588 2013-09-08 Jan Hubicka <jh@suse.cz>
2590 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
2593 2013-09-08 Jan Hubicka <jh@suse.cz>
2595 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
2598 2013-09-08 Jeff Law <law@redhat.com>
2601 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
2604 2013-09-08 Andi Kleen <ak@linux.intel.com>
2606 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
2608 2013-09-08 Jan Hubicka <jh@suse.cz>
2610 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
2612 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
2614 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
2615 for non-debug insns.
2616 * lra.c (new_insn_reg): Take the containing insn as a parameter.
2617 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
2618 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
2621 2013-09-08 Jan Hubicka <jh@suse.cz>
2623 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
2624 targets and devirtualize to BUILT_IN_UNREACHABLE.
2625 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
2626 * ipa.c (walk_polymorphic_call_targets): New function.
2627 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
2628 functions; use the new timevar.
2629 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
2630 was removed from the program.
2631 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
2632 not consider it in the walk when its vtable is dead.
2633 (possible_polymorphic_call_targets_1): Pass anonymous flag to
2635 (devirt_variable_node_removal_hook): New function.
2636 (possible_polymorphic_call_targets): Also register
2637 devirt_variable_node_removal_hook.
2638 (ipa_devirt): Do not do non-speculative devirtualization.
2639 (gate_ipa_devirt): One execute if devirtualizing speculatively.
2641 2013-09-08 Jan Hubicka <jh@suse.cz>
2643 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
2644 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
2645 varpool_remove_variable_insertion_hook): Declare.
2646 * varpool.c (varpool_node_hook_list): New structure.
2647 (first_varpool_node_removal_hook,
2648 first_varpool_variable_insertion_hook): New variables.
2649 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
2650 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
2651 varpool_remove_variable_insertion_hook,
2652 varpool_call_variable_insertion_hooks): New functions.
2653 (varpool_remove_node): Use it.
2655 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
2658 * diagnostic.c (diagnostic_build_prefix): When s.file is
2659 "<built-in>" don't output line and column numbers.
2661 2013-09-06 Jan Hubicka <jh@suse.cz>
2663 * cgraphunit.c (expand_thunk): Get body before touching arguments.
2664 * lto-streamer-out.c: Stream thunks, too.
2665 * lto-streamer-in.c (input_function): Pop cfun here
2666 (lto_read_body): Instead of here.
2668 2013-09-06 Caroline Tice <cmtice@google.com>
2670 * doc/install.texi: Add documentation for the --enable-vtable-verify
2671 and the --disable-libvtv configure options.
2673 2013-09-06 Jeff Law <law@redhat.com>
2675 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
2676 edge implied equivalences into successor phis.
2678 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
2680 * resource.c (mark_referenced_resources): Handle COND_EXEC.
2682 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
2684 * resource.c (mark_target_live_regs): Compute resources taking
2685 into account if a call is predicated or not.
2687 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
2689 * toplev.c (output_stack_usage): Be prepared for suffixes created by
2690 the compiler in the function names.
2692 2013-09-06 Jan Hubicka <jh@suse.cz>
2695 * ipa-inline.c (has_caller_p): New function.
2696 (want_inline_function_to_all_callers_p): Use it.
2697 (sum_callers, inline_to_all_callers): Break out from ...
2698 (ipa_inline): ... here.
2700 2013-09-06 Jan Hubicka <jh@suse.cz>
2702 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
2703 only when AVX is enabled.
2705 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2707 * config/aarch64/aarch64.md
2708 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
2709 is fpsimd_<load/store>2.
2710 (load_pair<mode>): Likewise.
2711 (store_pair<mode>): Likewise.
2713 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2715 * config/arm/types.md (type): Add "mrs" type.
2716 * config/aarch64/aarch64.md
2717 (aarch64_load_tp_hard): Make type "mrs".
2719 (load_tp_hard): Make type "mrs".
2720 * config/arm/cortex-a15.md: Update with new attributes.
2721 * config/arm/cortex-a5.md: Update with new attributes.
2722 * config/arm/cortex-a53.md: Update with new attributes.
2723 * config/arm/cortex-a7.md: Update with new attributes.
2724 * config/arm/cortex-a8.md: Update with new attributes.
2725 * config/arm/cortex-a9.md: Update with new attributes.
2726 * config/arm/cortex-m4.md: Update with new attributes.
2727 * config/arm/cortex-r4.md: Update with new attributes.
2728 * config/arm/fa526.md: Update with new attributes.
2729 * config/arm/fa606te.md: Update with new attributes.
2730 * config/arm/fa626te.md: Update with new attributes.
2731 * config/arm/fa726te.md: Update with new attributes.
2733 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2735 * config/aarch64/aarch64.md
2736 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
2737 (*movtf_aarch64): Likewise.
2739 (thumb1_movdi_insn): Use "multiple" for type where more than one
2740 instruction is used for a move.
2741 (*arm32_movhf): Likewise.
2742 (*thumb_movdf_insn): Likewise.
2744 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2746 * config/arm/types.md (type): Rename fcpys to fmov.
2748 (*arm_movsi_vfp): Rename type fcpys as fmov.
2749 (*thumb2_movsi_vfp): Likewise
2750 (*movhf_vfp_neon): Likewise
2751 (*movhf_vfp): Likewise
2752 (*movsf_vfp): Likewise
2753 (*thumb2_movsf_vfp): Likewise
2754 (*movsfcc_vfp): Likewise
2755 (*thumb2_movsfcc_vfp): Likewise
2756 * config/aarch64/aarch64-simd.md
2757 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
2758 * config/aarch64/aarch64.md
2759 (*movsi_aarch64): Replace type mov_reg with fmovs.
2760 (*movdi_aarch64): Likewise
2761 (*movsf_aarch64): Likewise
2762 (*movdf_aarch64): Likewise
2764 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
2765 * config/arm/iwmmxt.md
2766 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
2767 * config/arm/arm1020e.md: Update with new attributes.
2768 * config/arm/cortex-a15-neon.md: Update with new attributes.
2769 * config/arm/cortex-a5.md: Update with new attributes.
2770 * config/arm/cortex-a53.md: Update with new attributes.
2771 * config/arm/cortex-a7.md: Update with new attributes.
2772 * config/arm/cortex-a8-neon.md: Update with new attributes.
2773 * config/arm/cortex-a9.md: Update with new attributes.
2774 * config/arm/cortex-m4-fpu.md: Update with new attributes.
2775 * config/arm/cortex-r4f.md: Update with new attributes.
2776 * config/arm/marvell-pj4.md: Update with new attributes.
2777 * config/arm/vfp11.md: Update with new attributes.
2779 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2781 * config/aarch64/aarch64.md
2782 (*madd<mode>): Fix type attribute.
2783 (*maddsi_uxtw): Likewise.
2784 (*msub<mode>): Likewise.
2785 (*msubsi_uxtw): Likewise.
2786 (<su_optab>maddsidi4): Likewise.
2787 (<su_optab>msubsidi4): Likewise.
2789 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2791 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
2792 * config/arm/arm.md (core_cycles): Remove fdiv.
2793 * config/arm/vfp.md:
2794 (*sqrtsf2_vfp): Update for attribute changes.
2795 (*sqrtdf2_vfp): Likewise.
2796 * config/aarch64/aarch64.md:
2797 (sqrt<mode>2): Update for attribute changes.
2798 * config/arm/arm1020e.md: Update with new attributes.
2799 * config/arm/cortex-a15-neon.md: Update with new attributes.
2800 * config/arm/cortex-a5.md: Update with new attributes.
2801 * config/arm/cortex-a53.md: Update with new attributes.
2802 * config/arm/cortex-a7.md: Update with new attributes.
2803 * config/arm/cortex-a8-neon.md: Update with new attributes.
2804 * config/arm/cortex-a9.md: Update with new attributes.
2805 * config/arm/cortex-m4-fpu.md: Update with new attributes.
2806 * config/arm/cortex-r4f.md: Update with new attributes.
2807 * config/arm/marvell-pj4.md: Update with new attributes.
2808 * config/arm/vfp11.md: Update with new attributes.
2810 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2812 * config/arm/types.md
2813 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
2814 * config/aarch64/aarch64.md
2815 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
2817 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
2818 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
2819 (float<GPI:mode><GPF:mode>2): Likewise.
2821 (*truncsisf2_vfp): Update with new attributes.
2822 (*truncsidf2_vfp): Likewise.
2823 (fixuns_truncsfsi2): Likewise.
2824 (fixuns_truncdfsi2): Likewise.
2825 (*floatsisf2_vfp): Likewise.
2826 (*floatsidf2_vfp): Likewise.
2827 (floatunssisf2): Likewise.
2828 (floatunssidf2): Likewise.
2829 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
2830 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
2831 * config/arm/arm1020e.md: Update with new attributes.
2832 * config/arm/cortex-a15-neon.md: Update with new attributes.
2833 * config/arm/cortex-a5.md: Update with new attributes.
2834 * config/arm/cortex-a53.md: Update with new attributes.
2835 * config/arm/cortex-a7.md: Update with new attributes.
2836 * config/arm/cortex-a8-neon.md: Update with new attributes.
2837 * config/arm/cortex-a9.md: Update with new attributes.
2838 * config/arm/cortex-m4-fpu.md: Update with new attributes.
2839 * config/arm/cortex-r4f.md: Update with new attributes.
2840 * config/arm/marvell-pj4.md: Update with new attributes.
2841 * config/arm/vfp11.md: Update with new attributes.
2843 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2845 * config/aarch64/arm_neon.h
2846 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
2847 (vqtbx<1,2,3,4><q>_s8): Likewise.
2849 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2851 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
2852 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
2853 (add<mode>3): Add type attribute.
2854 (add<mode>3): Likewise.
2855 (usadd<mode>3): Likewise.
2856 (ssadd<mode>3): Likewise.
2857 (sub<mode>3): Likewise.
2858 (sub<mode>3): Likewise.
2859 (ussub<mode>3): Likewise.
2860 (sssub<mode>3): Likewise.
2861 (ssmulsa3): Likewise.
2862 (usmulusa3): Likewise.
2863 (arm_usatsihi): Likewise.
2865 (*movdi_vfp): Add types for all instructions.
2866 (*movdi_vfp_cortexa8): Likewise.
2867 (*movhf_vfp_neon): Likewise.
2868 (*movhf_vfp): Likewise.
2869 (*movdf_vfp): Likewise.
2870 (*thumb2_movdf_vfp): Likewise.
2871 (*thumb2_movdfcc_vfp): Likewise.
2872 * config/arm/arm.md: Add type attribute to all insn patterns.
2873 (*thumb1_adddi3): Add type attribute.
2874 (*arm_adddi3): Likewise.
2875 (*adddi_sesidi_di): Likewise.
2876 (*adddi_zesidi_di): Likewise.
2877 (*thumb1_addsi3): Likewise.
2878 (addsi3_compare0): Likewise.
2879 (*addsi3_compare0_scratch): Likewise.
2880 (*compare_negsi_si): Likewise.
2881 (cmpsi2_addneg): Likewise.
2882 (*addsi3_carryin_<optab>): Likewise.
2883 (*addsi3_carryin_alt2_<optab>): Likewise.
2884 (*addsi3_carryin_clobercc_<optab>): Likewise.
2885 (*subsi3_carryin): Likewise.
2886 (*subsi3_carryin_const): Likewise.
2887 (*subsi3_carryin_compare): Likewise.
2888 (*subsi3_carryin_compare_const): Likewise.
2889 (*arm_subdi3): Likewise.
2890 (*thumb_subdi3): Likewise.
2891 (*subdi_di_zesidi): Likewise.
2892 (*subdi_di_sesidi): Likewise.
2893 (*subdi_zesidi_di): Likewise.
2894 (*subdi_sesidi_di): Likewise.
2895 (*subdi_zesidi_ze): Likewise.
2896 (thumb1_subsi3_insn): Likewise.
2897 (*arm_subsi3_insn): Likewise.
2898 (*anddi3_insn): Likewise.
2899 (*anddi_zesidi_di): Likewise.
2900 (*anddi_sesdi_di): Likewise.
2901 (*ne_zeroextracts): Likewise.
2902 (*ne_zeroextracts): Likewise.
2903 (*ite_ne_zeroextr): Likewise.
2904 (*ite_ne_zeroextr): Likewise.
2905 (*anddi_notdi_di): Likewise.
2906 (*anddi_notzesidi): Likewise.
2907 (*anddi_notsesidi): Likewise.
2908 (andsi_notsi_si): Likewise.
2909 (thumb1_bicsi3): Likewise.
2910 (*iordi3_insn): Likewise.
2911 (*iordi_zesidi_di): Likewise.
2912 (*iordi_sesidi_di): Likewise.
2913 (*thumb1_iorsi3_insn): Likewise.
2914 (*xordi3_insn): Likewise.
2915 (*xordi_zesidi_di): Likewise.
2916 (*xordi_sesidi_di): Likewise.
2917 (*arm_xorsi3): Likewise.
2918 (*andsi_iorsi3_no): Likewise.
2919 (*smax_0): Likewise.
2920 (*smax_m1): Likewise.
2921 (*arm_smax_insn): Likewise.
2922 (*smin_0): Likewise.
2923 (*arm_smin_insn): Likewise.
2924 (*arm_umaxsi3): Likewise.
2925 (*arm_uminsi3): Likewise.
2926 (*minmax_arithsi): Likewise.
2927 (*minmax_arithsi_): Likewise.
2928 (*satsi_<SAT:code>): Likewise.
2929 (arm_ashldi3_1bit): Likewise.
2930 (arm_ashrdi3_1bit): Likewise.
2931 (arm_lshrdi3_1bit): Likewise.
2932 (*arm_negdi2): Likewise.
2933 (*thumb1_negdi2): Likewise.
2934 (*arm_negsi2): Likewise.
2935 (*thumb1_negsi2): Likewise.
2936 (*negdi_extendsid): Likewise.
2937 (*negdi_zero_extend): Likewise.
2938 (*arm_abssi2): Likewise.
2939 (*thumb1_abssi2): Likewise.
2940 (*arm_neg_abssi2): Likewise.
2941 (*thumb1_neg_abss): Likewise.
2942 (one_cmpldi2): Likewise.
2943 (extend<mode>di2): Likewise.
2944 (*compareqi_eq0): Likewise.
2945 (*arm_extendhisi2addsi): Likewise.
2946 (*arm_movdi): Likewise.
2947 (*thumb1_movdi_insn): Likewise.
2948 (*arm_movt): Likewise.
2949 (*thumb1_movsi_insn): Likewise.
2950 (pic_add_dot_plus_four): Likewise.
2951 (pic_add_dot_plus_eight): Likewise.
2952 (tls_load_dot_plus_eight): Likewise.
2953 (*thumb1_movhi_insn): Likewise.
2954 (*thumb1_movsf_insn): Likewise.
2955 (*movdf_soft_insn): Likewise.
2956 (*thumb_movdf_insn): Likewise.
2957 (cbranchsi4_insn): Likewise.
2958 (cbranchsi4_scratch): Likewise.
2959 (*negated_cbranchsi4): Likewise.
2960 (*tbit_cbranch): Likewise.
2961 (*tlobits_cbranch): Likewise.
2962 (*tstsi3_cbranch): Likewise.
2963 (*cbranchne_decr1): Likewise.
2964 (*addsi3_cbranch): Likewise.
2965 (*addsi3_cbranch_scratch): Likewise.
2966 (*arm_cmpdi_insn): Likewise.
2967 (*arm_cmpdi_unsig): Likewise.
2968 (*arm_cmpdi_zero): Likewise.
2969 (*thumb_cmpdi_zero): Likewise.
2970 (*deleted_compare): Likewise.
2971 (*mov_scc): Likewise.
2972 (*mov_negscc): Likewise.
2973 (*mov_notscc): Likewise.
2974 (*cstoresi_eq0_thumb1_insn): Likewise.
2975 (cstoresi_nltu_thumb1): Likewise.
2976 (cstoresi_ltu_thu): Likewise.
2977 (thumb1_addsi3_addgeu): Likewise.
2978 (*arm_jump): Likewise.
2979 (*thumb_jump): Likewise.
2980 (*check_arch2): Likewise.
2981 (arm_casesi_internal): Likewise.
2982 (thumb1_casesi_dispatch): Likewise.
2983 (*arm_indirect_jump): Likewise.
2984 (*thumb1_indirect_jump): Likewise.
2986 (*and_scc): Likewise.
2987 (*ior_scc): Likewise.
2988 (*compare_scc): Likewise.
2989 (*cond_move): Likewise.
2990 (*cond_arith): Likewise.
2991 (*cond_sub): Likewise.
2992 (*cmp_ite0): Likewise.
2993 (*cmp_ite1): Likewise.
2994 (*cmp_and): Likewise.
2995 (*cmp_ior): Likewise.
2996 (*ior_scc_scc): Likewise.
2997 (*ior_scc_scc_cmp): Likewise.
2998 (*and_scc_scc): Likewise.
2999 (*and_scc_scc_cmp): Likewise.
3000 (*and_scc_scc_nod): Likewise.
3001 (*negscc): Likewise.
3002 (movcond_addsi): Likewise.
3003 (movcond): Likewise.
3004 (*ifcompare_plus_move): Likewise.
3005 (*if_plus_move): Likewise.
3006 (*ifcompare_move_plus): Likewise.
3007 (*if_move_plus): Likewise.
3008 (*ifcompare_arith_arith): Likewise.
3009 (*if_arith_arith): Likewise.
3010 (*ifcompare_arith_move): Likewise.
3011 (*if_arith_move): Likewise.
3012 (*ifcompare_move_arith): Likewise.
3013 (*if_move_arith): Likewise.
3014 (*ifcompare_move_not): Likewise.
3015 (*if_move_not): Likewise.
3016 (*ifcompare_not_move): Likewise.
3017 (*if_not_move): Likewise.
3018 (*ifcompare_shift_move): Likewise.
3019 (*if_shift_move): Likewise.
3020 (*ifcompare_move_shift): Likewise.
3021 (*if_move_shift): Likewise.
3022 (*ifcompare_shift_shift): Likewise.
3023 (*ifcompare_not_arith): Likewise.
3024 (*ifcompare_arith_not): Likewise.
3025 (*if_arith_not): Likewise.
3026 (*ifcompare_neg_move): Likewise.
3027 (*if_neg_move): Likewise.
3028 (*ifcompare_move_neg): Likewise.
3029 (*if_move_neg): Likewise.
3030 (prologue_thumb1_interwork): Likewise.
3031 (*cond_move_not): Likewise.
3032 (*sign_extract_onebit): Likewise.
3033 (*not_signextract_onebit): Likewise.
3034 (stack_tie): Likewise.
3035 (align_4): Likewise.
3036 (align_8): Likewise.
3037 (consttable_end): Likewise.
3038 (consttable_1): Likewise.
3039 (consttable_2): Likewise.
3040 (consttable_4): Likewise.
3041 (consttable_8): Likewise.
3042 (consttable_16): Likewise.
3043 (*thumb1_tablejump): Likewise.
3044 (prefetch): Likewise.
3045 (force_register_use): Likewise.
3046 (thumb_eh_return): Likewise.
3047 (load_tp_hard): Likewise.
3048 (load_tp_soft): Likewise.
3049 (tlscall): Likewise.
3050 (*arm_movtas_ze): Likewise.
3051 (*arm_rev): Likewise.
3052 (*arm_revsh): Likewise.
3053 (*arm_rev16): Likewise.
3054 * config/arm/thumb2.md
3055 (*thumb2_smaxsi3): Likewise.
3056 (*thumb2_sminsi3): Likewise.
3057 (*thumb32_umaxsi3): Likewise.
3058 (*thumb2_uminsi3): Likewise.
3059 (*thumb2_negdi2): Likewise.
3060 (*thumb2_abssi2): Likewise.
3061 (*thumb2_neg_abss): Likewise.
3062 (*thumb2_movsi_insn): Likewise.
3063 (tls_load_dot_plus_four): Likewise.
3064 (*thumb2_movhi_insn): Likewise.
3065 (*thumb2_mov_scc): Likewise.
3066 (*thumb2_mov_negs): Likewise.
3067 (*thumb2_mov_negs): Likewise.
3068 (*thumb2_mov_nots): Likewise.
3069 (*thumb2_mov_nots): Likewise.
3070 (*thumb2_movsicc_): Likewise.
3071 (*thumb2_movsfcc_soft_insn): Likewise.
3072 (*thumb2_indirect_jump): Likewise.
3073 (*thumb2_and_scc): Likewise.
3074 (*thumb2_ior_scc): Likewise.
3075 (*thumb2_ior_scc_strict_it): Likewise.
3076 (*thumb2_cond_move): Likewise.
3077 (*thumb2_cond_arith): Likewise.
3078 (*thumb2_cond_ari): Likewise.
3079 (*thumb2_cond_sub): Likewise.
3080 (*thumb2_negscc): Likewise.
3081 (*thumb2_movcond): Likewise.
3082 (thumb2_casesi_internal): Likewise.
3083 (thumb2_casesi_internal_pic): Likewise.
3084 (*thumb2_alusi3_short): Likewise.
3085 (*thumb2_mov<mode>_shortim): Likewise.
3086 (*thumb2_addsi_short): Likewise.
3087 (*thumb2_subsi_short): Likewise.
3088 (thumb2_addsi3_compare0): Likewise.
3089 (*thumb2_cbz): Likewise.
3090 (*thumb2_cbnz): Likewise.
3091 (*thumb2_one_cmplsi2_short): Likewise.
3092 (*thumb2_negsi2_short): Likewise.
3093 (*orsi_notsi_si): Likewise.
3094 * config/arm/arm1020e.md: Update with new attributes.
3095 * config/arm/arm1026ejs.md: Update with new attributes.
3096 * config/arm/arm1136jfs.md: Update with new attributes.
3097 * config/arm/arm926ejs.md: Update with new attributes.
3098 * config/arm/cortex-a15.md: Update with new attributes.
3099 * config/arm/cortex-a5.md: Update with new attributes.
3100 * config/arm/cortex-a53.md: Update with new attributes.
3101 * config/arm/cortex-a7.md: Update with new attributes.
3102 * config/arm/cortex-a8.md: Update with new attributes.
3103 * config/arm/cortex-a9.md: Update with new attributes.
3104 * config/arm/cortex-m4.md: Update with new attributes.
3105 * config/arm/cortex-r4.md: Update with new attributes.
3106 * config/arm/fa526.md: Update with new attributes.
3107 * config/arm/fa606te.md: Update with new attributes.
3108 * config/arm/fa626te.md: Update with new attributes.
3109 * config/arm/fa726te.md: Update with new attributes.
3111 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
3113 * config/aarch64/aarch64-simd.md
3114 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
3115 <vwx> iterator to ensure correct register choice.
3116 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
3117 (aarch64_sqdmull_n<mode>): Likewise.
3118 (aarch64_sqdmull2_n<mode>_internal): Likewise.
3119 * config/aarch64/arm_neon.h
3120 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
3121 (vml<as><q>_n_<su>16): Likewise.
3122 (vml<as>l_high_lane<q>_<su>16): Likewise.
3123 (vml<as>l_high_n_<su>16): Likewise.
3124 (vml<as>l_lane<q>_<su>16): Likewise.
3125 (vml<as>l_n_<su>16): Likewise.
3126 (vmul<q>_lane<q>_<su>16): Likewise.
3127 (vmul<q>_n_<su>16): Likewise.
3128 (vmull_lane<q>_<su>16): Likewise.
3129 (vmull_n_<su>16): Likewise.
3130 (vmull_high_lane<q>_<su>16): Likewise.
3131 (vmull_high_n_<su>16): Likewise.
3132 (vqrdmulh<q>_n_s16): Likewise.
3134 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
3136 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
3137 have the correct lane parameter.
3139 2013-09-06 Richard Biener <rguenther@suse.de>
3141 * cfganal.c (control_dependences::~control_dependences):
3142 Properly free all of the vector.
3144 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
3147 * config/i386/i386.c (ix86_conditional_register_usage):
3148 Proper initialize extended SSE registers.
3150 2013-09-06 Jan Hubicka <jh@suse.cz>
3152 PR tree-optimization/58311
3153 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
3155 2013-09-06 Jan Hubicka <jh@suse.cz>
3157 * Makefile.in (tree-sra.o): Update dependencies.
3158 * tree-sra.c: Include ipa-utils.h
3159 (scan_function): Use recursive_call_p.
3160 (has_caller_p): New function.
3161 (cgraph_for_node_and_aliases): Count also callers of aliases.
3163 2013-09-06 Jan Hubicka <jh@suse.cz>
3166 * cgraph.h (symtab_semantically_equivalent_p): Declare.
3167 * tree-tailcall.c: Include ipa-utils.h.
3168 (find_tail_calls): Use it.
3169 * ipa-pure-const.c (check_call): Likewise.
3170 * ipa-utils.c (recursive_call_p): New function.
3171 * ipa-utils.h (recursive_call_p): Dclare.
3172 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
3173 (symtab_semantically_equivalent_p): New function.
3174 * Makefile.in (tree-tailcall.o): Update dependencies.
3176 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
3178 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
3181 2013-09-06 Richard Biener <rguenther@suse.de>
3183 * lto-streamer.h (lto_global_var_decls): Remove.
3184 * Makefile.in (OBJS): Remove lto-symtab.o.
3185 (lto-symtab.o): Remove.
3186 (GTFILES): Remove lto-symtab.c
3187 * lto-symtab.c: Move to lto/
3189 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3191 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
3192 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
3193 (UNSPEC_FPINT_RINT): New constant definitions.
3194 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
3195 definition with 2 attributes.
3196 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
3197 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
3200 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3202 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
3203 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
3204 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
3205 attribute to "z196_alone".
3206 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
3209 2013-09-06 Richard Biener <rguenther@suse.de>
3211 * basic-block.h (class control_dependences): New.
3212 * tree-ssa-dce.c (control_dependence_map): Remove.
3214 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
3215 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
3216 find_pdom, find_control_dependence, find_all_control_dependences):
3218 (mark_control_dependent_edges_necessary,
3219 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
3220 tree_dce_done, perform_tree_ssa_dce): Adjust.
3221 * cfganal.c (set_control_dependence_map_bit,
3222 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
3223 find_all_control_dependences): Move from tree-ssa-dce.c and
3224 implement as methods of control_dependences class.
3225 (control_dependences::control_dependences): New.
3226 (control_dependences::~control_dependences): Likewise.
3227 (control_dependences::get_edges_dependent_on): Likewise.
3228 (control_dependences::get_edge): Likewise.
3230 2013-09-04 Jan Hubicka <jh@suse.cz>
3232 * tree.c (types_same_for_odr): Drop overactive check.
3233 * ipa-devirt.c (hash_type_name): Likewise.
3235 2013-09-04 Jan Hubicka <jh@suse.cz>
3237 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
3238 (analyze_functions): ... here.
3240 2013-09-04 Jan Hubicka <jh@suse.cz>
3243 * cgraphunit.c (analyze_functions): Clear AUX fields
3244 after processing; initialize assembler name has.
3246 2013-09-05 Jeff Law <law@redhat.com>
3248 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
3249 from thread_around_empty_block. Record threading path into PATH.
3250 Recurse if threading through the initial block is successful.
3251 (thread_across_edge): Corresponding changes to slightly simplify.
3253 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
3255 * config/aarch64/aarch64.md
3256 (type): Remove frecpe, frecps, frecpx.
3257 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
3258 fix to be a TARGET_SIMD instruction.
3259 (aarch64_frecps): Remove.
3260 * config/aarch64/aarch64-simd.md
3261 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
3262 (aarch64_frecps<mode>): Handle all float/vector of float modes.
3264 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
3265 Sofiane Naci <sofiane.naci@arm.com>
3267 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
3268 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
3269 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
3270 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
3271 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
3272 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
3273 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
3274 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
3275 into "clz, "rbit". Rename "shift" to "shift_imm".
3276 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
3277 changes. Update for attribute changes all occurrences of arlo_* and
3279 * config/arm/arm-fixed.md: Update for attribute changes
3280 all occurrences of arlo_* types.
3281 * config/arm/thumb2.md: Update for attribute changes all occurrences
3283 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
3284 (cortexa7_older_only): Likewise.
3285 (cortexa7_younger): Likewise.
3286 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
3287 (1020alu_shift_op): Likewise.
3288 (1020alu_shift_reg_op): Likewise.
3289 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
3290 (alu_shift_op): Likewise.
3291 (alu_shift_reg_op): Likewise.
3292 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
3293 (11_alu_shift_op): Likewise.
3294 (11_alu_shift_reg_op): Likewise.
3295 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
3296 (9_alu_shift_reg_op): Likewise.
3297 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
3299 (cortex_a15_alu_shift): Likewise.
3300 (cortex_a15_alu_shift_reg): Likewise.
3301 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
3303 (cortex_a5_alu_shift): Likewise.
3304 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
3306 (cortex_a53_alu_shift): Likewise.
3307 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
3309 (cortex_a7_alu_reg): Likewise.
3310 (cortex_a7_alu_shift): Likewise.
3311 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
3313 (cortex_a8_alu_shift): Likewise.
3314 (cortex_a8_alu_shift_reg): Likewise.
3315 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
3316 (cortex_a9_dp_shift): Likewise.
3317 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
3319 * config/arm/cortex-r4.md
3320 (cortex_r4_alu): Update for attribute changes.
3321 (cortex_r4_mov): Likewise.
3322 (cortex_r4_alu_shift_reg): Likewise.
3323 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
3324 (526_alu_shift_op): Likewise.
3325 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
3326 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
3327 (626te_alu_shift_op): Likewise.
3328 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
3329 (726te_alu_shift_op): Likewise.
3330 (726te_alu_shift_reg_op): Likewise.
3331 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
3332 (mp626_alu_shift_op): Likewise.
3333 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
3334 (pj4_alu_conds): Likewise.
3335 (pj4_shift): Likewise.
3336 (pj4_shift_conds): Likewise.
3337 (pj4_alu_shift): Likewise.
3338 (pj4_alu_shift_conds): Likewise.
3339 * config/aarch64/aarch64.md: Update for attribute change
3340 all occurrences of arlo_* and shift* types.
3342 2013-09-05 Mike Stump <mikestump@comcast.net>
3344 * tree.h: Move documentation for tree_function_decl to tree-core.h
3345 with the declaration.
3347 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
3350 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
3351 looking for widest mode.
3353 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
3355 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
3357 2013-09-05 Richard Biener <rguenther@suse.de>
3359 PR tree-optimization/58137
3360 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
3361 Do not create vectors of pointers.
3362 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
3363 types for the components of the vector initializer.
3364 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
3365 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
3367 2013-09-05 Martin Jambor <mjambor@suse.cz>
3369 * ipa-prop.c (remove_described_reference): Accept missing references,
3370 return false if that hppens, otherwise return true.
3371 (cgraph_node_for_jfunc): New function.
3372 (try_decrement_rdesc_refcount): Likewise.
3373 (try_make_edge_direct_simple_call): Use them.
3374 (ipa_edge_removal_hook): Remove references from rdescs.
3375 (ipa_edge_duplication_hook): Clone rdescs and their references
3376 when the new edge has the same caller as the old one.
3377 * cgraph.c (cgraph_resolve_speculation): Remove speculative
3378 reference before removing any edges.
3380 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
3382 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
3384 * thumb2.md (thumb2_storewb_parisi): New pattern.
3386 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
3388 * config/aarch64/aarch64-option-extensions.def: Add
3389 AARCH64_OPT_EXTENSION of 'crc'.
3390 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
3391 (AARCH64_ISA_CRC): Ditto.
3392 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
3393 description of the CRC extension.
3395 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
3397 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
3398 * config/rs6000/linux.h: Ditto.
3399 * alpha/linux.h: Ditto.
3400 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
3401 no_c99_libc_has_function.
3402 * config/c6x/uclinux-elf.h: Ditto.
3403 * config/lm32/uclinux-elf.h: Ditto.
3404 * config/m68k/uclinux.h: Ditto.
3405 * config/moxie/uclinux.h: Ditto.
3406 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
3407 (crisv32-*-linux*, cris-*-linux*): Ditto.
3408 * config/bfin/bfin.c: Include "tm_p.h".
3410 2013-09-05 Richard Biener <rguenther@suse.de>
3412 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
3413 check for a definition without a basic-block.
3415 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
3416 Sofiane Naci <sofiane.naci@arm.com>
3418 * config/aarch64/aarch64.md
3419 (*movti_aarch64): Rename r_2_f and f_2_r.
3420 (*movsf_aarch64): Likewise.
3421 (*movdf_aarch64): Likewise.
3422 (*movtf_aarch64): Likewise.
3423 (aarch64_movdi_<mode>low): Likewise.
3424 (aarch64_movdi_<mode>high): Likewise.
3425 (aarch64_mov<mode>high_di): Likewise.
3426 (aarch64_mov<mode>low_di): Likewise.
3427 (aarch64_movtilow_tilow): Likewise.
3428 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
3429 values to config/arm/types.md
3430 (attribute "conds"): Update for attribute change.
3431 (anddi3_insn): Likewise.
3432 (iordi3_insn): Likewise.
3433 (xordi3_insn): Likewise.
3434 (one_cmpldi2): Likewise.
3435 * config/arm/types.md (type): Add Neon types.
3436 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
3437 use "type" attribute.
3438 (movmisalign<mode>_neon_store): Likewise.
3439 (movmisalign<mode>_neon_load): Likewise.
3440 (vec_set<mode>_internal): Likewise.
3441 (vec_setv2di_internal): Likewise.
3442 (vec_extract<mode>): Likewise.
3443 (vec_extractv2di): Likewise.
3444 (add<mode>3_neon): Likewise.
3445 (adddi3_neon): Likewise.
3446 (sub<mode>3_neon): Likewise.
3447 (subdi3_neon): Likewise.
3448 (mul<mode>3_neon): Likewise.
3449 (mul<mode>3add<mode>_neon): Likewise.
3450 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
3451 (fma<VCVTF:mode>4)): Likewise.
3452 (fma<VCVTF:mode>4_intrinsic): Likewise.
3453 (fmsub<VCVTF:mode>4)): Likewise.
3454 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
3455 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
3456 (ior<mode>3): Likewise.
3457 (and<mode>3): Likewise.
3458 (anddi3_neon): Likewise.
3459 (orn<mode>3_neon): Likewise.
3460 (orndi3_neon): Likewise.
3461 (bic<mode>3_neon): Likewise.
3462 (bicdi3_neon): Likewise.
3463 (xor<mode>3): Likewise.
3464 (one_cmpl<mode>2): Likewise.
3465 (abs<mode>2): Likewise.
3466 (neg<mode>2): Likewise.
3467 (umin<mode>3_neon): Likewise.
3468 (umax<mode>3_neon): Likewise.
3469 (smin<mode>3_neon): Likewise.
3470 (smax<mode>3_neon): Likewise.
3471 (vashl<mode>3): Likewise.
3472 (vashr<mode>3_imm): Likewise.
3473 (vlshr<mode>3_imm): Likewise.
3474 (ashl<mode>3_signed): Likewise.
3475 (ashl<mode>3_unsigned): Likewise.
3476 (neon_load_count): Likewise.
3477 (ashldi3_neon_noclobber): Likewise.
3478 (signed_shift_di3_neon): Likewise.
3479 (unsigned_shift_di3_neon): Likewise.
3480 (ashrdi3_neon_imm_noclobber): Likewise.
3481 (lshrdi3_neon_imm_noclobber): Likewise.
3482 (widen_ssum<mode>3): Likewise.
3483 (widen_usum<mode>3): Likewise.
3484 (quad_halves_<code>v4si): Likewise.
3485 (quad_halves_<code>v4sf): Likewise.
3486 (quad_halves_<code>v8hi): Likewise.
3487 (quad_halves_<code>v16qi): Likewise.
3488 (reduc_splus_v2di): Likewise.
3489 (neon_vpadd_internal<mode>): Likewise.
3490 (neon_vpsmin<mode>): Likewise.
3491 (neon_vpsmax<mode>): Likewise.
3492 (neon_vpumin<mode>): Likewise.
3493 (neon_vpumax<mode>): Likewise.
3494 (ss_add<mode>_neon): Likewise.
3495 (us_add<mode>_neon): Likewise.
3496 (ss_sub<mode>_neon): Likewise.
3497 (us_sub<mode>_neon): Likewise.
3498 (neon_vadd<mode>_unspec): Likewise.
3499 (neon_vaddl<mode>): Likewise.
3500 (neon_vaddw<mode>): Likewise.
3501 (neon_vhadd<mode>): Likewise.
3502 (neon_vqadd<mode>): Likewise.
3503 (neon_vaddhn<mode>): Likewise.
3504 (neon_vmul<mode>): Likewise.
3505 (neon_vmla<mode>): Likewise.
3506 (neon_vmlal<mode>): Likewise.
3507 (neon_vmls<mode>): Likewise.
3508 (neon_vmlsl<mode>): Likewise.
3509 (neon_vqdmulh<mode>): Likewise.
3510 (neon_vqdmlal<mode>): Likewise.
3511 (neon_vqdmlsl<mode>): Likewise.
3512 (neon_vmull<mode>): Likewise.
3513 (neon_vqdmull<mode>): Likewise.
3514 (neon_vsub<mode>_unspec): Likewise.
3515 (neon_vsubl<mode>): Likewise.
3516 (neon_vsubw<mode>): Likewise.
3517 (neon_vqsub<mode>): Likewise.
3518 (neon_vhsub<mode>): Likewise.
3519 (neon_vsubhn<mode>): Likewise.
3520 (neon_vceq<mode>): Likewise.
3521 (neon_vcge<mode>): Likewise.
3522 (neon_vcgeu<mode>): Likewise.
3523 (neon_vcgt<mode>): Likewise.
3524 (neon_vcgtu<mode>): Likewise.
3525 (neon_vcle<mode>): Likewise.
3526 (neon_vclt<mode>): Likewise.
3527 (neon_vcage<mode>): Likewise.
3528 (neon_vcagt<mode>): Likewise.
3529 (neon_vtst<mode>): Likewise.
3530 (neon_vabd<mode>): Likewise.
3531 (neon_vabdl<mode>): Likewise.
3532 (neon_vaba<mode>): Likewise.
3533 (neon_vabal<mode>): Likewise.
3534 (neon_vmax<mode>): Likewise.
3535 (neon_vmin<mode>): Likewise.
3536 (neon_vpaddl<mode>): Likewise.
3537 (neon_vpadal<mode>): Likewise.
3538 (neon_vpmax<mode>): Likewise.
3539 (neon_vpmin<mode>): Likewise.
3540 (neon_vrecps<mode>): Likewise.
3541 (neon_vrsqrts<mode>): Likewise.
3542 (neon_vqabs<mode>): Likewise.
3543 (neon_vqneg<mode>): Likewise.
3544 (neon_vcls<mode>): Likewise.
3545 (clz<mode>2): Likewise.
3546 (popcount<mode>2): Likewise.
3547 (neon_vrecpe): Likewise.
3548 (neon_vrsqrte): Likewise.
3549 (neon_vget_lane<mode>_sext_internal): Likewise.
3550 (neon_vget_lane<mode>_zext_internal): Likewise.
3551 (neon_vdup_n<mode>): Likewise.
3552 (neon_vdup_nv2di): Likewise.
3553 (neon_vdpu_lane<mode>_internal): Likewise.
3554 (neon_vswp<mode>): Likewise.
3555 (float<mode><V_cvtto>2): Likewise.
3556 (floatuns<mode><V_cvtto>2): Likewise.
3557 (fix_trunc<mode><V_cvtto>)2): Likewise
3558 (fixuns_trunc<mode><V_cvtto)2): Likewise.
3559 (neon_vcvt<mode>): Likewise.
3560 (neon_vcvtv4sfv4hf): Likewise.
3561 (neon_vcvtv4hfv4sf): Likewise.
3562 (neon_vcvt_n<mode>): Likewise.
3563 (neon_vmovn<mode>): Likewise.
3564 (neon_vqmovn<mode>): Likewise.
3565 (neon_vqmovun<mode>): Likewise.
3566 (neon_vmovl<mode>): Likewise.
3567 (neon_vmul_lane<mode>): Likewise.
3568 (neon_vmull_lane<mode>): Likewise.
3569 (neon_vqdmull_lane<mode>): Likewise.
3570 (neon_vqdmulh_lane<mode>): Likewise.
3571 (neon_vmla_lane<mode>): Likewise.
3572 (neon_vmlal_lane<mode>): Likewise.
3573 (neon_vqdmlal_lane<mode>): Likewise.
3574 (neon_vmls_lane<mode>): Likewise.
3575 (neon_vmlsl_lane<mode>): Likewise.
3576 (neon_vqdmlsl_lane<mode>): Likewise.
3577 (neon_vext<mode>): Likewise.
3578 (neon_vrev64<mode>): Likewise.
3579 (neon_vrev32<mode>): Likewise.
3580 (neon_vrev16<mode>): Likewise.
3581 (neon_vbsl<mode>_internal): Likewise.
3582 (neon_vshl<mode>): Likewise.
3583 (neon_vqshl<mode>): Likewise.
3584 (neon_vshr_n<mode>): Likewise.
3585 (neon_vshrn_n<mode>): Likewise.
3586 (neon_vqshrn_n<mode>): Likewise.
3587 (neon_vqshrun_n<mode>): Likewise.
3588 (neon_vshl_n<mode>): Likewise.
3589 (neon_vqshl_n<mode>): Likewise.
3590 (neon_vqshlu_n<mode>): Likewise.
3591 (neon_vshll_n<mode>): Likewise.
3592 (neon_vsra_n<mode>): Likewise.
3593 (neon_vsri_n<mode>): Likewise.
3594 (neon_vsli_n<mode>): Likewise.
3595 (neon_vtbl1v8qi): Likewise.
3596 (neon_vtbl2v8qi): Likewise.
3597 (neon_vtbl3v8qi): Likewise.
3598 (neon_vtbl4v8qi): Likewise.
3599 (neon_vtbx1v8qi): Likewise.
3600 (neon_vtbx2v8qi): Likewise.
3601 (neon_vtbx3v8qi): Likewise.
3602 (neon_vtbx4v8qi): Likewise.
3603 (neon_vtrn<mode>_internal): Likewise.
3604 (neon_vzip<mode>_internal): Likewise.
3605 (neon_vuzp<mode>_internal): Likewise.
3606 (neon_vld1<mode>): Likewise.
3607 (neon_vld1_lane<mode>): Likewise.
3608 (neon_vld1_dup<mode>): Likewise.
3609 (neon_vld1_dupv2di): Likewise.
3610 (neon_vst1<mode>): Likewise.
3611 (neon_vst1_lane<mode>): Likewise.
3612 (neon_vld2<mode>): Likewise.
3613 (neon_vld2_lane<mode>): Likewise.
3614 (neon_vld2_dup<mode>): Likewise.
3615 (neon_vst2<mode>): Likewise.
3616 (neon_vst2_lane<mode>): Likewise.
3617 (neon_vld3<mode>): Likewise.
3618 (neon_vld3qa<mode>): Likewise.
3619 (neon_vld3qb<mode>): Likewise.
3620 (neon_vld3_lane<mode>): Likewise.
3621 (neon_vld3_dup<mode>): Likewise.
3622 (neon_vst3<mode>): Likewise.
3623 (neon_vst3qa<mode>): Likewise.
3624 (neon_vst3qb<mode>): Likewise.
3625 (neon_vst3_lane<mode>): Likewise.
3626 (neon_vld4<mode>): Likewise.
3627 (neon_vld4qa<mode>): Likewise.
3628 (neon_vld4qb<mode>): Likewise.
3629 (neon_vld4_lane<mode>): Likewise.
3630 (neon_vld4_dup<mode>): Likewise.
3631 (neon_vst4<mode>): Likewise.
3632 (neon_vst4qa<mode>): Likewise.
3633 (neon_vst4qb<mode>): Likewise.
3634 (neon_vst4_lane<mode>): Likewise.
3635 (neon_vec_unpack<US>_lo_<mode>): Likewise.
3636 (neon_vec_unpack<US>_hi_<mode>): Likewise.
3637 (neon_vec_<US>mult_lo_<mode>): Likewise.
3638 (neon_vec_<US>mult_hi_<mode>): Likewise.
3639 (neon_vec_<US>shiftl_<mode>): Likewise.
3640 (neon_unpack<US>_<mode>): Likewise.
3641 (neon_vec_<US>mult_<mode>): Likewise.
3642 (vec_pack_trunc_<mode>): Likewise.
3643 (neon_vec_pack_trunk_<mode>): Likewise.
3644 (neon_vabd<mode>_2): Likewise.
3645 (neon_vabd<mode>_3): Likewise.
3646 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
3647 (thumb2_movsi_vfp): Likewise.
3648 (movdi_vfp): Likewise.
3649 (movdi_vfp_cortexa8): Likewise.
3650 (movhf_vfp_neon): Likewise.
3651 (movhf_vfp): Likewiwse.
3652 (movsf_vfp): Likewiwse.
3653 (thumb2_movsf_vfp): Likewiwse.
3654 (movdf_vfp): Likewise.
3655 (thumb2_movdf_vfp): Likewise.
3656 (movsfcc_vfp): Likewise.
3657 (thumb2_movsfcc_vfp): Likewise.
3658 (movdfcc_vfp): Likewise.
3659 (thumb2_movdfcc_vfp): Likewise.
3660 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
3661 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
3662 (v10_v2c): Likewise.
3663 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
3665 (cortex_a15_neon_int_2): Likewise.
3666 (cortex_a15_neon_int_3): Likewise.
3667 (cortex_a15_neon_int_4): Likewise.
3668 (cortex_a15_neon_int_5): Likewise.
3669 (cortex_a15_neon_vqneg_vqabs): Likewise.
3670 (cortex_a15_neon_vmov): Likewise.
3671 (cortex_a15_neon_vaba): Likewise.
3672 (cortex_a15_neon_vaba_qqq): Likewise.
3673 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3674 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3675 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
3676 scalar_64_32_long_scalar): Likewise.
3677 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3678 (cortex_a15_neon_mla_qqq_8_16): Likewise.
3679 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
3680 lotype_qdd_64_32_long): Likewise.
3681 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3682 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3683 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
3684 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3685 (cortex_a15_neon_shift_1): Likewise.
3686 (cortex_a15_neon_shift_2): Likewise.
3687 (cortex_a15_neon_shift_3): Likewise.
3688 (cortex_a15_neon_vshl_ddd): Likewise.
3689 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3690 (cortex_a15_neon_vsra_vrsra): Likewise.
3691 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
3692 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
3693 (cortex_a15_neon_fp_vmul_ddd): Likewise.
3694 (cortex_a15_neon_fp_vmul_qqd): Likewise.
3695 (cortex_a15_neon_fp_vmla_ddd): Likewise.
3696 (cortex_a15_neon_fp_vmla_qqq): Likewise.
3697 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
3698 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
3699 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3700 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3701 (cortex_a15_neon_bp_simple): Likewise.
3702 (cortex_a15_neon_bp_2cycle): Likewise.
3703 (cortex_a15_neon_bp_3cycle): Likewise.
3704 (cortex_a15_neon_vld1_1_2_regs): Likewise.
3705 (cortex_a15_neon_vld1_3_4_regs): Likewise.
3706 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3707 (cortex_a15_neon_vld2_4_regs): Likewise.
3708 (cortex_a15_neon_vld3_vld4): Likewise.
3709 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3710 (cortex_a15_neon_vst1_3_4_regs): Likewise.
3711 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
3712 (cortex_a15_neon_vst3_vst4): Likewise.
3713 (cortex_a15_neon_vld1_vld2_lane): Likewise.
3714 (cortex_a15_neon_vld3_vld4_lane" 10
3715 (cortex_a15_neon_vst1_vst2_lane): Likewise.
3716 (cortex_a15_neon_vst3_vst4_lane): Likewise.
3717 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
3718 (cortex_a15_neon_ldm_2): Likewise.0
3719 (cortex_a15_neon_stm_2): Likewise.
3720 (cortex_a15_neon_mcr): Likewise.
3721 (cortex_a15_neon_mcr_2_mcrr): Likewise.
3722 (cortex_a15_neon_mrc): Likewise.
3723 (cortex_a15_neon_mrrc): Likewise.
3724 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
3726 (cortex_a15_alu_shift): Likewise.
3727 (cortex_a15_alu_shift_reg): Likewise.
3728 (cortex_a15_mult32): Likewise.
3729 (cortex_a15_mult64): Likewise.
3730 (cortex_a15_block): Likewise.
3731 (cortex_a15_branch): Likewise.
3732 (cortex_a15_load1): Likewise.
3733 (cortex_a15_load3): Likewise.
3734 (cortex_a15_store1): Likewise.
3735 (cortex_a15_store3): Likewise.
3736 (cortex_a15_call): Likewise.
3737 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
3738 (cortex_a5_f2r): Likewise.
3739 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
3741 (cortex_a53_f2r): Likewise.
3742 * config/arm/cortex-a7.md
3743 (cortex_a7_branch): Update for attribute change.
3744 (cortex_a7_call): Likewise.
3745 (cortex_a7_alu_imm): Likewise.
3746 (cortex_a7_alu_reg): Likewise.
3747 (cortex_a7_alu_shift): Likewise.
3748 (cortex_a7_mul): Likewise.
3749 (cortex_a7_load1): Likewise.
3750 (cortex_a7_store1): Likewise.
3751 (cortex_a7_load2): Likewise.
3752 (cortex_a7_store2): Likewise.
3753 (cortex_a7_load3): Likewise.
3754 (cortex_a7_store3): Likewise.
3755 (cortex_a7_load4): Likewise.
3756 (cortex_a7_store4): Likewise.
3757 (cortex_a7_fpalu): Likewise.
3758 (cortex_a7_fconst): Likewise.
3759 (cortex_a7_fpmuls): Likewise.
3760 (cortex_a7_neon_mul): Likewise.
3761 (cortex_a7_fpmacs): Likewise.
3762 (cortex_a7_neon_mla: Likewise.
3763 (cortex_a7_fpmuld: Likewise.
3764 (cortex_a7_fpmacd: Likewise.
3765 (cortex_a7_fpfmad: Likewise.
3766 (cortex_a7_fdivs: Likewise.
3767 (cortex_a7_fdivd: Likewise.
3768 (cortex_a7_r2f: Likewise.
3769 (cortex_a7_f2r: Likewise.
3770 (cortex_a7_f_flags: Likewise.
3771 (cortex_a7_f_loads: Likewise.
3772 (cortex_a7_f_loadd: Likewise.
3773 (cortex_a7_f_stores: Likewise.
3774 (cortex_a7_f_stored: Likewise.
3775 (cortex_a7_neon): Likewise.
3776 * config/arm/cortex-a8-neon.md
3777 (cortex_a8_neon_mrc): Update for attribute change.
3778 (cortex_a8_neon_mrrc): Likewise.
3779 (cortex_a8_neon_int_1): Likewise.
3780 (cortex_a8_neon_int_2): Likewise.
3781 (cortex_a8_neon_int_3): Likewise.
3782 (cortex_a8_neon_int_4): Likewise.
3783 (cortex_a8_neon_int_5): Likewise.
3784 (cortex_a8_neon_vqneg_vqabs): Likewise.
3785 (cortex_a8_neon_vmov): Likewise.
3786 (cortex_a8_neon_vaba): Likewise.
3787 (cortex_a8_neon_vaba_qqq): Likewise.
3788 (cortex_a8_neon_vsma): Likewise.
3789 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3790 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3791 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
3792 long_scalar): Likewise.
3793 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3794 (cortex_a8_neon_mla_qqq_8_16): Likewise.
3795 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
3796 long_scalar_qdd_64_32_long): Likewise.
3797 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3798 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3799 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
3800 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3801 (cortex_a8_neon_shift_1): Likewise.
3802 (cortex_a8_neon_shift_2): Likewise.
3803 (cortex_a8_neon_shift_3): Likewise.
3804 (cortex_a8_neon_vshl_ddd): Likewise.
3805 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3806 (cortex_a8_neon_vsra_vrsra): Likewise.
3807 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
3808 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
3809 (cortex_a8_neon_fp_vsum): Likewise.
3810 (cortex_a8_neon_fp_vmul_ddd): Likewise.
3811 (cortex_a8_neon_fp_vmul_qqd): Likewise.
3812 (cortex_a8_neon_fp_vmla_ddd): Likewise.
3813 (cortex_a8_neon_fp_vmla_qqq): Likewise.
3814 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
3815 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
3816 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3817 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3818 (cortex_a8_neon_bp_simple): Likewise.
3819 (cortex_a8_neon_bp_2cycle): Likewise.
3820 (cortex_a8_neon_bp_3cycle): Likewise.
3821 (cortex_a8_neon_ldr): Likewise.
3822 (cortex_a8_neon_str): Likewise.
3823 (cortex_a8_neon_vld1_1_2_regs): Likewise.
3824 (cortex_a8_neon_vld1_3_4_regs): Likewise.
3825 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3826 (cortex_a8_neon_vld2_4_regs): Likewise.
3827 (cortex_a8_neon_vld3_vld4): Likewise.
3828 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3829 (cortex_a8_neon_vst1_3_4_regs): Likewise.
3830 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
3831 (cortex_a8_neon_vst3_vst4): Likewise.
3832 (cortex_a8_neon_vld1_vld2_lane): Likewise.
3833 (cortex_a8_neon_vld3_vld4_lane): Likewise.
3834 (cortex_a8_neon_vst1_vst2_lane): Likewise.
3835 (cortex_a8_neon_vst3_vst4_lane): Likewise.
3836 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
3837 (cortex_a8_neon_mcr): Likewise.
3838 (cortex_a8_neon_mcr_2_mcrr): Likewise.
3839 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
3840 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
3842 (ca9_neon_mrrc): Likewise.
3843 (cortex_a9_neon_int_1): Likewise.
3844 (cortex_a9_neon_int_2): Likewise.
3845 (cortex_a9_neon_int_3): Likewise.
3846 (cortex_a9_neon_int_4): Likewise.
3847 (cortex_a9_neon_int_5): Likewise.
3848 (cortex_a9_neon_vqneg_vqabs): Likewise.
3849 (cortex_a9_neon_vmov): Likewise.
3850 (cortex_a9_neon_vaba): Likewise.
3851 (cortex_a9_neon_vaba_qqq): Likewise.
3852 (cortex_a9_neon_vsma): Likewise.
3853 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3854 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3855 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
3856 long_scalar): Likewise.
3857 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3858 (cortex_a9_neon_mla_qqq_8_16): Likewise.
3859 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
3860 long_scalar_qdd_64_32_long): Likewise.
3861 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3862 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3863 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
3864 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3865 (cortex_a9_neon_shift_1): Likewise.
3866 (cortex_a9_neon_shift_2): Likewise.
3867 (cortex_a9_neon_shift_3): Likewise.
3868 (cortex_a9_neon_vshl_ddd): Likewise.
3869 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3870 (cortex_a9_neon_vsra_vrsra): Likewise.
3871 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
3872 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
3873 (cortex_a9_neon_fp_vsum): Likewise.
3874 (cortex_a9_neon_fp_vmul_ddd): Likewise.
3875 (cortex_a9_neon_fp_vmul_qqd): Likewise.
3876 (cortex_a9_neon_fp_vmla_ddd): Likewise.
3877 (cortex_a9_neon_fp_vmla_qqq): Likewise.
3878 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
3879 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
3880 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3881 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3882 (cortex_a9_neon_bp_simple): Likewise.
3883 (cortex_a9_neon_bp_2cycle): Likewise.
3884 (cortex_a9_neon_bp_3cycle): Likewise.
3885 (cortex_a9_neon_ldr): Likewise.
3886 (cortex_a9_neon_str): Likewise.
3887 (cortex_a9_neon_vld1_1_2_regs): Likewise.
3888 (cortex_a9_neon_vld1_3_4_regs): Likewise.
3889 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3890 (cortex_a9_neon_vld2_4_regs): Likewise.
3891 (cortex_a9_neon_vld3_vld4): Likewise.
3892 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3893 (cortex_a9_neon_vst1_3_4_regs): Likewise.
3894 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
3895 (cortex_a9_neon_vst3_vst4): Likewise.
3896 (cortex_a9_neon_vld1_vld2_lane): Likewise.
3897 (cortex_a9_neon_vld3_vld4_lane): Likewise.
3898 (cortex_a9_neon_vst1_vst2_lane): Likewise.
3899 (cortex_a9_neon_vst3_vst4_lane): Likewise.
3900 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
3901 (cortex_a9_neon_mcr): Likewise.
3902 (cortex_a9_neon_mcr_2_mcrr): Likewise.
3903 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
3904 (cortex_a9_fps): Likewise.
3905 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
3907 (cortex_m4_fmuls): Likewise.
3908 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
3910 (cortex_r4_mrc): Likewise.
3911 * config/arm/iterators.md: Update comment referring to neon_type.
3912 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
3913 (iwmmxt_movsi_insn): Likewise.
3914 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
3916 (pj4_core_to_vfp): Likewise.
3917 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
3919 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
3920 (vfp_fstore): Likewise.
3921 * doc/md.texi: Change references to neon_type to refer to type.
3923 2013-09-04 Dodji Seketeli <dodji@redhat.com>
3925 * tree.h (DECL_BUILT_IN): Fix typo in comment.
3927 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
3929 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
3932 2013-09-04 Easwaran Raman <eraman@google.com>
3935 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
3936 (build_and_add_sum): Use it.
3937 (appears_later_in_bb): Simplify code.
3939 2013-09-04 Teresa Johnson <tejohnson@google.com>
3941 * dumpfile.c (dump_finish): Don't close stderr/stdout.
3943 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
3945 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
3947 2013-09-04 Jan Hubicka <jh@suse.cz>
3949 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
3950 * ipa-devirt.c: Include diganostic.h
3951 (odr_type_d): Add types and types_set.
3952 (hash_type_name): Work for types with vtables during LTO.
3953 (odr_hasher::remove): Fix comment; destroy types_set.
3954 (add_type_duplicate): New function,
3955 (get_odr_type): Use it.
3956 (dump_type_inheritance_graph): Dump type duplicates.
3957 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
3959 * tree.c (types_same_for_odr): Give exact answers on types with
3962 2013-09-04 Dodji Seketeli <dodji@redhat.com>
3964 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
3965 explaining their differences.
3967 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
3969 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
3971 2013-09-03 Jeff Law <law@redhat.com>
3973 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
3974 when not threading through a joiner block. Pass joiner/no joiner
3975 state to register_jump_thread.
3976 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
3977 state from argument rather than implying on path length.
3978 Dump the entire jump thread path into debugging dump.
3979 * tree-flow.h (register_jump_thread): Update prototype.
3981 2013-08-29 Xinliang David Li <davidxl@google.com>
3983 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
3984 Remove a trivial gcc_assert.
3986 2013-08-29 Xinliang David Li <davidxl@google.com>
3988 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
3989 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
3990 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
3991 Delay base decl alignment adjustment.
3992 * tree-vectorizer.c (vect_destroy_datarefs): New function.
3993 * tree-vectorizer.h: New data structure.
3994 (set_dr_misalignment): New function.
3995 (dr_misalignment): Ditto.
3996 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
3997 (vectorizable_load): Ditto.
3998 (ensure_base_align): New function.
3999 (vectorize_loops): Add dbg_cnt support.
4000 (execute_vect_slp): Ditto.
4001 * dbgcnt.def: New debug counter.
4002 * Makefile: New dependency.
4004 2013-09-03 Meador Inge <meadori@codesourcery.com>
4008 2013-08-30 Meador Inge <meadori@codesourcery.com>
4010 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
4012 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
4014 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
4015 function descriptor.
4017 2013-09-03 Richard Biener <rguenther@suse.de>
4019 * tree-affine.c (add_elt_to_tree): Fix association issue,
4020 avoid useless converts and make sure to always return a
4021 properly typed result.
4023 2013-09-03 Richard Biener <rguenther@suse.de>
4026 * fold-const.c (negate_expr_p): Fix division case.
4027 (negate_expr): Likewise.
4029 2013-09-03 Richard Biener <rguenther@suse.de>
4032 * tree-streamer-out.c: Include tm.h.
4033 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
4035 2013-09-03 Jan Hubicka <jh@suse.cz>
4037 * tree-profile.c (tree_profiling): Cleanup CFG when done.
4039 2013-09-03 Alan Modra <amodra@gmail.com>
4041 * config.gcc (powerpc*-*-linux*): Add support for little-endian
4042 multilibs to big-endian target and vice versa.
4043 * config/rs6000/t-linux64: Use := assignment on all vars.
4044 (MULTILIB_EXTRA_OPTS): Remove fPIC.
4045 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
4046 * config/rs6000/t-linux64le: New file.
4047 * config/rs6000/t-linux64bele: New file.
4048 * config/rs6000/t-linux64lebe: New file.
4050 2013-09-02 Jan Hubicka <jh@suse.cz>
4052 * ipa-inline-transform.c (inline_transform): Do not
4053 optimize_inline_calls when not optimizing.
4055 2013-09-02 Jan Hubicka <jh@suse.cz>
4057 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
4058 duplicated nodes for assembler names.
4059 * symtab.c (symtab_unregister_node): Do not attempt to unlink
4060 hard registers from assembler name hash.
4062 2013-09-02 Jan Hubicka <jh@suse.cz>
4064 * ipa-split.c (execute_split_functions): Split externally visible
4065 functions called once.
4067 2013-09-02 Martin Jambor <mjambor@suse.cz>
4070 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
4071 linked list. When finding the correct duplicate, also consider also
4072 the caller in additon to its inlined_to node.
4074 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
4076 * config/aarch64/aarch64-simd-builtins.def
4077 (dup_lane_scalar): Remove.
4078 * config/aarch64/aarch64-simd.md
4079 (aarch64_simd_dup): Add 'w->w' alternative.
4080 (aarch64_dup_lane<mode>): Allow for VALL.
4081 (aarch64_dup_lane_scalar<mode>): Remove.
4082 (aarch64_dup_lane_<vswap_width_name><mode>): New.
4083 (aarch64_get_lane_signed<mode>): Add w->w altenative.
4084 (aarch64_get_lane_unsigned<mode>): Likewise.
4085 (aarch64_get_lane<mode>): Likewise.
4086 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
4087 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
4088 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
4089 (VCON): Change container of V2SF.
4090 (vswap_width_name): Likewise.
4091 * config/aarch64/arm_neon.h
4092 (__aarch64_vdup_lane_any): New.
4093 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
4094 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
4095 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
4097 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
4100 * expr.c (emit_move_complex): Do not move complex FP values as parts if
4101 the source or the destination is a single hard register.
4103 2013-09-02 Richard Biener <rguenther@suse.de>
4106 * tree-scalar-evolution.c (instantiate_scev_name): Allow
4109 2013-09-02 Richard Biener <rguenther@suse.de>
4111 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
4112 arithmetic to sizetype.
4114 2013-09-02 Bin Cheng <bin.cheng@arm.com>
4116 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
4117 Find auto-increment use both before and after candidate.
4119 2013-09-02 Marek Polacek <polacek@redhat.com>
4121 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
4123 2013-09-01 Jan Hubicka <jh@suse.cz>
4125 * Makefile.in: Add ipa-profile.o
4126 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
4127 * cgraph.c (struct cgraph_propagate_frequency_data,
4128 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
4129 ipa-profile.c; replace cgraph_ by ipa_ prefix.
4130 * cgraph.h (cgraph_propagate_frequency): Remove.
4131 * ipa-inline-analysis.c: Include ipa-utils.h;
4132 drop duplicated cfgloop.h.
4133 (inline_update_callee_summaries): Update.
4134 * ipa-profile.c: New file.
4135 * ipa-utils.h (ipa_propagate_frequency): Declare.
4136 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
4137 data-streamer.h, value-prof.h.
4138 (symtab_remove_unreachable_nodes): Update profile.
4139 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
4140 account_time_size, cmp_counts, dump_histogram,
4141 ipa_profile_generate_summary, ipa_profile_write_summary,
4142 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
4143 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
4144 Move to ipa-profile.c.
4146 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
4148 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
4150 2013-09-01 Jan Hubicka <jh@suse.cz>
4152 * common.opt (fdevirtualize-speculatively): New function.
4153 * invoke.texi (fdevirtualize-speculatively): Document.
4154 * ipa-devirt.c: Include ipa-inline.h
4155 (likely_target_p): New function.
4156 (ipa_devirt): New function.
4157 (gate_ipa_devirt): New function.
4158 (pass_data_ipa_devirt): New static var.
4159 (pass_ipa_devirt): Likewise.
4160 (make_pass_ipa_devirt): New function.
4161 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
4162 (common_handle_option): Disable devirtualization when
4163 value range profiling is available.
4164 * passes.def (pass_ipa_devirt): Add.
4165 * timever.def (TV_IPA_DEVIRT): New timevar.
4166 * tree-pass.h (make_pass_ipa_devirt):
4168 2013-09-01 Iain Sandoe <iain@codesourcery.com>
4170 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
4171 include sanitize(undefined).
4173 2013-08-31 Diego Novillo <dnovillo@google.com>
4175 * Makefile.in (TREE_CORE_H): Define.
4177 (GTFILES): Add tree-core.h.
4178 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
4180 * tree-core.h: New file.
4181 Move all data structures, enum, typedefs, global
4182 declarations and constants from ...
4185 2013-08-31 Jan Hubicka <jh@suse.cz>
4187 * bulitins.c (expand_builtin): Do not early exit for gcov
4188 instrumented functions.
4190 2013-08-31 Marek Polacek <polacek@redhat.com>
4192 * ubsan.c: Include tm_p.h.
4194 2013-08-31 Jan Hubicka <jh@suse.cz>
4196 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
4199 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
4200 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
4201 * tree-cfg.c (verify_gimple_label): ... here.
4202 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
4203 (ipa_merge_profiles): New function.
4204 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
4205 (lto_input_function_body): Likewise.
4206 * ipa-utils.h (ipa_merge_profiles): Declare.
4207 * lto-streamer.h (lto_input_function_body): Update prototype.
4208 (emit_label_in_global_context_p): Remove.
4209 * lto-symtab.c: Include ipa-utils.h
4210 (lto_cgraph_replace_node): Use ipa_merge_profiles.
4212 2013-08-31 Jan Hubicka <jh@suse.cz>
4214 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
4216 2013-08-31 Jan Hubicka <jh@suse.cz>
4218 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
4220 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
4222 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
4223 "cmp" RTX before signed_comparison_operator check to account
4226 2013-08-30 Jan Hubicka <jh@suse.cz>
4228 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
4229 (detect_type_change_1): Rename to ...
4230 (detect_type_change): ... this one; early return on non-polymorphic
4232 (detect_type_change_ssa): Add comp_type parameter; update
4233 use of detect_type_change.
4234 (compute_complex_assign_jump_func): Add param_type parameter;
4235 update use of detect_type_change_ssa.
4236 (compute_complex_ancestor_jump_func): Likewise.
4237 (ipa_get_callee_param_type): New function.
4238 (ipa_compute_jump_functions_for_edge): Compute parameter type;
4239 update calls to the jump function computation functions.
4241 2013-08-30 Teresa Johnson <tejohnson@google.com>
4242 Steven Bosscher <steven@gcc.gnu.org>
4244 * cfgrtl.c (fixup_new_cold_bb): New routine.
4245 (commit_edge_insertions): Invoke fixup_partitions.
4246 (find_partition_fixes): New routine.
4247 (fixup_partitions): Ditto.
4248 (verify_hot_cold_block_grouping): Update comments.
4249 (rtl_verify_edges): Invoke find_partition_fixes.
4250 (rtl_verify_bb_pointers): Update comments.
4251 (rtl_verify_bb_layout): Ditto.
4252 * basic-block.h (probably_never_executed_edge_p): Declare.
4253 (fixup_partitions): Ditto.
4254 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
4255 * bb-reorder.c (sanitize_hot_paths): New function.
4256 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
4258 * predict.c (probably_never_executed_edge_p): New routine.
4259 * cfg.c (check_bb_profile): Add partition insanity warnings.
4261 2013-08-30 Meador Inge <meadori@codesourcery.com>
4263 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
4265 2013-08-30 Marek Polacek <polacek@redhat.com>
4267 * Makefile.in (ubsan.o): Add.
4268 (c-family/c-ubsan.o): Add.
4269 (builtins.o): Add ubsan.h dependency.
4270 * ubsan.h: New file.
4271 * ubsan.c: New file.
4272 * common.opt: Add -fsanitize=undefined option.
4273 (flag_sanitize): Add variable.
4274 (fsanitize=): Add option. Add Driver.
4275 (fsanitize=thread): Remove option.
4276 (fsanitize=address): Likewise.
4277 (static-libubsan): New option.
4278 * doc/invoke.texi: Document the new flag and -static-libubsan.
4279 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
4280 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
4281 * builtin-attrs.def (ATTR_COLD): Define.
4282 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
4283 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
4284 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
4285 * flag-types.h (sanitize_code): New enum.
4286 * opts.c (common_handle_option): Parse command line arguments
4287 of -fsanitize=. Add -fsanitize=unreachable option.
4288 * varasm.c (get_variable_section): Adjust.
4289 (assemble_noswitch_variable): Likewise.
4290 (assemble_variable): Likewise.
4291 (output_constant_def_contents): Likewise.
4292 (categorize_decl_for_section): Likewise.
4293 (place_block_symbol): Likewise.
4294 (output_object_block): Likewise.
4295 * builtins.def: Likewise.
4296 * toplev.c (compile_file): Likewise.
4297 (process_options): Likewise.
4298 * cppbuiltin.c: Likewise.
4299 * tsan.c (tsan_pass): Likewise.
4300 (tsan_gate): Likewise.
4301 (tsan_gate_O0): Likewise.
4302 * cfgexpand.c (partition_stack_vars): Likewise.
4303 (expand_stack_vars): Likewise.
4304 (defer_stack_allocation): Likewise.
4305 (expand_used_vars): Likewise.
4306 * cfgcleanup.c (old_insns_match_p): Likewise.
4307 * asan.c (asan_finish_file): Likewise.
4308 (asan_instrument): Likewise.
4309 (gate_asan): Likewise.
4310 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
4311 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
4312 (asan_global_struct): Use pointer_sized_int_node instead
4313 calling build_nonstandard_integer_type.
4314 (initialize_sanitizer_builtins): Likewise.
4315 (asan_finish_file): Likewise.
4316 * gcc.c: Document %{%:function(args):X}.
4317 (static_spec_functions): Add sanitize.
4318 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
4319 store funcval != NULL there.
4320 (do_spec_1): Adjust handle_spec_function caller.
4321 (handle_braces): Allow %:function(args) as condition.
4322 (sanitize_spec_function): New function.
4323 (ADD_STATIC_LIBUBSAN_LIBS): Define.
4324 (LIBUBSAN_SPEC): Likewise.
4325 (LIBUBSAN_EARLY_SPEC): Likewise.
4326 (SANITIZER_SPEC): Handle libubsan.
4327 (SANITIZER_EARLY_SPEC): Likewise.
4328 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
4329 instead of fsanitize=address.
4330 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
4331 instead of fsanitize=address*.
4332 * builtins.c: Include ubsan.h.
4333 (fold_builtin_0): Instrument __builtin_unreachable.
4334 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
4335 instead of flag_asan.
4336 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
4337 (pointer_sized_int_node): Define.
4338 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
4340 2013-08-30 Mike Stump <mikestump@comcast.net>
4342 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
4345 2013-08-29 Jan Hubicka <jh@suse.cz>
4347 * cgraph.c (cgraph_function_body_availability): Handle weakref
4349 * passes.def: Remove pass_fixup_cfg.
4350 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
4351 track when we need to remove functions.
4352 (gate_ipa_inline): Execute inlining always; add comment why.
4353 (pass_data_ipa_inline): Remove TODO_remove_functions.
4354 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
4355 do not produce summaries.
4356 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
4357 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
4358 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
4359 constant pool and vtable.
4361 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
4363 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
4364 New arm_neon.h's internal macros to specify 64-bit constants.
4365 Avoid using stdint.h's macros.
4367 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
4369 * recog.c (verify_changes): Verify that changes[i].old is non-zero
4370 before applying REG_P.
4372 2013-08-30 Jakub Jelinek <jakub@redhat.com>
4374 PR tree-optimization/58277
4375 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
4376 after seeing too many stmts with vdef in between dombb and current
4377 bb, invalidate everything.
4379 2013-08-30 Richard Biener <rguenther@suse.de>
4381 * fold-const.c (fold_single_bit_test): Fix overflow test.
4383 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
4385 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
4386 and which can live in a register, always retrieve the value on entry.
4387 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
4388 passed by invisible reference specially.
4389 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
4390 (vt_add_function_parameter): Correctly deal with a parameter passed by
4391 invisible reference.
4393 2013-08-30 Jan Hubicka <jh@suse.cz>
4395 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
4397 2013-08-30 Richard Biener <rguenther@suse.de>
4399 PR tree-optimization/58228
4400 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
4401 allow invariant loads in nested loop vectorization.
4403 2013-08-30 Richard Biener <rguenther@suse.de>
4405 PR tree-optimization/58223
4406 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
4407 (has_anti_or_output_dependence): ... this and adjust to also
4408 look for output dependences.
4409 (mark_nodes_having_upstream_mem_writes): Adjust.
4410 (rdg_flag_uses): Likewise.
4412 2013-08-30 Richard Biener <rguenther@suse.de>
4414 PR tree-optimization/58010
4415 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
4416 assert that we have a loop-closed PHI.
4418 2013-08-29 Jan Hubicka <jh@suse.cz>
4420 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
4421 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
4422 * lto-section-in.c (lto_free_function_in_decl_state): New function.
4423 (lto_free_function_in_decl_state_for_node): New function.
4425 2013-08-29 Xinliang David Li <davidxl@google.com>
4427 * loop-unroll.c (report_unroll_peel): Minor message change.
4428 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
4429 Emit alignment peeling message with default -fopt-info.
4430 (vect_loop_versioning): Emit loop version info message.
4431 * tree-vectorizer.c (vectorize_loops): Minor message change.
4432 (execute_vect_slp): Ditto.
4434 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
4436 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
4437 of the clone from the DECL_NAME of the original function.
4439 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
4441 * passes.c (register_pass): Add overload.
4442 * tree-pass.h (register_pass): Forward declare it. Add comment.
4444 2013-08-29 Jan Hubicka <jh@suse.cz>
4446 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
4447 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
4448 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
4449 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
4450 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
4451 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
4452 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
4453 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
4455 2013-08-29 Teresa Johnson <tejohnson@google.com>
4457 * dumpfile.c (dump_loc): Output column number.
4458 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
4459 * doc/invoke.texi: Document optall -fopt-info flag.
4460 * profile.c (read_profile_edge_counts): Use new dump framework.
4461 (compute_branch_probabilities): Ditto.
4462 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
4463 when pass not in any opt group.
4464 * pass_manager.h (pass_manager::get_pass_profile): New method.
4465 * value-prof.c (check_counter): Use new dump framework.
4466 (check_ic_target): Ditto.
4467 * coverage.c (get_coverage_counts): Ditto.
4468 (coverage_init): Setup new dump framework.
4470 2013-08-29 Richard Biener <rguenther@suse.de>
4472 PR tree-optimization/58246
4473 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
4474 handle the dominance check inside a basic-block.
4476 2013-08-29 Richard Biener <rguenther@suse.de>
4479 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
4480 of default defs that appear in abnormal PHI nodes.
4482 2013-08-29 Richard Biener <rguenther@suse.de>
4484 PR tree-optimization/57685
4485 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
4486 single-use operands to avoid exponential complexity.
4488 2013-08-28 Dehao Chen <dehao@google.com>
4490 * ipa-inline.c (edge_badness): Fix integer underflow.
4492 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
4494 * gtm-builtins.def (_ITM_free): Declare leaf.
4496 2013-08-28 Jakub Jelinek <jakub@redhat.com>
4499 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
4500 (*tls_local_dynamic_base_64_largepic): Likewise.
4501 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
4502 Remove predicate from call operand.
4503 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
4504 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
4506 2013-08-28 Jeff Law <law@redhat.com>
4508 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
4509 checks for the number of predecessors and successors allowed.
4510 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
4511 which require copying a joiner block if there is a request which
4512 is a subpath that requires no joiner block copying.
4514 2013-08-28 Jan Hubicka <jh@suse.cz>
4516 * lto-streamer-out.c (DFS_write_tree_body): Drop
4517 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
4518 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
4519 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
4520 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
4521 Do not read DECL_ERROR_ISSUED.
4522 (unpack_ts_decl_with_vis_value_fields): Do not read
4524 (lto_input_ts_binfo_tree_pointers): Do not read
4525 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
4526 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
4527 write DECL_ERROR_ISSUED..
4528 (pack_ts_decl_with_vis_value_fields): Do not write
4530 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
4531 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
4532 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
4533 * tree.h (tree_decl_common): Update comment.
4534 (DECL_ERROR_ISSUED): Remove.
4536 2013-08-28 Jakub Jelinek <jakub@redhat.com>
4539 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
4541 2013-08-28 Jan Hubicka <jh@suse.cz>
4543 * builtins.def (free): Declare leaf.
4545 2013-08-27 David Malcolm <dmalcolm@redhat.com>
4548 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
4549 * configure: Regenerate.
4551 2013-08-27 Martin Jambor <mjambor@suse.cz>
4553 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
4554 (ipa_ancestor_jf_data): Likewise.
4555 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
4556 (ipa_get_jf_pass_through_type_preserved): New function.
4557 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
4558 (ipa_get_jf_ancestor_type_preserved): New function.
4559 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
4560 (ipa_get_jf_ancestor_result): Likewise.
4561 (propagate_vals_accross_pass_through): Use
4562 ipa_get_jf_pass_through_result to do all the value mappings.
4563 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
4564 type_preserved flag.
4565 (ipa_set_jf_cst_copy): New function.
4566 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
4567 (ipa_set_jf_arith_pass_through): Likewise.
4568 (ipa_set_ancestor_jf): Likewise.
4569 (compute_complex_assign_jump_func): Set type_preserved instead of
4571 (ipa_compute_jump_functions_for_edge): Likewise.
4572 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
4573 (update_jump_functions_after_inlining): Update type_preserved.
4574 Explicitely create jump functions when combining one with pass_through.
4575 (ipa_write_jump_function): Stream the type_preserved flags.
4576 (ipa_read_jump_function): Likewise.
4578 2013-08-27 Jakub Jelinek <jakub@redhat.com>
4579 Aldy Hernandez <aldyh@redhat.com>
4581 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
4582 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
4583 * function.h (struct function): Add has_force_vect_loops and
4585 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
4586 * gimple.c (gimple_build_omp_critical): Add KIND argument and
4588 * gimple.def: Update CLAUSES comments.
4589 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
4590 (gimple_build_omp_for): Add argument to prototype.
4591 (gimple_omp_for_kind): New.
4592 (gimple_omp_for_set_kind): New.
4593 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
4594 GOVD_DATA_SHARE_CLASS.
4595 (enum omp_region_type): Add ORT_SIMD.
4596 (gimple_add_tmp_var): Handle ORT_SIMD.
4597 (gimplify_var_or_parm_decl): Same.
4598 (is_gimple_stmt): Same.
4599 (omp_firstprivatize_variable): Same.
4600 (omp_add_variable): Only use splay_tree_insert if lookup failed.
4601 (omp_notice_variable): Handle ORT_SIMD.
4602 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
4603 (omp_check_private): Handle ORT_SIMD.
4604 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
4606 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
4607 Handle OMP_CLAUSE_LASTPRIVATE.
4608 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
4610 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
4611 (gimplify_expr): Handle OMP_SIMD.
4612 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
4613 (expand_GOMP_SIMD_VF): New.
4614 (expand_GOMP_SIMD_LAST_LANE): New.
4615 * internal-fn.def (GOMP_SIMD_LANE): New.
4616 (GOMP_SIMD_VF): New.
4617 (GOMP_SIMD_LAST_LANE): New.
4618 * omp-low.c: Include target.h.
4619 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
4621 (check_omp_nesting_restrictions): Same.
4623 (lower_rec_simd_input_clauses): New.
4624 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
4626 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
4627 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
4628 (expand_omp_build_assign): New.
4629 (expand_omp_for_init_counts): New.
4630 (expand_omp_for_init_vars): New.
4631 (extract_omp_for_update_vars): New.
4632 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
4633 and rewrite accordingly.
4634 (expand_omp_simd): New.
4635 (expand_omp_for): Use expand_omp_simd.
4636 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
4637 (lower_omp_for): Do not lower the body.
4638 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
4640 * tree-flow.h (find_omp_clause): Remove prototype.
4641 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
4642 forcing vectorization of the loop, or if flag_tree_vectorize.
4643 (gate_tree_if_conversion): Similarly.
4644 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
4645 gimple_build_omp_for.
4646 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
4647 * tree-parloops (create_parallel_loop): Pass kind argument to
4648 gimple_build_omp_for.
4649 * tree-pretty-print.c (dump_omp_clause): Add cases for
4650 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
4651 OMP_CLAUSE__SIMDUID_.
4652 (dump_generic_node): Handle OMP_SIMD.
4653 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
4654 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
4655 unroll OMP_SIMD loops here.
4656 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
4657 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
4659 (vect_analyze_data_refs): Handle simd loops.
4660 * tree-vect-loop.c (vectorizable_live_operation): Handle
4662 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
4663 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
4664 (vectorizable_load): Same.
4665 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
4666 (struct simduid_to_vf): New.
4667 (simduid_to_vf::hash): New.
4668 (simduid_to-vf::equal): New.
4669 (struct simd_array_to_simduid): New.
4670 (simd_array_to_simduid::hash): New.
4671 (simd_array_to_simduid::equal): New.
4672 (adjust_simduid_builtins): New.
4673 (struct note_simd_array_uses_struct): New.
4674 (note_simd_array_uses_cb): New.
4675 (note_simd_array_uses): New.
4676 (vectorize_loops): Handle simd hints and adjust simd builtins
4678 * tree-vectorizer.h (struct _stmt_vec_info): Add
4679 simd_lane_access_p field.
4680 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
4681 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
4682 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
4683 (omp_clause_code_name): Same.
4684 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
4685 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
4686 * tree.def (OMP_SIMD): New entry.
4687 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
4688 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
4689 (OMP_CLAUSE_DECL): Adjust range for new clauses.
4690 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
4691 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
4692 (OMP_CLAUSE_LINEAR_STEP): New.
4693 (OMP_CLAUSE_SAFELEN_EXPR): New.
4694 (OMP_CLAUSE__SIMDUID__DECL): New.
4695 (find_omp_clause): New prototype.
4697 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
4699 * config/i386/driver-i386.c (host_detect_local_cpu): Update
4700 Haswell processor detection.
4702 2013-08-27 Christian Widmer <shadow@umbrox.de>
4705 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
4706 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
4707 AVX2 capable processors.
4709 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
4711 * config/aarch64/arm_neon.h: Replace all inline asm implementations
4712 of vget_low_* with implementations in terms of other intrinsics.
4714 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
4717 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
4718 values to -1, 0 and 1.
4720 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
4722 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
4723 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
4724 (<optab><mode>3_insn): Remove.
4725 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
4726 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
4727 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
4728 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
4729 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
4730 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
4731 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
4732 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
4733 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
4734 (ror<mode>3_insn): Likewise.
4735 * config/aarch64/predicates.md (aarch64_simd_register): New.
4737 2013-08-27 Richard Biener <rguenther@suse.de>
4739 PR tree-optimization/57521
4740 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
4741 one edge is non-critical.
4742 (find_phi_replacement_condition): Make sure to use a non-critical
4743 edge. Cleanup and remove old bug workarounds.
4744 (bb_postdominates_preds): Remove.
4745 (if_convertible_loop_p_1): Do not compute post-dominators.
4746 (combine_blocks): Do not free post-dominators.
4747 (main_tree_if_conversion): Likewise.
4748 (pass_data_if_conversion): Add TODO_verify_ssa.
4750 2013-08-27 DJ Delorie <dj@redhat.com>
4752 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
4754 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
4756 * function.c (assign_parm_find_data_types): Set passed_mode and
4757 nominal_mode to the TYPE_MODE of nominal_type for the built
4758 pointer type in case of the struct-pass-by-reference.
4760 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
4762 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
4763 (UINT16_TYPE): Change default to "unsigned int".
4765 * config/avr/avr.opt (mfract-convert-truncate): New option.
4766 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
4767 is set, round negative fractional integers according to n1169
4768 when converting to integer types.
4770 2013-08-26 Jan Hubicka <jh@suse.cz>
4772 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
4773 methods can not be called indirectly when their address is not taken.
4775 2013-08-26 Jan Hubicka <jh@suse.cz>
4777 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
4780 2013-08-26 Jan Hubicka <jh@suse.cz>
4782 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
4785 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
4787 * reload.c (find_valid_class): Allow classes that do not include
4788 FIRST_PSEUDO_REGISTER - 1.
4790 2013-08-26 Jan Hubicka <jh@suse.cz>
4792 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
4793 fix edge count/frequency when speculation failed; fix type check
4794 for the direct call.
4796 2013-08-26 Jan Hubicka <jh@suse.cz>
4798 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
4800 2013-08-26 Jan Hubicka <jh@suse.cz>
4802 * ipa-inline-transform.c (inline_transform): Be ready for basic block
4803 to be changed by edge redirection.
4805 2013-08-26 Jan Hubicka <jh@suse.cz>
4807 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
4808 formating; add sanity check.
4809 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
4810 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
4811 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
4812 (resolve_noninline_speculation): Update callee keys, too.
4814 2013-08-26 Jan Hubicka <jh@suse.cz>
4816 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
4817 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
4819 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
4821 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
4824 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
4826 * config/i386/i386.c (ix86_debug_options): Remove prototype.
4827 (x86_64_elf_select_section): Ditto.
4828 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
4830 (ix86_pass_by_reference): Ditto.
4831 (output_set_got): Ditto.
4832 (ix86_unary_operator_ok): Ditto.
4833 (ix86_expand_builtin): Ditto.
4835 2013-08-23 Jan Hubicka <jh@suse.cz>
4837 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
4839 2013-08-23 Jan Hubicka <jh@suse.cz>
4841 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
4842 (tree_decl_with_vis): Add FINAL field.
4844 2013-08-23 Jeff Law <law@redhat.com>
4846 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
4847 the debugging dump when the expression is fully redundant.
4849 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
4851 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
4852 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
4853 * pretty-print.c (pp_formatted_text_data): Likewise.
4854 (pp_write_text_to_stream): Likewise.
4855 (pp_write_text_as_dot_label_to_stream): Likewise.
4856 (pp_append_r): Likewise.
4857 (pp_format): Likewise.
4858 (pp_flush): Likewise.
4859 (pp_clear_output_area): Likewise.
4860 (pp_append_text): Likewise.
4861 (pp_formatted_text): Likewise.
4862 (pp_remaining_character_count_for_line): Likewise.
4863 (pp_newline): Likewise.
4864 (pp_character): Likewise.
4865 (output_buffer::~output_buffer): Define.
4866 (pretty_printer::~pretty_printer): Destruct output buffer.
4867 * pretty-print.h (output_buffer::~output_buffer): Declare.
4868 (pretty_printer::~pretty_printer): Declare virtual.
4870 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
4873 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
4874 HOST_WIDE_INT_M1U): New macros.
4875 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
4876 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
4877 unsigned -1 for lshift.
4878 * cse.c (cse_insn): Likewise.
4879 * double-int.c (rshift_double, lshift_double): Likewise.
4880 * builtins.c (fold_builtin_bitop): Likewise.
4881 * combine.c (force_to_mode): Likewise.
4882 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
4883 * simplify-rtx.c (simplify_const_unary_operation,
4884 simplify_const_binary_operation): Likewise.
4885 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
4886 check_va_list_escapes): Likewise.
4887 * rtlanal.c (nonzero_bits1): Likewise.
4888 * expmed.c (expand_smod_pow2): Likewise.
4889 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
4891 2013-08-23 Jan Hubicka <jh@suse.cz>
4893 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
4894 as having address taken.
4896 2013-08-23 Jan Hubicka <jh@suse.cz>
4898 * ipa-utils.h (method_class_type): Declare.
4899 * ipa-devirt.c (method_class_type): Export.
4901 * cgraphunit.c (analyze_functions): Do basic devirtualization;
4902 do not walk base classes of anonymous types.
4904 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
4906 PR rtl-optimization/58220
4908 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
4909 handle SEQUENCE insns properly.
4911 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
4913 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
4915 (pp_newline_and_indent): Likewise.
4916 (pp_separate_with): Likewise.
4917 * pretty-print.c (pp_newline_and_flush): Define.
4918 (pp_newline_and_indent): Likewise.
4919 (pp_separate_with): Likewise.
4921 2013-08-23 Jakub Jelinek <jakub@redhat.com>
4924 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
4925 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
4927 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
4929 * config/i386/predicates.md (ext_sse_reg_operand): New.
4930 * config/i386/i386.md (*movti_internal): Use
4931 predicate to determine if EVEX is needed.
4932 (*movsi_internal): Ditto.
4933 (*movdf_internal): Ditto.
4934 (*movsf_internal): Ditto.
4935 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
4937 2013-08-23 Jakub Jelinek <jakub@redhat.com>
4939 PR tree-optimization/58209
4940 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
4941 (find_tail_calls): Give up for pointer result types if m is non-NULL.
4942 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
4943 emit POINTER_PLUS_EXPR.
4944 (create_tailcall_accumulator): For pointer result type accumulate in
4947 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
4949 * configure.ac: Add backslashes missing from the last change.
4950 * configure: Regenerate.
4952 2013-08-22 Jan Hubicka <jh@susue.cz>
4954 * ipa.c (function_and_variable_visibility): First remember function
4955 was global and then make it local.
4957 2013-08-22 Julian Brown <julian@codesourcery.com>
4959 * configure.ac: Add aarch64 to list of arches which use "nop" in
4961 * configure: Regenerate.
4963 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4965 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
4966 gnu_libc_has_function.
4967 * config/s390/tpf.h: Likewise.
4969 2013-08-22 Jan Hubicka <jh@susue.cz>
4971 * timevar.c (validate_phases): Add cast.
4973 2013-08-22 Jan Hubicka <jh@susue.cz>
4975 * timevar.c (validate_phases): Use size_t for memory.
4976 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
4978 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
4980 * pretty-print.h (output_buffer::output_buffer): Declare.
4981 (pretty_printer::pretty_printer): Likewise.
4982 (pp_construct): Remove.
4983 * pretty-print.c (output_buffer::output_buffer): Define.
4984 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
4985 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
4986 (print_gimple_expr): Likewise.
4987 (print_gimple_seq): Likewise.
4988 (gimple_dump_bb): Likewise.
4989 * sched-vis.c (dump_value_slim): Likewise.
4990 (dump_insn_slim): Likewise.
4991 (dump_rtl_slim): Likewise.
4992 (str_pattern_slim): Likewise.
4993 * tree-mudflap.c (mf_varname_tree): Likewise.
4994 * graph.c (print_graph_cfg): Likewise.
4995 (start_graph_dump): Likewise.
4996 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
4998 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
4999 pretty printer initialization.
5000 * coretypes.h (diagnostic_context): Remove superflous type alias
5002 (pretty_printer): Likewise. Declare directly as a class.
5003 (pretty_print_info): Remove declaration as class.
5004 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
5005 and pp_clear_output_area.
5006 (asan_add_global): Likewise.
5008 2013-08-22 Jan Hubicka <jh@suse.cz>
5010 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
5011 * ipa-utils.h (update_type_inheritance_graph): Declare.
5012 (possible_polymorphic_call_target_p): Declare.
5013 (possible_polymorphic_call_target_p): New.
5014 * ipa-devirt.c: Update toplevel comments.
5015 (cached_polymorphic_call_targets): Move up.
5016 (odr_type_d): Move ID down.
5017 (polymorphic_type_binfo_p): Update comment.
5018 (odr_hasher::remove): Likewise;
5019 (get_odr_type): Set anonymous_namespace.
5020 (dump_odr_type): Dump it.
5021 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
5022 (maybe_record_node): Record node in cached_polymorphic_call_targets.
5023 (record_binfo): Add comment.
5024 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
5026 (devirt_node_removal_hook): Do not iCE when cache is freed.
5027 (possible_polymorphic_call_target_p): New predicate.
5028 (update_type_inheritance_graph): New function.
5030 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
5031 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5032 Sergey Lega <sergey.s.lega@intel.com>
5033 Anna Tikhonova <anna.tikhonova@intel.com>
5034 Ilya Tocar <ilya.tocar@intel.com>
5035 Andrey Turetskiy <andrey.turetskiy@intel.com>
5036 Ilya Verbin <ilya.verbin@intel.com>
5037 Kirill Yukhin <kirill.yukhin@intel.com>
5038 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5040 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
5041 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
5042 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
5043 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
5044 (OPTION_MASK_ISA_AVX2_UNSET): Update.
5045 (OPTION_MASK_ISA_AVX512F_UNSET): New.
5046 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
5047 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
5048 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
5049 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
5050 OPT_mavx512pf, OPT_mavx512er cases.
5051 * config/i386/constraints.md (v): New constraint.
5052 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
5053 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
5054 (bit_AVX512CD): New.
5055 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
5056 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
5057 * config/i386/i386-c.c (ix86_target_macros_internal):
5058 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
5060 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
5061 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
5062 * config/i386/i386.c (regclass_map, dbx_register_map)
5063 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
5064 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
5065 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
5067 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
5068 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
5069 -mavx512pf options. Fix formatting.
5070 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
5071 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
5072 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
5073 -mavx512cd, -mavx512pf options.
5074 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
5075 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
5076 (ix86_preferred_output_reload_class): Replace SSE_REGS with
5078 (ix86_hard_regno_mode_ok): Support 512-bit registers.
5079 (ix86_set_reg_reg_cost): Ditto.
5080 (x86_order_regs_for_local_alloc): Ditto.
5081 (MAX_VECT_LEN): Extend to 64-byte.
5082 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
5083 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
5084 (TARGET_AVX512ER, TARGET_AVX512CD): New.
5085 (BIGGEST_ALIGNMENT): Extend to 512-bits.
5086 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
5087 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
5088 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
5089 (SSE_REG_MODE_P): Support new modes.
5090 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
5091 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
5092 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
5093 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
5094 (REG_CLASS_CONTENTS): Add new registers.
5095 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
5096 (EXT_REX_SSE_REGNO_P): New.
5097 (HI_REGISTER_NAMES): Add new registers.
5098 * config/i386/i386.md: Define constants for new registers.
5099 (mode): Add new 512-bit modes.
5100 (prefix): Support evex prefix.
5101 (isa): Support avx512f, noavx512f, fma_avx512f.
5102 (ssemodesuffix): Add new 512-bit modes.
5104 (*movxi_internal_avx512f): Ditto.
5105 (*movdi_internal): Replace constraint "x" with the new constraint "v".
5107 (*movsi_internal): Likewise.
5108 (*movdf_internal): Likewise.
5109 (*movsf_internal): Likewise.
5110 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
5111 (<code><mode>3): Likewise.
5112 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
5114 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
5115 with the new constraint "v".
5116 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
5118 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
5119 with the new constraint "v".
5120 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
5121 (<sse2>_storedqu<avxsizesuffix>): Likewise.
5122 (*<plusminus_insn><mode>3): Likewise.
5123 (<sse>_vm<plusminus_insn><mode>3): Likewise.
5124 (*mul<mode>3): Likewise.
5125 (<sse>_vmmul<mode>3): Likewise.
5126 (<sse>_div<mode>3): Likewise.
5127 (<sse>_vmdiv<mode>3): Likewise.
5128 (<sse>_sqrt<mode>2): Likewise.
5129 (<sse>_vmsqrt<mode>2): Likewise.
5130 (*<code><mode>3_finite): Likewise.
5131 (*<code><mode>3) <smaxmin>: Likewise.
5132 (<sse>_vm<code><mode>3): Likewise.
5133 (*<code><mode>3) <any_logic>: Likewise.
5134 (*fma_fmadd_<mode>): Likewise.
5135 (*fma_fmsub_<mode>): Likewise.
5136 (*fma_fnmadd_<mode>): Likewise.
5137 (*fma_fnmsub_<mode>): Likewise.
5138 (*fma_fmaddsub_<mode>): Likewise.
5139 (*fma_fmsubadd_<mode>): Likewise.
5140 (*fmai_fmadd_<mode>): Likewise.
5141 (*fmai_fmsub_<mode>): Likewise.
5142 (*fmai_fnmadd_<mode>): Likewise.
5143 (*fmai_fnmsub_<mode>): Likewise.
5144 (sse_cvtsi2ss): Likewise.
5145 (sse_cvtsi2ssq): Likewise.
5146 (sse_cvtss2si): Likewise.
5147 (sse_cvtss2si_2): Likewise.
5148 (sse_cvtss2siq): Likewise.
5149 (sse_cvtss2siq_2): Likewise.
5150 (sse_cvttss2si): Likewise.
5151 (sse_cvtss2siq_2): Likewise.
5152 (float<sseintvecmodelower><mode>2): Likewise.
5153 (sse2_cvtsd2si_2): Likewise.
5154 (sse2_cvtsd2siq_2): Likewise.
5155 (*<plusminus_insn><mode>3): Likewise.
5156 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
5157 (*<sse4_1_avx2>_mul<mode>3): Likewise.
5158 (ashr<mode>3): Likewise.
5159 (<shift_insn><mode>3): Likewise.
5160 (avx2_<code><mode>3): Likewise.
5161 (*avx2_<code><mode>3): Likewise.
5162 (*andnot<mode>3): Likewise.
5163 (*<code><mode>3) <any_logic>: Likewise.
5164 (abs<mode>2): Likewise.
5165 (avx2_permvar<mode>): Likewise.
5166 (avx2_perm<mode>_1): Likewise.
5167 (*avx_vpermilp<mode>): Likewise.
5168 (avx_vpermilvar<mode>3): Likewise.
5169 (avx2_ashrv<mode>): Likewise.
5170 (avx2_<shift_insn>v<mode>): Likewise.
5171 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
5173 * doc/rtl.texi: Document XImode.
5175 2013-08-21 Jeff Law <law@redhat.com>
5177 * tree-flow.h (register_jump_thread): Pass vector of edges
5178 instead of each important edge.
5179 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
5180 thread path into a vector and pass that to register_jump_thread.
5181 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
5182 passed in edge vector to the current 3-edge form.
5185 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
5187 * dce.c (fini_dce): Call df_analyze again just in case
5188 delete_unmarked_insns removed anything.
5190 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
5192 * reload.h (struct reg_equivs): Rename to ..
5193 (struct reg_equivs_s): .. this.
5195 2013-08-20 Martin Liska <marxin.liska@gmail.com>
5197 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
5199 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5201 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
5203 2013-08-21 Jeff Law <law@redhat.com>
5205 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
5206 simplify assignments too. If the RHS collapses to a singleton
5207 range, then return the value for the range.
5209 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
5211 * config/i386/sse.md (V16): Rename to...
5213 (mov<mode>): Update iterator name.
5214 (*mov<mode>_internal): Ditto.
5215 (push<mode>1): Ditto.
5216 (movmisalign<mode>): Ditto.
5218 2013-08-20 Jan Hubicka <jh@suse.cz>
5221 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
5222 entry for direct edges.
5223 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
5225 2013-08-20 David Malcolm <dmalcolm@redhat.com>
5227 Revert my last two changes, r201865 and r201864:
5230 2013-08-20 David Malcolm <dmalcolm@redhat.com>
5232 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
5233 instances can own GC refs.
5235 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
5236 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
5237 (gcc::context::gt_pch_nx): Likewise.
5238 (gcc::context::gt_pch_nx): Likewise.
5239 * ggc.h (gt_ggc_mx <T>): New.
5240 (gt_pch_nx_with_op <T>): New.
5241 (gt_pch_nx <T>): New.
5242 * passes.c (opt_pass::gt_ggc_mx): New.
5243 (opt_pass::gt_pch_nx): New.
5244 (opt_pass::gt_pch_nx_with_op): New.
5245 (pass_manager::gt_ggc_mx): New.
5246 (pass_manager::gt_pch_nx): New.
5247 (pass_manager::gt_pch_nx_with_op): New.
5248 (pass_manager::operator new): Use
5249 ggc_internal_cleared_alloc_stat rather than xcalloc.
5250 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
5251 (pass_manager::gt_ggc_mx): New.
5252 (pass_manager::gt_pch_nx): New.
5253 (pass_manager::gt_pch_nx_with_op): New.
5254 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
5255 (opt_pass::operator new): New.
5256 (opt_pass::gt_ggc_mx): New.
5257 (opt_pass::gt_pch_nx): New.
5258 (opt_pass::gt_pch_nx_with_op): New.
5261 2013-08-20 David Malcolm <dmalcolm@redhat.com>
5263 * Makefile.in (GTFILES): Add context.h.
5264 * context.c (gcc::context::operator new): New.
5265 (gcc::context::gt_ggc_mx): New.
5266 (gcc::context::gt_pch_nx): New.
5267 (gcc::context::gt_pch_nx): New.
5268 * context.h (gcc::context): Add GTY((user)) marking.
5269 (gcc::context::operator new): New.
5270 (gcc::context::gt_ggc_mx): New.
5271 (gcc::context::gt_pch_nx): New.
5272 (gcc::context::gt_pch_nx): New.
5273 (g): Add GTY marking.
5274 (gt_ggc_mx (gcc::context *)): New.
5275 (gt_pch_nx (gcc::context *)): New.
5276 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
5277 void *cookie)): New.
5278 * gengtype.c (open_base_files) <ifiles>: Add context.h.
5280 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
5282 * dce.c (fini_dce): Call df_analyze again just in case
5283 delete_unmarked_insns removed anything.
5285 2013-08-20 Teresa Johnson <tejohnson@google.com>
5287 PR rtl-optimizations/57451
5288 * final.c (reemit_insn_block_notes): Prevent lexical blocks
5289 from crossing split section boundaries.
5291 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
5293 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
5294 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
5295 with MULTLIB_DEFAULTS.
5297 2013-08-20 Nick Clifton <nickc@redhat.com>
5299 * target.def (narrow_volatile_bitfield): Note that the default
5300 value is false, not !TARGET_STRICT_ALIGN.
5301 * doc/tm.texi: Regenerate.
5303 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
5305 Fix LIB_SPEC for systems without libpthread.
5307 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
5308 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
5310 * config/i386/linux-common.h: Likewise.
5311 * config/mips/linux-common.h: Likewise.
5313 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
5315 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
5318 2013-08-20 David Malcolm <dmalcolm@redhat.com>
5320 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
5321 instances can own GC refs.
5323 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
5324 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
5325 (gcc::context::gt_pch_nx): Likewise.
5326 (gcc::context::gt_pch_nx): Likewise.
5327 * ggc.h (gt_ggc_mx <T>): New.
5328 (gt_pch_nx_with_op <T>): New.
5329 (gt_pch_nx <T>): New.
5330 * passes.c (opt_pass::gt_ggc_mx): New.
5331 (opt_pass::gt_pch_nx): New.
5332 (opt_pass::gt_pch_nx_with_op): New.
5333 (pass_manager::gt_ggc_mx): New.
5334 (pass_manager::gt_pch_nx): New.
5335 (pass_manager::gt_pch_nx_with_op): New.
5336 (pass_manager::operator new): Use
5337 ggc_internal_cleared_alloc_stat rather than xcalloc.
5338 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
5339 (pass_manager::gt_ggc_mx): New.
5340 (pass_manager::gt_pch_nx): New.
5341 (pass_manager::gt_pch_nx_with_op): New.
5342 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
5343 (opt_pass::operator new): New.
5344 (opt_pass::gt_ggc_mx): New.
5345 (opt_pass::gt_pch_nx): New.
5346 (opt_pass::gt_pch_nx_with_op): New.
5348 2013-08-20 David Malcolm <dmalcolm@redhat.com>
5350 * Makefile.in (GTFILES): Add context.h.
5351 * context.c (gcc::context::operator new): New.
5352 (gcc::context::gt_ggc_mx): New.
5353 (gcc::context::gt_pch_nx): New.
5354 (gcc::context::gt_pch_nx): New.
5355 * context.h (gcc::context): Add GTY((user)) marking.
5356 (gcc::context::operator new): New.
5357 (gcc::context::gt_ggc_mx): New.
5358 (gcc::context::gt_pch_nx): New.
5359 (gcc::context::gt_pch_nx): New.
5360 (g): Add GTY marking.
5361 (gt_ggc_mx (gcc::context *)): New.
5362 (gt_pch_nx (gcc::context *)): New.
5363 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
5364 void *cookie)): New.
5365 * gengtype.c (open_base_files) <ifiles>: Add context.h.
5367 2013-08-20 Alan Modra <amodra@gmail.com>
5370 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
5371 (rs6000_emit_epilogue): Likewise.
5373 2013-08-19 Dehao Chen <dehao@google.com>
5375 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
5377 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
5378 Jakub Jelinek <jakub@redhat.com>
5380 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
5381 (BUILT_IN_FABSD64): Likewise.
5382 (BUILT_IN_FABSD128): Likewise.
5383 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
5384 (fold_builtin_1): Likewise.
5385 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
5386 destination and source operands.
5387 (*abstd2_fpr): Likewise.
5388 (*nabstd2_fpr): Likewise.
5390 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
5392 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
5395 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
5397 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
5400 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
5402 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
5403 * builtins.c (default_libc_has_function): New.
5404 (gnu_libc_has_function): Ditto.
5405 (no_c99_libc_has_function): Ditto.
5406 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
5407 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
5408 (fold_builtin_sincos): Likewise.
5409 (fold_builtin_cexp): Likewise.
5410 * builtins.def (DEF_C94_BUILTIN): Likewise.
5411 (DEF_C99_BUILTIN): Likewise.
5412 (DEF_C99_C90RES_BUILTIN): Likewise.
5413 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
5414 definitions to using this define.
5415 * config/darwin-protos.h (darwin_libc_has_function): New.
5416 * config/darwin.c (darwin_libc_has_function): Ditto.
5417 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
5418 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
5419 * config/darwin.h: Ditto.
5420 * config/elfos.h: Ditto.
5421 * config/freebsd.h: Ditto.
5422 * config/i386/cygming.h: Ditto.
5423 * config/i386/djgpp.h: Ditto.
5424 * config/i386/i386-interix.h: Ditto.
5425 * config/microblaze/microblaze.h: Ditto.
5426 * config/mmix/mmix.h: Ditto.
5427 * config/gnu-user.h: Ditto.
5428 * config/ia64/hpux.h: Ditto.
5429 * config/pa/pa-hpux.h: Ditto.
5430 * config/pdp11/pdp11.h: Ditto.
5431 * config/picochip/picochip.h: Ditto.
5432 * config/linux.h: Ditto.
5433 * config/netbsd.h: Ditto.
5434 * config/openbsd.h: Ditto.
5435 * config/rs6000/aix43.h: Ditto.
5436 * config/rs6000/aix51.h: Ditto.
5437 * config/rs6000/aix52.h: Ditto.
5438 * config/rs6000/aix53.h: Ditto.
5439 * config/rs6000/aix61.h: Ditto.
5440 * config/rs6000/darwin.h: Ditto.
5441 * config/rs6000/linux.h: Ditto.
5442 * config/rs6000/linux64.h: Ditto.
5443 * config/s390/tpf.h: Ditto.
5444 * config/sol2-10.h: Ditto.
5445 * config/sol2.h: Ditto.
5446 * config/vms/vms.h: Ditto.
5447 * config/vxworks.h: Ditto.
5448 * config/linux-android.c (linux_android_libc_has_function):
5449 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
5450 * config/linux-protos.h (linux_android_libc_has_function):
5452 * config/i386/i386.c (ix86_libc_has_function): New.
5453 * config/i386/i386-protos.h
5454 (ix86_libc_has_function): New declaration.
5455 * config/i386/i386.md
5456 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
5457 ("isinf<mode>2): Likewise.
5458 * convert.c (convert_to_integer): Using new target hook
5459 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
5460 TARGET_C99_FUNCTIONS.
5461 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
5462 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
5463 * coretypes.h (function_class): New enum for different
5464 classes of functions.
5465 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
5466 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
5467 (TARGET_HAS_SINCOS): Likewise.
5468 (TARGET_LIBC_HAS_FUNCTION): New.
5469 * doc/tm.texi: Regenerated.
5470 * targhooks.h (default_libc_has_function): New declaration.
5471 (no_c99_libc_has_function): Ditto.
5472 (gnu_libc_has_function): Ditto.
5473 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
5474 and TARGET_HAS_SINCOS.
5476 2013-08-18 Jan Hubicka <jh@suse.cz>
5478 * Makeifle-in (ipa-devirt.o): New.
5479 (GTFILES): Add ipa-utils.h and ipa-devirt.c
5480 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
5481 (analyze_functions): Look into possible targets of polymorphic call.
5482 * dumpfile.c (dump_files): Add type-inheritance dump.
5483 * dumpfile.h (TDI_inheritance): New.
5484 * ipa-devirt.c: New file.
5485 * ipa-utils.h (odr_type_d): Forward declare.
5486 (odr_type): New type.
5487 (build_type_inheritance_graph): Declare.
5488 (possible_polymorphic_call_targets): Declare and introduce inline
5489 variant when only edge is pased.
5490 (dump_possible_polymorphic_call_targets): Likewise.
5491 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
5492 * tree.c (type_in_anonymous_namespace_p): Break out from ...
5493 (types_same_for_odr): ... here.
5494 * tree.h (type_in_anonymous_namespace_p): Declare.
5496 2013-08-18 Jakub Jelinek <jakub@redhat.com>
5498 PR tree-optimization/58006
5499 * tree-parloops.c (take_address_of): Don't ICE if get_name
5501 (eliminate_local_variables_stmt): Remove clobber stmts.
5503 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
5505 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
5506 error message is issued for an alias to undefined symbol.
5508 2013-08-18 Jan Hubicka <jh@suse.cz>
5510 * cgraph.c (cgraph_create_indirect_edge): Discover
5511 polymorphic calls and record basic info into indirect_info.
5512 * gimple-fold.c (gimple_fold_call): When doing BINFO based
5513 devirtualization, ignore objc function calls.
5514 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
5515 call with no parm index info.
5516 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
5517 * tree.c (virtual_method_call_p): New function.
5518 * tree.h (virtual_method_call_p): Declare.
5520 2013-08-16 Jan Hubicka <jh@suse.cz>
5523 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
5525 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
5527 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
5530 2013-08-16 David Malcolm <dmalcolm@redhat.com>
5532 * gengtype.c (type_for_name): Add special-case support for
5533 locating types within the "gcc::" namespace.
5534 (open_base_files): Emit a "using namespace gcc" directive.
5536 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
5539 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
5540 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
5542 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
5543 array instead of each individual operand as a separate argument.
5544 (emit_fusion_gpr_load): Likewise.
5545 (expand_fusion_gpr_load): Add new function declaration.
5547 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
5548 signature to have the operands passed as an array, instead of as
5549 separate arguments. Allow ZERO_EXTEND to be in the memory
5550 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
5551 depend on the register live/dead flags when peepholes are run.
5552 (expand_fusion_gpr_load): New function to be called from the
5553 peephole2 pass, to change the register that addis sets to be the
5555 (emit_fusion_gpr_load): Change the calling signature to have the
5556 operands passed as an array, instead of as separate arguments.
5557 Allow ZERO_EXTEND to be in the memory address, and also
5558 SIGN_EXTEND if -mpower8-fusion-sign.
5560 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
5562 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
5563 adjust the register addis loads up in the peephole2 pass. Do not
5564 depend on the register live/dead state when the peephole pass is done.
5566 2013-08-16 David Malcolm <dmalcolm@redhat.com>
5568 * gengtype.c (create_user_defined_type): Ensure that the kind
5569 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
5570 declaration is seen before the GTY((user)) marking.
5572 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
5575 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
5577 2013-08-16 Jan Hubicka <jh@suse.cz>
5579 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
5580 arugment expected_type.
5581 (gimple_fold_call): Use it.
5582 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
5583 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
5584 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
5585 (try_make_edge_direct_virtual_call): Likewise.
5586 * tree.c (obj_type_ref_class): New.
5587 * tree.h (obj_type_ref_class): Use it.
5589 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
5591 * sched-vis.c (rtl_slim_pp_initialized): Remove.
5592 (rtl_slim_pp): Likewise.
5593 (init_rtl_slim_pretty_print): Likewise.
5594 (dump_value_slim): Don't call it. Use local pretty printer.
5595 (dump_insn_slim): Likewise.
5596 (dump_rtl_slim): Likewise.
5597 (str_pattern_slim): Likewise.
5598 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
5601 2013-08-16 Jakub Jelinek <jakub@redhat.com>
5603 PR tree-optimization/58164
5604 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
5605 walk gimple_goto_dest of GIMPLE_GOTO.
5607 PR tree-optimization/58165
5608 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
5609 bi_call must be the last stmt in a bb, don't split_block, instead
5610 use fallthru edge from it and give up if there is none.
5611 Release conds vector when returning early.
5613 2013-08-14 Xinliang David Li <davidxl@google.com>
5615 * config/i386/i386.c (ix86_option_override_internal):
5616 Remove unused variable and field.
5618 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5621 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
5622 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
5623 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
5624 and Linux, correct BLKmode alignment when 128-bit alignment is
5625 required and compatibility flag is not set.
5626 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
5627 for zero-size arguments when compatibility flag is not set.
5629 2013-08-14 Jakub Jelinek <jakub@redhat.com>
5631 PR tree-optimization/58145
5632 * tree-sra.c (build_ref_for_offset): If prev_base has
5633 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
5635 2013-08-14 Xinliang David Li <davidxl@google.com>
5637 * config/i386/i386.c (ix86_option_override_internal):
5638 Fix uninitialized variable error.
5640 2013-08-14 Xinliang David Li <davidxl@google.com>
5642 * config/i386/i386.opt: Define two new options.
5643 * config/i386/x86-tune.def: Add arch selector field in macros.
5644 * config/i386/i386.h: Adjust macro definition.
5645 * config/i386/i386.c (ix86_option_override_internal):
5647 (parse_mtune_ctrl_str): New function.
5648 (set_ix86_tune_features): New function.
5649 (ix86_function_specific_restore): Call the new helper function.
5651 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
5653 PR rtl-optimization/57662
5654 * sel-sched.c (code_motion_process_successors): When the current insn
5655 is removed after the recursive traversal, break from the loop.
5656 Add comments and debug printouts.
5658 2013-08-14 Jakub Jelinek <jakub@redhat.com>
5659 Alexandre Oliva <aoliva@redhat.com>
5662 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
5663 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
5664 there also UNSPEC_PLTOFF.
5666 2013-08-14 Marek Polacek <polacek@redhat.com>
5668 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
5669 NUM_CONDITIONS bit positions.
5671 2013-08-13 Cary Coutant <ccoutant@google.com>
5673 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
5674 (attr_checksum): Hash vector contents instead of pointer.
5675 (attr_checksum_ordered): Likewise.
5677 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
5679 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
5680 when Pmode != word_mode. Add length_address attribute.
5681 (sse3_monitor_<mode>): Merge from sse3_monitor and
5682 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
5683 Pmode != word_mode. Update insn length attribute.
5684 * config/i386/i386.c (ix86_option_override_internal): Update
5685 ix86_gen_monitor selection for merged sse3_monitor insn.
5687 2013-08-13 Julian Brown <julian@codesourcery.com>
5689 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
5690 perform invalid legitimization on greater-than-word-size modes for
5693 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
5695 * ira.c (setup_class_translate_array): Use aclass instead of cl
5696 for classes not fully covered by allocno classes.
5698 2013-08-13 Jakub Jelinek <jakub@redhat.com>
5700 PR tree-optimization/57661
5701 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
5702 * tree-inline.c (tree_function_versioning): Initialize it.
5703 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
5704 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
5705 that is not being copied.
5708 * asan.c (instrument_strlen_call): Fix typo in comment.
5709 Use char * type even for the lhs of POINTER_PLUS_EXPR.
5711 2013-08-13 Steve Ellcey <sellcey@mips.com>
5713 * config/mips/mips.md (prefetch): Use lw instead of ld on
5714 loongson in 32bit mode.
5716 2013-08-13 Nick Clifton <nickc@redhat.com>
5718 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
5720 2013-08-13 Jan Hubicka <jh@suse.cz>
5722 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
5723 introduced edge; fix typo in sanity check.
5724 (cgraph_resolve_speculation): Export; improve diagnostic.
5725 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
5726 speculation at type mismatch.
5727 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
5728 (cgraph_resolve_speculation): Declare.
5729 (symtab_can_be_discarded): New function.
5730 * value-prof.c (gimple_ic_transform): Remove actual transform code.
5731 * ipa-inline-transform.c (speculation_removed): New global var.
5732 (clone_inlined_nodes): See if speculation can be removed.
5733 (inline_call): If speculations was removed, we growths may not match.
5734 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
5735 (speculation_useful_p): New function.
5736 (resolve_noninline_speculation): New function.
5737 (inline_small_functions): Resolve useless speculations.
5738 * ipa-inline.h (speculation_useful_p): Declare
5739 * ipa.c (can_replace_by_local_alias): Simplify.
5740 (ipa_profile): Produce speculative calls in non-lto, too;
5741 add simple cost model; produce local aliases.
5743 2013-08-13 David Malcolm <dmalcolm@redhat.com>
5745 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
5748 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
5750 * config/i386/i386.c (ix86_function_versions): Use error + inform.
5752 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
5754 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
5755 iterator instead of X87MODEF.
5757 2013-08-12 Perez Read <netfirewall@gmail.com>
5760 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
5761 operand 0 for intel asm alternative.
5762 (*movabs<mode>_2): Ditto for operand 1.
5764 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
5766 * config/aarch64/arm_none.h
5767 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
5769 2013-08-12 Nick Clifton <nickc@redhat.com>
5771 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
5773 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
5775 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
5776 expand for QI/HImode operand to produce more effictive code for
5777 unsigned char(short) --> float(double) conversion.
5779 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
5781 * doc/invoke.texi: Mention that -ftls-model does not force the final
5784 2013-08-12 Marek Polacek <polacek@redhat.com>
5785 Marc Glisse <marc.glisse@inria.fr>
5787 PR tree-optimization/57980
5788 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
5789 when creating -1 constant.
5791 2013-08-10 Jan Hubicka <jh@suse.cz>
5793 Workaround binutils PR14342.
5794 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
5795 (gimple_init_edge_profiler): Likewise.
5796 (gimple_gen_ic_func_profiler): Likewise.
5798 2013-08-09 Jan Hubicka <jh@suse.cz>
5800 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
5802 2013-08-09 Xinliang David Li <davidxl@google.com>
5804 * config/i386/stringop.def: New file.
5805 * config/i386/stringop.opt: New file.
5806 * config/i386/i386-opts.h: Include stringopt.def.
5807 * config/i386/i386.opt: Include stringopt.opt.
5808 * config/i386/i386.c (ix86_option_override_internal):
5809 Override default size based stringop inline strategies with options.
5810 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
5813 2013-08-09 Jan Hubicka <jh@suse.cz>
5815 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
5817 2013-08-09 Jan Hubicka <jh@suse.cz>
5819 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
5821 (dump_cgraph_node): Dump profile-id.
5822 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
5823 and common_target_probability.
5824 * lto-cgraph.c (lto_output_edge): Stream common targets.
5825 (lto_output_node): Stream profile ids.
5826 (input_node): Stream profile ids.
5827 (input_edge): Stream common targets.
5828 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
5829 * ipa.c: Include value-prof.h
5830 (ipa_profile_generate_summary): Turn indirect call statement histograms
5831 into common targets.
5832 (ipa_profile): Turn common targets into speculative edges.
5834 2013-08-09 Jan Hubicka <jh@suse.cz>
5836 * cgraph.h (cgraph_node): Add profile_id.
5837 * value-prof.c (cgraph_node_map): Turn into pointer_map.
5838 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
5839 (del_node_map): Update.
5840 (find_func_by_funcdef_no): Replace by ...
5841 (find_func_by_profile_id): ... this one.
5842 (gimple_ic_transform): Do not remove useful histograms when
5843 speculation is not done; dump info when indirect call removal
5845 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
5846 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
5847 (__gcov_indirect_call_profiler_v2): .. this one.
5848 * profile.h (init_node_map): Update.
5849 * coverage.c (coverage_compute_profile_id): New function.
5850 * coverage.h (coverage_compute_profile_id): Declare.
5851 * tree-profile.c (init_ic_make_global_vars): Make
5852 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
5853 (gimple_init_edge_profiler): Update prototype of
5854 __gcov_indirect_call_profiler.
5855 (gimple_gen_ic_func_profiler): Simplify.
5856 (tree_profiling): Use init_node_map
5858 2013-08-09 Jan Hubicka <jh@suse.cz>
5860 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
5861 non-speculative refs.
5862 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
5863 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
5864 (cgraph_set_call_stmt): Likewise.
5865 (cgraph_create_edge_1): Fix release checking compilatoin;
5867 (cgraph_free_edge): Free indirect info.
5868 (cgraph_turn_edge_to_speculative): New function.
5869 (cgraph_speculative_call_info): New function.
5870 (cgraph_make_edge_direct): Return direct edge; handle speculation.
5871 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
5872 (dump_cgraph_node): Dump speculation.
5873 (verify_edge_count_and_frequency): Accept speculative edges.
5874 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
5875 (verify_cgraph_node): Handle speculation.
5876 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
5877 (cgraph_set_call_stmt): Update prototype.
5878 (cgraph_make_edge_direct): Update prototype.
5879 (cgraph_speculative_call_info): Declare.
5880 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
5881 to change; update call of ipa_find_references.
5882 * ipa-ref.c (ipa_record_reference): Fix return value; clear
5883 lto_stmt_uid and speculative flags.
5884 (ipa_dump_references): Dump speculation.
5885 (ipa_clone_references): Clone speculative flag.
5886 (ipa_clone_referring): Likewise.
5887 (ipa_clone_ref): New function.
5888 (ipa_find_reference): Look into lto_stmt_uids
5889 (ipa_clear_stmts_in_references): Do not clear speculative calls.
5890 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
5891 (ipa_find_reference): Update declaration.
5892 (ipa_clone_ref): Declare.
5893 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
5894 stream speculative flag.
5895 (lto_output_ref): Stream statements uids and speculation.
5896 (input_ref): Likewise.
5897 (input_edge): Stream speuclation.
5898 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
5899 (cgraph_set_call_stmt_including_clones): Handle speculation.
5900 * ipa-inline.c (heap_edge_removal_hook): New function.
5901 (inline_small_functions): Register it.
5902 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
5903 also initialize refs.
5904 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
5906 (try_make_edge_direct_simple_call): Likewise.
5907 (try_make_edge_direct_simple_call): Likewise.
5908 (update_indirect_edges_after_inlining): Likewise.
5909 (remove_described_reference): Look proper lto_stmt_uid.
5910 (propagate_controlled_uses): Likewise.
5911 (propagate_controlled_uses): Liekwise.
5912 * tree-inline.c (copy_bb): Copy speculative edges.
5913 (redirect_all_calls): New function.
5914 (copy_cfg_body): Do redirection after loop info is updated.
5915 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
5917 2013-08-09 Jan Hubicka <jh@suse.cz>
5919 * lto-streamer-out.c (output_function): Renumber PHIs.
5920 * lto-streamer-in.c (input_function): Likewise.
5922 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
5924 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
5925 (get_lane_unsigned): Likewise.
5926 (dup_lane_scalar): Likewise.
5927 (get_lane): enable for VALL.
5928 * config/aarch64/aarch64-simd.md
5929 (aarch64_dup_lane_scalar<mode>): Remove.
5930 (aarch64_get_lane_signed<mode>): Likewise.
5931 (aarch64_get_lane_unsigned<mode>): Likewise.
5932 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
5933 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
5934 (aarch64_get_lane<mode>): Enable for all vector modes.
5935 (aarch64_get_lanedi): Remove misleading constraints.
5936 * config/aarch64/arm_neon.h
5937 (__aarch64_vget_lane_any): Define.
5938 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
5939 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
5940 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
5941 * config/aarch64/iterators.md (VDQQH): New.
5945 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
5947 * configure.ac: Add GAS check for LEON instructions on SPARC.
5948 * configure: Regenerate.
5949 * config.in: Likewise.
5950 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
5952 * config/sparc/sparc.opt (LEON, LEON3): New masks.
5953 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
5955 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
5956 (AS_LEON_FLAG): New macro.
5957 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
5958 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
5959 Deal with LEON and LEON3 for the memory model.
5960 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
5961 (atomic_compare_and_swap<mode>_1): Likewise.
5962 (*atomic_compare_and_swap<mode>_1): Likewise.
5964 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
5966 * config/arm/neon.md (vcond): Fix floating-point vector
5967 comparisons against 0.
5969 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
5971 * lra-constraints.c (emit_spill_move): Remove assert.
5972 (process_alt_operands): Add more debugging
5973 output. Increase reject for spilling into memory. Decrease
5974 reject for reloading scratch.
5975 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
5977 2013-08-08 Steve Ellcey <sellcey@mips.com>
5979 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
5980 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
5981 micromips incompatible. Add nan2008.
5982 (MULTILIB_DIRNAMES): Add nan2008.
5983 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
5984 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
5985 and micromips incompatible. Add nan2008.
5986 (MULTILIB_DIRNAMES): Add nan2008.
5987 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
5989 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
5991 PR rtl-optimization/58079
5992 * combine.c (combine_simplify_rtx): Avoid using SUBST if
5993 simplify_comparison has widened a comparison with an integer.
5995 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5997 * config/arm/neon.md (movmisalign<mode>): Disable when we
5998 don't allow unaligned accesses.
5999 (*movmisalign<mode>_neon_store): Likewise.
6000 (*movmisalign<mode>_neon_load): Likewise.
6001 (*movmisalign<mode>_neon_store): Likewise.
6002 (*movmisalign<mode>_neon_load): Likewise.
6004 2013-08-08 Jan Hubicka <jh@suse.cz>
6006 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
6007 (make_pass_rebuild_cgraph_edges): Also clear references.
6008 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
6009 * ipa-inline-transform.c (inline_transform): Remove all references
6011 * cgraphunit.c (expand_function): Remove all references after
6013 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
6014 (ipa_find_reference): Rewrite to iterator.
6015 (remove_stmt_references): Likewise.
6016 (ipa_clear_stmts_in_references): New function.
6017 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
6018 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
6020 * ipa-split.c (split_function): Remove references in split function.
6022 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
6025 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
6026 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
6028 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
6031 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
6032 suggested mode for the assignment isn't compatible with the
6035 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
6038 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
6040 2013-08-07 Xinliang David Li <davidxl@google.com>
6042 * config/i386/i386.opt: New option -mtune-ctrl=.
6043 * config/i386/x86-tune.def: New file.
6044 * config/i386/i386.h: include x86-tune.def.
6045 * config/i386/i386.c (ix86_option_override_internal):
6046 Parsing -mtune-ctrl= option and set tune features.
6048 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
6051 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
6052 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
6055 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
6057 * rtl.h (update_alignments): Declare.
6058 * final.c (grow_label_align): New function extracted from...
6059 (shorten_branches): ...here. Call it.
6060 (update_alignments): New function.
6061 * reorg.c (sibling_labels): New variable.
6062 (get_label_before): Add SIBLING parameter. If it is non-zero, push
6063 the new label along with it onto the sibling_labels vector.
6064 (fill_simple_delay_slots): Adjust call to get_label_before.
6065 (fill_slots_from_thread): Likewise.
6066 (relax_delay_slots): Likewise.
6067 (make_return_insns): Likewise.
6068 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
6070 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
6072 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
6073 document its semantics.
6074 (diagnostic_report_diagnostic): Adjust accordingly.
6076 2013-08-07 David Malcolm <dmalcolm@redhat.com>
6078 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
6079 (sparc_option_override): ...and port to new C++ pass API.
6080 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
6082 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
6084 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
6086 2013-08-06 Caroline Tice <cmtice@google.com>
6088 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
6089 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
6090 * tree-pass.h: Add pass_vtable_verify.
6091 * varasm.c (assemble_variable): Add code to properly set the comdat
6092 section and name for the .vtable_map_vars section.
6093 (assemble_vtyv_preinit_initializer): New function.
6094 (default_sectin_type_flags): Make sure .vtable_map_vars section has
6096 * output.h: Add function decl for assemble_vtv_preinit_initializer.
6097 * vtable-verify.c: New file.
6098 * vtable-verify.h: New file.
6099 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
6100 initialiation levels.
6101 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
6102 * passes.def: Insert pass_vtable_verify.
6103 * aclocal.m4: Reorder includes.
6104 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
6105 -fvtv-counts options.
6106 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
6107 as appropriate, if -fvtable-verify=... is used.
6108 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
6109 -fvtable-verify=... is used.
6110 * Makefile.in (OBJS): Add vtable-verify.o to list.
6111 (vtable-verify.o): Add new build rule.
6112 (GTFILES): Add vtable-verify.c to list.
6113 * common.opt (fvtable-verify=): New flag.
6114 (vtv_priority): Values for fvtable-verify= flag.
6115 (fvtv-counts): New flag.
6116 (fvtv-debug): New flag.
6117 * tree.h (save_vtable_map_decl): New extern function decl.
6119 2013-08-07 David Malcolm <dmalcolm@redhat.com>
6121 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
6122 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
6123 (pass_data_rl78_devirt): ...new pass_data instance and...
6124 (make_pass_rl78_devirt): ...new function.
6125 (rl78_asm_file_start): Port pass registration to new C++ API.
6127 2013-08-07 David Malcolm <dmalcolm@redhat.com>
6129 * coretypes.h (rtl_opt_pass): Add.
6130 (gcc::context): Add.
6131 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
6132 (epiphany_init): Port to new C++ pass API.
6133 (epiphany_optimize_mode_switching): Likewise.
6134 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
6135 (pass_manager::get_pass_mode_switching): New.
6136 (pass_manager::get_pass_peephole2): New.
6137 * mode-switching.c (pass_mode_switching): Add clone method.
6138 * recog.c (pass_peephole2): Add clone method.
6139 (pass_split_all_insns): Add clone method.
6141 2013-08-06 David Malcolm <dmalcolm@redhat.com>
6143 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
6144 (mips_option_override): ...here, porting to new C++ API for passes.
6146 2013-08-06 Jan Hubicka <jh@suse.cz>
6148 * cgraph.c (cgraph_get_body): New function based on lto.c
6150 * cgraph.h (cgraph_get_body): Declare.
6151 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
6153 * cgraphunit.c (expand_function): Get body prior expanding.
6154 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
6155 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
6157 * passes.c (do_per_function_toporder): Get body.
6158 * tree-inline.c (expand_call_inline): Get body prior inlining it.
6159 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
6161 2013-08-06 Martin Jambor <mjambor@suse.cz>
6164 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
6165 is not re-finalized. Rename second parameter to no_collect.
6167 2013-08-06 Martin Jambor <mjambor@suse.cz>
6170 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
6171 MEM_REF has proper alignment information.
6173 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
6176 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
6178 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
6179 rtx (*) (rtx, ...) with insn_gen_fn.
6180 * genoutput.c (output_insn_data): Cast gen_? function pointers to
6181 insn_gen_fn::stored_funcptr. Add initializer braces.
6183 2013-08-05 David Malcolm <dmalcolm@redhat.com>
6185 Rewrite how instances of passes are cloned to remove assumptions
6186 about their sizes (thus allowing pass subclasses to have
6187 additional data fields, albeit non-GC-managed ones at this point).
6189 * passes.c (make_pass_instance): Now that passes have clone
6190 methods, rewrite this function to eliminate XNEW and memcpy
6191 calls that used hardcoded sizes. Since this function no longer
6192 creates pass instances, rename it to...
6193 (add_pass_instance): ...this. Document the old way that passes were
6194 numbered and flagged, and rework this function to continue using it.
6195 (next_pass_1): Add an initial_pass argument for use by
6197 (position_pass): When adding multiple instances of a pass, use
6198 the pass's clone method, rather than relying on the XNEW/memcpy
6199 within the former make_pass_instance (now add_pass_instance).
6200 (pass_manager::pass_manager): When invoking next_pass_1, also supply
6201 the initial instance of the current pass within the pass manager.
6203 2013-08-05 David Malcolm <dmalcolm@redhat.com>
6205 This is the automated part of the conversion of passes from C
6206 structs to C++ classes.
6208 Patch autogenerated by refactor_passes.py from
6209 https://github.com/davidmalcolm/gcc-refactoring-scripts
6210 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
6212 * asan.c (pass_asan): Convert from a global struct to a subclass of
6213 gimple_opt_pass along with...
6214 (pass_data_asan): ...new pass_data instance and...
6215 (make_pass_asan): ...new function.
6216 (pass_asan_O0): Convert from a global struct to a subclass of
6217 gimple_opt_pass along with...
6218 (pass_data_asan_O0): ...new pass_data instance and...
6219 (make_pass_asan_O0): ...new function.
6220 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
6221 subclass of rtl_opt_pass along with...
6222 (pass_data_inc_dec): ...new pass_data instance and...
6223 (make_pass_inc_dec): ...new function.
6224 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
6225 a subclass of rtl_opt_pass along with...
6226 (pass_data_reorder_blocks): ...new pass_data instance and...
6227 (make_pass_reorder_blocks): ...new function.
6228 (pass_duplicate_computed_gotos): Convert from a global struct to a
6229 subclass of rtl_opt_pass along with...
6230 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
6231 (make_pass_duplicate_computed_gotos): ...new function.
6232 (pass_partition_blocks): Convert from a global struct to a subclass of
6233 rtl_opt_pass along with...
6234 (pass_data_partition_blocks): ...new pass_data instance and...
6235 (make_pass_partition_blocks): ...new function.
6236 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
6237 struct to a subclass of rtl_opt_pass along with...
6238 (pass_data_branch_target_load_optimize1): ...new pass_data instance
6240 (make_pass_branch_target_load_optimize1): ...new function.
6241 (pass_branch_target_load_optimize2): Convert from a global struct to a
6242 subclass of rtl_opt_pass along with...
6243 (pass_data_branch_target_load_optimize2): ...new pass_data instance
6245 (make_pass_branch_target_load_optimize2): ...new function.
6246 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
6247 of rtl_opt_pass along with...
6248 (pass_data_jump): ...new pass_data instance and...
6249 (make_pass_jump): ...new function.
6250 (pass_jump2): Convert from a global struct to a subclass of
6251 rtl_opt_pass along with...
6252 (pass_data_jump2): ...new pass_data instance and...
6253 (make_pass_jump2): ...new function.
6254 * cfgexpand.c (pass_expand): Convert from a global struct to a
6255 subclass of rtl_opt_pass along with...
6256 (pass_data_expand): ...new pass_data instance and...
6257 (make_pass_expand): ...new function.
6258 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
6259 of rtl_opt_pass along with...
6260 (pass_data_free_cfg): ...new pass_data instance and...
6261 (make_pass_free_cfg): ...new function.
6262 (pass_into_cfg_layout_mode): Convert from a global struct to a
6263 subclass of rtl_opt_pass along with...
6264 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
6265 (make_pass_into_cfg_layout_mode): ...new function.
6266 (pass_outof_cfg_layout_mode): Convert from a global struct to a
6267 subclass of rtl_opt_pass along with...
6268 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
6269 (make_pass_outof_cfg_layout_mode): ...new function.
6270 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
6271 struct to a subclass of gimple_opt_pass along with...
6272 (pass_data_build_cgraph_edges): ...new pass_data instance and...
6273 (make_pass_build_cgraph_edges): ...new function.
6274 (pass_rebuild_cgraph_edges): Convert from a global struct to a
6275 subclass of gimple_opt_pass along with...
6276 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
6277 (make_pass_rebuild_cgraph_edges): ...new function.
6278 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
6279 subclass of gimple_opt_pass along with...
6280 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
6282 (make_pass_remove_cgraph_callee_edges): ...new function.
6283 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
6284 struct to a subclass of rtl_opt_pass along with...
6285 (pass_data_stack_adjustments): ...new pass_data instance and...
6286 (make_pass_stack_adjustments): ...new function.
6287 * combine.c (pass_combine): Convert from a global struct to a subclass
6288 of rtl_opt_pass along with...
6289 (pass_data_combine): ...new pass_data instance and...
6290 (make_pass_combine): ...new function.
6291 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
6292 global struct to a subclass of rtl_opt_pass along with...
6293 (pass_data_compare_elim_after_reload): ...new pass_data instance
6295 (make_pass_compare_elim_after_reload): ...new function.
6296 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
6297 of rtl_opt_pass along with...
6298 (pass_data_rtl_cprop): ...new pass_data instance and...
6299 (make_pass_rtl_cprop): ...new function.
6300 * cse.c (pass_cse): Convert from a global struct to a subclass of
6301 rtl_opt_pass along with...
6302 (pass_data_cse): ...new pass_data instance and...
6303 (make_pass_cse): ...new function.
6304 (pass_cse2): Convert from a global struct to a subclass of
6305 rtl_opt_pass along with...
6306 (pass_data_cse2): ...new pass_data instance and...
6307 (make_pass_cse2): ...new function.
6308 (pass_cse_after_global_opts): Convert from a global struct to a
6309 subclass of rtl_opt_pass along with...
6310 (pass_data_cse_after_global_opts): ...new pass_data instance and...
6311 (make_pass_cse_after_global_opts): ...new function.
6312 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
6313 of rtl_opt_pass along with...
6314 (pass_data_ud_rtl_dce): ...new pass_data instance and...
6315 (make_pass_ud_rtl_dce): ...new function.
6316 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
6317 rtl_opt_pass along with...
6318 (pass_data_fast_rtl_dce): ...new pass_data instance and...
6319 (make_pass_fast_rtl_dce): ...new function.
6320 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
6321 a subclass of rtl_opt_pass along with...
6322 (pass_data_df_initialize_opt): ...new pass_data instance and...
6323 (make_pass_df_initialize_opt): ...new function.
6324 (pass_df_initialize_no_opt): Convert from a global struct to a
6325 subclass of rtl_opt_pass along with...
6326 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
6327 (make_pass_df_initialize_no_opt): ...new function.
6328 (pass_df_finish): Convert from a global struct to a subclass of
6329 rtl_opt_pass along with...
6330 (pass_data_df_finish): ...new pass_data instance and...
6331 (make_pass_df_finish): ...new function.
6332 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
6333 rtl_opt_pass along with...
6334 (pass_data_rtl_dse1): ...new pass_data instance and...
6335 (make_pass_rtl_dse1): ...new function.
6336 (pass_rtl_dse2): Convert from a global struct to a subclass of
6337 rtl_opt_pass along with...
6338 (pass_data_rtl_dse2): ...new pass_data instance and...
6339 (make_pass_rtl_dse2): ...new function.
6340 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
6341 subclass of rtl_opt_pass along with...
6342 (pass_data_dwarf2_frame): ...new pass_data instance and...
6343 (make_pass_dwarf2_frame): ...new function.
6344 * except.c (pass_set_nothrow_function_flags): Convert from a global
6345 struct to a subclass of rtl_opt_pass along with...
6346 (pass_data_set_nothrow_function_flags): ...new pass_data instance
6348 (make_pass_set_nothrow_function_flags): ...new function.
6349 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
6350 subclass of rtl_opt_pass along with...
6351 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
6353 (make_pass_convert_to_eh_region_ranges): ...new function.
6354 * final.c (pass_compute_alignments): Convert from a global struct to a
6355 subclass of rtl_opt_pass along with...
6356 (pass_data_compute_alignments): ...new pass_data instance and...
6357 (make_pass_compute_alignments): ...new function.
6358 (pass_final): Convert from a global struct to a subclass of
6359 rtl_opt_pass along with...
6360 (pass_data_final): ...new pass_data instance and...
6361 (make_pass_final): ...new function.
6362 (pass_shorten_branches): Convert from a global struct to a subclass of
6363 rtl_opt_pass along with...
6364 (pass_data_shorten_branches): ...new pass_data instance and...
6365 (make_pass_shorten_branches): ...new function.
6366 (pass_clean_state): Convert from a global struct to a subclass of
6367 rtl_opt_pass along with...
6368 (pass_data_clean_state): ...new pass_data instance and...
6369 (make_pass_clean_state): ...new function.
6370 * function.c (pass_instantiate_virtual_regs): Convert from a global
6371 struct to a subclass of rtl_opt_pass along with...
6372 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
6373 (make_pass_instantiate_virtual_regs): ...new function.
6374 (pass_leaf_regs): Convert from a global struct to a subclass of
6375 rtl_opt_pass along with...
6376 (pass_data_leaf_regs): ...new pass_data instance and...
6377 (make_pass_leaf_regs): ...new function.
6378 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
6379 subclass of rtl_opt_pass along with...
6380 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
6382 (make_pass_thread_prologue_and_epilogue): ...new function.
6383 (pass_match_asm_constraints): Convert from a global struct to a
6384 subclass of rtl_opt_pass along with...
6385 (pass_data_match_asm_constraints): ...new pass_data instance and...
6386 (make_pass_match_asm_constraints): ...new function.
6387 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
6388 subclass of rtl_opt_pass along with...
6389 (pass_data_rtl_fwprop): ...new pass_data instance and...
6390 (make_pass_rtl_fwprop): ...new function.
6391 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
6392 rtl_opt_pass along with...
6393 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
6394 (make_pass_rtl_fwprop_addr): ...new function.
6395 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
6396 rtl_opt_pass along with...
6397 (pass_data_rtl_pre): ...new pass_data instance and...
6398 (make_pass_rtl_pre): ...new function.
6399 (pass_rtl_hoist): Convert from a global struct to a subclass of
6400 rtl_opt_pass along with...
6401 (pass_data_rtl_hoist): ...new pass_data instance and...
6402 (make_pass_rtl_hoist): ...new function.
6403 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
6404 subclass of gimple_opt_pass along with...
6405 (pass_data_lower_cf): ...new pass_data instance and...
6406 (make_pass_lower_cf): ...new function.
6407 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
6408 from a global struct to a subclass of gimple_opt_pass along with...
6409 (pass_data_strength_reduction): ...new pass_data instance and...
6410 (make_pass_strength_reduction): ...new function.
6411 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
6412 of rtl_opt_pass along with...
6413 (pass_data_rtl_ifcvt): ...new pass_data instance and...
6414 (make_pass_rtl_ifcvt): ...new function.
6415 (pass_if_after_combine): Convert from a global struct to a subclass of
6416 rtl_opt_pass along with...
6417 (pass_data_if_after_combine): ...new pass_data instance and...
6418 (make_pass_if_after_combine): ...new function.
6419 (pass_if_after_reload): Convert from a global struct to a subclass of
6420 rtl_opt_pass along with...
6421 (pass_data_if_after_reload): ...new pass_data instance and...
6422 (make_pass_if_after_reload): ...new function.
6423 * init-regs.c (pass_initialize_regs): Convert from a global struct to
6424 a subclass of rtl_opt_pass along with...
6425 (pass_data_initialize_regs): ...new pass_data instance and...
6426 (make_pass_initialize_regs): ...new function.
6427 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
6428 of ipa_opt_pass_d along with...
6429 (pass_data_ipa_cp): ...new pass_data instance and...
6430 (make_pass_ipa_cp): ...new function.
6431 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
6432 global struct to a subclass of gimple_opt_pass along with...
6433 (pass_data_inline_parameters): ...new pass_data instance and...
6434 (make_pass_inline_parameters): ...new function.
6435 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
6436 subclass of gimple_opt_pass along with...
6437 (pass_data_early_inline): ...new pass_data instance and...
6438 (make_pass_early_inline): ...new function.
6439 (pass_ipa_inline): Convert from a global struct to a subclass of
6440 ipa_opt_pass_d along with...
6441 (pass_data_ipa_inline): ...new pass_data instance and...
6442 (make_pass_ipa_inline): ...new function.
6443 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
6444 struct to a subclass of gimple_opt_pass along with...
6445 (pass_data_local_pure_const): ...new pass_data instance and...
6446 (make_pass_local_pure_const): ...new function.
6447 (pass_ipa_pure_const): Convert from a global struct to a subclass of
6448 ipa_opt_pass_d along with...
6449 (pass_data_ipa_pure_const): ...new pass_data instance and...
6450 (make_pass_ipa_pure_const): ...new function.
6451 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
6452 to a subclass of ipa_opt_pass_d along with...
6453 (pass_data_ipa_reference): ...new pass_data instance and...
6454 (make_pass_ipa_reference): ...new function.
6455 * ipa-split.c (pass_split_functions): Convert from a global struct to
6456 a subclass of gimple_opt_pass along with...
6457 (pass_data_split_functions): ...new pass_data instance and...
6458 (make_pass_split_functions): ...new function.
6459 (pass_feedback_split_functions): Convert from a global struct to a
6460 subclass of gimple_opt_pass along with...
6461 (pass_data_feedback_split_functions): ...new pass_data instance and...
6462 (make_pass_feedback_split_functions): ...new function.
6463 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
6464 global struct to a subclass of simple_ipa_opt_pass along with...
6465 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
6467 (make_pass_ipa_function_and_variable_visibility): ...new function.
6468 (pass_ipa_free_inline_summary): Convert from a global struct to a
6469 subclass of simple_ipa_opt_pass along with...
6470 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
6471 (make_pass_ipa_free_inline_summary): ...new function.
6472 (pass_ipa_whole_program_visibility): Convert from a global struct to a
6473 subclass of ipa_opt_pass_d along with...
6474 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
6476 (make_pass_ipa_whole_program_visibility): ...new function.
6477 (pass_ipa_profile): Convert from a global struct to a subclass of
6478 ipa_opt_pass_d along with...
6479 (pass_data_ipa_profile): ...new pass_data instance and...
6480 (make_pass_ipa_profile): ...new function.
6481 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
6482 ipa_opt_pass_d along with...
6483 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
6484 (make_pass_ipa_cdtor_merge): ...new function.
6485 * ira.c (pass_ira): Convert from a global struct to a subclass of
6486 rtl_opt_pass along with...
6487 (pass_data_ira): ...new pass_data instance and...
6488 (make_pass_ira): ...new function.
6489 (pass_reload): Convert from a global struct to a subclass of
6490 rtl_opt_pass along with...
6491 (pass_data_reload): ...new pass_data instance and...
6492 (make_pass_reload): ...new function.
6493 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
6494 subclass of rtl_opt_pass along with...
6495 (pass_data_cleanup_barriers): ...new pass_data instance and...
6496 (make_pass_cleanup_barriers): ...new function.
6497 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
6498 of rtl_opt_pass along with...
6499 (pass_data_loop2): ...new pass_data instance and...
6500 (make_pass_loop2): ...new function.
6501 (pass_rtl_loop_init): Convert from a global struct to a subclass of
6502 rtl_opt_pass along with...
6503 (pass_data_rtl_loop_init): ...new pass_data instance and...
6504 (make_pass_rtl_loop_init): ...new function.
6505 (pass_rtl_loop_done): Convert from a global struct to a subclass of
6506 rtl_opt_pass along with...
6507 (pass_data_rtl_loop_done): ...new pass_data instance and...
6508 (make_pass_rtl_loop_done): ...new function.
6509 (pass_rtl_move_loop_invariants): Convert from a global struct to a
6510 subclass of rtl_opt_pass along with...
6511 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
6512 (make_pass_rtl_move_loop_invariants): ...new function.
6513 (pass_rtl_unswitch): Convert from a global struct to a subclass of
6514 rtl_opt_pass along with...
6515 (pass_data_rtl_unswitch): ...new pass_data instance and...
6516 (make_pass_rtl_unswitch): ...new function.
6517 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
6518 subclass of rtl_opt_pass along with...
6519 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
6521 (make_pass_rtl_unroll_and_peel_loops): ...new function.
6522 (pass_rtl_doloop): Convert from a global struct to a subclass of
6523 rtl_opt_pass along with...
6524 (pass_data_rtl_doloop): ...new pass_data instance and...
6525 (make_pass_rtl_doloop): ...new function.
6526 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
6527 a subclass of rtl_opt_pass along with...
6528 (pass_data_lower_subreg): ...new pass_data instance and...
6529 (make_pass_lower_subreg): ...new function.
6530 (pass_lower_subreg2): Convert from a global struct to a subclass of
6531 rtl_opt_pass along with...
6532 (pass_data_lower_subreg2): ...new pass_data instance and...
6533 (make_pass_lower_subreg2): ...new function.
6534 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
6535 struct to a subclass of ipa_opt_pass_d along with...
6536 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
6537 (make_pass_ipa_lto_gimple_out): ...new function.
6538 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
6539 of ipa_opt_pass_d along with...
6540 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
6541 (make_pass_ipa_lto_finish_out): ...new function.
6542 * mode-switching.c (pass_mode_switching): Convert from a global struct
6543 to a subclass of rtl_opt_pass along with...
6544 (pass_data_mode_switching): ...new pass_data instance and...
6545 (make_pass_mode_switching): ...new function.
6546 * modulo-sched.c (pass_sms): Convert from a global struct to a
6547 subclass of rtl_opt_pass along with...
6548 (pass_data_sms): ...new pass_data instance and...
6549 (make_pass_sms): ...new function.
6550 * omp-low.c (pass_expand_omp): Convert from a global struct to a
6551 subclass of gimple_opt_pass along with...
6552 (pass_data_expand_omp): ...new pass_data instance and...
6553 (make_pass_expand_omp): ...new function.
6554 (pass_lower_omp): Convert from a global struct to a subclass of
6555 gimple_opt_pass along with...
6556 (pass_data_lower_omp): ...new pass_data instance and...
6557 (make_pass_lower_omp): ...new function.
6558 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
6559 of gimple_opt_pass along with...
6560 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
6561 (make_pass_diagnose_omp_blocks): ...new function.
6562 * passes.c (pass_early_local_passes): Convert from a global struct to
6563 a subclass of simple_ipa_opt_pass along with...
6564 (pass_data_early_local_passes): ...new pass_data instance and...
6565 (make_pass_early_local_passes): ...new function.
6566 (pass_all_early_optimizations): Convert from a global struct to a
6567 subclass of gimple_opt_pass along with...
6568 (pass_data_all_early_optimizations): ...new pass_data instance and...
6569 (make_pass_all_early_optimizations): ...new function.
6570 (pass_all_optimizations): Convert from a global struct to a subclass
6571 of gimple_opt_pass along with...
6572 (pass_data_all_optimizations): ...new pass_data instance and...
6573 (make_pass_all_optimizations): ...new function.
6574 (pass_all_optimizations_g): Convert from a global struct to a subclass
6575 of gimple_opt_pass along with...
6576 (pass_data_all_optimizations_g): ...new pass_data instance and...
6577 (make_pass_all_optimizations_g): ...new function.
6578 (pass_rest_of_compilation): Convert from a global struct to a subclass
6579 of rtl_opt_pass along with...
6580 (pass_data_rest_of_compilation): ...new pass_data instance and...
6581 (make_pass_rest_of_compilation): ...new function.
6582 (pass_postreload): Convert from a global struct to a subclass of
6583 rtl_opt_pass along with...
6584 (pass_data_postreload): ...new pass_data instance and...
6585 (make_pass_postreload): ...new function.
6586 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
6587 subclass of rtl_opt_pass along with...
6588 (pass_data_gcse2): ...new pass_data instance and...
6589 (make_pass_gcse2): ...new function.
6590 * postreload.c (pass_postreload_cse): Convert from a global struct to
6591 a subclass of rtl_opt_pass along with...
6592 (pass_data_postreload_cse): ...new pass_data instance and...
6593 (make_pass_postreload_cse): ...new function.
6594 * predict.c (pass_profile): Convert from a global struct to a subclass
6595 of gimple_opt_pass along with...
6596 (pass_data_profile): ...new pass_data instance and...
6597 (make_pass_profile): ...new function.
6598 (pass_strip_predict_hints): Convert from a global struct to a subclass
6599 of gimple_opt_pass along with...
6600 (pass_data_strip_predict_hints): ...new pass_data instance and...
6601 (make_pass_strip_predict_hints): ...new function.
6602 * recog.c (pass_peephole2): Convert from a global struct to a subclass
6603 of rtl_opt_pass along with...
6604 (pass_data_peephole2): ...new pass_data instance and...
6605 (make_pass_peephole2): ...new function.
6606 (pass_split_all_insns): Convert from a global struct to a subclass of
6607 rtl_opt_pass along with...
6608 (pass_data_split_all_insns): ...new pass_data instance and...
6609 (make_pass_split_all_insns): ...new function.
6610 (pass_split_after_reload): Convert from a global struct to a subclass
6611 of rtl_opt_pass along with...
6612 (pass_data_split_after_reload): ...new pass_data instance and...
6613 (make_pass_split_after_reload): ...new function.
6614 (pass_split_before_regstack): Convert from a global struct to a
6615 subclass of rtl_opt_pass along with...
6616 (pass_data_split_before_regstack): ...new pass_data instance and...
6617 (make_pass_split_before_regstack): ...new function.
6618 (pass_split_before_sched2): Convert from a global struct to a subclass
6619 of rtl_opt_pass along with...
6620 (pass_data_split_before_sched2): ...new pass_data instance and...
6621 (make_pass_split_before_sched2): ...new function.
6622 (pass_split_for_shorten_branches): Convert from a global struct to a
6623 subclass of rtl_opt_pass along with...
6624 (pass_data_split_for_shorten_branches): ...new pass_data instance
6626 (make_pass_split_for_shorten_branches): ...new function.
6627 * ree.c (pass_ree): Convert from a global struct to a subclass of
6628 rtl_opt_pass along with...
6629 (pass_data_ree): ...new pass_data instance and...
6630 (make_pass_ree): ...new function.
6631 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
6632 subclass of rtl_opt_pass along with...
6633 (pass_data_stack_regs): ...new pass_data instance and...
6634 (make_pass_stack_regs): ...new function.
6635 (pass_stack_regs_run): Convert from a global struct to a subclass of
6636 rtl_opt_pass along with...
6637 (pass_data_stack_regs_run): ...new pass_data instance and...
6638 (make_pass_stack_regs_run): ...new function.
6639 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
6640 subclass of rtl_opt_pass along with...
6641 (pass_data_cprop_hardreg): ...new pass_data instance and...
6642 (make_pass_cprop_hardreg): ...new function.
6643 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
6644 subclass of rtl_opt_pass along with...
6645 (pass_data_reginfo_init): ...new pass_data instance and...
6646 (make_pass_reginfo_init): ...new function.
6647 * regmove.c (pass_regmove): Convert from a global struct to a subclass
6648 of rtl_opt_pass along with...
6649 (pass_data_regmove): ...new pass_data instance and...
6650 (make_pass_regmove): ...new function.
6651 * regrename.c (pass_regrename): Convert from a global struct to a
6652 subclass of rtl_opt_pass along with...
6653 (pass_data_regrename): ...new pass_data instance and...
6654 (make_pass_regrename): ...new function.
6655 * reorg.c (pass_delay_slots): Convert from a global struct to a
6656 subclass of rtl_opt_pass along with...
6657 (pass_data_delay_slots): ...new pass_data instance and...
6658 (make_pass_delay_slots): ...new function.
6659 (pass_machine_reorg): Convert from a global struct to a subclass of
6660 rtl_opt_pass along with...
6661 (pass_data_machine_reorg): ...new pass_data instance and...
6662 (make_pass_machine_reorg): ...new function.
6663 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
6664 of rtl_opt_pass along with...
6665 (pass_data_sched): ...new pass_data instance and...
6666 (make_pass_sched): ...new function.
6667 (pass_sched2): Convert from a global struct to a subclass of
6668 rtl_opt_pass along with...
6669 (pass_data_sched2): ...new pass_data instance and...
6670 (make_pass_sched2): ...new function.
6671 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
6672 to a subclass of rtl_opt_pass along with...
6673 (pass_data_stack_ptr_mod): ...new pass_data instance and...
6674 (make_pass_stack_ptr_mod): ...new function.
6675 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
6676 to a subclass of rtl_opt_pass along with...
6677 (pass_data_rtl_store_motion): ...new pass_data instance and...
6678 (make_pass_rtl_store_motion): ...new function.
6679 * tracer.c (pass_tracer): Convert from a global struct to a subclass
6680 of gimple_opt_pass along with...
6681 (pass_data_tracer): ...new pass_data instance and...
6682 (make_pass_tracer): ...new function.
6683 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
6684 to a subclass of gimple_opt_pass along with...
6685 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
6686 (make_pass_diagnose_tm_blocks): ...new function.
6687 (pass_lower_tm): Convert from a global struct to a subclass of
6688 gimple_opt_pass along with...
6689 (pass_data_lower_tm): ...new pass_data instance and...
6690 (make_pass_lower_tm): ...new function.
6691 (pass_tm_init): Convert from a global struct to a subclass of
6692 gimple_opt_pass along with...
6693 (pass_data_tm_init): ...new pass_data instance and...
6694 (make_pass_tm_init): ...new function.
6695 (pass_tm_mark): Convert from a global struct to a subclass of
6696 gimple_opt_pass along with...
6697 (pass_data_tm_mark): ...new pass_data instance and...
6698 (make_pass_tm_mark): ...new function.
6699 (pass_tm_edges): Convert from a global struct to a subclass of
6700 gimple_opt_pass along with...
6701 (pass_data_tm_edges): ...new pass_data instance and...
6702 (make_pass_tm_edges): ...new function.
6703 (pass_tm_memopt): Convert from a global struct to a subclass of
6704 gimple_opt_pass along with...
6705 (pass_data_tm_memopt): ...new pass_data instance and...
6706 (make_pass_tm_memopt): ...new function.
6707 (pass_ipa_tm): Convert from a global struct to a subclass of
6708 simple_ipa_opt_pass along with...
6709 (pass_data_ipa_tm): ...new pass_data instance and...
6710 (make_pass_ipa_tm): ...new function.
6711 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
6712 subclass of gimple_opt_pass along with...
6713 (pass_data_call_cdce): ...new pass_data instance and...
6714 (make_pass_call_cdce): ...new function.
6715 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
6716 subclass of gimple_opt_pass along with...
6717 (pass_data_build_cfg): ...new pass_data instance and...
6718 (make_pass_build_cfg): ...new function.
6719 (pass_split_crit_edges): Convert from a global struct to a subclass of
6720 gimple_opt_pass along with...
6721 (pass_data_split_crit_edges): ...new pass_data instance and...
6722 (make_pass_split_crit_edges): ...new function.
6723 (pass_warn_function_return): Convert from a global struct to a
6724 subclass of gimple_opt_pass along with...
6725 (pass_data_warn_function_return): ...new pass_data instance and...
6726 (make_pass_warn_function_return): ...new function.
6727 (pass_warn_function_noreturn): Convert from a global struct to a
6728 subclass of gimple_opt_pass along with...
6729 (pass_data_warn_function_noreturn): ...new pass_data instance and...
6730 (make_pass_warn_function_noreturn): ...new function.
6731 (pass_warn_unused_result): Convert from a global struct to a subclass
6732 of gimple_opt_pass along with...
6733 (pass_data_warn_unused_result): ...new pass_data instance and...
6734 (make_pass_warn_unused_result): ...new function.
6735 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
6736 a subclass of gimple_opt_pass along with...
6737 (pass_data_merge_phi): ...new pass_data instance and...
6738 (make_pass_merge_phi): ...new function.
6739 * tree-complex.c (pass_lower_complex): Convert from a global struct to
6740 a subclass of gimple_opt_pass along with...
6741 (pass_data_lower_complex): ...new pass_data instance and...
6742 (make_pass_lower_complex): ...new function.
6743 (pass_lower_complex_O0): Convert from a global struct to a subclass of
6744 gimple_opt_pass along with...
6745 (pass_data_lower_complex_O0): ...new pass_data instance and...
6746 (make_pass_lower_complex_O0): ...new function.
6747 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
6748 subclass of gimple_opt_pass along with...
6749 (pass_data_lower_eh): ...new pass_data instance and...
6750 (make_pass_lower_eh): ...new function.
6751 (pass_refactor_eh): Convert from a global struct to a subclass of
6752 gimple_opt_pass along with...
6753 (pass_data_refactor_eh): ...new pass_data instance and...
6754 (make_pass_refactor_eh): ...new function.
6755 (pass_lower_resx): Convert from a global struct to a subclass of
6756 gimple_opt_pass along with...
6757 (pass_data_lower_resx): ...new pass_data instance and...
6758 (make_pass_lower_resx): ...new function.
6759 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
6760 of gimple_opt_pass along with...
6761 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
6762 (make_pass_lower_eh_dispatch): ...new function.
6763 (pass_cleanup_eh): Convert from a global struct to a subclass of
6764 gimple_opt_pass along with...
6765 (pass_data_cleanup_eh): ...new pass_data instance and...
6766 (make_pass_cleanup_eh): ...new function.
6767 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
6768 to a subclass of simple_ipa_opt_pass along with...
6769 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
6770 (make_pass_ipa_lower_emutls): ...new function.
6771 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
6772 a subclass of gimple_opt_pass along with...
6773 (pass_data_if_conversion): ...new pass_data instance and...
6774 (make_pass_if_conversion): ...new function.
6775 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
6776 subclass of gimple_opt_pass along with...
6777 (pass_data_build_ssa): ...new pass_data instance and...
6778 (make_pass_build_ssa): ...new function.
6779 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
6780 global struct to a subclass of gimple_opt_pass along with...
6781 (pass_data_loop_distribution): ...new pass_data instance and...
6782 (make_pass_loop_distribution): ...new function.
6783 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
6784 subclass of gimple_opt_pass along with...
6785 (pass_data_mudflap_1): ...new pass_data instance and...
6786 (make_pass_mudflap_1): ...new function.
6787 (pass_mudflap_2): Convert from a global struct to a subclass of
6788 gimple_opt_pass along with...
6789 (pass_data_mudflap_2): ...new pass_data instance and...
6790 (make_pass_mudflap_2): ...new function.
6791 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
6792 subclass of gimple_opt_pass along with...
6793 (pass_data_mudflap_1): ...new pass_data instance and...
6794 (make_pass_mudflap_1): ...new function.
6795 (pass_mudflap_2): Convert from a global struct to a subclass of
6796 gimple_opt_pass along with...
6797 (pass_data_mudflap_2): ...new pass_data instance and...
6798 (make_pass_mudflap_2): ...new function.
6799 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
6800 gimple_opt_pass along with...
6801 (pass_data_nrv): ...new pass_data instance and...
6802 (make_pass_nrv): ...new function.
6803 (pass_return_slot): Convert from a global struct to a subclass of
6804 gimple_opt_pass along with...
6805 (pass_data_return_slot): ...new pass_data instance and...
6806 (make_pass_return_slot): ...new function.
6807 * tree-object-size.c (pass_object_sizes): Convert from a global struct
6808 to a subclass of gimple_opt_pass along with...
6809 (pass_data_object_sizes): ...new pass_data instance and...
6810 (make_pass_object_sizes): ...new function.
6811 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
6812 global struct to a subclass of gimple_opt_pass along with...
6813 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
6815 (make_pass_cleanup_cfg_post_optimizing): ...new function.
6816 (pass_fixup_cfg): Convert from a global struct to a subclass of
6817 gimple_opt_pass along with...
6818 (pass_data_fixup_cfg): ...new pass_data instance and...
6819 (make_pass_fixup_cfg): ...new function.
6820 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
6821 (make_pass_mudflap_1): ...new function.
6822 (pass_mudflap_2): Replace declaration with that of...
6823 (make_pass_mudflap_2): ...new function.
6824 (pass_asan): Replace declaration with that of...
6825 (make_pass_asan): ...new function.
6826 (pass_asan_O0): Replace declaration with that of...
6827 (make_pass_asan_O0): ...new function.
6828 (pass_tsan): Replace declaration with that of...
6829 (make_pass_tsan): ...new function.
6830 (pass_tsan_O0): Replace declaration with that of...
6831 (make_pass_tsan_O0): ...new function.
6832 (pass_lower_cf): Replace declaration with that of...
6833 (make_pass_lower_cf): ...new function.
6834 (pass_refactor_eh): Replace declaration with that of...
6835 (make_pass_refactor_eh): ...new function.
6836 (pass_lower_eh): Replace declaration with that of...
6837 (make_pass_lower_eh): ...new function.
6838 (pass_lower_eh_dispatch): Replace declaration with that of...
6839 (make_pass_lower_eh_dispatch): ...new function.
6840 (pass_lower_resx): Replace declaration with that of...
6841 (make_pass_lower_resx): ...new function.
6842 (pass_build_cfg): Replace declaration with that of...
6843 (make_pass_build_cfg): ...new function.
6844 (pass_early_tree_profile): Replace declaration with that of...
6845 (make_pass_early_tree_profile): ...new function.
6846 (pass_cleanup_eh): Replace declaration with that of...
6847 (make_pass_cleanup_eh): ...new function.
6848 (pass_sra): Replace declaration with that of...
6849 (make_pass_sra): ...new function.
6850 (pass_sra_early): Replace declaration with that of...
6851 (make_pass_sra_early): ...new function.
6852 (pass_early_ipa_sra): Replace declaration with that of...
6853 (make_pass_early_ipa_sra): ...new function.
6854 (pass_tail_recursion): Replace declaration with that of...
6855 (make_pass_tail_recursion): ...new function.
6856 (pass_tail_calls): Replace declaration with that of...
6857 (make_pass_tail_calls): ...new function.
6858 (pass_tree_loop): Replace declaration with that of...
6859 (make_pass_tree_loop): ...new function.
6860 (pass_tree_loop_init): Replace declaration with that of...
6861 (make_pass_tree_loop_init): ...new function.
6862 (pass_lim): Replace declaration with that of...
6863 (make_pass_lim): ...new function.
6864 (pass_tree_unswitch): Replace declaration with that of...
6865 (make_pass_tree_unswitch): ...new function.
6866 (pass_predcom): Replace declaration with that of...
6867 (make_pass_predcom): ...new function.
6868 (pass_iv_canon): Replace declaration with that of...
6869 (make_pass_iv_canon): ...new function.
6870 (pass_scev_cprop): Replace declaration with that of...
6871 (make_pass_scev_cprop): ...new function.
6872 (pass_empty_loop): Replace declaration with that of...
6873 (make_pass_empty_loop): ...new function.
6874 (pass_record_bounds): Replace declaration with that of...
6875 (make_pass_record_bounds): ...new function.
6876 (pass_graphite): Replace declaration with that of...
6877 (make_pass_graphite): ...new function.
6878 (pass_graphite_transforms): Replace declaration with that of...
6879 (make_pass_graphite_transforms): ...new function.
6880 (pass_if_conversion): Replace declaration with that of...
6881 (make_pass_if_conversion): ...new function.
6882 (pass_loop_distribution): Replace declaration with that of...
6883 (make_pass_loop_distribution): ...new function.
6884 (pass_vectorize): Replace declaration with that of...
6885 (make_pass_vectorize): ...new function.
6886 (pass_slp_vectorize): Replace declaration with that of...
6887 (make_pass_slp_vectorize): ...new function.
6888 (pass_complete_unroll): Replace declaration with that of...
6889 (make_pass_complete_unroll): ...new function.
6890 (pass_complete_unrolli): Replace declaration with that of...
6891 (make_pass_complete_unrolli): ...new function.
6892 (pass_parallelize_loops): Replace declaration with that of...
6893 (make_pass_parallelize_loops): ...new function.
6894 (pass_loop_prefetch): Replace declaration with that of...
6895 (make_pass_loop_prefetch): ...new function.
6896 (pass_iv_optimize): Replace declaration with that of...
6897 (make_pass_iv_optimize): ...new function.
6898 (pass_tree_loop_done): Replace declaration with that of...
6899 (make_pass_tree_loop_done): ...new function.
6900 (pass_ch): Replace declaration with that of...
6901 (make_pass_ch): ...new function.
6902 (pass_ccp): Replace declaration with that of...
6903 (make_pass_ccp): ...new function.
6904 (pass_phi_only_cprop): Replace declaration with that of...
6905 (make_pass_phi_only_cprop): ...new function.
6906 (pass_build_ssa): Replace declaration with that of...
6907 (make_pass_build_ssa): ...new function.
6908 (pass_build_alias): Replace declaration with that of...
6909 (make_pass_build_alias): ...new function.
6910 (pass_build_ealias): Replace declaration with that of...
6911 (make_pass_build_ealias): ...new function.
6912 (pass_dominator): Replace declaration with that of...
6913 (make_pass_dominator): ...new function.
6914 (pass_dce): Replace declaration with that of...
6915 (make_pass_dce): ...new function.
6916 (pass_dce_loop): Replace declaration with that of...
6917 (make_pass_dce_loop): ...new function.
6918 (pass_cd_dce): Replace declaration with that of...
6919 (make_pass_cd_dce): ...new function.
6920 (pass_call_cdce): Replace declaration with that of...
6921 (make_pass_call_cdce): ...new function.
6922 (pass_merge_phi): Replace declaration with that of...
6923 (make_pass_merge_phi): ...new function.
6924 (pass_split_crit_edges): Replace declaration with that of...
6925 (make_pass_split_crit_edges): ...new function.
6926 (pass_pre): Replace declaration with that of...
6927 (make_pass_pre): ...new function.
6928 (pass_profile): Replace declaration with that of...
6929 (make_pass_profile): ...new function.
6930 (pass_strip_predict_hints): Replace declaration with that of...
6931 (make_pass_strip_predict_hints): ...new function.
6932 (pass_lower_complex_O0): Replace declaration with that of...
6933 (make_pass_lower_complex_O0): ...new function.
6934 (pass_lower_complex): Replace declaration with that of...
6935 (make_pass_lower_complex): ...new function.
6936 (pass_lower_vector): Replace declaration with that of...
6937 (make_pass_lower_vector): ...new function.
6938 (pass_lower_vector_ssa): Replace declaration with that of...
6939 (make_pass_lower_vector_ssa): ...new function.
6940 (pass_lower_omp): Replace declaration with that of...
6941 (make_pass_lower_omp): ...new function.
6942 (pass_diagnose_omp_blocks): Replace declaration with that of...
6943 (make_pass_diagnose_omp_blocks): ...new function.
6944 (pass_expand_omp): Replace declaration with that of...
6945 (make_pass_expand_omp): ...new function.
6946 (pass_expand_omp_ssa): Replace declaration with that of...
6947 (make_pass_expand_omp_ssa): ...new function.
6948 (pass_object_sizes): Replace declaration with that of...
6949 (make_pass_object_sizes): ...new function.
6950 (pass_strlen): Replace declaration with that of...
6951 (make_pass_strlen): ...new function.
6952 (pass_fold_builtins): Replace declaration with that of...
6953 (make_pass_fold_builtins): ...new function.
6954 (pass_stdarg): Replace declaration with that of...
6955 (make_pass_stdarg): ...new function.
6956 (pass_early_warn_uninitialized): Replace declaration with that of...
6957 (make_pass_early_warn_uninitialized): ...new function.
6958 (pass_late_warn_uninitialized): Replace declaration with that of...
6959 (make_pass_late_warn_uninitialized): ...new function.
6960 (pass_cse_reciprocals): Replace declaration with that of...
6961 (make_pass_cse_reciprocals): ...new function.
6962 (pass_cse_sincos): Replace declaration with that of...
6963 (make_pass_cse_sincos): ...new function.
6964 (pass_optimize_bswap): Replace declaration with that of...
6965 (make_pass_optimize_bswap): ...new function.
6966 (pass_optimize_widening_mul): Replace declaration with that of...
6967 (make_pass_optimize_widening_mul): ...new function.
6968 (pass_warn_function_return): Replace declaration with that of...
6969 (make_pass_warn_function_return): ...new function.
6970 (pass_warn_function_noreturn): Replace declaration with that of...
6971 (make_pass_warn_function_noreturn): ...new function.
6972 (pass_cselim): Replace declaration with that of...
6973 (make_pass_cselim): ...new function.
6974 (pass_phiopt): Replace declaration with that of...
6975 (make_pass_phiopt): ...new function.
6976 (pass_forwprop): Replace declaration with that of...
6977 (make_pass_forwprop): ...new function.
6978 (pass_phiprop): Replace declaration with that of...
6979 (make_pass_phiprop): ...new function.
6980 (pass_tree_ifcombine): Replace declaration with that of...
6981 (make_pass_tree_ifcombine): ...new function.
6982 (pass_dse): Replace declaration with that of...
6983 (make_pass_dse): ...new function.
6984 (pass_nrv): Replace declaration with that of...
6985 (make_pass_nrv): ...new function.
6986 (pass_rename_ssa_copies): Replace declaration with that of...
6987 (make_pass_rename_ssa_copies): ...new function.
6988 (pass_sink_code): Replace declaration with that of...
6989 (make_pass_sink_code): ...new function.
6990 (pass_fre): Replace declaration with that of...
6991 (make_pass_fre): ...new function.
6992 (pass_check_data_deps): Replace declaration with that of...
6993 (make_pass_check_data_deps): ...new function.
6994 (pass_copy_prop): Replace declaration with that of...
6995 (make_pass_copy_prop): ...new function.
6996 (pass_vrp): Replace declaration with that of...
6997 (make_pass_vrp): ...new function.
6998 (pass_uncprop): Replace declaration with that of...
6999 (make_pass_uncprop): ...new function.
7000 (pass_return_slot): Replace declaration with that of...
7001 (make_pass_return_slot): ...new function.
7002 (pass_reassoc): Replace declaration with that of...
7003 (make_pass_reassoc): ...new function.
7004 (pass_rebuild_cgraph_edges): Replace declaration with that of...
7005 (make_pass_rebuild_cgraph_edges): ...new function.
7006 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
7007 (make_pass_remove_cgraph_callee_edges): ...new function.
7008 (pass_build_cgraph_edges): Replace declaration with that of...
7009 (make_pass_build_cgraph_edges): ...new function.
7010 (pass_local_pure_const): Replace declaration with that of...
7011 (make_pass_local_pure_const): ...new function.
7012 (pass_tracer): Replace declaration with that of...
7013 (make_pass_tracer): ...new function.
7014 (pass_warn_unused_result): Replace declaration with that of...
7015 (make_pass_warn_unused_result): ...new function.
7016 (pass_diagnose_tm_blocks): Replace declaration with that of...
7017 (make_pass_diagnose_tm_blocks): ...new function.
7018 (pass_lower_tm): Replace declaration with that of...
7019 (make_pass_lower_tm): ...new function.
7020 (pass_tm_init): Replace declaration with that of...
7021 (make_pass_tm_init): ...new function.
7022 (pass_tm_mark): Replace declaration with that of...
7023 (make_pass_tm_mark): ...new function.
7024 (pass_tm_memopt): Replace declaration with that of...
7025 (make_pass_tm_memopt): ...new function.
7026 (pass_tm_edges): Replace declaration with that of...
7027 (make_pass_tm_edges): ...new function.
7028 (pass_split_functions): Replace declaration with that of...
7029 (make_pass_split_functions): ...new function.
7030 (pass_feedback_split_functions): Replace declaration with that of...
7031 (make_pass_feedback_split_functions): ...new function.
7032 (pass_strength_reduction): Replace declaration with that of...
7033 (make_pass_strength_reduction): ...new function.
7034 (pass_ipa_lower_emutls): Replace declaration with that of...
7035 (make_pass_ipa_lower_emutls): ...new function.
7036 (pass_ipa_function_and_variable_visibility): Replace declaration with
7038 (make_pass_ipa_function_and_variable_visibility): ...new function.
7039 (pass_ipa_tree_profile): Replace declaration with that of...
7040 (make_pass_ipa_tree_profile): ...new function.
7041 (pass_early_local_passes): Replace declaration with that of...
7042 (make_pass_early_local_passes): ...new function.
7043 (pass_ipa_whole_program_visibility): Replace declaration with that
7045 (make_pass_ipa_whole_program_visibility): ...new function.
7046 (pass_ipa_lto_gimple_out): Replace declaration with that of...
7047 (make_pass_ipa_lto_gimple_out): ...new function.
7048 (pass_ipa_increase_alignment): Replace declaration with that of...
7049 (make_pass_ipa_increase_alignment): ...new function.
7050 (pass_ipa_inline): Replace declaration with that of...
7051 (make_pass_ipa_inline): ...new function.
7052 (pass_ipa_free_lang_data): Replace declaration with that of...
7053 (make_pass_ipa_free_lang_data): ...new function.
7054 (pass_ipa_free_inline_summary): Replace declaration with that of...
7055 (make_pass_ipa_free_inline_summary): ...new function.
7056 (pass_ipa_cp): Replace declaration with that of...
7057 (make_pass_ipa_cp): ...new function.
7058 (pass_ipa_reference): Replace declaration with that of...
7059 (make_pass_ipa_reference): ...new function.
7060 (pass_ipa_pure_const): Replace declaration with that of...
7061 (make_pass_ipa_pure_const): ...new function.
7062 (pass_ipa_pta): Replace declaration with that of...
7063 (make_pass_ipa_pta): ...new function.
7064 (pass_ipa_lto_finish_out): Replace declaration with that of...
7065 (make_pass_ipa_lto_finish_out): ...new function.
7066 (pass_ipa_tm): Replace declaration with that of...
7067 (make_pass_ipa_tm): ...new function.
7068 (pass_ipa_profile): Replace declaration with that of...
7069 (make_pass_ipa_profile): ...new function.
7070 (pass_ipa_cdtor_merge): Replace declaration with that of...
7071 (make_pass_ipa_cdtor_merge): ...new function.
7072 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
7074 (make_pass_cleanup_cfg_post_optimizing): ...new function.
7075 (pass_init_datastructures): Replace declaration with that of...
7076 (make_pass_init_datastructures): ...new function.
7077 (pass_fixup_cfg): Replace declaration with that of...
7078 (make_pass_fixup_cfg): ...new function.
7079 (pass_expand): Replace declaration with that of...
7080 (make_pass_expand): ...new function.
7081 (pass_instantiate_virtual_regs): Replace declaration with that of...
7082 (make_pass_instantiate_virtual_regs): ...new function.
7083 (pass_rtl_fwprop): Replace declaration with that of...
7084 (make_pass_rtl_fwprop): ...new function.
7085 (pass_rtl_fwprop_addr): Replace declaration with that of...
7086 (make_pass_rtl_fwprop_addr): ...new function.
7087 (pass_jump): Replace declaration with that of...
7088 (make_pass_jump): ...new function.
7089 (pass_jump2): Replace declaration with that of...
7090 (make_pass_jump2): ...new function.
7091 (pass_lower_subreg): Replace declaration with that of...
7092 (make_pass_lower_subreg): ...new function.
7093 (pass_cse): Replace declaration with that of...
7094 (make_pass_cse): ...new function.
7095 (pass_fast_rtl_dce): Replace declaration with that of...
7096 (make_pass_fast_rtl_dce): ...new function.
7097 (pass_ud_rtl_dce): Replace declaration with that of...
7098 (make_pass_ud_rtl_dce): ...new function.
7099 (pass_rtl_dce): Replace declaration with that of...
7100 (make_pass_rtl_dce): ...new function.
7101 (pass_rtl_dse1): Replace declaration with that of...
7102 (make_pass_rtl_dse1): ...new function.
7103 (pass_rtl_dse2): Replace declaration with that of...
7104 (make_pass_rtl_dse2): ...new function.
7105 (pass_rtl_dse3): Replace declaration with that of...
7106 (make_pass_rtl_dse3): ...new function.
7107 (pass_rtl_cprop): Replace declaration with that of...
7108 (make_pass_rtl_cprop): ...new function.
7109 (pass_rtl_pre): Replace declaration with that of...
7110 (make_pass_rtl_pre): ...new function.
7111 (pass_rtl_hoist): Replace declaration with that of...
7112 (make_pass_rtl_hoist): ...new function.
7113 (pass_rtl_store_motion): Replace declaration with that of...
7114 (make_pass_rtl_store_motion): ...new function.
7115 (pass_cse_after_global_opts): Replace declaration with that of...
7116 (make_pass_cse_after_global_opts): ...new function.
7117 (pass_rtl_ifcvt): Replace declaration with that of...
7118 (make_pass_rtl_ifcvt): ...new function.
7119 (pass_into_cfg_layout_mode): Replace declaration with that of...
7120 (make_pass_into_cfg_layout_mode): ...new function.
7121 (pass_outof_cfg_layout_mode): Replace declaration with that of...
7122 (make_pass_outof_cfg_layout_mode): ...new function.
7123 (pass_loop2): Replace declaration with that of...
7124 (make_pass_loop2): ...new function.
7125 (pass_rtl_loop_init): Replace declaration with that of...
7126 (make_pass_rtl_loop_init): ...new function.
7127 (pass_rtl_move_loop_invariants): Replace declaration with that of...
7128 (make_pass_rtl_move_loop_invariants): ...new function.
7129 (pass_rtl_unswitch): Replace declaration with that of...
7130 (make_pass_rtl_unswitch): ...new function.
7131 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
7132 (make_pass_rtl_unroll_and_peel_loops): ...new function.
7133 (pass_rtl_doloop): Replace declaration with that of...
7134 (make_pass_rtl_doloop): ...new function.
7135 (pass_rtl_loop_done): Replace declaration with that of...
7136 (make_pass_rtl_loop_done): ...new function.
7137 (pass_web): Replace declaration with that of...
7138 (make_pass_web): ...new function.
7139 (pass_cse2): Replace declaration with that of...
7140 (make_pass_cse2): ...new function.
7141 (pass_df_initialize_opt): Replace declaration with that of...
7142 (make_pass_df_initialize_opt): ...new function.
7143 (pass_df_initialize_no_opt): Replace declaration with that of...
7144 (make_pass_df_initialize_no_opt): ...new function.
7145 (pass_reginfo_init): Replace declaration with that of...
7146 (make_pass_reginfo_init): ...new function.
7147 (pass_inc_dec): Replace declaration with that of...
7148 (make_pass_inc_dec): ...new function.
7149 (pass_stack_ptr_mod): Replace declaration with that of...
7150 (make_pass_stack_ptr_mod): ...new function.
7151 (pass_initialize_regs): Replace declaration with that of...
7152 (make_pass_initialize_regs): ...new function.
7153 (pass_combine): Replace declaration with that of...
7154 (make_pass_combine): ...new function.
7155 (pass_if_after_combine): Replace declaration with that of...
7156 (make_pass_if_after_combine): ...new function.
7157 (pass_ree): Replace declaration with that of...
7158 (make_pass_ree): ...new function.
7159 (pass_partition_blocks): Replace declaration with that of...
7160 (make_pass_partition_blocks): ...new function.
7161 (pass_match_asm_constraints): Replace declaration with that of...
7162 (make_pass_match_asm_constraints): ...new function.
7163 (pass_regmove): Replace declaration with that of...
7164 (make_pass_regmove): ...new function.
7165 (pass_split_all_insns): Replace declaration with that of...
7166 (make_pass_split_all_insns): ...new function.
7167 (pass_fast_rtl_byte_dce): Replace declaration with that of...
7168 (make_pass_fast_rtl_byte_dce): ...new function.
7169 (pass_lower_subreg2): Replace declaration with that of...
7170 (make_pass_lower_subreg2): ...new function.
7171 (pass_mode_switching): Replace declaration with that of...
7172 (make_pass_mode_switching): ...new function.
7173 (pass_sms): Replace declaration with that of...
7174 (make_pass_sms): ...new function.
7175 (pass_sched): Replace declaration with that of...
7176 (make_pass_sched): ...new function.
7177 (pass_ira): Replace declaration with that of...
7178 (make_pass_ira): ...new function.
7179 (pass_reload): Replace declaration with that of...
7180 (make_pass_reload): ...new function.
7181 (pass_clean_state): Replace declaration with that of...
7182 (make_pass_clean_state): ...new function.
7183 (pass_branch_prob): Replace declaration with that of...
7184 (make_pass_branch_prob): ...new function.
7185 (pass_value_profile_transformations): Replace declaration with that
7187 (make_pass_value_profile_transformations): ...new function.
7188 (pass_postreload_cse): Replace declaration with that of...
7189 (make_pass_postreload_cse): ...new function.
7190 (pass_gcse2): Replace declaration with that of...
7191 (make_pass_gcse2): ...new function.
7192 (pass_split_after_reload): Replace declaration with that of...
7193 (make_pass_split_after_reload): ...new function.
7194 (pass_branch_target_load_optimize1): Replace declaration with that
7196 (make_pass_branch_target_load_optimize1): ...new function.
7197 (pass_thread_prologue_and_epilogue): Replace declaration with that
7199 (make_pass_thread_prologue_and_epilogue): ...new function.
7200 (pass_stack_adjustments): Replace declaration with that of...
7201 (make_pass_stack_adjustments): ...new function.
7202 (pass_peephole2): Replace declaration with that of...
7203 (make_pass_peephole2): ...new function.
7204 (pass_if_after_reload): Replace declaration with that of...
7205 (make_pass_if_after_reload): ...new function.
7206 (pass_regrename): Replace declaration with that of...
7207 (make_pass_regrename): ...new function.
7208 (pass_cprop_hardreg): Replace declaration with that of...
7209 (make_pass_cprop_hardreg): ...new function.
7210 (pass_reorder_blocks): Replace declaration with that of...
7211 (make_pass_reorder_blocks): ...new function.
7212 (pass_branch_target_load_optimize2): Replace declaration with that
7214 (make_pass_branch_target_load_optimize2): ...new function.
7215 (pass_leaf_regs): Replace declaration with that of...
7216 (make_pass_leaf_regs): ...new function.
7217 (pass_split_before_sched2): Replace declaration with that of...
7218 (make_pass_split_before_sched2): ...new function.
7219 (pass_compare_elim_after_reload): Replace declaration with that of...
7220 (make_pass_compare_elim_after_reload): ...new function.
7221 (pass_sched2): Replace declaration with that of...
7222 (make_pass_sched2): ...new function.
7223 (pass_stack_regs): Replace declaration with that of...
7224 (make_pass_stack_regs): ...new function.
7225 (pass_stack_regs_run): Replace declaration with that of...
7226 (make_pass_stack_regs_run): ...new function.
7227 (pass_df_finish): Replace declaration with that of...
7228 (make_pass_df_finish): ...new function.
7229 (pass_compute_alignments): Replace declaration with that of...
7230 (make_pass_compute_alignments): ...new function.
7231 (pass_duplicate_computed_gotos): Replace declaration with that of...
7232 (make_pass_duplicate_computed_gotos): ...new function.
7233 (pass_variable_tracking): Replace declaration with that of...
7234 (make_pass_variable_tracking): ...new function.
7235 (pass_free_cfg): Replace declaration with that of...
7236 (make_pass_free_cfg): ...new function.
7237 (pass_machine_reorg): Replace declaration with that of...
7238 (make_pass_machine_reorg): ...new function.
7239 (pass_cleanup_barriers): Replace declaration with that of...
7240 (make_pass_cleanup_barriers): ...new function.
7241 (pass_delay_slots): Replace declaration with that of...
7242 (make_pass_delay_slots): ...new function.
7243 (pass_split_for_shorten_branches): Replace declaration with that of...
7244 (make_pass_split_for_shorten_branches): ...new function.
7245 (pass_split_before_regstack): Replace declaration with that of...
7246 (make_pass_split_before_regstack): ...new function.
7247 (pass_convert_to_eh_region_ranges): Replace declaration with that
7249 (make_pass_convert_to_eh_region_ranges): ...new function.
7250 (pass_shorten_branches): Replace declaration with that of...
7251 (make_pass_shorten_branches): ...new function.
7252 (pass_set_nothrow_function_flags): Replace declaration with that of...
7253 (make_pass_set_nothrow_function_flags): ...new function.
7254 (pass_dwarf2_frame): Replace declaration with that of...
7255 (make_pass_dwarf2_frame): ...new function.
7256 (pass_final): Replace declaration with that of...
7257 (make_pass_final): ...new function.
7258 (pass_rtl_seqabstr): Replace declaration with that of...
7259 (make_pass_rtl_seqabstr): ...new function.
7260 (pass_release_ssa_names): Replace declaration with that of...
7261 (make_pass_release_ssa_names): ...new function.
7262 (pass_early_inline): Replace declaration with that of...
7263 (make_pass_early_inline): ...new function.
7264 (pass_inline_parameters): Replace declaration with that of...
7265 (make_pass_inline_parameters): ...new function.
7266 (pass_update_address_taken): Replace declaration with that of...
7267 (make_pass_update_address_taken): ...new function.
7268 (pass_convert_switch): Replace declaration with that of...
7269 (make_pass_convert_switch): ...new function.
7270 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
7271 to a subclass of simple_ipa_opt_pass along with...
7272 (pass_data_ipa_tree_profile): ...new pass_data instance and...
7273 (make_pass_ipa_tree_profile): ...new function.
7274 * tree-sra.c (pass_sra_early): Convert from a global struct to a
7275 subclass of gimple_opt_pass along with...
7276 (pass_data_sra_early): ...new pass_data instance and...
7277 (make_pass_sra_early): ...new function.
7278 (pass_sra): Convert from a global struct to a subclass of
7279 gimple_opt_pass along with...
7280 (pass_data_sra): ...new pass_data instance and...
7281 (make_pass_sra): ...new function.
7282 (pass_early_ipa_sra): Convert from a global struct to a subclass of
7283 gimple_opt_pass along with...
7284 (pass_data_early_ipa_sra): ...new pass_data instance and...
7285 (make_pass_early_ipa_sra): ...new function.
7286 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
7287 subclass of gimple_opt_pass along with...
7288 (pass_data_ccp): ...new pass_data instance and...
7289 (make_pass_ccp): ...new function.
7290 (pass_fold_builtins): Convert from a global struct to a subclass of
7291 gimple_opt_pass along with...
7292 (pass_data_fold_builtins): ...new pass_data instance and...
7293 (make_pass_fold_builtins): ...new function.
7294 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
7295 subclass of gimple_opt_pass along with...
7296 (pass_data_copy_prop): ...new pass_data instance and...
7297 (make_pass_copy_prop): ...new function.
7298 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
7299 global struct to a subclass of gimple_opt_pass along with...
7300 (pass_data_rename_ssa_copies): ...new pass_data instance and...
7301 (make_pass_rename_ssa_copies): ...new function.
7302 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
7303 subclass of gimple_opt_pass along with...
7304 (pass_data_dce): ...new pass_data instance and...
7305 (make_pass_dce): ...new function.
7306 (pass_dce_loop): Convert from a global struct to a subclass of
7307 gimple_opt_pass along with...
7308 (pass_data_dce_loop): ...new pass_data instance and...
7309 (make_pass_dce_loop): ...new function.
7310 (pass_cd_dce): Convert from a global struct to a subclass of
7311 gimple_opt_pass along with...
7312 (pass_data_cd_dce): ...new pass_data instance and...
7313 (make_pass_cd_dce): ...new function.
7314 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
7315 subclass of gimple_opt_pass along with...
7316 (pass_data_dominator): ...new pass_data instance and...
7317 (make_pass_dominator): ...new function.
7318 (pass_phi_only_cprop): Convert from a global struct to a subclass of
7319 gimple_opt_pass along with...
7320 (pass_data_phi_only_cprop): ...new pass_data instance and...
7321 (make_pass_phi_only_cprop): ...new function.
7322 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
7323 subclass of gimple_opt_pass along with...
7324 (pass_data_dse): ...new pass_data instance and...
7325 (make_pass_dse): ...new function.
7326 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
7327 a subclass of gimple_opt_pass along with...
7328 (pass_data_forwprop): ...new pass_data instance and...
7329 (make_pass_forwprop): ...new function.
7330 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
7331 struct to a subclass of gimple_opt_pass along with...
7332 (pass_data_tree_ifcombine): ...new pass_data instance and...
7333 (make_pass_tree_ifcombine): ...new function.
7334 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
7335 subclass of gimple_opt_pass along with...
7336 (pass_data_ch): ...new pass_data instance and...
7337 (make_pass_ch): ...new function.
7338 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
7339 subclass of gimple_opt_pass along with...
7340 (pass_data_tree_loop): ...new pass_data instance and...
7341 (make_pass_tree_loop): ...new function.
7342 (pass_tree_loop_init): Convert from a global struct to a subclass of
7343 gimple_opt_pass along with...
7344 (pass_data_tree_loop_init): ...new pass_data instance and...
7345 (make_pass_tree_loop_init): ...new function.
7346 (pass_lim): Convert from a global struct to a subclass of
7347 gimple_opt_pass along with...
7348 (pass_data_lim): ...new pass_data instance and...
7349 (make_pass_lim): ...new function.
7350 (pass_tree_unswitch): Convert from a global struct to a subclass of
7351 gimple_opt_pass along with...
7352 (pass_data_tree_unswitch): ...new pass_data instance and...
7353 (make_pass_tree_unswitch): ...new function.
7354 (pass_predcom): Convert from a global struct to a subclass of
7355 gimple_opt_pass along with...
7356 (pass_data_predcom): ...new pass_data instance and...
7357 (make_pass_predcom): ...new function.
7358 (pass_vectorize): Convert from a global struct to a subclass of
7359 gimple_opt_pass along with...
7360 (pass_data_vectorize): ...new pass_data instance and...
7361 (make_pass_vectorize): ...new function.
7362 (pass_graphite): Convert from a global struct to a subclass of
7363 gimple_opt_pass along with...
7364 (pass_data_graphite): ...new pass_data instance and...
7365 (make_pass_graphite): ...new function.
7366 (pass_graphite_transforms): Convert from a global struct to a subclass
7367 of gimple_opt_pass along with...
7368 (pass_data_graphite_transforms): ...new pass_data instance and...
7369 (make_pass_graphite_transforms): ...new function.
7370 (pass_check_data_deps): Convert from a global struct to a subclass of
7371 gimple_opt_pass along with...
7372 (pass_data_check_data_deps): ...new pass_data instance and...
7373 (make_pass_check_data_deps): ...new function.
7374 (pass_iv_canon): Convert from a global struct to a subclass of
7375 gimple_opt_pass along with...
7376 (pass_data_iv_canon): ...new pass_data instance and...
7377 (make_pass_iv_canon): ...new function.
7378 (pass_scev_cprop): Convert from a global struct to a subclass of
7379 gimple_opt_pass along with...
7380 (pass_data_scev_cprop): ...new pass_data instance and...
7381 (make_pass_scev_cprop): ...new function.
7382 (pass_record_bounds): Convert from a global struct to a subclass of
7383 gimple_opt_pass along with...
7384 (pass_data_record_bounds): ...new pass_data instance and...
7385 (make_pass_record_bounds): ...new function.
7386 (pass_complete_unroll): Convert from a global struct to a subclass of
7387 gimple_opt_pass along with...
7388 (pass_data_complete_unroll): ...new pass_data instance and...
7389 (make_pass_complete_unroll): ...new function.
7390 (pass_complete_unrolli): Convert from a global struct to a subclass of
7391 gimple_opt_pass along with...
7392 (pass_data_complete_unrolli): ...new pass_data instance and...
7393 (make_pass_complete_unrolli): ...new function.
7394 (pass_parallelize_loops): Convert from a global struct to a subclass
7395 of gimple_opt_pass along with...
7396 (pass_data_parallelize_loops): ...new pass_data instance and...
7397 (make_pass_parallelize_loops): ...new function.
7398 (pass_loop_prefetch): Convert from a global struct to a subclass of
7399 gimple_opt_pass along with...
7400 (pass_data_loop_prefetch): ...new pass_data instance and...
7401 (make_pass_loop_prefetch): ...new function.
7402 (pass_iv_optimize): Convert from a global struct to a subclass of
7403 gimple_opt_pass along with...
7404 (pass_data_iv_optimize): ...new pass_data instance and...
7405 (make_pass_iv_optimize): ...new function.
7406 (pass_tree_loop_done): Convert from a global struct to a subclass of
7407 gimple_opt_pass along with...
7408 (pass_data_tree_loop_done): ...new pass_data instance and...
7409 (make_pass_tree_loop_done): ...new function.
7410 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
7411 struct to a subclass of gimple_opt_pass along with...
7412 (pass_data_cse_reciprocals): ...new pass_data instance and...
7413 (make_pass_cse_reciprocals): ...new function.
7414 (pass_cse_sincos): Convert from a global struct to a subclass of
7415 gimple_opt_pass along with...
7416 (pass_data_cse_sincos): ...new pass_data instance and...
7417 (make_pass_cse_sincos): ...new function.
7418 (pass_optimize_bswap): Convert from a global struct to a subclass of
7419 gimple_opt_pass along with...
7420 (pass_data_optimize_bswap): ...new pass_data instance and...
7421 (make_pass_optimize_bswap): ...new function.
7422 (pass_optimize_widening_mul): Convert from a global struct to a
7423 subclass of gimple_opt_pass along with...
7424 (pass_data_optimize_widening_mul): ...new pass_data instance and...
7425 (make_pass_optimize_widening_mul): ...new function.
7426 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
7427 subclass of gimple_opt_pass along with...
7428 (pass_data_phiopt): ...new pass_data instance and...
7429 (make_pass_phiopt): ...new function.
7430 (pass_cselim): Convert from a global struct to a subclass of
7431 gimple_opt_pass along with...
7432 (pass_data_cselim): ...new pass_data instance and...
7433 (make_pass_cselim): ...new function.
7434 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
7435 subclass of gimple_opt_pass along with...
7436 (pass_data_phiprop): ...new pass_data instance and...
7437 (make_pass_phiprop): ...new function.
7438 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
7439 subclass of gimple_opt_pass along with...
7440 (pass_data_pre): ...new pass_data instance and...
7441 (make_pass_pre): ...new function.
7442 (pass_fre): Convert from a global struct to a subclass of
7443 gimple_opt_pass along with...
7444 (pass_data_fre): ...new pass_data instance and...
7445 (make_pass_fre): ...new function.
7446 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
7447 subclass of gimple_opt_pass along with...
7448 (pass_data_reassoc): ...new pass_data instance and...
7449 (make_pass_reassoc): ...new function.
7450 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
7451 subclass of gimple_opt_pass along with...
7452 (pass_data_sink_code): ...new pass_data instance and...
7453 (make_pass_sink_code): ...new function.
7454 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
7455 subclass of gimple_opt_pass along with...
7456 (pass_data_strlen): ...new pass_data instance and...
7457 (make_pass_strlen): ...new function.
7458 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
7459 struct to a subclass of gimple_opt_pass along with...
7460 (pass_data_build_alias): ...new pass_data instance and...
7461 (make_pass_build_alias): ...new function.
7462 (pass_build_ealias): Convert from a global struct to a subclass of
7463 gimple_opt_pass along with...
7464 (pass_data_build_ealias): ...new pass_data instance and...
7465 (make_pass_build_ealias): ...new function.
7466 (pass_ipa_pta): Convert from a global struct to a subclass of
7467 simple_ipa_opt_pass along with...
7468 (pass_data_ipa_pta): ...new pass_data instance and...
7469 (make_pass_ipa_pta): ...new function.
7470 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
7471 subclass of gimple_opt_pass along with...
7472 (pass_data_uncprop): ...new pass_data instance and...
7473 (make_pass_uncprop): ...new function.
7474 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
7475 global struct to a subclass of gimple_opt_pass along with...
7476 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
7477 (make_pass_late_warn_uninitialized): ...new function.
7478 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
7479 to a subclass of gimple_opt_pass along with...
7480 (pass_data_init_datastructures): ...new pass_data instance and...
7481 (make_pass_init_datastructures): ...new function.
7482 (pass_early_warn_uninitialized): Convert from a global struct to a
7483 subclass of gimple_opt_pass along with...
7484 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
7485 (make_pass_early_warn_uninitialized): ...new function.
7486 (pass_update_address_taken): Convert from a global struct to a
7487 subclass of gimple_opt_pass along with...
7488 (pass_data_update_address_taken): ...new pass_data instance and...
7489 (make_pass_update_address_taken): ...new function.
7490 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
7491 struct to a subclass of gimple_opt_pass along with...
7492 (pass_data_release_ssa_names): ...new pass_data instance and...
7493 (make_pass_release_ssa_names): ...new function.
7494 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
7495 subclass of gimple_opt_pass along with...
7496 (pass_data_stdarg): ...new pass_data instance and...
7497 (make_pass_stdarg): ...new function.
7498 * tree-switch-conversion.c (pass_convert_switch): Convert from a
7499 global struct to a subclass of gimple_opt_pass along with...
7500 (pass_data_convert_switch): ...new pass_data instance and...
7501 (make_pass_convert_switch): ...new function.
7502 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
7503 to a subclass of gimple_opt_pass along with...
7504 (pass_data_tail_recursion): ...new pass_data instance and...
7505 (make_pass_tail_recursion): ...new function.
7506 (pass_tail_calls): Convert from a global struct to a subclass of
7507 gimple_opt_pass along with...
7508 (pass_data_tail_calls): ...new pass_data instance and...
7509 (make_pass_tail_calls): ...new function.
7510 * tree-vect-generic.c (pass_lower_vector): Convert from a global
7511 struct to a subclass of gimple_opt_pass along with...
7512 (pass_data_lower_vector): ...new pass_data instance and...
7513 (make_pass_lower_vector): ...new function.
7514 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
7515 gimple_opt_pass along with...
7516 (pass_data_lower_vector_ssa): ...new pass_data instance and...
7517 (make_pass_lower_vector_ssa): ...new function.
7518 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
7519 to a subclass of gimple_opt_pass along with...
7520 (pass_data_slp_vectorize): ...new pass_data instance and...
7521 (make_pass_slp_vectorize): ...new function.
7522 (pass_ipa_increase_alignment): Convert from a global struct to a
7523 subclass of simple_ipa_opt_pass along with...
7524 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
7525 (make_pass_ipa_increase_alignment): ...new function.
7526 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
7527 gimple_opt_pass along with...
7528 (pass_data_vrp): ...new pass_data instance and...
7529 (make_pass_vrp): ...new function.
7530 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
7531 subclass of simple_ipa_opt_pass along with...
7532 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
7533 (make_pass_ipa_free_lang_data): ...new function.
7534 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
7535 gimple_opt_pass along with...
7536 (pass_data_tsan): ...new pass_data instance and...
7537 (make_pass_tsan): ...new function.
7538 (pass_tsan_O0): Convert from a global struct to a subclass of
7539 gimple_opt_pass along with...
7540 (pass_data_tsan_O0): ...new pass_data instance and...
7541 (make_pass_tsan_O0): ...new function.
7542 * var-tracking.c (pass_variable_tracking): Convert from a global
7543 struct to a subclass of rtl_opt_pass along with...
7544 (pass_data_variable_tracking): ...new pass_data instance and...
7545 (make_pass_variable_tracking): ...new function.
7546 * web.c (pass_web): Convert from a global struct to a subclass of
7547 rtl_opt_pass along with...
7548 (pass_data_web): ...new pass_data instance and...
7549 (make_pass_web): ...new function.
7550 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
7551 declaration with that of...
7552 (make_pass_mode_switch_use): ...new function.
7553 (pass_resolve_sw_modes): Replace declaration with that of...
7554 (make_pass_resolve_sw_modes): ...new function.
7555 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
7556 from a global struct to a subclass of rtl_opt_pass along with...
7557 (pass_data_mode_switch_use): ...new pass_data instance and...
7558 (make_pass_mode_switch_use): ...new function.
7559 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
7560 from a global struct to a subclass of rtl_opt_pass along with...
7561 (pass_data_resolve_sw_modes): ...new pass_data instance and...
7562 (make_pass_resolve_sw_modes): ...new function.
7563 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
7564 struct to a subclass of rtl_opt_pass along with...
7565 (pass_data_insert_vzeroupper): ...new pass_data instance and...
7566 (make_pass_insert_vzeroupper): ...new function.
7567 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
7568 global struct to a subclass of rtl_opt_pass along with...
7569 (pass_data_work_around_errata): ...new pass_data instance and...
7570 (make_pass_work_around_errata): ...new function.
7571 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
7572 struct to a subclass of rtl_opt_pass along with...
7573 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
7574 (make_pass_mips_machine_reorg2): ...new function.
7576 2013-08-05 David Malcolm <dmalcolm@redhat.com>
7578 * passes.c (pass_manager::operator new): New.
7580 2013-08-05 David Malcolm <dmalcolm@redhat.com>
7582 Handwritten part of conversion of passes to C++ classes.
7584 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
7585 (toplev.o): Add dep on PASS_MANAGER_H.
7586 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
7587 of early local pases to reflect this moving from a global to a
7588 member of gcc::pass_manager.
7589 (cgraph_add_new_function): Likewise.
7590 * lto-cgraph.c (lto_output_node): Update for conversion of
7591 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
7592 * passes.c (opt_pass::clone): New.
7593 (opt_pass::gate): New.
7594 (opt_pass::execute): New.
7595 (opt_pass::opt_pass): New.
7596 (pass_manager::execute_early_local_passes): New.
7597 (pass_manager::execute_pass_mode_switching): new.
7598 (finish_optimization_passes): Convert to...
7599 (pass_manager::finish_optimization_passes): ...this.
7600 (finish_optimization_passes): Update for conversion of passes to
7602 (register_dump_files_1): Use has_gate since we cannot portably
7603 check a vtable entry against NULL.
7604 (dump_one_pass): Likewise.
7605 (ipa_write_summaries_2): Likewise.
7606 (ipa_write_optimization_summaries_1): Likewise.
7607 (ipa_read_summaries_1): Likewise.
7608 (ipa_read_optimization_summaries_1): Likewise.
7609 (execute_ipa_stmt_fixups): Likewise.
7610 (pass_manager::pass_manager): Rewrite pass-creation, invoking
7611 pass-creation functions rather than wiring up globals, and
7612 storing the results in fields of pass_manager generated using
7614 (pass_manager::dump_profile_report): Update for conversion of
7615 passes to C++ classes.
7616 (pass_manager::execute_ipa_summary_passes): Likewise.
7617 (execute_one_ipa_transform_pass): Likewise.
7618 (execute_one_pass): Use has_gate and has_execute since we cannot
7619 portably check a vtable entry against NULL.
7620 * pass_manager.h (pass_manager::finish_optimization_passes): New.
7621 (pass_manager): Use pass-instances.def to add fields for the
7622 various pass instances.
7623 * toplev.c (finalize): Update for move of
7624 finish_optimization_passes to a method of gcc::pass_manager.
7625 * toplev.h (finish_optimization_passes): Move to method of class
7627 * tree-pass.h (struct pass_data): New.
7628 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
7629 (opt_pass::gate): Convert to virtual function.
7630 (opt_pass::~opt_pass): New.
7631 (opt_pass::clone): New.
7632 (opt_pass::execute): Convert to virtual function.
7633 (opt_pass::opt_pass): New.
7634 (opt_pass::ctxt_): new.
7635 (gimple_opt_pass): Convert to subclass of opt_pass.
7636 (gimple_opt_pass::gimple_opt_pass): New.
7637 (rtl_opt_pass): Convert to subclass of opt_pass.
7638 (rtl_opt_pass::rtl_opt_pass): New.
7639 (ipa_opt_pass_d): Convert to subclass of opt_pass.
7640 (ipa_opt_pass_d::ipa_opt_pass_d): New.
7641 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
7642 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
7643 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
7644 invocation of pass_mode_switching to reflect this moving from a
7645 global to a member of gcc::pass_manager.
7646 (ix86_option_override): Rework how pass_insert_vzeroupper is
7647 added to the pass_manager to reflect autogenerated changes.
7648 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
7650 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
7652 PR rtl-optimization/57708
7653 * recog.c (peep2_find_free_register): Validate all regs in a
7656 2013-08-05 Jan Hubicka <jh@suse.cz>
7659 * cgraph.c (verify_cgraph_node): Accept local flags from other
7661 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
7662 (function_and_variable_visibility): Likewise.
7663 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
7665 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
7667 * graph.c (init_graph_slim_pretty_print): Remove.
7668 (print_graph_cfg): Do not call it. Use local pretty printer.
7669 (start_graph_dump): Likewise.
7671 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
7673 * gimple-pretty-print.c (buffer): Remove.
7674 (initialized): Likewise.
7675 (maybe_init_pretty_print): Likewise.
7676 (print_gimple_stmt): Do not call it. Use non-static local
7677 pretty_printer variable.
7678 (print_gimple_expr): Likewise.
7679 (print_gimple_seq): Likewise.
7680 (gimple_dump_bb): Likewise.
7682 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
7684 * asan.c (asan_pp): Remove.
7685 (asan_pp_initialized): Likewise.
7686 (asan_pp_initialize): Likewise.
7687 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
7688 (asan_emit_stack_protection): Tidy. Use local pretty printer.
7689 (asan_add_global): Likewise.
7691 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
7693 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
7694 * diagnostic.h (diagnostic_flush_buffer): Adjust.
7695 * pretty-print.c (pp_formatted_text_data): Likewise.
7696 (pp_indent): Rename from pp_base_indent.
7697 (pp_format): Rename from pp_base_format.
7698 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
7699 (pp_format_verbatim): Rename from pp_base_format_verbatim.
7700 (pp_flush): Rename from pp_base_flush.
7701 (pp_set_line_maximum_length): Rename from
7702 pp_base_set_line_maximum_length.
7703 (pp_clear_output_area): Rename from pp_base_clear_output_area.
7704 (pp_set_prefix): Rename from pp_base_set_prefix.
7705 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
7706 (pp_emit_prefix): Rename from pp_base_emit_prefix.
7707 (pp_append_text): Rename from pp_base_append_text.
7708 (pp_formatted_text): Rename from pp_base_formatted_text.
7709 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
7710 (pp_remaining_character_count_for_line): Rename from
7711 pp_base_remaining_character_count_for_line.
7712 (pp_newline): Rename from pp_base_newline.
7713 (pp_character): Rename from pp_base_character.
7714 (pp_string): Rename from pp_base_string.
7715 (pp_maybe_space): Rename from pp_base_maybe_space.
7716 * asan.c (asan_pp_string): Adjust.
7717 (asan_emit_stack_protection): Likewise.
7718 (asan_add_global): Likewise.
7719 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
7720 * tree-mudflap.c (mf_varname_tree): Likewise.
7721 * tree-pretty-print.c (pp_tree_identifier): Rename from
7722 pp_base_tree_identifier.
7723 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
7724 Declare as function.
7726 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
7728 * pretty-print.h (pp_bar_bar): New.
7729 (pp_ampersand_ampersand): Likewise.
7730 (pp_less_equal): Likewise.
7731 (pp_greater_equal): Likewise.
7732 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
7733 printer functions instead of pp_string or operators and punctuators.
7734 (dump_gimple_call): Likewise.
7735 (dump_gimple_omp_for): Likewise.
7736 (dump_gimple_transaction): Likewise.
7737 (dump_gimple_phi): Likewise.
7738 (pp_gimple_stmt_1): Likewise.
7739 * sched-vis.c (print_insn): Likewise.
7740 * tree-mudflap.c (mf_varname_tree): Likewise.
7741 * tree-pretty-print.c (dump_block_node): Likewise.
7742 (dump_generic_node): Likewise.
7744 2013-08-02 Jan Hubicka <jh@suse.cz>
7746 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
7748 * lto-streamer-out.c (tree_is_indexable): Results decls and
7749 parm decls are not indexable.
7750 (DFS_write_tree_body): Do not follow args and results.
7751 (hash_tree): Likewise.
7752 (output_functions): Rearrange so struct function is needed
7753 only when real body is output; be able to also ouptut abstract
7754 functions; output DECL_ARGUMENTS and DECL_RESULT.
7755 (lto_output): When not in WPA, ale store abstract functions.
7756 (write_symbol): Do not care about RESULT_DECL.
7757 (output_symbol_p): Handle correctly sbtract decls.
7758 * lto-streamer-in.c (input_function): Rearrange so struct
7759 function can be NULL at entry; allow streaming of
7760 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
7761 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
7762 sanity check during LTO.
7763 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
7764 RESULT_DECl and DECL_ARGUMENTS.
7765 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
7768 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
7770 * pretty-print.h (pp_underscore): New.
7772 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
7773 printer functions instead of pp_character.
7774 (dump_binary_rhs): Likewise.
7775 (dump_ternary_rhs): Likewise.
7776 (dump_gimple_call_args): Likewise.
7777 (pp_points_to_solution): Likewise.
7778 (dump_gimple_call): Likewise.
7779 (dump_gimple_switch): Likewise.
7780 (dump_gimple_cond): Likewise.
7781 (dump_gimple_bind): Likewise.
7782 (dump_gimple_try): Likewise.
7783 (dump_gimple_omp_for): Likewise.
7784 (dump_gimple_omp_continue): Likewise.
7785 (dump_gimple_omp_single): Likewise.
7786 (dump_gimple_omp_sections): Likewise.
7787 (dump_gimple_omp_block): Likewise.
7788 (dump_gimple_omp_critical): Likewise.
7789 (dump_gimple_transaction): Likewise.
7790 (dump_gimple_asm): Likewise.
7791 (dump_gimple_phi): Likewise.
7792 (dump_gimple_omp_parallel): Likewise.
7793 (dump_gimple_omp_task): Likewise.
7794 (dump_gimple_omp_atomic_load): Likewise.
7795 (dump_gimple_omp_atomic_store): Likewise.
7796 (dump_gimple_mem_ops): Likewise.
7797 (pp_gimple_stmt_1): Likewise.
7798 (pp_cfg_jump): Likewise.
7799 (dump_implicit_edges): Likewise.
7800 (gimple_dump_bb_for_graph): Likewise.
7801 * graph.c (draw_cfg_node): Likewise.
7802 * langhooks.c (lhd_print_error_function): Likewise.
7803 * sched-vis.c (print_exp): Likewise.
7804 (print_value): Likewise.
7805 (print_pattern): Likewise.
7806 (print_insn): Likewise.
7807 (rtl_dump_bb_for_graph): Likewise.
7808 * tree-pretty-print.c (dump_function_declaration): Likewise.
7809 (dump_array_domain): Likewise.
7810 (dump_omp_clause): Likewise.
7811 (dump_location): Likewise.
7812 (dump_generic_node): Likewise.
7813 (print_struct_decl): Likewise.
7814 * diagnostic.c (diagnostic_show_locus): Use pp_space.
7816 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
7818 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
7819 candidate table when replacing a candidate statement.
7820 (replace_rhs_if_not_dup): Likewise.
7821 (replace_one_candidate): Likewise.
7823 2013-08-02 Jan Hubicka <jh@suse.cz>
7824 Martin Liska <marxin.liska@gmail.com>
7826 * cgraphunit.c (add_new_function): Fix logic when adding from
7828 (assemble_thunk): Rename to ...
7829 (expand_thunk); .. this one; export; get it working with
7830 general functions; make produced gimple valid.
7831 * cgraph.h (expand_thunk): Declare.
7833 2013-08-02 Jan Hubicka <jh@suse.cz>
7835 * ipa-cp.c (gather_context_independent_values): Use
7836 ipa_get_param_move_cost.
7837 (get_replacement_map): Remove PARAM; move parameter folding
7839 (create_specialized_node): Update.
7840 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
7841 assert that we have gimple body; update move_cost.
7842 (count_formal_params): Assert that we have gimple body.
7843 (ipa_dump_param): New function.
7844 (ipa_alloc_node_params): Break out from ...
7845 (ipa_initialize_node_params): ... here.
7846 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
7847 (ipa_write_node_info): Stream move costs.
7848 (ipa_read_node_info): Read move costs.
7849 (ipa_update_after_lto_read): Do not recompute node params.
7850 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
7851 (ipa_get_param): Check we are not in WPA.
7852 (ipa_get_param_move_cost): New.
7853 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
7854 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
7855 parm numbers to be present.
7857 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
7859 PR rtl-optimization/58048
7860 * lra-constraints.c (process_alt_operands): Don't check asm
7861 operand on register.
7863 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
7865 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
7866 the implied StoreLoad barrier for atomic operations if before.
7868 2013-08-02 Jan Hubicka <jh@suse.cz>
7869 Martin Liska <marxin.liska@gmail.com>
7871 * cgraph.c (cgraph_function_body_availability): Do not check
7873 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
7874 symtab_node_availability): Declare.
7875 * ipa.c (can_replace_by_local_alias): New.
7876 (function_and_variable_visibility): Use it.
7877 * symtab.c (symtab_for_node_and_aliases,
7878 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
7880 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
7882 PR rtl-optimization/57963
7883 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
7884 (lra_constraints): Use them.
7886 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
7888 * config/arm/types.md (define_attr "type"): Add "load_acq"
7890 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
7892 (cortex_a53_store1): Likewise.
7894 2013-08-01 Jan Hubicka <jh@suse.cz>
7896 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
7897 partitions are not needed.
7899 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
7901 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
7902 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
7903 MAYBE_NON_Q_CLASS_P where appropriate.
7905 2013-08-01 Jan Hubicka <jh@suse.cz>
7907 * cgraph.h (release_function_body): Declare.
7908 * tree.c (free_lang_data_in_decl): Free, parameters and return values
7909 of unused delcarations.
7911 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7913 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
7914 RTL form when subtracting a constant.
7916 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7918 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
7919 Generate canonical plus rtx with negated immediate instead of minus
7921 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
7923 2013-08-01 Jan Hubicka <jh@suse.cz>
7925 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
7926 (cgraph_release_function_body): Likewise.
7927 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
7928 * cgraph.h (cgrpah_node): Rename abstract_and_needed
7929 to used_as_abstract_origin.
7930 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
7931 symbols used as abstract origins.
7932 * cgraphunit.c (analyze_functions): Update.
7933 * ipa.c (symtab_remove_unreachable_nodes): Recompute
7934 used_as_abstract_origin.
7935 * tree-inline.c (tree_function_versioning): Update
7936 used_as_abstract_origin; be ready for DECL_RESULT and
7937 DECL_ARGUMENTS to be NULL.
7939 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
7940 for abstract functions.
7941 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
7944 2013-08-01 Jan Hubicka <jh@suse.cz>
7946 * profile.c (compute_value_histograms): Fix thinko.
7948 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
7950 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
7951 aarch-common-protos.h to extra_headers.
7952 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
7953 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
7954 * config/aarch64/t-aarch64 (aarch-common.o): Define.
7956 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
7958 * config/aarch64/aarch64.md (define_attr "type"): Delete.
7959 Include "../arm/types.md". Define "type" attribute for all patterns.
7960 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
7963 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
7965 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
7966 to support power8 load fusion.
7967 (fusion_gpr_mem_load): Likewise.
7969 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
7971 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
7972 declarations for power8 load fusion.
7973 (emit_fusion_gpr_load): Likewise.
7975 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
7976 tuning for power8, turn on fusion mode by default. Turn on sign
7977 extending fusion mode if normal fusion mode is on, and we are at
7979 (fusion_gpr_load_p): New function, return true if we can fuse an
7980 addis instruction with a dependent load to a GPR.
7981 (emit_fusion_gpr_load): Emit the instructions for power8 load
7984 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
7985 (VSX load fusion peepholes): New peepholes to fuse together an
7986 addi instruction with a VSX load instruction.
7988 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
7989 peepholes to fuse an addis instruction with a load to a GPR base
7990 register. If we are supporting sign extending fusions, convert
7991 sign extending loads to zero extending loads and add an explicit
7994 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
7996 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
7997 aarch-common-protos.h to extra_headers.
7998 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
7999 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
8000 (arm_early_store_addr_dep): Likewise.
8001 (arm_no_early_alu_shift_dep): Likewise.
8002 (arm_no_early_alu_shift_value_dep): Likewise.
8003 (arm_no_early_mul_dep): Likewise.
8004 (arm_no_early_store_addr_dep): Likewise.
8005 (arm_mac_accumulator_is_mul_result): Likewise.
8006 (arm_mac_accumulator_is_result): Likewise.
8007 * config/arm/aarch-common.c: ... here. New file.
8008 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
8010 (arm_early_store_addr_dep): Likewise.
8011 (arm_no_early_alu_shift_dep): Likewise.
8012 (arm_no_early_alu_shift_value_dep): Likewise.
8013 (arm_no_early_mul_dep): Likewise.
8014 (arm_no_early_store_addr_dep): Likewise.
8015 (arm_mac_accumulator_is_mul_result): Likewise.
8016 (arm_mac_accumulator_is_result): Likewise.
8017 * config/arm/aarch-common-protos.h: ... here. New file.
8018 * config/arm/t-arm (aarch-common.o): Define.
8020 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
8022 * config/arm/arm.md: Include new file "types.md".
8023 (define_attr "type"): Move from here to ...
8024 (define_attr "mul32"): Likewise.
8025 (define_attr "mul64"): Likewise.
8026 * config/arm/types.md: ... here. New file.
8028 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
8030 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
8031 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
8033 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8035 * gen-pass-instances.awk: Fix offset of substr().
8037 2013-07-31 David Malcolm <dmalcolm@redhat.com>
8039 * Makefile.in (pass-instances.def): New.
8040 (passes.o): Replace dependency on passes.def with one on
8043 * gen-pass-instances.awk: New.
8045 * passes.c (pass_manager::pass_manager): Use pass-instances.def
8046 rather than passes.def, updating local definition of NEXT_PASS
8047 macro to add an extra NUM parameter (currently unused).
8049 2013-07-30 David Malcolm <dmalcolm@redhat.com>
8051 * Makefile.in (PASS_MANAGER_H): New.
8052 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
8053 (passes.o): Likewise.
8054 (statistics.o): Likewise.
8055 (cgraphunit.o): Likewise.
8056 (context.o): Depend on PASS_MANAGER_H.
8058 * pass_manager.h: New.
8060 * cgraphunit.c (cgraph_add_new_function): Update for moves
8061 of globals to fields of pass_manager.
8062 (analyze_function): Likewise.
8063 (expand_function): Likewise.
8064 (ipa_passes): Likewise.
8065 (compile): Likewise.
8067 * context.c (context::context): New.
8068 * context.h (context::context): New.
8069 (context::get_passes): New.
8070 (context::passes_): New.
8072 * lto-cgraph.c (input_node): Update for moves of globals to
8073 fields of pass_manager.
8075 * passes.c (all_passes): Remove, in favor of a field of the
8076 same name within the new class pass_manager.
8077 (all_small_ipa_passes): Likewise.
8078 (all_lowering_passes): Likewise.
8079 (all_regular_ipa_passes): Likewise.
8080 (all_late_ipa_passes): Likewise.
8081 (all_lto_gen_passes): Likewise.
8082 (passes_by_id): Likewise.
8083 (passes_by_id_size): Likewise.
8084 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
8085 the new class pass_manager.
8086 (set_pass_for_id): Convert to...
8087 (pass_manager::set_pass_for_id): ...method.
8088 (get_pass_for_id): Convert to...
8089 (pass_manager::get_pass_for_id): ...method.
8090 (register_one_dump_file): Move body of implementation into...
8091 (pass_manager::register_one_dump_file): ...here.
8092 (register_dump_files_1): Convert to...
8093 (pass_manager::register_dump_files_1): ...method.
8094 (register_dump_files): Convert to...
8095 (pass_manager::register_dump_files): ...method.
8096 (create_pass_tab): Update for moves of globals to fields of
8098 (dump_passes): Move body of implementation into...
8099 (pass_manager::dump_passes): ...here.
8100 (register_pass): Move body of implementation into...
8101 (pass_manager::register_pass): ...here.
8102 (init_optimization_passes): Convert into...
8103 (pass_manager::pass_manager): ...constructor for new
8104 pass_manager class, and initialize the pass_lists array.
8105 (check_profile_consistency): Update for moves of globals to
8106 fields of pass_manager.
8107 (dump_profile_report): Move body of implementation into...
8108 (pass_manager::dump_profile_report): ...here.
8109 (ipa_write_summaries_1): Update for moves of pass lists from
8110 being globals to fields of pass_manager.
8111 (ipa_write_optimization_summaries): Likewise.
8112 (ipa_read_summaries): Likewise.
8113 (ipa_read_optimization_summaries): Likewise.
8114 (execute_all_ipa_stmt_fixups): Likewise.
8116 * statistics.c (statistics_fini): Update for moves of globals to
8117 fields of pass_manager.
8119 * toplev.c (general_init): Replace call to
8120 init_optimization_passes with construction of the pass_manager
8123 * tree-pass.h (all_passes): Remove, in favor of a field of the
8124 same name within the new class pass_manager.
8125 (all_small_ipa_passes): Likewise.
8126 (all_lowering_passes): Likewise.
8127 (all_regular_ipa_passes): Likewise.
8128 (all_lto_gen_passes): Likewise.
8129 (all_late_ipa_passes): Likewise.
8130 (passes_by_id): Likewise.
8131 (passes_by_id_size): Likewise.
8132 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
8133 the new class pass_manager.
8134 (get_pass_for_id): Remove.
8136 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
8138 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
8141 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
8143 * arm.md (mulhi3): New expand pattern.
8145 2013-07-30 Jan Hubicka <jh@suse.cz>
8146 Martin Liska <marxin.liska@gmail.com>
8148 * profile.c (compute_value_histograms): Do not ICE when
8149 there is mismatch only on some counters.
8151 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
8153 PR rtl-optimization/57637
8154 * function.c (move_insn_for_shrink_wrap): Also check the
8155 GEN set of the LIVE problem for the liveness analysis
8156 if it exists, otherwise give up.
8158 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
8160 PR tree-optimization/57993
8161 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
8162 replaced statement in the candidate table.
8163 (phi_add_costs): Return infinite cost when the hidden basis does
8164 not dominate all phis on which the candidate is dependent.
8165 (replace_one_candidate): Record replaced statement in the
8168 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
8170 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
8171 (ashlv2si3): New expander.
8172 (*ashlv2si3_i): New define_insn_and_split.
8173 * predicates.md (float_operation): Allow patterns with three
8176 PR rtl-optimization/58021
8177 * mode-switching.c (create_pre_exit): Always split off preceding
8178 insns if we are not at the basic block head.
8180 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
8182 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
8183 (UCLIBC_DYNAMIC_LINKER): New macro.
8184 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
8186 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
8187 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
8189 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
8190 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
8191 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
8192 for SF and DF modes. Use ieee_quad_format for TF mode.
8193 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
8194 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
8195 (mips_option_override): Handle `-mnan=legacy'.
8196 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
8197 `-mabs=2008' and `-mnan=2008'.
8198 (OPTION_DEFAULT_SPECS): Add "nan" default.
8199 (ASM_SPEC): Handle `-mnan='.
8200 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
8201 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
8202 comment accordingly.
8203 (neg<mode>2): Likewise.
8204 * config/mips/mips.opt (mabs, mnan): New options.
8205 * doc/install.texi (Configuration): Document `--with-nan=' option.
8206 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
8208 (MIPS Options): Document them.
8209 * config.gcc <mips*-*-*>: Handle `--with-nan='.
8210 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
8211 * configure: Regenerate.
8212 * config.in: Regenerate.
8214 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
8216 * config/i386/i386.md (float post-reload splitters): Do not check
8217 for subregs of SSE registers.
8219 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
8220 H.J. Lu <hongjiu.lu@intel.com>
8224 * config/i386/i386.md (post-reload splitter
8225 to avoid partial SSE reg dependency stalls): New pattern.
8227 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
8229 * config/s390/s390.md ("movcc"): Swap load and store instructions.
8231 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
8233 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
8234 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
8236 2013-07-26 Cary Coutant <ccoutant@google.com>
8238 * dwarf2out.c (die_checksum_ordered): Don't include template
8239 instantiations in signature.
8240 (is_template_parameter): New function.
8241 (is_template_instantiation): New function.
8242 (generate_skeleton_bottom_up): Don't include template instantiations
8244 (generate_skeleton): Likewise.
8245 (break_out_comdat_types): Move recursive call to break out nested
8247 (prune_unused_types_mark_generic_parms_dies): Call
8248 is_template_parameter.
8250 2013-07-26 Ian Bolton <ian.bolton@arm.com>
8252 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
8253 uses vector registers.
8254 * config/aarch64/iterators.md: Add attributes rtn and vas.
8256 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8257 Richard Earnshaw <richard.earnshaw@arm.com>
8259 * combine.c (simplify_comparison): Re-canonicalize operands
8261 * config/arm/arm.md (movcond_addsi): New splitter.
8263 2013-07-25 Sterling Augustine <saugustine@google.com>
8265 * dwarf2out.c (size_of_pubnames): Move code to...
8266 (include_pubname_in_output): ...here. New.
8267 (want_pubnames): Rearrange.
8268 (output_pubnames): Call include_pubname_in_output. Move assertion.
8270 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
8272 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
8274 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
8277 * doc/extend.texi: Remove obsolete builtins. Fix
8278 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
8280 2013-07-25 Jan Hubicka <jh@suse.cz>
8282 * cgraph.c (release_function_body): Break out from ...
8283 (cgraph_release_function_body): ... this one; also release DECL_RESULT
8285 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
8286 old_tree in the map.
8287 (create_specialized_node): Update.
8288 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
8290 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
8291 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
8292 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
8293 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
8296 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8298 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
8299 addsi3_carryin_alt2_<optab>): Correct output template.
8301 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8303 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
8304 Adjust for arm_restrict_it.
8305 Remove trailing whitespace.
8307 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
8309 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
8310 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
8312 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
8314 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
8316 PR rtl-optimization/57960
8317 * lra-constraints.c (process_alt_operands): Use the right mode
8318 when checking strict_low.
8320 2013-07-25 Jan Hubicka <jh@suse.cz>
8322 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
8323 * cgraph.c (cgraph_remove_node): Do not release function body
8324 when in cgraph streaming.
8325 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
8326 in other partitions are not considered reachable; fix handling of
8329 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
8331 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
8333 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
8338 * config/arm/arm.md ("*sibcall_insn): Replace use of
8339 Ss with US. Adjust output for v5 and v4t.
8340 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
8342 * config/arm/constraints.md ("Ss"): Rename to US.
8344 2013-07-25 Terry Guo <terry.guo@arm.com>
8346 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
8347 shift_add/shift_sub0/shift_sub1 RTXs.
8349 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
8350 Anton Blanchard <anton@au1.ibm.com>
8352 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
8353 (altivec_vpks<VI_char>ss): Likewise.
8354 (altivec_vpks<VI_char>us): Likewise.
8355 (altivec_vpku<VI_char>us): Likewise.
8356 (altivec_vpku<VI_char>um): Likewise.
8358 2013-07-24 David Malcolm <dmalcolm@redhat.com>
8360 Introduce context class.
8362 * Makefile.in (CONTEXT_H): New.
8363 (OBJS): Add context.o.
8364 (toplev.o): Add CONTEXT_H to dependencies.
8367 * toplev.c (general_init): Create the singleton gcc::context instance.
8373 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
8375 PR rtl-optimization/57968
8376 * mode-switching.c (create_pre_exit): Allow instructions that
8377 don't set a return register to need a non-exit mode.
8379 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
8380 Anton Blanchard <anton@au1.ibm.com>
8382 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
8383 operands to vperm for little endian.
8384 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
8385 of lvsl to create the control mask for a vperm for little endian.
8387 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8388 Anton Blanchard <anton@au1.ibm.com>
8390 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
8391 two operands for little-endian.
8393 2013-07-23 Steve Ellcey <sellcey@mips.com>
8395 * config/mips/mips.c (mips_case_values_threshold): New.
8396 (TARGET_CASE_VALUES_THRESHOLD): Define.
8398 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8399 Anton Blanchard <anton@au1.ibm.com>
8401 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
8402 selection of field for vector splat in little endian mode.
8404 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
8406 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
8407 expanders to rs6000.md.
8408 (ior<mode>3): Likewise.
8409 (and<mode>3): Likewise.
8410 (one_cmpl<mode>2): Likewise.
8411 (nor<mode>3): Likewise.
8412 (andc<mode>3): Likewise.
8413 (eqv<mode>3): Likewise.
8414 (nand<mode>3): Likewise.
8415 (orc<mode>3): Likewise.
8417 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
8420 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
8421 to split multi-word logical operations.
8422 (rs6000_split_logical_di): Likewise.
8423 (rs6000_split_logical): Likewise.
8425 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
8426 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
8427 and allow TImode operations in 32-bit.
8428 (vsx_and<mode>3_64bit): Likewise.
8429 (vsx_ior<mode>3_32bit): Likewise.
8430 (vsx_ior<mode>3_64bit): Likewise.
8431 (vsx_xor<mode>3_32bit): Likewise.
8432 (vsx_xor<mode>3_64bit): Likewise.
8433 (vsx_one_cmpl<mode>2_32bit): Likewise.
8434 (vsx_one_cmpl<mode>2_64bit): Likewise.
8435 (vsx_nor<mode>3_32bit): Likewise.
8436 (vsx_nor<mode>3_64bit): Likewise.
8437 (vsx_andc<mode>3_32bit): Likewise.
8438 (vsx_andc<mode>3_64bit): Likewise.
8439 (vsx_eqv<mode>3_32bit): Likewise.
8440 (vsx_eqv<mode>3_64bit): Likewise.
8441 (vsx_nand<mode>3_32bit): Likewise.
8442 (vsx_nand<mode>3_64bit): Likewise.
8443 (vsx_orc<mode>3_32bit): Likewise.
8444 (vsx_orc<mode>3_64bit): Likewise.
8446 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
8447 logical types in GPRs.
8449 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
8450 logical insns to rs6000.md, and allow TImode operations in
8452 (altivec_ior<mode>3): Likewise.
8453 (altivec_xor<mode>3): Likewise.
8454 (altivec_one_cmpl<mode>2): Likewise.
8455 (altivec_nor<mode>3): Likewise.
8456 (altivec_andc<mode>3): Likewise.
8458 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
8459 attributes for moving the 128-bit logical operations into
8461 (BOOL_REGS_OUTPUT): Likewise.
8462 (BOOL_REGS_OP1): Likewise.
8463 (BOOL_REGS_OP2): Likewise.
8464 (BOOL_REGS_UNARY): Likewise.
8465 (BOOL_REGS_AND_CR0): Likewise.
8466 (one_cmpl<mode>2): Add support for DI logical operations on
8467 32-bit, splitting the operations to 32-bit.
8471 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
8472 changes to combine the 32/64-bit code, allow logical operations on
8473 TI mode in 32-bit, and to use similar match_operator patterns like
8474 scalar mode uses. Combine the Altivec and VSX code for logical
8475 operations, and move it here.
8476 (ior<mode>3, 128-bit types): Likewise.
8477 (xor<mode>3, 128-bit types): Likewise.
8478 (one_cmpl<mode>3, 128-bit types): Likewise.
8479 (nor<mode>3, 128-bit types): Likewise.
8480 (andc<mode>3, 128-bit types): Likewise.
8481 (eqv<mode>3, 128-bit types): Likewise.
8482 (nand<mode>3, 128-bit types): Likewise.
8483 (orc<mode>3, 128-bit types): Likewise.
8484 (and<mode>3_internal): Likewise.
8485 (bool<mode>3_internal): Likewise.
8486 (boolc<mode>3_internal1): Likewise.
8487 (boolc<mode>3_internal2): Likewise.
8488 (boolcc<mode>3_internal1): Likewise.
8489 (boolcc<mode>3_internal2): Likewise.
8490 (eqv<mode>3_internal1): Likewise.
8491 (eqv<mode>3_internal2): Likewise.
8492 (one_cmpl1<mode>3_internal): Likewise.
8494 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
8496 * config/microblaze/microblaze.c (microblaze_expand_prologue):
8497 Rename flag_stack_usage to flag_stack_usage_info.
8499 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
8501 * config/microblaze/sync.md: New file.
8502 * config/microblaze/microblaze.md: Include sync.md
8503 * config/microblaze/microblaze.c: Add print_operand 'y'.
8504 * config/microblaze/constraints.md: Add memory_contraint
8505 'Q' which is a single register.
8507 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
8509 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
8511 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
8513 * reload.c (find_reloads): Exit loop once we find this operand
8514 cannot be reloaded somehow for this alternative.
8516 * reload.c (find_reloads): Exit loop once we find a hard register.
8518 * rtlanal.c (computed_jump_p): Exit loop once we find label
8521 * i386.c (ix86_pad_returns): Exit loop after setting replace.
8523 * cfgloopmanip.c (remove_path): Exit loop after setting
8526 * gensupport.c (subst_dup): Avoid loop if code is not
8527 MATCH_DUP nor MATCH_OP_DUP.
8529 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
8531 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
8533 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
8535 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
8536 true for SP_REGNUM if mode == ptr_mode.
8537 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
8538 with value R0_REGNUM + 31.
8540 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
8542 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
8543 pad pointer-typed argument downward.
8545 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
8547 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
8548 and __ILP32__ when the ILP32 model is in use.
8550 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
8552 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
8553 (aarch64_load_symref_appropriately): In the case of
8554 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
8555 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
8556 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
8557 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
8558 if 'mode' doesn't equal to 'ptr_mode'.
8559 (aarch64_output_mi_thunk): Add an assertion on the alignment of
8560 'vcall_offset'; change to call aarch64_emit_move differently depending
8561 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
8562 to calculate the upper bound of 'vcall_offset'.
8563 (aarch64_cannot_force_const_mem): Change to also return true if
8565 (aarch64_legitimize_reload_address): In the case of large
8566 displacements, add new local variable 'xmode' and an assertion
8567 based on it; change to use 'xmode' to generate the new rtx and
8569 (aarch64_asm_trampoline_template): Change to generate the template
8570 differently depending on TARGET_ILP32 or not; change to use
8571 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
8572 (aarch64_trampoline_size): Removed.
8573 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
8574 and replace immediate literals with it. Change to use 'ptr_mode'
8575 instead of 'DImode' and call convert_memory_address if the mode
8576 of 'fnaddr' doesn't equal to 'ptr_mode'.
8577 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
8579 (aarch64_elf_asm_destructor): Likewise.
8580 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
8581 on TARGET_ILP32 instead of aarch64_trampoline_size.
8582 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
8583 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
8584 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
8585 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
8586 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
8587 (storewb_pair<GPI:mode>_<P:mode>): ... this.
8588 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
8589 depending on the value of 'mode'.
8590 (add_losym_<mode>): New.
8591 (ldr_got_small_<mode>): New, based on ldr_got_small.
8592 (ldr_got_small): Remove.
8593 (ldr_got_small_sidi): New.
8594 * config/aarch64/iterators.md (P): New.
8595 (PTR): Change to 'ptr_mode' in the condition.
8597 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
8599 * config.gcc (aarch64*-*-*): Support --with-abi.
8600 (aarch64*-*-elf): Support --with-multilib-list.
8601 (aarch64*-*-linux*): Likewise.
8602 (supported_defaults): Add abi to aarch64*-*-*.
8603 * configure.ac: Mention AArch64 for --with-multilib-list.
8604 * configure: Re-generated.
8605 * config/aarch64/biarchilp32.h: New file.
8606 * config/aarch64/biarchlp64.h: New file.
8607 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
8609 (MULTILIB_DEFAULTS): Ditto.
8610 (DRIVER_SELF_SPECS): Ditto.
8611 (ASM_SPEC): Update to also substitute -mabi.
8612 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
8613 file whose name depends on -mabi= and -mbig-endian.
8614 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
8616 (POINTER_SIZE): New define.
8617 (POINTERS_EXTEND_UNSIGNED): Ditto.
8618 (enum aarch64_abi_type): New enumeration tag.
8619 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
8620 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
8621 (TARGET_ILP32): New define.
8622 * config/aarch64/aarch64.opt (mabi): New.
8624 (ilp32, lp64): New values for -mabi.
8625 * config/aarch64/t-aarch64 (comma): New define.
8626 (MULTILIB_OPTIONS): Ditto.
8627 (MULTILIB_DIRNAMES): Ditto.
8628 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
8629 * doc/invoke.texi: Document -mabi for AArch64.
8631 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
8633 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
8635 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
8636 Anton Blanchard <anton@au1.ibm.com>
8638 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
8639 endianness when selecting field to splat.
8641 2013-07-22 Eric Christopher <echristo@gmail.com>
8643 * dwarf2out.c (die_odr_checksum): New function to use
8644 CHECKSUM_ macros and ULEB128 for DIE tag.
8645 (generate_type_signature): Use.
8647 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
8649 * config.gcc (sparc*-*-*): Accept leon3 processor.
8650 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
8651 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
8652 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
8653 * config/sparc/sparc.opt (enum processor_type): Add leon3.
8654 (mfix-ut699): Adjust comment.
8655 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
8656 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
8657 (CPP_CPU_SPEC): Likewise.
8658 (ASM_CPU_SPEC): Likewise.
8659 * config/sparc/sparc.c (leon3_cost): New constant.
8660 (sparc_option_override): Add leon3 support.
8661 (mem_ref): New function.
8662 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
8663 (sparc_do_work_around_errata): Look into the instruction in the delay
8664 slot and adjust accordingly. Add fix for the data cache nullify issues
8665 of the UT699. Change insertion position for the NOP.
8666 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
8667 (leon3_load): New reservation.
8668 (leon_store): Bump latency to 2.
8669 (grfpu): New automaton.
8670 (grfpu_alu): New unit.
8671 (grfpu_ds): Likewise.
8672 (leon_fp_alu): Adjust.
8673 (leon_fp_mult): Delete.
8674 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
8675 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
8676 * config/sparc/sparc.md (cpu): Add leon3.
8677 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
8679 (atomic_test_and_set): Likewise.
8682 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
8684 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
8685 default for R5900 targets.
8686 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
8687 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
8688 * config/mips/mips.c (mips_option_override): Report an error for
8689 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
8690 for -march=r5900 -mhard-float.
8692 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
8694 * df-problems.c (can_move_insns_across): Exit loop once we
8695 find a non-fixed, non-global register.
8697 * ipa-pure-const.c (propagate_nothrow): Exit loop after
8700 * omega.c (omega_eliminate_red): Break after setting red_found.
8701 (omega_problem_has_red_equations): Similarly after setting found.
8702 (omega_query_variable): Similarly after setting coupled.
8704 2013-07-22 Marek Polacek <polacek@redhat.com>
8706 * gimplify.c: Don't include gimple.h twice.
8708 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8710 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
8711 instead of TARGET_THUMB1.
8712 (Pz): New constraint.
8713 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
8715 (compare_negsi_si): Likewise.
8716 (compare_addsi2_op0): Likewise.
8717 (compare_addsi2_op1): Likewise.
8718 (addsi3_carryin_<optab>): Likewise.
8719 (addsi3_carryin_alt2_<optab>): Likewise.
8720 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
8721 for arm_restrict_it.
8722 (subsi3_carryin): Likewise.
8723 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
8724 (minmax_arithsi): Disable for arm_restrict_it.
8725 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
8726 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
8727 (satsi_<SAT:code>_shift): Likewise.
8728 (arm_shiftsi3): Add alternative for 16-bit encoding.
8729 (arm32_movhf): Disable for arm_restrict_it.
8730 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
8731 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
8733 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
8735 * config/arm/arm.md (attribute "insn"): Delete.
8736 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
8737 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
8738 (not_shiftsi): Update for attribute change.
8739 (not_shiftsi_compare0): Likewise.
8740 (not_shiftsi_compare0_scratch): Likewise.
8741 (arm_one_cmplsi2): Likewise.
8742 (thumb1_one_cmplsi2): Likewise.
8743 (notsi_compare0): Likewise.
8744 (notsi_compare0_scratch): Likewise.
8745 (thumb1_movdi_insn): Likewise.
8746 (arm_movsi_insn): Likewise.
8747 (movhi_insn_arch4): Likewise.
8748 (movhi_bytes): Likewise.
8749 (arm_movqi_insn): Likewise.
8750 (thumb1_movqi_insn): Likewise.
8751 (arm32_movhf): Likewise.
8752 (thumb1_movhf): Likewise.
8753 (arm_movsf_soft_insn): Likewise.
8754 (thumb1_movsf_insn): Likewise.
8755 (thumb_movdf_insn): Likewise.
8756 (movsicc_insn): Likewise.
8757 (movsfcc_soft_insn): Likewise.
8758 (and_scc): Likewise.
8759 (cond_move): Likewise.
8760 (if_move_not): Likewise.
8761 (if_not_move): Likewise.
8762 (if_shift_move): Likewise.
8763 (if_move_shift): Likewise.
8764 (if_shift_shift): Likewise.
8765 (if_not_arith): Likewise.
8766 (if_arith_not): Likewise.
8767 (cond_move_not): Likewise.
8768 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
8769 (neon_mov<mode>): Likewise.
8770 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
8771 (thumb2_movsi_vfp): Likewise.
8772 (movsf_vfp): Likewise.
8773 (thumb2_movsf_vfp): Likewise.
8774 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
8776 (cortexa7_older_only): Likewise.
8777 (cortexa7_younger): Likewise.
8778 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
8779 (1020alu_shift_op): Likewise.
8780 (1020alu_shift_reg_op): Likewise.
8781 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
8782 (alu_shift_op): Likewise.
8783 (alu_shift_reg_op): Likewise.
8784 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
8785 (11_alu_shift_op): Likewise.
8786 (11_alu_shift_reg_op): Likewise.
8787 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
8788 (9_alu_shift_reg_op): Likewise.
8789 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
8791 (cortex_a15_alu_shift): Likewise.
8792 (cortex_a15_alu_shift_reg): Likewise.
8793 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
8794 (cortex_a5_alu_shift): Likewise.
8795 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
8797 (cortex_a53_alu_shift): Likewise.
8798 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
8800 (cortex_a7_alu_reg): Likewise.
8801 (cortex_a7_alu_shift): Likewise.
8802 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
8803 (cortex_a8_alu_shift): Likewise.
8804 (cortex_a8_alu_shift_reg): Likewise.
8805 (cortex_a8_mov): Likewise.
8806 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
8807 (cortex_a9_dp_shift): Likewise.
8808 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
8809 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
8810 (cortex_r4_mov): Likewise.
8811 (cortex_r4_alu_shift): Likewise.
8812 (cortex_r4_alu_shift_reg): Likewise.
8813 * config/arm/fa526.md (526_alu_op): Update for attribute change.
8814 (526_alu_shift_op): Likewise.
8815 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
8816 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
8817 (626te_alu_shift_op): Likewise.
8818 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
8819 (726te_alu_op): Likewise.
8820 (726te_alu_shift_op): Likewise.
8821 (726te_alu_shift_reg_op): Likewise.
8822 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
8823 (mp626_alu_shift_op): Likewise.
8824 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
8825 (pj4_alu_e1_conds): Likewise.
8826 (pj4_alu): Likewise.
8827 (pj4_alu_conds): Likewise.
8828 (pj4_shift): Likewise.
8829 (pj4_shift_conds): Likewise.
8830 (pj4_alu_shift): Likewise.
8831 (pj4_alu_shift_conds): Likewise.
8833 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8835 * config/arm/predicates.md (shiftable_operator_strict_it):
8837 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
8838 Disable cond_exec version for arm_restrict_it.
8839 (thumb2_smaxsi3): Convert to generate cond_exec.
8840 (thumb2_sminsi3): Likewise.
8841 (thumb32_umaxsi3): Likewise.
8842 (thumb2_uminsi3): Likewise.
8843 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
8844 (thumb2_neg_abssi2): Likewise.
8845 (thumb2_mov_scc): Add alternative for 16-bit encoding.
8846 (thumb2_movsicc_insn): Adjust alternatives.
8847 (thumb2_mov_negscc): Disable for arm_restrict_it.
8848 (thumb2_mov_negscc_strict_it): New pattern.
8849 (thumb2_mov_notscc_strict_it): New pattern.
8850 (thumb2_mov_notscc): Disable for arm_restrict_it.
8851 (thumb2_ior_scc): Likewise.
8852 (thumb2_ior_scc_strict_it): New pattern.
8853 (thumb2_cond_move): Adjust for arm_restrict_it.
8854 (thumb2_cond_arith): Disable for arm_restrict_it.
8855 (thumb2_cond_arith_strict_it): New pattern.
8856 (thumb2_cond_sub): Adjust for arm_restrict_it.
8857 (thumb2_movcond): Likewise.
8858 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
8859 (thumb2_zero_extendhisi2_v6): Likewise.
8860 (thumb2_zero_extendqisi2_v6): Likewise.
8861 (orsi_notsi_si): Likewise.
8862 (orsi_not_shiftsi_si): Likewise.
8864 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
8866 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
8867 instruction sequence is 1 byte shorter.
8869 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
8871 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
8872 it is not needed after split.
8874 2013-07-20 Iain Sandoe <iain@codesourcery.com>
8877 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
8878 second label for nonlocal goto receivers. Don't output pic base labels
8879 unless we're producing PIC; mark that action unreachable().
8880 (ix86_save_reg): If the function contains a nonlocal label, save the
8882 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
8883 * config/darwin.c (emitted_pic_label_num): New GTY.
8884 (update_pic_label_number_if_needed): New.
8885 (machopic_output_function_base_name): Adjust for nonlocal receiver
8887 (machopic_should_output_picbase_label): New.
8888 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
8889 (nonlocal_goto_receiver): New insn and split.
8891 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
8893 * config/aarch64/aarch64-builtins.c
8894 (aarch64_fold_builtin): Fold abs in all modes.
8895 * config/aarch64/aarch64-simd-builtins.def
8896 (abs): Enable for all modes.
8897 * config/aarch64/arm_neon.h
8898 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
8899 (vabs_f64): Add missing intrinsic.
8901 2013-07-19 Ian Bolton <ian.bolton@arm.com>
8903 * config/aarch64/arm_neon.h (vabs_s64): New function
8905 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
8908 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
8909 * config/avr/avr.md (adjust_len): Add `round'.
8910 * config/avr/avr-protos.h (avr_out_round): New prototype.
8911 (avr_out_plus): Add `out_label' argument.
8912 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
8913 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
8914 Handle the case where `insn' is just a pattern.
8915 (avr_out_bitop): Handle the case where `insn' is just a pattern.
8916 (avr_out_round): New function.
8917 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
8919 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
8921 * config/microblaze/microblaze.c (microblaze_expand_prologue):
8922 Add check for flag_stack_usage to handle -fstack-usage support
8924 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
8926 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
8927 interaction for new Power8 flags and VSX.
8929 2013-07-18 Sriraman Tallam <tmsriram@google.com>
8932 * tree-inline.c (expand_call_inline): Emit errors during
8933 early_inlining only if optimization is not turned on.
8935 2013-07-18 David Malcolm <dmalcolm@redhat.com>
8939 * passes.c (init_optimization_passes): Move the construction of
8940 the pass hierarchy into a new passes.def file.
8942 * Makefile.in (passes.o): Add dependency on passes.def.
8944 2013-07-18 David Malcolm <dmalcolm@redhat.com>
8946 * passes.c (init_optimization_passes): Introduce macros for
8947 constructing the tree of passes (INSERT_PASSES_AFTER,
8948 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
8949 TERMINATE_PASS_LIST).
8951 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
8952 Wei Mi <wmi@google.com>
8954 PR rtl-optimization/57878
8955 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
8957 (reload_pseudo_compare_func): Check nregs first for reload
8960 2013-07-18 David Malcolm <dmalcolm@redhat.com>
8962 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
8964 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
8966 * read-rtl.c (validate_const_int): Once an invalid character is
8967 seen, quit the loop.
8969 * gengtype.c (write_roots): Similarly once we find the "deletable"
8970 or "if_marked" option.
8972 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
8974 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
8975 "xtab" and "sat". Move value "clz" from here to ...
8976 (attriubte "type"): ... here.
8977 (satsi_<SAT:code>): Delete "insn" attribute.
8978 (satsi_<SAT:code>_shift): Likewise.
8979 (arm_zero_extendqisi2addsi): Likewise.
8980 (arm_extendqisi2addsi): Likewise.
8981 (clzsi2): Update for attribute changes.
8982 (rbitsi2): Likewise.
8983 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
8985 (arm_usatsihi): Likewise.
8986 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
8988 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
8990 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
8991 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
8992 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
8993 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
8994 in alphabetical order.
8995 (attribute "core_cycles"): Update for attribute changes.
8996 (arm_addsi3): Likewise.
8997 (addsi3_compare0): Likewise.
8998 (addsi3_compare0_scratch): Likewise.
8999 (addsi3_compare_op1): Likewise.
9000 (addsi3_compare_op2): Likewise.
9001 (compare_addsi2_op0): Likewise.
9002 (compare_addsi2_op1): Likewise.
9003 (addsi3_carryin_shift_<optab>): Likewise.
9004 (subsi3_carryin_shift): Likewise.
9005 (rsbsi3_carryin_shift): Likewise.
9006 (arm_subsi3_insn): Likewise.
9007 (subsi3_compare0): Likewise.
9008 (subsi3_compare): Likewise.
9009 (arm_andsi3_insn): Likewise.
9010 (thumb1_andsi3_insn): Likewise.
9011 (andsi3_compare0): Likewise.
9012 (andsi3_compare0_scratch): Likewise.
9013 (zeroextractsi_compare0_scratch
9014 (andsi_not_shiftsi_si): Likewise.
9015 (iorsi3_insn): Likewise.
9016 (iorsi3_compare0): Likewise.
9017 (iorsi3_compare0_scratch): Likewise.
9018 (arm_xorsi3): Likewise.
9019 (thumb1_xorsi3_insn): Likewise.
9020 (xorsi3_compare0): Likewise.
9021 (xorsi3_compare0_scratch): Likewise.
9022 (satsi_<SAT:code>_shift): Likewise.
9024 (arm_shiftsi3): Likewise.
9025 (shiftsi3_compare0): Likewise.
9026 (not_shiftsi): Likewise.
9027 (not_shiftsi_compare0): Likewise.
9028 (not_shiftsi_compare0_scratch): Likewise.
9029 (arm_one_cmplsi2): Likewise.
9030 (thumb_one_complsi2): Likewise.
9031 (notsi_compare0): Likewise.
9032 (notsi_compare0_scratch): Likewise.
9033 (thumb1_zero_extendhisi2): Likewise.
9034 (arm_zero_extendhisi2): Likewise.
9035 (arm_zero_extendhisi2_v6): Likewise.
9036 (arm_zero_extendhisi2addsi): Likewise.
9037 (thumb1_zero_extendqisi2): Likewise.
9038 (thumb1_zero_extendqisi2_v6): Likewise.
9039 (arm_zero_extendqisi2): Likewise.
9040 (arm_zero_extendqisi2_v6): Likewise.
9041 (arm_zero_extendqisi2addsi): Likewise.
9042 (thumb1_extendhisi2): Likewise.
9043 (arm_extendhisi2): Likewise.
9044 (arm_extendhisi2_v6): Likewise.
9045 (arm_extendqisi): Likewise.
9046 (arm_extendqisi_v6): Likewise.
9047 (arm_extendqisi2addsi): Likewise.
9048 (thumb1_extendqisi2): Likewise.
9049 (thumb1_movdi_insn): Likewise.
9050 (arm_movsi_insn): Likewise.
9051 (movsi_compare0): Likewise.
9052 (movhi_insn_arch4): Likewise.
9053 (movhi_bytes): Likewise.
9054 (arm_movqi_insn): Likewise.
9055 (thumb1_movqi_insn): Likewise.
9056 (arm32_movhf): Likewise.
9057 (thumb1_movhf): Likewise.
9058 (arm_movsf_soft_insn): Likewise.
9059 (thumb1_movsf_insn): Likewise.
9060 (movdf_soft_insn): Likewise.
9061 (thumb_movdf_insn): Likewise.
9062 (arm_cmpsi_insn): Likewise.
9063 (cmpsi_shiftsi): Likewise.
9064 (cmpsi_shiftsi_swp): Likewise.
9065 (arm_cmpsi_negshiftsi_si): Likewise.
9066 (movsicc_insn): Likewise.
9067 (movsfcc_soft_insn): Likewise.
9068 (arith_shiftsi): Likewise.
9069 (arith_shiftsi_compare0
9070 (arith_shiftsi_compare0_scratch
9071 (sub_shiftsi): Likewise.
9072 (sub_shiftsi_compare0
9073 (sub_shiftsi_compare0_scratch
9074 (and_scc): Likewise.
9075 (cond_move): Likewise.
9076 (if_plus_move): Likewise.
9077 (if_move_plus): Likewise.
9078 (if_move_not): Likewise.
9079 (if_not_move): Likewise.
9080 (if_shift_move): Likewise.
9081 (if_move_shift): Likewise.
9082 (if_shift_shift): Likewise.
9083 (if_not_arith): Likewise.
9084 (if_arith_not): Likewise.
9085 (cond_move_not): Likewise.
9086 (thumb1_ashlsi3): Set type attribute.
9087 (thumb1_ashrsi3): Likewise.
9088 (thumb1_lshrsi3): Likewise.
9089 (thumb1_rotrsi3): Likewise.
9090 (shiftsi3_compare0_scratch): Likewise.
9091 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
9092 (neon_mov<mode>): Likewise.
9093 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
9095 (thumb2_movsi_insn): Likewise.
9096 (thumb2_cmpsi_neg_shiftsi): Likewise.
9097 (thumb2_extendqisi_v6): Likewise.
9098 (thumb2_zero_extendhisi2_v6): Likewise.
9099 (thumb2_zero_extendqisi2_v6): Likewise.
9100 (thumb2_shiftsi3_short): Likewise.
9101 (thumb2_addsi3_compare0_scratch): Likewise.
9102 (orsi_not_shiftsi_si): Likewise.
9103 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
9104 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
9106 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
9107 (1020alu_shift_op): Likewise.
9108 (1020alu_shift_reg_op): Likewise.
9109 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
9110 (alu_shift_op): Likewise.
9111 (alu_shift_reg_op): Likewise.
9112 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
9113 (11_alu_shift_op): Likewise.
9114 (11_alu_shift_reg_op): Likewise.
9115 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
9116 (9_alu_shift_reg_op): Likewise.
9117 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
9119 (cortex_a15_alu_shift): Likewise.
9120 (cortex_a15_alu_shift_reg): Likewise.
9121 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
9123 (cortex_a5_alu_shift): Likewise.
9124 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
9126 (cortex_a53_alu_shift): Likewise.
9127 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
9129 (cortex_a7_alu_reg): Likewise.
9130 (cortex_a7_alu_shift): Likewise.
9131 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
9133 (cortex_a8_alu_shift): Likewise.
9134 (cortex_a8_alu_shift_reg): Likewise.
9135 (cortex_a8_mov): Likewise.
9136 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
9137 (cortex_a9_dp_shift): Likewise.
9138 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
9140 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
9142 (cortex_r4_mov): Likewise.
9143 (cortex_r4_alu_shift): Likewise.
9144 (cortex_r4_alu_shift_reg): Likewise.
9145 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
9146 (526_alu_shift_op): Likewise.
9147 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
9148 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
9149 (626te_alu_shift_op): Likewise.
9150 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
9151 (726te_alu_op): Likewise.
9152 (726te_alu_shift_op): Likewise.
9153 (726te_alu_shift_reg_op): Likewise.
9154 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
9155 (mp626_alu_shift_op): Likewise.
9156 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
9157 (pj4_alu_e1_conds): Likewise.
9158 (pj4_alu): Likewise.
9159 (pj4_alu_conds): Likewise.
9160 (pj4_shift): Likewise.
9161 (pj4_shift_conds): Likewise.
9162 (pj4_alu_shift): Likewise.
9163 (pj4_alu_shift_conds): Likewise.
9164 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
9166 (cortexa7_older_only): Likewise.
9167 (cortexa7_younger): Likewise.
9169 2013-07-18 David Malcolm <dmalcolm@redhat.com>
9171 * ipa-pure-const.c (generate_summary): Rename to...
9172 (pure_const_generate_summary): ... this.
9174 2013-07-17 Iain Sandoe <iain@codesourcery.com>
9176 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
9178 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
9181 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
9184 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9186 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
9187 enabled without -march=zEC12.
9188 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
9191 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
9193 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
9194 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
9195 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
9196 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
9197 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
9198 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
9199 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
9201 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
9202 and ISA_HAS_NMADD3_NMSUB3.
9203 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
9204 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
9205 (nmsub4<mode>, nmsub3<mode>): Likewise.
9206 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
9208 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
9210 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
9211 TARGET_MIPS5400 checking.
9213 2013-07-16 Jakub Jelinek <jakub@redhat.com>
9214 Peter Bergner <bergner@vnet.ibm.com>
9216 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
9217 registers in the comment.
9218 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
9219 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
9220 rather than FIRST_PSEUDO_REGISTERS.
9222 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
9224 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
9225 enable extra ISA flags with TARGET_HTM.
9227 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
9229 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
9232 2013-07-15 Cong Hou <congh@google.com>
9234 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
9235 in compare function for sorting.
9237 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
9239 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
9240 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
9241 * config/rs6000/rs6000.opt: Add -mhtm option.
9242 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
9243 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
9244 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
9245 __HTM__ if the HTM instructions are available.
9246 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
9247 htm_spr_reg_operand): New define_predicates.
9248 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
9249 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
9251 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
9252 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
9253 HTM builtin functions.
9254 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
9255 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
9256 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
9257 (rs6000_builtin_mask_calculate): Likewise.
9258 (rs6000_option_override_internal): Likewise.
9259 (bdesc_htm): Add new HTM builtin support.
9260 (htm_spr_num): New function.
9261 (htm_spr_regno): Likewise.
9262 (rs6000_htm_spr_icode): Likewise.
9263 (htm_expand_builtin): Likewise.
9264 (htm_init_builtins): Likewise.
9265 (rs6000_expand_builtin): Add support for HTM builtin functions.
9266 (rs6000_init_builtins): Likewise.
9267 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
9269 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
9270 (TARGET_HTM, MASK_HTM): Define macros.
9271 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
9272 (FIXED_REGISTERS): Likewise.
9273 (CALL_USED_REGISTERS): Likewise.
9274 (CALL_REALLY_USED_REGISTERS): Likewise.
9275 (REG_ALLOC_ORDER): Likewise.
9276 (enum reg_class): Likewise.
9277 (REG_CLASS_NAMES): Likewise.
9278 (REG_CLASS_CONTENTS): Likewise.
9279 (REGISTER_NAMES): Likewise.
9280 (ADDITIONAL_REGISTER_NAMES): Likewise.
9281 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
9282 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
9283 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
9284 * config/rs6000/htm.md: New file.
9285 * config/rs6000/htmintrin.h: New file.
9286 * config/rs6000/htmxlintrin.h: New file.
9288 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
9290 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
9291 Define SYMBOL_TINY_GOT, update comment.
9292 * config/aarch64/aarch64.c
9293 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
9294 (aarch64_expand_mov_immediate): Likewise.
9295 (aarch64_print_operand): Likewise.
9296 (aarch64_classify_symbol): Likewise.
9297 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
9298 (ldr_got_tiny): Define.
9300 2013-07-13 Tobias Grosser <tobias@grosser.es>
9302 PR tree-optimization/54094
9303 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
9304 scheduling dimension for the parallelism check from the polyhedral
9305 information in the AST.
9306 * graphite-dependences.c (carries_deps): Do not assume the schedule is
9309 2013-07-13 Jason Merrill <jason@redhat.com>
9311 * print-tree.c (debug_vec_tree): Use debug_raw.
9312 (debug_raw (vec<tree, va_gc> &)): New.
9313 (debug_raw (vec<tree, va_gc> *)): New.
9314 * tree.h: Declare them.
9316 2013-07-13 Bin Cheng <bin.cheng@arm.com>
9318 * ifcvt.c (ifcvt_after_combine): New static variable.
9319 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
9321 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
9322 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
9323 rest_of_handle_if_after_reload): Pass new argument for if_convert.
9325 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
9327 * config/mips/mips.c (mips_expand_call): Remove empty statement.
9329 2013-07-12 Michael Matz <matz@suse.de>
9332 * convert.c (convert_to_real): Reject non-float inner types.
9334 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
9336 * config/aarch64/aarch64-protos.h
9337 (aarch64_simd_immediate_valid_for_move): Remove.
9338 * config/aarch64/aarch64.c (simd_immediate_info): New member.
9339 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
9341 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
9343 2013-07-11 Steve Ellcey <sellcey@mips.com>
9345 * config/mips/mips.c (mips_conditional_register_usage): Do not
9346 use t[0-7] registers in MIPS16 mode when optimizing for size.
9348 2013-07-11 Sriraman Tallam <tmsriram@google.com>
9350 * config/i386/i386.c (dispatch_function_versions): Fix array
9351 indexing of function_version_info to match actual_versions.
9353 2013-07-11 Teresa Johnson <tejohnson@google.com>
9355 * vec.h (struct va_gc): Move release out-of-line.
9356 (va_gc::release): Call ggc_free on released vec.
9358 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
9360 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
9361 Require GOT register as additional operand in UNSPEC.
9362 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
9363 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
9364 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
9365 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
9366 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
9367 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
9368 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
9370 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
9373 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
9374 name seen by assembler/linker rather if available.
9376 2013-07-11 Andreas Schwab <schwab@suse.de>
9378 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
9380 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
9382 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
9384 2013-07-10 Joseph Myers <joseph@codesourcery.com>
9386 * doc/tm.texi.in: Move hook documentation to ....
9387 * target.def: ... here.
9389 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
9391 * doc/tm.texi: Regenerate.
9393 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
9396 * doc/invoke.texi: Document -Wconditionally-supported.
9398 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
9401 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
9404 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
9407 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
9408 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
9409 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
9410 Remove duplicate devices.
9411 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
9412 * config/avr/t-multilib: Regenerate.
9413 * config/avr/avr-tables.opt: Regenerate.
9414 * doc/avr-mmcu.texi: Regenerate.
9416 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
9419 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
9421 2013-07-10 Graham Stott <graham.stott@btinternet.com>
9423 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
9424 the cost of MULT when optimizing for size.
9426 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9428 * config/cr16/cr16-protos.h: Don't include target.h.
9430 2013-07-09 Joseph Myers <joseph@codesourcery.com>
9432 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
9433 adjust register size for TDmode and TFmode for VSX registers.
9435 2013-07-08 Kai Tietz <ktietz@redhat.com>
9438 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
9439 hook_bool_const_tree_true.
9441 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9443 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
9444 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
9445 * config/s390/s390.md: Define FPR*_REGNUM constants.
9446 Fix FPR2_REGNUM constant (18 -> 17).
9447 ("*trunc<BFP:mode><DFP_ALL:mode>2")
9448 ("*trunc<DFP_ALL:mode><BFP:mode>2")
9449 ("trunc<BFP:mode><DFP_ALL:mode>2")
9450 ("trunc<DFP_ALL:mode><BFP:mode>2")
9451 ("*extend<BFP:mode><DFP_ALL:mode>2")
9452 ("*extend<DFP_ALL:mode><BFP:mode>2")
9453 ("extend<BFP:mode><DFP_ALL:mode>2")
9454 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
9457 2013-07-08 Graham Stott <graham.stott@btinternet.com>
9459 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
9461 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9463 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
9464 and cfun_fpr_bit_p to cfun_fpr_save_p.
9465 (s390_frame_area, s390_register_info, s390_frame_info)
9466 (s390_emit_prologue, s390_emit_epilogue)
9467 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
9469 * config/s390/s390.h: Define *_REGNUM macros for floating point
9472 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
9474 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
9476 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
9478 PR rtl-optimization/57786
9479 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
9480 and break out of the loop when it is set to false.
9482 2013-07-08 Jakub Jelinek <jakub@redhat.com>
9485 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
9486 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
9487 (const_int 63)) 0)).
9488 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
9489 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
9490 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
9492 PR rtl-optimization/57829
9493 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
9494 mask bits outside of mode are just sign-extension from mode to HWI.
9496 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
9498 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
9499 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
9500 adjust_address instead of change_address to keep info about alignment.
9501 (emit_strmov): Remove.
9502 (emit_memmov): New function.
9503 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
9504 (expand_movmem_epilogue): Likewise and return updated rtx for
9506 (expand_constant_movmem_prologue): Likewise and return updated rtx for
9507 destination and source.
9508 (decide_alignment): Refactor, handle vector_loop.
9509 (ix86_expand_movmem): Likewise.
9510 (ix86_expand_setmem): Likewise.
9511 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
9513 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
9515 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
9516 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
9518 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
9520 * config/i386/sse.md (sse_movlhps): Change alternative 3
9521 of operand 2 to "m".
9523 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
9526 * config/i386/sse.md (iptr): New mode attribute.
9527 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
9528 (<sse>_vm<plusminus_insn><mode>3): Ditto.
9529 (<sse>_vmmul<mode>3): Ditto.
9530 (<sse>_vmdiv<mode>3): Ditto.
9531 (sse_vmrcpv4sf2): Ditto.
9532 (<sse>_vmsqrt<mode>2): Ditto.
9533 (sse_vmrsqrtv4sf2): Ditto.
9534 (<sse>_vm<code><mode>3): Ditto.
9535 (avx_vmcmp<mode>3): Ditto.
9536 (<sse>_vmmaskcmp<mode>3): Ditto.
9537 (<sse>_comi): Ditto.
9538 (<sse>_ucomi): Ditto.
9539 (*xop_vmfrcz_<mode>): Ditto.
9540 (*fmai_fmadd_<mode>): Ditto.
9541 (*fmai_fmsub_<mode>): Ditto.
9542 (*fmai_fnmadd_<mode>): Ditto.
9543 (*fmai_fnmsub_<mode>): Ditto.
9544 (*fma4i_vmfmadd_<mode>): Ditto.
9545 (*fma4i_vmfmsub_<mode>): Ditto.
9546 (*fma4i_vmfnmadd_<mode>): Ditto.
9547 (*fma4i_vmfnmsub_<mode>): Ditto.
9548 (*xop_vmfrcz_<mode>): Ditto.
9549 (sse_cvtps2pi): Ditto.
9550 (sse_cvttps2pi): Ditto.
9551 (sse_cvtss2si): Ditto.
9552 (sse_cvtss2si_2): Ditto.
9553 (sse_cvtss2siq_2): Ditto.
9554 (sse_cvttss2si): Ditto.
9555 (sse_cvttss2siq): Ditto.
9556 (sse_cvtsd2si): Ditto.
9557 (sse_cvtsd2si_2): Ditto.
9558 (sse_cvtsd2siq_2): Ditto.
9559 (sse_cvttsd2si): Ditto.
9560 (sse_cvttsd2siq): Ditto.
9561 (sse_cvtsd2ss): Ditto.
9562 (sse_cvtss2sd): Ditto.
9563 (avx2_pbroadcast<mode>): Ditto.
9564 (avx2_pbroadcast<mode>_1): Ditto.
9565 (*avx_vperm_broadcast_v4sf): Ditto.
9567 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
9568 (sse_movlhps): Ditto.
9569 (sse_storehps): Ditto.
9570 (sse_loadhps): Ditto.
9571 (sse_storelps): Ditto.
9572 (sse_loadlps): Ditto.
9573 (*vec_concatv4sf): Ditto.
9574 (*vec_interleave_highv2df): Ditto.
9575 (*vec_interleave_lowv2df): Ditto.
9576 (*vec_extractv2df_1_sse): Ditto.
9577 (*vec_extractv2df_0_sse): Ditto.
9578 (sse2_storelpd): Ditto.
9579 (sse2_loadlpd): Ditto.
9580 (sse2_movsd): Ditto.
9581 (*vec_concatv4si): Ditto.
9582 (vec_concatv2di): Ditto.
9584 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
9585 for Intel asm dialect.
9586 (mmx_punpcklwd): Ditto.
9587 (mmx_punpckldq): Ditto.
9589 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
9590 for intel assembler dialect.
9592 2013-07-06 Jakub Jelinek <jakub@redhat.com>
9595 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
9596 for BUILT_IN_C{LZ,LRSB}*.
9597 * tree.h (CASE_INT_FN): Add FN##IMAX case.
9598 * tree-vrp.c (extract_range_basic): Handle
9599 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
9600 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
9601 fall thru to code calling set_value*.
9602 * builtins.c (expand_builtin): Remove *IMAX cases.
9603 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
9604 if width is bigger than 2*HWI.
9606 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
9608 PR rtl-optimization/55342
9609 * lra-int.h (lra_subreg_reload_pseudos): New.
9610 * lra.c: Add undoing optional reloads to the block diagram.
9611 (lra_subreg_reload_pseudos): New.
9612 (lra_optional_reload_pseudos): Change comments.
9613 (lra): Init and clear lra_subreg_reload_pseudos. Clear
9614 lra_optional_reload_pseudos after undo transformations.
9615 * lra-assigns.c (pseudo_prefix_title): New.
9616 (lra_setup_reg_renumber): Use it.
9617 (spill_for): Ditto. Check subreg reload pseudos too.
9618 (assign_by_spills): Consider subreg reload pseudos too.
9619 * lra-constraints.c (simplify_operand_subreg): Use
9620 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
9621 (curr_insn_transform): Recognize and do optional reloads.
9622 (undo_optional_reloads): New.
9623 (lra_undo_inheritance): Call undo_optional_reloads.
9625 2013-07-05 Thomas Quinot <quinot@adacore.com>
9627 * tree-complex.c (expand_complex_operations_1): Fix typo.
9629 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
9631 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
9632 (tune_params): New member 'const vec_costs'.
9633 * config/aarch64/aarch64.c (generic_vector_cost): New.
9634 (generic_tunings): New member 'generic_vector_cost'.
9635 (aarch64_builtin_vectorization_cost): New.
9636 (aarch64_add_stmt_cost): New.
9637 (TARGET_VECTORIZE_ADD_STMT_COST): New.
9638 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
9640 2013-07-03 Jakub Jelinek <jakub@redhat.com>
9643 * config/i386/predicates.md (vsib_address_operand): Disallow
9644 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
9646 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
9649 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
9650 expand_builtin_setjmp_receiver.
9651 (expand_label): Adjust, call expand_builtin_setjmp_receiver
9652 with NULL for the label parameter.
9653 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
9654 the frame-pointer. Adjust comments.
9655 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
9656 only if LABEL is non-NULL.
9658 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
9660 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
9661 (ARM_ABI_AAPCS64): Ditto.
9663 (ARM_DEFAULT_ABI): Ditto.
9665 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
9667 * config/aarch64/aarch64-builtins.c
9668 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
9669 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
9671 * config/aarch64/aarch64-simd.md
9672 (aarch64_ld1<VALL:mode>): New.
9673 (aarch64_st1<VALL:mode>): Likewise.
9674 * config/aarch64/arm_neon.h
9675 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
9677 2013-07-02 Sriraman Tallam <tmsriram@google.com>
9679 * config/i386/i386.c (gate_insert_vzeroupper): Check if
9681 (ix86_option_override_internal):Turn on all -mavx target flags by
9682 default as they are dependent on AVX anyway.
9684 2013-07-02 Cary Coutant <ccoutant@google.com>
9686 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
9688 (loc_checksum_ordered): Likewise.
9689 (hash_loc_operands): Remove inline keyword.
9691 2013-07-02 Jakub Jelinek <jakub@redhat.com>
9693 PR tree-optimization/57741
9694 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
9695 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
9696 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
9697 Allow REAL_CST step_exprs if flag_associative_math.
9698 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
9700 2013-07-02 Ian Bolton <ian.bolton@arm.com>
9702 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
9704 2013-07-02 Ian Bolton <ian.bolton@arm.com>
9706 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
9708 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9710 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
9712 (iorsi3_insn): Likewise.
9713 (arm_xorsi3): Likewise.
9715 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
9717 * arm.md (attribute "wtype"): Delete. Move attribute values from here
9719 (attribute "type"): ... here, and prefix with "wmmx_".
9720 (attribute "core_cycles"): Update for attribute changes.
9721 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
9722 (tbcstv4hi): Likewise.
9723 (tbcstv2si): Likewise.
9724 (iwmmxt_iordi3): Likewise.
9725 (iwmmxt_xordi3): Likewise.
9726 (iwmmxt_anddi3): Likewise.
9727 (iwmmxt_nanddi3): Likewise.
9728 (iwmmxt_arm_movdi): Likewise.
9729 (iwmmxt_movsi_insn): Likewise.
9730 (mov<mode>_internal): Likewise.
9731 (and<mode>3_iwmmxt): Likewise.
9732 (ior<mode>3_iwmmxt): Likewise.
9733 (xor<mode>3_iwmmxt): Likewise.
9734 (add<mode>3_iwmmxt): Likewise.
9735 (ssaddv8qi3): Likewise.
9736 (ssaddv4hi3): Likewise.
9737 (ssaddv2si3): Likewise.
9738 (usaddv8qi3): Likewise.
9739 (usaddv4hi3): Likewise.
9740 (usaddv2si3): Likewise.
9741 (sub<mode>3_iwmmxt): Likewise.
9742 (sssubv8qi3): Likewise.
9743 (sssubv4hi3): Likewise.
9744 (sssubv2si3): Likewise.
9745 (ussubv8qi3): Likewise.
9746 (ussubv4hi3): Likewise.
9747 (ussubv2si3): Likewise.
9748 (mulv4hi3_iwmmxt): Likewise.
9749 (smulv4hi3_highpart): Likewise.
9750 (umulv4hi3_highpart): Likewise.
9751 (iwmmxt_wmacs): Likewise.
9752 (iwmmxt_wmacsz): Likewise.
9753 (iwmmxt_wmacu): Likewise.
9754 (iwmmxt_wmacuz): Likewise.
9755 (iwmmxt_clrdi): Likewise.
9756 (iwmmxt_clrv8qi): Likewise.
9757 (iwmmxt_clr4hi): Likewise.
9758 (iwmmxt_clr2si): Likewise.
9759 (iwmmxt_uavgrndv8qi3): Likewise.
9760 (iwmmxt_uavgrndv4hi3): Likewise.
9761 (iwmmxt_uavgv8qi3): Likewise.
9762 (iwmmxt_uavgv4hi3): Likewise.
9763 (iwmmxt_tinsrb): Likewise.
9764 (iwmmxt_tinsrh): Likewise.
9765 (iwmmxt_tinsrw): Likewise.
9766 (iwmmxt_textrmub): Likewise.
9767 (iwmmxt_textrmsb): Likewise.
9768 (iwmmxt_textrmuh): Likewise.
9769 (iwmmxt_textrmsh): Likewise.
9770 (iwmmxt_textrmw): Likewise.
9771 (iwmxxt_wshufh): Likewise.
9772 (eqv8qi3): Likewise.
9773 (eqv4hi3): Likewise.
9774 (eqv2si3): Likewise.
9775 (gtuv8qi3): Likewise.
9776 (gtuv4hi3): Likewise.
9777 (gtuv2si3): Likewise.
9778 (gtv8qi3): Likewise.
9779 (gtv4hi3): Likewise.
9780 (gtv2si3): Likewise.
9781 (smax<mode>3_iwmmxt): Likewise.
9782 (umax<mode>3_iwmmxt): Likewise.
9783 (smin<mode>3_iwmmxt): Likewise.
9784 (umin<mode>3_iwmmxt): Likewise.
9785 (iwmmxt_wpackhss): Likewise.
9786 (iwmmxt_wpackwss): Likewise.
9787 (iwmmxt_wpackdss): Likewise.
9788 (iwmmxt_wpackhus): Likewise.
9789 (iwmmxt_wpackwus): Likewise.
9790 (iwmmxt_wpackdus): Likewise.
9791 (iwmmxt_wunpckihb): Likewise.
9792 (iwmmxt_wunpckihh): Likewise.
9793 (iwmmxt_wunpckihw): Likewise.
9794 (iwmmxt_wunpckilb): Likewise.
9795 (iwmmxt_wunpckilh): Likewise.
9796 (iwmmxt_wunpckilw): Likewise.
9797 (iwmmxt_wunpckehub): Likewise.
9798 (iwmmxt_wunpckehuh): Likewise.
9799 (iwmmxt_wunpckehuw): Likewise.
9800 (iwmmxt_wunpckehsb): Likewise.
9801 (iwmmxt_wunpckehsh): Likewise.
9802 (iwmmxt_wunpckehsw): Likewise.
9803 (iwmmxt_wunpckelub): Likewise.
9804 (iwmmxt_wunpckeluh): Likewise.
9805 (iwmmxt_wunpckeluw): Likewise.
9806 (iwmmxt_wunpckelsb): Likewise.
9807 (iwmmxt_wunpckelsh): Likewise.
9808 (iwmmxt_wunpckelsw): Likewise.
9809 (ror<mode>3): Likewise.
9810 (ashr<mode>3_iwmmxt): Likewise.
9811 (lshr<mode>3_iwmmxt): Likewise.
9812 (ashl<mode>3_iwmmxt): Likewise.
9813 (ror<mode>3_di): Likewise.
9814 (ashr<mode>3_di): Likewise.
9815 (lshr<mode>3_di): Likewise.
9816 (ashl<mode>3_di): Likewise.
9817 (iwmmxt_wmadds): Likewise.
9818 (iwmmxt_wmaddu): Likewise.
9819 (iwmmxt_tmia): Likewise.
9820 (iwmmxt_tmiaph): Likewise.
9821 (iwmmxt_tmiabb): Likewise.
9822 (iwmmxt_tmiatb): Likewise.
9823 (iwmmxt_tmiabt): Likewise.
9824 (iwmmxt_tmiatt): Likewise.
9825 (iwmmxt_tmovmskb): Likewise.
9826 (iwmmxt_tmovmskh): Likewise.
9827 (iwmmxt_tmovmskw): Likewise.
9828 (iwmmxt_waccb): Likewise.
9829 (iwmmxt_wacch): Likewise.
9830 (iwmmxt_waccw): Likewise.
9831 (iwmmxt_waligni): Likewise.
9832 (iwmmxt_walignr): Likewise.
9833 (iwmmxt_walignr0): Likewise.
9834 (iwmmxt_walignr1): Likewise.
9835 (iwmmxt_walignr2): Likewise.
9836 (iwmmxt_walignr3): Likewise.
9837 (iwmmxt_wsadb): Likewise.
9838 (iwmmxt_wsadh): Likewise.
9839 (iwmmxt_wsadbz): Likewise.
9840 (iwmmxt_wsadhz): Likewise.
9841 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
9842 (iwmmxt_wabsdiffb): Likewise.
9843 (iwmmxt_wabsdiffh): Likewise.
9844 (iwmmxt_wabsdiffw): Likewise.
9845 (iwmmxt_waddsubhx): Likewise
9846 (iwmmxt_wsubaddhx): Likewise.
9847 (addc<mode>3): Likewise.
9848 (iwmmxt_avg4): Likewise.
9849 (iwmmxt_avg4r): Likewise.
9850 (iwmmxt_wmaddsx): Likewise.
9851 (iwmmxt_wmaddux): Likewise.
9852 (iwmmxt_wmaddsn): Likewise.
9853 (iwmmxt_wmaddun): Likewise.
9854 (iwmmxt_wmulwsm): Likewise.
9855 (iwmmxt_wmulwum): Likewise.
9856 (iwmmxt_wmulsmr): Likewise.
9857 (iwmmxt_wmulumr): Likewise.
9858 (iwmmxt_wmulwsmr): Likewise.
9859 (iwmmxt_wmulwumr): Likewise.
9860 (iwmmxt_wmulwl): Likewise.
9861 (iwmmxt_wqmulm): Likewise.
9862 (iwmmxt_wqmulwm): Likewise.
9863 (iwmmxt_wqmulmr): Likewise.
9864 (iwmmxt_wqmulwmr): Likewise.
9865 (iwmmxt_waddbhusm): Likewise.
9866 (iwmmxt_waddbhusl): Likewise.
9867 (iwmmxt_wqmiabb): Likewise.
9868 (iwmmxt_wqmiabt): Likewise.
9869 (iwmmxt_wqmiatb): Likewise.
9870 (iwmmxt_wqmiatt): Likewise.
9871 (iwmmxt_wqmiabbn): Likewise.
9872 (iwmmxt_wqmiabtn): Likewise.
9873 (iwmmxt_wqmiatbn): Likewise.
9874 (iwmmxt_wqmiattn): Likewise.
9875 (iwmmxt_wmiabb): Likewise.
9876 (iwmmxt_wmiabt): Likewise.
9877 (iwmmxt_wmiatb): Likewise.
9878 (iwmmxt_wmiatt): Likewise.
9879 (iwmmxt_wmiabbn): Likewise.
9880 (iwmmxt_wmiabtn): Likewise.
9881 (iwmmxt_wmiatbn): Likewise.
9882 (iwmmxt_wmiattn): Likewise.
9883 (iwmmxt_wmiawbb): Likewise.
9884 (iwmmxt_wmiawbt): Likewise.
9885 (iwmmxt_wmiawtb): Likewise.
9886 (iwmmxt_wmiawtt): Likewise.
9887 (iwmmxt_wmiawbbn): Likewise.
9888 (iwmmxt_wmiawbtn): Likewise.
9889 (iwmmxt_wmiawtbn): Likewise.
9890 (iwmmxt_wmiawttn): Likewise.
9891 (iwmmxt_wmerge): Likewise.
9892 (iwmmxt_tandc<mode>3): Likewise.
9893 (iwmmxt_torc<mode>3): Likewise.
9894 (iwmmxt_torvsc<mode>3): Likewise.
9895 (iwmmxt_textrc<mode>3): Likewise.
9896 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
9897 (wmmxt_pack): Likewise.
9898 (wmmxt_mult_c1): Likewise.
9899 (wmmxt_mult_c2): Likewise.
9900 (wmmxt_alu_c1): Likewise.
9901 (wmmxt_alu_c2): Likewise.
9902 (wmmxt_alu_c3): Likewise.
9903 (wmmxt_transfer_c1): Likewise.
9904 (wmmxt_transfer_c2): Likewise.
9905 (wmmxt_transfer_c3): Likewise.
9906 (marvell_f_iwmmxt_wstr): Likewise.
9907 (marvell_f_iwmmxt_wldr): Likewise.
9909 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
9911 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
9913 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
9916 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
9917 * lra-constraints.c (need_for_split_p): Check call used hard regs
9918 living through calls.
9920 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
9921 call used regs for call insn.
9923 2013-06-28 Jakub Jelinek <jakub@redhat.com>
9926 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
9927 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
9929 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
9931 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
9933 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
9935 * lra-constraints.c (need_for_split_p): Check call used hard regs
9936 living through calls.
9938 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
9941 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
9942 to tie with any other modes. Eliminate Altivec vector mode tests,
9943 since these are a subset of ALTIVEC or VSX vector modes. Simplify
9944 code, to return 0 if testing MODE2 for a condition, if we've
9945 already tested MODE1 for the same condition.
9947 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
9949 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
9952 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
9954 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
9955 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
9957 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
9959 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
9961 (aarch64_symbolic_constant_p): Remove.
9962 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
9963 static. Fix line length and white space.
9964 (aarch64_symbolic_constant_p): Remove.
9965 * config/aarch64/predicates.md (aarch64_valid_symref):
9966 Use aarch64_classify_symbol_expression.
9968 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9970 * config/arm/constraints.md (Ts): New constraint.
9971 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
9973 (compare_scc): Use "Ts" constraint for operand 0.
9974 (ior_scc_scc): Likewise.
9975 (and_scc_scc): Likewise.
9976 (and_scc_scc_nodom): Likewise.
9977 (ior_scc_scc_cmp): Likewise for operand 7.
9978 (and_scc_scc_cmp): Likewise.
9979 * config/arm/thumb2.md (thumb2_movsi_insn):
9980 Add alternatives for 16-bit encodings.
9981 (thumb2_movhi_insn): Likewise.
9982 (thumb2_movsicc_insn): Likewise.
9983 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
9984 (thumb2_negscc): Use "Ts" constraint.
9985 Move mvn instruction outside cond_exec block.
9986 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
9987 for 16-bit encodings.
9989 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9991 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
9993 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
9994 (mulsi3subsi): Likewise.
9995 (mulsidi3adddi): Likewise.
9996 (mulsidi3_v6): Likewise.
9997 (umulsidi3_v6): Likewise.
9998 (umulsidi3adddi_v6): Likewise.
9999 (smulsi3_highpart_v6): Likewise.
10000 (umulsi3_highpart_v6): Likewise.
10001 (mulhisi3tb): Likewise.
10002 (mulhisi3bt): Likewise.
10003 (mulhisi3tt): Likewise.
10004 (maddhisi4): Likewise.
10005 (maddhisi4tb): Likewise.
10006 (maddhisi4tt): Likewise.
10007 (maddhidi4): Likewise.
10008 (maddhidi4tb): Likewise.
10009 (maddhidi4tt): Likewise.
10010 (zeroextractsi_compare0_scratch): Likewise.
10011 (insv_zero): Likewise.
10012 (insv_t2): Likewise.
10013 (anddi_notzesidi_di): Likewise.
10014 (anddi_notsesidi_di): Likewise.
10015 (andsi_notsi_si): Likewise.
10016 (iordi_zesidi_di): Likewise.
10017 (xordi_zesidi_di): Likewise.
10018 (andsi_iorsi3_notsi): Likewise.
10019 (smax_0): Likewise.
10020 (smax_m1): Likewise.
10021 (smin_0): Likewise.
10022 (not_shiftsi): Likewise.
10023 (unaligned_loadsi): Likewise.
10024 (unaligned_loadhis): Likewise.
10025 (unaligned_loadhiu): Likewise.
10026 (unaligned_storesi): Likewise.
10027 (unaligned_storehi): Likewise.
10028 (extv_reg): Likewise.
10029 (extzv_t2): Likewise.
10030 (divsi3): Likewise.
10031 (udivsi3): Likewise.
10032 (arm_zero_extendhisi2addsi): Likewise.
10033 (arm_zero_extendqisi2addsi): Likewise.
10034 (compareqi_eq0): Likewise.
10035 (arm_extendhisi2_v6): Likewise.
10036 (arm_extendqisi2addsi): Likewise.
10037 (arm_movt): Likewise.
10038 (thumb2_ldrd): Likewise.
10039 (thumb2_ldrd_base): Likewise.
10040 (thumb2_ldrd_base_neg): Likewise.
10041 (thumb2_strd): Likewise.
10042 (thumb2_strd_base): Likewise.
10043 (thumb2_strd_base_neg): Likewise.
10044 (arm_negsi2): Add alternative for 16-bit encoding.
10045 (arm_one_cmplsi2): Likewise.
10047 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10049 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
10050 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
10051 (movdfcc): Likewise.
10052 * config/arm/vfp.md (*thumb2_movsf_vfp):
10053 Disable predication for arm_restrict_it.
10054 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
10055 (*thumb2_movdfcc_vfp): Likewise.
10056 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
10057 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
10058 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
10059 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
10060 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
10061 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
10062 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
10063 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
10064 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
10065 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
10066 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
10067 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
10068 Disable predication for arm_restrict_it.
10070 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
10072 * config/i386/bmiintrin.h (_bextr_u32): New.
10073 (_bextr_u64): Ditto.
10075 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
10077 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
10078 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
10079 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
10080 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
10081 * config/mips/n32-elf.h: ...this new file.
10083 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
10086 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
10087 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
10089 2013-06-27 Catherine Moore <clm@codesourcery.com>
10091 * config/mips/mips-tables.opt: Regenerate.
10092 * config/mips/mips-cpus.def: Add m14ke and m14kec.
10093 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
10094 * doc/invoke.texi: Add -m14kc.
10096 2013-06-27 Jakub Jelinek <jakub@redhat.com>
10099 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
10100 constraints of operand 1 and 2.
10103 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
10104 to match RTL canonicalization. Swap predicates and
10105 constraints of operand 1 and 2.
10107 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
10109 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
10110 Process OP_INOUT regs for splitting too.
10112 2013-06-27 Jakub Jelinek <jakub@redhat.com>
10114 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
10115 decl before the loop, initialize to NULL.
10116 (vectorizable_load): Initialize ptr_incr to NULL.
10118 2013-06-27 Martin Jambor <mjambor@suse.cz>
10121 * ipa-ref.h (ipa_maybe_record_reference): Declare.
10122 * ipa-ref.c (ipa_maybe_record_reference): New function.
10123 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
10124 * ipa-cp.c (create_specialized_node): Record potential references from
10126 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
10128 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
10130 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
10131 parameter 'mode' of type 'enum machine_mode mode'; change to pass
10132 'mode' to force_reg.
10133 (aarch64_add_offset): Update calls to aarch64_force_temporary.
10134 (aarch64_expand_mov_immediate): Likewise.
10136 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
10138 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
10139 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
10141 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10143 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
10144 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
10145 (struct machine_function): Add tbegin_p.
10146 (s390_canonicalize_comparison): Fold CC mode compares to
10147 conditional jump if possible.
10148 (s390_emit_jump): Return the emitted jump.
10149 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
10150 Handle CCRAWmode compares.
10151 (s390_option_override): Default to -mhtm if available.
10152 (s390_reg_clobbered_rtx): Handle floating point regs as well.
10153 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
10154 FPRs instead of df_regs_ever_live_p.
10155 (s390_optimize_nonescaping_tx): New function.
10156 (s390_init_frame_layout): Extend clobbered_regs array to cover
10158 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
10159 (s390_expand_tbegin): New function.
10160 (enum s390_builtin): New enum definition.
10161 (code_for_builtin): New array definition.
10162 (s390_init_builtins): New function.
10163 (s390_expand_builtin): New function.
10164 (TARGET_INIT_BUILTINS): Define.
10165 (TARGET_EXPAND_BUILTIN): Define.
10166 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
10167 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
10168 (s390_alc_comparison): Likewise.
10169 * config/s390/s390-modes.def: Add CCRAWmode.
10170 * config/s390/s390.h (processor_flags): Add PF_TX.
10171 (TARGET_CPU_HTM): Define macro.
10172 (TARGET_HTM): Define macro.
10173 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
10174 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
10175 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
10176 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
10177 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
10178 (TBEGIN_MASK, TBEGINC_MASK): New constants.
10179 ("*cc_to_int"): Move up.
10180 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
10181 constants other than 0.
10182 ("*ccraw_to_int"): New insn and splitter definition.
10183 ("tbegin", "tbegin_nofloat", "tbegin_retry")
10184 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
10185 ("tx_assist"): New expander.
10186 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
10187 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
10188 * config/s390/s390.opt: Add -mhtm option.
10189 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
10190 * config/s390/htmxlintrin.h: New file.
10191 * config/s390/htmintrin.h: New file.
10192 * config/s390/s390intrin.h: New file.
10193 * doc/extend.texi: Document htm builtins.
10194 * config.gcc: Add the new header files to extra_headers.
10196 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
10198 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
10199 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
10201 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
10202 Pat Haugen <pthaugen@us.ibm.com>
10203 Peter Bergner <bergner@vnet.ibm.com>
10205 * config/rs6000/power8.md: New.
10206 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
10207 setting for power8 entry.
10208 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
10209 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
10210 test for Power4/Power5 only.
10211 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
10213 (force_new_group): Adjust comment.
10214 * config/rs6000/rs6000.md: Include power8.md.
10216 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
10218 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
10219 * config/arm/arm-protos.h (arm_max_conditional_execute): New
10221 (tune_params): Update comment.
10222 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
10223 (arm_max_conditional_execute): New function.
10224 (thumb2_final_prescan_insn): Use max_insn_skipped and
10225 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
10227 2013-06-25 Jakub Jelinek <jakub@redhat.com>
10229 PR tree-optimization/57705
10230 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
10231 SSA_NAME step, provided that it is not defined inside the loop.
10232 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
10233 (get_initial_def_for_induction): Handle SSA_NAME IV step.
10235 2013-06-25 Martin Jambor <mjambor@suse.cz>
10237 PR middle-end/57670
10238 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
10239 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
10241 (ipa_note_param_call): Initialize member_ptr flag.
10242 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
10243 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
10244 (ipa_write_indirect_edge_info): Stream member_ptr flag.
10245 (ipa_read_indirect_edge_info): Likewise.
10247 2013-06-25 Richard Biener <rguenther@suse.de>
10249 PR middle-end/56977
10250 * passes.c (init_optimization_passes): Move pass_fold_builtins
10251 and pass_dce earlier with -Og.
10253 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
10255 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
10256 <BIT_FIELD_REF>: Remove trailing TAB.
10257 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
10260 2013-06-24 Martin Jambor <mjambor@suse.cz>
10262 PR tree-optimization/57358
10263 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
10264 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
10265 (ipa_analyze_params_uses): Generate pessimistic info when true.
10267 2013-06-24 Martin Jambor <mjambor@suse.cz>
10269 PR tree-optimization/57539
10270 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
10271 global.inlined_to of the new node to it. All callers changed.
10272 * ipa-inline-transform.c (clone_inlined_nodes): New variable
10273 inlining_into, pass it to cgraph_clone_node.
10274 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
10275 ipa_free_edge_args_substructures.
10276 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
10277 rdesc linked list. Do not assert rdesc edges have inlined caller.
10278 Assert we have found an rdesc in the rdesc list.
10280 2013-06-24 Richard Biener <rguenther@suse.de>
10282 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
10283 (pointer_set_lookup): Declare.
10284 (class pointer_map): New template class implementing a
10285 generic pointer to T map.
10286 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
10287 pointer_map<T>::contains, pointer_map<T>::insert,
10288 pointer_map<T>::traverse): New functions.
10289 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
10290 (pointer_set_lookup): New function.
10291 (pointer_set_contains): Use pointer_set_lookup.
10292 (pointer_set_insert): Likewise.
10293 (insert_aux): Remove.
10294 (struct pointer_map_t): Embed a pointer_set_t.
10295 (pointer_map_create): Adjust.
10296 (pointer_map_destroy): Likewise.
10297 (pointer_map_contains): Likewise.
10298 (pointer_map_insert): Likewise.
10299 (pointer_map_traverse): Likewise.
10300 * tree-streamer.h (struct streamer_tree_cache_d): Use a
10301 pointer_map<unsigned> instead of a pointer_map_t.
10302 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
10303 (streamer_tree_cache_lookup): Likewise.
10304 (streamer_tree_cache_create): Likewise.
10305 (streamer_tree_cache_delete): Likewise.
10306 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
10307 pointer_map<unsigned> instead of a pointer_map_t.
10308 (lto_init_tree_ref_encoder): Adjust.
10309 (lto_destroy_tree_ref_encoder): Likewise.
10310 * lto-section-out.c (lto_output_decl_index): Likewise.
10311 (lto_record_function_out_decl_state): Likewise.
10312 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
10314 2013-06-24 Richard Biener <rguenther@suse.de>
10316 PR tree-optimization/57488
10317 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
10319 2013-06-24 Alan Modra <amodra@gmail.com>
10321 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
10322 (gen_easy_altivec_constant): Likewise.
10323 * config/rs6000/predicates.md (easy_vector_constant_add_self,
10324 easy_vector_constant_msb): Likewise.
10326 2013-06-23 Jakub Jelinek <jakub@redhat.com>
10329 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
10330 add missing return true.
10332 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
10335 * config/sh/predicates.md (general_extend_operand): Invoke
10336 general_movsrc_operand for memory operands.
10337 (general_movsrc_operand): Allow reg+reg addressing, do not use
10338 general_operand for memory operands.
10340 2013-06-23 Sriraman Tallam <tmsriram@google.com>
10342 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
10343 when current target options does not apply.
10344 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
10345 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
10346 * config/i386/bmiintrin.h: Pass appropriate target
10347 attributes to header.
10348 * config/i386/mmintrin.h: Ditto.
10349 * config/i386/nmmintrin.h: Ditto.
10350 * config/i386/avx2intrin.h: Ditto.
10351 * config/i386/fxsrintrin.h: Ditto.
10352 * config/i386/tbmintrin.h: Ditto.
10353 * config/i386/xsaveintrin.h: Ditto.
10354 * config/i386/f16cintrin.h: Ditto.
10355 * config/i386/xtestintrin.h: Ditto.
10356 * config/i386/xsaveoptintrin.h: Ditto.
10357 * config/i386/bmi2intrin.h: Ditto.
10358 * config/i386/lzcntintrin.h: Ditto.
10359 * config/i386/smmintrin.h: Ditto.
10360 * config/i386/wmmintrin.h: Ditto.
10361 * config/i386/x86intrin.h: Remove all header include guards.
10362 * config/i386/prfchwintrin.h: Ditto.
10363 * config/i386/pmmintrin.h: Ditto.
10364 * config/i386/tmmintrin.h: Ditto.
10365 * config/i386/xmmintrin.h: Ditto.
10366 * config/i386/popcntintrin.h: Ditto.
10367 * config/i386/rdseedintrin.h: Ditto.
10368 * config/i386/ammintrin.h: Ditto.
10369 * config/i386/emmintrin.h: Ditto.
10370 * config/i386/immintrin.h: Remove all header include guards.
10371 * config/i386/fma4intrin.h: Ditto.
10372 * config/i386/lwpintrin.h: Ditto.
10373 * config/i386/xopintrin.h: Ditto.
10374 * config/i386/ia32intrin.h: Ditto.
10375 * config/i386/avxintrin.h: Ditto.
10376 * config/i386/rtmintrin.h: Ditto.
10377 * config/i386/fmaintrin.h: Ditto.
10378 * config/i386/mm3dnow.h: Ditto.
10380 2013-06-22 Sriraman Tallam <tmsriram@google.com>
10382 * common/config/i386/i386-common.c: Handle LZCNT.
10384 2013-06-22 Andi Kleen <ak@linux.intel.com>
10386 * doc/extend.texi: Use __atomic_store_n instead of
10387 __atomic_store in HLE example.
10389 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
10391 * config/sh/sh.c: Remove <cstdlib> workaround.
10393 2013-06-21 Andi Kleen <ak@linux.intel.com>
10395 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
10397 2013-06-21 Andi Kleen <ak@linux.intel.com>
10399 * doc/extend.texi: Document that __atomic_clear and
10400 __atomic_test_and_set should only be used with bool.
10402 2013-06-20 Jan Hubicka <jh@suse.cz>
10404 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
10405 types_same_for_odr.
10406 * tree.c (decls_same_for_odr): New function.
10407 (same_for_edr): New function.
10408 (types_same_for_odr): New function.
10409 (get_binfo_at_offset): Use it.
10410 * tree.h (types_same_for_odr): Declare.
10412 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
10413 Jason Merrill <jason@redhat.com>
10415 * system.h: Include <cstdlib> as well as <stdlib.h>.
10417 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
10420 * config/i386/i386.c (construct_container): Report error if
10421 long double is used with disabled x87 float returns.
10423 2013-06-20 Jan Hubicka <jh@suse.cz>
10425 * lto-cgraph.c (input_symtab): Do not set cgraph state.
10427 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
10429 PR rtl-optimization/57425
10430 PR rtl-optimization/57569
10431 * alias.c (write_dependence_p): Remove parameters mem_mode and
10432 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
10433 Changed all callers.
10434 (canon_anti_dependence): Get comments and semantics in sync.
10435 Add parameter mem_canonicalized. Changed all callers.
10436 * rtl.h (canon_anti_dependence): Update prototype.
10438 2013-06-20 Richard Biener <rguenther@suse.de>
10440 * data-streamer-in.c (streamer_read_uhwi): Optimize single
10441 byte case, inline streamer_read_uchar and defer section
10444 2013-06-20 Richard Biener <rguenther@suse.de>
10446 PR tree-optimization/57584
10447 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
10448 SSA names into the expanded expression that take part in
10449 abnormal coalescing.
10451 2013-06-19 Sharad Singhai <singhai@google.com>
10453 * gcov.c (print_usage): Handle new option.
10454 (process_args): Ditto.
10455 (get_gcov_intermediate_filename): New function.
10456 (output_intermediate_file): New function.
10457 (output_gcov_file): New function
10458 (generate_results): Handle new option.
10459 (release_function): Relase demangled name.
10460 (read_graph_file): Handle demangled name.
10461 (output_lines): Ditto.
10462 * doc/gcov.texi: Document gcov intermediate format.
10464 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
10467 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
10468 (lra_emit_add): Use the functions. Add comment about Y as an
10471 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
10474 * collect2.c (collect_atexit): New.
10475 (collect_exit): Delete.
10476 (main): Register collect_atexit with atexit.
10477 (collect_wait): Change collect_exit to exit.
10479 * collect2.h (collect_exit): Delete.
10480 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
10482 2013-06-19 Wei Mi <wmi@google.com>
10484 PR rtl-optimization/57518
10485 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
10486 if regno is used in paradoxical subreg.
10487 (update_equiv_regs): Check pdx_subregs[regno] before
10488 set a reg to be equivalent with a mem.
10490 2013-06-19 Matthias Klose <doko@ubuntu.com>
10493 * file-find.h (find_a_file): Add a mode parameter.
10494 * file-find.c (find_a_file): Likewise.
10495 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
10496 with X_OK for the executables.
10497 * collect2.c (main): Call find_a_file with X_OK.
10499 2013-06-19 Steve Ellcey <sellcey@mips.com>
10502 * config/mips/mips.md (casesi_internal_mips16_<mode>):
10503 Use NEXT_INSN instead of next_real_insn.
10505 2013-06-19 Jan Hubicka <jh@suse.cz>
10507 * cgraph.h (const_value_known_p): Replace by ...
10508 (ctor_for_folding): .. this one.
10509 * cgraphunit.c (process_function_and_variable_attributes): Use it.
10510 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
10511 * expr.c (expand_expr_real_1): Likewise.
10512 (string_constant): Likewise.
10513 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
10514 * ipa.c (process_references): Likewise.
10515 (symtab_remove_unreachable_nodes): Likewise.
10516 * ipa-inline-analysis.c (param_change_prob): Likewise.
10517 * gimple-fold.c (canonicalize_constructor_val): Likewise.
10518 (get_base_constructor): Likwise.
10519 * varpool.c (varpool_remove_node): Likewise.
10520 (varpool_remove_initializer): LIkewise.
10521 (dump_varpool_node): LIkwise.
10522 (const_value_known_p): Rewrite to ...
10523 (ctor_for_folding): ... this one.
10525 2013-06-19 Jakub Jelinek <jakub@redhat.com>
10528 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
10529 PERSONALITY in $PATH derived prefixes.
10531 2013-06-19 Jeff Law <law@redhat.com>
10533 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
10536 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
10537 (simplify_bitwise_binary): Use it to simpify certain binary ops on
10540 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
10542 * config/arm/vfp.md: Move VFP instruction classification documentation
10544 * config/arm/arm.md: ... here. Update instruction classification
10547 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
10549 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
10550 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
10551 pattern. Use more efficient sequences on ARMv5 and Thumb2.
10553 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
10556 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
10557 with NEXT_INSN. Use tablejump_p to check for jump table data
10560 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
10563 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
10564 that now in C++ the value is correct per the C++ standards.
10566 2013-06-19 Richard Biener <rguenther@suse.de>
10568 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
10569 for global context.
10571 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10574 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10577 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
10578 with next_active_insn.
10580 2013-06-18 Sriraman Tallam <tmsriram@google.com>
10582 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
10583 functions are inlined during failures to flag an error.
10584 * tree-inline.c (expand_call_inline): Allow the error to be flagged
10585 in early inline pass.
10587 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
10589 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
10592 2013-06-18 Julian Brown <julian@codesourcery.com>
10594 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
10595 Permit virtual register pre-reload if !strict.
10596 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
10598 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
10600 * config/arm/neon.md (movmisalign<mode>): Use
10601 neon_perm_struct_or_reg_operand instead of
10602 neon_struct_or_register_operand.
10603 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
10604 neon_permissive_struct_operand instead of neon_struct_operand.
10605 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
10606 neon_vector_mem_operand.
10607 * config/arm/predicates.md (neon_struct_operand): Adjust call to
10608 neon_vector_mem_operand.
10609 (neon_permissive_struct_operand): New.
10610 (neon_struct_or_register_operand): Rename to...
10611 (neon_perm_struct_or_reg_operand): This. Adjust call to
10612 neon_vector_mem_operand.
10614 2013-06-18 Richard Biener <rguenther@suse.de>
10616 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
10617 * lto-streamer.h: Include pointer-set.h.
10618 (struct lto_decl_slot): Remove.
10619 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
10620 Remove next_index entry.
10621 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
10622 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
10623 (lto_init_tree_ref_encoder): Adjust.
10624 (lto_destroy_tree_ref_encoder): Likewise.
10625 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
10626 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
10627 (lto_output_decl_index): Adjust.
10628 (lto_new_out_decl_state): Likewise.
10629 (lto_record_function_out_decl_state): Likewise.
10630 * lto-streamer-out.c (copy_function): Likewise.
10632 2013-06-18 Richard Biener <rguenther@suse.de>
10634 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
10635 * cgraphunit.c: Include cfgloop.h.
10636 (init_lowered_empty_function): Initialize the loop tree.
10637 (assemble_thunk): Insert new BBs into loops.
10639 2013-06-18 Richard Biener <rguenther@suse.de>
10641 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
10642 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
10643 the map from cache entry to cache index optional.
10644 (streamer_tree_cache_replace_tree): Adjust accordingly.
10645 (streamer_tree_cache_append): Likewise.
10646 (streamer_tree_cache_delete): Likewise.
10647 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
10648 streamer cache map from cache entry to cache index.
10649 * lto-streamer-out.c (create_output_block): Adjust.
10651 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
10653 * config/arm/arm.md (attribute "insn"): Move multiplication and
10654 division attributes to...
10655 (attribute "type"): ... here. Remove mult.
10656 (attribute "mul32"): New attribute.
10657 (attribute "mul64"): Add umaal.
10658 (*arm_mulsi3): Update attributes.
10659 (*arm_mulsi3_v6): Likewise.
10660 (*thumb_mulsi3): Likewise.
10661 (*thumb_mulsi3_v6): Likewise.
10662 (*mulsi3_compare0): Likewise.
10663 (*mulsi3_compare0_v6): Likewise.
10664 (*mulsi_compare0_scratch): Likewise.
10665 (*mulsi_compare0_scratch_v6): Likewise.
10666 (*mulsi3addsi): Likewise.
10667 (*mulsi3addsi_v6): Likewise.
10668 (*mulsi3addsi_compare0): Likewise.
10669 (*mulsi3addsi_compare0_v6): Likewise.
10670 (*mulsi3addsi_compare0_scratch): Likewise.
10671 (*mulsi3addsi_compare0_scratch_v6): Likewise.
10672 (*mulsi3subsi): Likewise.
10673 (*mulsidi3adddi): Likewise.
10674 (*mulsi3addsi_v6): Likewise.
10675 (*mulsidi3adddi_v6): Likewise.
10676 (*mulsidi3_nov6): Likewise.
10677 (*mulsidi3_v6): Likewise.
10678 (*umulsidi3_nov6): Likewise.
10679 (*umulsidi3_v6): Likewise.
10680 (*umulsidi3adddi): Likewise.
10681 (*umulsidi3adddi_v6): Likewise.
10682 (*smulsi3_highpart_nov6): Likewise.
10683 (*smulsi3_highpart_v6): Likewise.
10684 (*umulsi3_highpart_nov6): Likewise.
10685 (*umulsi3_highpart_v6): Likewise.
10686 (mulhisi3): Likewise.
10687 (*mulhisi3tb): Likewise.
10688 (*mulhisi3bt): Likewise.
10689 (*mulhisi3tt): Likewise.
10690 (maddhisi4): Likewise.
10691 (*maddhisi4tb): Likewise.
10692 (*maddhisi4tt): Likewise.
10693 (maddhidi4): Likewise.
10694 (*maddhidi4tb): Likewise.
10695 (*maddhidi4tt): Likewise.
10696 (divsi3): Likewise.
10697 (udivsi3): Likewise.
10698 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
10699 (thumb2_mulsi_short_compare0): Likewise.
10700 (thumb2_mulsi_short_compare0_scratch): Likewise.
10701 * config/arm/arm1020e.md (1020mult1): Update attribute change.
10702 (1020mult2): Likewise.
10703 (1020mult3): Likewise.
10704 (1020mult4): Likewise.
10705 (1020mult5): Likewise.
10706 (1020mult6): Likewise.
10707 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
10709 (cortex_a15_mult64): Likewise.
10710 (cortex_a15_sdiv): Likewise.
10711 (cortex_a15_udiv): Likewise.
10712 * config/arm/arm1026ejs.md (mult1): Update attribute change.
10718 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
10719 (pj4_ir_div): Likewise.
10720 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
10721 (11_mult2): Likewise.
10722 (11_mult3): Likewise.
10723 (11_mult4): Likewise.
10724 (11_mult5): Likewise.
10725 (11_mult6): Likewise.
10726 (11_mult7): Likewise.
10727 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
10728 (cortex_a8_mla): Likewise.
10729 (cortex_a8_mull): Likewise.
10730 (cortex_a8_smulwy): Likewise.
10731 (cortex_a8_smlald): Likewise.
10732 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
10733 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
10734 (cortex_r4_mul_3): Likewise.
10735 (cortex_r4_mla_4): Likewise.
10736 (cortex_r4_mla_3): Likewise.
10737 (cortex_r4_smlald): Likewise.
10738 (cortex_r4_mull): Likewise.
10739 (cortex_r4_sdiv): Likewise.
10740 (cortex_r4_udiv): Likewise.
10741 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
10742 (cortex_a7_idiv): Likewise.
10743 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
10744 (9_mult2): Likewise.
10745 (9_mult3): Likewise.
10746 (9_mult4): Likewise.
10747 (9_mult5): Likewise.
10748 (9_mult6): Likewise.
10749 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
10750 (cortex_a53_sdiv): Likewise.
10751 (cortex_a53_udiv): Likewise.
10752 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
10753 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
10754 (mp626_mult2): Likewise.
10755 (mp626_mult3): Likewise.
10756 (mp626_mult4): Likewise.
10757 * config/arm/fa526.md (526_mult1): Update attribute change.
10758 (526_mult2): Likewise.
10759 * config/arm/arm-generic.md (mult): Update attribute change.
10760 (mult_ldsched_strongarm): Likewise.
10761 (mult_ldsched): Likewise.
10762 (multi_cycle): Likewise.
10763 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
10764 * config/arm/fa606te.md (606te_mult1): Update attribute change.
10765 (606te_mult2): Likewise.
10766 (606te_mult3): Likewise.
10767 (606te_mult4): Likewise.
10768 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
10769 (cortex_a9_mac16): Likewise.
10770 (cortex_a9_multiply): Likewise.
10771 (cortex_a9_mac): Likewise.
10772 (cortex_a9_multiply_long): Likewise.
10773 * config/arm/fa626te.md (626te_mult1): Update attribute change.
10774 (626te_mult2): Likewise.
10775 (626te_mult3): Likewise.
10776 (626te_mult4): Likewise.
10778 2013-06-18 Richard Biener <rguenther@suse.de>
10781 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
10783 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10786 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
10787 with next_active_insn.
10789 2013-06-18 Alan Modra <amodra@gmail.com>
10791 * config/rs6000/rs6000.h (enum data_align): New.
10792 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
10793 (DATA_ABI_ALIGNMENT): Define.
10794 (CONSTANT_ALIGNMENT): Correct comment.
10795 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
10796 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
10798 2013-06-17 David Malcolm <dmalcolm@redhat.com>
10800 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
10801 ATTRIBUTE_UNUSED marking.
10803 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
10805 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
10806 alternative and update.
10807 (aarch64_dup_lanedi): Delete.
10808 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
10809 * config/aarch64/aarch64-simd-builtins.def: Update.
10811 2013-06-17 Richard Biener <rguenther@suse.de>
10813 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
10814 (lto_input_scc): Declare.
10815 (lto_input_tree_1): Likewise.
10816 (struct lto_stats_d): Add num_tree_bodies_output and
10817 num_pickle_refs_output.
10818 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
10819 (lto_read_tree_1): Split out from ...
10820 (lto_read_tree): ... this.
10821 (lto_input_scc): New function.
10822 (lto_input_tree_1): Split out from ...
10823 (lto_input_tree): ... this. Handle LTO_tree_scc.
10824 (lto_data_in_create): Create the streamer cache without hashes.
10825 * lto-streamer-out.c (create_output_block): Create the streamer
10826 cache with hashes when not doing WPA.
10827 (lto_write_tree_1): Split out from ...
10828 (lto_write_tree): ... this.
10829 (get_symbol_initial_value): New function.
10830 (lto_output_tree_1): Split out from ...
10831 (lto_output_tree): ... this. Write trees as series of SCCs
10832 using a DFS walk via DFS_write_tree.
10833 (struct sccs, struct scc_entry): New types.
10834 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
10835 (DFS_write_tree_body): New function.
10836 (DFS_write_tree): Likewise.
10837 (hash_tree): Likewise.
10838 (scc_entry_compare): Likewise.
10839 (hash_scc): Likewise.
10840 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
10841 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
10842 TREE_CHAIN as regular reference.
10843 (streamer_read_integer_cst): Remove.
10844 (streamer_get_pickled_tree): Adjust.
10845 * tree-streamer-out.c (streamer_write_chain): Disable streaming
10846 of DECL_EXTERNALs in BLOCK_VARS for now.
10847 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
10849 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
10850 Add hash value argument and record that if hashes are recorded
10852 (streamer_tree_cache_insert_1): Adjust.
10853 (streamer_tree_cache_insert): Likewise.
10854 (streamer_tree_cache_insert_at): Rename to ...
10855 (streamer_tree_cache_replace_tree): ... this and adjust.
10856 (streamer_tree_cache_append): Adjust.
10857 (record_common_node): Likewise.
10858 (streamer_tree_cache_create): Add argument whether to
10859 record hash values together with trees.
10860 (streamer_tree_cache_delete): Adjust.
10861 * tree-streamer.h (struct streamer_tree_cache_d): Add
10863 (streamer_read_integer_cst): Remove.
10864 (streamer_tree_cache_insert): Adjust.
10865 (streamer_tree_cache_append): Likewise.
10866 (streamer_tree_cache_insert_at): Rename to ...
10867 (streamer_tree_cache_replace_tree): ... this and adjust.
10868 (streamer_tree_cache_create): Add argument whether to record hashes.
10869 (streamer_tree_cache_get): Rename to ...
10870 (streamer_tree_cache_get_tree): ... this.
10871 (streamer_tree_cache_get_hash): New function.
10872 * tree.c (cache_integer_cst): New function.
10873 * tree.h (cache_integer_cst): Declare.
10874 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
10875 * lto-symtab.c (lto_varpool_replace_node): Only release
10876 DECL_INITIAL of non-prevailing decls.
10877 * varpool.c (varpool_remove_initializer): Do not release
10878 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
10880 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
10882 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
10883 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
10884 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
10885 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
10886 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
10887 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
10888 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
10889 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
10890 instead of TARGET_64BIT.
10891 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
10892 Require ISA_HAS_<D>DIV.
10894 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
10896 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
10897 (mips*-*-linux*): Move default with_llsc setting to where other
10899 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
10901 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
10902 Likewise. Remove default with_tune setting. Move default float
10903 setting to its own block. Handle with_llsc in the same block as above.
10905 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
10907 PR rtl-optimization/57425
10908 PR rtl-optimization/57569
10909 * alias.c (write_dependence_p): Add new parameters mem_mode,
10910 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
10911 Changed all callers.
10912 (canon_anti_dependence): New function.
10913 * cse.c (check_dependence): Use canon_anti_dependence.
10914 * cselib.c (cselib_invalidate_mem): Likewise.
10915 * rtl.h (canon_anti_dependence): Declare.
10917 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
10919 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
10920 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
10921 ".set mips3" for 64-bit targets.
10923 2013-06-15 Dehao Chen <dehao@google.com>
10925 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
10926 * gimple-low.c (gimple_check_call_matching_types): Likewise.
10927 (gimple_check_call_args): Likewise.
10928 * value-prof.c (check_ic_target): Likewise.
10929 * ipa-inline.c (early_inliner): Likewise.
10930 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
10931 * cgraph.c (cgraph_create_edge_1): Likewise.
10932 (cgraph_make_edge_direct): Likewise.
10934 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
10937 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
10938 rs6000_output_move_128bit to handle emitting quad memory
10939 operations. Set attribute length to 8 bytes.
10941 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
10943 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
10945 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
10946 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
10947 (aarch64_<su>mlsl<mode>): Likewise.
10949 2013-06-14 Mike Stump <mikestump@comcast.net>
10951 * Makefile.in (TARGET_H): Add insn-codes.h.
10953 2013-06-14 Alan Modra <amodra@gmail.com>
10955 PR middle-end/57134
10956 PR middle-end/57586
10957 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
10958 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
10959 bitfield expansion when EXPAND_MEMORY.
10960 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
10962 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
10964 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
10965 test for clearing quad memory on 32-bit later.
10967 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
10969 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
10970 (fold_negate_expr): Likewise.
10971 (fold_real_zero_addition_p): Handle vectors.
10972 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
10974 2013-06-14 Alan Modra <amodra@gmail.com>
10976 * varasm.c (force_const_mem): Revert 2013-06-07 change.
10978 2013-06-13 Jan Hubicka <jh@suse.cz>
10980 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
10981 Local comdats are not externally visible.
10982 * symtab.c (dump_symtab_base): Dump externally visible.
10983 (verify_symtab_base): Verify back links in the symtab hash.
10985 2013-06-13 Bin Cheng <bin.cheng@arm.com>
10987 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
10988 CONVERT_EXPR as equal nodes.
10990 2013-06-13 Bin Cheng <bin.cheng@arm.com>
10992 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
10994 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
10996 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
10997 Generalize to complex and vector.
10998 * tree.c (build_all_ones_cst): New function.
10999 * tree.h (build_all_ones_cst): Declare it.
11001 2013-06-13 Alan Modra <amodra@gmail.com>
11003 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
11004 * config/rs6000/rs6000.md (signbittf2): New insn.
11005 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
11006 (abstf2_internal, cmptf_internal2): Likewise.
11007 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
11009 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
11010 Pat Haugen <pthaugen@us.ibm.com>
11011 Peter Bergner <bergner@vnet.ibm.com>
11013 * config/rs6000/rs6000.c (emit_load_locked): Add support for
11014 power8 byte, half-word, and quad-word atomic instructions.
11015 (emit_store_conditional): Likewise.
11016 (rs6000_expand_atomic_compare_and_swap): Likewise.
11017 (rs6000_expand_atomic_op): Likewise.
11019 * config/rs6000/sync.md (larx): Add new modes for power8.
11021 (AINT): New mode iterator to include TImode as well as normal
11022 integer modes on power8.
11023 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
11024 that VSX registers are not considered. Use AINT mode iterator
11025 instead of INT1 to allow inclusion of quad word atomic operations
11027 (load_locked<mode>): Likewise.
11028 (store_conditional<mode>): Likewise.
11029 (atomic_compare_and_swap<mode>): Likewise.
11030 (atomic_exchange<mode>): Likewise.
11031 (atomic_nand<mode>): Likewise.
11032 (atomic_fetch_<fetchop_name><mode>): Likewise.
11033 (atomic_nand_fetch<mode>): Likewise.
11034 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
11036 (ATOMIC): On power8, add QImode, HImode modes.
11037 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
11038 modes that promote to SImode.
11039 (load_lockedti): Convert TImode arguments to PTImode, so that we
11040 get a guaranteed even/odd register pair.
11041 (load_lockedpti): Likewise.
11042 (store_conditionalti): Likewise.
11043 (store_conditionalpti): Likewise.
11045 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
11046 atomic load/store instructions.
11049 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
11051 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
11053 (insn_count): New attribute, with most cases extracted from...
11054 (length): ...here. Redefine most cases in terms of insn_count.
11055 (single_insn): Delete.
11056 (can_delay): Use insn_count to check for single instructions.
11057 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
11058 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
11059 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
11060 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
11061 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
11062 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
11063 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
11064 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
11065 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
11066 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
11067 rather than "length".
11068 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
11069 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
11070 Use "insn_count" rather than "length".
11071 * config/mips/mips-dsp.md
11072 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
11073 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
11076 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
11078 PR tree-optimization/57361
11079 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
11081 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
11083 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
11085 (aarch64_simd_combine<mode>): New instruction expansion.
11086 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
11087 function prototype.
11088 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
11089 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
11091 2013-06-12 Jan Hubicka <jh@suse.cz>
11093 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
11094 decl has when in streaming stage.
11095 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
11096 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
11098 2013-06-12 Roland Stigge <stigge@antcom.de>
11101 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
11103 2013-06-12 Jakub Jelinek <jakub@redhat.com>
11105 PR tree-optimization/57537
11106 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
11107 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
11109 2013-06-12 Richard Biener <rguenther@suse.de>
11111 * data-streamer.h (streamer_write_char_stream): CSE
11112 obs->current_pointer.
11113 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
11114 streamer_write_char_stream manually and optimize the resulting loop.
11115 (streamer_write_hwi_stream): Likewise.
11117 2013-06-12 Jan Hubicka <jh@suse.cz>
11119 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
11120 * cgraph.h (varpool_create_empty_node): Declare.
11121 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
11123 * symtab.c (symtab_unregister_node): Be lax about missin entries
11125 (symtab_get_node): Update comment.
11126 * varpool.c (varpool_create_empty_node): Break out from ...
11127 (varpool_node_for_decl): ... here.
11128 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
11130 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
11132 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
11133 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
11134 part. Use straight-line flow at the end.
11135 <COMPONENT_REF>: Remove superfluous else.
11136 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
11138 2013-06-12 Jakub Jelinek <jakub@redhat.com>
11141 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
11142 target hook even for !TREE_PUBLIC decls. If no resolution info
11143 is available, return false for common and external decls.
11145 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
11147 * config/rl78/constraints.md (U): New constraint.
11148 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
11151 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
11154 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
11155 to allow returning address to AT_PLATFORM name.
11157 2013-06-11 Jan Hubicka <jh@suse.cz>
11159 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
11160 * cgraph.h (symtab_node_base): Add weakref flag.
11161 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
11162 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
11163 (output_weakrefs): Use weakref flag.
11164 * fold-const.c (simple_operand_p): Handle WEAK.
11165 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
11166 * ipa.c (varpool_externally_visible_p): Drop weakref.
11167 (function_and_variable_visibility): Update comment; fix weakref
11168 sanity checks; do not clear DECL_WEAK on them.
11169 * lto-cgraph.c (lto_output_node): update.
11170 (lto_output_varpool_node): Update.
11171 (input_overwrite_node): Update.
11172 (input_node): Update.
11173 (input_varpool_node): Update.
11174 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
11175 (lto_symtab_merge_symbols): Add sanity check.
11176 (lto_symtab_prevailing_decl): Do not special case weakrefs.
11177 * passes.c (rest_of_decl_compilation): Set static flag, too.
11178 * symtab.c (dump_symtab_base): Dump weakref.
11179 (verify_symtab_base): Sanity check weakrefs.
11180 (symtab_make_decl_local): Remove duplicated code.
11181 (symtab_alias_ultimate_target): Simplify.
11182 * varpool.c (varpool_create_variable_alias): Set weakref flag.
11184 2013-06-11 Tom de Vries <tom@codesourcery.com>
11186 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
11187 sequence_vect == NULL.
11189 2013-06-11 DJ Delorie <dj@redhat.com>
11191 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
11192 (rl78_unwind_word_mode): New.
11194 2013-06-11 David Malcolm <dmalcolm@redhat.com>
11196 * final.c (debug_prefix_maps): Make static.
11198 2013-06-11 David Malcolm <dmalcolm@redhat.com>
11200 * function.c (initial_trampoline): Remove stray copy.
11202 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
11204 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
11206 2013-06-11 Martin Jambor <mjambor@suse.cz>
11208 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
11209 within bounds at the beginning of the function.
11211 2013-06-11 Alan Modra <amodra@gmail.com>
11213 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
11215 (get_named_section): Don't NULL !DECL_P decl.
11217 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
11219 * doc/invoke.texi (core-avx2): Document.
11221 (atom): Updated with MOVBE.
11223 2013-06-11 Richard Biener <rguenther@suse.de>
11225 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
11227 2013-06-11 Anton Blanchard <anton@samba.org>
11229 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
11230 correct shift value in little-endian mode.
11232 2013-06-11 Jakub Jelinek <jakub@redhat.com>
11235 * varasm.c (get_variable_align): Move #endif to the right place.
11237 2013-06-10 Cary Coutant <ccoutant@google.com>
11239 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
11240 for hash so that hash table traversal order is deterministic.
11242 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
11243 Pat Haugen <pthaugen@us.ibm.com>
11244 Peter Bergner <bergner@vnet.ibm.com>
11246 * config/rs6000/vector.md (GPR move splitter): Do not split moves
11247 of vectors in GPRS if they are direct moves or quad word load or
11250 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
11252 (direct_move_p): Likewise.
11253 (quad_load_store_p): Likewise.
11255 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
11256 classes into bins based on the physical register type.
11257 (reg_class_to_reg_type): Likewise.
11258 (IS_STD_REG_TYPE): Likewise.
11259 (IS_FP_VECT_REG_TYPE): Likewise.
11260 (reload_fpr_gpr): Arrays to determine what insn to use if we can
11261 use direct move instructions.
11262 (reload_gpr_vsx): Likewise.
11263 (reload_vsx_gpr): Likewise.
11264 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
11265 information that is a simplification of register classes. Also
11266 precalculate direct move reload helpers.
11267 (direct_move_p): New function to return true if the operation can
11268 be done as a direct move instruciton.
11269 (quad_load_store_p): New function to return true if the operation
11270 is a quad memory operation.
11271 (rs6000_legitimize_address): If quad memory, only allow register
11272 indirect for TImode addresses.
11273 (rs6000_legitimate_address_p): Likewise.
11274 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
11275 (rs6000_reload_register_type): Likewise.
11276 (register_to_reg_type): Return register type.
11277 (rs6000_secondary_reload_simple_move): New helper function for
11278 secondary reload and secondary memory needed to identify anything
11279 that is a simple move, and does not need reloading.
11280 (rs6000_secondary_reload_direct_move): New helper function for
11281 secondary reload to identify cases that can be done with several
11282 instructions via the direct move instructions.
11283 (rs6000_secondary_reload_move): New helper function for secondary
11284 reload to identify moves between register types that can be done.
11285 (rs6000_secondary_reload): Add support for quad memory operations
11286 and for direct move.
11287 (rs6000_secondary_memory_needed): Likewise.
11288 (rs6000_debug_secondary_memory_needed): Change argument names.
11289 (rs6000_output_move_128bit): New function to return the move to
11290 use for 128-bit moves, including knowing about the various
11291 limitations of quad memory operations.
11293 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
11294 memory operations. call rs6000_output_move_128bit for the actual
11295 instruciton(s) to generate.
11296 (vsx_movti_64bit): Likewise.
11298 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
11299 (UNSPEC_P8V_MTVSRWZ): Likewise.
11300 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
11301 (UNSPEC_P8V_MTVSRD): Likewise.
11302 (UNSPEC_P8V_XXPERMDI): Likewise.
11303 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
11304 (UNSPEC_FUSION_GPR): Likewise.
11305 (FMOVE128_GPR): New iterator for direct move.
11306 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
11307 (f32_sv): Likewise.
11308 (f32_dm): Likewise.
11309 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
11310 loads and direct move instructions.
11311 (zero_extendsidi2_lfiwzx): Likewise.
11312 (extendsidi2_lfiwax): Likewise.
11313 (extendsidi2_nocell): Likewise.
11314 (floatsi<mode>2_lfiwax): Likewise.
11315 (lfiwax): Likewise.
11316 (floatunssi<mode>2_lfiwzx): Likewise.
11317 (lfiwzx): Likewise.
11318 (fix_trunc<mode>_stfiwx): Likewise.
11319 (fixuns_trunc<mode>_stfiwx): Likewise.
11320 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
11321 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
11322 (parity<mode>2_cmpb): Set length/type attr.
11323 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
11324 for 'mr.' to fast_compare.
11325 (bpermd_<mode>): Change type attr to popcnt.
11326 (p8_fmrgow_<mode>): New insns for power8 direct move support.
11327 (p8_mtvsrwz_1): Likewise.
11328 (p8_mtvsrwz_2): Likewise.
11329 (reload_fpr_from_gpr<mode>): Likewise.
11330 (p8_mtvsrd_1): Likewise.
11331 (p8_mtvsrd_2): Likewise.
11332 (p8_xxpermdi_<mode>): Likewise.
11333 (reload_vsx_from_gpr<mode>): Likewise.
11334 (reload_vsx_from_gprsf): Likewise.
11335 (p8_mfvsrd_3_<mode>): LIkewise.
11336 (reload_gpr_from_vsx<mode>): Likewise.
11337 (reload_gpr_from_vsxsf): Likewise.
11338 (p8_mfvsrd_4_disf): Likewise.
11339 (multi-word GPR splits): Do not split direct moves or quad memory
11342 2013-06-10 David Malcolm <dmalcolm@redhat.com>
11344 * tree-into-ssa.c (interesting_blocks): Make static.
11346 2013-06-10 Jakub Jelinek <jakub@redhat.com>
11349 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
11350 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
11351 Use DATA_ABI_ALIGNMENT for that case instead if defined.
11352 (get_variable_align): New function.
11353 (get_variable_section, emit_bss, emit_common,
11354 assemble_variable_contents, place_block_symbol): Use
11355 get_variable_align instead of DECL_ALIGN.
11356 (assemble_noswitch_variable): Add align argument, use it
11357 instead of DECL_ALIGN.
11358 (assemble_variable): Adjust caller. Use get_variable_align
11359 instead of DECL_ALIGN.
11360 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
11362 (DATA_ABI_ALIGNMENT): Define.
11363 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
11364 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
11365 opt is false, only return the psABI mandated alignment increase.
11366 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
11367 (DATA_ABI_ALIGNMENT): ... this.
11368 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
11369 (DATA_ABI_ALIGNMENT): ... this.
11370 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
11371 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
11372 (DATA_ABI_ALIGNMENT): ... this.
11373 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
11374 * doc/tm.texi: Regenerated.
11376 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
11378 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
11379 cmp_code to construct REG_EQUAL note.
11381 2013-06-09 Jakub Jelinek <jakub@redhat.com>
11384 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
11385 that operands[2] doesn't overlap with operands[0].
11387 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
11388 Jan Hubicka <jh@suse.cz>
11390 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
11391 hack to mark symbols as used.
11393 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
11395 PR rtl-optimization/57559
11396 * lra-constraints.c (process_alt_operands): Don't discourage
11397 memory with known offset for offsetable memory constraint.
11398 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
11400 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
11402 * varasm.c (struct oc_local_state): Reorder fields.
11403 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
11404 and adjust accordingly.
11405 (output_constructor): Reorder initialization code and adjust call to
11406 output_constructor_bitfield.
11408 2013-06-07 Jan Hubicka <jh@suse.cz>
11410 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
11412 2013-06-07 David Malcolm <dmalcolm@redhat.com>
11414 * tree-object-size.c (unknown): Make const.
11416 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11418 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
11419 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
11420 for last alternative in the cpu_facility attribute.
11422 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11425 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
11426 (xordi3): Change operand 2 constraint to arm_xordi_operand.
11427 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
11428 * config/arm/constraints.md (Dg): New constraint.
11429 * config/arm/neon.md (xordi3_neon): Remove.
11430 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
11431 * config/arm/predicates.md (arm_xordi_operand): New predicate.
11433 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11435 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
11436 Clean up alternatives.
11438 2013-06-07 Alan Modra <amodra@gmail.com>
11440 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
11443 2013-06-07 Alan Modra <amodra@gmail.com>
11445 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
11447 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11449 * config/arm/constraints.md (Df): New constraint.
11450 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
11451 Correct length attribute for last two alternatives.
11453 2013-06-07 Alan Modra <amodra@gmail.com>
11455 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
11456 override user -mfp-in-toc.
11457 (offsettable_ok_by_alignment): Consider just the current access
11458 rather than the whole object, unless BLKmode. Handle
11459 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
11460 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
11461 for -mcmodel=medium.
11462 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
11463 override user -mfp-in-toc or -msum-in-toc. Default to
11464 -mno-fp-in-toc for -mcmodel=medium.
11466 2013-06-06 DJ Delorie <dj@redhat.com>
11468 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
11469 TARGET_VALID_POINTER_MODE.
11471 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
11472 Pat Haugen <pthaugen@us.ibm.com>
11473 Peter Bergner <bergner@vnet.ibm.com>
11475 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
11476 Document new power8 builtins.
11478 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
11479 condition code register, to allow 128-bit logical operations to be
11480 done in the VSX or GPR registers.
11481 (nor<mode>3): Use the canonical form for nor.
11482 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
11483 vclz*, and vpopcnt* vector instructions.
11484 (nand<mode>3): Likewise.
11485 (orc<mode>3): Likewise.
11486 (clz<mode>2): LIkewise.
11487 (popcount<mode>2): Likewise.
11489 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
11490 that only the GPRs are recognized.
11492 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
11493 support for new power8 builtins.
11495 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
11497 (xscvdpspn): Likewise.
11503 (vpopcnt): Likewise.
11504 (vpopcntb): Likewise.
11505 (vpopcnth): Likewise.
11506 (vpopcntw): Likewise.
11507 (vpopcntd): Likewise.
11509 (vmrgew): Likewise.
11510 (vmrgow): Likewise.
11512 (eqv_v16qi3): Likewise.
11513 (eqv_v8hi3): Likewise.
11514 (eqv_v4si3): Likewise.
11515 (eqv_v2di3): Likewise.
11516 (eqv_v4sf3): Likewise.
11517 (eqv_v2df3): Likewise.
11519 (nand_v16qi3): Likewise.
11520 (nand_v8hi3): Likewise.
11521 (nand_v4si3): Likewise.
11522 (nand_v2di3): Likewise.
11523 (nand_v4sf3): Likewise.
11524 (nand_v2df3): Likewise.
11526 (orc_v16qi3): Likewise.
11527 (orc_v8hi3): Likewise.
11528 (orc_v4si3): Likewise.
11529 (orc_v2di3): Likewise.
11530 (orc_v4sf3): Likewise.
11531 (orc_v2df3): Likewise.
11533 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
11534 allow power8 quad mode in 64-bit.
11535 (rs6000_builtin_vectorized_function): Add support to vectorize
11536 ISA 2.07 count leading zeros, population count builtins.
11537 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
11538 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
11539 (builtin_function_type): Add vgbbd builtin function which takes an
11541 (altivec_expand_vec_perm_const): Add support for new power8 merge
11544 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
11545 that does not include TImdoe for use with 32-bit.
11546 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
11548 (UNSPEC_VSX_CVDPSPN): Likewise.
11549 (vsx_xscvdpspn): Likewise.
11550 (vsx_xscvspdpn): Likewise.
11551 (vsx_xscvdpspn_scalar): Likewise.
11552 (vsx_xscvspdpn_directmove): Likewise.
11553 (vsx_and<mode>3): Split logical operations into 32-bit and
11554 64-bit. Add support to do logical operations on TImode as well as
11555 VSX vector types. Allow logical operations to be done in either
11556 VSX registers or in general purpose registers in 64-bit mode. Add
11557 splitters if GPRs were used. For AND, add clobber of CCmode to
11558 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
11560 (vsx_and<mode>3_32bit): Likewise.
11561 (vsx_and<mode>3_64bit): Likewise.
11562 (vsx_ior<mode>3): Likewise.
11563 (vsx_ior<mode>3_32bit): Likewise.
11564 (vsx_ior<mode>3_64bit): Likewise.
11565 (vsx_xor<mode>3): Likewise.
11566 (vsx_xor<mode>3_32bit): Likewise.
11567 (vsx_xor<mode>3_64bit): Likewise.
11568 (vsx_one_cmpl<mode>2): Likewise.
11569 (vsx_one_cmpl<mode>2_32bit): Likewise.
11570 (vsx_one_cmpl<mode>2_64bit): Likewise.
11571 (vsx_nor<mode>3): Likewise.
11572 (vsx_nor<mode>3_32bit): Likewise.
11573 (vsx_nor<mode>3_64bit): Likewise.
11574 (vsx_andc<mode>3): Likewise.
11575 (vsx_andc<mode>3_32bit): Likewise.
11576 (vsx_andc<mode>3_64bit): Likewise.
11577 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
11578 and xxlorc instructions.
11579 (vsx_eqv<mode>3_64bit): Likewise.
11580 (vsx_nand<mode>3_32bit): Likewise.
11581 (vsx_nand<mode>3_64bit): Likewise.
11582 (vsx_orc<mode>3_32bit): Likewise.
11583 (vsx_orc<mode>3_64bit): Likewise.
11585 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
11587 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
11589 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
11590 (p8_vmrgow): Likewise.
11591 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
11592 GPRs to be split under VSX.
11593 (p8v_clz<mode>2): Add power8 count leading zero support.
11594 (p8v_popcount<mode>2): Add power8 population count support.
11595 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
11598 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
11601 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
11603 (vec_nand): Likewise.
11604 (vec_vclz): Likewise.
11605 (vec_vclzb): Likewise.
11606 (vec_vclzd): Likewise.
11607 (vec_vclzh): Likewise.
11608 (vec_vclzw): Likewise.
11609 (vec_vgbbd): Likewise.
11610 (vec_vmrgew): Likewise.
11611 (vec_vmrgow): Likewise.
11612 (vec_vpopcnt): Likewise.
11613 (vec_vpopcntb): Likewise.
11614 (vec_vpopcntd): Likewise.
11615 (vec_vpopcnth): Likewise.
11616 (vec_vpopcntw): Likewise.
11618 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
11620 PR rtl-optimization/57468
11621 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
11624 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
11626 PR rtl-optimization/57459
11627 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
11628 type when setting live regs.
11630 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
11632 * config/s390/s390.opt (mlra): New option.
11633 * config/s390/s390.c (s390_decompose_address): Check displacement
11634 for all registers for LRA.
11635 (s390_secondary_reload): Don't used secondary reloads for LRA.
11636 (s390_lra_p): New function.
11637 (TARGET_LRA_P): Define.
11638 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
11639 of attribute cpu_facility to zarch for the last alternative.
11640 (*cmpmem_short): Ditto.
11642 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
11644 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
11645 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
11646 (arm_expand_prologue): Likewise.
11648 2013-06-06 Teresa Johnson <tejohnson@google.com>
11651 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
11652 as this is now done by redirect_edge_and_branch_force.
11653 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
11654 barriers, and fix interaction with splitting.
11655 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
11656 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
11657 reflect changes made in the routine.
11658 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
11659 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
11660 since this is called in cfglayout mode, and replace partition fixup
11661 with assert as that is now done by force_nonfallthru_and_redirect.
11662 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
11663 already be marked with region crossing note.
11664 (insert_section_boundary_note): Make non-static, gate on flag
11665 has_bb_partition, rewrite to also check for multiple partitions.
11666 (rest_of_handle_reorder_blocks): Remove call to
11667 insert_section_boundary_note, now done later during free_cfg.
11668 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
11669 * bb-reorder.h (insert_section_boundary_note): Declare.
11670 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
11671 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
11672 invoke insert_section_boundary_note.
11673 (try_redirect_by_replacing_jump): Remove unnecessary
11674 check for region crossing note.
11675 (fixup_partition_crossing): New function.
11676 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
11677 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
11678 in non-cfglayout mode.
11679 (force_nonfallthru_and_redirect): Fixup partition boundaries,
11680 remove old code that tried to do this. Emit barrier correctly
11681 when we are in cfglayout mode.
11682 (last_bb_in_partition): New function.
11683 (rtl_split_edge): Correctly fixup partition boundaries.
11684 (commit_one_edge_insertion): Remove old code that tried to
11685 fixup region crossing edge since this is now handled in
11686 split_block, and set up insertion point correctly since
11687 block may now end in a jump.
11688 (verify_hot_cold_block_grouping): Guard against checking when not in
11689 linearized RTL mode.
11690 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
11692 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
11693 rtl_verify_flow_info, so not called in cfglayout mode.
11694 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
11695 (fixup_reorder_chain): Remove old code that attempted to fixup region
11696 crossing note as this is now handled in force_nonfallthru_and_redirect.
11697 (duplicate_insn_chain): Don't duplicate switch section notes.
11698 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
11700 * basic-block.h (emit_barrier_after_bb): Declare.
11702 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11704 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
11705 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
11706 arm_usatsihi): Adjust alternatives for arm_restrict_it.
11708 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11710 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
11712 * config/arm/ldmstm.md: Regenerate.
11714 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11716 * config/arm/sync.md (atomic_loaddi_1):
11717 Disable predication for arm_restrict_it.
11718 (arm_load_exclusive<mode>): Likewise.
11719 (arm_load_exclusivesi): Likewise.
11720 (arm_load_exclusivedi): Likewise.
11721 (arm_load_acquire_exclusive<mode>): Likewise.
11722 (arm_load_acquire_exclusivesi): Likewise.
11723 (arm_load_acquire_exclusivedi): Likewise.
11724 (arm_store_exclusive<mode>): Likewise.
11725 (arm_store_exclusive<mode>): Likewise.
11726 (arm_store_release_exclusivedi): Likewise.
11727 (arm_store_release_exclusive<mode>): Likewise.
11729 2013-06-06 Richard Biener <rguenther@suse.de>
11731 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
11733 (lto_tag_is_tree_code_p): Adjust.
11734 (lto_tag_is_gimple_code_p): Likewise.
11735 (lto_gimple_code_to_tag): Likewise.
11736 (lto_tag_to_gimple_code): Likewise.
11737 (lto_tree_code_to_tag): Likewise.
11738 (lto_tag_to_tree_code): Likewise.
11739 * data-streamer.h (streamer_write_hwi_in_range): Use
11740 uhwi streaming to stream the normalized range.
11741 (streamer_read_hwi_in_range): Likewise.
11743 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11745 * config/arm/arm.md (enabled_for_depr_it): New attribute.
11746 (predicable_short_it): Likewise.
11747 (predicated): Likewise.
11748 (enabled): Handle above.
11749 (define_cond_exec): Set predicated attribute to yes.
11751 2013-06-05 Mike Stump <mikestump@comcast.net>
11753 * gdbinit.in (__FUNCTION__): Add.
11755 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
11757 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
11758 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
11760 2013-06-05 Jan Hubicka <jh@suse.cz>
11762 * varasm.c (mark_decl_referenced): Revert the removal until targets
11765 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
11767 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
11768 instead of mark_decl_referenced.
11770 2013-06-05 Jan Hubicka <jh@suse.cz>
11772 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
11773 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
11774 and symtab_used_from_object_file_p.
11775 (cgraph_make_node_local_1): Clear forced_by_abi.
11776 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
11777 * cgraph.h (symtab_node_base): Add forced_by_abi.
11778 (decide_is_variable_needed): Remove.
11779 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
11780 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
11781 (decide_is_symbol_needed): ... this one; handle symbols in general;
11782 always analyze virtuals; honnor forced_by_abi.
11783 (cgraph_finalize_function): Update.
11784 (varpool_finalize_decl): Update.
11785 (symbol_defined_and_needed): Remove.
11786 (analyze_functions): Update.
11787 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
11788 output_refs, input_overwrite_node): Handle forced_by_abi.
11789 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
11790 (address_taken_from_non_vtable_p): ... this one.
11791 (comdat_can_be_unshared_p_1): New function.
11792 (cgraph_comdat_can_be_unshared_p): Rename to ...
11793 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
11794 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
11795 (function_and_variable_visibility): Clear forced_by_abi as needed.
11796 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
11797 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
11798 * symtab.c (dump_symtab_base): Dump forced_by_abi.
11799 * varpool.c (decide_is_variable_needed): Remove.
11801 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11803 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
11804 (arm_option_override): Override arm_restrict_it where appropriate.
11805 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
11806 * config/arm/arm.opt (mrestrict-it): New command-line option.
11807 * doc/invoke.texi: Document -mrestrict-it.
11809 2013-06-05 David Malcolm <dmalcolm@redhat.com>
11811 * tsan.c (tsan_atomic_table): Make const.
11813 2013-06-05 Richard Biener <rguenther@suse.de>
11815 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
11816 index associated with the tree we are supposed to replace.
11817 * tree-streamer-out.c (pack_ts_base_value_fields): Output
11818 TREE_ASM_WRITTEN as zero for everything but SSA names.
11820 2013-06-05 David Malcolm <dmalcolm@redhat.com>
11822 * tree-ssa-structalias.c (call_stmt_vars): Make static.
11824 2013-06-04 Jan Hubicka <jh@suse.cz>
11826 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
11827 (input_node, input_varpool_node): Handle correctly external same
11829 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
11830 nodes at ltrans stage.
11832 2013-06-04 Jan Hubicka <jh@suse.cz>
11834 * ipa-inline.c (update_caller_keys): Fix availability test.
11835 (update_callee_keys): Likewise.
11836 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
11837 to follow ELF standard.
11839 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
11841 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
11842 (mips64r5900el-*-elf*): New configurations.
11843 * config/mips/mips-cpus.def (r5900): New processor.
11844 * config/mips/mips-tables.opt: Regenerate.
11845 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
11846 (mips_issue_rate): Handle PROCESSOR_R5900.
11847 (mips_reorg_process_insns): Force reorder mode for the R5900.
11848 * config/mips/mips.h (TARGET_MIPS5900): Define.
11849 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
11851 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
11853 * config/mips/mips.md (processor): Add r5900.
11854 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
11856 2013-06-04 Ian Bolton <ian.bolton@arm.com>
11858 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
11859 into function to generate MOVI instruction.
11860 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
11861 (aarch64_preferred_simd_mode): Turn into wrapper.
11862 (aarch64_output_scalar_simd_mov_immediate): New function.
11863 * config/aarch64/aarch64-protos.h: Add prototype for above.
11865 2013-06-04 Ian Bolton <ian.bolton@arm.com>
11867 * config/aarch64/aarch64.c (simd_immediate_info): Remove
11868 element_char member.
11869 (sizetochar): Return signed char.
11870 (aarch64_simd_valid_immediate): Remove elchar and other
11871 unnecessary variables.
11872 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
11873 Calculate element_char as required.
11874 * config/aarch64/aarch64-protos.h: Update and move prototype
11875 for aarch64_output_simd_mov_immediate.
11876 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
11879 2013-06-04 Ian Bolton <ian.bolton@arm.com>
11881 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
11882 information completed by aarch64_simd_valid_immediate.
11883 (aarch64_legitimate_constant_p): Update arguments.
11884 (aarch64_simd_valid_immediate): Work with struct rather than many
11886 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
11887 (aarch64_simd_make_constant): Update arguments.
11888 (aarch64_output_simd_mov_immediate): Work with struct rather than
11889 many pointers. Output immediate directly rather than as operand.
11890 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
11892 * config/aarch64/constraints.md (Dn): Update arguments.
11894 2013-06-04 Ian Bolton <ian.bolton@arm.com>
11896 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
11898 (aarch64_simd_immediate_valid_for_move): Remove.
11899 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
11900 (aarch64_simd_make_constant): Update call.
11901 (aarch64_output_simd_mov_immediate): Update call.
11902 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
11904 * config/aarch64/constraints.md (Dn): Update call.
11906 2013-06-04 Ian Bolton <ian.bolton@arm.com>
11908 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
11909 return type to bool for prototype.
11910 (aarch64_legitimate_constant_p): Check for true instead of not -1.
11911 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
11912 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
11914 2013-06-04 Catherine Moore <clm@codesourcery.com>
11916 * config/mips/mips.opt (meva): New.
11917 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
11918 (ASM_SPEC): Handle -meva.
11919 * doc/invoke.texi (meva): Document.
11921 2013-06-04 Alan Modra <amodra@gmail.com>
11923 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
11926 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11928 * rtl.def: Add extra fourth optional field to define_cond_exec.
11929 * gensupport.c (process_one_cond_exec): Process attributes from
11931 * doc/md.texi: Document fourth field in define_cond_exec.
11933 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
11935 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
11936 out the processing order as in store_bit_field_1.
11938 2013-06-04 Jan Hubicka <jh@suse.cz>
11940 PR middle-end/57500
11941 * cgraphunit.c (cgraph_process_same_body_aliases): Create
11942 non-VAR_DECL node if it does not exist yet.
11944 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
11946 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
11947 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
11948 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
11949 target_cpu_default setting.
11951 2013-06-03 Teresa Johnson <tejohnson@google.com>
11953 * dumpfile.c (opt_info_switch_p): Change -fopt-info
11954 default to -fopt-info=optimized instead of all.
11955 * doc/invoke.texi: Ditto.
11956 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
11957 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
11958 (execute_vect_slp): Emit BB vectorization success under
11959 MSG_OPTIMIZED_LOCATIONS.
11960 * tree-vect-slp.c (vect_slp_transform_bb): Change
11961 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
11962 * tree-vect-loop.c (vect_transform_loop): Ditto.
11964 2013-06-03 Jason Merrill <jason@redhat.com>
11967 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
11968 Use TARGET_EXPR for C++.
11970 2013-06-03 Jakub Jelinek <jakub@redhat.com>
11972 PR rtl-optimization/57268
11973 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
11974 if DEBUG_INSN_P (insn).
11977 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
11979 PR rtl-optimization/57268
11980 * sched-deps.c (sched_analyze_2): Flush dependence lists if
11981 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
11983 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
11985 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
11986 (ix86_avoid_lea_for_addr): Likewise.
11987 (exact_dependency_1): Likewise.
11988 (ix86_adjust_cost): Likewise.
11989 (swap_top_of_ready_list): Fix formatting and !reload_completed check
11991 (do_reorder_for_imul): Fix typo, formatting and
11992 !reload_completed check removed.
11993 (ix86_sched_reorder): Fix typo and formatting.
11994 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
11997 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
11999 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
12001 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
12003 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
12004 <STRING_CST>: Likewise.
12005 <VECTOR_CST>: Likewise.
12007 2013-06-01 Janus Weil <janus@gcc.gnu.org>
12008 Mikael Morin <mikael@gcc.gnu.org>
12010 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
12011 * config.in: Regenerated.
12012 * configure: Regenerated.
12014 2013-06-01 Jan Hubicka <jh@suse.cz>
12016 PR middle-end/57366
12017 * cgraphunit.c (compile): When weakref is not supported,
12018 set up transparent aliases before final output pass.
12019 * varasm.c (assemble_alias): Do not try to do it here.
12021 2013-06-01 Jan Hubicka <jh@suse.cz>
12023 PR middle-end/57467
12024 * passes.c (for_per_function): Skip unanalyzed functions.
12026 2013-06-01 Jan Hubicka <jh@suse.cz>
12028 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
12029 (lto_symtab_merge_symbols_1): ... this one.
12030 (lto_symtab_merge_cgraph_nodes): Rename to ...
12031 (lto_symtab_merge_symbols): ... this one; simplify.
12032 * cgraph.c (same_body_aliases_done): Rename to ...
12033 (cpp_implicit_aliases_done): ... this one.
12034 (cgraph_create_function_alias): Update.
12035 (cgraph_same_body_alias): Update.
12036 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
12037 (verify_edge_corresponds_to_fndecl): Simplify.
12038 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
12039 (cgraph_node): Remove same_body_alias.
12040 (varpool_node): Remove alias_of and extra_name_alias.
12041 (same_body_aliases_done): Rename to ..
12042 (cpp_implicit_aliases_done): ... this one.
12043 (symtab_alias_ultimate_target): Add default parameter.
12044 (symtab_resolve_alias): New function.
12045 (fixup_same_cpp_alias_visibility): Declare.
12046 (cgraph_function_node): Add default parameter.
12047 (cgraph_node_asm_name): Likewise.
12048 (cgraph_function_or_thunk_node): Add default parameter; do
12049 not ICE when it is NULL.
12050 (varpool_variable_node): Likewise.
12051 * tree-emutls.c (create_emultls_var): Update.
12052 (ipa_lower_emutls): Update.
12053 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
12054 (cgraph_reset_node): Reset alias info.
12055 (cgraph_finalize_function): Update.
12056 (fixup_same_cpp_alias_visibility): Move to symtab.c.
12057 (analyze_function): Simplify.
12058 (cgraph_process_same_body_aliases): Simplify.
12059 (analyze_functions): Fixup same body aliases.
12060 (handle_alias_pairs): Simplify.
12061 (assemble_thunk): Update.
12062 (assemble_thunks_and_aliases): Update.
12063 (output_weakrefs): Rewrite.
12064 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
12065 (lto_output_varpool_node): Likewise.
12066 (compute_ltrans_boundary): Remve assert.
12067 (get_alias_symbol): New functoin.
12068 (input_node): Rewrite alias handling.
12069 (input_varpool_node): Likewise.
12070 * ipa-pure-const.c (propagate_pure_const): Fix formating.
12071 * ipa.c (process_references): Handle weakrefs correctly.
12072 (symtab_remove_unreachable_nodes): Likewise.
12073 * trans-mem.c (get_cg_data): Update.
12074 (ipa_tm_create_version_alias): Update.
12075 (ipa_tm_execute): Update.
12076 * symtab.c (dump_symtab_base): Dump aliases.
12077 (verify_symtab_base): Verify aliases.
12078 (symtab_node_availability): New function.
12079 (symtab_alias_ultimate_target): Simplify.
12080 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
12081 handle all the fixup cases.
12082 (symtab_resolve_alias): New function.
12083 * passes.c (ipa_write_summaries): Handle weakrefs.
12084 * varpool.c (varpool_analyze_node): Simplify.
12085 (assemble_aliases): Update.
12086 (varpool_create_variable_alias): Simplify.
12087 (varpool_extra_name_alias): Simplify.
12088 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
12089 (lto_symtab_merge_symbols): ... this one.
12091 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
12094 PR rtl-optimization/57268
12095 * sched-deps.c (sched_analyze_2): Flush dependence lists if
12096 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
12098 2013-06-01 Tobias Burnus <burnus@net-b.de>
12100 Partially reverted:
12101 2013-05-31 Tobias Burnus <burnus@net-b.de>
12103 PR middle-end/57073
12104 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
12107 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
12109 PR rtl-optimization/57268
12110 * sched-deps.c (sched_analyze_2): Flush dependence lists if
12111 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
12113 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
12115 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
12116 unordered comparison operators when -fno-trapping-math is in effect
12118 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
12119 and implement unordered comparison operators properly on the e500.
12121 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
12123 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
12124 for constant scalar integers.
12125 (simplify_relational_operation_1): Likewise.
12127 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
12129 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
12130 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
12133 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
12134 Igor Zamyatin <igor.zamyatin@intel.com>
12136 Silvermont (SLM) architecture performance tuning.
12137 * config/i386/i386.h (enum ix86_tune_indices): Add
12138 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
12139 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
12141 * config/i386/i386.c (initial_ix86_tune_features)
12142 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
12143 (ix86_lea_outperforms): Handle Silvermont tuning.
12144 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
12146 (ix86_use_lea_for_mov): Likewise.
12147 (ix86_avoid_lea_for_addr): Likewise.
12148 (ix86_lea_for_add_ok): Likewise.
12149 (exact_dependency_1): New function.
12150 (exact_store_load_dependency): Likewise.
12151 (ix86_adjust_cost): Handle Silvermont tuning.
12152 (do_reoder_for_imul): Likewise.
12153 (swap_top_of_ready_list): New function.
12154 (ix86_sched_reorder): Changed to handle Silvermont tuning.
12156 * config/i386/i386.md (peepholes that split memory operand in fp
12159 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
12161 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
12162 Remove un-necessary braces.
12164 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
12166 * config/aarch64/aarch64.c (aarch64_classify_symbol):
12167 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
12169 2013-05-31 Tobias Burnus <burnus@net-b.de>
12171 PR middle-end/57073
12172 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
12174 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12177 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
12178 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
12179 * config/arm/neon.md (iordi3_neon): Remove.
12180 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
12181 * config/arm/predicates.md (imm_for_neon_logic_operand):
12182 Move to earlier in the file.
12183 (neon_logic_op2): Likewise.
12184 (arm_iordi_operand_neon): New predicate.
12186 2013-05-31 Richard Biener <rguenther@suse.de>
12188 PR tree-optimization/57478
12189 PR tree-optimization/57453
12190 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
12193 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
12195 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
12196 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
12198 2013-05-30 Tobias Burnus <burnus@net-b.de>
12199 Thomas Koenig <tkoenig@gcc.gnu.org>
12201 PR middle-end/57073
12202 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
12203 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
12205 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
12207 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
12209 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
12211 * target.def (register_usage_leveling_p): New hook.
12212 * targhooks.c (default_register_usage_leveling_p): New.
12213 * targhooks.h (default_register_usage_leveling_p): New prototype.
12214 * lra-assigns.c (register_usage_leveling_p): Use the hook.
12215 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
12216 * doc/tm.texi: Update.
12217 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
12219 2013-05-30 Ian Bolton <ian.bolton@arm.com>
12221 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
12222 (*insv_reg<mode>): New define_insn.
12224 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
12226 PR rtl-optimization/57439
12227 * postreload.c (move2add_valid_value_p): Check that we have
12228 a zero subreg_regno_offset when accessing the register in
12229 the requested mode.
12231 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
12232 Igor Zamyatin <igor.zamyatin@intel.com>
12234 Silvermont (SLM) architecture pipeline model, tuning and
12236 * config.gcc: Add slm config options and target.
12238 * config/i386/slm.md: New.
12240 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
12242 * config/i386/i386-c.c (ix86_target_macros_internal): New case
12244 (ix86_target_macros_internal): Likewise.
12246 * config/i386/i386.c (slm_cost): New cost.
12247 (m_SLM): New macro flag.
12248 (initial_ix86_tune_features): Set m_SLM.
12249 (x86_accumulate_outgoing_args): Likewise.
12250 (x86_arch_always_fancy_math_387): Likewise.
12251 (processor_target_table): Add slm cost.
12252 (cpu_names): Add slm cpu name.
12253 (x86_option_override_internal): Set SLM ISA.
12254 (ix86_issue_rate): New case PROCESSOR_SLM.
12255 (ia32_multipass_dfa_lookahead): Likewise.
12256 (fold_builtin_cpu): Add slm.
12258 * config/i386/i386.h (TARGET_SLM): New target macro.
12259 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
12260 (processor_type): Add PROCESSOR_SLM.
12262 * config/i386/i386.md (cpu): Add new value "slm".
12263 (slm.md): Include slm.md.
12265 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
12266 Zhenqiang Chen <zhenqiang.chen@linaro.org>
12268 * config/arm/arm-protos.h: Add and update function protos.
12269 * config/arm/arm.c (use_simple_return_p): New added.
12270 (thumb2_expand_return): Check simple_return flag.
12271 * config/arm/arm.md: Add simple_return and conditional simple_return.
12272 * config/arm/iterators.md: Add iterator for return and simple_return.
12274 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
12276 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
12277 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
12278 (arm_emit_vfp_multi_reg_pop): Likewise.
12279 (thumb2_emit_ldrd_pop): Likewise.
12280 (arm_expand_epilogue): Add misc REG_CFA notes.
12281 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
12283 2013-05-29 Lawrence Crowl <crowl@google.com>
12285 * config/arm/t-arm: Update for below.
12287 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
12288 Change type to hash_table. Update dependent calls and types.
12290 * config/i386/t-cygming: Update for below.
12292 * config/i386/t-interix: Update for below.
12294 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
12295 Change type to hash_table. Update dependent calls and types.
12296 (i386_find_on_wrapper_list::wrappers): Likewise.
12298 * config/ia64/t-ia64: Update for below.
12300 * config/ia64/ia64.c (bundle_state_table):
12301 Change type to hash_table. Update dependent calls and types.
12303 * config/mips/mips.c (mips_reorg_process_insns::htab):
12304 Change type to hash_table. Update dependent calls and types.
12306 * config/sol2.c (solaris_comdat_htab):
12307 Change type to hash_table. Update dependent calls and types.
12309 * config/t-sol2: Update for above.
12311 2013-05-29 Teresa Johnson <tejohnson@google.com>
12313 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
12314 functions are not yet marked as defined.
12316 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
12317 Pat Haugen <pthaugen@us.ibm.com>
12318 Peter Bergner <bergner@vnet.ibm.com>
12320 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
12324 (vrotl<mode>3): Likewise.
12325 (vashl<mode>3): Likewise.
12326 (vlshr<mode>3): Likewise.
12327 (vashr<mode>3): Likewise.
12329 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
12330 support for power8 V2DI builtins.
12332 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
12333 power8 V2DI builtins.
12334 (vupkhsw): Likewise.
12335 (vupklsw): Likewise.
12336 (vaddudm): Likewise.
12337 (vminsd): Likewise.
12338 (vmaxsd): Likewise.
12339 (vminud): Likewise.
12340 (vmaxud): Likewise.
12341 (vpkudum): Likewise.
12342 (vpksdss): Likewise.
12343 (vpkudus): Likewise.
12344 (vpksdus): Likewise.
12349 (vsubudm): Likewise.
12350 (vcmpequd): Likewise.
12351 (vcmpgtsd): Likewise.
12352 (vcmpgtud): Likewise.
12353 (vcmpequd_p): Likewise.
12354 (vcmpgtsd_p): Likewise.
12355 (vcmpgtud_p): Likewise.
12356 (vupkhsw): Likewise.
12357 (vupklsw): Likewise.
12358 (vaddudm): Likewise.
12359 (vmaxsd): Likewise.
12360 (vmaxud): Likewise.
12361 (vminsd): Likewise.
12362 (vminud): Likewise.
12363 (vpksdss): Likewise.
12364 (vpksdus): Likewise.
12365 (vpkudum): Likewise.
12366 (vpkudus): Likewise.
12371 (vsubudm): Likewise.
12373 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
12374 support for power8 V2DI instructions.
12376 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
12377 power8 V2DI instructions. Combine pack and unpack insns to use an
12378 iterator for each mode. Check whether a particular mode supports
12379 Altivec instructions instead of just checking TARGET_ALTIVEC.
12380 (UNSPEC_VPKUWUM): Likewise.
12381 (UNSPEC_VPKSHSS): Likewise.
12382 (UNSPEC_VPKSWSS): Likewise.
12383 (UNSPEC_VPKUHUS): Likewise.
12384 (UNSPEC_VPKSHUS): Likewise.
12385 (UNSPEC_VPKUWUS): Likewise.
12386 (UNSPEC_VPKSWUS): Likewise.
12387 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
12388 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
12389 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
12390 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
12391 (UNSPEC_VUPKHSB): Likewise.
12392 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
12393 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
12394 (UNSPEC_VUPKHSH): Likewise.
12395 (UNSPEC_VUPKLSB): Likewise.
12396 (UNSPEC_VUPKLSH): Likewise.
12398 (VI_char): Likewise.
12399 (VI_scalar): Likewise.
12400 (VI_unit): Likewise.
12402 (VP_small): Likewise.
12403 (VP_small_lc): Likewise.
12404 (VU_char): Likewise.
12405 (add<mode>3): Likewise.
12406 (altivec_vaddcuw): Likewise.
12407 (altivec_vaddu<VI_char>s): Likewise.
12408 (altivec_vadds<VI_char>s): Likewise.
12409 (sub<mode>3): Likewise.
12410 (altivec_vsubcuw): Likewise.
12411 (altivec_vsubu<VI_char>s): Likewise.
12412 (altivec_vsubs<VI_char>s): Likewise.
12413 (altivec_vavgs<VI_char>): Likewise.
12414 (altivec_vcmpbfp): Likewise.
12415 (altivec_eq<mode>): Likewise.
12416 (altivec_gt<mode>): Likewise.
12417 (altivec_gtu<mode>): Likewise.
12418 (umax<mode>3): Likewise.
12419 (smax<mode>3): Likewise.
12420 (umin<mode>3): Likewise.
12421 (smin<mode>3): Likewise.
12422 (altivec_vpkuhum): Likewise.
12423 (altivec_vpkuwum): Likewise.
12424 (altivec_vpkshss): Likewise.
12425 (altivec_vpkswss): Likewise.
12426 (altivec_vpkuhus): Likewise.
12427 (altivec_vpkshus): Likewise.
12428 (altivec_vpkuwus): Likewise.
12429 (altivec_vpkswus): Likewise.
12430 (altivec_vpks<VI_char>ss): Likewise.
12431 (altivec_vpks<VI_char>us): Likewise.
12432 (altivec_vpku<VI_char>us): Likewise.
12433 (altivec_vpku<VI_char>um): Likewise.
12434 (altivec_vrl<VI_char>): Likewise.
12435 (altivec_vsl<VI_char>): Likewise.
12436 (altivec_vsr<VI_char>): Likewise.
12437 (altivec_vsra<VI_char>): Likewise.
12438 (altivec_vsldoi_<mode>): Likewise.
12439 (altivec_vupkhsb): Likewise.
12440 (altivec_vupkhs<VU_char>): Likewise.
12441 (altivec_vupkls<VU_char>): Likewise.
12442 (altivec_vupkhsh): Likewise.
12443 (altivec_vupklsb): Likewise.
12444 (altivec_vupklsh): Likewise.
12445 (altivec_vcmpequ<VI_char>_p): Likewise.
12446 (altivec_vcmpgts<VI_char>_p): Likewise.
12447 (altivec_vcmpgtu<VI_char>_p): Likewise.
12448 (abs<mode>2): Likewise.
12449 (vec_unpacks_hi_v16qi): Likewise.
12450 (vec_unpacks_hi_v8hi): Likewise.
12451 (vec_unpacks_lo_v16qi): Likewise.
12452 (vec_unpacks_hi_<VP_small_lc>): Likewise.
12453 (vec_unpacks_lo_v8hi): Likewise.
12454 (vec_unpacks_lo_<VP_small_lc>): Likewise.
12455 (vec_pack_trunc_v8h): Likewise.
12456 (vec_pack_trunc_v4si): Likewise.
12457 (vec_pack_trunc_<mode>): Likewise.
12459 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
12461 (vec_vmaxsd): Likewise.
12462 (vec_vmaxud): Likewise.
12463 (vec_vminsd): Likewise.
12464 (vec_vminud): Likewise.
12465 (vec_vpksdss): Likewise.
12466 (vec_vpksdus): Likewise.
12467 (vec_vpkudum): Likewise.
12468 (vec_vpkudus): Likewise.
12469 (vec_vrld): Likewise.
12470 (vec_vsld): Likewise.
12471 (vec_vsrad): Likewise.
12472 (vec_vsrd): Likewise.
12473 (vec_vsubudm): Likewise.
12474 (vec_vupkhsw): Likewise.
12475 (vec_vupklsw): Likewise.
12477 2013-05-29 Jan Hubicka <jh@suse.cz>
12479 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
12480 flags; reorder rest of fields in more consistent way.
12481 (varpool_node): Remove analyzed, finalized and alias.
12482 (cgraph_ndoe): Likewise.
12483 (symtab_alias_ultimate_target): New function.
12484 (cgraph_function_node): Move offline.
12485 (cgraph_reset_node): Declare.
12486 (cgraph_comdat_can_be_unshared_p): Remove.
12487 (varpool_remove_initializer): Declare.
12488 (varpool_first_defined_variable, varpool_next_defined_variable
12489 cgraph_first_defined_function, cgraph_next_defined_function): Update.
12490 (cgraph_function_with_gimple_body_p): Update.
12491 (varpool_all_refs_explicit_p): Update.
12492 (symtab_alias_target): New function.
12493 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
12494 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
12495 (cgraph_function_or_thunk_node): Simplify using
12496 symtab_alias_ultimate_target.
12497 (varpool_variable_node): Likewise.
12498 * cgraph.c (cgraph_create_function_alias): Update.
12499 (cgraph_add_thunk): Update.
12500 (cgraph_remove_node): Update.
12501 (dump_cgraph_node): Do not dump removed flags.
12502 (cgraph_function_body_availability): Update.
12503 (cgraph_propagate_frequency): Update.
12504 (verify_cgraph_node): Check sanity of local flag.
12505 (cgraph_function_node): Move here from cgraph.h; revamp for
12506 cgraph_function_or_thunk_node.
12507 * lto-symtab.c (lto_varpool_replace_node): Update.
12508 (lto_symtab_resolve_can_prevail_p): Update.
12509 (lto_symtab_merge_cgraph_nodes): Update.
12510 * ipa-cp.c (determine_versionability, initialize_node_lattices,
12511 propagate_constants_accross_call, devirtualization_time_bonus,
12512 ipcp_propagate_stage): Update.
12513 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
12514 * ipa-inline-transform.c (clone_inlined_nodes,
12515 preserve_function_body_p): Update.
12516 * ipa-reference.c (propagate): Update.
12517 (write_node_summary_p): Update.
12518 * toplev.c (wrapup_global_declaration_2): Update.
12519 * cgraphunit.c (cgraph_analyze_function): Rename to ...
12520 (analyze_function) ... this one.
12521 (cgraph_process_new_functions): Update.
12522 (cgraph_reset_node): Export.
12523 (cgraph_finalize_function): Update.
12524 (cgraph_add_new_function): Update.
12525 (process_function_and_variable_attributes): Update.
12526 (varpool_finalize_decl): Update.
12527 (symbol_finalized): Remove.
12528 (symbol_finalized_and_needed): Rename to ...
12529 (symbol_defined_and_needed): ... update.
12530 (cgraph_analyze_functions): Update.
12531 (handle_alias_pairs): Update.
12532 (mark_functions_to_output): Update.
12533 (assemble_thunk): Update.
12534 (output_in_order): Update.
12535 (output_weakrefs): Update.
12536 (finalize_compilation_unit): Update.
12537 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
12538 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
12539 input_node, input_varpool_node): Update.
12540 * dbxout.c (dbxout_expand_expr): Update.
12541 * cgraphclones.c (cgraph_clone_node): Update.
12542 (cgraph_copy_node_for_versioning): Update.
12543 (cgraph_materialize_clone): Update.
12544 (cgraph_materialize_all_clones): Update.
12545 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
12546 propagate_pure_const, propagate_nothrow): Update.
12547 * lto-streamer-out.c (lto_output, write_symbol): Update.
12548 * ipa-utils.c (ipa_reverse_postorder): Update.
12549 * ipa-inline.c (can_inline_edge_p): Update.
12550 (update_caller_keys, ipa_inline): Update.
12551 * dwarf2out.c (reference_to_unused,
12552 premark_types_used_by_global_vars_helper): Update.
12553 * tree-eh.c (tree_could_trap_p): Update.
12554 * ipa-split.c (consider_split, execute_split_functions): Update.
12555 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
12556 has_addr_references_p): Update; move ahead in file for better
12558 (process_references): Simplify.
12559 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
12560 bodies are removed.
12561 (cgraph_comdat_can_be_unshared_p): Make static.
12562 (cgraph_externally_visible_p): Update.
12563 (varpool_externally_visible_p): Update.
12564 (function_and_variable_visibility): Update.
12565 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
12566 ipa_tm_mark_force_output_node): Update.
12567 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
12568 estimate_edge_devirt_benefit, inline_generate_summary,
12569 inline_write_summary): Update.
12570 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
12571 * ipa-prop.c (ipa_compute_jump_functions): Update.
12572 (ipa_print_node_params, ipa_prop_read_section,
12573 ipa_update_after_lto_read, read_replacements_section): Update.
12574 * varasm.c (mark_decl_referenced): Update.
12575 (assemble_alias, dump_tm_clone_pairs): Update.
12576 * tree-inline.c (copy_bb): Update.
12577 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
12579 * symtab.c (dump_symtab_base): Print new flags.
12580 (verify_symtab_base): Verify new flags.
12581 (symtab_alias_ultimate_target): New function.
12582 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
12583 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
12585 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
12587 * i386.c (ix86_get_function_versions_dispatcher,
12588 ix86_generate_version_dispatcher_body): Update.
12589 (fold_builtin_cpu): Use varpool_add_new_variable.
12590 * varpool.c (varpool_remove_initializer): Break out from ...
12591 (varpool_remove_node): ... this one.
12592 (dump_varpool_node, varpool_node_for_asm,
12593 cgraph_variable_initializer_availability, varpool_analyze_node,
12594 varpool_assemble_decl, varpool_remove_unreferenced_decls,
12595 varpool_finalize_named_section_flags, varpool_create_variable_alias):
12598 2013-05-29 Jan Hubicka <jh@suse.cz>
12600 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
12602 2013-05-29 Easwaran Raman <eraman@google.com>
12604 PR tree-optimization/57442
12605 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
12606 when control exits the main loop.
12608 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
12610 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
12612 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
12613 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
12614 * rx/t-rx: Add rx100 under multi library matches option for nofpu
12617 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12619 PR tree-optimization/57441
12620 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
12621 Don't limit size of incr_vec to number of candidates.
12623 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
12625 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
12626 and mips16 directories.
12627 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
12628 (MULTILIB_DIRNAMES): Ditto.
12629 (MULTILIB_EXCEPTIONS): Add new exceptions.
12630 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
12631 (MULTILIB_DIRNAMES): Ditto.
12632 (MULTILIB_EXCEPTIONS): Add new exceptions.
12634 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
12635 Marcus Shawcroft <marcus.shawcroft@arm.com>
12637 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
12638 SYMBOL_TINY_ABSOLUTE.
12639 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
12640 SYMBOL_TINY_ABSOLUTE.
12641 (aarch64_expand_mov_immediate): Likewise.
12642 (aarch64_classify_symbol): Likewise.
12643 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
12644 Permit SYMBOL_TINY_ABSOLUTE.
12645 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
12647 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
12648 Marcus Shawcroft <marcus.shawcroft@arm.com>
12650 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
12651 Refactor if/switch. Replace gcc_assert with if.
12653 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
12655 * config/i386/i386.c (initial_ix86_tune_features): Enable
12656 FP Reassociation for AMD bdver1 and bdver2.
12658 2013-05-29 Martin Jambor <mjambor@suse.cz>
12660 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
12661 and IMAGPART_EXPR do not occur within other handled_components.
12663 2013-05-29 Richard Biener <rguenther@suse.de>
12665 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
12666 access on whether the use is in the BB we currently try to
12668 (vect_bb_vectorization_profitable_p): Pass the BB we currently
12669 vectorize to vect_bb_slp_scalar_cost.
12671 2013-05-29 Richard Biener <rguenther@suse.de>
12673 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
12674 computing scalar cost offsetted by stmts that are kept live
12676 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
12677 for computation of scalar cost.
12679 2013-05-28 Steve Ellcey <sellcey@mips.com>
12681 * config/mips/mips-cpus.def (mips32r2): Change processor type.
12683 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
12685 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
12686 array notation built-in reduction functions.
12687 * doc/passes.texi (Passes): Added documentation about changes done
12689 * doc/invoke.texi (C Dialect Options): Added documentation about
12690 the -fcilkplus flag.
12691 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
12692 (BUILTINS_DEF): Depend on cilkplus.def.
12693 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
12694 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
12695 * cilkplus.def: New file.
12697 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
12699 PR rtl-optimization/57439
12700 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
12702 2013-05-28 Easwaran Raman <eraman@google.com>
12704 PR tree-optimization/57337
12705 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
12706 (find_insert_point): Correctly identify the insertion point
12707 when two statements with the same UID is compared.
12709 2013-05-28 Richard Biener <rguenther@suse.de>
12711 PR tree-optimization/56787
12712 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
12713 from the list of data references.
12714 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
12716 (vect_analyze_loop_operations): Likewise.
12717 (vect_transform_loop): Remove clobbers.
12719 2013-05-28 Martin Jambor <mjambor@suse.cz>
12721 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
12722 and REALPART_EXPRs have scalar type.
12724 2013-05-28 Richard Biener <rguenther@suse.de>
12726 PR tree-optimization/57411
12727 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
12729 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
12730 virtual operand propagation.
12732 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
12734 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
12735 destination register for bmasksi_vis.
12736 (vector_init_bshuffle): Likewise.
12737 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
12739 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
12741 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
12742 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
12743 mode if the instruction isn't available in the original mode.
12744 * config/sparc/sparc.opt (mfix-ut699): New option.
12745 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
12746 (divdf3): Turn into expander.
12747 (divdf3_nofix): New insn.
12748 (divdf3_fix): Likewise.
12749 (divsf3): Disable if -mfix-ut699.
12750 (sqrtdf2): Turn into expander.
12751 (sqrtdf2_nofix): New insn.
12752 (sqrtdf2_fix): Likewise.
12753 (sqrtsf2): Disable if -mfix-ut699.
12755 2013-05-27 Richard Biener <rguenther@suse.de>
12757 PR middle-end/57412
12758 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
12759 block for the new loop.
12761 2013-05-27 Richard Biener <rguenther@suse.de>
12763 PR tree-optimization/57343
12764 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
12765 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
12766 (number_of_iterations_cond): Do not build the folded tree.
12768 2013-05-27 Richard Biener <rguenther@suse.de>
12771 PR middle-end/57381
12772 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
12773 OEP_CONSTANT_ADDRESS_OF retained.
12775 PR tree-optimization/57417
12776 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
12777 for unchanged base.
12778 (set_ssa_val_to): Compare addresses using
12779 get_addr_base_and_unit_offset.
12781 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
12783 PR rtl-optimization/56833
12784 * postreload.c (move2add_record_mode): New function.
12785 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
12786 (move2add_use_add2_insn): Use move2add_record_sym_value.
12787 (move2add_use_add3_insn): Likewise.
12788 (reload_cse_move2add): Use move2add_valid_value_p and
12789 move2add_record_mode. Invalidate call-clobbered and REG_INC
12790 affected regs by setting reg_mode to VOIDmode.
12791 (move2add_note_store): Don't pretend the inside of a SUBREG is
12792 the actual destination. Invalidate single/leading registers by
12793 setting reg_mode to VOIDmode.
12794 Use move2add_record_sym_value, move2add_valid_value_p and
12795 move2add_record_mode.
12797 2013-05-27 Richard Biener <rguenther@suse.de>
12799 PR tree-optimization/57396
12800 * tree-affine.c (double_int_constant_multiple_p): Properly
12801 return false for val == 0 and div != 0.
12803 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
12805 * config/mips/mips.h: Use #elif in preprocessor conditions.
12807 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
12810 * config/mips/constraints.md (kl): New constraint.
12811 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
12812 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
12813 constraint for operand 0. Split after CSE for MIPS16. Emit a move
12814 from LO for MIPS16.
12815 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
12817 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
12820 * config/mips/mips.c (mips_can_inline_p): New function.
12821 (TARGET_CAN_INLINE_P): Define.
12823 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
12825 * sched-int.h (ds_t, dw_t): Make unsigned int.
12826 Fix documentation that describes how all the ds_t bits are used.
12827 Reserve the last bit for delayed-branch scheduling.
12828 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
12829 (BITS_PER_DEP_WEAK): Fix definition and documentation.
12830 (gen_dep_weak_1): Remove prototype.
12831 * sched-deps.c (get_dep_weak_1): Make static.
12832 * target.def (speculate_insn, needs_block_p, gen_spec_check,
12833 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
12834 * doc/tm.texi: Regenerate.
12835 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
12837 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
12840 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
12842 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
12843 Sandra Loosemore <sandra@codesourcery.com>
12845 * config.gcc (powerpc-*): Allow native for with-cpu.
12847 2013-05-24 Jeff Law <law@redhat.com>
12849 PR tree-optimization/57124
12850 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
12851 conversion feeding a condition if the range has an overflow
12852 if -fstrict-overflow. Add warnings for when we do make the
12855 2013-05-24 Dehao Chen <dehao@google.com>
12857 * tree-cfg.c (locus_discrim_map): Fix the typo.
12858 (locus_discrim_hasher): Likewise.
12859 (locus_discrim_hasher::hash): Likewise.
12860 (locus_discrim_hasher::equal): Likewise.
12862 2013-05-24 Martin Jambor <mjambor@suse.cz>
12864 PR tree-optimization/57294
12865 * cgraph.h (ipa_record_stmt_references): Declare.
12866 * cgraphbuild.c (ipa_record_stmt_references): New function.
12867 (build_cgraph_edges): Use ipa_record_stmt_references.
12868 (rebuild_cgraph_edges): Likewise.
12869 (cgraph_rebuild_references): Likewise.
12870 * ipa-prop.c (ipa_modify_call_arguments): Discard references
12871 associated with the old statement and build references from the
12872 newly built statements.
12873 * ipa-ref.c (ipa_remove_stmt_references): New function.
12874 * ipa-ref.h (ipa_remove_stmt_references): Declare.
12876 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
12878 * lra-constraints.c (emit_spill_move): Use smaller mode for
12880 (check_and_process_move): Consider mem-reg moves for secondary
12882 (curr_insn_transform): Don't lose insns emitted before for
12883 secondary memory moves.
12884 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
12885 reg set up in the current insn.
12887 2013-05-24 Dehao Chen <dehao@google.com>
12889 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
12891 (locus_descrim_hasher::equal): Likewise.
12892 (build_gimple_cfg): New discriminator assignment algorithm.
12893 (make_edges): Likewise.
12894 (next_discriminator_for_locus): Likewise.
12895 (same_line_p): Likewise.
12896 (assign_discriminators): Likewise.
12897 (make_cond_expr_edges): Likewise.
12898 (make_gimple_switch_edges): Likewise.
12899 (make_goto_expr_edges): Likewise.
12900 (make_gimple_asm_edges): Likewise.
12902 2013-05-24 Ian Bolton <ian.bolton@arm.com>
12904 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
12905 X format specifier to only display bottom 16 bits.
12906 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
12907 immediate to match for operand 2, since it will be masked.
12909 2013-05-24 Richard Biener <rguenther@suse.de>
12911 PR tree-optimization/57287
12912 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
12913 all SSA names that occur in abnormal PHIs.
12915 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
12918 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
12919 that index is not negative.
12921 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
12923 PR rtl-optimization/55177
12924 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
12925 (simplify_byte_swapping_operation): New.
12926 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
12927 (simplify_relational_operation_1): Deal with BSWAP.
12929 2013-05-23 Richard Henderson <rth@redhat.com>
12932 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
12933 (ix86_reorg): Call it.
12935 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
12938 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
12939 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
12940 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
12942 2013-05-23 Christian Bruel <christian.bruel@st.com>
12945 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
12947 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
12948 Marcus Shawcroft <marcus.shawcroft@arm.com>
12950 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
12951 * config/aarch64/constraints.md (Usa): Remove.
12952 * doc/md.texi (AArch64 Usa): Remove.
12954 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
12955 Marcus Shawcroft <marcus.shawcroft@arm.com>
12957 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
12958 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
12959 * config/aarch64/predicates.md (aarch64_const_address): Remove.
12960 (aarch64_mov_operand): Use aarch64_mov_operand_p.
12962 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
12964 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
12965 instruction (AdvSIMD).
12966 * config/aarch64/aarch64-builtins.c
12967 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
12968 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
12970 2013-05-23 Martin Jambor <mjambor@suse.cz>
12972 PR middle-end/57347
12973 * tree.h (contains_bitfld_component_ref_p): Declare.
12974 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
12975 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
12977 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
12978 not access a bit-field. Assert all final offsets are byte-aligned.
12980 2013-05-23 Richard Biener <rguenther@suse.de>
12982 PR tree-optimization/57380
12983 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
12984 least one invariant or re-used load.
12985 * passes.c (init_optimization_passes): Move pass_phiprop before
12988 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
12990 * config/aarch64/aarch64-simd.md
12991 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
12993 2013-05-23 Richard Biener <rguenther@suse.de>
12995 PR middle-end/57381
12996 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
12997 OEP_CONSTANT_ADDRESS_OF retained.
12999 2013-05-23 Jakub Jelinek <jakub@redhat.com>
13001 PR middle-end/57344
13002 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
13003 don't lower unit. Handle unit not being always BITS_PER_WORD.
13005 2013-05-23 Richard Biener <rguenther@suse.de>
13007 PR rtl-optimization/57341
13008 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
13009 instead of true_dependence.
13011 2013-05-22 David Malcolm <dmalcolm@redhat.com>
13013 * bb-reorder.c (branch_threshold): Make const.
13014 (exec_threshold): Ditto.
13016 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
13017 Pat Haugen <pthaugen@us.ibm.com>
13018 Peter Bergner <bergner@vnet.ibm.com>
13020 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
13021 documentation for the power8 crypto builtins.
13023 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
13025 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
13026 macros for defining power8 builtin functions.
13027 (BU_P8V_AV_2): Likewise.
13028 (BU_P8V_AV_P): Likewise.
13029 (BU_P8V_VSX_1): Likewise.
13030 (BU_P8V_OVERLOAD_1): Likewise.
13031 (BU_P8V_OVERLOAD_2): Likewise.
13032 (BU_CRYPTO_1): Likewise.
13033 (BU_CRYPTO_2): Likewise.
13034 (BU_CRYPTO_3): Likewise.
13035 (BU_CRYPTO_OVERLOAD_1): Likewise.
13036 (BU_CRYPTO_OVERLOAD_2): Likewise.
13037 (XSCVSPDP): Fix typo, point to the correct instruction.
13038 (VCIPHER): Add power8 crypto builtins.
13039 (VCIPHERLAST): Likewise.
13040 (VNCIPHER): Likewise.
13041 (VNCIPHERLAST): Likewise.
13042 (VPMSUMB): Likewise.
13043 (VPMSUMH): Likewise.
13044 (VPMSUMW): Likewise.
13045 (VPERMXOR_V2DI): Likewise.
13046 (VPERMXOR_V4SI: Likewise.
13047 (VPERMXOR_V8HI: Likewise.
13048 (VPERMXOR_V16QI: Likewise.
13049 (VSHASIGMAW): Likewise.
13050 (VSHASIGMAD): Likewise.
13051 (VPMSUM): Likewise.
13052 (VPERMXOR): Likewise.
13053 (VSHASIGMA): Likewise.
13055 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
13056 __CRYPTO__ if the crypto instructions are available.
13057 (altivec_overloaded_builtins): Add support for overloaded power8
13060 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
13061 support for power8 crypto builtins.
13062 (builtin_function_type): Likewise.
13063 (altivec_init_builtins): Add support for builtins that take vector
13064 long long (V2DI) arguments.
13066 * config/rs6000/crypto.md: New file, define power8 crypto
13069 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
13070 Pat Haugen <pthaugen@us.ibm.com>
13071 Peter Bergner <bergner@vnet.ibm.com>
13073 * doc/invoke.texi (Option Summary): Add power8 options.
13074 (RS/6000 and PowerPC Options): Likewise.
13076 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
13077 constraints.md instead of rs6000.h. Reorder w* constraints. Add
13078 wm, wn, wr documentation.
13080 * config/rs6000/constraints.md (wm): New constraint for VSX
13081 registers if direct move instructions are enabled.
13082 (wn): New constraint for no registers.
13083 (wq): New constraint for quad word even GPR registers.
13084 (wr): New constraint if 64-bit instructions are enabled.
13085 (wv): New constraint if power8 vector instructions are enabled.
13086 (wQ): New constraint for quad word memory locations.
13088 * config/rs6000/predicates.md (const_0_to_15_operand): New
13089 constraint for 0..15 for crypto instructions.
13090 (gpc_reg_operand): If VSX allow registers in VSX registers as well
13091 as GPR and floating point registers.
13092 (int_reg_operand): New predicate to match only GPR registers.
13093 (base_reg_operand): New predicate to match base registers.
13094 (quad_int_reg_operand): New predicate to match even GPR registers
13095 for quad memory operations.
13096 (vsx_reg_or_cint_operand): New predicate to allow vector logical
13097 operations in both GPR and VSX registers.
13098 (quad_memory_operand): New predicate for quad memory operations.
13099 (reg_or_indexed_operand): New predicate for direct move support.
13101 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
13102 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
13103 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
13104 (POWERPC_MASKS): Add power8 options.
13105 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
13108 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
13109 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
13111 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
13112 (-mpower8-fusion): New power8 options.
13113 (-mpower8-fusion-sign): Likewise.
13114 (-mpower8-vector): Likewise.
13115 (-mcrypto): Likewise.
13116 (-mdirect-move): Likewise.
13117 (-mquad-memory): Likewise.
13119 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
13120 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
13122 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
13123 (rs6000_debug_vector_unit): Add p8_vector.
13124 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
13125 definitions. Also print fusion state.
13126 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
13127 (rs6000_builtin_mask_calculate): Add power8 builtin support.
13128 (rs6000_option_override_internal): Add support for power8.
13129 (rs6000_common_init_builtins): Add debugging for skipped builtins
13130 if -mdebug=builtin.
13131 (rs6000_adjust_cost): Add power8 support.
13132 (rs6000_issue_rate): Likewise.
13133 (insn_must_be_first_in_group): Likewise.
13134 (insn_must_be_last_in_group): Likewise.
13135 (force_new_group): Likewise.
13136 (rs6000_register_move_cost): Likewise.
13137 (rs6000_opt_masks): Likewise.
13139 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
13140 power8 capable assembler, default to power7 options.
13141 (TARGET_DIRECT_MOVE): Likewise.
13142 (TARGET_CRYPTO): Likewise.
13143 (TARGET_P8_VECTOR): Likewise.
13144 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
13145 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
13146 (VECTOR_MEM_P8_VECTOR_P): Likewise.
13147 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
13148 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
13149 (TARGET_XSCVDPSPN): Likewise.
13150 (TARGET_XSCVSPDPN): Likewsie.
13151 (TARGET_SYNC_HI_QI): Likewise.
13152 (TARGET_SYNC_TI): Likewise.
13153 (MASK_CRYPTO): Likewise.
13154 (MASK_DIRECT_MOVE): Likewise.
13155 (MASK_P8_FUSION): Likewise.
13156 (MASK_P8_VECTOR): Likewise.
13157 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
13158 temporary used by some of the direct move instructions to get two FP
13159 temporary registers does not force creation of a stack frame.
13160 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
13161 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
13162 that any VSX registers are tieable, even if they are also an
13163 Altivec vector mode.
13164 (r6000_reg_class_enum): Add wm, wr, wv constraints.
13165 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
13166 (RS6000_BTM_CRYPTO): Likewise.
13167 (RS6000_BTM_COMMON): Likewise.
13169 * config/rs6000/rs6000.md (cpu attribute): Add power8.
13170 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
13171 (enum rs6000_vector): Add power8 vector support.
13173 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
13177 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
13178 (any_sibcall_could_use_r3): this and handle indirect calls.
13179 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
13181 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13183 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
13185 2013-05-22 Richard Biener <rguenther@suse.de>
13187 PR middle-end/57349
13188 * profile.c (branch_prob): Do not split blocks that are
13189 abnormally receiving from ECF_RETURNS_TWICE functions.
13191 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
13193 * recog.c (offsettable_address_addr_space_p): Fix calculation of
13194 address mode. Move pointer mode initialization to the same place.
13196 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
13198 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
13199 while it has any effect.
13201 2013-05-21 Easwaran Raman <eraman@google.com>
13203 PR tree-optimization/57322
13204 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
13205 UID of the statement added to the BB to be 1.
13207 2013-05-21 Jakub Jelinek <jakub@redhat.com>
13209 PR tree-optimization/57331
13210 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
13211 of conversion from pointer type to integral type with integer.
13213 2013-05-21 Martin Jambor <mjambor@suse.cz>
13216 * ipa-prop.c (ipa_read_node_info): Process param_used and
13217 controlled_uses in the same order as when writing.
13219 2013-05-21 Magnus Granberg <baldrick@free.fr>
13222 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
13224 2013-05-21 Richard Biener <rguenther@suse.de>
13226 PR tree-optimization/57318
13227 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
13228 estimate stmts with side-effects as likely eliminated.
13230 2013-05-21 Richard Biener <rguenther@suse.de>
13232 PR tree-optimization/57330
13233 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
13234 preserve the call stmts fntype.
13236 2013-05-21 Richard Biener <rguenther@suse.de>
13238 PR tree-optimization/57303
13239 * tree-ssa-sink.c (statement_sink_location): Improve killing
13240 stmt detection and properly handle self-assignments.
13242 2013-05-21 Christian Bruel <christian.bruel@st.com>
13244 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
13245 spanning registers. LEAF_REG_REMAP is supported only for contiguous
13246 registers. Set register size out of the PARALLEL loop.
13248 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
13251 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
13252 (*fmasf4, *fmasf4_media): New insns.
13254 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
13256 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
13257 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
13258 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
13259 (mips_idiv_insns): Update the comments to say that the returned
13260 instruction counts are in units of BASE_INSN_LENGTH.
13261 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
13262 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
13263 using 2 rather than 4 as the length of indirect MIPS16 and
13264 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
13265 length of a NOP. Don't divide MIPS16 lengths by 2.
13266 (mips16_split_long_branches): Assume a branch is long if the
13267 length is greater than 4 rather than 8.
13268 * config/mips/mips.md (length): Give MIPS16 lengths directly,
13269 rather than multiplying them by 2. Multiply instruction counts
13270 by BASE_INSN_LENGTH rather than 4.
13271 (*jump_mips16, tls_get_tp_mips16_<mode>)
13272 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
13274 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
13276 * config/mips/mips.md (extended_mips16): Remove branch case.
13277 (length): Remove duplicated extended_mips16 test.
13279 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
13281 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
13283 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
13285 * recog.h (Recog_data): Rename to...
13286 (recog_data_d): ...this.
13287 (recog_data): Update accordingly.
13288 * recog.c (recog_data): Likewise.
13289 * reload.c (save_recog_data): Likewise.
13290 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
13291 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
13293 2013-05-17 Julian Brown <julian@codesourcery.com>
13295 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
13296 found in a REG_EQUAL note, invalidate it.
13298 2013-05-17 Easwaran Raman <eraman@google.com>
13300 * tree-ssa-reassoc.c (find_insert_point): New function.
13301 (insert_stmt_after): Likewise.
13302 (get_def_stmt): Likewise.
13303 (ensure_ops_are_available): Likewise.
13304 (not_dominated_by): Likewise.
13305 (rewrite_expr_tree): Do not move statements beyond what is
13306 necessary. Remove call to swap_ops_for_binary_stmt...
13307 (reassociate_bb): ... and move it here.
13308 (build_and_add_sum): Assign UIDs for new statements.
13309 (linearize_expr): Likewise.
13310 (do_reassoc): Renumber gimple statement UIDs.
13312 2013-05-17 Jan Hubicka <jh@suse.cz>
13314 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
13316 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
13317 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
13319 (output_weakrefs): Update.
13321 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
13322 Martin Jambor <mjambor@suse.cz>
13324 PR middle-end/57276
13325 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
13326 value that corresponds to the given aggval is found in values vector.
13328 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
13330 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
13331 sse, sse2, sse3, ssse3 and sse4a flags to options.
13333 2013-05-17 David Malcolm <dmalcolm@redhat.com>
13335 * gengtype-state.c: (s_expr_writer): New class, to handle
13336 prettifying of output layout of s-expressions.
13337 (state_writer): New class, to write out gtype.state.
13338 (state_written_type_count): Move this variable into member data of
13340 (s_expr_writer::s_expr_writer): New code: constructor for new class
13341 (state_writer::state_writer(): ditto
13342 (s_expr_writer::write_new_line): New function
13343 (s_expr_writer::write_any_indent): ditto
13344 (s_expr_writer::begin_s_expr): ditto
13345 (s_expr_writer::end_s_expr): ditto
13346 (write_state_fileloc): convert to method of state_writer...
13347 (state_writer:: write_state_fileloc): ...and use methods of
13348 s_expr_writer to write indentation into the gtype.state output file
13349 to visually represent the hierarchical structure of the list
13351 (write_state_fields): ditto, renaming to...
13352 (state_writer::write_state_fields)
13353 (write_state_a_string): ditto, renaming to...
13354 (state_writer::write_state_a_string)
13355 (write_state_string_option): ditto, renaming to...
13356 (state_writer::write_state_string_option)
13357 (write_state_type_option): ditto, renaming to...
13358 (state_writer::write_state_type_option)
13359 (write_state_nested_option): ditto, renaming to...
13360 (state_writer::write_state_nested_option)
13361 (write_state_option): ditto, renaming to...
13362 (state_writer::write_state_option)
13363 (write_state_options): ditto, renaming to...
13364 (state_writer::write_state_options)
13365 (write_state_lang_bitmap): ditto, renaming to...
13366 (state_writer::write_state_lang_bitmap)
13367 (write_state_version): ditto, renaming to...
13368 (state_writer::write_state_version)
13369 (write_state_scalar_type): ditto, renaming to...
13370 (state_writer::write_state_scalar_type)
13371 (write_state_string_type): ditto, renaming to...
13372 (state_writer::write_state_string_type)
13373 (write_state_undefined_type): ditto, renaming to...
13374 (state_writer::write_state_undefined_type)
13375 (write_state_struct_union_type): ditto, renaming to...
13376 (state_writer::write_state_struct_union_type)
13377 (write_state_struct_type): ditto, renaming to...
13378 (state_writer::write_state_struct_type)
13379 (write_state_user_struct_type): ditto, renaming to...
13380 (state_writer::write_state_user_struct_type)
13381 (write_state_lang_struct_type): ditto, renaming to...
13382 (state_writer::write_state_lang_struct_type)
13383 (write_state_param_struct_type): ditto, renaming to...
13384 (state_writer::write_state_param_struct_type)
13385 (write_state_pointer_type): ditto, renaming to...
13386 (state_writer::write_state_pointer_type)
13387 (write_state_array_type): ditto, renaming to...
13388 (state_writer::write_state_array_type)
13389 (write_state_gc_used): ditto, renaming to...
13390 (state_writer::write_state_gc_used)
13391 (write_state_common_type_content): ditto, renaming to...
13392 (state_writer::write_state_common_type_content)
13393 (write_state_type): ditto, renaming to...
13394 (state_writer::write_state_type)
13395 (write_state_pair_list): ditto, renaming to...
13396 (state_writer::write_state_pair_list)
13397 (write_state_pair): ditto, renaming to...
13398 (state_writer::write_state_pair)
13399 (write_state_typedefs): ditto, renaming to...
13400 (state_writer::write_state_typedefs)
13401 (write_state_structures): ditto, renaming to...
13402 (state_writer::write_state_structures)
13403 (write_state_param_structs): ditto, renaming to...
13404 (state_writer::write_state_param_structs)
13405 (write_state_variables): ditto, renaming to...
13406 (state_writer::write_state_variables)
13407 (write_state_srcdir): ditto, renaming to...
13408 (state_writer::write_state_srcdir)
13409 (write_state_files_list): ditto, renaming to...
13410 (state_writer::write_state_files_list)
13411 (write_state_languages): ditto, renaming to...
13412 (state_writer::write_state_languages)
13413 (write_state): create a state_writer instance and use it when
13414 writing out the state file
13416 2013-05-17 Mike Stump <mikestump@comcast.net>
13418 PR rtl-optimization/57304
13419 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
13420 accessing DF_REF_REAL_LOC.
13422 2013-05-17 Jakub Jelinek <jakub@redhat.com>
13424 PR rtl-optimization/57281
13425 PR rtl-optimization/57300
13426 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
13427 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
13428 what the other splitter did if the registers are dead.
13430 2013-05-17 Richard Biener <rguenther@suse.de>
13432 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
13435 2013-05-17 Jakub Jelinek <jakub@redhat.com>
13437 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
13440 2013-05-17 Marek Polacek <polacek@redhat.com>
13442 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
13443 length when doing non-zero store of storing '\0' to '\0'.
13445 2013-05-17 Jakub Jelinek <jakub@redhat.com>
13447 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
13448 vect_external_def oprnd1 with loop_vinfo, try to emit
13449 optional cast, negation and and stmts on the loop preheader
13450 edge instead of into the pattern def seq.
13452 PR tree-optimization/57051
13453 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
13454 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
13456 2013-05-16 Nick Clifton <nickc@redhat.com>
13458 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
13459 (rl78_is_naked_func): New function.
13460 (rl78_expand_prologue): Skip prologue generation for naked functions.
13461 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
13462 * doc/extend.texi (naked): Add RL78 to the list of processors
13463 that supports this attribute.
13465 2013-05-16 Jeff Law <law@redhat.com>
13467 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
13469 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
13471 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
13472 cache parameters using detect_caches_amd also for CYRIX,
13473 NSC and TM2 signatures.
13475 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
13476 Dzianis Kahanovich <mahatma@eu.by>
13480 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
13481 VIA/Centaur processors and determine their cache parameters
13482 using detect_caches_amd.
13484 2013-05-16 Teresa Johnson <tejohnson@google.com>
13486 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
13487 (rtl_verify_edges): New function.
13488 (rtl_verify_bb_insns): Ditto.
13489 (rtl_verify_bb_pointers): Ditto.
13490 (rtl_verify_bb_insn_chain): Ditto.
13491 (rtl_verify_fallthru): Ditto.
13492 (rtl_verify_bb_layout): Ditto.
13493 (rtl_verify_flow_info_1): Outline checks into new functions.
13494 (rtl_verify_flow_info): Ditto.
13496 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
13498 * cfghooks.c (copy_bbs): Add update_dominance argument.
13499 * cfghooks.h (copy_bbs): Update prototype.
13500 * tree-cfg.c (gimple_duplicate_sese_region):
13501 Add update_dominance argument.
13502 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
13503 * tree-ssa-loop-ch.c (copy_loop_headers): Update
13504 gimple_duplicate_sese_region call.
13505 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
13506 Update copy_bbs call.
13507 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
13508 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
13510 2013-05-16 Jakub Jelinek <jakub@redhat.com>
13512 * tree-vectorizer.h (NUM_PATTERNS): Increment.
13513 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
13514 vect_recog_rotate_pattern.
13515 (vect_recog_rotate_pattern): New function.
13517 2013-05-16 Jason Merrill <jason@redhat.com>
13519 * Makefile.in (LLINKER): New variable.
13520 (mostlyclean): Remove link mutex.
13521 * configure.ac: Handle --enable-link-mutex.
13522 * lock-and-run.sh: New script.
13524 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
13527 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
13530 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
13532 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
13534 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
13536 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
13537 * config/arm/arm.c (next_consecutive_mem): New function.
13538 (gen_movmem_ldrd_strd): Likewise.
13539 * config/arm/arm.md (movmemqi): Update condition and code.
13540 (unaligned_loaddi, unaligned_storedi): New patterns.
13542 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
13544 * config.gcc: Obsolete *-*-solaris2.9*.
13545 * doc/install.texi (Specific, *-*-solaris2*): Document it.
13547 2013-05-16 Richard Biener <rguenther@suse.de>
13549 * passes.c (init_optimization_passes): Move pass_parallelize_loops
13550 earlier, after GRAPHITE transforms and IV canonicalization.
13552 2013-05-16 Jakub Jelinek <jakub@redhat.com>
13554 * omp-low.c (extract_omp_for_data): For collapsed loops,
13555 if at least one of the loops is known at compile time to
13556 iterate zero times, set count to 0.
13557 (expand_omp_regimplify_p): New function.
13558 (expand_omp_for_generic): For collapsed loops, if at least
13559 one of the loops isn't known to iterate at least once,
13560 add runtime check with setting count to 0.
13561 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
13562 For unsigned types if it isn't known at compile time that
13563 the loop will iterate at least once, add runtime check to bypass
13564 the whole loop if initial condition isn't true.
13566 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
13568 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
13570 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
13572 PR middle-end/57286
13573 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
13574 transformations to avoid an infinite loop.
13576 2013-05-16 Marek Polacek <polacek@redhat.com>
13578 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
13580 2013-05-15 Leif Ekblad <leif@rdos.net>
13582 * config/i386/i386.c (ix86_decompose_address): Use
13583 DEFAULT_TLS_SEG_REG to access TLS segment register.
13584 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
13585 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
13586 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
13588 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
13591 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
13592 sibling calls to functions that would normally be lazily bound,
13593 unless $gp is call-clobbered.
13595 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
13597 * config/i386/i386.c (ix86_option_override_internal): Update
13598 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
13599 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
13600 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
13601 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
13603 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
13605 2013-05-15 Andreas Schwab <schwab@suse.de>
13607 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
13608 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
13611 2013-05-15 Teresa Johnson <tejohnson@google.com>
13613 * loop-unroll.c (report_unroll_peel): Check decision before
13614 emitting unroll/peel message.
13616 2013-05-15 Teresa Johnson <tejohnson@google.com>
13618 * function.h (has_bb_partition): New rtl_data flag.
13619 (bb_reorder_complete): Ditto.
13620 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
13621 instead of flag_reorder_blocks_and_partition.
13622 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
13623 with some enhancements.
13624 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
13625 * bb-reorder.c (connect_traces): Check for has_bb_partition
13626 instead of flag_reorder_blocks_and_partition.
13627 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
13628 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
13629 verify_hot_cold_block_grouping.
13630 (partition_hot_cold_basic_blocks): Set has_bb_partition.
13632 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
13635 * config/arm/predicates.md (call_insn_operand): New predicate.
13636 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
13637 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
13638 if insn is not a tail call.
13639 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
13641 * config/arm/arm.h (enum reg_class): New caller save register class.
13642 (REG_CLASS_NAMES): Likewise.
13643 (REG_CLASS_CONTENTS): Likewise.
13644 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
13647 2013-05-15 Richard Biener <rguenther@suse.de>
13649 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
13650 of MSG_OPTIMIZED_LOCATIONS.
13651 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
13652 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
13654 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
13655 of MSG_OPTIMIZED_LOCATIONS.
13656 (execute_vect_slp): Likewise.
13657 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
13658 (vect_create_cond_for_alias_checks): Likewise.
13659 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
13660 (vect_recog_widen_mult_pattern): Likewise.
13661 (vect_recog_widen_sum_pattern): Likewise.
13662 (vect_recog_over_widening_pattern): Likewise.
13663 (vect_recog_widen_shift_pattern): Likewise.
13664 (vect_recog_vector_vector_shift_pattern): Likewise.
13665 (vect_recog_divmod_pattern): Likewise.
13666 (vect_recog_mixed_size_cond_pattern): Likewise.
13667 (vect_recog_bool_pattern): Likewise.
13668 (vect_pattern_recog_1): Likewise.
13670 2013-05-15 Martin Jambor <mjambor@suse.cz>
13672 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
13673 non-functions to builtin_unreachable.
13674 * ipa-inline-transform.c (inline_call): Do not assert estimates were
13675 correct when new direct edges were discovered.
13677 2013-05-15 Martin Jambor <mjambor@suse.cz>
13679 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
13680 header, print symbol order instead of node uid, print more information
13681 about indirect edge targets.
13682 (ipa_make_edge_direct_to_target): Print symbol order instead of node
13684 (ipa_make_edge_direct_to_target): Likewise.
13685 (remove_described_reference): Likewise.
13686 (propagate_controlled_uses): Likewise.
13687 (ipa_print_node_params): Also print symbol order.
13688 (ipcp_transform_function): Print symbol order instead of node uids.
13689 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
13690 (cgraph_get_create_real_symbol_node): Likewise.
13691 * ipa-cp.c (print_lattice): Likewise.
13692 (print_all_lattices): Likewise.
13693 (determine_versionability): Likewise.
13694 (initialize_node_lattices): Likewise.
13695 (estimate_local_effects): Likewise.
13696 (update_profiling_info): Likewise.
13697 (create_specialized_node): Likewise.
13698 (perhaps_add_new_callers): Likewise.
13699 (decide_about_value): Likewise.
13700 (decide_whether_version_node): Likewise.
13701 (identify_dead_nodes): Likewise.
13702 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
13703 (dump_inline_summary): Likewise.
13704 (estimate_node_size_and_time): Likewise.
13705 (inline_analyze_function): Likewise.
13706 * ipa-inline.c (report_inline_failed_reason): Likewise.
13707 (want_early_inline_function_p): Likewise.
13708 (edge_badness): Likewise.
13709 (update_edge_key): Likewise.
13710 (inline_small_functions): Likewise. Add dumping of order to two other
13712 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
13713 instead of node uids.
13714 (propagate_pure_const): Likewise.
13715 (propagate_pure_const): Likewise.
13716 * ipa-utils.c (dump_cgraph_node_set): Likewise.
13717 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
13718 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
13720 * tree-pretty-print.c (dump_function_header): Likewise.
13721 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
13722 Print symbol order instead of node uids.
13724 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13726 * config/s390/s390.c (s390_register_move_cost): Don't impose the
13727 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
13729 2013-05-15 Richard Biener <rguenther@suse.de>
13731 PR tree-optimization/57275
13732 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
13733 return value for fail to do runtime alias checks for gather loads.
13735 2013-05-15 Jan Hubicka <jh@suse.cz>
13739 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
13740 weakrefs are not external.
13741 (lto_symtab_merge_decls): Fix thinko when dealing with
13742 non-lto_symtab decls.
13743 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
13744 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
13745 * varpool.c (dump_varpool_node): Dump more flags.
13747 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
13749 * config/i386/i386.c (processor_alias_table): Add instruction
13750 FSGSBASE for AMD bdver3 architecture.
13752 2013-05-14 Jakub Jelinek <jakub@redhat.com>
13754 * tree.c (warn_deprecated_use): Print file:line using locus color.
13755 * diagnostic.c (diagnostic_report_current_module): Print file:line
13756 and file:line:column using locus color.
13758 2013-05-14 Mike Stump <mikestump@comcast.net>
13760 * gdbinit.in: Add __null.
13762 2013-05-14 Mike Stump <mikestump@comcast.net>
13764 * recog.h: Rename struct recog_data to Recog_data.
13765 * recog.c: Likewise.
13766 * reload.c (can_reload_into): Likewise.
13767 * config/picochip/picochip.c: Likewise.
13769 2013-05-14 Mike Stump <mikestump@comcast.net>
13771 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
13773 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
13775 * resource.h (struct resources): Remove unch_memory member.
13776 (CLEAR_RESOURCE): Don't clear unch_memory.
13777 * resource.c (mark_referenced_resources): Don't set it.
13778 (mark_set_resources): Likewise.
13779 (mark_target_live_regs): Don't clear it.
13780 (init_resource_info): Likewise.
13781 * reorg.c (resource_conflicts_p): Don't compare it.
13782 (redundant_insn): Don't set it.
13784 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
13786 * emit-rtl.c (next_label): Remove unused function.
13787 (skip_consecutive_labels, link_cc0_insns): Move to ...
13788 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
13789 only place where these functions are used, and make them static.
13791 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
13793 * fold-const.c (fold_negate_expr): Handle vectors.
13794 (fold_truth_not_expr): Make it static.
13795 (fold_invert_truthvalue): New static function.
13796 (invert_truthvalue_loc): Handle vectors. Do not call
13797 fold_truth_not_expr directly.
13798 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
13799 <TRUTH_NOT_EXPR>: Do not cast to boolean.
13800 (fold_comparison): Handle vector constants.
13801 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
13802 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
13803 * tree.h (fold_truth_not_expr): Remove declaration.
13805 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
13807 * config/aarch64/aarch64-simd.md
13808 (aarch64_vcond_internal<mode>): Rename to...
13809 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
13810 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
13811 float modes. Clarify all iterator modes.
13812 (vcond<mode><mode>): Use new name for vcond expanders.
13813 (vcond<v_cmp_result><mode>): Likewise.
13814 (vcondu<mode><mode>: Likewise.
13815 * config/aarch64/iterators.md (VDQF_COND): New.
13817 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
13820 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
13821 variable for the shift amount. Check that we shift by non-negative
13824 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
13827 * config/arm/arm.h (EPILOGUE_USES): Only return true
13828 for LR_REGNUM after epilogue_completed.
13830 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
13832 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
13833 is error_mark_node.
13835 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
13838 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
13839 and Solaris 11+/x86 with gld.
13840 * configure: Regenerate.
13842 2013-05-14 Jakub Jelinek <jakub@redhat.com>
13844 * expmed.c (expand_shift_1): Canonicalize rotates by
13845 constant bitsize / 2 to bitsize - 1.
13846 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
13847 case ROTATERT>: Likewise.
13850 2013-05-10 Jakub Jelinek <jakub@redhat.com>
13852 * config/i386/i386.md (rotateinv): New code attr.
13853 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
13854 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
13855 roll $31, %eax, etc.
13857 2013-05-14 Richard Biener <rguenther@suse.de>
13859 PR middle-end/57235
13860 * tree-eh.c (sink_clobbers): Give up for successors with
13861 multiple predecessors and no virtual uses.
13863 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
13865 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
13866 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
13868 2013-05-14 Jakub Jelinek <jakub@redhat.com>
13870 PR middle-end/57251
13871 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
13872 the case when both op0 and op1 have VOIDmode.
13874 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
13876 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
13877 in multiply-accumulate mode.
13879 2013-05-13 Guozhi Wei <carrot@google.com>
13881 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
13883 2013-05-13 Kai Tietz <ktietz@redhat.com>
13886 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
13887 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
13888 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
13889 * config/i386/i386.c (ix86_option_override_internal): Likewise.
13890 (ix86_expand_prologue): Likewise.
13891 (ix86_expand_split_stack_prologue): Likewise.
13892 (legitimate_pic_address_disp_p): Likewise.
13893 (legitimize_pic_address): Likewise.
13894 (legitimize_tls_address): Likewise.
13895 (legitimize_pe_coff_symbol): Likewise.
13896 (output_pic_addr_const): Likewise.
13897 (construct_plt_address): Likewise.
13898 (ix86_expand_call): Likewise.
13899 (x86_output_mi_thunk): Likewise.
13900 (x86_function_profiler): Likewise.
13902 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
13904 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
13905 similar switch cases.
13906 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
13907 (aarch64_simd_mov_to_<mode>low): Delete.
13908 (aarch64_simd_mov_to_<mode>high): Delete.
13909 (move_lo_quad_<mode>): Add w<-r alternative.
13910 (aarch64_simd_move_hi_quad_<mode>): Likewise.
13911 (aarch64_simd_mov_from_*): Update type attribute.
13912 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
13915 2013-05-13 Jan Hubicka <jh@suse.cz>
13917 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
13918 * config/i386/i386.c (ix86_compute_frame_layout,
13919 ix86_expand_epilogue, emit_i387_cw_initialization,
13920 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
13921 ix86_local_alignment): Fix use of size/speed predicates.
13923 2013-05-13 Jakub Jelinek <jakub@redhat.com>
13925 PR tree-optimization/45216
13926 PR tree-optimization/57157
13927 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
13928 the (-Y) & (B - 1) variant if OP is |.
13929 * expmed.c (expand_shift_1): For rotations by const0_rtx just
13930 return shifted. Use (-op1) & (prec - 1) as other_amount
13931 instead of prec - op1.
13933 2013-05-13 Martin Jambor <mjambor@suse.cz>
13935 PR middle-end/42371
13936 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
13937 (ipa_constant_data): New type.
13938 (ipa_jump_func): Use ipa_constant_data to hold information about
13939 constant jump functions.
13940 (ipa_get_jf_constant): Adjust to jump function type changes.
13941 (ipa_get_jf_constant_rdesc): New function.
13942 (ipa_param_descriptor): New field controlled_uses.
13943 (ipa_get_controlled_uses): New function.
13944 (ipa_set_controlled_uses): Likewise.
13945 * ipa-ref.h (ipa_find_reference): Declare.
13946 * ipa-prop.c (ipa_cst_ref_desc): New type.
13947 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
13949 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
13950 New parameter cs. Adjust all callers.
13951 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
13952 (remove_described_reference): New function.
13953 (jfunc_rdesc_usable): Likewise.
13954 (try_make_edge_direct_simple_call): Decrement controlled use count,
13955 attempt to remove reference if it hits zero.
13956 (combine_controlled_uses_counters): New function.
13957 (propagate_controlled_uses): Likewise.
13958 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
13959 (ipa_edge_duplication_hook): Duplicate reference descriptions.
13960 (ipa_print_node_params): Print described use counter.
13961 (ipa_write_jump_function): Adjust to jump function type changes.
13962 (ipa_read_jump_function): New parameter CS, pass it to
13963 ipa_set_jf_constant. Adjust caller.
13964 (ipa_write_node_info): Stream controlled use count
13965 (ipa_read_node_info): Likewise.
13966 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
13968 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
13969 count. Remove cloning-added reference if it reaches zero.
13970 * ipa-ref.c (ipa_find_reference): New function.
13972 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
13974 * config/i386/i386.c (processor_target_table): Modified default
13975 alignment values for AMD BD and BT architectures.
13977 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
13979 * tree-vect-generic.c (uniform_vector_p): Move ...
13980 * tree.c (uniform_vector_p): ... here.
13981 * tree.h (uniform_vector_p): Declare it.
13982 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
13985 2013-05-13 Jakub Jelinek <jakub@redhat.com>
13987 PR tree-optimization/57230
13988 * tree-ssa-strlen.c (handle_char_store): Record length for
13989 array store from STRING_CST.
13991 PR tree-optimization/57230
13992 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
13995 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
13997 * config/epiphany/epiphany.c (epiphany_init): Check size of
13998 NUM_MODES_FOR_MODE_SWITCHING.
13999 (epiphany_expand_prologue):
14000 Remove CONFIG_REGNUM initial value handling code.
14001 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
14002 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
14003 (emit_set_fp_mode, epiphany_mode_after): Likewise.
14004 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
14005 Don't return 1 for FP_MODE_NONE.
14006 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
14007 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
14008 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
14009 * config/epiphany/epiphany.md (save_config): New pattern.
14011 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
14013 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
14015 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
14017 * config/i386/i386.md (memory): Handle sseishft1.
14018 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
14019 (*vec_extractv2di_1): Ditto.
14021 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
14023 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
14026 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
14028 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
14029 Add mthumb/march=armv7-a multilib.
14030 Add mthumb/march=armv7-r multilib.
14031 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
14033 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
14035 * config/v850/t-rtems: Add more multilibs.
14037 2013-05-10 Richard Biener <rguenther@suse.de>
14039 PR tree-optimization/57214
14040 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
14041 not propagate from SSA names that occur in abnormal PHI nodes.
14043 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
14045 * stor-layout.c (element_precision): New function.
14046 * machmode.h (element_precision): Declare it.
14047 * tree.c (build_minus_one_cst): New function.
14048 (element_precision): Likewise.
14049 * tree.h (build_minus_one_cst): Declare new function.
14050 (element_precision): Likewise.
14051 * fold-const.c (operand_equal_p): Use element_precision.
14052 (fold_binary_loc): Handle vector types.
14053 * convert.c (convert_to_integer): Use element_precision.
14054 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
14057 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
14059 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
14060 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
14061 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
14062 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
14063 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
14064 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
14065 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
14066 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
14067 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
14068 (Uuw8): New constraints.
14069 (Usb4): Move into alphabetical order.
14070 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
14071 (sd8_operand, ub8_operand, uw8_operand): New predicates.
14072 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
14073 previously unnamed patterns.
14074 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
14075 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
14076 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
14077 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
14078 of set_attr_alternative/if_then_else. Use extended_mips16 instead
14079 of specific lengths.
14081 2013-05-10 Jakub Jelinek <jakub@redhat.com>
14083 * config/i386/i386.md (rotateinv): New code attr.
14084 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
14085 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
14086 roll $31, %eax, etc.
14088 PR tree-optimization/45216
14089 PR tree-optimization/57157
14090 * tree-ssa-forwprop.c (simplify_rotate): New function.
14091 (ssa_forward_propagate_and_combine): Call it.
14093 2013-05-10 Richard Biener <rguenther@suse.de>
14095 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
14096 disable peeling when we version for aliasing.
14097 (vector_alignment_reachable_p): Honor explicit user alignment.
14098 (vect_supportable_dr_alignment): Likewise.
14099 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
14100 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
14101 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
14102 then peeling to arrange for the cost-model check to come first.
14104 2013-05-10 Alan Modra <amodra@gmail.com>
14106 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
14107 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
14108 * configure: Regenerate.
14110 2013-05-10 Alan Modra <amodra@gmail.com>
14113 * varasm.c (default_elf_select_section): Move !DECL_P check..
14114 (get_named_section): ..to here before calling get_section_name.
14116 (default_section_type_flags): Add DECL_P check.
14117 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
14118 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
14120 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
14122 * config/epiphany/epiphany.c (epiphany_expand_prologue):
14123 When using gen_stack_adjust_str with a register offset, add a
14124 REG_FRAME_RELATED_EXPR note.
14126 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
14128 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
14129 (*vec_extractv4si_zext_mem): Ditto.
14130 (*vec_extractv2di): Add 0->x and x->x alternatives.
14131 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
14132 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
14134 2013-05-09 Jason Merrill <jason@redhat.com>
14136 N3639 C++1y VLA support
14137 * gimplify.c (gimplify_vla_decl): Don't touch an existing
14140 * tree.c (build_constructor_va): New.
14141 * tree.h: Declare it.
14143 2013-05-09 Martin Jambor <mjambor@suse.cz>
14146 * gimple-fold.c (canonicalize_constructor_val): Call
14147 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
14149 2013-05-09 Jan Hubicka <jh@suse.cz>
14150 Richard Biener <rguenther@suse.de>
14153 * symtab.c (symtab_make_decl_local): Do not add private names.
14155 2013-05-09 Jan Hubicka <jh@suse.cz>
14158 * symtab.c (insert_to_assembler_name_hash): Handle clones.
14159 (unlink_from_assembler_name_hash): Likewise.
14160 (symtab_prevail_in_asm_name_hash, symtab_register_node,
14161 symtab_unregister_node, symtab_initialize_asm_name_hash,
14162 change_decl_assembler_name): Update.
14164 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
14166 * config/aarch64/aarch64.md: New movtf split.
14167 (*movtf_aarch64): Update.
14168 (aarch64_movdi_tilow): Handle TF modes and rename to
14169 aarch64_movdi_<mode>low.
14170 (aarch64_movdi_tihigh): Handle TF modes and rename to
14171 aarch64_movdi_<mode>high
14172 (aarch64_movtihigh_di): Handle TF modes and rename to
14173 aarch64_mov<mode>high_di
14174 (aarch64_movtilow_di): Handle TF modes and rename to
14175 aarch64_mov<mode>low_di
14176 (aarch64_movtilow_tilow): Remove spurious whitespace.
14177 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
14179 (aarch64_print_operand): Update.
14181 2013-05-09 Alan Modra <amodra@gmail.com>
14183 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
14185 * configure: Regenerate.
14187 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
14189 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
14190 splitter preparation statements.
14191 * config/i386/sse.md (*vec_extract* splitters): Ditto.
14192 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
14195 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14197 * gimple-ssa-strength-reduction.c (count_candidates): Change
14198 return value to int.
14199 (analyze_candidates_and_replace): Change type of length to int.
14201 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
14203 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
14204 (*vec_extract<mode>): Use VI12_128 mode iterator.
14205 (*vec_extract<mode>_mem): Ditto.
14206 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
14209 2013-05-08 Diego Novillo <dnovillo@google.com>
14214 2012-08-17 Diego Novillo <dnovillo@google.com>
14217 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
14218 * config.in: Regenerate.
14219 * configure: Regenerate.
14220 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
14222 2013-05-08 Jan Hubicka <jh@suse.cz>
14225 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
14226 * cgraph.h (symtab_node_base): Add unique_name.
14227 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
14228 input_overwrite_node, input_varpool_node): Stream unique_name.
14229 * cgraphclones.c (cgraph_create_virtual_clone,
14230 cgraph_function_versioning): Set unique_name.
14231 * ipa.c (function_and_variable_visibility): Set unique_name.
14233 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14235 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
14236 (alloc_cand_and_find_basis): Restrict conditional candidate
14237 processing to CAND_MULTs.
14239 2013-05-08 Jan Hubicka <jh@suse.cz>
14242 lto-symtab.c (lto_symtab_symbol_p): New function.
14243 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
14244 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
14245 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
14246 Skip static symbols.
14248 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
14250 PR tree-optimization/57200
14251 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
14252 Only call inform if the preceding warning_at returns true.
14254 2013-05-07 Han Shen <shenhan@google.com>
14256 * cfgexpand.c (record_or_union_type_has_array_p): New function.
14257 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
14258 * common.opt (fstack-protector-strong): New option.
14259 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
14260 * doc/invoke.texi (Optimization Options): Document
14261 "-fstack-protector-strong".
14262 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
14264 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
14266 * config/mips/mips.c (mips_machine_reorg2): Return 0.
14268 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
14270 * ira.c (update_equiv_regs): Add insn having equiv memory even if
14271 it is not lhs of the insn.
14272 (setup_reg_equiv): Remove insn having equiv memory which it is not
14274 * lra-constraints.c (process_address): Try to improve generation
14275 code for address base + disp.
14276 (lra_constraints): Make correct the code for checking insn setting
14277 up backward equivalence. Remove insn only if it is in the init
14279 * lra-eliminations.c (update_reg_eliminate): Change return value.
14280 (lra_eliminate): Use the result.
14282 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
14284 * config/i386/sse.md (ssescalarnummask): New mode attribute.
14285 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
14286 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
14287 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
14288 register target operands.
14289 (*vec_extractv8hi_sse2): New pattern.
14290 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
14291 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
14292 (*vec_extract<mode>_mem): New insn and split pattern.
14294 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
14296 * config/arm/arm.c (arm_asan_shadow_offset): New function.
14297 (TARGET_ASAN_SHADOW_OFFSET): Define.
14298 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
14299 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
14301 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14303 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
14304 (incr_vec_index): Return -1 if increment not found.
14305 (create_add_on_incoming_edge): Assert if increment not found.
14306 (record_increment): Limit number of increments recorded.
14307 (all_phi_incrs_profitable): Return false if an increment not found.
14308 (replace_profitable_candidates): Don't process increments that were
14310 (analyze_candidates_and_replace): Limit size of incr_vec.
14312 2013-05-07 Richard Biener <rguenther@suse.de>
14314 * calls.c (special_function_p): setjmp-like functions are leaf.
14315 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
14316 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
14318 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
14320 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
14321 (aarch64_simd_mov<mode>): New expander.
14322 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
14323 (aarch64_simd_mov_to_<mode>high): Likewise.
14324 (aarch64_simd_mov_from_<mode>low): Likewise.
14325 (aarch64_simd_mov_from_<mode>high): Likewise.
14326 (aarch64_dup_lane<mode>): Update.
14327 (aarch64_dup_lanedi): New instruction pattern.
14328 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
14329 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
14331 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14333 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
14334 (replace_mult_candidate): Remove unnecessary argument; remove
14335 unnecessary parameter from call to introduce_cast_before_cand.
14336 (replace_unconditional_candidate): Remove unnecessary parameter
14337 from call to replace_mult_candidate.
14338 (replace_conditional_candidate): Likewise.
14339 (insert_initializers): Use make_temp_ssa_name.
14340 (introduce_cast_before_cand): Remove unnecessary argument; use
14341 make_temp_ssa_name.
14342 (replace_one_candidate): Remove unnecessary argument; remove
14343 unnecessary parameter from calls to introduce_cast_before_cand.
14344 (replace_profitable_candidates): Remove unnecessary parameters
14345 from calls to replace_one_candidate.
14347 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14349 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
14350 phi def as possibly hiding a basis for a CAND_ADD whose operands
14351 have been commuted in the analysis.
14352 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
14354 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14356 * config/aarch64/aarch64.md
14357 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
14358 shift value between 0-4.
14360 2013-05-07 Richard Biener <rguenther@suse.de>
14362 * double-int.h (rshift): New overload.
14363 * double-int.c (rshift): New function.
14364 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
14365 (create_reference_ops_from_ref): Remove.
14366 (vn_reference_insert): Use shared ops for constructing the
14367 reference and copy it.
14369 2013-05-07 Richard Biener <rguenther@suse.de>
14371 PR middle-end/57190
14372 * tree-eh.c (sink_clobbers): Properly propagate
14373 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
14375 2013-05-07 Jakub Jelinek <jakub@redhat.com>
14377 PR tree-optimization/57149
14378 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
14379 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
14380 collect_phi_def_edges, execute_late_warn_uninitialized): Use
14381 uninit_undefined_value_p instead of ssa_undefined_value_p.
14384 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
14385 for modifier == EXPAND_INITIALIZER.
14387 2013-05-07 Anton Blanchard <anton@samba.org>
14389 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
14390 for powerpc64 little endian.
14391 * configure: Regenerate.
14393 2013-05-06 Graham Stott <grahams@btinternet.com>
14395 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
14396 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
14397 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
14398 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
14400 2013-05-06 Graham Stott <grahams@btinternet.com>
14402 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
14403 codes which allow non-lvalues.
14405 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
14407 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
14408 components are all 1s.
14409 (integer_minus_onep): New function.
14410 * tree.h (integer_minus_onep): Declare it.
14411 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
14412 integer_minus_onep instead of integer_all_onesp.
14414 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
14417 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
14418 variations of these patterns.
14420 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
14422 * config/i386/i386.md (isa): Add x64_sse4 member.
14423 (enabled): Handle x64_sse4.
14424 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
14425 instruction for 64bit SSE4_1 targets. Update insn attributes.
14426 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
14427 instruction for SSE4_1 targets. Update insn attributes.
14428 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
14429 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
14430 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
14432 (*vec_extractv4si): Rename from *sse4_1_pextrd.
14433 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
14434 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
14436 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
14439 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
14441 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
14443 * final.c (do_assembler_dialects): Don't handle curly braces and
14444 vertical bar escaped by % as dialect delimiters.
14445 (output_asm_insn): Print curly braces and vertical bar if escaped
14446 by % and ASSEMBLER_DIALECT defined.
14447 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
14448 * doc/tm.texi: Regenerated.
14450 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
14452 * config/mips/mips.c: Include tree-pass.h.
14453 (mips_reorg): Split in pre- and post-dbr_schedule parts.
14454 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
14455 (pass_mips_machine_reorg2): New machine specific pass.
14456 (insert_pass_mips_machine_reorg2): New pass plugin definition.
14457 (mips_option_override): Register the new pass.
14458 * rtl.h (cleanup_barriers): Remove prototype.
14459 (dbr_schedule): Likewise.
14460 * jump.c (cleanup_barriers): Make static.
14461 * reorg.c (dbr_schedule): Likewise.
14463 2013-05-06 Richard Biener <rguenther@suse.de>
14465 PR tree-optimization/57185
14466 * tree-parloops.c (add_field_for_reduction): Handle anonymous
14467 SSA names properly.
14469 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
14472 * config/i386/i386.c (add_parameter_dependencies): Add dependence
14473 between "first_arg" and "insn", not "last" and "insn".
14475 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14477 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
14478 (find_candidates_in_block): Re-enable slsr_process_phi.
14479 (create_phi_basis): Fix double counting of candidate adjustment.
14481 2013-05-06 Richard Biener <rguenther@suse.de>
14483 PR middle-end/57147
14484 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
14485 the edge is also fallthru, preserve it and just clear the
14487 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
14488 also complex, preserve that and just clear the fallthru flag.
14489 * tree-inline.c (update_ssa_across_abnormal_edges): Also
14490 update virtual operands.
14492 2013-05-06 Alan Modra <amodra@gmail.com>
14494 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
14495 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
14496 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
14497 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
14498 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
14499 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
14501 2013-05-06 Alan Modra <amodra@gmail.com>
14503 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
14504 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
14505 (DEFAULT_ASM_ENDIAN): Define.
14506 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
14507 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
14508 Update -K PIC clause from sysv4.h.
14509 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
14510 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
14512 2013-05-06 Alan Modra <amodra@gmail.com>
14514 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
14515 twice for little-endian.
14516 (ashrdi3_no_power, ashrdi3): Support little-endian.
14518 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
14521 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
14522 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
14524 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
14525 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
14526 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
14529 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
14530 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
14532 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
14533 * config/pa/pa.opt: Make mbig-switch a no-op.
14534 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
14535 (CASE_VECTOR_MODE): Always return SImode.
14536 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
14537 for the !TARGET_BIG_SWITCH case.
14538 * config/pa/pa-linux.h: Likewise.
14539 * config/pa/pa-openbsd.h: Likewise.
14540 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
14541 * config/pa/pa.md (short_jump): Remove define_insn.
14542 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
14543 (casesi0): Remove define_insn.
14544 (type): Remove btable_branch.
14545 (pa_combine_type): Likewise.
14546 (in_nullified_branch_delay): Likewise.
14547 (in_call_delay): Likewise.
14548 (define_delay): Likewise.
14549 (define_insn_reservation "Z3"): Likewise.
14550 (define_insn_reservation "Z4"): Likewise.
14551 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
14552 (pa_adjust_insn_length): Remove adjustment for btable branches.
14553 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
14556 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
14558 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
14559 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
14560 Add m->r,x alternatives.
14561 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
14562 splitters using SWI48x mode iterator.
14563 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
14564 TARGET_64BIT. Add m->x alternative.
14565 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
14566 Add o->x alternative. Enable for TARGET_SSE.
14567 (sse_storeq): Remove expander.
14568 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
14569 with memory input operand.
14570 (*vec_extractv2di_1 splitter): New.
14571 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
14572 * config/i386/i386.md (ssevecmodelower): New mode attribute.
14574 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
14576 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
14577 (INT_LOWPART): Delete.
14578 (extract_MB): Adjust.
14579 (extract_ME): Adjust.
14580 (print_operand): Adjust.
14582 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
14584 * config/rs6000/predicates.md (reg_or_add_cint_operand,
14585 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
14586 (reg_or_logical_cint_operand, easy_fp_constant,
14587 logical_const_operand): Delete "CONST_DOUBLE" case.
14588 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
14589 "HOST_BITS_PER_WIDE_INT == 64" test.
14590 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
14591 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
14592 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
14593 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
14595 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
14596 CONST_DOUBLE DImode/VOIDmode case.
14597 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
14598 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
14599 CONST_DOUBLE VOIDmode case.
14600 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
14601 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
14602 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
14603 Delete CONST_DOUBLE case.
14604 (splitters for mov FMOVE64 const_double): Delete
14605 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
14606 "HOST_BITS_PER_WIDE_INT >= 64" test.
14607 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
14609 (mov DI const_double): Delete.
14611 2013-05-04 Jakub Jelinek <jakub@redhat.com>
14613 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
14614 on op shows all bits zero in mode of a lowpart subreg, return zero.
14616 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
14619 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
14620 to save TFmode registers and DImode to save TImode registers for
14621 caller save operations.
14622 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
14623 mark being partially clobbered since they only use the first
14626 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
14627 and TDmode only use the upper 64-bits of each VSX register.
14629 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14631 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
14632 (find_candidates_in_block): Disable slsr_process_phi.
14634 2013-05-03 Guozhi Wei <carrot@google.com>
14636 * coverage.c (coverage_obj_init): Move the construction of gcov
14638 (build_init_ctor): ... here.
14640 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14642 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
14643 (slsr_cand_d): Redefine def_phi.
14644 (stride_status, phi_adjust_status, count_phis_status): New enums.
14645 (find_phi_def): New.
14646 (find_basis_for_base_expr): New.
14647 (find_basis_for_candidate): Handle hidden bases.
14648 (alloc_cand_and_find_basis): Handle phi candidates.
14649 (slsr_process_phi): New.
14650 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
14651 (create_mul_imm_cand): Likewise.
14652 (create_add_ssa_cand): Exclude phi base candidates.
14653 (create_add_imm_cand): Likewise.
14654 (slsr_process_cast): Likewise.
14655 (slsr_process_copy): Likewise.
14656 (find_candidates_in_block): Handle phi candidates.
14657 (dump_candidate): Likewise.
14658 (unconditional_cands): Delete.
14659 (unconditional_cands_with_known_stride_p): Delete.
14660 (phi_dependent_cand_p): New.
14661 (cand_increment): Handle phi-dependent candidates.
14662 (replace_dependent): Delete.
14663 (replace_mult_candidate): New.
14664 (replace_unconditional_candidate): New.
14665 (incr_vec_index): Move to avoid forward reference.
14666 (create_add_on_incoming_edge): New.
14667 (create_phi_basis): New.
14668 (replace_dependents): Delete.
14669 (replace_conditional_candidate): New.
14670 (phi_add_costs): New.
14671 (replace_uncond_cands_and_profitable_phis): New.
14672 (record_increment): Handle phi adjustments.
14673 (record_phi_increments): New.
14674 (record_increments): Handle phi adjustments.
14675 (phi_incr_cost): New.
14676 (lowest_cost_path): Handle phis.
14677 (total_savings): Likewise.
14678 (analyze_increments): Likewise.
14679 (ncd_with_phi): New.
14680 (ncd_of_cand_and_phis): New.
14681 (nearest_common_dominator_for_cands): Handle phi increments.
14682 (all_phi_incrs_profitable): New.
14683 (replace_profitable_candidates): Handle phi-dependent candidates.
14684 (analyze_candidates_and_replace): Likewise.
14686 2013-05-03 Teresa Johnson <tejohnson@google.com>
14689 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
14690 do not exceed REG_BR_PROB_BASE.
14692 2013-05-03 Jeff Law <law@redhat.com>
14694 PR tree-optimization/57144
14695 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
14696 operand of the condition will bit into the new type when eliminating
14697 a cast feeding a condition.
14699 2013-05-03 Jakub Jelinek <jakub@redhat.com>
14701 PR rtl-optimization/57130
14702 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
14703 of COMPARE as in_code to the recursive call if needed.
14705 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
14707 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
14708 (enabled): Handle new members.
14709 * config/i386/sse.md (*vec_concatv2si): Merge from
14710 *vec_concatv2si_sse2 and vec_concatv2si_sse.
14711 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
14713 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
14715 PR tree-optimization/57027
14716 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
14717 for fnms opportunity, check we got the prerequisite kind
14718 of tree / gimple before using accessor functions.
14720 2013-05-03 Richard Biener <rguenther@suse.de>
14722 * double-int.h (lshift): New overload without precision
14723 and arith argument.
14724 (operator *=, operator +=, operator -=): Move ...
14725 * double-int.c (operator *=, operator +=, operator -=): ... here
14726 and implement more efficiently.
14727 (mul_double_with_sign): Remove.
14728 (lshift_double): Adjust to take unsinged shift argument, push
14729 dispatching code to callers.
14730 (mul_double_wide_with_sign): Add early out for callers that
14731 are not interested in high parts or overflow.
14732 (lshift): New function.
14733 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
14734 dispatch code here.
14735 (lrotate, rrotate): Use logical shifts.
14736 * expr.c (get_inner_reference): Use lshift.
14737 * fixed-value.c (do_fixed_divide): Likewise.
14738 * tree-dfa.c (get_ref_base_and_extent): Likewise.
14739 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
14740 (indirect_refs_may_alias_p): Likewise.
14741 (stmt_kills_ref_p_1): Likewise.
14743 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
14745 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
14747 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
14749 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
14750 scalar form of FABD instruction.
14752 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
14754 * lra-constraints.c (process_alt_operands): Add checking alt
14755 number to choose the best alternative.
14757 2013-05-02 Richard Biener <rguenther@suse.de>
14759 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
14760 bitmap and its handling.
14761 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
14763 2013-05-02 Richard Biener <rguenther@suse.de>
14765 PR middle-end/57140
14766 * tree-inline.c (copy_loops): Properly handle removed loops.
14767 (copy_cfg_body): Mark destination loops for fixup if source
14768 loops needed fixup.
14770 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
14773 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
14774 generating simple_return for naked functions.
14776 2013-05-02 Martin Jambor <mjambor@suse.cz>
14778 PR middle-end/56988
14779 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
14780 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
14782 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
14783 ipa_agg_replacement_value structures.
14784 (known_aggs_to_agg_replacement_list): Likewise.
14785 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
14786 (read_agg_replacement_chain): Likewise.
14787 (ipcp_transform_function): Also check that by_ref flags match.
14789 2013-05-02 Richard Biener <rguenther@suse.de>
14791 * graphds.h (struct graph): Add obstack member.
14792 * graphds.c (new_graph): Initialize obstack and allocate
14794 (add_edge): Allocate edge from the obstack.
14795 (free_graph): Free the obstack instead of all edges and vertices.
14797 2013-05-02 Teresa Johnson <tejohnson@google.com>
14799 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
14801 * cfg.c (update_bb_profile_for_threading): Ditto.
14802 * tree-inline.c (copy_bb): Ditto.
14803 (copy_edges_for_bb): Ditto.
14804 (initialize_cfun): Ditto.
14805 (copy_cfg_body): Ditto.
14806 (expand_call_inline): Ditto.
14807 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
14808 (estimate_node_size_and_time): Ditto.
14809 (inline_merge_summary): Ditto.
14810 * cgraphclones.c (cgraph_clone_edge): Ditto.
14811 (cgraph_clone_node): Ditto.
14812 * sched-rgn.c (compute_dom_prob_ps): Ditto.
14813 (compute_trg_info): Ditto.
14815 2013-05-02 Ian Bolton <ian.bolton@arm.com>
14817 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
14818 S reg when fp attribute set.
14819 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
14821 2013-05-02 Ian Bolton <ian.bolton@arm.com>
14823 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
14825 (*and_one_cmplsi3_compare0_uxtw): Likewise.
14826 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
14827 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
14829 2013-05-02 Richard Biener <rguenther@suse.de>
14831 * tree-scalar-evolution.c (scev_info_hasher): Remove.
14832 (struct instantiate_cache_entry): New type.
14833 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
14834 (struct instantiate_cache_type): New type.
14835 (set_instantiated_value, get_instantiated_value): Remove.
14836 (get_instantiated_value_entry): New function.
14837 (instantiate_scev_name): Use the new cache and adjust.
14838 (instantiate_scev_poly): Adjust.
14839 (instantiate_scev_binary): Likewise.
14840 (instantiate_array_ref): Likewise.
14841 (instantiate_scev_convert): Likewise.
14842 (instantiate_scev_not): Likewise.
14843 (instantiate_scev_3): Likewise.
14844 (instantiate_scev_2): Likewise.
14845 (instantiate_scev_r): Likewise.
14846 (instantiate_scev): Likewise.
14847 (resolve_mixers): Likewise.
14849 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
14852 * lra-constraints.c (best_small_class_operands_num): Remove.
14853 (process_alt_operands): Remove small_class_operands_num. Take
14854 small classes operands into losers and only if the operand is not
14855 matched. Modify debugging output.
14856 (curr_insn_transform): Remove best_small_class_operands_num.
14859 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14861 * config/aarch64/aarch64-builtins.c
14862 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
14863 * config/aarch64/aarch64-simd-builtins.def
14864 (reduc_splus_): Add new modes.
14865 (reduc_uplus_): New.
14866 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
14867 (reduc_uplus_v4sf): Likewise.
14868 (reduc_splus_v4sf): Likewise.
14869 (aarch64_addv<mode>): Likewise.
14870 (reduc_uplus_<mode>): Likewise.
14871 (reduc_splus_<mode>): Likewise.
14872 (aarch64_addvv2di): Likewise.
14873 (reduc_uplus_v2di): Likewise.
14874 (reduc_splus_v2di): Likewise.
14875 (aarch64_addvv2si): Likewise.
14876 (reduc_uplus_v2si): Likewise.
14877 (reduc_splus_v2si): Likewise.
14878 (reduc_<sur>plus_<mode>): New.
14879 (reduc_<sur>plus_v2di): Likewise.
14880 (reduc_<sur>plus_v2si): Likewise.
14881 (reduc_<sur>plus_v4sf): Likewise.
14882 (aarch64_addpv4sf): Likewise.
14883 * config/aarch64/arm_neon.h
14884 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
14885 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
14886 add UNSPEC_SADDV, UNSPEC_UADDV.
14888 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
14890 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14892 * config/aarch64/arm_neon.h
14893 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
14895 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14897 * config/aarch64/aarch64-builtins
14898 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
14900 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14902 * config/aarch64/aarch64-simd-builtins.def
14903 (reduc_smax_): New.
14904 (reduc_smin_): Likewise.
14905 (reduc_umax_): Likewise.
14906 (reduc_umin_): Likewise.
14907 (reduc_smax_nan_): Likewise.
14908 (reduc_smin_nan_): Likewise.
14911 (smax): Update for V2SF, V4SF and V2DF modes.
14914 (smin_nan): Likewise.
14915 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
14916 (<su><maxmin><mode>3): ...This, refactor.
14917 (s<maxmin><mode>3): New.
14918 (<maxmin_uns><mode>3): Likewise.
14919 (reduc_<maxmin_uns>_<mode>): Refactor.
14920 (reduc_<maxmin_uns>_v4sf): Likewise.
14921 (reduc_<maxmin_uns>_v2si): Likewise.
14922 (aarch64_<fmaxmin><mode>: Remove.
14923 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
14925 (vmin<q>_f<32,64>): Likewise.
14926 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
14928 (su): Add mappings for smax, smin, umax, umin.
14930 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
14931 (FMAXMIN): Rename as...
14932 (FMAXMIN_UNS): ...This.
14934 (fmaxminv): Likewise.
14935 (fmaxmin): Likewise.
14937 (maxmin_uns_op): Likewise.
14939 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14941 * config/aarch64/arm_neon.h
14942 (vac<ge, gt><sd>_f<32, 64>): Rename to...
14943 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
14944 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
14946 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14948 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
14949 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
14951 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14953 * config/aarch64/aarch64-simd.md
14954 (vcond<mode>_internal): Handle special cases for constant masks.
14955 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
14956 (vcondu<mode><mode>): Likewise.
14957 (vcond<v_cmp_result><mode>): New.
14959 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14961 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
14962 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
14963 * config/aarch64/aarch64-simd-builtins.def
14964 (cmeq): Update to BUILTIN_VALLDI.
14969 * config/aarch64/arm_neon.h
14970 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
14971 to builtins or C as appropriate.
14973 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14975 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
14977 (cmhi): Rename to...
14979 * config/aarch64/aarch64-simd.md
14980 (simd_mode): Add SF.
14981 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
14982 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
14983 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
14984 (cstore<mode>_neg): ...This.
14985 * config/aarch64/iterators.md
14987 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
14988 (COMPARISONS): New.
14989 (UCOMPARISONS): Likewise.
14990 (optab): Add missing comparisons.
14996 (VCMP_S): Likewise.
14997 (VCMP_U): Likewise.
14998 (V_cmp_result): Add DF, SF modes.
14999 (v_cmp_result): Likewise.
15001 (vmtype): Likewise.
15002 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
15004 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
15006 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
15007 define_insn to define_insn_and_split.
15008 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
15009 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
15010 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
15011 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
15012 (thumb2_negscc): Likewise.
15014 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
15016 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
15018 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
15020 * config/arm/thumb2.md: Remove trailing whitespaces.
15022 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
15024 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
15025 Use gen_int_mode rather than GEN_INT.
15027 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
15029 * value-prof.c (stream_in_histogram_value): Remove the strayed
15032 2013-04-30 Richard Biener <rguenther@suse.de>
15034 PR middle-end/57122
15035 * cfghooks.c (split_edge): Properly check for the loop latch edge.
15037 2013-04-30 Richard Biener <rguenther@suse.de>
15039 PR middle-end/57107
15040 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
15042 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
15044 PR rtl-optimization/56957
15045 PR rtl-optimization/57105
15046 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
15047 variable. Use just INSN_UID for determining whether an insn
15048 should be only disconnected from the insn stream.
15049 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
15051 2013-04-30 Jakub Jelinek <jakub@redhat.com>
15053 PR tree-optimization/57104
15054 * tsan.c (instrument_expr): Don't instrument accesses to
15055 DECL_HARD_REGISTER VAR_DECLs.
15057 2013-04-30 Richard Biener <rguenther@suse.de>
15059 * function.h (loops_for_fn): New inline function.
15060 (set_loops_for_fn): Likewise.
15061 * cfgloop.h (place_new_loop): Add struct function parameter.
15062 (get_loop): Likewise.
15063 (get_loops): Likewise.
15064 (number_of_loops): Likewise.
15065 (fel_next): Adjust.
15066 (fel_init): Likewise.
15067 * cfg.c (get_loop_copy): Adjust.
15068 * cfgloop.c (flow_loops_dump): Likewise.
15069 (record_loop_exits): Likewise.
15070 (verify_loop_structure): Likewise.
15071 * cfgloopanal.c (mark_irreducible_loops): Likewise.
15072 (estimate_reg_pressure_cost): Likewise.
15073 (mark_loop_exit_edges): Likewise.
15074 * cfgloopmanip.c (place_new_loop): Likewise.
15075 (add_loop): Likewise.
15076 (duplicate_loop): Likewise.
15077 * graph.c (draw_cfg_nodes): Likewise.
15078 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
15079 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
15080 (extract_affine_chrec): Likewise.
15081 (build_scop_iteration_domain): Likewise.
15082 * graphite.c (graphite_initialize): Likewise.
15083 * ira-build.c (create_loop_tree_nodes): Likewise.
15084 (more_one_region_p): Likewise.
15085 (rebuild_regno_allocno_maps): Likewise.
15086 (mark_loops_for_removal): Likewise.
15087 (mark_all_loops_for_removal): Likewise.
15088 (remove_unnecessary_regions): Likewise.
15089 (ira_build): Likewise.
15090 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
15091 * loop-init.c (fix_loop_structure): Likewise.
15092 (gate_rtl_move_loop_invariants): Likewise.
15093 (gate_rtl_unswitch): Likewise.
15094 (gate_rtl_unroll_and_peel_loops): Likewise.
15095 (rtl_doloop): Likewise.
15096 * lto-streamer-in.c (input_cfg): Likewise.
15097 * lto-streamer-out.c (output_cfg): Likewise.
15098 * modulo-sched.c (sms_schedule): Likewise.
15099 * predict.c (tree_estimate_probability): Likewise.
15100 (tree_estimate_probability_driver): Likewise.
15101 (estimate_loops): Likewise.
15102 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
15103 (move_sese_region_to_fn): Likewise.
15104 (debug_loop_num): Likewise.
15105 * tree-chrec.c (chrec_evaluate): Likewise.
15106 (hide_evolution_in_other_loops_than_loop): Likewise.
15107 (chrec_component_in_loop_num): Likewise.
15108 (reset_evolution_in_loop): Likewise.
15109 (evolution_function_is_invariant_rec_p): Likewise.
15110 * tree-if-conv.c (main_tree_if_conversion): Likewise.
15111 * tree-inline.c (copy_loops): Likewise.
15112 (copy_cfg_body): Likewise.
15113 (tree_function_versioning): Likewise.
15114 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
15115 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
15117 (add_to_evolution_1): Likewise.
15118 (scev_const_prop): Likewise.
15119 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
15120 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
15121 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
15122 (tree_ssa_lim_initialize): Likewise.
15123 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
15124 (verify_loop_closed_ssa): Likewise.
15125 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
15126 (tree_ssa_loop_im): Likewise.
15127 (tree_ssa_loop_unswitch): Likewise.
15128 (tree_vectorize): Likewise.
15129 (check_data_deps): Likewise.
15130 (tree_ssa_loop_ivcanon): Likewise.
15131 (tree_ssa_loop_bounds): Likewise.
15132 (tree_complete_unroll): Likewise.
15133 (tree_complete_unroll_inner): Likewise.
15134 (tree_parallelize_loops): Likewise.
15135 (tree_ssa_loop_prefetch): Likewise.
15136 (tree_ssa_loop_ivopts): Likewise.
15137 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
15138 * tree-vectorizer.c (vectorize_loops): Likewise.
15140 2013-04-29 Mike Frysinger <vapier@gentoo.org>
15142 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
15143 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
15144 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
15145 with EABI_LINK_SPEC.
15147 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
15150 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
15153 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
15156 * lra-constraints.c (process_alt_operands): Discourage a bit more
15157 using memory for pseudos. Print cost dump for alternatives.
15158 Modify cost values for conflicts with early clobbers.
15159 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
15161 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
15164 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
15166 2013-04-29 Ian Bolton <ian.bolton@arm.com>
15168 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
15169 from/to S register.
15170 (movdi_aarch64): Support LDR/STR from/to D register.
15172 2013-04-29 Ian Bolton <ian.bolton@arm.com>
15174 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
15175 or higher by default.
15177 2013-04-29 Richard Biener <rguenther@suse.de>
15179 PR middle-end/57075
15180 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
15181 even if not adding abnormal edges for calls that can make
15184 2013-04-29 Richard Biener <rguenther@suse.de>
15186 PR middle-end/57103
15187 * tree-cfg.c (move_stmt_op): Fix condition under which to update
15189 (move_stmt_r): Remove redundant checking.
15191 2013-04-29 Teresa Johnson <tejohnson@google.com>
15194 * basic-block.h (apply_scale): New function.
15195 (apply_probability): Use apply_scale.
15196 * gimple-streamer-in.c (input_bb): Ditto.
15197 * lto-streamer-in.c (input_cfg): Ditto.
15198 * lto-cgraph.c (merge_profile_summaries): Ditto.
15199 * tree-optimize.c (execute_fixup_cfg): Ditto.
15200 * tree-inline.c (copy_bb): Update comment to use apply_scale.
15201 (copy_edges_for_bb): Ditto.
15202 (copy_cfg_body): Ditto.
15204 2013-04-29 Tom de Vries <tom@codesourcery.com>
15206 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
15207 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
15208 (tail_merge_optimize): Handle current_loops == NULL.
15210 2013-04-26 Jeff Law <law@redhat.com>
15212 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
15213 (simplify_cond_using_ranges): Generalize code to simplify
15214 COND_EXPRs where one argument is a constant and the other
15215 is an SSA_NAME created by an integral type conversion.
15217 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15219 * config/arm/arm.md (store_minmaxsi): Use only when
15220 optimize_insn_for_size_p.
15222 2013-04-29 Christian Bruel <christian.bruel@st.com>
15225 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
15227 2013-04-29 Richard Biener <rguenther@suse.de>
15229 PR middle-end/57089
15230 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
15231 loop tree make sure to schedule a fixup for the child as well.
15232 (expand_omp_for_generic): Properly add loops.
15233 (expand_omp_for_static_nochunk): Likewise.
15234 (expand_omp_for_static_chunk): Likewise.
15235 (expand_omp_for): For the degenerate case fixup loops.
15236 (expand_omp_sections): Fix default bb placement in loops.
15237 (expand_omp_atomic_pipeline): Properly add loops.
15239 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15241 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
15243 2013-04-29 Tom de Vries <tom@codesourcery.com>
15245 * tree-ssa-tail-merge.c: Update header comment.
15247 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15249 * config/aarch64/arm_neon.h
15250 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
15251 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
15252 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
15253 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
15254 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
15255 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
15256 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
15257 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
15259 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15261 * config/aarch64/aarch64-simd.md
15262 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
15263 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
15264 fix_trunc, fixuns_trunc.
15265 (ftrunc<VDQF:mode>2): New.
15266 * config/aarch64/iterators.md (optab): Add fix, fixuns.
15267 (fix_trunc_optab): New.
15269 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15271 * config/aarch64/aarch64-builtins.c
15272 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
15273 iceilf, lround, iroundf.
15275 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
15278 * config/i386/i386.h (enum ix86_tune_indices)
15279 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
15280 New, split from X86_TUNE_INTER_UNIT_MOVES.
15281 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
15282 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
15283 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
15284 (TARGET_INTER_UNIT_MOVES): Remove.
15285 * config/i386/i386.c (initial_ix86_tune_features): Update.
15286 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
15287 (ix86_expand_convert_uns_didf_sse): Use
15288 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
15289 (ix86_expand_vector_init_one_nonzero): Ditto.
15290 (ix86_expand_vector_init_interleave): Ditto.
15291 (inline_secondary_memory_needed): Return true for moves from SSE class
15292 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
15293 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
15294 * config/i386/constraints.md (Yi, Ym): Depend on
15295 TARGET_INTER_UNIT_MOVES_TO_VEC.
15296 (Yj, Yn): New constraints.
15297 * config/i386/i386.md (*movdi_internal): Change constraints of
15298 operand 1 from Yi to Yj and from Ym to Yn.
15299 (*movsi_internal): Ditto.
15300 (*movdf_internal): Ditto.
15301 (*movsf_internal): Ditto.
15302 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
15303 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
15304 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
15305 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
15306 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
15307 * config/i386/sse.md (movdi_to_sse): Ditto.
15308 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
15309 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
15310 TARGET_INTER_UNIT_MOVES.
15311 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
15312 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
15313 instead of TARGET_INTER_UNIT_MOVES.
15314 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
15315 operand 1 from Yi to Yj and from Ym to Yn.
15317 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15319 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
15320 (float_truncate_hi_): Likewise.
15321 (float_extend_lo_): Likewise.
15322 (float_truncate_lo_): Likewise.
15323 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
15324 (aarch64_float_extend_lo_v2df): Likewise.
15325 (vec_unpacks_hi_v4sf): Likewise.
15326 (aarch64_float_truncate_lo_v2sf): Likewise.
15327 (aarch64_float_truncate_hi_v4sf): Likewise.
15328 (vec_pack_trunc_v2df): Likewise.
15329 (vec_pack_trunc_df): Likewise.
15331 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15333 * config/aarch64/aarch64-builtins.c
15334 (aarch64_fold_builtin): Fold float conversions.
15335 * config/aarch64/aarch64-simd-builtins.def
15336 (floatv2si, floatv4si, floatv2di): New.
15337 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
15338 * config/aarch64/aarch64-simd.md
15339 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
15340 * config/aarch64/iterators.md (FLOATUORS): New.
15341 (optab): Add float, floatuns.
15342 (su_optab): Likewise.
15344 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15346 * config/aarch64/aarch64-builtins.c
15347 (aarch64_builtin_vectorized_function): Use new names for
15349 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
15350 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
15351 (fcvtzu): Split as...
15352 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
15353 (fcvtas): Split as...
15354 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
15355 (fcvtau): Split as...
15356 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
15357 (fcvtps): Split as...
15358 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
15359 (fcvtpu): Split as...
15360 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
15361 (fcvtms): Split as...
15362 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
15363 (fcvtmu): Split as...
15364 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
15365 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
15366 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
15367 (lfrintnusf, lfrintnudf): Likewise.
15368 * config/aarch64/aarch64-simd.md
15369 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
15371 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
15372 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
15373 (fcvt_pattern): Likewise.
15375 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15377 * config/aarch64/aarch64-simd.md
15378 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
15379 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
15381 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15383 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
15384 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
15385 (vrnd<a,m,n,p>_f32): Implement using builtins.
15386 (vrnd<i,x><q>_f<32, 64>): New.
15388 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15390 * config/aarch64/aarch64-builtins.c
15391 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
15392 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
15393 (frintz): Rename to...
15395 (frintp): Rename to...
15397 (frintm): Rename to...
15399 (frinti): Rename to...
15400 (nearbyint): ...this.
15401 (frintx): Rename to...
15403 (frinta): Rename to...
15405 * config/aarch64/aarch64-simd.md
15406 (aarch64_frint<frint_suffix><mode>): Delete.
15407 (<frint_pattern><mode>2): Convert to insn.
15408 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
15409 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
15410 (frint_pattern): Likewise.
15411 (frint_suffix): Likewise.
15413 2013-04-29 Richard Biener <rguenther@suse.de>
15415 PR tree-optimization/57081
15416 * loop-init.c: Include tree-flow.h.
15417 (loop_optimizer_finalize): Free number of iteration estimates.
15418 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
15420 2013-04-29 Jakub Jelinek <jakub@redhat.com>
15422 PR tree-optimization/57083
15423 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
15424 non-singleton shift count range, zero extend low_bound for uns case.
15426 * config/i386/predicates.md (general_vector_operand): New predicate.
15427 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
15428 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
15429 if they aren't nonimmediate operands. If their original values
15430 satisfy const_vector_equal_evenodd_p, don't shift them.
15431 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
15432 predicates. For the SSE4.1 case force operands[{1,2}] into registers
15433 if not nonimmediate_operand.
15434 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
15435 instead of register_operand.
15436 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
15438 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
15440 * stor-layout.c (finalize_size_functions): Allocate a structure and
15441 reset cfun before dumping the functions.
15443 2013-04-27 Jakub Jelinek <jakub@redhat.com>
15445 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
15448 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
15449 use xop_pmacsdqh if uns_p.
15450 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
15451 the immediate rotate count.
15453 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
15455 * rtl.h (struct rtx_def): Add comment for field jump.
15456 (LRA_SUBREG_P): New macro.
15457 * recog.c (register_operand): Check LRA_SUBREG_P.
15458 * lra.c (lra): Add note at the end of RTL code. Align non-empty
15460 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
15461 (lra_final_code_change): Skip subreg change for operators.
15462 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
15463 if there are no operand changes.
15464 * lra-constraints.c (curr_insn_set): New.
15465 (match_reload): Set LRA_SUBREG_P.
15466 (emit_spill_move): Ditto.
15467 (check_and_process_move): Use curr_insn_set. Process only single
15468 set insns. Don't initialize sec_mem_p and change_p.
15469 (simplify_operand_subreg): Use LRA_SUBREG_P.
15470 (reg_in_class_p): New function.
15471 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
15472 of #ifdef. Add code to remove cycling.
15473 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
15474 non-null disp. Reload inner instead of disp when base and index
15475 are null. Try to put lo_sum into register.
15476 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
15477 (check_and_process_move): Move code for move cost check to
15478 simple_move_p. Remove equiv_substitution.
15479 (simple_move_p): New function.
15480 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
15481 curr_insn_set. Call check_and_process_move only for single set
15482 insns. Use the new function. Move call of check_and_process_move
15483 after operand equiv substitution and address process.
15485 2013-04-26 Jakub Jelinek <jakub@redhat.com>
15488 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
15489 with nonlocal goto receivers or returns twice calls, ignore
15490 unininitialized values from abnormal edges to nl goto receiver
15491 or returns twice call.
15493 2013-04-26 Jakub Jelinek <jakub@redhat.com>
15495 PR tree-optimization/57051
15496 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
15497 and VEC_RSHIFT_EXPR if shift count is a multiple of element
15500 2013-04-26 Richard Biener <rguenther@suse.de>
15502 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
15503 (expand_omp_taskreg): Likewise. Mark loops for fixup.
15504 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
15505 (fixup_loop_arrays_after_move): New function.
15506 (move_sese_region_to_fn): Properly outline the loop tree parts
15507 of the SESE region.
15509 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
15511 * config/i386/i386.md (type, unit): Fix long lines.
15513 2013-04-26 Richard Biener <rguenther@suse.de>
15515 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
15516 (lto-streamer-out.o): Likewise.
15517 * cfgloop.c (init_loops_structure): Export, add struct function
15518 argument and adjust.
15519 (flow_loops_find): Adjust.
15520 * cfgloop.h (enum loop_estimation): Add EST_LAST.
15521 (init_loops_structure): Declare.
15522 * lto-streamer-in.c: Include cfgloop.h.
15523 (input_cfg): Input the loop tree.
15524 * lto-streamer-out.c: Include cfgloop.h.
15525 (output_cfg): Output the loop tree.
15526 (output_struct_function_base): Do not drop PROP_loops.
15528 2013-03-26 Richard Biener <rguenther@suse.de>
15530 * tree-cfg.c (execute_build_cfg): Build the loop tree.
15531 (pass_build_cfg): Provide PROP_loops.
15532 (move_sese_region_to_fn): Remove loops that are outlined into fn
15534 * tree-inline.c: Include cfgloop.h.
15535 (initialize_cfun): Do not drop PROP_loops.
15536 (copy_loops): New function.
15537 (copy_cfg_body): Copy loop structure.
15538 (tree_function_versioning): Initialize destination loop tree.
15539 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
15540 (pass_parallelize_loops): Do IL verification.
15541 * loop-init.c (loop_optimizer_init): Fixup loops if required.
15542 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
15543 the CFG make sure we fixup loops as well.
15544 * tree-ssa-tail-merge.c: Include cfgloop.h.
15545 (replace_block_by): When merging loop latches mark loops for fixup.
15546 * lto-streamer-out.c (output_struct_function_base): Drop
15547 PROP_loops for now.
15548 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
15549 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
15550 * ipa-split.c: Include cfgloop.h.
15551 (split_function): Add the new return block to the loop tree root.
15552 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
15553 whether we have removed the forwarder block.
15554 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
15555 * cfgloop.h (place_new_loop): Declare.
15556 * cfgloopmanip.c (place_new_loop): Export.
15557 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
15558 (tree-switch-conversion.o): Likewise.
15559 (tree-complex.o): Likewise.
15560 (tree-inline.o): Likewise.
15561 (tree-ssa-tailmerge.o): Likewise.
15562 (ipa-split.o): Likewise.
15563 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
15564 (tree-ssa-copy.o): Likewise.
15565 * tree-switch-conversion.c: Include cfgloop.h
15566 (process_switch): If we emit a bit-test cascade, schedule loops
15568 * tree-complex.c: Include cfgloop.h.
15569 (expand_complex_div_wide): Properly add new basic-blocks to loops.
15570 * asan.c: Include cfgloop.h.
15571 (create_cond_insert_point): Properly add new basic-blocks to
15572 loops, schedule loop fixup.
15573 * cfgloop.c (verify_loop_structure): Check that looks are not
15575 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
15577 (expand_omp_for_generic): Likewise.
15578 (expand_omp_sections): Likewise.
15579 (expand_omp_atomic_pipeline): Schedule loops for fixup.
15580 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
15581 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
15582 is initialized, not when loops are present.
15583 * tree-parloops.c (parallelize_loops): Remove checking here.
15584 * passes.c (init_optimization_passes): Schedule a copy-propagation
15585 pass before complete unrolling of inner loops.
15587 2013-04-26 Jakub Jelinek <jakub@redhat.com>
15589 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
15590 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
15591 (colorize_init): Add argument to _WIN32 version.
15592 * toplev.c: Include diagnostic-color.h.
15593 (process_options): Default to -fdiagnostics-color=auto if
15594 GCC_COLORS env var is in the environment.
15595 * common.opt (fdiagnostics-color=): Add Var and Init.
15596 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
15597 env var is in the environment, the default is auto rather than never.
15599 * diagnostic.h (file_name_as_prefix): Add context argument.
15600 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
15601 the string as locus.
15602 * langhooks.c (lhd_print_error_function): Adjust caller.
15604 2013-04-25 Lawrence Crowl <crowl@google.com>
15606 * var-tracking.c (shared_hash_def::htab):
15607 Change type to hash_table. Update dependent calls and types.
15609 2013-04-25 Lawrence Crowl <crowl@google.com>
15611 * Makefile.in: Update as needed below.
15613 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
15614 Move declaration to after the type's method definitons.
15616 * attribs.c (htab_t scoped_attributes::attribute_hash):
15617 Change type to hash_table. Update dependent calls and types.
15619 * bitmap.c (htab_t bitmap_desc_hash):
15620 Change type to hash_table. Update dependent calls and types.
15622 * cselib.c (htab_t cselib_hash_table):
15623 Change type to hash_table. Update dependent calls and types.
15625 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
15626 (hash_string_slot_node): Move implementation into lto-streamer.h
15627 struct string_slot_hasher.
15628 (eq_string_slot_node): Likewise.
15630 * data-streamer-out.c: Update output_block::string_hash_table
15631 dependent calls and types.
15633 * dwarf2cfi.c (htab_t trace_index):
15634 Change type to hash_table. Update dependent calls and types.
15636 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
15637 Change type to hash_table. Update dependent calls and types.
15638 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
15639 (htab_t optimize_external_refs::map): Likewise.
15640 (htab_t output_comp_unit::extern_map): Likewise.
15641 (htab_t output_comdat_type_unit::extern_map): Likewise.
15642 (htab_t output_macinfo::macinfo_htab): Likewise.
15643 (htab_t optimize_location_lists::htab): Likewise.
15644 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
15646 * except.c (htab_t ehspec_hash_type):
15647 Change type to hash_table. Update dependent calls and types.
15648 (assign_filter_values::ttypes): Likewise.
15649 (assign_filter_values::ehspec): Likewise.
15650 (sjlj_assign_call_site_values::ar_hash): Likewise.
15651 (convert_to_eh_region_ranges::ar_hash): Likewise.
15653 * gcse.c (htab_t pre_ldst_table):
15654 Change type to hash_table. Update dependent calls and types.
15656 * ggc-common.c (htab_t saving_htab):
15657 Change type to hash_table. Update dependent calls and types.
15658 (htab_t loc_hash): Likewise.
15659 (htab_t ptr_hash): Likewise.
15660 (call_count): Rename ggc_call_count.
15661 (call_alloc): Rename ggc_call_alloc.
15662 (loc_descriptor): Rename make_loc_descriptor.
15663 (add_statistics): Rename ggc_add_statistics.
15665 * ggc-common.c (saving_htab):
15666 Change type to hash_table. Update dependent calls and types.
15668 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
15669 (push_gimplify_context): Likewise.
15670 (pop_gimplify_context): Likewise.
15671 (struct gimple_temp_hash_elt): Added.
15672 (struct gimplify_hasher): Likewise.
15673 (struct gimplify_ctx.temp_htab):
15674 Change type to hash_table. Update dependent calls and types.
15676 * gimple-fold.c: Include gimplify-ctx.h.
15678 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
15679 Change type to hash_table. Update dependent calls and types.
15680 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
15681 avoid potential global name collision.
15683 * gimplify.c: Include gimplify-ctx.h.
15684 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
15685 (htab_t gimplify_ctx::temp_htab):
15686 Update dependent calls and types for new type hash_table.
15687 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
15688 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
15690 * gimplify-ctx.h: New.
15691 (struct gimple_temp_hash_elt): Move from gimplify.c.
15692 (class gimplify_hasher): New.
15693 (struct gimplify_ctx): Move from gimple.h.
15694 (htab_t gimplify_ctx::temp_htab):
15695 Change type to hash_table. Update dependent calls and types.
15697 * graphite-clast-to-gimple.c: Include graphite-htab.h.
15698 (htab_t ivs_params::newivs_index):
15699 Change type to hash_table. Update dependent calls and types.
15700 (htab_t ivs_params::params_index): Likewise.
15701 (htab_t print_generated_program::params_index): Likewise.
15702 (htab_t gloog::newivs_index): Likewise.
15703 (htab_t gloog::params_index): Likewise.
15705 * graphite.c: Include graphite-htab.h.
15706 4htab_t graphite_transform_loops::bb_pbb_mapping):
15707 Change type to hash_table. Update dependent calls and types.
15709 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
15710 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
15711 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
15713 * graphite-dependences.c: Include graphite-htab.h.
15714 (loop_is_parallel_p): Change hash table type of parameter.
15716 * graphite-htab.h: New.
15717 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
15718 (extern find_pbb_via_hash): Move from graphite-poly.h.
15719 (extern loop_is_parallel_p): Move from graphite-poly.h.
15720 (extern get_loop_body_pbbs): Move from graphite-poly.h.
15722 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
15723 (extern loop_is_parallel_p): Move to graphite-htab.h.
15724 (extern get_loop_body_pbbs): Move to graphite-htab.h.
15726 * haifa-sched.c (htab_t delay_htab):
15727 Change type to hash_table. Update dependent calls and types.
15728 (htab_t delay_htab_i2): Likewise.
15730 * ira-color.c (htab_t allocno_hard_regs_htab):
15731 Change type to hash_table. Update dependent calls and types.
15733 * ira-costs.c (htab_t cost_classes_htab):
15734 Change type to hash_table. Update dependent calls and types.
15736 * loop-invariant.c (htab_t merge_identical_invariants::eq):
15737 Change type to hash_table. Update dependent calls and types.
15739 * loop-iv.c (htab_t bivs):
15740 Change type to hash_table. Update dependent calls and types.
15742 * loop-unroll.c (htab_t opt_info::insns_to_split):
15743 Change type to hash_table. Update dependent calls and types.
15744 (htab_t opt_info::insns_with_var_to_expand): Likewise.
15746 * lto-streamer.h (struct string_slot): Move from data-streamer.h
15747 (struct string_slot_hasher): New.
15748 (htab_t output_block::string_hash_table):
15749 Change type to hash_table. Update dependent calls and types.
15751 * lto-streamer-in.c (freeing_string_slot_hasher): New.
15752 (htab_t file_name_hash_table):
15753 Change type to hash_table. Update dependent calls and types.
15755 * lto-streamer-out.c: Update output_block::string_hash_table dependent
15758 * lto-streamer.c (htab_t tree_htab):
15759 Change type to hash_table. Update dependent calls and types.
15761 * omp-low.c: Include gimplify-ctx.h.
15763 * passes.c (htab_t name_to_pass_map):
15764 Change type to hash_table. Update dependent calls and types.
15765 (pass_traverse): Rename to passes_pass_traverse.
15767 * plugin.c (htab_t event_tab):
15768 Change type to hash_table. Update dependent calls and types.
15770 * postreload-gcse.c (htab_t expr_table):
15771 Change type to hash_table. Update dependent calls and types.
15772 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
15774 * sese.c (debug_rename_map_1): Make extern.
15775 (htab_t copy_bb_and_scalar_dependences::rename_map):
15776 Change type to hash_table. Update dependent calls and types.
15778 * sese.h (extern debug_rename_map): Move to .c file.
15780 * store-motion.c (htab_t store_motion_mems_table):
15781 Change type to hash_table. Update dependent calls and types.
15783 * trans-mem.c (htab_t tm_new_mem_hash):
15784 Change type to hash_table. Update dependent calls and types.
15786 * tree-browser.c (htab_t TB_up_ht):
15787 Change type to hash_table. Update dependent calls and types.
15789 * tree-cfg.c (htab_t discriminator_per_locus):
15790 Change type to hash_table. Update dependent calls and types.
15792 * tree-complex.c: Include tree-hasher.h
15793 (htab_t complex_variable_components):
15794 Change type to hash_table. Update dependent calls and types.
15796 * tree-eh.c (htab_t finally_tree):
15797 Change type to hash_table. Update dependent calls and types.
15799 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
15800 struct int_tree_hasher.
15801 (extern int_tree_map_eq): Likewise.
15802 (uid_decl_map_hash): Removed.
15803 (extern decl_tree_map_eq): Likewise.
15805 * tree-hasher.h: New.
15806 (struct int_tree_hasher): New.
15807 (typedef int_tree_htab_type): New.
15809 * tree-inline.c: Include gimplify-ctx.h.
15811 * tree-mudflap.c: Include gimplify-ctx.h.
15813 * tree-parloops.c: Include tree-hasher.h.
15814 (htab_t eliminate_local_variables_stmt::decl_address):
15815 Change type to hash_table. Update dependent calls and types.
15816 (htab_t separate_decls_in_region::decl_copies): Likewise.
15818 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
15819 Change type to hash_table. Update dependent calls and types.
15821 * tree-sra.c (candidates):
15822 Change type to hash_table. Update dependent calls and types.
15824 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
15826 (int_tree_map_hash): Likewise.
15828 * tree-ssa-dom.c (htab_t avail_exprs):
15829 Change type to hash_table. Update dependent calls and types.
15831 * tree-ssa-live.c (var_map_base_init::tree_to_index):
15832 Change type to hash_table. Update dependent calls and types.
15834 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
15835 Change type to hash_table. Update dependent calls and types.
15837 * tree-ssa-phiopt.c (seen_ssa_names):
15838 Change type to hash_table. Update dependent calls and types.
15840 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
15841 Change type to hash_table. Update dependent calls and types.
15843 * tree-ssa-uncprop.c (equiv):
15844 Change type to hash_table. Update dependent calls and types.
15846 2013-04-25 Jakub Jelinek <jakub@redhat.com>
15848 PR rtl-optimization/57003
15849 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
15850 call note_stores with kill_clobbered_value callback again after
15851 killing regs_invalidated_by_call.
15853 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
15855 * config/aarch64/aarch64-simd.md
15856 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
15857 (aarch64_simd_bsl<mode>): Likewise.
15858 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
15860 2013-04-25 Marek Polacek <polacek@redhat.com>
15862 PR tree-optimization/57066
15863 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
15865 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
15867 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
15869 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
15871 * config/aarch64/aarch64-builtins.c
15872 (aarch64_fold_builtin): New.
15873 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
15874 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
15875 * config/aarch64/aarch64-simd-builtins.def (abs): New.
15876 * config/aarch64/arm_neon.h
15877 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
15879 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
15880 Tejas Belagod <tejas.belagod@arm.com>
15882 * config/aarch64/aarch64-builtins.c
15883 (aarch64_gimple_fold_builtin): New.
15884 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
15885 * config/aarch64/aarch64-simd-builtins.def (addv): New.
15886 * config/aarch64/aarch64-simd.md (addpv4sf): New.
15887 (addvv4sf): Update.
15888 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
15890 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
15892 * config/aarch64/aarch64.md
15893 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
15895 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
15897 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
15898 (*ngcsi_uxtw): New pattern.
15900 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15901 Julian Brown <julian@codesourcery.com>
15903 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
15904 (TB_DREG): Add T_V4HF.
15905 (v4hf_UP): New macro.
15906 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
15907 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
15908 Handle initialisation of V4HF. Adjust initialisation of reinterpret
15910 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
15911 (arm_vector_mode_supported_p): Handle V4HF.
15912 (arm_mangle_map): Handle V4HFmode.
15913 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
15914 * config/arm/arm_neon_builtins.def: Add entries for
15915 vcvtv4hfv4sf, vcvtv4sfv4hf.
15916 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
15917 (neon_vcvtv4hfv4sf): Likewise.
15918 * config/arm/neon-gen.ml: Handle half-precision floating point
15920 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
15921 * config/arm/arm_neon.h: Regenerate.
15922 * config/arm/neon.ml (type elts): Add F16.
15923 (type vectype): Add T_float16x4, T_floatHF.
15924 (type vecmode): Add V4HF.
15925 (type features): Add Requires_FP_bit feature.
15926 (elt_width): Handle F16.
15927 (elt_class): Likewise.
15928 (elt_of_class_width): Likewise.
15929 (mode_of_elt): Refactor.
15930 (type_for_elt): Handle F16, fix error messages.
15931 (vectype_size): Handle T_float16x4.
15932 (vcvt_sh): New function.
15933 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
15934 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
15935 (string_of_mode): Handle V4HF.
15936 * doc/arm-neon-intrinsics.texi: Regenerate.
15938 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
15940 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
15941 format specifier in 'X' case.
15943 2013-04-25 Alan Modra <amodra@gmail.com>
15946 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
15947 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
15948 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
15949 Repeat for many other rotate/shift and mask patterns using subregs.
15950 Name lshiftrt insns.
15951 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
15952 on WORDS_BIG_ENDIAN.
15954 2013-04-25 Alan Modra <amodra@gmail.com>
15956 * config.gcc: Support little-endian powerpc-linux targets.
15957 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
15958 (LINK_OS_LINUX_SPEC): Define.
15959 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
15960 Preserve MASK_LITTLE_ENDIAN.
15961 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
15962 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
15963 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
15964 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
15965 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
15966 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
15967 Correct fp word order for little-endian. Don't shift toc entries
15968 smaller than a word for little-endian.
15969 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
15970 (bswapdi2 splits): Correct low-part subreg for little-endian.
15971 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
15972 low/high where such is correct only for be.
15973 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
15974 little-endian for -mcall-aixdesc.
15976 2013-04-25 Alan Modra <amodra@gmail.com>
15978 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
15979 replace_equiv_address_nv.
15981 2013-04-25 Alan Modra <amodra@gmail.com>
15983 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
15985 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
15988 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
15989 * rtl.h (struct rtx_def): ...
15991 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
15993 PR rtl-optimizations/57046
15994 * lra-constraints (split_reg): Set up lra_risky_transformations_p
15995 for multi-reg splits.
15997 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
15999 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
16001 2013-04-24 Sterling Augustine <saugustine@google.com>
16003 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
16004 (comp_dir_string, debug_str_dwo_section): New.
16005 (DEBUG_STR_DWO_SECTION): Rename to ...
16006 (DEBUG_DWO_STR_SECTION): ... this.
16007 (DEBUG_NORM_STR_SECTION): Delete.
16008 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
16009 (DEBUG_STR_DWO_SECTION_FLAGS): New.
16010 (find_AT_string): Move most logic to ...
16011 (find_AT_string_in_table): ... here. New.
16012 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
16013 add_skeleton_AT_string. Delete logic.
16014 (output_skeleton_debug_sections): Remove call to
16015 add_top_level_skeleton_die_attrs.
16016 (add_comp_dir_attribute): Move logic to comp_dir_string.
16017 (dwarf2out_init): Initialize debug_str_dwo_section.
16018 (output_indirect_string): Call find_string_form.
16019 (output_indirect_strings): Rewrite.
16020 (prune_unused_types): Empty skeleton_debug_str_hash.
16021 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
16022 (dwarf2out_finish): Call output_indirect_strings.
16024 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
16026 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
16028 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
16030 * rtl.h (struct rtx_def): Add comment for field jump.
16031 (LRA_SUBREG_P): New macro.
16032 * recog.c (register_operand): Check LRA_SUBREG_P.
16033 * lra.c (lra): Add note at the end of RTL code. Align non-empty
16035 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
16036 (lra_final_code_change): Skip subreg change for operators.
16037 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
16038 if there are no operand changes.
16039 * lra-constraints.c (curr_insn_set): New.
16040 (match_reload): Set LRA_SUBREG_P.
16041 (emit_spill_move): Ditto.
16042 (check_and_process_move): Use curr_insn_set. Process only single
16043 set insns. Don't initialize sec_mem_p and change_p.
16044 (simplify_operand_subreg): Use LRA_SUBREG_P.
16045 (reg_in_class_p): New function.
16046 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
16047 of #ifdef. Add code to remove cycling.
16048 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
16049 non-null disp. Reload inner instead of disp when base and index
16050 are null. Try to put lo_sum into register.
16051 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
16052 (check_and_process_move): Move code for move cost check to
16053 simple_move_p. Remove equiv_substitution.
16054 (simple_move_p): New function.
16055 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
16056 curr_insn_set. Call check_and_process_move only for single set
16057 insns. Use the new function. Move call of check_and_process_move
16058 after operand equiv substitution and address process.
16060 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
16062 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
16063 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
16064 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
16066 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
16068 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
16070 2013-04-24 Marek Polacek <polacek@redhat.com>
16072 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
16073 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
16074 (select_loops_exit_conditions): Likewise.
16075 (number_of_iterations_for_all_loops): Likewise.
16076 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
16077 (scev_analysis): Likewise.
16079 2013-04-02 Catherine Moore <clm@codesourcery.com>
16080 Chao-ying Fu <fu@mips.com>
16082 * config/mips/micromips.md (jraddiusp): New pattern.
16083 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
16084 instruction if possible.
16086 2013-04-24 Alan Modra <amodra@gmail.com>
16088 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
16090 2013-04-24 Julian Brown <julian@codesourcery.com>
16091 Chung-Lin Tang <cltang@codesourcery.com>
16093 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
16094 dependency behavior in enumeration type DIE generation. Add TODO note
16095 to comments about future DW_FORM_sdata/udata re-work of related code.
16097 2013-04-23 Lawrence Crowl <crowl@google.com>
16099 * Makefile.in: Update as needed below.
16101 * hash-table.h (class hash_table):
16102 Correct many methods with parameter types compare_type to the correct
16103 value_type. (Correct code was unlikely to notice the change.)
16104 (hash_table::elements_with_deleted) New.
16105 (class hashtable::iterator): New.
16106 (hashtable::begin()): New.
16107 (hashtable::end()): New.
16108 (FOR_EACH_HASH_TABLE_ELEMENT): New.
16110 * statistics.c (statistics_hashes):
16111 Change type to hash_table. Update dependent calls and types.
16113 * tree-into-ssa.c (var_infos):
16114 Change type to hash_table. Update dependent calls and types.
16116 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
16117 Change type to hash_table. Update dependent calls and types.
16119 * tree-ssa-loop-im.c (struct mem_ref.refs):
16120 Change type to hash_table. Update dependent calls and types.
16122 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
16123 Change type to hash_table. Update dependent calls and types.
16125 * tree-ssa-sccvn.c (vn_tables_s::nary):
16126 Change type to hash_table. Update dependent calls and types.
16127 (vn_tables_s::phis): Likewise.
16128 (vn_tables_s::references): Likewise.
16130 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
16131 (vn_reference_eq): Update parameter and return types.
16133 * tree-ssa-structalias.c (pointer_equiv_class_table):
16134 Change type to hash_table. Update dependent calls and types.
16135 (location_equiv_class_table): Likewise.
16137 * tree-vect-data-refs.c: Consequential changes for making
16138 peeling a hash_table.
16140 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
16141 (destroy_loop_vec_info): Dependent hash_table update.
16143 * tree-vectorizer.h (peeling_htab):
16144 Change type to hash_table. Update dependent calls and types.
16146 2013-04-23 Shiva Chen <shiva0217@gmail.com>
16148 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
16149 to check the register content is equal or not.
16150 * lra-constraints.c (match_reload): Use lra_assign_reg_val
16151 to assign register content record.
16152 * lra-eliminations.c (update_reg_eliminate): Use
16153 lra_update_reg_val_offset to update register content offset.
16154 * lra-int.h (struct lra_reg): Add offset member.
16155 (lra_reg_val_equal_p): New static inline function.
16156 (lra_update_reg_val_offset): New static inline function.
16157 (lra_assign_reg_val): New static inline function.
16158 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
16159 to assign register content record.
16160 (initialize_lra_reg_info_element): Initial offset to zero.
16162 2013-04-23 Catherine Moore <clm@codesourcery.com>
16164 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
16165 operands. Record compression.
16167 2013-04-23 Xinliang David Li <davidxl@google.com>
16169 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
16171 2013-04-23 Richard Biener <rguenther@suse.de>
16173 PR middle-end/57036
16174 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
16175 parameter, only add abnormal goto edges from the copied body
16176 if the call could perform abnormal gotos.
16177 (copy_cfg_body): Adjust.
16179 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
16181 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
16183 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
16185 * coretypes.h (gimple_stmt_iterator): Add struct to make
16188 2013-04-23 Richard Biener <rguenther@suse.de>
16190 PR tree-optimization/57026
16191 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
16192 from SSA names occuring in abnormal PHI nodes.
16194 2013-04-22 Andi Kleen <ak@linux.intel.com>
16196 * lto/lto.c (print_lto_report_1): Fix LTO report names.
16198 2013-04-22 Andi Kleen <ak@linux.intel.com>
16200 * lto/lto.c (print_lto_report_1): Declare early.
16201 (read_cgraph_and_symbols): Call print_lto_report_1 early.
16203 2013-04-22 Andi Kleen <ak@linux.intel.com>
16205 * common.opt (-flto-report-wpa): Add.
16206 * doc/invoke.texi (-flto-report-wpa): Add.
16207 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
16210 2013-04-22 Xinliang David Li <davidxl@google.com>
16212 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
16213 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
16214 * Makefile.in: New dependency
16216 David Daney <ddaney.cavm@gmail.com>
16218 * configure.ac (gcc_cv_as_micromips_support): Use the
16219 --fatal-warnings option.
16220 * configure: Regenerate.
16222 2013-04-22 Marek Polacek <polacek@redhat.com>
16225 * tsan.c (instrument_expr): Don't instrument expression
16226 in case its size is zero.
16228 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
16232 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
16234 * config/alpha/alpha.c (TARGET_LRA_P): New define.
16236 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
16238 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
16239 (gimple_stmt_iterator): New typedef.
16240 * gimple.h (gimple_stmt_iterator): Rename to...
16241 (gimple_stmt_iterator_d): ... This.
16242 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
16243 trees be valid for GIMPLE and GENERIC.
16244 (TARGET_GIMPLE_FOLD_BUILTIN): New.
16245 * gimple-fold.c (gimple_fold_call): Call target hook
16246 gimple_fold_builtin.
16247 * hooks.c (hook_bool_gsiptr_false): New.
16248 * hooks.h (hook_bool_gsiptr_false): New.
16249 * target.def (fold_stmt): New.
16250 * doc/tm.texi: Regenerate.
16252 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
16255 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
16256 a set sp if no stack realignment.
16258 2013-04-22 Nick Clifton <nickc@redhat.com>
16260 * config.gcc (tilegx-linux): Extend extra_objs rather than
16262 (tilepro-linux): Likewise.
16264 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
16266 * config/aarch64/aarch64-builtins.c
16268 (CF0, CF1, CF2, CF3, CF4, CF10): New.
16269 (VAR<1-12>): Add MAP parameter.
16270 (BUILTIN_*): Likewise.
16271 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
16272 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
16273 (aarch64_ushl_n<mode>): Likewise.
16274 (aarch64_sshr_n<mode>): Likewise.
16275 (aarch64_ushr_n<mode>): Likewise.
16276 (aarch64_<maxmin><mode>): Likewise.
16277 (aarch64_sqrt<mode>): Likewise.
16278 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
16279 (vshr<q>_n_*): Likewise.
16281 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
16283 * config/aarch64/aarch64-builtins.c
16284 (aarch64_simd_builtin_type_mode): Handle SF types.
16286 (BUILTIN_GPF): Define.
16287 (aarch64_init_simd_builtins): Handle SF types.
16288 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
16289 (frecps): Likewise.
16290 (frecpx): Likewise.
16291 * config/aarch64/aarch64-simd.md
16292 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
16293 (aarch64_frecpe<mode>): New.
16294 (aarch64_frecps<mode>): Likewise.
16295 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
16296 (v8type): Add frecp<esx>.
16297 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
16298 (aarch64_frecps<mode>): Likewise.
16299 * config/aarch64/iterators.md (FRECP): New.
16300 (frecp_suffix): Likewise.
16301 * config/aarch64/arm_neon.h
16302 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
16304 2013-04-22 Christian Bruel <christian.bruel@st.com>
16307 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
16308 (REG_CLASS_NAMES): Idem.
16309 (REG_CLASS_CONTENTS): Idem.
16310 (REGCLASS_HAS_FP_REG): Idem.
16311 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
16312 (sh_conditional_register_usage): Idem.
16314 2013-04-21 Jeff Law <law@redhat.com>
16316 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
16317 (ssa_forward_propagate_and_combine): Use it.
16319 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
16321 * lra.c: Update the flow chart diagram.
16323 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
16325 PR rtl-optimization/56847
16326 * lra-constraints.c (process_alt_operands): Discourage alternative
16327 with non-matche doffsettable memory constraint fro memory with
16330 2013-04-19 Richard Biener <rguenther@suse.de>
16332 PR tree-optimization/56982
16333 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
16335 * gimplify.c (gimplify_call_expr): Notice special calls.
16336 (gimplify_modify_expr): Likewise.
16337 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
16338 abnormal control flow receivers.
16339 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
16340 in the same way as cfun->has_nonlocal_labels.
16341 (gimple_purge_dead_abnormal_call_edges): Likewise.
16342 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
16343 receivers start a basic-block.
16345 2013-04-19 Richard Biener <rguenther@suse.de>
16347 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
16349 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
16350 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
16351 (SLP_TREE_LOAD_PERMUTATION): Add.
16352 (vect_transform_slp_perm_load): Adjust prototype.
16353 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
16354 (vect_free_slp_instance): Likewise.
16355 (vect_create_new_slp_node): Likewise.
16356 (vect_supported_slp_permutation_p): Remove.
16357 (vect_slp_rearrange_stmts): Adjust.
16358 (vect_supported_load_permutation_p): Likewise. Inline
16359 vect_supported_slp_permutation_p here.
16360 (vect_analyze_slp_instance): Compute load permutations per
16361 slp node instead of per instance.
16362 (vect_get_slp_defs): Adjust.
16363 (vect_transform_slp_perm_load): Likewise.
16364 (vect_schedule_slp_instance): Remove redundant code.
16365 (vect_schedule_slp): Remove hack for PR56270, add it ...
16366 * tree-vect-stmts.c (vectorizable_load): ... here, do not
16367 CSE loads for SLP. Adjust.
16369 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
16371 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
16372 spelling in two comments.
16374 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
16377 * config/arm/arm.c (load_multiple_sequence): Require SP
16378 as base register for loads if SP is in the register list.
16380 2013-04-19 Martin Jambor <mjambor@suse.cz>
16382 PR tree-optimization/56718
16383 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
16384 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
16385 and made public. Adjusted all callers.
16386 (ipa_intraprocedural_devirtualization): New function.
16387 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
16388 (ipa_intraprocedural_devirtualization): Likewise.
16389 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
16391 2013-04-19 Richard Biener <rguenther@suse.de>
16393 PR tree-optimization/57000
16394 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
16396 2013-04-19 Terry Guo <terry.guo@arm.com>
16398 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
16400 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
16401 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
16402 (cortex_m4_fmacs): Use new reservations.
16403 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
16405 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
16407 PR rtl-optimization/56999
16408 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
16410 (lra_coalesce): Remove split_origin_bitmap and related code.
16411 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
16412 ranges if necessary.
16414 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
16416 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
16418 (ix86_expand_call): Remove clobbered_registers array and use
16419 x86_64_ms_sysv_extra_clobbered_registers instead.
16420 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
16422 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
16424 * config/i386/i386.md (*call_rex64_ms_sysv): Use
16425 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
16426 (*call_value_rex64_ms_sysv): Ditto.
16428 2013-04-18 Cary Coutant <ccoutant@google.com>
16430 * dwarf2out.c (output_pubnames): Check die_perennial_p of
16431 parent instead of die_mark.
16433 2013-04-18 Diego Novillo <dnovillo@google.com>
16435 * gimple.c (create_gimple_tmp): New.
16436 (get_expr_type): New.
16437 (build_assign): New.
16438 (build_type_cast): New.
16439 * gimple.h (enum ssa_mode): Define.
16440 (gimple_seq_set_location): New.
16441 * asan.c (build_check_stmt): Change some gimple_build_* calls
16442 to use build_assign and build_type_cast.
16444 2013-04-18 Richard Biener <rguenther@suse.de>
16446 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
16447 handle negative step. Remove redundant checks.
16448 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
16449 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
16450 for negative step and grouped loads fail to vectorize.
16452 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
16454 * emit-rtl.c (reset_insn_used_flags): New function.
16455 (reset_all_used_flags): Use it.
16456 (verify_insn_sharing): New function.
16457 (verify_rtl_sharing): Fix verification for SEQUENCEs.
16459 2013-04-18 Jakub Jelinek <jakub@redhat.com>
16461 PR tree-optimization/56984
16462 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
16463 and (x >> M) >= N don't register any assertion if N << M is the
16466 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
16468 * lower-subreg.c (resolve_simple_move): If called self-recursive,
16469 do not delete_insn insns that have not yet been emitted, only
16470 unlink them with remove_insn.
16471 * df-scan.c (df_insn_delete): Revert r197492.
16473 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
16475 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
16476 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
16478 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
16480 * config/arm/arm.md (movsicc_insn): Convert define_insn into
16481 define_insn_and_split.
16482 (and_scc,ior_scc,negscc): Likewise.
16483 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
16485 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
16487 * config/arm/arm.c (use_return_insn): Return 0 for targets that
16488 can benefit from using a sequence of LDRD instructions in epilogue
16489 instead of a single LDM instruction.
16491 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
16494 * doc/extend.texi: Fix typo.
16496 2013-04-17 Richard Biener <rguenther@suse.de>
16498 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
16499 (vect_build_slp_tree): ... here.
16500 (vect_build_slp_tree_1): Compute which stmts of the SLP group
16501 match. Remove special-casing of mismatched complex loads.
16502 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
16503 re-try the match with swapped commutative operands.
16504 (vect_supported_load_permutation_p): Remove special-casing of
16505 mismatched complex loads.
16506 (vect_analyze_slp_instance): Adjust.
16508 2013-04-17 Richard Biener <rguenther@suse.de>
16510 PR rtl-optimization/56921
16511 * cfgloop.h (struct loop): Add simple_loop_desc member.
16512 (struct niter_desc): Mark with GTY(()).
16513 (simple_loop_desc): Do not use aux field but simple_loop_desc.
16514 * loop-iv.c (get_simple_loop_desc): Likewise.
16515 (free_simple_loop_desc): Likewise.
16518 2013-04-16 Richard Biener <rguenther@suse.de>
16520 PR rtl-optimization/56921
16521 * loop-init.c (pass_rtl_move_loop_invariants): Add
16522 TODO_do_not_ggc_collect to todo_flags_finish.
16523 (pass_rtl_unswitch): Same.
16524 (pass_rtl_unroll_and_peel_loops): Same.
16525 (pass_rtl_doloop): Same.
16527 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
16529 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
16530 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
16531 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
16533 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
16534 * tree-streamer.c (record_common_node): Adjust reference in comment.
16536 2013-04-17 Terry Guo <terry.guo@arm.com>
16538 * config/arm/cortex-m4.md: Add a new bypass.
16540 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16542 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
16544 (*subs_<optab><mode>_multp2): New pattern.
16545 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
16546 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
16548 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16550 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
16551 (*subs_mul_imm_<mode>): New pattern.
16553 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
16556 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
16557 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
16558 (vsx_movti_32bit): Change j->wa to O->wa.
16560 2013-04-16 Richard Biener <rguenther@suse.de>
16562 PR rtl-optimization/56921
16563 * loop-init.c (pass_rtl_move_loop_invariants): Add
16564 TODO_do_not_ggc_collect to todo_flags_finish.
16565 (pass_rtl_unswitch): Same.
16566 (pass_rtl_unroll_and_peel_loops): Same.
16567 (pass_rtl_doloop): Same.
16569 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
16571 * config/arm/arm.c (emit_multi_reg_push): New declaration
16572 for an existing function.
16573 (arm_emit_strd_push): New function.
16574 (arm_expand_prologue): Used here.
16575 (arm_emit_ldrd_pop): New function.
16576 (arm_expand_epilogue): Used here.
16577 (arm_get_frame_offsets): Update condition.
16578 (arm_emit_multi_reg_pop): Add a special case for load of a single
16579 register with writeback.
16581 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
16583 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
16586 2013-04-16 Richard Biener <rguenther@suse.de>
16588 PR tree-optimization/56756
16589 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
16590 (first_mem_ref_loc): New.
16591 (execute_sm): Place the load temporarily before a previous
16592 access instead of in the latch edge to ensure its SSA dependencies
16593 are defined at points dominating the load.
16595 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
16597 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
16598 correct fix by moving header and footer insn to the footer of
16599 the merged basic block. Clear BB_END of the merged-away block.
16601 PR middle-end/43631
16602 * emit-rtl.c (make_note_raw): New function.
16603 (link_insn_into_chain): New static inline function.
16604 (add_insn): Use it.
16605 (add_insn_before, add_insn_after): Factor insn chain linking code...
16606 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
16607 using link_insn_into_chain.
16608 (note_outside_basic_block_p): New helper function for emit_note_after
16609 and emit_note_before.
16610 (emit_note_after): Use nobb variant of add_insn_after if the note
16611 should not be contained in a basic block.
16612 (emit_note_before): Use nobb variant of add_insn_before if the note
16613 should not be contained in a basic block.
16614 (emit_note_copy): Use make_note_raw.
16615 (emit_note): Likewise.
16616 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
16617 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
16618 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
16619 the moved barrier the tail of the basic block it follows.
16620 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
16622 2013-04-15 Jakub Jelinek <jakub@redhat.com>
16624 PR tree-optimization/56962
16625 * gimple-ssa-strength-reduction.c (record_increment): Only set
16626 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
16627 either rhs1 or rhs2 is equal to c->base_expr.
16629 2013-04-15 Richard Biener <rguenther@suse.de>
16631 PR tree-optimization/56933
16632 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
16634 (GROUP_READ_WRITE_DEPENDENCE): Remove.
16635 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
16636 * tree-vect-data-refs.c (vect_analyze_group_access): Move
16637 dependence check ...
16638 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
16640 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
16641 GROUP_READ_WRITE_DEPENDENCE.
16643 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16645 * emit-rtl.c (reset_all_used_flags): New function.
16646 (verify_rtl_sharing): Call reset_all_used_flags before and after
16647 performing the checks.
16649 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16651 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
16652 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
16653 * config/arm/constraints.md (De): New constraint.
16654 * config/arm/neon.md (anddi3_neon): Delete.
16655 (neon_vand<mode>): Expand to standard anddi3 pattern.
16656 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
16657 Move earlier in the file.
16658 (neon_inv_logic_op2): Likewise.
16659 (arm_anddi_operand_neon): New predicate.
16661 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16663 * configure.ac (gcc_cv_ld_as_needed): Set
16664 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
16665 Use -z ignore, -z record on *-*-solaris2*.
16666 (HAVE_LD_AS_NEEDED): Update comment.
16667 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
16668 * configure: Regenerate.
16669 * config.in: Regenerate.
16670 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
16671 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
16672 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
16673 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
16674 equivalents. Fix markup.
16675 * doc/tm.texi: Regenerate.
16677 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
16679 * config/i386/i386.opt: New option mstack-protector-guard=.
16680 * config/i386/i386-opts.h: Add enum stack_protector_guard.
16681 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
16682 TARGET_SSP_TLS_GUARD.
16683 * config/i386/i386.c (ix86_option_override_internal): Set
16684 ix86_stack_protector_guard.
16685 * config/i386/i386.md (stack_protect_set): Enable for
16686 TARGET_SSP_TLS_GUARD only.
16687 (stack_protect_set_<mode>): Ditto.
16688 (stack_protect_test): Ditto.
16689 (stack_protect_test_<mode>): Ditto.
16690 * doc/invoke.texi (i386 Option): Document.
16692 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
16695 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
16696 (S_MODES): Set H_MODE bit.
16697 (SF_MODES): Set only S_MODE and SF_MODE bits.
16698 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
16699 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
16700 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
16701 <MODE_FLOAT>: Likewise.
16703 2013-04-15 Joey Ye <joey.ye@arm.com>
16705 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
16707 2013-04-15 Joey Ye <joey.ye@arm.com>
16709 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
16711 (thumb_far_jump_used_p): Count instruction size and set
16714 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
16716 * reorg.c (fill_simple_delay_slots): Reindent block of code.
16717 * resource.c (mark_target_live_regs): Reformat conditional block.
16719 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
16721 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
16722 notes, they are emitted only just before final.
16723 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
16725 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
16727 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
16728 * cfgrtl.c (delete_insn): Call it here instead.
16729 * lra-spills.c (lra_final_code_change): Use delete_insn.
16730 * haifa-sched.c (sched_remove_insn): Likewise.
16731 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
16732 returning to the nop pool.
16733 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
16734 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
16736 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
16738 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
16739 * doc/tm.texi: Regenerated.
16741 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
16743 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
16746 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
16748 * df-core.c (df_find_def): Compare register numbers.
16749 (df_find_use): Likewise.
16751 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
16754 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
16755 lra_in_progress for return.
16757 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
16759 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
16760 define_insn into define_insn_and_split and emit movsicc patterns.
16762 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
16764 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
16766 2013-04-12 Richard Biener <rguenther@suse.de>
16768 * tree-pass.h (TODO_do_not_ggc_collect): New.
16769 * passes.c (execute_one_ipa_transform_pass): Honor
16770 TODO_do_not_ggc_collect.
16771 (execute_one_pass): Likewise.
16774 2013-04-10 Richard Biener <rguenther@suse.de>
16776 * passes.c (init_optimization_passes): Remove reload pass.
16777 * ira.c (do_reload): Merge into ...
16779 (rest_of_handle_reload): Remove.
16780 (pass_reload): Likewise.
16781 * config/i386/i386.c (ix86_option_override): Refer to ira instead
16782 of reload for vzeroupper pass placement.
16784 2013-04-12 Jakub Jelinek <jakub@redhat.com>
16786 PR tree-optimization/56918
16787 PR tree-optimization/56920
16788 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
16789 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
16790 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
16791 use wide_mul_with_sign method.
16793 2013-04-12 Richard Biener <rguenther@suse.de>
16795 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
16796 not be considered a gimple constant.
16798 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
16800 * fold-const.c (const_binop): Handle vector shifts by a scalar.
16801 (fold_binary_loc): Call const_binop also for mixed vector-scalar
16804 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
16805 Jakub Jelinek <jakub@redhat.com>
16807 * opts.c: Include diagnostic-color.h.
16808 (common_handle_option): Handle OPT_fdiagnostics_color_.
16809 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
16810 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
16811 (diagnostic-color.o): New.
16812 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
16813 (diagnostic_color_rule): New enum.
16814 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
16815 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
16816 the location string.
16817 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
16818 either NULL, or color kind.
16819 * diagnostic-color.c: New file.
16820 * diagnostic-color.h: New file.
16821 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
16823 * doc/invoke.texi (-fdiagnostics-color): Document.
16824 * pretty-print.h (pp_show_color): Define.
16825 (struct pretty_print_info): Add show_color field.
16826 * diagnostic.c: Include diagnostic-color.h.
16827 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
16828 macros. Colorize error:, warning: etc. strings and also the location
16830 (diagnostic_show_locus): Colorize the caret line.
16831 * pretty-print.c: Include diagnostic-color.h.
16832 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
16833 inside of %< %> quotes or quoted through q format modifier.
16835 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16837 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
16839 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16841 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
16842 code in CC_NZ mode.
16843 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
16846 2013-04-11 Marek Polacek <polacek@redhat.com>
16848 PR tree-optimization/48184
16849 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
16851 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
16853 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
16854 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
16855 (skip_simple_arithmetic): Tidy up.
16856 * tree.h (skip_simple_constant_arithmetic): Declare.
16858 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16860 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
16862 2013-04-11 Richard Biener <rguenther@suse.de>
16864 * tree-vect-loop.c (get_initial_def_for_induction): Properly
16865 generate vector constants.
16867 2013-04-11 Richard Biener <rguenther@suse.de>
16869 PR tree-optimization/56878
16870 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
16871 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
16873 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
16874 Prefer to align the DR with the most invariant base address.
16876 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
16878 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
16881 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
16883 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
16884 floating-point vector comparisons against 0.
16886 2013-04-11 Jakub Jelinek <jakub@redhat.com>
16888 PR tree-optimization/56899
16889 * fold-const.c (extract_muldiv_1): Apply distributive law
16890 only if TYPE_OVERFLOW_WRAPS (ctype).
16892 2013-04-11 Bin Cheng <bin.cheng@arm.com>
16895 * ira-costs.c (scan_one_insn): Check whether the source rtx of
16896 loading has side effect.
16898 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
16900 * config/sparc/sparc.c: Include tree-pass.h.
16901 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
16902 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
16903 head of file. Change return type. Split off gate function.
16904 (sparc_gate_work_around_errata): New function.
16905 (pass_work_around_errata): New pass definition.
16906 (insert_pass_work_around_errata) New pass insert definition to
16907 insert pass_work_around_errata just after delayed-branch scheduling.
16908 (sparc_option_override): Insert the pass.
16909 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
16911 2013-04-10 David S. Miller <davem@davemloft.net>
16913 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
16914 or -mcpu=hypersparc.
16916 * target.def (cstore_mode): New hook.
16917 * target.h: Include insn-codes.h
16918 * targhooks.c: Likewise.
16919 (default_cstore_mode): New function.
16920 * targhooks.h: Declare it.
16921 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
16922 * doc/tm.texi: Rebuild.
16923 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
16924 target hook, rather than inspecting the insn_data.
16925 * config/sparc/sparc.c (sparc_cstore_mode): New function.
16926 (TARGET_CSTORE_MODE): Redefine.
16927 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
16929 * config/sparc/predicates.md (cstore_result_operand): New special
16931 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
16932 Use it for operand 0.
16933 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
16934 (*snesi_special): Likewise.
16935 (*snesi_zero): Likewise.
16936 (*seqsi_zero): Likewise.
16937 (*sltu_insn): Likewise.
16938 (*sgeu_insn): Likewise.
16939 (*seqdi_special): Make operand 0 and comparison operation be of
16941 (*snedi_special): Likewise.
16942 (*snedi_special_vis3): Likewise.
16943 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
16944 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
16945 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
16946 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
16947 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
16948 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
16949 (*sltu_extend_sp64): Likewise.
16950 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
16951 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
16952 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
16953 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
16954 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
16956 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
16958 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
16959 (aarch64_start_file): Use the new function.
16961 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
16962 Jason Merrill <jason@redhat.com>
16964 * common.opt: Add -gdwarf.
16965 * opts.c (common_handle_option): Handle it.
16966 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
16968 2013-04-10 Richard Biener <rguenther@suse.de>
16970 * passes.c (execute_todo): Do not call ggc_collect conditional here.
16971 (execute_one_ipa_transform_pass): But unconditionally here.
16972 (execute_one_pass): And here.
16973 (init_optimization_passes): Remove reload pass.
16974 * tree-pass.h (TODO_ggc_collect): Remove.
16975 (pass_reload): Likewise.
16976 * ira.c (do_reload): Merge into ...
16978 (rest_of_handle_reload): Remove.
16979 (pass_reload): Likewise.
16980 * config/i386/i386.c (ix86_option_override): Refer to ira instead
16981 of reload for vzeroupper pass placement.
16982 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
16983 and todo_flags_finish of all passes.
16985 2013-04-10 Richard Biener <rguenther@suse.de>
16987 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
16988 first_const_oprnd field, rename first_def_type to first_op_type.
16989 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
16990 (vect_get_and_check_slp_defs): Always use the type of the
16991 operand. Allow mixed vect_external_def, vect_constant_def types.
16992 (vect_get_constant_vectors): Handle mixed vect_external_def,
16993 vect_constant_def types.
16995 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
16997 PR tree-optimization/55524
16998 * tree-ssa-math-opts.c
16999 (convert_mult_to_fma): Don't use an fms construct
17000 when we don't have an fms operation, but fnma, and it looks
17001 likely that we'll be able to use the latter.
17003 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
17005 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
17007 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
17008 inline fail caused by overwritable functions.
17010 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
17012 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
17013 unnecessary bits in the constant power of two case.
17015 2013-04-10 Richard Biener <rguenther@suse.de>
17017 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
17018 broken code swapping operands.
17019 (vect_build_slp_tree): Do not compute load permutations here.
17020 (vect_analyze_slp_instance): Compute load permutations here,
17021 after building the SLP tree.
17023 2013-04-09 Christian Bruel <christian.bruel@st.com>
17025 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
17026 of next/prev_real_insn.
17028 2013-04-09 Jan Hubicka <jh@suse.cz>
17030 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
17031 Drop aliased parameter.
17032 (function_and_variable_visibility): Do not handle alias pairs.
17033 * cgraph.c (varpool_externally_visible_p): Update prototype.
17034 * varpool.c (varpool_add_new_variable): Update.
17036 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17038 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
17040 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
17042 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
17044 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
17046 2013-04-09 Marek Polacek <polacek@redhat.com>
17048 PR tree-optimization/48762
17049 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
17051 2013-04-09 Richard Biener <rguenther@suse.de>
17053 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
17055 (vect_build_slp_tree): Likewise.
17056 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
17057 calculating the cost of a SLP instance.
17058 (vect_analyze_slp_instance): Use it from here, after building
17061 2013-04-09 Jakub Jelinek <jakub@redhat.com>
17063 PR middle-end/56883
17064 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
17065 expand_omp_for_static_chunk): Use simple_p = true in
17066 force_gimple_operand_gsi calls when assigning to addressable decls.
17068 2013-04-09 Jeff Law <law@redhat.com>
17070 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
17071 when the boolean was created by converting a wider object which
17072 had a boolean range.
17074 2013-04-09 Richard Biener <rguenther@suse.de>
17076 * tree-vectorizer.h (slp_void_p): Remove.
17077 (slp_tree): Typedef before _slp_tree declaration.
17078 (struct _slp_tree): Use a vector of slp_tree as children.
17079 (vect_get_place_in_interleaving_chain): Remove.
17080 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
17082 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
17084 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
17085 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
17086 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
17087 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
17088 Use slp_node instead of slp_void_p and adjust.
17090 2013-04-09 Richard Biener <rguenther@suse.de>
17092 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
17093 work that is not necessary.
17095 2013-04-09 Jakub Jelinek <jakub@redhat.com>
17097 PR tree-optimization/56854
17098 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
17099 forward into clobber stmts if it would change MEM_REF lhs into
17102 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
17104 * tree.c (type_hash_lookup, type_hash_add): Make static.
17105 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
17107 2013-04-09 Richard Biener <rguenther@suse.de>
17109 * tree.h (unsave_expr_now): Remove.
17110 * tree-inline.c (mark_local_for_remap_r): Remove.
17111 (unsave_expr_1): Likewise.
17112 (unsave_r): Likewise.
17113 (unsave_expr_now): Likewise.
17114 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
17115 (propagate_tree_value): Likewise.
17117 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
17119 * doc/rtl.texi (sequence): Rewrite documentation to match the
17120 current use of SEQUENCE rtl objects.
17121 * rtl.def (SEQUENCE): Likewise.
17123 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
17124 Update documentation.
17125 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
17126 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
17128 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
17130 2013-04-08 Teresa Johnson <tejohnson@google.com>
17132 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
17133 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
17135 (estimate_edge_size_and_time): Add comment to suggest using rounding
17137 (estimate_node_size_and_time): Ditto.
17138 (remap_edge_change_prob): Use helper rounding divide methods.
17139 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
17140 (gimple_mod_pow2_value_transform): Ditto.
17141 (gimple_mod_subtract_transform): Ditto.
17142 (gimple_ic_transform): Ditto.
17143 (gimple_stringops_transform): Ditto.
17144 * stmt.c (conditional_probability): Ditto.
17145 (emit_case_dispatch_table): Ditto.
17146 * lto-cgraph.c (merge_profile_summaries): Ditto.
17147 * tree-optimize.c (execute_fixup_cfg): Ditto.
17148 * cfgcleanup.c (try_forward_edges): Ditto.
17149 * cfgloopmanip.c (scale_loop_profile): Ditto.
17151 (duplicate_loop_to_header_edge): Ditto.
17152 (lv_adjust_loop_entry_edge): Ditto.
17153 * tree-vect-loop.c (vect_transform_loop): Ditto.
17154 * profile.c (compute_branch_probabilities): Ditto.
17155 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
17156 * lto-streamer-in.c (input_cfg): Ditto.
17157 * gimple-streamer-in.c (input_bb): Ditto.
17158 * ipa-cp.c (update_profiling_info): Ditto.
17159 (update_specialized_profile): Ditto.
17160 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
17161 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
17163 * sched-rgn.c (compute_dom_prob_ps): Ditto.
17164 (compute_trg_info): Ditto.
17165 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
17166 (purge_dead_edges): Ditto.
17167 * loop-unswitch.c (unswitch_loop): Ditto.
17168 * cgraphclones.c (cgraph_clone_edge): Ditto.
17169 (cgraph_clone_node): Ditto.
17170 * tree-inline.c (copy_bb): Ditto.
17171 (copy_edges_for_bb): Ditto.
17172 (initialize_cfun): Ditto.
17173 (copy_cfg_body): Ditto.
17174 (expand_call_inline): Ditto.
17176 2013-04-08 Kai Tietz <ktietz@redhat.com>
17178 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
17179 TARGET_CYGWIN64 by TARGET_64BIT.
17181 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
17183 * config/epiphany/epiphany.md (GPR_1): New constant.
17184 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
17185 * config/epiphany/epiphany.c (gen_compare_reg):
17186 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
17187 is already in place.
17188 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
17189 Don't require being called during rtl expansion; If y operlaps r0,
17191 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
17192 (epiphany_expand_epilogue): Likewise.
17194 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
17195 Don't use CC_FPmode for ORDERED / UNORDERED.
17196 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
17198 * config/epiphany/constraints.md (CnL): New constraint.
17199 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
17200 * config/epiphany/predicates.md (add_operand): Allow 1024.
17202 * config/epiphany/epiphany.md (logical_op): New code iterator.
17203 (op_mnc): New code attribute.
17204 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
17205 (mov_f+1, mov_f+2): New peephole2 patterns.
17207 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
17208 (cstoresi4): Also allow re-use of zero result when doing a NE
17209 comparison to a non-zero operand.
17210 Use (clobber (scratch)) for first insn if the gpr output is not needed.
17212 * config/epiphany/epiphany.md (<insn_opname>v2si3):
17213 Use gen_addsi3_i / gen_subsi3_i.
17215 2013-04-08 Jakub Jelinek <jakub@redhat.com>
17219 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
17220 contain anything but clobbers, at most one __builtin_stack_restore,
17221 optionally debug stmts and final resx, and if it has at least one
17222 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
17223 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
17224 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
17225 which isn't defaut definition, remove them.
17226 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
17227 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
17228 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
17229 with MEM_REF LHS with SSA_NAME address.
17231 2013-04-08 Jeff Law <law@redhat.com>
17233 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
17235 2013-04-08 Richard Biener <rguenther@suse.de>
17237 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
17239 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
17240 determined vector type.
17241 (vect_analyze_data_refs): Likewise.
17242 (vect_get_new_vect_var): Adjust.
17243 (vect_create_destination_var): Preserve SSA name versions.
17244 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
17245 not dump anything here.
17247 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
17249 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
17250 Add member lr_slot_known.
17251 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
17253 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
17254 Remove code that sets lr_slot_offset according to what a previous
17255 version of epiphany_emit_save_restore used to do.
17256 (epiphany_emit_save_restore): When doing an lr save or restore,
17257 set/verify lr_slot_known and lr_slot_offset.
17259 2013-04-08 Xinyu Qi <xyqi@marvell.com>
17262 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
17265 2013-04-08 Richard Biener <rguenther@suse.de>
17267 * alias.c (find_base_term): Fix thinko in previous change.
17269 2013-04-08 Jakub Jelinek <jakub@redhat.com>
17271 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
17272 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
17273 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
17274 if possible to compute val.
17275 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
17276 For QImode integers don't require anything about precision. Use
17277 const_with_all_bytes_same to find out if the constant doesn't have
17278 repeated bytes in it.
17280 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17282 * config/s390/s390.c (s390_expand_insv): Only accept insertions
17285 2013-04-08 Marek Polacek <polacek@redhat.com>
17287 PR rtl-optimization/48182
17288 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
17291 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17294 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
17295 nuses, make sure we have a label.
17297 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17300 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
17301 (rs6000_emit_swdiv_low_precision): Remove.
17302 (rs6000_emit_swdiv): Rewrite to handle between one and four
17303 iterations of Newton-Raphson generally; modify required number of
17304 iterations for some cases.
17305 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
17307 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
17309 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
17310 set-but-unused variable.
17312 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
17313 basic blocks of released function bodies garbage-collectable.
17315 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
17316 (struct rtl_opt_pass): Add TODO_df_finish.
17318 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
17320 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17322 * config/arm/constraints.md (q): New constraint.
17323 * config/arm/ldrdstrd.md: New file.
17324 * config/arm/arm.md (ldrdstrd.md) New include.
17325 (arm_movdi): Use "q" instead of "r" constraint
17326 for double-word memory access.
17327 (movdf_soft_insn): Likewise.
17328 * config/arm/vfp.md (movdi_vfp): Likewise.
17329 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
17330 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
17331 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
17332 (mem_ok_for_ldrd_strd): Likewise.
17333 (output_move_double): Update assertion.
17335 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17337 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
17339 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17341 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
17342 define_insn_and_split.
17343 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
17345 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17347 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
17348 define_insn_and_split.
17349 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
17350 (shiftsi3_compare): New pattern.
17351 (rrx): New pattern.
17352 * config/arm/unspecs.md (UNSPEC_RRX): New.
17354 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17356 * config/arm/arm.md (negdi_extendsidi): New pattern.
17357 (negdi_zero_extendsidi): Likewise.
17359 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17361 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
17362 define_insn_and_split.
17363 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
17364 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
17366 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17368 * config/arm/arm.md (arm_subdi3): Convert define_insn into
17369 define_insn_and_split.
17370 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
17371 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
17373 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17375 * config/arm/arm.md (subsi3_carryin): New pattern.
17376 (subsi3_carryin_const): Likewise.
17377 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
17378 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
17380 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17382 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
17384 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17386 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
17387 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
17389 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17391 * config/arm/arm.c (arm_expand_builtin): Change fcode
17392 type to unsigned int.
17394 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17396 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
17398 2013-04-04 Ian Lance Taylor <iant@google.com>
17400 * doc/standards.texi (Standards): The Go frontend supports the Go 1
17403 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
17405 PR middle-end/56729
17406 * df-scan.c (df_insn_delete): Disable failing assert.
17408 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17410 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
17411 New function prototype.
17412 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
17413 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
17414 (arm_builtin_vectorized_function): New function.
17416 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17418 * config/arm/arm_neon_builtins.def: New file.
17419 * config/arm/arm.c (neon_builtin_data): Move contents to
17420 arm_neon_builtins.def.
17421 (enum arm_builtins): Include neon builtin definitions.
17422 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
17423 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
17425 2013-04-04 Marek Polacek <polacek@redhat.com>
17427 PR tree-optimization/48186
17428 * predict.c (maybe_hot_frequency_p): Return false if
17429 HOT_BB_FREQUENCY_FRACTION is 0.
17430 (cgraph_maybe_hot_edge_p): Likewise.
17432 2013-04-04 Richard Biener <rguenther@suse.de>
17434 PR tree-optimization/56826
17435 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
17438 2013-04-04 Richard Biener <rguenther@suse.de>
17440 PR tree-optimization/56213
17441 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
17442 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
17444 2013-04-04 Richard Biener <rguenther@suse.de>
17446 PR tree-optimization/56837
17447 * tree-loop-distribution.c (classify_partition): For non-zero
17448 values require that the value has the same precision as its
17449 mode to be useful as memset value.
17451 2013-04-03 Nick Clifton <nickc@redhat.com>
17453 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
17454 (fmssf4): Use fmsf.s on E3V5 architectures.
17455 (fnmasf4): Use fnmaf.s on E3V5 architectures.
17456 (fnmssf4): Use fnmsf.s on E3V5 architectures.
17458 2013-04-03 Jeff Law <law@redhat.com>
17460 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
17461 (lra-eliminations.o): Likewise.
17463 2013-04-03 Teresa Johnson <tejohnson@google.com>
17465 * gcov-io.c (compute_working_sets): Moved most of body of old
17466 compute_working_sets here from profile.c.
17467 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
17468 (gcov_working_set_t): Moved typedef here from basic-block.h
17469 (compute_working_set): Declare.
17470 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
17471 (get_working_sets): Renamed from compute_working_set,
17472 replace most of body with call to new compute_working_sets.
17473 (get_exec_counts): Replace call to compute_working_sets
17474 to get_working_sets.
17475 * profile.h (get_working_sets): Renamed from compute_working_set.
17476 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
17477 to get_working_sets.
17478 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
17479 * gcov-dump.c (dump_working_sets): New function.
17481 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
17483 * hwint.c (sext_hwi, zext_hwi): New functions.
17484 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
17485 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
17486 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
17487 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
17488 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
17489 (sext_hwi, zext_hwi): New functions.
17491 2013-04-03 Jeff Law <law@redhat.com>
17493 PR tree-optimization/56799
17494 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
17495 back test for widening conversion erroneously dropped in prior change.
17497 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17500 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
17501 instead of next_real_insn.
17503 2013-04-03 Marek Polacek <polacek@redhat.com>
17506 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
17508 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17511 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
17513 (thumb1_output_casesi): Likewise.
17514 (thumb2_output_casesi): Likewise.
17516 2013-04-03 Richard Biener <rguenther@suse.de>
17518 PR tree-optimization/56817
17519 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
17521 (tree_unroll_loops_completely_1): ... new function to manually
17522 walk the loop tree, properly defering outer loops of unrolled
17523 loops to later iterations.
17525 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
17527 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
17528 (vectorizable_load): Likewise.
17529 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
17530 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
17532 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
17534 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
17537 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
17539 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
17541 2013-04-03 Bin Cheng <bin.cheng@arm.com>
17543 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
17545 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
17547 PR tree-optimization/56790
17548 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
17551 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
17553 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
17555 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
17556 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
17559 2013-04-03 Jakub Jelinek <jakub@redhat.com>
17562 * tree.h (force_folding_builtin_constant_p): New decl.
17563 * builtins.c (force_folding_builtin_constant_p): New variable.
17564 (fold_builtin_constant_p): Fold immediately also if
17565 force_folding_builtin_constant_p.
17567 2013-04-03 Richard Biener <rguenther@suse.de>
17569 PR tree-optimization/56812
17570 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
17571 DRs of the same interleaving chain are independent.
17573 2013-04-02 Jason Merrill <jason@redhat.com>
17575 * gdbinit.in (pbb): Use debug fn.
17577 2013-04-02 Lawrence Crowl <crowl@google.com>
17579 * sese.h (struct ivtype_map_elt_s): Remove unused.
17580 (extern debug_ivtype_map): Remove unused.
17581 (extern eq_ivtype_map_elts): Remove unused.
17582 * sese.c (debug_ivtype_map): Removed unused.
17583 (debug_ivtype_map_1): Removed unused.
17584 (debug_ivtype_elt): Remove unused.
17585 (eq_ivtype_map_elts): Remove unused.
17588 2013-04-02 Kai Tietz <ktietz@redhat.com>
17591 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
17592 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
17593 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
17595 (legitimize_pe_coff_symbol): Likewise.
17596 (is_imported_p): New helper-function.
17597 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
17598 for Windows x64 targets.
17599 (ix86_expand_prologue): Optimize for pe-coff targets.
17600 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
17601 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
17602 medium/large code-model.
17603 (legitimize_pic_address): Likewise.
17604 (legitimize_tls_address): Likewise.
17605 (ix86_expand_call): Likewise.
17606 (x86_output_mi_thunk): Likewise.
17607 (get_dllimport_decl): Add new beimport argument.
17608 (construct_plt_address): Don't assert for x64 pe-coff targets.
17609 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
17611 (SYMBOL_FLAG_STUBVAR): New macro.
17612 (SYMBOL_REF_STUBVAR_P): Likewise.
17613 * config/i386/winnt.c (stub_list): New structure.
17614 (stub_head): New local variable.
17615 (i386_pe_record_stub): New function.
17616 (i386_pe_file_end): Emit refptr-stubs.
17618 2013-04-02 Jakub Jelinek <jakub@redhat.com>
17620 PR rtl-optimization/56745
17621 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
17622 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
17625 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
17626 and both of them are MEM_REFs, just compare first argument for
17627 equality and attempt to deal even with differing offsets.
17630 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
17631 of gimple_clobber_p to be MEM_REF.
17632 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
17633 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
17634 after gimplification.
17635 * asan.c (get_mem_ref_of_assignment): Don't instrument
17636 gimple_clobber_p stmts.
17637 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
17638 gimple_clobber_p stmt if they have MEM_REF lhs and
17639 are dead because of another gimple_clobber_p stmt.
17640 * tree-ssa-live.c (clear_unused_block_pointer): Treat
17641 gimple_clobber_p stmts like debug stmts.
17642 (remove_unused_locals): Remove clobbers with MEM_REF lhs
17643 that refer to unused VAR_DECLs or uninitialized values.
17644 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
17645 gimple_clobber_p stmts if they refer to removed parameters.
17646 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
17649 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
17651 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
17652 using SWI48 mode attribute.
17654 2013-04-02 Wei Mi <wmi@google.com>
17656 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
17657 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
17658 *<rotate_insn><mode>3_mask in i386.md.
17660 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
17662 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
17664 2013-04-02 Richard Biener <rguenther@suse.de>
17666 PR tree-optimization/56778
17667 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
17668 Runtime alias tests are not supported for gather loads.
17669 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
17670 stmts referenced from SSA operands before updating SSA form.
17672 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
17673 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17675 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
17676 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
17677 * config/arm/cortex-a53.md: New file.
17678 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
17679 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
17680 * config/arm/arm.c (arm_issue_rate): Likewise.
17681 * config/arm/arm-tune.md: Regenerate
17682 * config/arm/arm-tables.opt: Regenerate.
17683 * config/arm/arm-cores.def: Add cortex-a53.
17685 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
17687 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
17690 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
17692 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
17693 scalar load/store operations using B/H registers.
17694 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
17696 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
17698 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
17700 * config/aarch64/aarch64.c
17701 (aarch64_simd_scalar_immediate_valid_for_move): New.
17702 * config/aarch64/aarch64-protos.h
17703 (aarch64_simd_scalar_immediate_valid_for_move): New.
17704 * config/aarch64/constraints.md (Dh, Dq): New.
17705 * config/aarch64/iterators.md (hq): New.
17707 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
17709 * reorg.c (get_branch_condition): Deal with conditional returns.
17710 (fill_simple_delay_slots): Remove dead code dealing with jumps.
17712 2013-04-01 Wei Mi <wmi@google.com>
17714 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
17715 Truncate operand 2 using %b asm operand modifier.
17716 (*<shift_insn><mode>3_mask): Ditto.
17717 (*<rotate_insn><mode>3_mask): Ditto.
17719 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
17721 PR middle-end/56798
17722 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
17724 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
17726 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
17728 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
17730 2013-03-30 Lawrence Crowl <crowl@google.com>
17732 * dse.c (clear_alias_sets): Remove never set.
17733 (disqualified_clear_alias_sets): Remove never set.
17734 (clear_alias_mode_pool): Remove never set.
17735 (dse_step0): Remove condition that is never true.
17736 (canon_address): Remove condition that is never true.
17737 (dse_step7): Remove condition that is never true.
17738 (rest_of_handle_dse): Remove condition that is never true.
17739 (rest_of_handle_dse::did_global): Remove never read from above.
17740 (dse_step2_spill): Remove never called from above.
17741 (dse_step5_spill): Remove never called from above.
17743 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
17745 * doc/md.texi (Standard Names) <casesi>: Update documentation for
17746 JUMP_TABLE_DATA changes.
17747 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
17748 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
17749 (Insns) <jump_table_data>: New entry.
17750 * doc/tm.texi: Regenerate.
17752 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
17754 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
17755 for table jump at the end of a basic block using tablejump_p.
17756 * targhooks.c (default_invalid_within_doloop): Likewise.
17757 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
17758 target hook implementation that is identical to the default hook.
17759 (rs6000_invalid_within_doloop): Remove.
17761 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
17762 unused variable from tablejump_p call.
17764 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
17765 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
17766 (INSN_DELETED_P): Likewise.
17767 (emit_jump_table_data): New prototype.
17768 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
17769 after 4th as unused.
17770 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
17771 * sched-vis.c (print_insn): Likewise.
17772 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
17773 insn for compatibility with back ends that use next_active_insn to
17774 identify jump table data.
17775 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
17776 (remove_insn): Likewise.
17777 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
17779 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
17780 (emit_jump_table_data): New function.
17782 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
17783 basic block, a JUMP_TABLE_DATA never is.
17784 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
17785 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
17786 off from code handling real insns.
17787 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
17788 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
17789 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
17790 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
17791 is not a NONDEBUG_INSN_P.
17792 * ira-costs.c (scan_one_insn): Likewise.
17793 * jump.c (mark_all_labels): Likewise.
17794 (mark_jump_label_1): Likewise.
17795 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
17796 * lra.c (get_insn_freq): Expect all insns reaching here to be in
17798 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
17799 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
17800 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
17801 JUMP_TABLE_DATA_P insns.
17802 (calculate_elim_costs_all_insns): Likewise.
17803 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
17804 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
17805 (delete_output_reload): Code style fixups.
17806 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
17807 insn flags on this non-insn.
17808 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
17809 as scheduling barriers, for pre-change compatibility.
17810 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
17811 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
17813 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
17814 redundant JUMP_TABLE_DATA_P test.
17815 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
17816 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
17817 (frv_for_each_packet): Likewise.
17818 * config/i386/i386.c (min_insn_size): Likewise.
17819 (ix86_avoid_jump_mispredicts): Likewise.
17820 * config/m32r/m32r.c (m32r_is_insn): Likewise.
17821 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
17822 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
17823 (mips16_insn_length): Robustify.
17824 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
17825 (mips16_split_long_branches): Likewise.
17826 * config/pa/pa.c (pa_combine_instructions): Likewise.
17827 * config/rs6000/rs6000.c (get_next_active_insn): Treat
17828 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
17829 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
17830 as contributing to pool range lengths.
17831 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
17832 Remove redundant JUMP_TABLE_DATA_P test.
17833 (sh_loop_align): Likewise.
17834 (split_branches): Likewise.
17835 (sh_insn_length_adjustment): Likewise.
17836 * config/spu/spu.c (get_branch_target): Likewise.
17838 2013-03-29 Jan Hubicka <jh@suse.cz>
17840 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
17841 gcov streaming; stream hot bb threshold to ltrans.
17842 * predict.c (get_hot_bb_threshold): Break out from ....
17843 (maybe_hot_count_p): ... here.
17844 (set_hot_bb_threshold): New function.
17845 * lto-section-in.c (lto_section_name): Add profile.
17846 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
17847 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
17848 and data-streamer.h
17849 (histogram_entry): New structure.
17850 (histogram, histogram_pool): New global vars.
17851 (histogram_hash): New structure.
17852 (histogram_hash::hash): New method.
17853 (histogram_hash::equal): Likewise.
17854 (account_time_size): New function.
17855 (cmp_counts): New function.
17856 (dump_histogram): New function.
17857 (ipa_profile_generate_summary): New function.
17858 (ipa_profile_write_summary): New function.
17859 (ipa_profile_read_summary): New function.
17860 (ipa_profile): Decide on threshold.
17861 (pass_ipa_profile): Add ipa_profile_write_summary and
17862 ipa_profile_read_summary.
17863 * Makefile.in (ipa.o): Update dependencies.
17864 * lto-streamer.h (LTO_section_ipa_profile): New section.
17866 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
17868 * tree.h (VAR_P): New.
17870 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
17873 * doc/invoke.texi ([-fwhole-program]): Fix typo.
17875 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
17877 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
17878 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
17879 (control_flow_insn_p): Likewise.
17880 * cfgrtl.c (duplicate_insn_chain): Likewise.
17881 * final.c (get_attr_length_1): Likewise.
17882 (shorten_branches): Likewise.
17883 (final_scan_insn): Likewise.
17884 * function.c (instantiate_virtual_regs): Likewise.
17885 * gcse.c (insert_insn_end_basic_block): Likewise.
17886 * ira-costs.c (scan_one_insn): Likewise.
17887 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
17888 * lra.c (check_rtl): Likewise.
17889 * reload1.c (elimination_costs_in_insn): Likewise.
17890 * reorg.c (follow_jumps): Likewise.
17892 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
17893 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
17894 (thumb_far_jump_used_p): Likewise.
17895 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
17896 (workaround_speculation): Likewise.
17897 (add_sched_insns_for_speculation): Likewise.
17898 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
17899 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
17900 (frv_for_each_packet): Likewise.
17901 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
17902 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
17903 (final_emit_insn_group_barriers): Likewise.
17904 * config/m32r/m32r.c (m32r_is_insn): Likewise.
17905 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
17906 (mips16_insn_length): Likewise.
17907 * config/pa/pa.c (pa_reorg): Likewise.
17908 (pa_combine_instructions): Likewise.
17909 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
17910 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
17911 (sh_reorg): Likewise.
17912 (split_branches): Likewise.
17913 * config/spu/spu.c (get_branch_target): Likewise.
17915 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
17918 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
17920 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
17921 Fix declaration name.
17923 2013-03-28 Lawrence Crowl <crowl@google.com>
17925 * graphds.h (struct graph.indicies): Remove unused.
17926 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
17927 (SCOP_ORIGINAL_PDDRS): Remove unused.
17928 * sese.h (extern insert_loop_close_phis): Removed unused.
17929 (extern insert_guard_phis): Removed unused.
17930 (extern ivtype_map_elt_info): Removed unused.
17931 (new_ivtype_map_elt): Removed unused.
17932 * sese.c (ivtype_map_elt_info): Removed unused.
17934 2013-03-28 Lawrence Crowl <crowl@google.com>
17936 * Makefile.in: Add several missing include dependences.
17938 (test-dump.o): New. This object is not added to any executable,
17939 but is present for ad-hoc testing.
17941 (debug (const bitmap_head_def &)): New.
17942 (debug (const bitmap_head_def *)): New.
17944 (extern debug (const bitmap_head_def &)): New.
17945 (extern debug (const bitmap_head_def *)): New.
17947 (debug (edge_def &)): New.
17948 (debug (edge_def *)): New.
17950 (debug (basic_block_def &)): New.
17951 (debug (basic_block_def *)): New.
17953 (dump_node (const_tree, int, FILE *)): Correct source file.
17955 (debug (die_struct &)): New.
17956 (debug (die_struct *)): New.
17958 (extern debug (die_struct &)): New.
17959 (extern debug (die_struct *)): New.
17960 * gimple-pretty-print.c
17961 (debug (gimple_statement_d &)): New.
17962 (debug (gimple_statement_d *)): New.
17963 * gimple-pretty-print.h
17964 (extern debug (gimple_statement_d &)): New.
17965 (extern debug (gimple_statement_d *)): New.
17967 (debug (ira_allocno_copy &)): New.
17968 (debug (ira_allocno_copy *)): New.
17969 (debug (ira_allocno &)): New.
17970 (debug (ira_allocno *)): New.
17972 (extern debug (ira_allocno_copy &)): New.
17973 (extern debug (ira_allocno_copy *)): New.
17974 (extern debug (ira_allocno &)): New.
17975 (extern debug (ira_allocno *)): New.
17977 (debug (live_range &)): New.
17978 (debug (live_range *)): New.
17980 (debug (lra_live_range &)): New.
17981 (debug (lra_live_range *)): New.
17983 (debug (lra_live_range &)): New.
17984 (debug (lra_live_range *)): New.
17986 (debug (omega_pb_d &)): New.
17987 (debug (omega_pb_d *)): New.
17989 (extern debug (omega_pb_d &)): New.
17990 (extern debug (omega_pb_d *)): New.
17992 (debug (const rtx_def &)): New.
17993 (debug (const rtx_def *)): New.
17995 (debug_tree (tree): Move within file.
17996 (debug_raw (const tree_node &)): New.
17997 (debug_raw (const tree_node *)): New.
17998 (dump_tree_via_hooks (const tree_node *, int)): New.
17999 (debug (const tree_node &)): New.
18000 (debug (const tree_node *)): New.
18001 (debug_verbose (const tree_node &)): New.
18002 (debug_verbose (const tree_node *)): New.
18003 (debug_head (const tree_node &)): New.
18004 (debug_head (const tree_node *)): New.
18005 (debug_body (const tree_node &)): New.
18006 (debug_body (const tree_node *)): New.
18007 (debug_vec_tree (tree): Move and reimplement in terms of dump.
18008 (debug (vec<tree, va_gc> &)): New.
18009 (debug (vec<tree, va_gc> *)): New.
18011 (extern debug (const rtx_def &)): New.
18012 (extern debug (const rtx_def *)): New.
18014 (debug_raw (simple_bitmap_def &)): New.
18015 (debug_raw (simple_bitmap_def *)): New.
18016 (debug (simple_bitmap_def &)): New.
18017 (debug (simple_bitmap_def *)): New.
18019 (extern debug (simple_bitmap_def &)): New.
18020 (extern debug (simple_bitmap_def *)): New.
18021 (extern debug_raw (simple_bitmap_def &)): New.
18022 (extern debug_raw (simple_bitmap_def *)): New.
18024 (debug (vinsn_def &)): New.
18025 (debug (vinsn_def *)): New.
18026 (debug_verbose (vinsn_def &)): New.
18027 (debug_verbose (vinsn_def *)): New.
18028 (debug (expr_def &)): New.
18029 (debug (expr_def *)): New.
18030 (debug_verbose (expr_def &)): New.
18031 (debug_verbose (expr_def *)): New.
18032 (debug (vec<rtx> &)): New.
18033 (debug (vec<rtx> *)): New.
18035 (extern debug (vinsn_def &)): New.
18036 (extern debug (vinsn_def *)): New.
18037 (extern debug_verbose (vinsn_def &)): New.
18038 (extern debug_verbose (vinsn_def *)): New.
18039 (extern debug (expr_def &)): New.
18040 (extern debug (expr_def *)): New.
18041 (extern debug_verbose (expr_def &)): New.
18042 (extern debug_verbose (expr_def *)): New.
18043 (extern debug (vec<rtx> &)): New.
18044 (extern debug (vec<rtx> *)): New.
18046 (_list_iter_cond_expr): Make inline instead of static.
18048 (debug (sreal &)): New.
18049 (debug (sreal *)): New.
18051 (extern debug (sreal &)): New.
18052 (extern debug (sreal *)): New.
18054 (extern debug_raw (const tree_node &)): New.
18055 (extern debug_raw (const tree_node *)): New.
18056 (extern debug (const tree_node &)): New.
18057 (extern debug (const tree_node *)): New.
18058 (extern debug_verbose (const tree_node &)): New.
18059 (extern debug_verbose (const tree_node *)): New.
18060 (extern debug_head (const tree_node &)): New.
18061 (extern debug_head (const tree_node *)): New.
18062 (extern debug_body (const tree_node &)): New.
18063 (extern debug_body (const tree_node *)): New.
18064 (extern debug (vec<tree, va_gc> &)): New.
18065 (extern debug (vec<tree, va_gc> *)): New.
18067 (debug (struct loop &)): New.
18068 (debug (struct loop *)): New.
18069 (debug_verbose (struct loop &)): New.
18070 (debug_verbose (struct loop *)): New.
18071 * tree-dump.c: Add header dependence.
18073 (extern debug (struct loop &)): New.
18074 (extern debug (struct loop *)): New.
18075 (extern debug_verbose (struct loop &)): New.
18076 (extern debug_verbose (struct loop *)): New.
18078 (debug (data_reference &)): New.
18079 (debug (data_reference *)): New.
18080 (debug (vec<data_reference_p> &)): New.
18081 (debug (vec<data_reference_p> *)): New.
18082 (debug (vec<ddr_p> &)): New.
18083 (debug (vec<ddr_p> *)): New.
18085 (extern debug (data_reference &)): New.
18086 (extern debug (data_reference *)): New.
18087 (extern debug (vec<data_reference_p> &)): New.
18088 (extern debug (vec<data_reference_p> *)): New.
18089 (extern debug (vec<ddr_p> &)): New.
18090 (extern debug (vec<ddr_p> *)): New.
18092 (debug (pt_solution &)): New.
18093 (debug (pt_solution *)): New.
18095 (extern debug (pt_solution &)): New.
18096 (extern debug (pt_solution *)): New.
18098 (debug (_var_map &)): New.
18099 (debug (_var_map *)): New.
18100 (debug (tree_live_info_d &)): New.
18101 (debug (tree_live_info_d *)): New.
18103 (extern debug (_var_map &)): New.
18104 (extern debug (_var_map *)): New.
18105 (extern debug (tree_live_info_d &)): New.
18106 (extern debug (tree_live_info_d *)): New.
18108 2013-03-28 Jan Hubicka <jh@suse.cz>
18110 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
18112 2013-03-28 Ian Bolton <ian.bolton@arm.com>
18114 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
18115 record only when desired or required.
18117 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
18119 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
18120 *vec_extractv2di_1_rex64. Use x64 isa attribute.
18122 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18124 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
18125 (*andsi3_compare0_uxtw): New pattern.
18126 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
18127 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
18129 2013-03-28 Jan Hubicka <jh@suse.cz>
18131 * data-streamer-in.c (streamer_read_gcov_count): New function.
18132 * gimple-streamer-out.c: Include value-prof.h.
18133 (output_gimple_stmt): Output histogram.
18134 (output_bb): Use streamer_write_gcov_count.
18135 * value-prof.c: Include data-streamer.h
18136 (dump_histogram_value): Add HIST_TYPE_MAX.
18137 (stream_out_histogram_value): New function.
18138 (stream_in_histogram_value): New function.
18139 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
18140 (stream_out_histogram_value, stream_in_histogram_value): Declare.
18141 * data-streamer-out.c (streamer_write_gcov_count): New function.
18142 (streamer_write_gcov_count_stream): New function.
18143 * lto-cgraph.c (lto_output_edge): Update counter streaming.
18144 (lto_output_node): Likewise.
18145 (input_node, input_edge): Likewise.
18146 * lto-streamer-out.c (output_cfg): Update streaming.
18147 * lto-streamer-in.c (input_cfg): Likewise.
18148 * data-streamer.h (streamer_write_gcov_count,
18149 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
18150 * gimple-streamer-in.c: Include value-prof.h
18151 (input_gimple_stmt): Input histograms.
18152 (input_bb): Update profile streaming.
18154 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
18156 * genmodes.c (emit_max_int): New function.
18157 (emit_insn_modes_h): Added call to emit_max_function.
18158 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
18160 * machmode.def: Fixed comment.
18162 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
18164 * combine.c (try_combine): Removed useless assert.
18165 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
18167 2013-03-28 Marek Polacek <polacek@redhat.com>
18168 Richard Biener <rguenther@suse.de>
18170 PR tree-optimization/56695
18171 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
18172 build signed result of a vector comparison.
18173 * tree-cfg.c (verify_gimple_comparison): Check that a result
18174 of a vector comparison has signed type.
18176 2013-03-28 Richard Biener <rguenther@suse.de>
18178 PR tree-optimization/37021
18179 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
18180 do not restrict gaps between groups.
18181 * tree-vect-stmts.c (vectorizable_load): Properly account for
18182 a gap between groups.
18184 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
18186 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
18187 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
18190 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
18192 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
18193 * doc/extend.texi (Named Address Spaces): Ditto.
18194 (Variable Attributes): Ditto.
18196 2013-03-27 Kai Tietz <ktietz@redhat.com>
18198 * config.build: Add support for cygwin x64 target.
18199 * config.gcc: Likewise.
18200 * config.host: Likewise.
18201 * configure.ac: Likewise
18202 * configure: Regenerated.
18204 2013-03-27 Kai Tietz <ktietz@redhat.com>
18206 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
18207 * config/i386/t-cygwin-w64: New file.
18208 * config/i386/cygwin-w64.h: New file.
18209 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
18210 and add support for x64-cygwin target.
18211 (CPP_SPEC): Likewise.
18212 (CXX_WRAP_SPEC_LIST): Undefine before define.
18213 (LIBGCJ_SONAME): Use 15 as version.
18215 2013-03-27 Richard Biener <rguenther@suse.de>
18217 PR tree-optimization/56716
18218 * tree-ssa-structalias.c (perform_var_substitution): Adjust
18219 dumping for ref nodes.
18221 2013-03-27 Martin Jambor <mjambor@suse.cz>
18223 PR tree-optimization/55334
18224 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
18225 restricted pointers to arrays.
18227 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
18229 * Makefile.in (.SUFFIXES): Add .cc.
18230 (.c.o): Apply same recipe for implicit rule .cc.o.
18232 2013-03-27 Richard Biener <rguenther@suse.de>
18234 PR tree-optimization/37021
18235 * tree-vect-data-refs.c (vect_check_strided_load): Allow
18236 REALPART/IMAGPART_EXPRs around the supported refs.
18237 * tree-ssa-structalias.c (find_func_aliases): Assume that
18238 floating-point values are not used to transfer pointers.
18240 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
18242 * target.def (TARGET_HAS_IFUNC_P): New target hook.
18243 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
18244 * doc/tm.texi: Regenerate.
18245 * targhooks.h (default_has_ifunc_p): New.
18246 * targhooks.c (default_has_ifunc_p): Ditto.
18247 * config/linux-protos.h: New file.
18248 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
18249 hook for linux which disables support of indirect functions in android.
18250 * config/linux-android.c: New file.
18251 * config/t-linux-android.c: Ditto.
18252 * config.gcc: Added new object file linux-android.o.
18253 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
18254 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
18255 * varasm.c (do_assemble_alias): Likewise.
18256 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
18257 doesn't support indirect functions.
18258 * configure: Regenerate.
18260 2013-03-27 Bin Cheng <bin.cheng@arm.com>
18263 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
18264 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
18267 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18269 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
18271 2013-03-27 Terry Guo <terry.guo@arm.com>
18273 * config/arm/arm-cores.def: Added core cortex-r7.
18274 * config/arm/arm-tune.md: Regenerated.
18275 * config/arm/arm-tables.opt: Regenerated.
18276 * doc/invoke.texi: Added entry for core cortex-r7.
18278 2013-03-27 Walter Lee <walt@tilera.com>
18280 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
18281 double-decrement of next_scratch_regno.
18283 2013-03-27 Walter Lee <walt@tilera.com>
18285 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
18287 (insn_v1mulus): Ditto.
18288 (insn_v2muls): Ditto.
18290 2013-03-27 Walter Lee <walt@tilera.com>
18292 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
18293 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
18295 2013-03-27 Walter Lee <walt@tilera.com>
18297 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
18298 (*sibcall_value): Ditto.
18300 2013-03-27 Walter Lee <walt@tilera.com>
18302 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
18303 (insn_mnz_v8qi): ... this ...
18304 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
18306 (insn_v<n>mnz): Replaced by ...
18307 (insn_v1mnz): ... this ...
18308 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
18310 (insn_mz_<mode>): Replaced by ...
18311 (insn_mz_v8qi): ... this ...
18312 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
18314 (insn_v<n>mz): Replaced by ...
18315 (insn_v1mz): ... this ...
18316 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
18319 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
18321 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
18323 2013-03-26 Roland McGrath <mcgrathr@google.com>
18325 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
18326 than fprintf with a non-constant, non-format string.
18328 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
18330 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
18331 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
18332 operand 0 predicate.
18333 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
18334 attribute. Use general_x64nomem_operand as operand 1 predicate.
18335 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
18336 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
18337 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
18338 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
18339 (mov<mode>_insv_1): Remove expander. Merge insn with
18340 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
18341 Use general_x64nomem_operand as operand 1 predicate.
18342 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
18343 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
18344 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
18345 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
18346 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
18347 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
18348 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
18349 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
18350 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
18351 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
18352 (general_x64nomem_operand): Ditto.
18354 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
18356 * config/rtems.opt: Add -pthread option.
18358 2013-03-26 Richard Biener <rguenther@suse.de>
18360 * alias.c (find_base_term): Avoid redundant and not used recursion.
18361 (base_alias_check): Get the initial base term from the caller.
18362 (true_dependence_1): Compute and pass base terms to base_alias_check.
18363 (write_dependence_p): Likewise.
18364 (may_alias_p): Likewise.
18366 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
18368 * config/aarch64/aarch64.c (aarch64_classify_address): Support
18369 PC-relative load in SI modes and above only.
18371 2013-03-26 Xinyu Qi <xyqi@marvell.com>
18373 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
18374 * config/arm/iwmmxt.md (WCGR0): Update.
18375 (WCGR1, WCGR2, WCGR3): Likewise.
18377 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
18379 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
18380 Use x64 and nox64 isa attributes.
18382 2013-03-26 Richard Biener <rguenther@suse.de>
18384 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
18385 alignment computations and rely on get_object_alignment_1
18386 for the !TYPE_P case.
18387 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
18389 2013-03-26 Walter Lee <walt@tilera.com>
18391 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
18392 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
18394 2013-03-25 Jeff Law <law@redhat.com>
18396 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
18397 check for INTEGRAL_TYPE_P that was missing due to checking in
18398 wrong version of prior patch.
18400 2013-03-25 Walter Lee <walt@tilera.com>
18402 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
18403 TILEGX_INSN_SHUFFLEBYTES1.
18404 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
18406 (tilegx_builtins): Ditto.
18407 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
18409 2013-03-25 Walter Lee <walt@tilera.com>
18411 * config/tilegx/tilegx.md (floatsisf2): New pattern.
18412 (floatunssisf2): New pattern.
18413 (floatsidf2): New pattern.
18414 (floatunssidf2): New pattern.
18416 2013-03-25 Walter Lee <walt@tilera.com>
18418 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
18419 tests for constraint J, K, N, P.
18421 2013-03-25 Walter Lee <walt@tilera.com>
18423 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
18424 Use indirect/pcrel encoding.
18425 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
18428 2013-03-25 Steve Ellcey <sellcey@mips.com>
18430 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
18431 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
18432 * config/mips/mips.c (mips_option_override): Set IMADD default.
18433 * config/mips/mips.h (PTF_AVOID_IMADD): New.
18434 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
18435 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
18436 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
18437 * doc/invoke.texi (-mimadd/-mno-imadd): New.
18439 2013-03-25 Jeff Law <law@redhat.com>
18441 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
18442 slightly to avoid creating and folding useless trees. Simplify
18443 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
18445 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
18447 * config/i386/i386.md (*zero_extendsidi2): Merge with
18448 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
18449 * config/i386/predicates.md (x86_64_zext_operand): Rename from
18450 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
18451 targets. Clarify comment.
18453 2013-03-25 Martin Jambor <mjambor@suse.cz>
18455 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
18456 pass-through jump functions differently.
18457 (ipa_read_jump_function): Likewise. Also use setter functions to set
18460 2013-03-25 Martin Jambor <mjambor@suse.cz>
18462 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
18463 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
18465 (ipa_get_indirect_edge_target): New function.
18466 (devirtualization_time_bonus): New parameter known_aggs, pass it to
18467 ipa_get_indirect_edge_target. Update all callers.
18468 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
18469 ipa_get_indirect_edge_target_1 instead of calling
18470 ipa_get_indirect_edge_target.
18471 (create_specialized_node): Pass aggvlas to
18472 ipcp_discover_new_direct_edges.
18474 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18476 * config/arm/arm.md (f_sels, f_seld): New types.
18477 (*cmov<mode>): New pattern.
18478 * config/arm/predicates.md (arm_vsel_comparison_operator): New
18481 2013-03-25 Kai Tietz <ktietz@redhat.com>
18483 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
18484 POSIX-printf for mingw-hosted builds.
18486 2013-03-25 Richard Biener <rguenther@suse.de>
18488 PR middle-end/56694
18489 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
18490 must-not-throw stmt location.
18492 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18494 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
18495 Emit load-acquire versions when acq is true.
18496 (arm_emit_store_exclusive): Add rel parameter.
18497 Emit store-release versions when rel is true.
18498 (arm_split_compare_and_swap): Use acquire-release instructions
18500 of barriers when appropriate.
18501 (arm_split_atomic_op): Likewise.
18502 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
18503 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
18504 (VUNSPEC_SLX): Likewise.
18505 (VUNSPEC_LDA): Likewise.
18506 (VUNSPEC_STL): Likewise.
18507 * config/arm/sync.md (atomic_load<mode>): New pattern.
18508 (atomic_store<mode>): Likewise.
18509 (arm_load_acquire_exclusive<mode>): Likewise.
18510 (arm_load_acquire_exclusivesi): Likewise.
18511 (arm_load_acquire_exclusivedi): Likewise.
18512 (arm_store_release_exclusive<mode>): Likewise.
18514 2013-03-25 Catherine Moore <clm@codesourcery.com>
18516 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
18517 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
18518 * config/mip/predicates.md (lwsp_swsp_operand,
18519 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
18520 sb16_operand, db4_operand, db7_operand, ib3_operand,
18521 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
18522 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
18523 andi16_operand): New predicates.
18524 * config/mips/mips.md (compression): New attribute.
18525 (enabled): New attribute.
18526 (length): Consider compression in computing length.
18527 (shift_compression): New code attribute.
18528 (*add<mode>3): New operands. Record compression.
18529 (sub<mode>3): Likewise.
18530 (one_cmpl<mode>2): Likewise.
18531 (*and<mode>3): Likewise.
18532 (*ior<mode>3): Likewise.
18533 (unnamed pattern for xor): Likewise.
18534 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
18535 (*<optab><mode>3): Likewise.
18536 (*mov<mode>_internal: Likewise.
18537 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
18538 (mips_unsigned_immediate_p): New.
18539 (umips_lwsp_swsp_address_p): New.
18540 (m16_based_address_p): New.
18541 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
18542 (mips_unsigned_immediate_p): New prototype.
18543 (lwsp_swsp_address_p): New prototype.
18544 (m16_based_address_p): New prototype.
18545 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
18546 (mips_signed_immediate_p): New function.
18547 (m16_based_address_p): New function.
18548 (lwsp_swsp_address_p): New function.
18549 (mips_print_operand_punctuation): Recognize short delay slot insns
18550 for microMIPS.add<mode>3"
18552 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18555 * config/arm/iterators.md (v_cmp_result): New mode attribute.
18556 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
18558 2013-03-25 Richard Biener <rguenther@suse.de>
18560 PR tree-optimization/56689
18561 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
18564 2013-03-25 Richard Biener <rguenther@suse.de>
18566 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
18568 (memory_references): Likewise.
18569 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
18570 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
18571 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
18572 (gather_mem_refs_in_loops): Fold into ...
18573 (analyze_memory_references): ... this. Move initialization
18574 to tree_ssa_lim_initialize.
18575 (fill_always_executed_in): Rename to ...
18576 (fill_always_executed_in_1): ... this.
18577 (fill_always_executed_in): Move contains_call computation to
18578 this new function from ...
18579 (tree_ssa_lim_initialize): ... here.
18580 (tree_ssa_lim): Call fill_always_executed_in.
18582 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
18584 * postreload.c (reload_combine): Fix code detecting returns.
18586 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
18588 * function.c (emit_use_return_register_into_block): On cc0 targets,
18589 do not emit the sequence between cc0 setter and user.
18591 2013-03-25 Kai Tietz <ktietz@redhat.com>
18593 * config/i386/predicates.md (local_symbolic_operand): Interpret
18594 dll-imported symbols as none-local.
18596 2013-03-25 Richard Biener <rguenther@suse.de>
18598 * tree-ssa-loop-im.c (struct depend): Remove.
18599 (struct lim_aux_data): Make depends a vec of gimples.
18600 (free_lim_aux_data): Adjust.
18601 (add_dependency): Likewise.
18602 (set_level): Likewise.
18604 2013-03-25 Richard Biener <rguenther@suse.de>
18606 PR middle-end/56434
18607 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
18608 the pointer returned by calls with ECF_MALLOC set.
18610 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
18612 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
18614 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
18616 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
18617 using MMXMODE mode iterator.
18618 (*move<mode>_internal): Merge with *movv2sf_internal and
18619 *movv2sf_internal_rex64 using MMXMODE mode iterator.
18621 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
18623 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
18624 (record_last_mem_set_info): Likewise.
18626 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
18627 of XNEWVEC followed by memset.
18628 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
18630 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
18632 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
18633 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
18634 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
18635 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
18636 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
18637 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
18638 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
18639 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
18640 BARRIER_P instead of GET_CODE.
18642 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
18644 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
18645 inaccuracy in the probing code.
18647 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
18648 (ctrapdi4): Likewise.
18650 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
18652 * calls.c (expand_call): Add missing guard to code handling return
18653 of non-BLKmode structures in MSB.
18654 * function.c (expand_function_end): Likewise.
18656 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
18658 * combine.c (try_combine): Adjust comment. Do not add the set of
18659 insn #0 if the destination indirectly is set or dies in insn #2.
18660 Tidy up code to distribute a new note.
18662 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
18664 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
18665 also for alternatives 16 and 17.
18667 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
18669 * config/i386/sse.md (*mov<mode>_internal): Merge with
18670 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
18671 Emit insn template depending on type attribute. Use
18672 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
18673 movd instead of movq mnemonic for interunit moves. Rewrite mode
18674 attribute calculation. Remove unit attribute calculation.
18675 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
18676 Set prefix_data16 attribute for DImode ssemov types.
18677 Use Ym instead of y for SSE-MMX conversion alternatives.
18678 Reorder operand constraints.
18680 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
18682 * df.h (df_insn_delete): Adjust prototype.
18683 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
18684 and let it decide whether mark the basic block dirty.
18685 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
18686 * df-scan.c (df_insn_info_delete): New helper function, split
18687 off from df_insn_delete.
18688 (df_scan_free_bb_info): Use it.
18689 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
18691 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
18692 that the insn is actually an insn and it has a non-NULL basic block.
18693 Do not mark basic block dirty if only deleting a DEBUG_INSN.
18695 2013-03-22 Richard Biener <rguenther@suse.de>
18697 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
18699 (mem_ref_alloc): Do not allocate them.
18700 (refs_independent_p): Do not query or maintain a cache.
18702 2013-03-22 Richard Biener <rguenther@suse.de>
18704 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
18705 (gather_mem_refs_in_loops): Do not compute it.
18706 (analyze_memory_references): Do not allocate it.
18707 (tree_ssa_lim_finalize): Do not free it.
18708 (for_all_locs_in_loop): Do not query all_refs_in_loop.
18710 2013-03-22 Richard Biener <rguenther@suse.de>
18712 * is-a.h (as_a): Use gcc_checking_assert.
18714 2013-03-22 Ian Bolton <ian.bolton@arm.com>
18716 * config/aarch64/aarch64.c (aarch64_print_operand): New
18717 format specifier for printing a constant in hex.
18718 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
18719 format specifier for printing second operand.
18721 2013-03-22 Richard Biener <rguenther@suse.de>
18723 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
18725 (gather_mem_refs_in_loops): Perform store accumulation here.
18726 (create_vop_ref_mapping_loop): Remove.
18727 (create_vop_ref_mapping): Likewise.
18728 (analyze_memory_references): Initialize refs_stored_in_loop.
18729 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
18730 (record_indep_loop): Remove.
18731 (record_dep_loop): New function.
18732 (ref_indep_loop_p_1): Adjust to only walk over references
18733 in the loop, not its subloops.
18734 (ref_indep_loop_p): Rename to ...
18735 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
18736 maintaining a more fine-grained cache.
18737 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
18738 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
18740 2013-03-22 Richard Biener <rguenther@suse.de>
18742 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
18743 (struct mem_ref): Make accesses_in_loop a vec of a vec of
18744 aggregate mem_ref_loc.
18745 (free_mem_ref_locs): Inline into ...
18746 (memref_free): ... this and adjust.
18747 (mem_ref_alloc): Adjust.
18748 (mem_ref_locs_alloc): Remove.
18749 (record_mem_ref_loc): Adjust.
18750 (get_all_locs_in_loop): Rewrite into ...
18751 (for_all_locs_in_loop): ... this iterator.
18752 (rewrite_mem_ref_loc): New functor.
18753 (rewrite_mem_refs): Use for_all_locs_in_loop.
18754 (sm_set_flag_if_changed): New functor.
18755 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
18756 (ref_always_accessed): New functor.
18757 (ref_always_accessed_p): Use for_all_locs_in_loop.
18759 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
18761 * tree-pass.h (PROP_gimple_lvec): New.
18762 * passes.c (dump_properties): Handle PROP_gimple_lvec.
18763 (init_optimization_passes): Move pass_lower_vector.
18764 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
18766 (pass_lower_vector): Provide PROP_gimple_lvec.
18767 (pass_lower_vector_ssa): Likewise.
18768 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
18770 2013-03-21 Mark Wielaard <mjw@redhat.com>
18772 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
18774 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
18776 * config/i386/i386.md (*movdi_internal): Disparage slightly
18777 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
18778 conversion alternatives.
18780 2013-03-21 Jakub Jelinek <jakub@redhat.com>
18782 PR middle-end/48087
18783 * diagnostic.def (DK_WERROR): New kind.
18784 * diagnostic.h (werrorcount): Define.
18785 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
18786 promoted to DK_ERROR, increment DK_WERROR counter instead of
18788 * toplev.c (toplev_main): Call print_ignored_options even if
18789 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
18790 even if just werrorcount is non-zero.
18793 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
18795 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
18796 (string_cst_pool_decl): New function.
18797 (optimize_one_addr_into_implicit_ptr): New function.
18798 (resolve_addr_in_expr): Optimize DWARF location expression
18799 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
18800 which doesn't live in memory, but has DW_AT_location or
18801 DW_AT_const_value, or refers to a string literal, into
18802 DW_OP_GNU_implicit_pointer.
18803 (optimize_location_into_implicit_ptr): New function.
18804 (resolve_addr): If removing DW_AT_location of a variable because
18805 it was DW_OP_addr of address of the variable, but the variable doesn't
18806 live in memory, try to emit const value attribute for the initializer.
18808 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
18810 * tree.h (VECTOR_TYPE_P): New macro.
18811 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
18812 TYPE_MODE): Use it.
18813 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
18814 VEC_COND_EXPR cannot be lvalues.
18815 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
18817 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
18819 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
18820 Restrict the transformation to equal modes.
18822 2013-03-21 Richard Biener <rguenther@suse.de>
18824 PR tree-optimization/39326
18825 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
18826 (MEM_ANALYZABLE): Adjust.
18827 (record_mem_ref_loc): Move bitmap ops ...
18828 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
18829 unanalyzable refs, do not record locations for it.
18830 (analyze_memory_references): Allocate ref zero as shared
18832 (refs_independent_p): Do not test for unanalyzed mems here.
18833 (ref_indep_loop_p_1): Special-case disambiguation against
18834 the unanalyzed ref.
18835 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
18837 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
18839 * config/arm/arm-protos.h (tune_params): Add
18840 prefer_neon_for_64bits field.
18841 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
18842 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
18843 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
18844 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
18845 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
18846 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
18847 (arm_option_override): Handle -mneon-for-64bits new option.
18848 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
18849 (prefer_neon_for_64bits): Declare new variable.
18850 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
18851 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
18852 (arch_enabled): Handle new arch types. Remove support for onlya8
18854 (one_cmpldi2): Use new arch names.
18855 (zero_extend<mode>di2, extend<mode>di2): Ditto.
18856 * config/arm/arm.opt (mneon-for-64bits): Add option.
18857 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
18858 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
18859 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
18861 * doc/invoke.texi (-mneon-for-64bits): Document.
18863 2013-03-21 Richard Biener <rguenther@suse.de>
18865 PR tree-optimization/39326
18866 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
18867 (sort_bbs_in_loop_postorder_cmp): New function.
18868 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
18870 2013-03-21 Richard Biener <rguenther@suse.de>
18872 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
18873 (vect_insert_into_interleaving_chain): Likewise.
18874 (vect_drs_dependent_in_basic_block): Inline ...
18875 (vect_slp_analyze_data_ref_dependence): ... here. New function,
18877 (vect_analyze_data_ref_dependence): ... here. Simplify.
18878 (vect_check_interleaving): Simplify.
18879 (vect_analyze_data_ref_dependences): Likewise. Split out ...
18880 (vect_slp_analyze_data_ref_dependences): ... this new function.
18881 (dr_group_sort_cmp): New function.
18882 (vect_analyze_data_ref_accesses): Compute data-reference groups
18883 here instead of in vect_analyze_data_ref_dependence. Use
18884 a more efficient algorithm.
18885 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
18886 vect_slp_analyze_data_ref_dependences. Call
18887 vect_analyze_data_ref_accesses earlier.
18888 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
18889 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
18890 (vect_slp_analyze_data_ref_dependences): New prototype.
18892 2013-03-21 Richard Biener <rguenther@suse.de>
18894 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
18895 ref is stored in the loop.
18896 (find_refs_for_sm): Walk only over all stores.
18897 (store_motion_loop): Allocate from lim_bitmap_obstack.
18898 (store_motion): Likewise.
18900 2013-03-21 Richard Biener <rguenther@suse.de>
18902 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
18903 Update virtual SSA form.
18905 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
18907 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
18908 * configure: Regenerate.
18909 * config.in: Regenerate.
18910 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
18911 if !HAVE_LD_EH_FRAME_CIEV3.
18913 2013-03-21 Richard Biener <rguenther@suse.de>
18915 * tree-cfg.c (verify_expr_no_block): New function.
18916 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
18917 nor DECL_VALUE_EXPR have locations with associated blocks.
18918 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
18919 (clear_unused_block_pointer): Remove code dealing with
18920 blocks in DECL_DEBUG_EXPR locations.
18922 2013-03-21 Richard Biener <rguenther@suse.de>
18924 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
18925 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
18926 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
18927 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
18928 instead of DECL_DEBUG_EXPR_IS_FROM.
18929 * gimplify.c (gimplify_modify_expr): Likewise.
18930 * tree-cfg.c (verify_expr_location_1): Likewise.
18931 * tree-complex.c (create_one_component_var): Likewise.
18932 * tree-sra.c (create_access_replacement): Likewise.
18933 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
18934 (clear_unused_block_pointer): Likewise.
18935 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
18936 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
18937 * var-tracking.c (var_debug_decl): Likewise.
18938 (track_expr_p): Likewise.
18939 * tree-inline.c (add_local_variables): Likewise. Set
18940 DECL_HAS_DEBUG_EXPR_P after copying it.
18941 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
18942 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
18944 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
18947 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
18948 * configure: Regenerate.
18949 * config.in: Regenerate.
18950 * config/i386/i386.md (*movdf_internal): Use
18951 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
18952 movd instead of movq mnemonic for interunit moves.
18953 (*movdi_internal): Ditto.
18955 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18957 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
18958 (abd<mode>_3): New pattern.
18959 (aba<mode>_3): New pattern.
18960 (fabd<mode>_3): New pattern.
18962 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18964 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
18965 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
18966 occurrence of REGISTER_PREFIX as its empty string.
18968 2013-03-20 Jeff Law <law@redhat.com>
18970 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
18971 addititional equivalences for equality comparisons between an SSA_NAME
18972 and a constant where the SSA_NAME was set from a widening conversion.
18974 2013-03-20 Walter Lee <walt@tilera.com>
18976 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
18978 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
18980 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
18981 depending on type attribute.
18982 (*movti_internal): Ditto.
18983 (*movtf_internal): Ditto.
18984 (*movxf_internal): Ditto.
18985 (*movdf_internal): Ditto.
18986 (*movsf_internal): Ditto.
18988 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
18990 * config/i386/i386.md (*movti_internal): Set prefix attribute to
18991 maybe_vex for sselog1 and ssemov types.
18992 (*movdi_internal): Reorder operand constraints.
18993 (*movsi_internal): Ditto. Set prefix attribute to
18994 maybe_vex for sselog1 and ssemov types.
18995 (*movtf_internal): Set prefix attribute to maybe_vex
18996 for sselog1 and ssemov types.
18997 (*movdf_internal): Ditto. Set prefix_data16 attribute for
18998 DImode ssemov types. Reorder operand constraints.
18999 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
19000 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
19001 attribute for SImode ssemov types. Reorder operand constraints.
19003 2013-03-20 Martin Jambor <mjambor@suse.cz>
19005 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
19006 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
19008 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
19010 * config/rs6000/predicates.md (indexed_address, update_address_mem
19011 update_indexed_address_mem): New predicates.
19012 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
19013 attribute for load/store instructions.
19014 * config/rs6000/dfp.md (movsd_store): Likewise.
19015 (movsd_load): Likewise.
19016 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
19017 (unnamed HI->DI extend define_insn): Likewise.
19018 (unnamed SI->DI extend define_insn): Likewise.
19019 (unnamed QI->SI extend define_insn): Likewise.
19020 (unnamed QI->HI extend define_insn): Likewise.
19021 (unnamed HI->SI extend define_insn): Likewise.
19022 (unnamed HI->SI extend define_insn): Likewise.
19023 (extendsfdf2_fpr): Likewise.
19024 (movsi_internal1): Likewise.
19025 (movsi_internal1_single): Likewise.
19026 (movhi_internal): Likewise.
19027 (movqi_internal): Likewise.
19028 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
19029 attribute for load/store instructions.
19030 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
19032 (mov<mode>_softfloat): Likewise.
19033 (mov<mode>_hardfloat32): Likewise.
19034 (mov<mode>_hardfloat64): Likewise.
19035 (mov<mode>_softfloat64): Likewise.
19036 (movdi_internal32): Likewise.
19037 (movdi_internal64): Likewise.
19038 (probe_stack_<mode>): Likewise.
19040 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
19042 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
19043 floating point, and decimal floating point to reload iterator.
19045 * config/rs6000/constraints.md (wl constraint): New constraints to
19046 return FLOAT_REGS if certain options are used to reduce the number
19047 of separate patterns that exist in the file.
19048 (wx constraint): Likewise.
19049 (wz constraint): Likewise.
19051 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
19052 -mdebug=reg, print wg, wl, wx, and wz constraints.
19053 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
19054 Initialize the reload functions for 64-bit binary/decimal floating
19056 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
19057 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
19058 create the buffer on the stack to overcome not having a 32-bit
19060 (rs6000_emit_move): Likewise.
19061 (rs6000_secondary_memory_needed_rtx): Likewise.
19062 (rs6000_alloc_sdmode_stack_slot): Likewise.
19063 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
19064 via xxlxor, just like DFmode 0.0.
19066 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
19067 define as 1 if we are running on a power7 or newer.
19068 (enum r6000_reg_class_enum): Add new constraints.
19070 * config/rs6000/dfp.md (movsd): Delete, combine with binary
19071 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
19072 with other moves by using conditional constraits (wg). Use LFIWZX
19073 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
19074 (movsd splitter): Likewise.
19075 (movsd_hardfloat): Likewise.
19076 (movsd_softfloat): Likewise.
19078 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
19079 binary and decimal floating point moves.
19080 (fmove_ok): New attributes to combine binary and decimal floating
19081 point moves, and to combine power6x (mfpgpr) moves along normal
19083 (real_value_to_target): Likewise.
19084 (f32_lr): Likewise.
19085 (f32_lm): Likewise.
19086 (f32_li): Likewise.
19087 (f32_sr): Likewise.
19088 (f32_sm): Likewise.
19089 (f32_si): Likewise.
19090 (movsf): Combine binary and decimal floating point moves. Combine
19091 power6x (mfpgpr) moves with other moves by using conditional
19092 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
19093 (mov<mode> for SFmode/SDmode); Likewise.
19094 (SFmode/SDmode splitters): Likewise.
19095 (movsf_hardfloat): Likewise.
19096 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
19097 (movsf_softfloat): Likewise.
19098 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
19100 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
19101 wx and wz constraints.
19103 * config/rs6000/constraints.md (wg constraint): New constraint to
19104 return FLOAT_REGS if -mmfpgpr (power6x) was used.
19106 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
19109 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
19110 -mdebug=reg, print wg, wl, wx, and wz constraints.
19111 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
19112 Initialize the reload functions for 64-bit binary/decimal floating
19114 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
19115 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
19116 create the buffer on the stack to overcome not having a 32-bit
19118 (rs6000_emit_move): Likewise.
19119 (rs6000_secondary_memory_needed_rtx): Likewise.
19120 (rs6000_alloc_sdmode_stack_slot): Likewise.
19121 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
19122 via xxlxor, just like DFmode 0.0.
19124 * config/rs6000/dfp.md (movdd): Delete, combine with binary
19125 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
19126 with other moves by using conditional constraits (wg). Use LFIWZX
19127 and STFIWX for loading SDmode on power7.
19128 (movdd splitters): Likewise.
19129 (movdd_hardfloat32): Likewise.
19130 (movdd_softfloat32): Likewise.
19131 (movdd_hardfloat64_mfpgpr): Likewise.
19132 (movdd_hardfloat64): Likewise.
19133 (movdd_softfloat64): Likewise.
19135 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
19136 64-bit binary and decimal floating point moves.
19137 (FMOVE64X): Likewise.
19138 (movdf): Combine 64-bit binary and decimal floating point moves.
19139 Combine power6x (mfpgpr) moves with other moves by using
19140 conditional constraits (wg).
19141 (mov<mode> for DFmode/DDmode): Likewise.
19142 (DFmode/DDmode splitters): Likewise.
19143 (movdf_hardfloat32): Likewise.
19144 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
19145 (movdf_softfloat32): Likewise.
19146 (movdf_hardfloat64_mfpgpr): Likewise.
19147 (movdf_hardfloat64): Likewise.
19148 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
19149 (movdf_softfloat64): Likewise.
19150 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
19151 (reload_<mode>_load): Move to later in the file so they aren't in
19152 the middle of the floating point move insns.
19153 (reload_<mode>_store): Likewise.
19155 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
19158 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
19159 constraint if -mdebug=reg.
19160 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
19161 Enable using dd reload support if needed.
19163 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
19164 binary and decimal floating point moves in rs6000.md.
19165 (movtd_internal): Likewise.
19167 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
19168 decimal floating point moves.
19170 (movtf_internal): Likewise.
19171 (mov<mode>_internal, TDmode/TFmode): Likewise.
19172 (movtf_softfloat): Likewise.
19173 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
19175 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
19176 movdi_internal64, using wg constraint for move direct operations.
19177 (movdi_internal64): Likewise.
19179 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
19180 MODES_TIEABLE_P for selected modes. Print the numerical value of
19181 the various virtual registers. Use GPR/FPR first/last values,
19182 instead of hard coding the register numbers. Print which modes
19183 have reload functions registered.
19184 (rs6000_option_override_internal): If -mdebug=reg, trace the options
19185 settings before/after setting cpu, target and subtarget settings.
19186 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
19187 and for secondary reload failures in rs6000_secondary_reload_inner.
19188 (rs6000_secondary_reload_fail): Likewise.
19189 (rs6000_secondary_reload_inner): Likewise.
19191 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
19192 macros for first/last GPR and FPR registers.
19193 (LAST_GPR_REGNO): Likewise.
19194 (FIRST_FPR_REGNO): Likewise.
19195 (LAST_FPR_REGNO): Likewise.
19197 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
19198 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
19199 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
19200 (vcond<mode><mode>): Likewise.
19201 (vcondu<mode><mode>): Likewise.
19202 (vector_gtu<mode>): Likewise.
19203 (vector_gte<mode>): Likewise.
19204 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
19205 to prevent the compiler from converting DImode operations to TImode.
19206 (ior<mode>3): Likewise.
19207 (and<mode>3): Likewise.
19208 (one_cmpl<mode>2): Likewise.
19209 (nor<mode>3): Likewise.
19210 (andc<mode>3): Likewise.
19212 * config/rs6000/constraints.md (wt constraint): New constraint
19213 that returns VSX_REGS if TImode is allowed in VSX registers.
19215 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
19216 constant under VSX.
19218 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
19219 similar to TImode, but it is restricted to being in the GPRs.
19221 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
19222 TImode to occupy a single VSX register.
19224 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
19225 -mvsx-timode for power7/power8.
19226 (power7 cpu): Likewise.
19227 (power8 cpu): Likewise.
19229 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
19230 sure that TFmode/TDmode take up two registers if they are ever
19231 allowed in the upper VSX registers.
19232 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
19234 (rs6000_init_hard_regno_mode_ok): Likewise.
19235 (rs6000_debug_reg_global): Add debugging for PTImode and wt
19236 constraint. Print if LRA is turned on.
19237 (rs6000_option_override_internal): Give an error if -mvsx-timode
19238 and VSX is not enabled.
19239 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
19240 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
19241 to reg+offset addressing. Use PTImode when checking offset
19242 addresses for validity.
19243 (reg_offset_addressing_ok_p): Likewise.
19244 (rs6000_legitimate_offset_address_p): Likewise.
19245 (rs6000_legitimize_address): Likewise.
19246 (rs6000_legitimize_reload_address): Likewise.
19247 (rs6000_legitimate_address_p): Likewise.
19248 (rs6000_eliminate_indexed_memrefs): Likewise.
19249 (rs6000_emit_move): Likewise.
19250 (rs6000_secondary_reload): Likewise.
19251 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
19252 reloads to fpr registers to continue to use reg+offset addressing,
19253 but 64-bit reloads to altivec registers need reg+reg addressing.
19254 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
19255 it. Treat LO_SUM like a PLUS operation.
19256 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
19257 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
19258 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
19259 registers to share a register with a smaller sized type, since VSX
19260 puts scalars in the upper 64-bits.
19261 (print_operand): Add support for PTImode.
19262 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
19263 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
19264 registers, but don't have arithmetic support.
19265 (rs6000_memory_move_cost): Add test for VSX.
19266 (rs6000_opt_masks): Add -mvsx-timode.
19268 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
19271 (VSr): Use wt constraint for TImode.
19272 (VSv): Drop TImode support.
19273 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
19274 (vsx_movti_64bit): Likewise.
19275 (vsx_movti_32bit): Likewise.
19276 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
19277 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
19278 one '?' on the appropriate output constraint. Do not allow TImode
19279 logical operations on 32-bit systems.
19280 (vsx_ior<mode>3): Likewise.
19281 (vsx_xor<mode>3): Likewise.
19282 (vsx_one_cmpl<mode>2): Likewise.
19283 (vsx_nor<mode>3): Likewise.
19284 (vsx_andc<mode>3): Likewise.
19285 (vsx_concat_<mode>): Likewise.
19286 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
19288 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
19289 OPTION_MASK_VSX_TIMODE.
19290 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
19291 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
19293 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
19294 (TI2 iterator): New iterator for TImode, PTImode.
19295 (wd mode attribute): Add values for vector types.
19296 (movti_string): Replace TI move operations with operations for TImode
19297 and PTImode. Add support for TImode being allowed in VSX registers.
19298 (mov<mode>_string, TImode/PTImode): Likewise.
19299 (movti_ppc64): Likewise.
19300 (mov<mode>_ppc64, TImode/PTImode): Likewise.
19301 (TI mode splitters): Likewise.
19303 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
19306 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
19308 PR tree-optimization/56355
19309 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
19310 Also handle integers with undefined overflow.
19312 2013-03-20 Catherine Moore <clm@codesourcery.com>
19313 Maciej W. Rozycki <macro@codesourcery.com>
19314 Tom de Vries <tom@codesourcery.com>
19315 Nathan Sidwell <nathan@codesourcery.com>
19316 Iain Sandoe <iain@codesourcery.com>
19317 Nathan Froyd <froydnj@codesourcery.com>
19318 Chao-ying Fu <fu@mips.com>
19320 * doc/extend.texi: (micromips, nomicromips, nocompression):
19321 Document new function attributes.
19322 * doc/invoke.texi (minterlink-compressed, mmicromips,
19323 m14k, m14ke, m14kec): Document new options.
19324 (minterlink-mips16): Update documentation.
19325 * doc/md.texi (ZC, ZD): Document new constraints.
19326 * configure.ac (gcc_cv_as_micromips): Check if linker
19327 supports the .set micromips directive.
19328 * configure: Regenerate.
19329 * config.in: Regenerate.
19330 * config/mips/mips-tables.opt: Regenerate.
19331 * config/mips/micromips.md: New file.
19332 * constraints.md (ZC, ZD): New constraints.
19333 * config/mips/predicates.md (movep_src_register): New predicate.
19334 (movep_src_operand): New predicate.
19335 (non_volatile_mem_operand): New predicate.
19336 * config/mips/mips.md (multimem): New type.
19337 (length): Differentiate between 17-bit and 18-bit branch offsets.
19338 (MOVEP1, MOVEP2): New mode iterator.
19339 (mov_<load>l): Use ZC constraint.
19340 (mov_<load>r): Likewise.
19341 (mov_<store>l): Likewise.
19342 (mov_<store>r): Likewise.
19343 (*branch_equality<mode>_inverted): Add microMIPS support.
19344 (*branch_equality<mode>): Likewise.
19345 (*jump_absolute): Likewise.
19346 (indirect_jump_<mode>): Likewise.
19347 (tablejump_<mode>): Likewise.
19348 (<optab>_internal): Likewise.
19349 (sibcall_internal): Likewise.
19350 (sibcall_value_internal): Likewise.
19351 (prefetch): Use constraint ZD.
19352 * config/mips/mips.opt (minterlink-compressed): New option.
19353 (minterlink-mips16): Now an alias for minterlink-compressed.
19354 (mmicromips): New option.
19355 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
19356 (compare_and_swap_12): Likewise.
19357 (sync_add<mode>): Likewise.
19358 (sync_<optab>_12): Likewise.
19359 (sync_old_<optab>_12): Likewise.
19360 (sync_new_<optab>_12): Likewise.
19361 (sync_nand_12): Likewise.
19362 (sync_old_nand_12): Likewise.
19363 (sync_new_nand_12): Likewise.
19364 (sync_sub<mode>): Likewise.
19365 (sync_old_add<mode>): Likewise.
19366 (sync_old_sub<mode>): Likewise.
19367 (sync_new_add<mode>): Likewise.
19368 (sync_new_sub<mode>): Likewise.
19369 (sync_<optab><mode>): Likewise.
19370 (sync_old_<optab><mode>): Likewise.
19371 (sync_new_<optab><mode>): Likewise.
19372 (sync_nand<mode>): Likewise.
19373 (sync_old_nand<mode>): Likewise.
19374 (sync_new_nand<mode>): Likewise.
19375 (sync_lock_test_and_set<mode>): Likewise.
19376 (test_and_set_12): Likewise.
19377 (atomic_compare_and_swap<mode>): Likewise.
19378 (atomic_exchange<mode>_llsc): Likewise.
19379 (atomic_fetch_add<mode>_llsc): Likewise.
19380 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
19381 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
19382 (umips_save_restore_pattern_p): Likewise.
19383 (umips_load_store_pair_p): Likewise.
19384 (umips_output_load_store_pair): Likewise.
19385 (umips_movep_target_p): Likewise.
19386 (umips_12bit_offset_address_p): Likewise.
19387 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
19388 (mips_base_mips16): Rename this...
19389 (mips_base_compression_flags): ...to this. Update all uses.
19390 (mips_attribute_table): Add micromips, nomicromips and nocompression.
19391 (mips_mips16_decl_p): Delete.
19392 (mips_nomips16_decl_p): Delete.
19393 (mips_get_compress_on_flags): New function.
19394 (mips_get_compress_off_flags): New function.
19395 (mips_get_compress_mode): New function.
19396 (mips_get_compress_on_name): New function.
19397 (mips_get_compress_off_name): New function.
19398 (mips_insert_attributes): Support multiple compression types.
19399 (mips_merge_decl_attributes): Likewise.
19400 (umips_12bit_offset_address_p): New function.
19401 (mips_start_function_definition): Emit .set micromips directive.
19402 (mips_call_may_need_jalx_p): New function.
19403 (mips_function_ok_for_sibcall): Add microMIPS support.
19404 (mips_print_operand_punctuation): Support short delay slots and
19406 (umips_swm_mask, umips_swm_encoding): New.
19407 (umips_build_save_restore): New function.
19408 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
19409 (was_mips16_p): Remove.
19410 (old_compression_mode): New.
19411 (mips_set_compression_mode): New function.
19412 (mips_set_current_function): Add microMIPS support.
19413 (mips_option_override): Likewise.
19414 (umips_save_restore_pattern_p): New function.
19415 (umips_output_save_restore): New function.
19416 (umips_load_store_pair_p_1): New function.
19417 (umips_load_store_pair_p): New function.
19418 (umips_output_load_store_pair_1): New function.
19419 (umips_output_load_store_pair): New function.
19420 (umips_movep_target_p) New function.
19421 (mips_prepare_pch_save): Add microMIPS support.
19422 * config/mips/mips.h (TARGET_COMPRESSION): New.
19423 (TARGET_CPU_CPP_BUILTINS): Update macro
19424 to use new compression flags and to support microMIPS.
19425 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
19426 (MIPS_ARCH_FLOAT_SPEC): Likewise.
19427 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
19428 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
19429 (ASM_SPEC): Support mmicromips and mno-micromips.
19430 (M16STORE_REG_P): New macro.
19431 (MIPS_CALL): Support TARGET_MICROMIPS.
19432 (MICROMIPS_J): New macro.
19433 (mips_base_mips16): Rename this...
19434 (mips_base_compression_flags): ...to this.
19435 (UMIPS_12BIT_OFFSET_P): New macro.
19436 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
19437 (MULTILIB_DIRNAMES): Likewise.
19438 2013-03-20 Richard Biener <rguenther@suse.de>
19440 PR tree-optimization/56661
19441 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
19442 the result does not have to be distinct.
19444 2013-03-20 Richard Biener <rguenther@suse.de>
19446 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
19449 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
19450 Steven Bosscher <steven@gcc.gnu.org>
19452 PR rtl-optimization/56605
19453 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
19455 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
19458 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
19459 that require movd instead of movq.
19461 2013-03-20 Richard Biener <rguenther@suse.de>
19463 * tree-ssa-structalias.c (struct variable_info): Add pointer
19464 to the first field of an aggregate with sub-vars. Make
19465 this and the pointer to the next subfield its ID.
19466 (vi_next): New function.
19467 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
19468 storedanything_id, integer_id): Increment by one.
19469 (new_var_info, get_call_vi, lookup_call_clobber_vi,
19470 get_call_clobber_vi): Adjust.
19471 (solution_set_expand): Simplify and speedup.
19472 (solution_set_add): Inline into ...
19473 (set_union_with_increment): ... this. Adjust accordingly.
19474 (do_sd_constraint): Likewise.
19475 (do_ds_constraint): Likewise.
19476 (do_complex_constraint): Simplify.
19477 (build_pred_graph): Adjust.
19478 (solve_graph): Likewise. Simplify and speedup.
19479 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
19480 get_constraint_for_component_ref, get_constraint_for_1,
19481 first_vi_for_offset, first_or_preceding_vi_for_offset,
19482 create_function_info_for, create_variable_info_for_1,
19483 create_variable_info_for, intra_create_variable_infos): Adjust.
19484 (init_base_vars): Push NULL for ID zero.
19485 (compute_points_to_sets): Adjust.
19487 2013-03-20 Richard Biener <rguenther@suse.de>
19489 * cfgloop.c (verify_loop_structure): Streamline and avoid
19490 ICEing on corrupt loop tree.
19491 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
19494 2013-03-20 Richard Biener <rguenther@suse.de>
19496 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
19497 check whether an SSA update is needed.
19499 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
19501 * config/mips/constraints.md (T): Rename to...
19505 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
19506 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
19508 2013-03-19 Ian Bolton <ian.bolton@arm.com>
19510 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
19511 (*subsi3_carryin_uxtw): Likewise.
19513 2013-03-19 Ian Bolton <ian.bolton@arm.com>
19515 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
19516 (*rorsi3_insn_uxtw): Likewise.
19518 2013-03-19 Ian Bolton <ian.bolton@arm.com>
19520 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
19521 (*extrsi5_insn_uxtw): Likewise.
19523 2013-03-19 Richard Biener <rguenther@suse.de>
19525 PR tree-optimization/56273
19526 * passes.c (init_optimization_passes): Move second VRP after DOM.
19528 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
19530 * config/i386/i386.md (*movti_internal): Merge from
19531 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
19532 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
19533 nox64 isa attributes.
19535 2013-03-18 Richard Biener <rguenther@suse.de>
19537 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
19539 (merge_node_constraints): Likewise.
19540 (build_succ_graph): Likewise.
19541 (valid_graph_edge): Inline into single caller.
19542 (unify_nodes): Likewise. Use bitmap_set_bit return value
19544 (scc_visit): Fix formatting and variable use.
19545 (do_sd_constraint): Use gcc_checking_assert.
19546 (do_ds_constraint): Likewise.
19547 (do_complex_constraint): Likewise.
19548 (condense_visit): Likewise. Cleanup.
19549 (dump_pred_graph): New function.
19550 (perform_var_substitution): Dump the pred-graph before
19551 variable substitution.
19552 (find_equivalent_node): Use gcc_checking_assert.
19553 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
19555 2013-03-18 Richard Biener <rguenther@suse.de>
19557 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
19558 Remove cond_expr_stmt_list argument and do not gimplify the
19560 (vect_loop_versioning): Adjust.
19561 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
19562 Cleanup to use less temporaries.
19563 (vect_create_data_ref_ptr): Cleanup.
19565 2013-03-18 Jakub Jelinek <jakub@redhat.com>
19567 PR tree-optimization/56635
19568 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
19569 require types_compatible_p types.
19571 2013-03-18 Nick Clifton <nickc@redhat.com>
19573 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
19574 spurious backslash.
19576 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
19577 Add missing line to comment describing function.
19579 2013-03-18 Richard Biener <rguenther@suse.de>
19581 PR tree-optimization/56210
19582 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
19583 Handle string / character search functions.
19584 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
19586 2013-03-18 Richard Biener <rguenther@suse.de>
19588 PR middle-end/56483
19589 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
19590 and implement properly.
19591 * gimple.h (gimple_cond_single_var_p): Remove.
19593 2013-03-18 Richard Biener <rguenther@suse.de>
19595 * tree-data-ref.h (find_data_references_in_loop): Declare.
19596 * tree-data-ref.c (get_references_in_stmt): Use a stack
19597 vector pre-allocated in the callers.
19598 (find_data_references_in_stmt): Adjust.
19599 (graphite_find_data_references_in_stmt): Likewise.
19600 (create_rdg_vertices): Likewise.
19601 (find_data_references_in_loop): Export.
19602 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
19603 Compute dependences here...
19604 (vect_analyze_data_refs): ...not here. When we encounter
19605 a non-vectorizable data reference in basic-block vectorization
19606 truncate the data reference vector. Do not bother to
19607 fixup data-dependence information for gather loads.
19608 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
19609 of data references, as reported.
19611 2013-03-18 Richard Biener <rguenther@suse.de>
19613 PR tree-optimization/3713
19614 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
19615 has_constants and expr.
19616 (stmt_has_constants): Properly valueize SSA names when deciding
19617 whether the stmt has constants.
19619 2013-03-18 Richard Biener <rguenther@suse.de>
19621 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
19622 whole function when there is nothing to do.
19623 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
19624 * tree-vectorizer.c (vectorize_loops): Update virtual and
19625 loop-closed SSA once.
19626 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
19628 2013-03-18 Richard Biener <rguenther@suse.de>
19630 PR middle-end/56113
19631 * domwalk.c (bb_postorder): New global static.
19632 (cmp_bb_postorder): New function.
19633 (walk_dominator_tree): Replace scheme imposing an order for
19634 visiting dominator sons by one sorting them at the time they
19635 are pushed on the stack.
19637 2013-03-18 Richard Biener <rguenther@suse.de>
19639 PR tree-optimization/39326
19640 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
19641 (struct mem_ref): Replace mem member with ao_ref typed member.
19642 (MEM_ANALYZABLE): Adjust.
19643 (memref_eq): Likewise.
19644 (mem_ref_alloc): Likewise.
19645 (gather_mem_refs_stmt): Likewise.
19646 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
19647 (execute_sm_if_changed_flag_set): Adjust.
19648 (execute_sm): Likewise.
19649 (ref_always_accessed_p): Likewise.
19650 (refs_independent_p): Likewise.
19651 (can_sm_ref_p): Likewise.
19653 2013-03-18 Jakub Jelinek <jakub@redhat.com>
19656 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
19657 return 1 even for !unsignedp.
19659 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
19661 * config/i386/i386.md (isa): Add x64 and nox64.
19662 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
19663 (*pushtf): Enable *roF alternative for x64 isa only.
19664 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
19665 mode attribute of integer alternatives to DImode for TARGET_64BIT.
19666 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
19667 (*movtf_internal): Merge from *movtf_internal_rex64 and
19668 *movtf_internal_sse. Use x64 and nox64 isa attributes.
19669 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
19670 nox64 isa attributes.
19671 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
19672 nox64 isa attributes.
19673 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
19675 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
19677 * config/alpha/alpha.c (TARGET_LRA_P): New define.
19679 2013-03-17 Jakub Jelinek <jakub@redhat.com>
19682 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
19683 class names. Remove trailing comma after "ALL_REGS".
19685 2013-03-16 Jan Hubicka <jh@suse.cz>
19687 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
19688 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
19689 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
19690 of cgraph_get_create_node.
19691 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
19693 2013-03-16 Jason Merrill <jason@redhat.com>
19696 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
19697 with DW_AT_default_value.
19699 2013-03-16 Jakub Jelinek <jakub@redhat.com>
19701 * BASE-VER: Set to 4.9.0.
19703 2013-03-14 Andi Kleen <ak@linux.intel.com>
19706 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
19707 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
19708 Document _x* TSX intrinsics.
19710 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
19711 David Holsgrove <david.holsgrove@xilinx.com>
19713 * configure.ac: Add MicroBlaze TLS support detection.
19714 * configure: Regenerate.
19715 * config/microblaze/microblaze-protos.h
19716 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
19717 symbol_mentioned_p, label_mentioned_p): Add prototypes.
19718 * config/microblaze/microblaze.c (microblaze_address_type): Add
19719 ADDRESS_TLS and tls_reloc address types.
19720 (microblaze_address_info): Add tls_reloc.
19721 (TARGET_HAVE_TLS): Define.
19722 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
19723 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
19724 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
19725 load_tls_operand, microblaze_call_tls_get_addr,
19726 microblaze_legitimize_tls_address): New functions.
19727 (microblaze_classify_unspec): Handle UNSPEC_TLS.
19728 (get_base_reg): Use microblaze_tls_symbol_p.
19729 (microblaze_classify_address): Handle TLS.
19730 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
19731 label_mentioned_p and microblaze_tls_referenced_p.
19732 (microblaze_legitimize_address): Handle TLS.
19733 (microblaze_address_insns): Handle ADDRESS_TLS.
19734 (pic_address_needs_scratch): Handle TLS.
19735 (print_operand_address): Handle TLS.
19736 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
19737 (microblaze_expand_move): Handle TLS.
19738 (microblaze_legitimate_constant_p): Check
19739 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
19740 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
19741 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
19742 (PIC_OFFSET_TABLE_REGNUM): Set.
19743 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
19744 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
19745 (addsi3, movsi_internal2, movdf_internal): Update constraints
19746 * config/microblaze/predicates.md (arith_plus_operand): Define
19747 (move_operand): Redefine as move_src_operand,
19748 check microblaze_tls_referenced_p.
19750 2013-03-14 Ian Bolton <ian.bolton@arm.com>
19752 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
19753 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
19755 2013-03-14 Ian Bolton <ian.bolton@arm.com>
19757 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
19760 2013-03-14 Jakub Jelinek <jakub@redhat.com>
19762 PR tree-optimization/53265
19763 * common.opt (Waggressive-loop-optimizations): New option.
19764 * tree-ssa-loop-niter.c: Include tree-pass.h.
19765 (do_warn_aggressive_loop_optimizations): New function.
19766 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
19767 if number_of_latch_executions returned constant.
19768 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
19769 early. If number_of_latch_executions returned constant, set
19770 nb_iterations_upper_bound back to it.
19771 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
19773 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
19774 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
19776 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
19777 (MULTILIB_OSDIRNAMES): Set.
19778 * genmultilib: If defaultosdirname doesn't start with :: , set
19779 defaultosdirname2 instead, clear it and emit two . multilib_raw
19780 entries instead of just one.
19782 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
19784 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
19785 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
19786 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
19787 (SUBTARGET_OVERRIDE_OPTIONS): New.
19789 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
19792 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
19793 (musermode): Convert to Var(TARGET_USERMODE).
19794 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
19795 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
19796 * config/sh/sh.c (sh_option_override): Use
19797 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
19798 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
19800 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
19802 (udivsi3_i4_single, divsi3_i4_single): Use
19803 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
19805 2013-03-13 Dave Korn <dave.korn.cygwin@....>
19807 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
19810 2013-03-13 Richard Biener <rguenther@suse.de>
19812 PR tree-optimization/56608
19813 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
19814 calls when vectorizing basic-blocks.
19816 2013-03-13 Jakub Jelinek <jakub@redhat.com>
19819 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
19822 2013-03-12 Jakub Jelinek <jakub@redhat.com>
19824 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
19826 2013-03-11 Jan Hubicka <jh@suse.cz>
19829 * lto-streamer-out.c (output_symbol_p): Skip references from
19830 constructors of external variables.
19832 2013-03-11 Jan Hubicka <jh@suse.cz>
19834 PR middle-end/56571
19835 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
19837 * emit-rtl.c (verify_rtx_sharing): Likewise.
19838 (copy_insn_1): Likewise.
19839 * rtl.c (copy_rtx): Likewise.
19841 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
19844 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
19845 output_operand_lossage message.
19847 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
19850 * arm.c (shift_op): Validate RTL pattern on the fly.
19851 (arm_print_operand, case 'S'): Don't use shift_operator to validate
19854 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
19857 * config/pa/pa.md (call_value): Check for calls to powf and direct to
19858 new call patterns that clobber %fr12.
19859 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
19860 split and postreload patterns.
19861 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
19862 registers %fr12 and %fr12R as call used.
19864 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
19866 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
19867 (canon_address, record_store, replace_read, check_mem_read_rtx,
19868 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
19869 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
19870 rest_of_handle_dse): Likewise.
19872 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
19874 PR middle-end/56524
19875 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
19877 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
19878 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
19879 (save_optabs_if_changed): Replace with...
19880 (init_tree_optimization_optabs): ...this.
19881 * optabs.c (save_optabs_if_changed): Rename to...
19882 (init_tree_optimization_optabs): ...this. Take the optimization node
19883 as argument. Do nothing if the base optabs are already correct.
19884 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
19885 to recompute optabs.
19886 * function.h (function): Remove optabs field.
19887 * function.c (invoke_set_current_function_hook): Call
19888 init_tree_optimization_optabs. Use the result to initialize
19891 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
19893 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
19894 if GTMA_HAS_NO_INSTRUMENTATION.
19895 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
19896 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
19897 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
19898 * gimple-pretty-print.c (dump_gimple_transaction): Handle
19899 GTMA_HAS_NO_INSTRUMENTATION.
19901 2013-03-08 Jakub Jelinek <jakub@redhat.com>
19903 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
19906 2013-03-08 Marek Polacek <polacek@redhat.com>
19907 Jakub Jelinek <jakub@redhat.com>
19909 PR tree-optimization/56478
19910 * predict.c (is_comparison_with_loop_invariant_p): Change the
19911 type of loop_step to tree.
19912 (predict_loops): Adjust.
19913 (predict_iv_comparison): Perform the computations on double_ints.
19915 2013-03-08 Richard Biener <rguenther@suse.de>
19917 PR tree-optimization/56570
19918 * tree-cfg.c (verify_expr_location_1): Verify locations for
19920 * tree-sra.c (create_access_replacement): Strip locations
19921 from DECL_DEBUG_EXPRs.
19923 2013-03-08 Richard Biener <rguenther@suse.de>
19925 * tree-inline.c (expand_call_inline): Do not associate
19926 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
19927 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
19929 2013-03-08 Richard Biener <rguenther@suse.de>
19931 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
19932 or block changes with -Og. Fix for location / block encoding
19933 changes and PHI arguments with locations.
19935 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
19937 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
19939 (struct output_info): Likewise.
19940 (register_overhead): Remove bad gcc_assert.
19941 (bitmap_find_bit): If there is only a single bitmap element, do not
19942 count a miss as a search.
19943 (print_statistics): Update for counter type changes.
19944 (dump_bitmap_statistics): Likewise. Print headers such that they
19945 are properly lined up with the printed counters.
19947 2013-03-07 Jakub Jelinek <jakub@redhat.com>
19949 PR tree-optimization/56559
19950 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
19951 check that it has only a single use.
19953 2013-03-07 Richard Biener <rguenther@suse.de>
19955 * doc/invoke.texi (fwhole-program): Discourage use in combination
19958 2013-03-06 Jakub Jelinek <jakub@redhat.com>
19960 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
19962 PR tree-optimization/56539
19963 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
19964 instead of GSI_CONTINUE_LINKING as last argument to
19965 force_gimple_operand_gsi. Adjust function comment.
19967 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
19970 PR middle-end/56548
19971 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
19972 promoted mode, convert the result back to the original mode.
19974 2013-03-06 Richard Biener <rguenther@suse.de>
19976 PR middle-end/56294
19977 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
19978 (insert_updated_phi_nodes_compare_uids): New function.
19979 (update_ssa): Sort symbols_to_rename after UID before
19980 traversing it to insert PHI nodes.
19982 2013-03-06 Richard Biener <rguenther@suse.de>
19984 PR middle-end/50494
19985 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
19986 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
19989 2013-02-13 Richard Biener <rguenther@suse.de>
19992 * varasm.c (output_constant_def_1): Get the decl representing
19993 the constant as argument.
19994 (output_constant_def): Wrap output_constant_def_1.
19995 (make_decl_rtl): Use output_constant_def_1 with the decl
19996 representing the constant.
19997 (build_constant_desc): Optionally re-use a decl already
19998 representing the constant.
19999 (tree_output_constant_def): Adjust.
20001 2013-03-06 Joey Ye <joey.ye@arm.com>
20004 * gcc.c (convert_white_space): New function.
20005 (main): Handles white space in function name.
20007 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
20010 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
20011 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
20012 to SH_DIV_CALL_TABLE for TARGET_SH2.
20013 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
20015 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
20016 call-table options.
20018 2013-03-05 Sterling Augustine <saugustine@google.com>
20019 Cary Coutant <ccoutant@google.com>
20022 * dwarf2out.c (resolve_addr): Don't call
20023 remove_loc_list_addr_table_entries a second time for the same
20026 2013-03-05 Jakub Jelinek <jakub@redhat.com>
20029 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
20030 (avoid_complex_debug_insns): New function.
20031 (expand_debug_locations): Call it.
20033 PR rtl-optimization/56484
20034 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
20035 lifetimes of hard registers on small register class machines.
20037 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
20039 * config/microblaze/microblaze-protos.h: Rename
20040 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
20041 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
20043 (microblaze_fast_interrupt_function_p): New function.
20044 (microblaze_is_interrupt_handler): Rename to
20045 microblaze_is_interrupt_variant and add fast_interrupt check.
20046 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
20047 (save_restore_insns): Likewise.
20048 (compute_frame_size): Likewise.
20049 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
20050 (microblaze_globalize_label): Likewise.
20051 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
20052 * config/microblaze/microblaze.md: Use wrapper
20053 microblaze_is_interrupt_variant.
20055 2013-03-05 Kai Tietz <ktietz@redhat.com>
20057 * sdbout.c (sdbout_one_type): Switch to current function's section
20058 supporting cold/hot.
20060 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
20062 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
20065 2013-03-05 Jakub Jelinek <jakub@redhat.com>
20067 PR middle-end/56461
20068 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
20069 if VALGRIND_GET_VBITS is defined, temporarily make object
20070 memory all defined, and restore previous valgrind addressability
20071 and definability afterwards. Free this_object at the end.
20073 PR middle-end/56461
20074 * lra.c (lra): Call lra_clear_live_ranges if live_p,
20075 right before calling lra_create_live_ranges, also call it
20076 when clearing live_p. Only call lra_clear_live_ranges
20077 at the end if live_p.
20079 PR middle-end/56461
20080 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
20082 2013-03-05 Richard Biener <rguenther@suse.de>
20084 PR tree-optimization/56521
20085 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
20088 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
20091 * except.h (remove_unreachable_eh_regions): New prototype.
20092 * except.c (remove_eh_handler_splicer): New function, split out
20093 of remove_eh_handler.
20094 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
20095 warning about running it on many EH regions one at a time.
20096 (remove_unreachable_eh_regions_worker): New function, walk the
20097 EH tree in depth-first order and remove non-marked regions.
20098 (remove_unreachable_eh_regions): New function.
20099 * tree-eh.c (mark_reachable_handlers): New function, split out
20100 from remove_unreachable_handlers.
20101 (remove_unreachable_handlers): Use mark_reachable_handlers and
20102 remove_unreachable_eh_regions.
20103 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
20104 and remove_unreachable_eh_regions.
20106 2013-03-05 Richard Biener <rguenther@suse.de>
20108 PR middle-end/56525
20109 * loop-init.c (fix_loop_structure): Remove loops in two stages,
20110 not freeing them until the end.
20112 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20114 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
20116 2013-03-05 Richard Biener <rguenther@suse.de>
20118 PR tree-optimization/56270
20119 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
20120 of loads after scheduling an SLP instance.
20122 2013-03-05 Jakub Jelinek <jakub@redhat.com>
20124 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
20126 (check_gcc_parallelize): Run guality.exp as a separate job from
20127 vect.exp with unsorted.exp and $(dg_target_exps) separately from
20128 struct-layout-1.exp with stackalign.exp.
20130 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
20132 PR middle-end/56461
20133 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
20134 load_index sbitmap even if some bit in it isn't set.
20136 PR middle-end/56461
20137 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
20138 (discover_iteration_bound_by_body_walk): Change queues to
20139 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
20140 spelling in comment. Call safe_push on queues[bound_index] directly.
20141 Release queues[queue_index] in every iteration unconditionally.
20142 Release bounds vector.
20144 PR middle-end/56461
20145 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
20146 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
20147 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
20150 2013-03-05 Richard Biener <rguenther@suse.de>
20153 * tree-inline.c (remap_blocks_to_null): New function.
20154 (expand_call_inline): When expanding a call stmt without
20155 an associated block inline remap all callee blocks to NULL.
20157 2013-03-05 Jakub Jelinek <jakub@redhat.com>
20159 PR rtl-optimization/56494
20160 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
20161 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
20162 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
20164 PR middle-end/56461
20165 * sel-sched-ir.c (free_sched_pools): Release
20166 succs_info_pool.stack[succs_info_pool.max_top] vectors too
20167 if succs_info_pool.max_top isn't -1.
20170 * opts.c (opts_obstack, opts_concat): Moved to...
20171 * opts-common.c (opts_obstack, opts_concat): ... here.
20173 2013-03-04 Jakub Jelinek <jakub@redhat.com>
20175 PR middle-end/56461
20176 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
20178 2013-03-04 Martin Jambor <mjambor@suse.cz>
20180 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
20181 all appropriate places.
20183 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
20185 PR tree-optimization/56424
20186 * ipa-split.c (split_function): Do not set the RSO flag if result is
20187 not by reference and its type is a register type.
20189 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
20191 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
20192 (microblaze_legitimate_pic_operand): Likewise
20193 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
20194 new function microblaze_legitimate_pic_operand
20195 * config/microblaze/microblaze-protos.h
20196 (microblaze_legitimate_pic_operand): Declare.
20198 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20200 * config/microblaze/predicates.md (call_insn_simple_operand):
20201 New predicate for supported rtx code types.
20202 * config/microblaze/microblaze.md (call_internal1): Use
20203 call_insn_simple_operand predicate.
20205 2013-03-04 Jakub Jelinek <jakub@redhat.com>
20207 PR middle-end/56461
20208 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
20209 partitions.ordered_remove.
20211 PR middle-end/56461
20212 * tree-vect-stmts.c (vectorizable_conversion): Don't call
20213 vec_oprnds0.create (1) for modifier == NONE.
20215 PR middle-end/56461
20216 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
20217 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
20218 vec_oprnds1 right before pushing anything to it for
20221 PR middle-end/56461
20222 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
20223 set nbbs to 0 instead of having separate code path.
20224 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
20225 instead of false as last argument if returning NULL.
20227 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
20229 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
20230 the attribute is now called "target" instead of "option".
20231 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
20232 * doc/tm.texi.in (Target Attributes): Likewise document the correct
20233 attribute/pragma name for TARGET_OPTION_VALID_P and
20234 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
20235 * doc/tm.texi: Regenerated.
20237 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
20239 * config/microblaze/microblaze.c:
20240 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
20241 * config/microblaze/microblaze.h: Add -mxl-reorder to
20243 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
20244 instructions emitted if TARGET_REORDER.
20245 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
20246 or 0 for -m/-mno case, but initialises as 2 to detect default use case
20249 2013-03-01 Xinliang David Li <davidxl@google.com>
20251 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
20254 2013-03-01 Jakub Jelinek <jakub@redhat.com>
20256 PR middle-end/56461
20257 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
20258 vector even when returning true. Fix up function comment formatting.
20260 PR middle-end/56461
20261 * ira-build.c (ira_loop_nodes_count): New variable.
20262 (create_loop_tree_nodes): Initialize it.
20263 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
20265 PR middle-end/56461
20266 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
20267 method on dr_chain and result_chain.
20268 * tree-vect-stmts.c (vectorizable_store): Only call
20269 result_chain.create if j == 0.
20271 PR middle-end/56461
20272 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
20273 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
20274 before overwriting it.
20276 2013-03-01 Tobias Burnus <burnus@net-b.de>
20278 * doc/extended.texi (C Extensions): Change order in @menu
20280 (Other MIPS Built-in Functions): Move last MIPS entry before
20281 "picoChip Built-in Functions".
20282 (SH Built-in Functions): Move after RX Built-in Functions.
20283 * doc/gcc.texi (Introduction): Change order in @menu
20285 * doc/md.texi (Constraints): Ditto.
20286 * gty.texi (Type Information): Ditto.
20287 (User-provided marking routines for template types): Make
20289 * doc/invoke.texi (AArch64 Options): Move before
20290 "Adapteva Epiphany Options".
20292 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
20293 Jakub Jelinek <jakub@redhat.com>
20296 * asan.c (gate_asan): Lookup no_sanitize_address instead of
20297 no_address_safety_analysis attribute.
20298 * doc/extend.texi (no_address_safety_attribute): Rename to
20299 no_sanitize_address attribute, mention no_address_safety_analysis
20300 attribute as deprecated alias.
20302 2013-02-28 Jakub Jelinek <jakub@redhat.com>
20304 PR middle-end/56461
20305 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
20306 type to vec<vec<tree> > *.
20307 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
20308 to be vec<tree> instead of vec<tree> *, set vec_defs
20309 to vNULL and call vec_defs.create (number_of_vects), adjust other
20311 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
20312 vectorizable_condition): Adjust vect_get_slp_defs callers.
20314 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
20316 * config/aarch64/aarch64.c
20317 (aarch64_float_const_representable): Remove unused variable.
20319 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
20321 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
20323 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
20325 * config/aarch64/aarch64-builtins.c
20326 (aarch64_init_simd_builtins): Make static.
20328 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
20330 * config/aarch64/aarch64.c
20331 (aarch64_simd_make_constant): Make static.
20333 2013-02-28 Martin Jambor <mjambor@suse.cz>
20335 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
20336 with no initialization to the RHS of debug statements.
20338 2013-02-28 Martin Jambor <mjambor@suse.cz>
20340 PR tree-optimization/56294
20341 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
20343 (get_access_replacement): Do not call create_access_replacement.
20344 Assert a replacement exists.
20345 (get_repl_default_def_ssa_name): Create the replacement declaration
20348 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20350 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
20351 final_end_function.
20353 2013-02-28 Marek Polacek <polacek@redhat.com>
20355 PR rtl-optimization/56466
20356 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
20357 if we're changing a loop.
20358 (peel_loops_completely): Likewise.
20360 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
20363 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
20365 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
20368 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
20369 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
20370 INTX_FTYPE_FX, FX_FTYPE_INTX.
20371 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
20373 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
20375 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
20376 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
20377 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
20378 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
20379 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
20380 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
20381 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
20382 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
20383 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
20384 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
20385 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
20386 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
20387 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
20388 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
20389 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
20390 (avrxmega6): Increase max flash segments from 5 to 6.
20391 * config/avr/t-multilib: Regenerate.
20392 * config/avr/avr-tables.opt: Regenerate.
20393 * doc/avr-mmcu.texi: Regenerate.
20395 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
20397 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
20398 (avr_device_to_arch): Rename to avr_device_to_ld.
20399 (avr_device_to_as): New prototype.
20400 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
20401 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
20402 * config/avr/driver-avr.c (avr_device_to_as): New.
20403 (avr_device_to_arch): Rename to avr_device_to_ld.
20405 2013-02-27 Jakub Jelinek <jakub@redhat.com>
20407 PR middle-end/56461
20408 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
20409 method on dr_chain and result_chain.
20411 PR middle-end/56461
20412 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
20413 pointer_set_destroy on not_executed_last_iteration.
20415 PR middle-end/56461
20416 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
20418 PR middle-end/56461
20419 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
20420 FOR_EACH_DEFINED_FUNCTION when freeing state.
20422 PR middle-end/56461
20423 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
20425 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
20428 PR middle-end/56461
20429 * ipa-cp.c (decide_whether_version_node): Call vec_free on
20430 known_aggs[i].items and release known_aggs vector.
20432 PR middle-end/56461
20433 * ipa-reference.c (propagate): Free node_info even for alias nodes.
20435 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20437 * config/microblaze/microblaze.c (microblaze_emit_compare):
20438 Use xor for EQ/NE comparisions.
20439 * config/microblaze/microblaze.md (cstoresf4): Add constraints
20440 (cbranchsf4): Adjust operator to comparison_operator.
20442 2013-02-27 Jakub Jelinek <jakub@redhat.com>
20444 PR middle-end/56461
20445 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
20447 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
20448 vec_safe_push, always update *slot.
20449 (redirect_edge_var_map_clear): Use vec_free.
20450 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
20451 (free_var_map_entry): Use vec_free.
20452 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
20453 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
20455 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
20457 PR middle-end/45472
20458 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
20459 when the may_trap_p bit of the exprs being merged differs.
20460 Reorder tests for speculativeness in the logical and operator.
20462 2013-02-27 Jakub Jelinek <jakub@redhat.com>
20464 * incpath.c (add_standard_paths): Use reconcat instead of concat
20465 where appropriate and avoid leaking memory.
20467 * opts.h: Include obstack.h.
20468 (opts_concat): New prototype.
20469 (opts_obstack): New declaration.
20470 * opts.c (opts_concat): New function.
20471 (opts_obstack): New variable.
20472 (init_options_struct): Call gcc_init_obstack on opts_obstack.
20473 (finish_options): Use opts_concat instead of concat
20474 and XOBNEWVEC instead of XNEWVEC.
20475 * opts-common.c (generate_canonical_option, decode_cmdline_option,
20476 generate_option): Likewise.
20477 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
20478 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
20481 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
20482 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
20484 2013-02-26 Jakub Jelinek <jakub@redhat.com>
20486 PR middle-end/56461
20487 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
20489 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
20491 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
20492 (arm_block_move_unaligned_straight): Likewise.
20493 (arm_adjust_block_mem): Likewise.
20495 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
20498 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
20499 temp, cond and label.
20500 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
20503 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
20504 * config/c6x/c6x.h (dbx_register_map): Update declaration.
20507 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
20508 of prologue/epilogue functions.
20511 * config/tilegx/tilegx.c (tilegx_expand_prologue):
20512 Remove unused variable cfa_offset.
20513 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
20516 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
20517 type promotion to unsigned.
20520 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
20521 for HOST_WIDE_INT of 32 bit / same size as int.
20522 (arm_block_move_unaligned_straight): Likewise.
20523 (arm_adjust_block_mem): Likewise.
20526 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
20529 2013-02-26 Marek Polacek <polacek@redhat.com>
20531 PR tree-optimization/56426
20532 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
20534 2013-02-26 Richard Biener <rguenther@suse.de>
20537 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
20538 unused variable loops.
20540 2013-02-26 Jakub Jelinek <jakub@redhat.com>
20542 PR tree-optimization/56448
20543 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
20544 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
20545 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
20546 later operands of the references, or even first operand for
20547 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
20549 PR tree-optimization/56443
20550 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
20551 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
20552 to type_for_mode langhook.
20554 2013-02-25 Matt Turner <mattst88@gmail.com>
20556 * doc/invoke.texi: Document r4700.
20558 2013-02-25 Richard Biener <rguenther@suse.de>
20560 PR tree-optimization/56175
20561 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
20563 (simplify_bitwise_binary): ... here. Also guard the conversion
20564 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
20566 2013-02-25 Catherine Moore <clm@codesourcery.com>
20569 2013-02-24 Catherine Moore <clm@codesourcery.com>
20570 Maciej W. Rozycki <macro@codesourcery.com>
20571 Tom de Vries <tom@codesourcery.com>
20572 Nathan Sidwell <nathan@codesourcery.com>
20573 Iain Sandoe <iain@codesourcery.com>
20574 Nathan Froyd <froydnj@codesourcery.com>
20575 Chao-ying Fu <fu@mips.com>
20577 * doc/extend.texi: (micromips, nomicromips, nocompression):
20578 Document new function attributes.
20579 * doc/invoke.texi (minterlink-compressed, mmicromips,
20580 m14k, m14ke, m14kec): Document new options.
20581 (minterlink-mips16): Update documentation.
20582 * doc/md.texi (ZC, ZD): Document new constraints.
20583 * configure.ac (gcc_cv_as_micromips): Check if linker
20584 supports the .set micromips directive.
20585 * configure: Regenerate.
20586 * config.in: Regenerate.
20587 * config/mips/mips-tables.opt: Regenerate.
20588 * config/mips/micromips.md: New file.
20589 * constraints.md (ZC, AD): New constraints.
20590 * config/mips/predicates.md (movep_src_register): New predicate.
20591 (movep_src_operand): New predicate.
20592 (non_volatile_mem_operand): New predicate.
20593 * config/mips/mips.md (multimem): New type.
20594 (length): Differentiate between 17-bit and 18-bit branch offsets.
20595 (MOVEP1, MOVEP2): New mode iterator.
20596 (mov_<load>l): Use ZC constraint.
20597 (mov_<load>r): Likewise.
20598 (mov_<store>l): Likewise.
20599 (mov_<store>r): Likewise.
20600 (*branch_equality<mode>_inverted): Add microMIPS support.
20601 (*branch_equality<mode>): Likewise.
20602 (*jump_absolute): Likewise.
20603 (indirect_jump_<mode>): Likewise.
20604 (tablejump_<mode>): Likewise.
20605 (<optab>_internal): Likewise.
20606 (sibcall_internal): Likewise.
20607 (sibcall_value_internal): Likewise.
20608 (prefetch): Use constraint ZD.
20609 * config/mips/mips.opt (minterlink-compressed): New option.
20610 (minterlink-mips16): Now an alias for minterlink-compressed.
20611 (mmicromips): New option.
20612 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
20613 (compare_and_swap_12): Likewise.
20614 (sync_add<mode>): Likewise.
20615 (sync_<optab>_12): Likewise.
20616 (sync_old_<optab>_12): Likewise.
20617 (sync_new_<optab>_12): Likewise.
20618 (sync_nand_12): Likewise.
20619 (sync_old_nand_12): Likewise.
20620 (sync_new_nand_12): Likewise.
20621 (sync_sub<mode>): Likewise.
20622 (sync_old_add<mode>): Likewise.
20623 (sync_old_sub<mode>): Likewise.
20624 (sync_new_add<mode>): Likewise.
20625 (sync_new_sub<mode>): Likewise.
20626 (sync_<optab><mode>): Likewise.
20627 (sync_old_<optab><mode>): Likewise.
20628 (sync_new_<optab><mode>): Likewise.
20629 (sync_nand<mode>): Likewise.
20630 (sync_old_nand<mode>): Likewise.
20631 (sync_new_nand<mode>): Likewise.
20632 (sync_lock_test_and_set<mode>): Likewise.
20633 (test_and_set_12): Likewise.
20634 (atomic_compare_and_swap<mode>): Likewise.
20635 (atomic_exchange<mode>_llsc): Likewise.
20636 (atomic_fetch_add<mode>_llsc): Likewise.
20637 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
20638 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
20639 (umips_save_restore_pattern_p): Likewise.
20640 (umips_load_store_pair_p): Likewise.
20641 (umips_output_load_store_pair): Likewise.
20642 (umips_movep_target_p): Likewise.
20643 (umips_12bit_offset_address_p): Likewise.
20644 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
20645 (mips_base_mips16): Rename this...
20646 (mips_base_compression_flags): ...to this. Update all uses.
20647 (mips_attribute_table): Add micromips, nomicromips and nocompression.
20648 (mips_mips16_decl_p): Delete.
20649 (mips_nomips16_decl_p): Delete.
20650 (mips_get_compress_on_flags): New function.
20651 (mips_get_compress_off_flags): New function.
20652 (mips_get_compress_mode): New function.
20653 (mips_get_compress_on_name): New function.
20654 (mips_get_compress_off_name): New function.
20655 (mips_insert_attributes): Support multiple compression types.
20656 (mips_merge_decl_attributes): Likewise.
20657 (umips_12bit_offset_address_p): New function.
20658 (mips_start_function_definition): Emit .set micromips directive.
20659 (mips_call_may_need_jalx_p): New function.
20660 (mips_function_ok_for_sibcall): Add microMIPS support.
20661 (mips_print_operand_punctuation): Support short delay slots and
20663 (umips_swm_mask, umips_swm_encoding): New.
20664 (umips_build_save_restore): New function.
20665 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
20666 (was_mips16_p): Remove.
20667 (old_compression_mode): New.
20668 (mips_set_compression_mode): New function.
20669 (mips_set_current_function): Add microMIPS support.
20670 (mips_option_override): Likewise.
20671 (umips_save_restore_pattern_p): New function.
20672 (umips_output_save_restore): New function.
20673 (umips_load_store_pair_p_1): New function.
20674 (umips_load_store_pair_p): New function.
20675 (umips_output_load_store_pair_1): New function.
20676 (umips_output_load_store_pair): New function.
20677 (umips_movep_target_p) New function.
20678 (mips_prepare_pch_save): Add microMIPS support.
20679 * config/mips/mips.h (TARGET_COMPRESSION): New.
20680 (TARGET_CPU_CPP_BUILTINS): Update macro
20681 to use new compression flags and to support microMIPS.
20682 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
20683 (MIPS_ARCH_FLOAT_SPEC): Likewise.
20684 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
20685 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
20686 (ASM_SPEC): Support mmicromips and mno-micromips.
20687 (M16STORE_REG_P): New macro.
20688 (MIPS_CALL): Support TARGET_MICROMIPS.
20689 (MICROMIPS_J): New macro.
20690 (mips_base_mips16): Rename this...
20691 (mips_base_compression_flags): ...to this.
20692 (UMIPS_12BIT_OFFSET_P): New macro.
20693 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
20694 (MULTILIB_DIRNAMES): Likewise.
20696 2013-02-25 Tom de Vries <tom@codesourcery.com>
20698 PR rtl-optimization/56131
20699 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
20700 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
20701 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
20703 2013-02-25 Tobias Burnus <burnus@net-b.de>
20705 * doc/invoke.texi (-fsanitize=): Move from optimization
20706 to debugging options.
20708 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
20710 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
20712 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
20713 Alexander Monakov <amonakov@ispras.ru>
20715 PR middle-end/56077
20716 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
20717 flush pending lists also on non-jumps. Adjust comment.
20719 2013-02-24 Catherine Moore <clm@codesourcery.com>
20720 Maciej W. Rozycki <macro@codesourcery.com>
20721 Tom de Vries <tom@codesourcery.com>
20722 Nathan Sidwell <nathan@codesourcery.com>
20723 Iain Sandoe <iain@codesourcery.com>
20724 Nathan Froyd <froydnj@codesourcery.com>
20725 Chao-ying Fu <fu@mips.com>
20727 * doc/extend.texi: (micromips, nomicromips, nocompression):
20728 Document new function attributes.
20729 * doc/invoke.texi (minterlink-compressed, mmicromips,
20730 m14k, m14ke, m14kec): Document new options.
20731 (minterlink-mips16): Update documentation.
20732 * doc/md.texi (ZC, ZD): Document new constraints.
20733 * configure.ac (gcc_cv_as_micromips): Check if linker
20734 supports the .set micromips directive.
20735 * configure: Regenerate.
20736 * config.in: Regenerate.
20737 * config/mips/mips-tables.opt: Regenerate.
20738 * config/mips/micromips.md: New file.
20739 * constraints.md (ZC, AD): New constraints.
20740 * config/mips/predicates.md (movep_src_register): New predicate.
20741 (movep_src_operand): New predicate.
20742 (non_volatile_mem_operand): New predicate.
20743 * config/mips/mips.md (multimem): New type.
20744 (length): Differentiate between 17-bit and 18-bit branch offsets.
20745 (MOVEP1, MOVEP2): New mode iterator.
20746 (mov_<load>l): Use ZC constraint.
20747 (mov_<load>r): Likewise.
20748 (mov_<store>l): Likewise.
20749 (mov_<store>r): Likewise.
20750 (*branch_equality<mode>_inverted): Add microMIPS support.
20751 (*branch_equality<mode>): Likewise.
20752 (*jump_absolute): Likewise.
20753 (indirect_jump_<mode>): Likewise.
20754 (tablejump_<mode>): Likewise.
20755 (<optab>_internal): Likewise.
20756 (sibcall_internal): Likewise.
20757 (sibcall_value_internal): Likewise.
20758 (prefetch): Use constraint ZD.
20759 * config/mips/mips.opt (minterlink-compressed): New option.
20760 (minterlink-mips16): Now an alias for minterlink-compressed.
20761 (mmicromips): New option.
20762 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
20763 (compare_and_swap_12): Likewise.
20764 (sync_add<mode>): Likewise.
20765 (sync_<optab>_12): Likewise.
20766 (sync_old_<optab>_12): Likewise.
20767 (sync_new_<optab>_12): Likewise.
20768 (sync_nand_12): Likewise.
20769 (sync_old_nand_12): Likewise.
20770 (sync_new_nand_12): Likewise.
20771 (sync_sub<mode>): Likewise.
20772 (sync_old_add<mode>): Likewise.
20773 (sync_old_sub<mode>): Likewise.
20774 (sync_new_add<mode>): Likewise.
20775 (sync_new_sub<mode>): Likewise.
20776 (sync_<optab><mode>): Likewise.
20777 (sync_old_<optab><mode>): Likewise.
20778 (sync_new_<optab><mode>): Likewise.
20779 (sync_nand<mode>): Likewise.
20780 (sync_old_nand<mode>): Likewise.
20781 (sync_new_nand<mode>): Likewise.
20782 (sync_lock_test_and_set<mode>): Likewise.
20783 (test_and_set_12): Likewise.
20784 (atomic_compare_and_swap<mode>): Likewise.
20785 (atomic_exchange<mode>_llsc): Likewise.
20786 (atomic_fetch_add<mode>_llsc): Likewise.
20787 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
20788 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
20789 (umips_save_restore_pattern_p): Likewise.
20790 (umips_load_store_pair_p): Likewise.
20791 (umips_output_load_store_pair): Likewise.
20792 (umips_movep_target_p): Likewise.
20793 (umips_12bit_offset_address_p): Likewise.
20794 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
20795 (mips_base_mips16): Rename this...
20796 (mips_base_compression_flags): ...to this. Update all uses.
20797 (mips_attribute_table): Add micromips, nomicromips and nocompression.
20798 (mips_mips16_decl_p): Delete.
20799 (mips_nomips16_decl_p): Delete.
20800 (mips_get_compress_on_flags): New function.
20801 (mips_get_compress_off_flags): New function.
20802 (mips_get_compress_mode): New function.
20803 (mips_get_compress_on_name): New function.
20804 (mips_get_compress_off_name): New function.
20805 (mips_insert_attributes): Support multiple compression types.
20806 (mips_merge_decl_attributes): Likewise.
20807 (umips_12bit_offset_address_p): New function.
20808 (mips_start_function_definition): Emit .set micromips directive.
20809 (mips_call_may_need_jalx_p): New function.
20810 (mips_function_ok_for_sibcall): Add microMIPS support.
20811 (mips_print_operand_punctuation): Support short delay slots and
20813 (umips_swm_mask, umips_swm_encoding): New.
20814 (umips_build_save_restore): New function.
20815 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
20816 (was_mips16_p): Remove.
20817 (old_compression_mode): New.
20818 (mips_set_compression_mode): New function.
20819 (mips_set_current_function): Add microMIPS support.
20820 (mips_option_override): Likewise.
20821 (umips_save_restore_pattern_p): New function.
20822 (umips_output_save_restore): New function.
20823 (umips_load_store_pair_p_1): New function.
20824 (umips_load_store_pair_p): New function.
20825 (umips_output_load_store_pair_1): New function.
20826 (umips_output_load_store_pair): New function.
20827 (umips_movep_target_p) New function.
20828 (mips_prepare_pch_save): Add microMIPS support.
20829 * config/mips/mips.h (TARGET_COMPRESSION): New.
20830 (TARGET_CPU_CPP_BUILTINS): Update macro
20831 to use new compression flags and to support microMIPS.
20832 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
20833 (MIPS_ARCH_FLOAT_SPEC): Likewise.
20834 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
20835 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
20836 (ASM_SPEC): Support mmicromips and mno-micromips.
20837 (M16STORE_REG_P): New macro.
20838 (MIPS_CALL): Support TARGET_MICROMIPS.
20839 (MICROMIPS_J): New macro.
20840 (mips_base_mips16): Rename this...
20841 (mips_base_compression_flags): ...to this.
20842 (UMIPS_12BIT_OFFSET_P): New macro.
20843 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
20844 (MULTILIB_DIRNAMES): Likewise.
20846 2013-02-24 Jakub Jelinek <jakub@redhat.com>
20849 * target-globals.c (save_target_globals): For init_reg_sets and
20850 target_reinit remporarily set this_fn_optabs to this_target_optabs.
20852 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
20854 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
20855 * config/aarch64/t-aarch64
20856 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
20858 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
20860 PR inline-asm/56148
20861 * lra-constraints.c (process_alt_operands): Reload operand
20862 conflicting with earlier clobber only if no more other conflicting
20865 2013-02-22 Jakub Jelinek <jakub@redhat.com>
20868 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
20869 if not linking a shared library.
20871 2013-02-22 Seth LaForge <sethml@google.com>
20873 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
20875 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
20877 * config/arm/arm.md (split for extendsidi): Update condition.
20878 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
20879 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
20880 (qhs_zextenddi_cstr): Likewise.
20882 2013-02-21 Jakub Jelinek <jakub@redhat.com>
20884 PR middle-end/56420
20885 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
20886 avoid signed wrapping.
20887 (expand_mult): Handle properly multiplication by
20888 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
20889 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
20890 in the compiler if coeff is HOST_WIDE_INT_MIN.
20891 (expand_divmod): Don't make ext_op1 static, change it's type to
20892 uhwi. Avoid undefined behavior in -INTVAL (op1).
20894 PR rtl-optimization/50339
20895 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
20897 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
20898 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
20899 into splitting_ashiftrt field.
20900 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
20902 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
20905 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
20907 PR middle-end/56108
20908 * trans-mem.c (execute_tm_mark): Do not expand transactions that
20909 are sure to go irrevocable.
20911 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
20913 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
20914 scalars are valid operands.
20916 2013-02-21 Martin Jambor <mjambor@suse.cz>
20918 PR tree-optimization/56310
20919 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
20920 only matching indices and non-negative final offsets.
20921 (intersect_aggregates_with_edge): Pass src_idx to
20922 agg_replacements_to_vector. Pass src_idx insstead of index to
20923 intersect_with_agg_replacements.
20925 2013-02-21 Martin Jambor <mjambor@suse.cz>
20927 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
20928 instead of hard-wired defaults.
20930 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
20932 * doc/invoke.texi (MIPS Options): Update documentation of the
20933 floating-point multiply-accumulate instruction restrictions.
20935 2013-02-21 Kostya Serebryany <kcc@google.com>
20937 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
20938 asan_shadow_offset on x86_64 linux.
20940 2013-02-21 Richard Biener <rguenther@suse.de>
20942 PR tree-optimization/56415
20944 2013-02-11 Richard Biener <rguenther@suse.de>
20946 PR tree-optimization/56273
20947 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
20950 2013-02-21 Jakub Jelinek <jakub@redhat.com>
20953 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
20956 PR inline-asm/56405
20957 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
20958 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
20960 2013-02-20 Jan Hubicka <jh@suse.cz>
20962 PR tree-optimization/56265
20963 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
20964 when target is referenced for first time.
20966 2013-02-20 Richard Biener <rguenther@suse.de>
20968 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
20969 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
20970 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
20971 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
20972 not return anything.
20973 (rename_ssa_copies): Do not remove unused locals.
20974 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
20975 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
20976 * passes.c (execute_function_todo): Do not schedule unused locals
20977 removal if cleanup_tree_cfg did something.
20978 * tree-ssa-live.c (remove_unused_locals): Dump statistics
20979 about the number of removed locals.
20981 2013-02-20 Richard Biener <rguenther@suse.de>
20983 PR tree-optimization/56398
20984 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
20986 2013-02-20 Martin Jambor <mjambor@suse.cz>
20988 PR tree-optimization/55334
20989 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
20990 restricted pointers to arrays.
20992 2013-02-20 Richard Biener <rguenther@suse.de>
20993 Jakub Jelinek <jakub@redhat.com>
20995 PR tree-optimization/56396
20996 * tree-ssa-ccp.c (n_const_val): New static variable.
20997 (get_value): Return NULL for SSA names we don't have a lattice
20999 (ccp_initialize): Initialize n_const_val.
21000 * tree-ssa-copy.c (n_copy_of): New static variable.
21001 (init_copy_prop): Initialize n_copy_of.
21002 (get_value): Return NULL_TREE for SSA names we don't have a
21005 2013-02-20 Martin Jambor <mjambor@suse.cz>
21007 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
21009 2013-02-20 Richard Biener <rguenther@suse.de>
21011 * genpreds.c (write_lookup_constraint): Do not compare first
21012 letter of the constraint again.
21014 2013-02-20 Richard Biener <rguenther@suse.de>
21016 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
21018 (get_use_iv_cost): Terminate hashtable walk when coming across
21021 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
21023 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
21024 reassociation for avx2 targets.
21026 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
21028 * config/microblaze/microblaze.c: microblaze_has_clz = 0
21029 Add version check for v8.10.a to enable microblaze_has_clz
21030 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
21031 version and TARGET_PATTERN_COMPARE check
21032 * config/microblaze/microblaze.md: New clzsi2 instruction
21034 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
21036 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
21037 function before branching.
21039 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
21041 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
21043 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
21045 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
21047 PR middle-end/55889
21048 * sel-sched.c: Include ira.h.
21049 (implicit_clobber_conflict_p): New function.
21050 (moveup_expr): Use it.
21051 * Makefile.in (sel-sched.o): Depend on ira.h.
21053 2013-02-19 Richard Biener <rguenther@suse.de>
21055 PR tree-optimization/56384
21056 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
21057 (vn_hash_type): Split out from ...
21058 (vn_hash_constant_with_type): ... here.
21059 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
21060 (vn_phi_eq): Compare types from vn_phi_s structure.
21061 (vn_phi_lookup): Populate vn_phi_s type.
21062 (vn_phi_insert): Likewise.
21064 2013-02-19 Jakub Jelinek <jakub@redhat.com>
21066 PR tree-optimization/56350
21067 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
21068 if haven't found reduction or nested cycle operand, rather than
21069 asserting we must find it.
21071 PR tree-optimization/56381
21072 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
21075 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
21076 Jakub Jelinek <jakub@redhat.com>
21079 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
21080 (swap_optab_enable): Same.
21081 (init_all_optabs): Use argument instead of global.
21082 * tree.h (struct tree_optimization_option): New field target_optabs.
21083 * expr.h (init_all_optabs): Add argument to prototype.
21084 (TREE_OPTIMIZATION_OPTABS): New.
21085 (save_optabs_if_changed): Protoize.
21086 * optabs.h: Declare this_fn_optabs.
21087 * optabs.c (save_optabs_if_changed): New.
21088 Declare this_fn_optabs.
21089 (init_optabs): Add argument to init_all_optabs() call.
21090 * function.c (invoke_set_current_function_hook): Handle per
21092 * function.h (struct function): New field optabs.
21093 * config/mips/mips.c (mips_set_mips16_mode): Handle when
21094 optimization_current_node has changed.
21095 * target-globals.h (save_target_globals_default_opts): Protoize.
21096 * target-globals.c (save_target_globals_default_opts): New.
21098 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21101 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
21102 registers %fr12 and %fr12R as call used.
21105 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
21106 and HImode, require all displacements to be an integer multiple of
21108 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
21109 only allow QImode and HImode when reload is in progress and strict is
21110 true. Likewise for symbolic addresses. Use base14_operand to check
21111 displacements in REG+BASE addresses.
21113 2013-02-18 Richard Biener <rguenther@suse.de>
21115 PR tree-optimization/56366
21116 * tree-vect-loop.c (get_initial_def_for_induction): Properly
21117 handle sign-conversion of outer-loop initial induction value.
21119 2013-02-18 Richard Biener <rguenther@suse.de>
21121 PR middle-end/56349
21122 * cfghooks.c (merge_blocks): If we merge a latch into another
21123 block adjust references to it.
21124 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
21125 (verify_loop_structure): Verify that a recorded latch is in fact
21128 2013-02-18 Richard Biener <rguenther@suse.de>
21130 PR tree-optimization/56321
21131 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
21132 order SSA name release and virtual operand unlinking.
21134 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
21136 * config/microblaze/microblaze.md (save_stack_block): Define.
21137 (restore_stack_block): Likewise.
21139 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
21141 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
21142 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
21143 * config/microblaze/microblaze.c (microblaze_option_override):
21144 Bail out early for PIC modes when target does not support PIC.
21146 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
21148 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
21149 Replace with a microblaze version.
21150 (microblaze_trampoline_init): Adapt for microblaze.
21151 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
21154 2013-02-16 Jakub Jelinek <jakub@redhat.com>
21155 Dodji Seketeli <dodji@redhat.com>
21158 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
21159 (instrument_mem_region_access): Do not forget to always put
21160 instrumentation of the of 'base' and 'base + len' in a "if (len !=
21161 0) statement, even for cases where either 'base' or 'base + len'
21162 are not instrumented -- because they have been previously
21163 instrumented. Simplify the logic by putting all the statements
21164 instrument 'base + len' inside a sequence, and then insert that
21165 sequence right before the current insertion point. Then, to
21166 instrument 'base + len', just get an iterator on that statement.
21167 And do not forget to update the pointer to iterator the function
21168 received as argument.
21170 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
21172 PR rtl-optimization/56348
21173 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
21175 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
21177 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
21178 (clean_graph_dump_file): Pass base to start_graph_dump.
21180 2013-02-14 Richard Henderson <rth@redhat.com>
21183 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
21185 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
21187 * collect2-aix.h: Define F_LOADONLY.
21189 2013-02-14 Richard Biener <rguenther@suse.de>
21192 * varasm.c (output_constant_def_1): Get the decl representing
21193 the constant as argument.
21194 (output_constant_def): Wrap output_constant_def_1.
21195 (make_decl_rtl): Use output_constant_def_1 with the decl
21196 representing the constant.
21197 (build_constant_desc): Optionally re-use a decl already
21198 representing the constant.
21199 (tree_output_constant_def): Adjust.
21201 2013-02-14 Dodji Seketeli <dodji@redhat.com>
21204 * asan.c (instrument_builtin_call): Really put the length of the
21205 second source argument into src1_len.
21207 2013-02-13 Jakub Jelinek <jakub@redhat.com>
21209 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
21210 argument. If it is false, don't create edge from then_bb to
21212 (insert_if_then_before_iter): Pass true to it.
21213 (build_check_stmt): Pass false to it.
21214 (transform_statements): Flush hash table only on extended basic
21215 block boundaries, rather than at the beginning of every bb.
21216 Don't flush hash table on nonfreeing_call_p calls.
21217 * tree-flow.h (nonfreeing_call_p): New prototype.
21218 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
21220 2013-02-13 David S. Miller <davem@davemloft.net>
21222 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
21224 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
21227 * ira.c (max_regno_before_ira): Move from ...
21229 (fix_reg_equiv_init): Use max_regno_before_ira instead of
21232 2013-02-13 Jakub Jelinek <jakub@redhat.com>
21234 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
21236 2013-02-13 Richard Biener <rguenther@suse.de>
21239 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
21240 globals in MEM_REFs.
21242 2013-02-13 Richard Biener <rguenther@suse.de>
21244 * loop-init.c (loop_optimizer_init): Clear loop state when
21245 re-initializing preserved loops.
21246 * loop-unswitch.c (unswitch_single_loop): Return whether
21247 we unswitched the loop. Do not verify loop state here.
21248 (unswitch_loops): When we unswitched a loop discover new loops.
21250 2013-02-13 Kostya Serebryany <kcc@google.com>
21252 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
21254 * sanitizer.def: Rename __asan_init to __asan_init_v1.
21256 2013-02-12 Dodji Seketeli <dodji@redhat.com>
21258 Avoid instrumenting duplicated memory access in the same basic block
21259 * Makefile.in (asan.o): Add new dependency on hash-table.h
21260 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
21261 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
21262 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
21263 (free_mem_ref_resources, has_mem_ref_been_instrumented)
21264 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
21265 (get_mem_ref_of_assignment): New functions.
21266 (get_mem_refs_of_builtin_call): Extract from
21267 instrument_builtin_call and tweak a little bit to make it fit with
21269 (instrument_builtin_call): Use the new
21270 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
21271 of is_gimple_builtin_call.
21272 (instrument_derefs, instrument_mem_region_access): Insert the
21273 instrumented memory reference into the hash table.
21274 (maybe_instrument_assignment): Renamed instrument_assignment into
21275 this, and change it to advance the iterator when instrumentation
21276 actually happened and return true in that case. This makes it
21277 homogeneous with maybe_instrument_assignment, and thus give a
21278 chance to callers to be more 'regular'.
21279 (transform_statements): Clear the memory reference hash table
21280 whenever we enter a new BB, when we cross a function call, or when
21281 we are done transforming statements. Use
21282 maybe_instrument_assignment instead of instrumentation. No more
21283 need to special case maybe_instrument_assignment and advance the
21284 iterator after calling it; it's now handled just like
21285 maybe_instrument_call. Update comment.
21287 2013-02-13 Richard Biener <rguenther@suse.de>
21289 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
21290 Fix loop discovery code.
21292 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
21294 PR inline-asm/56148
21295 * lra-constraints.c (process_alt_operands): Match early clobber
21296 operand with itself. Check conflicts with earlyclobber only if
21297 the operand is not reloaded. Prefer to reload conflicting operand
21298 if earlyclobber and matching operands are the same.
21300 2013-02-12 Richard Biener <rguenther@suse.de>
21303 * lto-streamer-out.c (write_symbol): Do not output symbols
21304 for hard register variables.
21306 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
21309 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
21310 (umulsidi3_insn, mulsidi3_insn): New insns.
21312 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
21314 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
21315 (struct tune_params): Add vec_costs field.
21316 * config/arm/arm.c (arm_builtin_vectorization_cost)
21317 (arm_add_stmt_cost): New functions.
21318 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
21319 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
21320 (arm_default_vec_cost): New struct of type cpu_vec_costs.
21321 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
21322 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
21323 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
21324 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
21326 2013-02-12 Richard Biener <rguenther@suse.de>
21329 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
21330 decls again if possible.
21332 2013-02-12 Richard Biener <rguenther@suse.de>
21334 PR middle-end/56288
21335 * tree-ssa.c (verify_ssa_name): Fix check, move
21336 SSA_NAME_IN_FREE_LIST check up.
21338 2013-02-12 Jakub Jelinek <jakub@redhat.com>
21339 Steven Bosscher <steven@gcc.gnu.org>
21341 PR rtl-optimization/56151
21342 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
21343 equal to op0 or op1, and last_insn pattern is CODE operation
21344 with MEM dest and one of the operands matches that MEM.
21346 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
21348 * doc/extend.texi: Document Function Multiversioning and "default"
21349 parameter string to target attribute.
21350 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
21351 target attribute parameter is "default".
21352 (ix86_compare_version_priority): Remove checks for target attribute.
21353 (ix86_mangle_function_version_assembler_name): Change error to sorry.
21354 Remove check for target attribute equal to NULL. Add assert.
21355 (ix86_generate_version_dispatcher_body): Change error to sorry.
21357 2013-02-11 Iain Sandoe <iain@codesourcery.com>
21358 Jack Howarth <howarth@bromo.med.uc.edu>
21359 Patrick Marlier <patrick.marlier@gmail.com>
21362 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
21363 define ENDFILE_SPEC as TM_DESTRUCTOR.
21364 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
21366 2013-02-11 Alexander Potapenko <glider@google.com>
21367 Jack Howarth <howarth@bromo.med.uc.edu>
21368 Jakub Jelinek <jakub@redhat.com>
21371 * config/darwin.c (cdtor_record): Rename ctor_record.
21372 (sort_cdtor_records): Rename sort_ctor_records.
21373 (finalize_dtors): New routine to sort destructors by
21374 priority before use in assemble_integer.
21375 (machopic_asm_out_destructor): Use finalize_dtors if needed.
21377 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
21379 PR rtl-optimization/56275
21380 * simplify-rtx.c (avoid_constant_pool_reference): Check that
21381 offset is non-negative and less than cmode size before
21382 calling simplify_subreg.
21384 2013-02-11 Richard Biener <rguenther@suse.de>
21386 PR tree-optimization/56264
21387 * cfgloop.h (fix_loop_structure): Adjust prototype.
21388 * loop-init.c (fix_loop_structure): Return the number of
21389 newly discovered loops.
21390 * tree-cfgcleanup.c (repair_loop_structures): When new loops
21391 are discovered, do a full loop-closed SSA rewrite.
21393 2013-02-11 Richard Biener <rguenther@suse.de>
21395 PR tree-optimization/56273
21396 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
21398 (check_array_ref): Fix missing newline in dumps.
21399 (search_for_addr_array): Likewise.
21401 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
21403 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
21405 2013-02-09 Jakub Jelinek <jakub@redhat.com>
21408 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
21410 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
21412 PR rtl-optimization/56246
21413 * lra-constraints.c (simplify_operand_subreg): Try to reuse
21415 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
21416 constraints are satisfied.
21418 2013-02-08 Jeff Law <law@redhat.com>
21421 * emit-rtl.c (reg_is_parm_p): New function.
21422 * regs.h (reg_is_parm_p): New prototype.
21423 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
21424 callee-clobbered registers.
21426 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
21429 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
21430 If there is no implicit builtin declaration, just return NULL.
21432 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
21434 * config/i386/sse.md (FMAMODEM): New mode iterator.
21435 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
21436 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
21438 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
21440 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
21441 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
21442 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
21444 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
21446 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
21447 (microblaze*-*-elf): Likewise.
21448 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
21450 * config/microblaze/microblaze-c.c: Add builtin defines for
21451 _LITTLE_ENDIAN and _BIG_ENDIAN.
21452 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
21453 add to TARGET_DEFAULT flags.
21454 Expand ASM_SPEC and LINK_SPEC.
21455 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
21456 * config/microblaze/microblaze.md: Update extendsidi2 and
21457 movdi_internal instructions to use low-order / high-order reg
21459 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
21460 options and inversemask / mask of LITTLE_ENDIAN.
21461 * config/microblaze/t-microblaze: Expand multilib options to
21462 include mlittle-endian (le) and update exceptions patterns.
21464 2013-02-08 Jakub Jelinek <jakub@redhat.com>
21466 PR rtl-optimization/56195
21467 * lra-constraints.c (get_reload_reg): Don't reuse regs
21468 if they have smaller mode than requested, if they have
21469 wider mode than requested, try to return a SUBREG.
21471 PR tree-optimization/56250
21472 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
21473 if type is unsigned and code isn't MULT_EXPR.
21475 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
21477 PR tree-optimization/56064
21478 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
21479 bits according to mode.
21480 * fixed-value.h (fixed_from_double_int)
21481 (const_fixed_from_double_int): Adjust comments.
21483 2013-02-08 Richard Biener <rguenther@suse.de>
21486 * lto-streamer.h (struct data_in): Remove current_file, current_line
21487 and current_col members.
21488 * lto-streamer-out.c (lto_output_location): Stream changed bits
21489 en-block for efficiency.
21490 * lto-streamer-in.c (clear_line_info): Remove.
21491 (lto_input_location): Cache current file, line and column
21492 globally via local statics. Read changed bits en-block.
21493 (input_function): Do not call clear_line_info.
21494 (lto_read_body): Likewise.
21495 (lto_input_toplevel_asms): Likewise.
21497 2013-02-08 Michael Matz <matz@suse.de>
21499 PR tree-optimization/52448
21500 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
21501 (nt_call_phase): New static.
21502 (add_or_mark_expr): Only mark accesses with newer phase than any
21504 (nonfreeing_call_p): New.
21505 (nt_init_block): Update nt_call_phase, mark blocks as visited.
21506 (nt_fini_block): Keep blocks marked as visited.
21507 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
21509 2013-02-08 Richard Biener <rguenther@suse.de>
21511 * ira.c (ira): Free broken dominator information.
21513 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
21515 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
21517 2013-02-08 Marek Polacek <polacek@redhat.com>
21519 * cfgloop.c (verify_loop_structure): Add more checking of headers.
21521 2013-02-08 Richard Biener <rguenther@suse.de>
21523 PR middle-end/56181
21524 * cfgloop.h (flow_loops_find): Adjust.
21525 (bb_loop_header_p): Declare.
21526 * cfgloop.c (bb_loop_header_p): New function split out from ...
21527 (flow_loops_find): ... here. Adjust function signature,
21528 support incremental loop structure update.
21529 (verify_loop_structure): Cleanup. Verify a loop is a loop.
21530 * cfgloopmanip.c (fix_loop_structure): Move ...
21531 * loop-init.c (fix_loop_structure): ... here.
21532 (apply_loop_flags): Split out from ...
21533 (loop_optimizer_init): ... here.
21534 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
21535 in incremental mode, only remove dead loops here.
21537 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
21540 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
21541 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
21542 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
21543 (*round<mode>3.libgcc): New insns for fixed-modes.
21544 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
21545 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
21546 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
21547 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
21548 implementations. Define to __builtin_avr_absFX,
21549 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
21550 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
21551 __builtin_avr_countlsFX, respectively.
21552 * config/avr/avr-c.c (target.h): Include it.
21553 (enum avr_builtin_id): New enum.
21554 (avr_resolve_overloaded_builtin): New static function.
21555 (avr_register_target_pragmas): Use it to set
21556 targetm.resolve_overloaded_builtin.
21557 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
21558 tree nodes used by DEF_BUILTIN.
21559 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
21560 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
21561 <AVR_BUILTIN_xxBITS>: Same.
21563 2013-02-08 Richard Biener <rguenther@suse.de>
21565 * cfgloop.c (verify_loop_structure): Properly handle
21566 a loop exiting to another loop header.
21567 * ira-int.h (ira_loops): Remove.
21568 * ira.c (ira_loops): Remove.
21569 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
21570 (do_reload): Use loop_optimizer_finalize.
21571 * ira-build.c (create_loop_tree_nodes): Use get_loops and
21572 number_of_loops to access the loop tree.
21573 (more_one_region_p): Likewise.
21574 (finish_loop_tree_nodes): Likewise.
21575 (rebuild_regno_allocno_maps): Likewise.
21576 (mark_loops_for_removal): Likewise.
21577 (mark_all_loops_for_removal): Likewise.
21578 (remove_unnecessary_regions): Likewise.
21579 (ira_build): Likewise.
21580 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
21582 2013-02-08 Richard Biener <rguenther@suse.de>
21584 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
21585 * ipa-pure-const.c (analyze_function): Avoid calling
21586 mark_irreducible_loops twice.
21587 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
21589 2013-02-07 David S. Miller <davem@davemloft.net>
21591 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
21593 * var-tracking.c (vt_add_function_parameter): Test the presence of
21594 HAVE_window_save properly and do not remap argument registers when
21595 we have a leaf function.
21597 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
21600 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
21602 * config/i386/i386.c (ix86_print_operand): Ditto.
21604 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
21606 * lra-constraints.c (process_alt_operands): Fix recently added comment.
21608 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
21610 PR rtl-optimization/56225
21611 * lra-constraints.c (process_alt_operands): Check that reload hard
21612 reg can hold value for strict_low_part.
21614 2013-02-07 Jakub Jelinek <jakub@redhat.com>
21617 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
21618 dwarf2out_end_function.
21619 (in_first_function_p, maybe_at_text_label_p,
21620 first_loclabel_num_not_at_text_label): New variables.
21621 (dwarf2out_var_location): In the first function find out
21622 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
21623 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
21626 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
21628 PR rtl-optimization/56178
21629 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
21630 SUBREG of a register. Tidy up related block of code.
21631 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
21632 note if the source is a register or a SUBREG of a register.
21634 2013-02-07 Jakub Jelinek <jakub@redhat.com>
21637 * config/rs6000/rs6000.md (ptrm): New mode attr.
21638 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
21639 call_value_indirect_aix<pttrsize>,
21640 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
21643 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
21645 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
21646 if -bnortl. Convert to strcmp and strncmp.
21648 2013-02-07 Alan Modra <amodra@gmail.com>
21651 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
21652 addresses won't wrap when offsetting.
21653 (rs6000_secondary_reload): Provide secondary reloads needed for
21654 wrapping LO_SUM addresses.
21656 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
21658 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
21659 MACH, just __MACH__.
21661 2013-02-06 Richard Biener <rguenther@suse.de>
21663 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
21664 instead of calling fix_loop_structure.
21666 2013-02-06 Jakub Jelinek <jakub@redhat.com>
21668 PR middle-end/56217
21669 * omp-low.c (use_pointer_for_field): Return false if
21670 lower_send_shared_vars doesn't generate any copy-out code.
21672 2013-02-06 Tom de Vries <tom@codesourcery.com>
21674 PR rtl-optimization/56131
21675 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
21676 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
21677 of the label is NULL. Add comment.
21679 2013-02-05 Jakub Jelinek <jakub@redhat.com>
21681 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
21684 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
21685 (STATIC_LIBTSAN_LIBS): Likewise.
21686 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
21687 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
21688 is defined, don't add anything else beyond that.
21689 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
21690 (LINK_COMMAND_SPEC): Use them.
21692 PR tree-optimization/56205
21693 * tree-stdarg.c (check_all_va_list_escapes): Return true if
21694 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
21695 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
21697 2013-02-05 Richard Biener <rguenther@suse.de>
21699 PR tree-optimization/53342
21700 PR tree-optimization/53185
21701 * tree-vectorizer.h (vect_check_strided_load): Remove.
21702 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
21703 not disallow peeling for vectorized strided loads.
21704 (vect_check_strided_load): Make static and simplify.
21705 (vect_analyze_data_refs): Adjust.
21706 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
21707 correctly when vectorizing strided loads.
21709 2013-02-05 Richard Biener <rguenther@suse.de>
21711 * doc/install.texi: Refer to ISL, not PPL.
21713 2013-02-05 Jan Hubicka <jh@suse.cz>
21715 PR tree-optimization/55789
21716 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
21718 2013-02-05 Jan Hubicka <jh@suse.cz>
21720 PR tree-optimization/55789
21721 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
21722 the dead call anyway.
21724 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
21727 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
21729 2013-02-04 Alexander Potapenko <glider@google.com>
21730 Jack Howarth <howarth@bromo.med.uc.edu>
21731 Jakub Jelinek <jakub@redhat.com>
21734 * config/darwin.c (sort_ctor_records): Stabilized qsort
21735 on constructor priority by using original position.
21736 (finalize_ctors): New routine to sort constructors by
21737 priority before use in assemble_integer.
21738 (machopic_asm_out_constructor): Use finalize_ctors if needed.
21740 2013-02-04 Jakub Jelinek <jakub@redhat.com>
21743 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
21744 about visibility on artificial decls.
21745 * config/sol2.c (solaris_assemble_visibility): Likewise.
21747 2013-02-04 Kai Tietz <ktietz@redhat.com>
21750 * config/i386/i386.c (function_value_ms_64): Add additional valtype
21751 argument and improve checking of return-argument types for 16-byte
21753 (ix86_function_value_1): Add additional valtype argument on call
21754 of function_value_64.
21755 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
21756 handling infunction_value_64 function.
21758 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
21760 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
21762 2013-02-04 Richard Biener <rguenther@suse.de>
21764 PR tree-optimization/56188
21765 * tree-ssa-structalias.c (label_visit): Consider case with
21766 initially non-empty points-to set.
21767 (perform_var_substitution): Dump node mapping and clean up.
21769 2013-02-04 Richard Guenther <rguenther@suse.de>
21772 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
21773 node prevail as last resort.
21774 (lto_symtab_merge_decls): Remove guard on LTRANS here.
21775 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
21777 2013-02-04 Richard Biener <rguenther@suse.de>
21779 PR tree-optimization/56113
21780 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
21782 (equiv_class_lookup_or_add): ... this.
21783 (label_visit): Adjust and fix error in previous patch.
21784 (perform_var_substitution): Adjust.
21786 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
21788 * config/sh/divtab.c: Fix formatting and comments throughout the file.
21789 * config/sh/sh4-300.md: Likewise.
21790 * config/sh/sh4a.md: Likewise.
21791 * config/sh/constraints.md: Likewise.
21792 * config/sh/sh.md: Likewise.
21793 * config/sh/netbsd-elf.h: Likewise.
21794 * config/sh/predicates.md: Likewise.
21795 * config/sh/sh-protos.h: Likewise.
21796 * config/sh/ushmedia.h: Likewise.
21797 * config/sh/linux.h: Likewise.
21798 * config/sh/sh.c: Likewise.
21799 * config/sh/superh.h: Likewise.
21800 * config/sh/elf.h: Likewise.
21801 * config/sh/sh4.md: Likewise.
21802 * config/sh/sh.h: Likewise.
21804 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21806 * config/pa/constraints.md: Adjust unused letters. Change "T"
21807 constraint to match_test floating_point_store_memory_operand().
21808 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
21809 (base14_operand): New.
21810 (floating_point_store_memory_operand): New.
21811 (integer_store_memory_operand): Revise to use base14_operand and
21812 reg_plus_base_memory_operand.
21813 (move_dest_operand): Allow symbolic_memory_operands.
21814 (symbolic_memory_operand): Check for LO_SOM.
21815 (symbolic_operand): Change default case to break.
21816 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
21817 CONST_DOUBLE values to be reloaded by putting them into memory when
21818 the destination is a floating point register.
21819 (movdf): Remove code to handle CONST_DOUBLE.
21821 (reload_indf_r1): New.
21822 (reload_insf_r1): New.
21823 Consistently use "Q" and "T" constraints with integer and floating
21824 point move instructions, respectively.
21825 (movdi): Remove FAIL.
21826 Change predicate for source operand unamed DImode move from
21827 general_operand to move_src_operand.
21828 (umulsidi3): Change predicate for destination operand to
21830 Likewise for similar unamed patterns.
21831 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
21832 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
21833 (hppa_legitimize_address): Simplify mask calculation.
21834 (pa_emit_move_sequence): Revised handling of secondary reloads from
21835 REG+D addresses for floating point loads and stores. Directly handle
21836 loading CONST0_RTX (mode) to a floating point register.
21837 (pa_secondary_reload): Handle reloading DF and SFmode constant values
21838 to floating point registers. Don't restrict secondary reloads to
21839 floating point registers to integer modes. Revise some comments and
21841 (TARGET_LEGITIMATE_ADDRESS_P): Define.
21842 (pa_legitimate_address_p): New.
21843 (pa_legitimize_reload_address): New.
21844 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
21845 (STRICT_REG_OK_FOR_BASE_P): New.
21846 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
21847 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
21849 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
21850 Andrew Dixie <andrewd@gentrack.com>
21852 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
21855 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
21857 * expmed.c (extract_bit_field_1): Pass the full width of the
21858 structure to get_best_reg_extraction_insn.
21860 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
21863 * configure.ac (use_cxa_atexit): Add AIX.
21864 * configure: Regenerate.
21866 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
21868 2013-02-01 Jakub Jelinek <jakub@redhat.com>
21871 * final.c (need_profile_function): New variable.
21872 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
21873 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
21874 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
21875 notes, targetm.asm_out.function_prologue doesn't emit anything,
21876 HAVE_prologue and profiler should be emitted before prologue,
21877 set need_profile_function instead of emitting it.
21878 (final_scan_insn): If need_profile_function, emit
21879 profile_function on the first NOTE_INSN_BASIC_BLOCK or
21880 NOTE_INSN_FUNCTION_BEG note.
21882 2013-02-01 Richard Henderson <rth@redhat.com>
21884 * config/rs6000/rs6000.md (smulditi3): New.
21887 * config/alpha/alpha.md (umulditi3): New.
21889 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
21891 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
21892 (ASM_OUTPUT_ALIGNED_LOCAL): New.
21894 2013-02-01 Richard Biener <rguenther@suse.de>
21896 PR tree-optimization/56113
21897 * tree-ssa-structalias.c (label_visit): Reduce work for
21898 single-predecessor nodes.
21900 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
21902 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
21903 range isn't testing for zero.
21905 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
21907 PR middle-end/56113
21908 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
21910 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
21911 Nick Clifton <nickc@redhat.com>
21913 * config/v850/constraints.md (Q): Define as a memory constraint.
21914 * config/v850/predicates.md (label_ref_operand): New predicate.
21915 (e3v5_shift_operand): New predicate.
21916 (ior_operator): New predicate.
21917 * config/v850/t-v850: Add e3v5 multilib.
21918 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
21919 (v850_gen_movdi): Prototype.
21920 * config/v850/v850.c: Add support for e3v5 architecture.
21921 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
21923 (construct_save_jarl): Add e3v5 long JARL support.
21924 (v850_adjust_insn_length): New function. Adjust length of call
21925 insns when using e3v5 instructions.
21926 (v850_gen_movdi): New function: Generate instructions to move a
21928 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
21929 (CPP_SPEC): Define __v850e3v5__ as appropriate.
21930 (TARGET_USE_FPU): Enable for e3v5.
21931 (CONST_OK_FOR_W): New macro.
21932 (ADJUST_INSN_LENGTH): Define.
21933 * config/v850/v850.md (UNSPEC_LOOP): Define.
21934 (attr cpu): Add v850e3v5.
21935 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
21936 (movdi): New pattern.
21937 (movdi_internal): New pattern.
21938 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
21939 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
21940 (cstoresf4): Likewise.
21941 (cstoredf4): Likewise.
21942 (insv): New pattern.
21943 (rotlso3_a): New pattern.
21944 (rotlsi3_b): New pattern
21945 (rotlsi3_v850e3v5): New pattern.
21946 (doloop_begin): New pattern.
21947 (fix_loop_counter): New pattern.
21948 (doloop_end): New pattern.
21949 (branch_normal): Add e3v5 long branch support.
21950 (branch_invert): Likewise.
21951 (branch_z_normal): Likewise.
21952 (branch_z_invert): Likewise.
21953 (branch_nz_normal): Likewise.
21954 (branch_nz_invert): Likewise.
21955 (call_internal_short): Add e3v5 register-indirect JARL support.
21956 (call_internal_long): Likewise.
21957 (call_value_internal_short): Likewise.
21958 (call_value_internal_long): Likewise.
21959 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
21960 (mloop): New option.
21961 * config.gcc: Add support for configuring v840e3v5 target.
21962 * doc/invoke.texi: Document new v850 specific command line options.
21964 2013-01-31 Paul Koning <ni1d@arrl.net>
21968 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
21969 children if parent is a class.
21970 (prune_unused_types_prune): Don't add DW_AT_declaration.
21972 2013-01-31 Richard Biener <rguenther@suse.de>
21974 PR tree-optimization/56157
21975 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
21976 match up operand with SLP child.
21978 2013-01-31 Jason Merrill <jason@redhat.com>
21981 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
21982 parameters the first time.
21983 (gen_scheduled_generic_parms_dies): Check completeness here.
21985 2013-01-31 Richard Biener <rguenther@suse.de>
21987 PR middle-end/53073
21988 * common.opt (faggressive-loop-optimizations): New flag,
21989 enabled by default.
21990 * doc/invoke.texi (faggressive-loop-optimizations): Document.
21991 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
21992 infer_loop_bounds_from_undefined by it.
21994 2013-01-31 Richard Biener <rguenther@suse.de>
21996 PR tree-optimization/56150
21997 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
21998 visit virtual operands.
21999 (find_uses_to_rename_bb): Likewise.
22001 2013-01-31 Richard Biener <rguenther@suse.de>
22003 PR tree-optimization/56150
22004 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
22005 mixed store non-store stmts.
22007 2013-01-30 Jakub Jelinek <jakub@redhat.com>
22010 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
22011 LIBASAN_EARLY_SPEC is defined.
22012 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
22013 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
22015 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
22018 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
22019 invalid args instead of ICEing on it.
22020 (ix86_valid_target_attribute_tree): Return error_mark_node if
22021 ix86_valid_target_attribute_inner_p failed.
22022 (ix86_valid_target_attribute_p): Return false only if
22023 ix86_valid_target_attribute_tree returned error_mark_node. Allow
22024 target("default") attribute.
22025 (sorted_attr_string): Change argument from const char * to tree,
22026 merge in all target attribute arguments rather than just one.
22027 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
22028 instead of free. Avoid using strcat.
22029 (ix86_mangle_function_version_assembler_name): Mangle
22030 target("default") as if no target attribute is present. Adjust
22031 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
22032 instead of xmalloc and XDELETEVEC instead of free.
22033 (ix86_function_versions): Don't return true if one of the decls
22034 doesn't have target attribute. If they don't and one of the decls
22035 is DECL_FUNCTION_VERSIONED, report an error. Adjust
22036 sorted_attr_string caller. Use XDELETEVEC instead of free.
22037 (ix86_supports_function_versions): Remove.
22038 (make_name): Fix up formatting.
22039 (make_dispatcher_decl): Remove resolver_name and its initialization.
22040 Avoid leaking memory.
22041 (is_function_default_version): Return true if there is
22042 target("default") attribute rather than no target attribute at all.
22043 (make_resolver_func): Avoid leaking memory.
22044 (ix86_generate_version_dispatcher_body): Likewise.
22045 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
22046 * target.def (supports_function_versions): Remove.
22047 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
22048 * doc/tm.texi: Regenerated.
22050 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
22052 PR rtl-optimization/56144
22053 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
22054 for values with side effects.
22056 2013-01-30 Richard Biener <rguenther@suse.de>
22058 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
22059 (sparseset_pop): Likewise.
22060 * cfganal.c (compute_idf): Likewise. Increase work-stack size
22061 to be able to use quick_push in the worker loop.
22063 2013-01-30 Marek Polacek <polacek@redhat.com>
22065 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
22067 2013-01-30 Richard Biener <rguenther@suse.de>
22070 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
22072 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
22074 PR tree-optimization/56064
22075 * fixed-value.c (fixed_from_double_int): New function.
22076 * fixed-value.h (fixed_from_double_int): New prototype.
22077 (const_fixed_from_double_int): New static inline function.
22078 * fold-const.c (native_interpret_fixed): New static function.
22079 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
22080 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
22081 (native_encode_fixed): New static function.
22082 (native_encode_expr) <FIXED_CST>: Use it.
22083 (native_interpret_int): Move double_int worker code to...
22084 * double-int.c (double_int::from_buffer): ...this new static method.
22085 * double-int.h (double_int::from_buffer): Prototype it.
22087 2013-01-30 Richard Biener <rguenther@suse.de>
22089 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
22090 New pointer-map and obstack.
22091 (init_alias_vars): Allocate pointer-map and obstack.
22092 (delete_points_to_sets): Free them.
22093 (find_what_var_points_to): Cache result.
22094 (find_what_p_points_to): Adjust for changed interface of
22095 find_what_var_points_to.
22096 (compute_points_to_sets): Likewise.
22097 (ipa_pta_execute): Likewise.
22099 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
22101 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
22102 * configure: Regenerate.
22103 * config.in: Regenerate.
22104 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
22105 #nobits/#progbits if supported.
22107 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
22110 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
22111 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
22112 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
22114 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
22116 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
22117 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
22119 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
22121 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
22122 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
22124 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
22126 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
22128 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
22129 * config/arm/cortex-a7.md: New bypasses using
22130 arm_mac_accumulator_is_result.
22132 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
22134 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
22135 (cortex_a7_neon_mla): Likewise.
22136 (cortex_a7_fpfmad): New reservation.
22137 (cortex_a7_fpmacs): Use ffmas and update required units.
22138 (cortex_a7_fpmuld): Update required units and latency.
22139 (cortex_a7_fpmacd): Likewise.
22140 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
22141 (cortex_a7_neon). Likewise.
22142 (bypass) Update participating units.
22144 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
22146 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
22147 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
22149 * config/arm/vfp11.md (vfp_farith): Use ffmas.
22150 (vfp_fmul): Use ffmad.
22151 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
22152 (cortex_r4_fmacd): Use ffmad.
22153 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
22154 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
22155 (cortex_a9_fmacd): Use ffmad.
22156 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
22157 (cortex_a8_vfp_macd): Use ffmad.
22158 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
22159 (cortex_a5_fpmacd): Use ffmad.
22160 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
22161 (cortex_a15_vfp_macd): Use ffmad.
22162 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
22164 2013-01-29 Jason Merrill <jason@redhat.com>
22167 * varasm.c (default_assemble_visibility): Don't warn about
22168 visibility on artificial decls.
22170 2013-01-29 Richard Biener <rguenther@suse.de>
22172 PR tree-optimization/56113
22173 * tree-ssa-structalias.c (equiv_class_lookup): Also return
22175 (label_visit): Free duplicate bitmaps and record the leader instead.
22176 (perform_var_substitution): Adjust.
22178 2013-01-29 Richard Biener <rguenther@suse.de>
22180 PR tree-optimization/55270
22181 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
22182 the CFG, schedule loops for fixup.
22184 2013-01-29 Nick Clifton <nickc@redhat.com>
22186 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
22189 2013-01-28 Leif Ekblad <leif@rdos.net>
22191 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
22192 * config/i386/i386.h (TARGET_RDOS): New macro.
22193 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
22194 * config/i386/i386.c (ix86_option_override_internal): For 64bit
22195 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
22196 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
22197 DEFAULT_LARGE_SECTION_THRESHOLD.
22198 * config/i386/i386.md (R14_REG, R15_REG): New constants.
22199 * config/i386/rdos.h: New file.
22200 * config/i386/rdos64.h: New file.
22202 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
22205 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
22208 2013-01-28 Jakub Jelinek <jakub@redhat.com>
22210 PR rtl-optimization/56117
22211 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
22212 call cselib_lookup_from_insn on the MEM before calling
22213 add_insn_mem_dependence.
22215 2013-01-28 Richard Biener <rguenther@suse.de>
22217 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
22218 to a stmt that didn't have one.
22219 (copy_phis_for_bb): Likewise for PHI arguments.
22220 (copy_debug_stmt): Likewise for debug stmts.
22222 2013-01-28 Richard Biener <rguenther@suse.de>
22224 PR tree-optimization/56034
22225 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
22226 (partition_builtin_p): Adjust.
22227 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
22228 it is the last partition.
22229 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
22230 up the vertex for the definition.
22231 (classify_partition): Classify whether a partition is a
22232 PKIND_REDUCTION, thus has uses outside of the loop.
22233 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
22234 Merge all PKIND_REDUCTION partitions into the last partition.
22235 (tree_loop_distribution): Seed partitions from reductions as well.
22237 2013-01-28 Jakub Jelinek <jakub@redhat.com>
22239 PR tree-optimization/56125
22240 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
22241 pow(x,c) into sqrt(x) * powi(x, n/2) or
22242 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
22243 optimizing for size.
22244 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
22245 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
22248 PR tree-optimization/56094
22249 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
22250 to UNKNOWN_LOCATION while gimplifying expr.
22252 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
22255 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
22256 operand 0 in movabs insn template for -masm=intel asm alternative.
22257 (*movabs<mode>_2): Ditto for operand 1.
22259 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
22262 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
22265 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
22267 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
22270 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22272 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
22273 Undef to avoid warning.
22275 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
22277 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
22278 * configure: Regenerate.
22280 2013-01-25 Jakub Jelinek <jakub@redhat.com>
22282 PR tree-optimization/56098
22283 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
22284 for stmts with volatile ops.
22285 (cond_store_replacement): Don't optimize if assign has volatile ops.
22286 (cond_if_else_store_replacement_1): Don't optimize if either
22287 then_assign or else_assign have volatile ops.
22288 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
22291 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
22293 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
22295 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
22297 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
22298 missing ':' in asm example.
22300 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
22302 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
22303 entries into lane and laneq entries.
22304 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
22305 Remove AdvSIMD scalar modes.
22306 (aarch64_sq<r>dmulh_laneq<mode>): New.
22307 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
22309 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
22310 builtin implementations to relfect changes in RTL in aarch64-simd.md.
22311 * config/aarch64/iterators.md (VCOND): New.
22314 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
22317 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
22318 Add NULL LIBNAME argument to existing definitions.
22319 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
22320 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
22321 * config/avr/avr.c (DEF_BUILTIN): Same.
22322 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
22323 (avr_expand_builtin): Expand to a vanilla call if a libgcc
22324 implementation is available (DECL_ASSEMBLER_NAME is set).
22325 (avr_fold_absfx): New static function.
22326 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
22327 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
22328 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
22329 AVR_BUILTIN_ABSLLK.
22330 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
22331 (abshk, absk, abslk, absllk): Provide as static inline functions.
22333 2013-01-25 Marek Polacek <polacek@redhat.com>
22335 PR tree-optimization/56035
22336 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
22338 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
22340 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
22341 (*movtf_internal_rex64): Add (!o,C) alternative
22342 (*movxf_internal_rex64): Ditto.
22343 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
22345 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
22347 * doc/invoke.texi: fix typo.
22348 * doc/objc.texi: fix typo.
22350 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
22352 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
22353 for the first two alternatives.
22355 2013-01-24 Diego Novillo <dnovillo@google.com>
22357 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
22358 (ggc-zone.o): Remove.
22359 * configure.ac: Remove option --with-gc.
22360 * configure: Re-generate.
22361 * doc/install.texi: Remove documentation for --with-gc.
22362 * gengtype.c (write_enum_defn): Remove. Update all users.
22363 (write_Types_process_field): Remove generation of gt_e_* argument.
22364 (output_type_enum): Remove. Update all users.
22365 (write_enum_defn): Remove. Update all users.
22366 (enum alloc_zone): Remove. Update all users.
22367 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
22368 * ggc-common.c (ggc_splay_alloc): Remove first argument.
22369 Update all callers.
22370 (struct ptr_data): Remove field TYPE. Update all users.
22371 (gt_pch_note_object): Remove argument TYPE. Update all users.
22372 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
22374 * ggc-none.c (ggc_alloc_typed_stat): Remove.
22375 (struct alloc_zone): Remove.
22376 (ggc_internal_alloc_zone_stat): Remove.
22377 (ggc_internal_cleared_alloc_zone_stat): Remove.
22378 * ggc-page.c (ggc_alloc_typed_stat): Remove.
22379 (ggc_pch_count_object): Remove last argument. Update all users.
22380 (ggc_pch_alloc_object): Remove last argument. Update all users.
22381 (struct alloc_zone): Remove.
22382 * ggc-zone.c: Remove.
22383 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
22384 (struct alloc_zone): Remove.
22385 (ggc_alloc_typed_stat): Remove.
22386 (ggc_alloc_typed): Remove.
22387 (ggc_splay_alloc): Remove first argument.
22388 (rtl_zone): Remove. Update all users.
22389 (tree_zone): Remove. Update all users.
22390 (tree_id_zone): Remove. Update all users.
22391 (ggc_internal_zone_alloc_stat): Remove. Update all users.
22392 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
22393 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
22394 * tree-ssanames.c: Remove references to zone allocator in comments.
22396 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
22398 * config/avr/avr.c (avr_out_fract): Make register numbers that
22399 might be outside of source operand signed.
22401 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
22403 * config/i386/constraints.md (Yf): New constraint.
22404 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
22405 of f constraint to conditionaly disable x87 register preferences.
22406 (*movdf_internal): Ditto.
22407 (*movsf_internal): Ditto.
22409 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
22411 PR inline-asm/55934
22412 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
22413 that have operands with impossible constraints.
22414 Add a FIXME for a speed-up opportunity.
22415 * lra-constraints.c (process_alt_operands): Verify that a class
22416 selected from constraints on asms is valid for the operand mode.
22417 (curr_insn_transform): Remove incorrect comment.
22419 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
22421 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
22422 TOC operand is a valid symbol ref in the constant pool.
22424 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
22426 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
22428 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
22431 * config/avr/stdfix.h: New file.
22432 * t-avr (stdfix-gcc.h): New rule to build it.
22433 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
22435 2013-01-23 Kostya Serebryany <kcc@google.com>
22437 * config/darwin.h: remove dependency on
22438 CoreFoundation (asan on Mac OS).
22440 2013-01-23 Jakub Jelinek <jakub@redhat.com>
22443 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
22444 instead of cmpdi_operand for first comparison operand.
22445 Don't assert that comparison operands aren't both constants.
22447 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
22449 * doc/install.texi (Downloading the Source): Update references to
22450 downloading separate components.
22452 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
22454 * doc/extend.texi (__int128): Improve grammar.
22456 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
22459 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
22460 alternative to (o,r).
22461 (*movdi_internal_rex64): Remove (!o,n) alternative.
22462 (DImode immediate->memory splitter): Remove.
22463 (DImode immediate->memory peephole2): Remove.
22464 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
22465 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
22466 alternative to (!o,*r).
22467 (*movtf_internal_sse): New pattern.
22468 (*movxf_internal_rex64): New pattern.
22469 (*movxf_internal): Disable for TARGET_64BIT.
22470 (*movdf_internal_rex64): Remove (!o,F) alternative.
22472 2013-01-22 Jakub Jelinek <jakub@redhat.com>
22474 PR middle-end/56074
22475 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
22476 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
22477 * tree-vect-loop-manip.c (find_loop_location): Also ignore
22478 stmt locations where LOCATION_LOCUS of the stmt location is
22479 UNKNOWN_LOCATION or BUILTINS_LOCATION.
22482 * config/i386/i386.md (UNSPEC_STOS): New.
22483 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
22484 *strsetqi_1): Add UNSPEC_STOS.
22486 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
22489 * doc/invoke.texi: Remove left over -Wsynth example.
22491 2013-01-21 Jakub Jelinek <jakub@redhat.com>
22493 PR tree-optimization/56051
22494 * fold-const.c (fold_binary_loc): Don't fold
22495 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
22496 a narrowing conversion, or widening conversion from signed
22499 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
22501 PR rtl-optimization/56023
22502 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
22503 dependent on debug instruction.
22505 2013-01-21 Martin Jambor <mjambor@suse.cz>
22507 PR middle-end/56022
22508 * function.c (allocate_struct_function): Call
22509 invoke_set_current_function_hook earlier.
22511 2013-01-21 Jakub Jelinek <jakub@redhat.com>
22513 * reload1.c (init_reload): Only initialize reload_obstack
22514 during the first call.
22516 2013-01-21 Marek Polacek <polacek@redhat.com>
22518 * cfgloop.c (verify_loop_structure): Fix up grammar.
22520 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
22522 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
22523 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
22525 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
22528 * config/arm/marvell-pj4.md: Update copyright year.
22529 Fix up use of alu to alu_reg and simple_alu_imm.
22531 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
22533 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
22535 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
22538 * lra-constraints.c (curr_insn_transform): Don't reuse original
22539 insn for secondary memory move when memory mode should be different.
22541 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
22543 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
22544 atomic_storedi_1): New patterns.
22546 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
22548 btver2 pipeline descriptions.
22549 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
22551 * config/i386/i386.md (btver2_decode): New type attributes.
22552 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
22554 * config/i386/btver2.md: New file describing btver2 pipelines.
22556 2013-01-19 Andrew Pinski <apinski@cavium.com>
22558 PR tree-optimization/52631
22559 * tree-ssa-sccvn (visit_use): Before looking up the original
22560 statement, try looking up the simplified expression.
22562 2013-01-19 Anthony Green <green@moxielogic.com>
22564 * config/moxie/moxie.c (moxie_expand_prologue): Set
22565 current_function_static_stack_size.
22567 2013-01-18 Jakub Jelinek <jakub@redhat.com>
22569 PR tree-optimization/56029
22570 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
22571 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
22573 2013-01-18 Sharad Singhai <singhai@google.com>
22575 PR tree-optimization/55995
22576 * dumpfile.c (dump_loc): Print location only if available.
22577 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
22579 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
22582 * lra-constraints.c (curr_insn_transform): Reuse original insn for
22583 secondary memory move.
22584 (inherit_reload_reg): Use rclass instead of cl for
22585 check_secondary_memory_needed_p.
22587 2013-01-18 Jakub Jelinek <jakub@redhat.com>
22589 PR middle-end/56015
22590 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
22591 the case where writing real complex part of target modifies op1.
22593 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
22595 * config/aarch64/aarch64-simd.md
22596 (aarch64_vcond_internal<mode>): Handle unordered cases.
22597 * config/aarch64/iterators.md (v_cmp_result): New.
22599 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
22600 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
22602 * config/arm/marvell-pj4.md: New file.
22603 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
22604 * config/arm/arm.md (generic_sched): Add marvell_pj4.
22605 (generic_vfp): Likewise.
22606 * config/arm/arm-cores.def: Add marvell-pj4.
22607 * config/arm/arm-tune.md: Regenerate.
22608 * config/arm/arm-tables.opt: Regenerate.
22609 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
22610 * doc/invoke.texi: Document marvell-pj4.
22612 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
22614 * config/aarch64/arm_neon.h: Map scalar types to standard types.
22616 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
22621 * var-tracking.c (negative_power_of_two_p): New.
22622 (global_get_addr_cache, local_get_addr_cache): New.
22623 (get_addr_from_global_cache, get_addr_from_local_cache): New.
22624 (vt_canonicalize_addr): Rewrite using the above. Adjust the
22626 (vt_stack_offset_p): Remove.
22627 (vt_canon_true_dep): Always canonicalize loc's address.
22628 (clobber_overlapping_mems): Make sure we have a MEM.
22629 (local_get_addr_clear_given_value): New.
22630 (val_reset): Clear local cached entries.
22631 (compute_bb_dataflow): Create and release the local cache.
22632 Disable duplicate MEMs clobbering.
22633 (emit_notes_in_bb): Clobber MEMs likewise.
22634 (vt_emit_notes): Create and release the local cache.
22635 (vt_initialize, vt_finalize): Create and release the global
22636 cache, respectively.
22637 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
22639 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
22641 PR libmudflap/53359
22642 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
22643 not found in the symtab.
22645 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
22648 PR rtl-optimization/55547
22649 PR rtl-optimization/53827
22652 * alias.c (offset_overlap_p): New, factored out of...
22653 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
22654 the conservative special case for symbolic constants. Don't
22655 adjust zero sizes on alignment.
22657 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
22659 PR rtl-optimization/52573
22660 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
22661 REG_UNUSED for the same register.
22663 2013-01-17 Richard Biener <rguenther@suse.de>
22664 Marek Polacek <polacek@redhat.com>
22666 PR rtl-optimization/55833
22667 * loop-unswitch.c (unswitch_loops): Move loop verification...
22668 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
22669 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
22670 Set it to true when we're removing a loop from hierarchy tree in
22671 an irreducible region.
22672 (fix_bb_placements): Adjust caller.
22673 (fix_loop_placements): Likewise.
22675 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
22677 * config/avr/builtins.def (DEF_BUILTIN): Factor out
22678 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
22679 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
22680 Remove ID. Adjust comments.
22681 * config/avr/avr-c.c (avr_builtin_name): Remove.
22682 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
22683 * config/avr/avr.c (avr_tolower): New static function.
22684 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
22685 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
22686 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
22689 2013-01-17 Jan Hubicka <jh@suse.cz>
22691 PR tree-optimization/55273
22692 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
22694 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
22697 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
22698 store through atomic_store<mode>_1.
22699 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
22701 2013-01-17 Martin Jambor <mjambor@suse.cz>
22703 PR tree-optimizations/55264
22704 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
22705 for virtual methods.
22706 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
22707 virtual methods before inlining is over.
22708 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
22710 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
22713 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
22715 PR rtl-optimization/56005
22716 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
22717 pending reads for prefetch.
22719 2013-01-16 Ian Bolton <ian.bolton@arm.com>
22721 * config/aarch64/aarch64.md
22722 (*cstoresi_neg_uxtw): New pattern.
22723 (*cmovsi_insn_uxtw): New pattern.
22724 (*<optab>si3_uxtw): New pattern.
22725 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
22726 (*<optab>si3_insn_uxtw): New pattern.
22727 (*bswapsi2_uxtw): New pattern.
22729 2013-01-16 Richard Biener <rguenther@suse.de>
22731 * tree-inline.c (tree_function_versioning): Remove set but
22732 never used variable.
22734 2013-01-16 Richard Biener <rguenther@suse.de>
22736 PR tree-optimization/55964
22737 * tree-flow.h (rename_variables_in_loop): Remove.
22738 (rename_variables_in_bb): Likewise.
22739 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
22740 (copy_loop_before): Adjust and delete update-ssa status.
22741 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
22742 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
22743 (rename_variables_in_loop): Remove.
22744 (slpeel_update_phis_for_duplicate_loop): Likewise.
22745 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
22746 use available cfg machinery instead of duplicating it.
22747 Update PHI nodes and perform poor-mans SSA update here.
22748 (slpeel_tree_peel_loop_to_edge): Adjust.
22750 2013-01-16 Richard Biener <rguenther@suse.de>
22752 PR tree-optimization/54767
22753 PR tree-optimization/53465
22754 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
22755 (vrp_visit_phi_node): For PHI arguments coming via backedges
22756 drop all symbolical range information.
22757 (execute_vrp): Compute backedges.
22759 2013-01-16 Richard Biener <rguenther@suse.de>
22761 * doc/install.texi: Update CLooG and ISL requirements to
22764 2013-01-16 Christian Bruel <christian.bruel@st.com>
22767 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
22768 (broken_move): Handle UNSPECV_SP_SWITCH_B.
22769 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
22771 2013-01-16 DJ Delorie <dj@redhat.com>
22773 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
22774 (UNSPECV_SP_SWITCH_E): New.
22775 (sp_switch_1): Change to an unspec.
22776 (sp_switch_2): Change to an unspec. Don't use post-inc when we
22779 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
22781 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
22782 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
22783 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
22784 (expand_mem_thread_fence): Ditto.
22785 (expand_mem_signal_fence): Ditto.
22786 (expand_atomic_load): Ditto.
22787 (expand_atomic_store): Ditto.
22789 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
22791 PR rtl-optimization/55547
22792 PR rtl-optimization/53827
22795 * alias.c (memrefs_conflict_p): Set sizes to negative after
22798 2013-01-15 Jakub Jelinek <jakub@redhat.com>
22801 * function.c (thread_prologue_and_epilogue_insns): Always
22802 add crtl->drap_reg to set_up_by_prologue.set, even if
22803 stack_realign_drap is false.
22805 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
22807 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
22808 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
22809 *call): Fix indention.
22811 2013-01-15 Tom de Vries <tom@codesourcery.com>
22814 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
22817 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
22819 PR rtl-optimization/55153
22820 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
22822 2013-01-15 Martin Jambor <mjambor@suse.cz>
22824 PR tree-optimization/55920
22825 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
22826 accesses as grp_to_be_debug_replaced.
22828 2013-01-15 Jakub Jelinek <jakub@redhat.com>
22830 PR tree-optimization/55920
22831 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
22832 there is non-useless type conversion needed from debug rhs to lhs,
22833 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
22835 2013-01-15 Joseph Myers <joseph@codesourcery.com>
22836 Mikael Pettersson <mikpe@it.uu.se>
22839 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
22841 (ASM_OUTPUT_CASE_LABEL): Remove.
22842 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
22843 * final.c (shorten_branches): Update alignment of labels before
22844 jump tables if CASE_VECTOR_SHORTEN_MODE.
22846 2013-01-15 Richard Biener <rguenther@suse.de>
22849 * system.h: Do not include gmp.h for building host tools.
22851 2013-01-15 Richard Biener <rguenther@suse.de>
22853 PR middle-end/55882
22854 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
22855 account for bitpos when computing alignment.
22857 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
22859 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
22860 (ix86_target_macros_internal): Likewise.
22862 * config/i386/i386.c (m_CORE2I7): Removed.
22863 (m_CORE_HASWELL): New macro.
22864 (m_CORE_ALL): Likewise.
22865 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
22866 (initial_ix86_arch_features): Likewise.
22867 (processor_target_table): Initializations for Core avx2.
22868 (cpu_names): New names "core-avx2".
22869 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
22870 PROCESSOR_CORE_HASWELL.
22871 (ix86_issue_rate): New case.
22872 (ia32_multipass_dfa_lookahead): Likewise.
22873 (ix86_sched_init_global): Likewise.
22875 * config/i386/i386.h (TARGET_HASWELL): New macro.
22876 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
22877 (processor_type): New PROCESSOR_HASWELL.
22879 2013-01-15 Jakub Jelinek <jakub@redhat.com>
22881 PR tree-optimization/55955
22882 * tree-vect-loop.c (vectorizable_reduction): Give up early on
22883 *SHIFT_EXPR and *ROTATE_EXPR codes.
22885 PR tree-optimization/48766
22886 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
22887 -ftrapv disable -fwrapv.
22889 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
22892 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
22893 etc. to 1 and not to __flash.
22894 Use LL suffix for __INT24_MAX__ with -mint8.
22895 Use ULL suffix for __UINT24_MAX__ with -mint8.
22897 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
22899 * config/avr/avr-arch.h
22900 (struct base_arch_s): Use typedef avr_arch_t instead.
22901 (struct arch_info_s): Use typedef avr_arch_info_t instead.
22902 (struct mcu_type_s): Use typedef avr_mcu_t instead.
22903 * config/avr/avr.c: Same.
22904 * config/avr/avr-devices.c: Same.
22905 * config/avr/driver-avr.c: Same.
22906 * config/avr/gen-avr-mmcu-texi.c: Same.
22907 * config/avr/avr-mcus.def: Adjust comment.
22909 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
22911 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
22912 * config/aarch64/iterators.md (VALLDI): New.
22914 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
22915 Andi Kleen <ak@linux.intel.com>
22918 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
22919 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
22922 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
22924 * config/avr/avr-stdint.h: Remove trailing blanks.
22925 * config/avr/avr-log.h: Same.
22926 * config/avr/avr-arch.h: Same.
22927 * config/avr/avr-devices.c: Same.
22928 * config/avr/avr-dimode.md: Same.
22929 * config/avr/predicates.md: Same.
22930 * config/avr/avr-c.c: Same. And fix typo.
22932 * config/avr/avr-protos.h: Same. And:
22933 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
22934 (init_cumulative_args): Rename to avr_init_cumulative_args.
22935 (expand_prologue): Rename to avr_expand_prologue.
22936 (expand_epilogue): Rename to avr_expand_epilogue.
22937 (adjust_insn_length): Rename to avr_adjust_insn_length.
22938 (notice_update_cc): Rename to avr_notice_update_cc.
22939 (final_prescan_insn): Rename to avr_final_prescan_insn.
22940 * config/avr/avr.c: Same.
22941 * config/avr/avr.h: Same.
22942 * config/avr/avr.md: Remove trailing blanks.
22943 (prologue): Use avr_expand_prologue.
22944 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
22946 2013-01-14 Richard Biener <rguenther@suse.de>
22948 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
22949 verify_location, collect_subblocks): New functions.
22950 (verify_gimple_in_cfg): Verify that locations only reference
22951 BLOCKs in the functions BLOCK tree.
22953 2013-01-14 Richard Biener <rguenther@suse.de>
22955 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
22957 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
22959 (insert_out_of_ssa_copy_on_edge): Likewise.
22960 (rewrite_close_phi_out_of_ssa): Likewise.
22961 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
22963 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
22964 propagated constants.
22965 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
22968 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
22970 * config/avr/avr-modes.def: Add GPL copyright notice.
22972 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
22974 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
22975 MEMMODEL_MASK to determine memory model.
22976 (atomic_store<mode>): Ditto from operands[2].
22977 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
22979 2013-01-13 Jakub Jelinek <jakub@redhat.com>
22982 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
22983 (fold_gimple_assign): Don't call unshare_expr here.
22984 (fold_ctor_reference): Call unshare_expr.
22986 2013-01-13 Terry Guo <terry.guo@arm.com>
22988 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
22989 * doc/fragments.texi: Document MULTILIB_REUSE.
22990 * gcc.c (multilib_reuse): New internal spec.
22991 (set_multilib_dir): Also search multilib from multilib_reuse.
22992 * genmultilib (tmpmultilib3): Refactor code.
22993 (tmpmultilib4): Ditto.
22994 (multilib_reuse): New multilib argument.
22996 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
22998 * Makefile.in: Update copyright.
23000 2013-01-12 Tom de Vries <tom@codesourcery.com>
23002 PR middle-end/55890
23003 * calls.c (expand_call): Check if arg_nr is valid.
23005 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
23007 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
23008 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
23009 documentation. Add missing '__' in front of
23010 __builtin_ia32_packssdw256.
23012 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23015 * config/s390/s390.c (s390_preferred_reload_class): Do not return
23016 NO_REGS for larl operands.
23017 (s390_reload_larl_operand): Use s390_load_address instead of
23020 2013-01-11 Richard Biener <rguenther@suse.de>
23022 * tree-cfg.c (verify_node_sharing_1): Split out from ...
23023 (verify_node_sharing): ... here.
23024 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
23026 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
23028 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
23029 Substitute TREECHECKING.
23030 * configure: Regenerate.
23031 * Makefile.in (TREECHECKING): New.
23033 2013-01-11 Richard Guenther <rguenther@suse.de>
23035 PR tree-optimization/44061
23036 * tree-vrp.c (extract_range_basic): Compute zero as
23037 value-range for __builtin_constant_p of function parameters.
23039 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
23041 Update copyright years.
23043 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
23045 PR rtl-optimization/55672
23046 * lra-eliminations.c (mark_not_eliminable): Permit addition with
23047 const to be eliminable.
23049 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
23051 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
23052 * configure: Regenerate.
23054 2013-01-10 Richard Biener <rguenther@suse.de>
23056 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
23058 2013-01-10 Richard Biener <rguenther@suse.de>
23061 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
23062 locations for virtual PHI arguments.
23063 (rewrite_update_phi_arguments): Likewise.
23065 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
23067 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
23070 2013-01-10 Jakub Jelinek <jakub@redhat.com>
23072 PR tree-optimization/55921
23073 * tree-complex.c (expand_complex_asm): New function.
23074 (expand_complex_operations_1): Call it for GIMPLE_ASM.
23076 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23079 * config/s390/s390.c (s390_symref_operand_p)
23080 (s390_loadrelative_operand_p): Merge the two functions.
23081 (s390_check_qrst_address, print_operand_address): Add parameters
23082 to s390_loadrelative_operand_p invokation.
23083 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
23084 (s390_reload_larl_operand, s390_secondary_reload): Use
23085 s390_loadrelative_operand_p instead of s390_symref_operand_p.
23086 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
23088 2013-01-09 Mike Stump <mikestump@comcast.net>
23090 * dse.c (record_store): Remove unnecessary assert.
23092 2013-01-09 Jan Hubicka <jh@suse.cz>
23094 PR tree-optimization/55569
23095 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
23096 * cfgloop.h (scale_loop_profile): Likewise.
23098 2013-01-09 Jan Hubicka <jh@suse.cz>
23101 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
23103 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
23105 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
23107 PR middle-end/55114
23108 * expr.h (maybe_emit_group_store): Declare.
23109 * expr.c (maybe_emit_group_store): New function.
23110 * builtins.c (expand_builtin_int_roundingfn): Call it.
23111 (expand_builtin_int_roundingfn_2): Likewise.
23113 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
23115 PR rtl-optimization/55829
23116 * lra-constraints.c (match_reload): Add code for absent output.
23117 (curr_insn_transform): Add code for reloads of matched inputs
23120 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
23122 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
23123 attribute of movddup insn to DF.
23124 (*vec_interleave_lowv2df): Ditto.
23125 (vec_dupv2df): Ditto.
23127 2013-01-09 Jan Hubicka <jh@suse.cz>
23129 PR tree-optimiation/55875
23130 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
23131 EVERY_ITERATION parameter.
23132 (number_of_iterations_exit): Check if exit is executed every iteration.
23133 (idx_infer_loop_bounds): Similarly here.
23134 (n_of_executions_at_most): Simplify
23135 to only test for cases where statement is dominated by the
23136 particular bound; handle correctly the "postdominance" test.
23137 (scev_probably_wraps_p): Use max loop iterations info
23138 as a global bound first.
23140 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
23141 Nick Clifton <nickc@redhat.com>
23143 * config/v850/v850.md (cbranchsf4): New pattern.
23144 (cstoresf4): New pattern.
23145 (cbranchdf4): New pattern.
23146 (cstoredf4): New pattern.
23147 (movsicc): Disallow floating point comparisons.
23148 (cmpsf_le_insn): Fix order of operators.
23149 (cmpsf_lt_insn): Likewise.
23150 (cmpsf_eq_insn): Likewise.
23151 (cmpdf_le_insn): Likewise.
23152 (cmpdf_lt_insn): Likewise.
23153 (cmpdf_eq_insn): Likewise.
23154 (cmpsf_ge_insn): Use LE comparison.
23155 (cmpdf_ge_insn): Likewise.
23156 (cmpsf_gt_insn): Use LT comparison.
23157 (cmpdf_gt_insn): Likewise.
23158 (cmpsf_ne_insn): Delete pattern.
23159 (cmpdf_ne_insn): Delete pattern.
23160 * config/v850/v850.c (v850_gen_float_compare): Use
23161 gen_cmpdf_eq_insn for NE comparison.
23162 (v850_float_z_comparison_operator)
23163 (v850_float_nz_comparison_operator): Move from here ...
23164 * config/v850/predicates.md: ... to here. Move GT and GE
23165 comparisons into v850_float_z_comparison_operator.
23166 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
23168 (v850_float_nz_comparison_operator): Likewise.
23170 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23172 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
23173 with calls to gen_insvsi/gen_insvdi.
23175 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
23177 * config/i386/i386.c (initial_ix86_tune_features): Set up
23178 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
23180 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
23181 Jakub Jelinek <jakub@redhat.com>
23183 PR tree-optimization/48189
23184 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
23185 If nitercst is 0, don't predict the exit edge.
23187 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23189 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
23190 in asm_fprintf with reg_names.
23191 (aarch64_print_operand_address): Likewise.
23192 (aarch64_return_addr): Likewise.
23193 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
23195 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23197 * config/pa/pa.h (VAL_U6_BITS_P): Define.
23198 (INT_U6_BITS): Likewise.
23199 * config/pa/predicates.md (uint6_operand): New predicate.
23200 (shift5_operand, shift6_operand): Likewise.
23201 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
23203 (lshrdi3): Use shift6_operand.
23204 (shrpsi4, shrpdi4): New insn patterns.
23205 (extzv): Delete expander.
23206 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
23207 predicates in unamed zero extract patterns. Tighten common constraint.
23208 (extv): Delete expander.
23209 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
23210 predicates in unamed sign extract patterns. Tighten common constraint.
23211 (insv): Delete expander.
23212 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
23213 predicates in unamed insert patterns. Tighten common constraint.
23214 Change uint32_operand predicate to uint6_operand predicate in unamed
23215 DImode pattern to insert constant values of type 1...1xxxx.
23217 2013-01-04 Jan Hubicka <jh@suse.cz>
23219 PR tree-optimization/55823
23220 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
23223 2013-01-08 Jakub Jelinek <jakub@redhat.com>
23224 Uros Bizjak <ubizjak@gmail.com>
23226 PR rtl-optimization/55845
23227 * df-problems.c (can_move_insns_across): Stop scanning at
23228 volatile_insn_p source instruction or give up if
23229 across_from .. across_to range contains any volatile_insn_p
23232 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
23234 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
23235 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
23237 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
23238 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
23240 2013-01-08 Jakub Jelinek <jakub@redhat.com>
23243 * asan.c (asan_clear_shadow): New function.
23244 (asan_emit_stack_protection): Use it.
23246 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
23248 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
23249 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
23250 with tab instead of space.
23252 2013-01-08 Nick Clifton <nickc@redhat.com>
23254 * config/rl78/rl78.c (rl78_expand_prologue): Always select
23255 register bank 0 at the start of an interrupt handler.
23256 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
23259 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
23261 * config/aarch64/aarch64-simd.md
23262 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
23263 (aarch64_simd_bsl): Likewise.
23264 (aarch64_vcond_internal<mode>): Likewise.
23265 (vcond<mode><mode>): Likewise.
23266 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
23267 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
23269 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
23271 * config/aarch64/aarch64-builtins.c
23272 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
23274 2013-01-08 Martin Jambor <mjambor@suse.cz>
23277 * tree-sra.c (analyze_access_subtree): Return true also after
23278 potentially creating a debug-only replacement.
23280 2013-01-08 Jakub Jelinek <jakub@redhat.com>
23282 PR middle-end/55890
23283 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
23285 PR tree-optimization/54120
23286 * tree-vrp.c (range_fits_type_p): Don't allow
23287 src_precision < precision from signed vr to unsigned_p
23288 if vr->min or vr->max is negative.
23289 (simplify_float_conversion_using_ranges): Test can_float_p
23290 against CODE_FOR_nothing.
23292 2013-01-08 Jakub Jelinek <jakub@redhat.com>
23293 Richard Biener <rguenther@suse.de>
23295 PR middle-end/55851
23296 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
23297 types instead of just INTEGER_TYPE types.
23299 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
23301 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
23304 2013-01-07 Steve Ellcey <sellcey@mips.com>
23307 * config/mips/mips.opt: Change mad to mmad to match documentation.
23309 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
23312 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
23313 .progmemx.data now.
23315 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
23318 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
23319 (avr_addrspace_t): Add .section_name field.
23320 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
23322 (avr_addrspace): Same. Initialize .section_name. Remove last
23323 NULL entry. Put __memx into .progmemx.data.
23324 (progmem_section_prefix): Remove.
23325 (avr_asm_init_sections): No need to initialize progmem_section.
23326 (avr_asm_named_section): Use avr_addrspace[].section_name to get
23327 section name prefix.
23328 (avr_asm_select_section): Ditto. And use get_unnamed_section to
23329 retrieve the progmem section.
23330 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
23331 boundary to run over avr_addrspace[].
23332 (avr_register_target_pragmas): Ditto.
23334 2013-01-06 Jakub Jelinek <jakub@redhat.com>
23336 * varasm.c (output_constant_def_contents): For asan_protect_global
23337 protected strings, adjust DECL_ALIGN if needed, before testing for
23339 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
23340 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
23342 (output_object_block): For asan protected decls, emit asan padding
23343 after their contents.
23344 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
23345 (asan_finish_file): Test it here instead.
23347 2013-01-07 Nick Clifton <nickc@redhat.com>
23348 Matthias Klose <doko@debian.org>
23349 Doug Kwan <dougkwan@google.com>
23350 H.J. Lu <hongjiu.lu@intel.com>
23353 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
23355 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
23357 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
23359 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
23361 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
23363 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
23366 * doc/install.texi (Cross-Compiler-Specific Options): Document
23369 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
23371 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
23372 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
23373 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
23374 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
23375 vqmovun_high_s64): Fix source operand number and update copyright.
23377 2013-01-07 Richard Biener <rguenther@suse.de>
23379 PR middle-end/55890
23380 * gimple.h (gimple_call_builtin_p): New overload.
23381 * gimple.c (validate_call): New function.
23382 (gimple_call_builtin_p): Likewise.
23383 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
23384 Use gimple_call_builtin_p.
23385 (find_func_clobbers): Likewise.
23386 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
23387 (strlen_optimize_stmt): Likewise.
23389 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
23391 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
23392 (vld1q_dup_*): Likewise.
23393 (vld1_*): Likewise.
23394 (vld1q_*): Likewise.
23395 (vld1_lane_*): Likewise.
23396 (vld1q_lane_*): Likewise.
23398 2013-01-07 Richard Biener <rguenther@suse.de>
23400 * lto-streamer.h (LTO_minor_version): Bump to 2.
23402 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
23404 * config/aarch64/aarch64-protos.h
23405 (aarch64_const_double_zero_rtx_p): Rename to...
23406 (aarch64_float_const_zero_rtx_p): ...this.
23407 (aarch64_float_const_representable_p): New.
23408 (aarch64_output_simd_mov_immediate): Likewise.
23409 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
23410 move immediate case.
23411 * config/aarch64/aarch64.c
23412 (aarch64_const_double_zero_rtx_p): Rename to...
23413 (aarch64_float_const_zero_rtx_p): ...this.
23414 (aarch64_print_operand): Allow printing of new constants.
23415 (aarch64_valid_floating_const): New.
23416 (aarch64_legitimate_constant_p): Check for valid floating-point
23418 (aarch64_simd_valid_immediate): Likewise.
23419 (aarch64_vect_float_const_representable_p): New.
23420 (aarch64_float_const_representable_p): Likewise.
23421 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
23422 (aarch64_output_simd_mov_immediate): New.
23423 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
23424 (*movdf_aarch64): Likewise.
23425 * config/aarch64/constraints.md (Ufc): New.
23426 (Y): call aarch64_float_const_zero_rtx.
23427 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
23429 2013-01-07 Richard Biener <rguenther@suse.de>
23431 PR tree-optimization/55888
23432 PR tree-optimization/55862
23433 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
23434 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
23435 not if it is contained therein.
23437 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
23439 * config/avr/t-avr: Typo.
23441 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
23444 * config/avr/t-avr: Don't automatically rebuild
23445 $(srcdir)/config/avr/t-multilib
23446 $(srcdir)/config/avr/avr-tables.opt
23447 $(srcdir)/doc/avr-mmcu.texi
23448 (avr-mcus): New phony target to build them on request.
23449 (s-avr-mlib, s-avr-mmcu-texi): Remove.
23450 * avr/avr-mcus.def: Adjust comments.
23452 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
23454 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
23456 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
23458 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
23460 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
23462 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
23464 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
23466 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
23467 to generate profiling.
23468 * config/rs6000/aix64.h (LIB_SPEC): Same.
23470 2013-01-04 Andrew Pinski <apinski@cavium.com>
23472 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
23474 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
23476 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
23478 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
23480 (ix86_expand_move): Ditto.
23481 (ix86_zero_extend_to_Pmode): Ditto.
23482 (ix86_expand_call): Ditto.
23483 (ix86_expand_special_args_builtin): Ditto.
23484 (ix86_expand_builtin): Ditto.
23486 2013-01-04 Richard Biener <rguenther@suse.de>
23488 PR tree-optimization/55862
23489 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
23490 translating them through PHI nodes.
23492 2013-01-04 Martin Jambor <mjambor@suse.cz>
23494 PR tree-optimization/55755
23495 * tree-sra.c (sra_modify_assign): Do not check that an access has no
23496 children when trying to avoid producing a VIEW_CONVERT_EXPR.
23498 2013-01-04 Marek Polacek <polacek@redhat.com>
23500 PR middle-end/55859
23501 * opts.c (default_options_optimization): Clarify error message.
23503 2013-01-04 Richard Biener <rguenther@suse.de>
23505 PR middle-end/55863
23506 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
23509 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23512 * config/pa/pa.md (movsi): Revert previous change.
23513 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
23516 2013-01-03 Richard Henderson <rth@redhat.com>
23518 * config/i386/i386.c (ix86_expand_move): Always assign to op1
23519 after eliminating TLS symbols.
23521 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
23524 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
23525 * graphite-poly.c (debug_gmp_value): Likewise.
23527 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
23530 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
23531 selected code model, define __code_mode_small__, __code_model_medium__,
23532 __code_model_large__, __code_model_32__ or __code_model_kernel__.
23533 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
23534 xchg temporary register with %k. Declare temporary register as
23536 [__x86_64__]: For medium and large code models, preserve %rbx register.
23538 2013-01-03 Richard Biener <rguenther@suse.de>
23540 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
23541 (dump_subscript): Adjust.
23542 (finalize_ddr_dependent): Do not dump redundant info.
23543 (analyze_siv_subscript): Adjust.
23544 (subscript_dependence_tester): Likewise.
23545 (compute_affine_dependence): Likewise.
23547 2013-01-03 Richard Biener <rguenther@suse.de>
23550 2013-01-03 Richard Biener <rguenther@suse.de>
23552 PR tree-optimization/55857
23553 * tree-vect-stmts.c (vectorizable_load): Do not setup
23554 re-alignment for invariant loads.
23556 2013-01-02 Richard Biener <rguenther@suse.de>
23558 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
23559 invariant load do not generate a vector load from the scalar location.
23561 2013-01-03 Richard Biener <rguenther@suse.de>
23563 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
23564 for not vectorizing.
23565 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
23566 not build INDIRECT_REFs, call get_name once only.
23567 (vect_create_data_ref_ptr): Likewise. Dump base object kind
23568 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
23570 2013-01-03 Richard Biener <rguenther@suse.de>
23572 PR tree-optimization/55857
23573 * tree-vect-stmts.c (vectorizable_load): Do not setup
23574 re-alignment for invariant loads.
23576 2013-01-03 Richard Biener <rguenther@suse.de>
23579 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
23580 prefer a built-in decl.
23582 2013-01-03 Jakub Jelinek <jakub@redhat.com>
23584 * gcc.c (process_command): Update copyright notice dates.
23585 * gcov.c (print_version): Likewise.
23586 * gcov-dump.c (print_version): Likewise.
23588 PR rtl-optimization/55838
23589 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
23590 iv0.step, iv1.step and step.
23592 2013-01-03 Jakub Jelinek <jakub@redhat.com>
23593 Marc Glisse <marc.glisse@inria.fr>
23595 PR tree-optimization/55832
23596 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
23597 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
23598 integer_{one,zero}_node.
23600 2013-01-03 Jakub Jelinek <jakub@redhat.com>
23603 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
23604 * var-tracking.c (reverse_op): Don't add reverse ops to
23605 VALUEs that have already
23606 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
23608 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
23610 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
23612 2013-01-02 Teresa Johnson <tejohnson@google.com>
23614 * dumpfile.c (dump_loc): Print filename with location.
23615 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
23616 new location_t parameter to emit complete unroll message with
23617 new dump framework.
23618 (canonicalize_loop_induction_variables): Compute loops location
23619 and pass to try_unroll_loop_completely.
23620 * loop-unroll.c (report_unroll_peel): New function.
23621 (peel_loops_completely): Use new dump format with location
23622 for main dumpfile message, and invoke report_unroll_peel on success.
23623 (decide_unrolling_and_peeling): Ditto.
23624 (decide_peel_once_rolling): Remove old dumpfile message subsumed
23625 by report_unroll_peel.
23626 (decide_peel_completely): Ditto.
23627 (decide_unroll_constant_iterations): Ditto.
23628 (decide_unroll_runtime_iterations): Ditto.
23629 (decide_peel_simple): Ditto.
23630 (decide_unroll_stupid): Ditto.
23631 * cfgloop.c (get_loop_location): New function.
23632 * cfgloop.h (get_loop_location): Declare.
23634 2013-01-02 Sriraman Tallam <tmsriram@google.com>
23636 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
23639 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23641 PR middle-end/55198
23642 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
23643 BLKmode objects when EXPAND_MEMORY is specified.
23645 2013-01-02 Sriraman Tallam <tmsriram@google.com>
23647 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
23649 (fold_builtin_cpu): Do not share cpu model decls across statements.
23651 2013-01-02 Jason Merrill <jason@redhat.com>
23654 * tree.c (build_array_type_1): Revert earlier change.
23656 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
23658 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
23660 * config/aarch64/aarch64-tune.md: Re-generate.
23662 2013-01-02 Richard Biener <rguenther@suse.de>
23664 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
23665 invariant load do not generate a vector load from the scalar location.
23667 2013-01-02 Richard Biener <rguenther@suse.de>
23670 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
23671 * configure: Regenerate.
23673 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
23675 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
23676 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
23677 (expand_builtin_int_roundingfn_2): Keep the original target around
23678 for the fallback case.
23680 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
23682 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
23683 to be clear for sign changes.
23685 2013-01-01 Jan Hubicka <jh@suse.cz>
23687 * ipa-inline-analysis.c: Fix formatting.
23689 2013-01-01 Jakub Jelinek <jakub@redhat.com>
23691 PR tree-optimization/55831
23692 * tree-vect-loop.c (get_initial_def_for_induction): Use
23693 gsi_after_labels instead of gsi_start_bb.
23695 Copyright (C) 2013 Free Software Foundation, Inc.
23697 Copying and distribution of this file, with or without modification,
23698 are permitted in any medium without royalty provided the copyright
23699 notice and this notice are preserved.