1 2013-11-15 Aldy Hernandez <aldyh@redhat.com>
3 * Makefile.in (C_COMMON_OBJS): Depend on c-cilkplus.o.
4 * gimple-pretty-print.c (dump_omp_for): Add case for
5 GF_OMP_FOR_KIND_CILKSIMD.
6 * gimple.h (enum gf_mask): Restructure entries to add
7 GF_OMP_FOR_KIND_CILKSIMD.
8 * gimplify.c (is_gimple_stmt): Add case for CILK_SIMD.
9 (gimplify_omp_for): Handle CILK_SIMD.
10 (gimplify_expr): Add ccase for CILK_SIMD.
11 * omp-low.c (extract_omp_for_data): Handle CILK_SIMD.
12 (build_outer_var_ref): Same.
13 (check_omp_nesting_restrictions): Same.
14 (lower_rec_input_clauses): Same.
15 (lower_lastprivate_clauses): Same.
16 (expand_omp_for): Same.
17 (execute_expand_omp): Check flag_enable_cilkplus.
18 (execute_lower_omp): Same.
19 (diagnose_sb_0): Handle CILK_SIMD.
20 (diagnose_omp_structured_block_errors): Check
22 (setjmp_or_longjmp_p): New.
23 (scan_omp_1_stmt): Error on setjmp/longjmp in a simd construct.
24 * tree-pretty-print.c (dump_generic_node): Add case for CILK_SIMD.
25 * tree.def: Add tree code for CILK_SIMD.
27 2013-11-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
29 * config/rs6000/altivec.md (UNSPEC_VPERM_X, UNSPEC_VPERM_UNS_X):
31 (altivec_vperm_<mode>): Revert earlier little endian change.
32 (*altivec_vperm_<mode>_internal): Remove.
33 (altivec_vperm_<mode>_uns): Revert earlier little endian change.
34 (*altivec_vperm_<mode>_uns_internal): Remove.
35 * config/rs6000/vector.md (vec_realign_load_<mode>): Revise
38 2013-11-15 Jeff Law <law@redhat.com>
40 * basic-block.h (has_abnormal_or_eh_outgoing_edge): Renamed from
41 has_abnormal_or_outgoing_edge. Check for EH edges as well.
42 * gimple-ssa-isolate-paths.c
43 (find_implicit_erroneous_behaviour): Corresponding changes.
44 Do not check stmt_ends_bb_p or GIMPLE_RETURN anymore.
45 (find_explicit_erroneous_behaviour): Likewise.
47 2013-11-15 Jeff Law <law@redhat.com>
49 * ifcvt.c (find_cond_trap): Properly handle case where
52 2013-11-15 Andreas Schwab <schwab@linux-m68k.org>
54 * configure: Regenerate.
56 2013-11-15 James Greenhalgh <james.greenhalgh@arm.com>
58 * config/aarch64/aarch64-simd.md: Remove simd_type from all
60 * config/aarch64/aarch64.md: Likewise, correct "type" attribute
61 where it is incorrect or missing.
63 2013-11-15 Richard Sandiford <rdsandiford@googlemail.com>
65 * dwarf2out.c (gen_enumeration_type_die): Remove unnecessary
67 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
68 Use TREE_INT_CST_LOW rather than tree_low_cst when reading the
70 * fold-const.c (fold_binary_loc): Replace a host_integerp/tree_low_cst
71 pair with a TREE_CODE test and TREE_INT_CST_LOW.
72 * tree-vect-generic.c (expand_vector_divmod): Likewise.
74 2013-11-15 Richard Biener <rguenther@suse.de>
76 PR tree-optimization/50262
77 * tree-ssa-alias.h (struct pt_solution): Split
78 vars_contains_global into vars_contains_nonlocal,
79 vars_contains_escaped and vars_contains_escaped_heap.
80 * tree-ssa-structalias.c (label_visit): Expand comment.
81 (handle_lhs_call): Adjust comment.
82 (set_uids_in_ptset): Set the new flags appropriately.
83 (pt_solution_set): Adjust.
84 (pt_solution_set_var): Likewise.
85 (pt_solution_ior_into): Likewise.
86 (pt_solution_includes_global): Likewise.
87 (pt_solutions_intersect_1): Optimize escaped handling.
88 (compute_points_to_sets): Remove heap variable globalization.
89 (ipa_escaped_pt): Adjust initializer.
90 (pass_data_ipa_pta): Do not run TODO_update_ssa.
91 * gimple-pretty-print.c (pp_points_to_solution): Print split
93 * tree-ssa-alias.c (dump_points_to_solution): Likewise.
95 2013-11-15 Richard Biener <rguenther@suse.de>
97 * tree-loop-distribution.c (tree_loop_distribution): Make sure
98 to distribute all stores.
100 2013-11-15 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
102 * doc/invoke.texi (-mabi=elfv1, -mabi=elfv2): Document.
104 2013-11-15 Joseph Myers <joseph@codesourcery.com>
106 * acinclude.m4 (GCC_GLIBC_VERSION_GTE_IFELSE): New configure
108 * configure.ac: Determine target_header_dir earlier.
109 (--with-glibc-version): New configure option.
110 Use GCC_GLIBC_VERSION_GTE_IFELSE in enable_gnu_unique_object,
111 gcc_cv_libc_provides_ssp and gcc_cv_target_ldbl128 tests.
112 * configure: Regenerate.
113 * doc/install.texi (--enable-gnu-unique-object): Don't refer to
114 native toolchains for default.
115 (--with-glibc-version): Document.
117 2013-11-15 Eric Botcazou <ebotcazou@adacore.com>
119 * fold-const.c (fold_binary_loc) <comparisons>: Reuse local variable.
121 2013-11-15 Uros Bizjak <ubizjak@gmail.com>
123 * lto-streamer-in.c (input function): Call cgraph_create_node if
124 cgraph_get_node failed.
126 2013-11-14 Olivier Hainque <hainque@adacore.com>
128 * cfgexpand.c (defer_stack_allocation): When optimization is enabled,
129 defer allocation of DECL_IGNORED_P variables at toplevel unless really
130 small. Factorize size threshold computation from the existing one.
131 (expand_used_vars): Refine comment.
133 2013-11-14 Cong Hou <congh@google.com>
135 * tree-vectorizer.h (struct dr_with_seg_len): Remove the base
136 address field as it can be obtained from dr. Rename the struct.
137 * tree-vect-data-refs.c (comp_dr_with_seg_len_pair): Consider
138 steps of data references during sort.
139 (vect_prune_runtime_alias_test_list): Adjust with the change to
140 struct dr_with_seg_len.
141 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
142 Adjust with the change to struct dr_with_seg_len.
144 2013-11-14 Jeff Law <law@redhat.com>
147 * basic-block.h (has_abnormal_outgoing_edge_p): Moved here from...
148 * tree-inline.c (has_abnormal_outgoing_edge_p): Remove.
149 * gimple-ssa-isolate-paths.c: Include tree-cfg.h.
150 (find_implicit_erroneous_behaviour): If a block has abnormal outgoing
151 edges, then ignore it. If the statement exhibiting erroneous
152 behaviour ends basic blocks, with the exception of GIMPLE_RETURNs,
153 then we can not optimize.
154 (find_explicit_erroneous_behaviour): Likewise.
156 2013-11-14 Andrew MacLeod <amacleod@redhat.com>
158 * gimplify-me.h: New file. Add prototypes.
159 * gimplify.h: Don't include gimple.h.
160 (struct gimplify_hasher, struct gimplify_ctx, is_gimple_sizepos):
161 Relocate from gimple.h.
162 * gimple.h (struct gimplify_hasher, struct gimplify_ctx,
163 is_gimple_sizepos): Move to gimplify.h.
164 (gimplify_hasher::hash, gimplify_hasher::equal): Move to gimplify.c.
165 (enum gsi_iterator_update): Move to gimple-iterator.h.
166 * gimple-iterator.h (enum gsi_iterator_update): Relocate from gimple.h.
167 * gimplify-me.c: New File.
168 (force_gimple_operand_1, force_gimple_operand,
169 force_gimple_operand_gsi_1, force_gimple_operand_gsi,
170 gimple_regimplify_operands): Relocate from gimplify.c.
171 * gimplify.c (force_gimple_operand_1, force_gimple_operand,
172 force_gimple_operand_gsi_1, force_gimple_operand_gsi,
173 gimple_regimplify_operands): Move to gimplify-me.c.
174 (gimplify_hasher::hash, gimplify_hasher::equal): Relocate
176 * Makefile.in (OBJS): Add gimplify-me.o
177 * asan.c: Include only gimplify.h, gimplify-me.h, and/or gimple.h as
179 * cfgloopmanip.c: Likewise.
180 * cgraphunit.c: Likewise.
181 * cilk-common.c: Likewise.
182 * fold-const.c: Likewise.
183 * function.c: Likewise.
184 * gimple-expr.c: Likewise.
185 * gimple-fold.c: Likewise.
186 * gimple-ssa-strength-reduction.c: Likewise.
187 * gimple.c: Likewise.
188 * graphite-clast-to-gimple.c: Likewise.
189 * graphite-sese-to-poly.c: Likewise.
190 * ipa-prop.c: Likewise.
191 * ipa-split.c: Likewise.
193 * langhooks.c: Likewise.
194 * omp-low.c: Likewise.
196 * stor-layout.c: Likewise.
197 * targhooks.c: Likewise.
198 * trans-mem.c: Likewise.
199 * tree-affine.c: Likewise.
200 * tree-cfg.c: Likewise.
201 * tree-cfgcleanup.c: Likewise.
202 * tree-complex.c: Likewise.
203 * tree-if-conv.c: Likewise.
204 * tree-inline.c: Likewise.
205 * tree-loop-distribution.c: Likewise.
206 * tree-nested.c: Likewise.
207 * tree-parloops.c: Likewise.
208 * tree-predcom.c: Likewise.
209 * tree-profile.c: Likewise.
210 * tree-scalar-evolution.c: Likewise.
211 * tree-sra.c: Likewise.
212 * tree-ssa-address.c: Likewise.
213 * tree-ssa-ccp.c: Likewise.
214 * tree-ssa-dce.c: Likewise.
215 * tree-ssa-forwprop.c: Likewise.
216 * tree-ssa-ifcombine.c: Likewise.
217 * tree-ssa-loop-im.c: Likewise.
218 * tree-ssa-loop-ivopts.c: Likewise.
219 * tree-ssa-loop-manip.c: Likewise.
220 * tree-ssa-loop-niter.c: Likewise.
221 * tree-ssa-loop-prefetch.c: Likewise.
222 * tree-ssa-loop-unswitch.c: Likewise.
223 * tree-ssa-math-opts.c: Likewise.
224 * tree-ssa-phiopt.c: Likewise.
225 * tree-ssa-phiprop.c: Likewise.
226 * tree-ssa-pre.c: Likewise.
227 * tree-ssa-propagate.c: Likewise.
228 * tree-ssa-reassoc.c: Likewise.
229 * tree-ssa-sccvn.c: Likewise.
230 * tree-ssa-strlen.c: Likewise.
231 * tree-ssa.c: Likewise.
232 * tree-switch-conversion.c: Likewise.
233 * tree-tailcall.c: Likewise.
234 * tree-vect-data-refs.c: Likewise.
235 * tree-vect-generic.c: Likewise.
236 * tree-vect-loop-manip.c: Likewise.
237 * tree-vect-loop.c: Likewise.
238 * tree-vect-patterns.c: Likewise.
239 * tree-vect-stmts.c: Likewise.
242 * value-prof.c: Likewise.
243 * config/aarch64/aarch64.c: Likewise.
244 * config/alpha/alpha.c: Likewise.
245 * config/darwin.c: Likewise.
246 * config/i386/i386.c: Likewise.
247 * config/ia64/ia64.c: Likewise.
248 * config/mep/mep.c: Likewise.
249 * config/mips/mips.c: Likewise.
250 * config/rs6000/rs6000.c: Likewise.
251 * config/s390/s390.c: Likewise.
252 * config/sh/sh.c: Likewise.
253 * config/sparc/sparc.c: Likewise.
254 * config/spu/spu.c: Likewise.
255 * config/stormy16/stormy16.c: Likewise.
256 * config/tilegx/tilegx.c: Likewise.
257 * config/tilepro/tilepro.c: Likewise.
258 * config/xtensa/xtensa.c: Likewise.
260 2013-11-14 Joern Rennecke <joern.rennecke@embecosm.com>
262 * config/arc/arc.md (doloop_begin_i): Remove extra alignment;
265 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
267 * config/rs6000/sysv4le.h (LINUX64_DEFAULT_ABI_ELFv2): Define.
269 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
270 Alan Modra <amodra@gmail.com>
272 * config/rs6000/rs6000.h (RS6000_SAVE_AREA): Handle ABI_ELFv2.
273 (RS6000_SAVE_TOC): Remove.
274 (RS6000_TOC_SAVE_SLOT): New macro.
275 * config/rs6000/rs6000.c (rs6000_parm_offset): New function.
276 (rs6000_parm_start): Use it.
277 (rs6000_function_arg_advance_1): Likewise.
278 (rs6000_emit_prologue): Use RS6000_TOC_SAVE_SLOT.
279 (rs6000_emit_epilogue): Likewise.
280 (rs6000_call_aix): Likewise.
281 (rs6000_output_function_prologue): Do not save/restore r11
282 around calling _mcount for ABI_ELFv2.
284 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
285 Alan Modra <amodra@gmail.com>
287 * config/rs6000/rs6000-protos.h (rs6000_reg_parm_stack_space):
289 * config/rs6000/rs6000.h (RS6000_REG_SAVE): Remove.
290 (REG_PARM_STACK_SPACE): Call rs6000_reg_parm_stack_space.
291 * config/rs6000/rs6000.c (rs6000_parm_needs_stack): New function.
292 (rs6000_function_parms_need_stack): Likewise.
293 (rs6000_reg_parm_stack_space): Likewise.
294 (rs6000_function_arg): Do not replace BLKmode by Pmode when
295 returning a register argument.
297 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
298 Michael Gschwind <mkg@us.ibm.com>
300 * config/rs6000/rs6000.h (FP_ARG_MAX_RETURN): New macro.
301 (ALTIVEC_ARG_MAX_RETURN): Likewise.
302 (FUNCTION_VALUE_REGNO_P): Use them.
303 * config/rs6000/rs6000.c (TARGET_RETURN_IN_MSB): Define.
304 (rs6000_return_in_msb): New function.
305 (rs6000_return_in_memory): Handle ELFv2 homogeneous aggregates.
306 Handle aggregates of up to 16 bytes for ELFv2.
307 (rs6000_function_value): Handle ELFv2 homogeneous aggregates.
309 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
310 Michael Gschwind <mkg@us.ibm.com>
312 * config/rs6000/rs6000.h (AGGR_ARG_NUM_REG): Define.
313 * config/rs6000/rs6000.c (rs6000_aggregate_candidate): New function.
314 (rs6000_discover_homogeneous_aggregate): Likewise.
315 (rs6000_function_arg_boundary): Handle homogeneous aggregates.
316 (rs6000_function_arg_advance_1): Likewise.
317 (rs6000_function_arg): Likewise.
318 (rs6000_arg_partial_bytes): Likewise.
319 (rs6000_psave_function_arg): Handle BLKmode arguments.
321 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
322 Michael Gschwind <mkg@us.ibm.com>
324 * config/rs6000/rs6000.h (AGGR_ARG_NUM_REG): Define.
325 * config/rs6000/rs6000.c (rs6000_aggregate_candidate): New function.
326 (rs6000_discover_homogeneous_aggregate): Likewise.
327 (rs6000_function_arg_boundary): Handle homogeneous aggregates.
328 (rs6000_function_arg_advance_1): Likewise.
329 (rs6000_function_arg): Likewise.
330 (rs6000_arg_partial_bytes): Likewise.
331 (rs6000_psave_function_arg): Handle BLKmode arguments.
333 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
335 * config/rs6000/rs6000.c (machine_function): New member
337 (rs6000_emit_prologue): Set r2_setup_needed if necessary.
338 (rs6000_output_mi_thunk): Set r2_setup_needed.
339 (rs6000_output_function_prologue): Output global entry point
340 prologue and local entry point marker if needed for ABI_ELFv2.
341 Output -mprofile-kernel code here.
342 (output_function_profiler): Do not output -mprofile-kernel
343 code here; moved to rs6000_output_function_prologue.
344 (rs6000_file_start): Output ".abiversion 2" for ABI_ELFv2.
346 (rs6000_emit_move): Do not handle dot symbols for ABI_ELFv2.
347 (rs6000_output_function_entry): Likewise.
348 (rs6000_assemble_integer): Likewise.
349 (rs6000_elf_encode_section_info): Likewise.
350 (rs6000_elf_declare_function_name): Do not create dot symbols
351 or .opd section for ABI_ELFv2.
353 (rs6000_trampoline_size): Update for ABI_ELFv2 trampolines.
354 (rs6000_trampoline_init): Likewise.
355 (rs6000_elf_file_end): Call file_end_indicate_exec_stack for ABI_ELFv2.
357 (rs6000_call_aix): Handle ELFv2 indirect calls. Do not check
358 for function descriptors in ABI_ELFv2.
360 * config/rs6000/rs6000.md ("*call_indirect_aix<mode>"): Support
361 on ABI_AIX only, not ABI_ELFv2.
362 ("*call_value_indirect_aix<mode>"): Likewise.
363 ("*call_indirect_elfv2<mode>"): New pattern.
364 ("*call_value_indirect_elfv2<mode>"): Likewise.
366 * config/rs6000/predicates.md ("symbol_ref_operand"): Do not
367 check for function descriptors in ABI_ELFv2.
368 ("current_file_function_operand"): Likewise.
370 * config/rs6000/ppc-asm.h [__powerpc64__ && _CALL_ELF == 2]:
372 (FUNC_NAME): Define ELFv2 variant.
373 (JUMP_TARGET): Likewise.
374 (FUNC_START): Likewise.
375 (HIDDEN_FUNC): Likewise.
376 (FUNC_END): Likeiwse.
378 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
380 * config.gcc [powerpc*-*-* | rs6000-*-*]: Support --with-abi=elfv1
381 and --with-abi=elfv2.
382 * config/rs6000/option-defaults.h (OPTION_DEFAULT_SPECS): Add "abi".
383 * config/rs6000/rs6000.opt (mabi=elfv1): New option.
384 (mabi=elfv2): Likewise.
385 * config/rs6000/rs6000-opts.h (enum rs6000_abi): Add ABI_ELFv2.
386 * config/rs6000/linux64.h (DEFAULT_ABI): Do not hard-code to AIX_ABI
388 (ELFv2_ABI_CHECK): New macro.
389 (SUBSUBTARGET_OVERRIDE_OPTIONS): Use it to decide whether to set
390 rs6000_current_abi to ABI_AIX or ABI_ELFv2.
391 (GLIBC_DYNAMIC_LINKER64): Support ELFv2 ld.so version.
392 * config/rs6000/rs6000-c.c (rs6000_cpu_cpp_builtins): Predefine
393 _CALL_ELF and __STRUCT_PARM_ALIGN__ if appropriate.
395 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Handle ABI_ELFv2.
396 (debug_stack_info): Likewise.
397 (rs6000_file_start): Treat ABI_ELFv2 the same as ABI_AIX.
398 (rs6000_legitimize_tls_address): Likewise.
399 (rs6000_conditional_register_usage): Likewise.
400 (rs6000_emit_move): Likewise.
401 (init_cumulative_args): Likewise.
402 (rs6000_function_arg_advance_1): Likewise.
403 (rs6000_function_arg): Likewise.
404 (rs6000_arg_partial_bytes): Likewise.
405 (rs6000_output_function_entry): Likewise.
406 (rs6000_assemble_integer): Likewise.
407 (rs6000_savres_strategy): Likewise.
408 (rs6000_stack_info): Likewise.
409 (rs6000_function_ok_for_sibcall): Likewise.
410 (rs6000_emit_load_toc_table): Likewise.
411 (rs6000_savres_routine_name): Likewise.
412 (ptr_regno_for_savres): Likewise.
413 (rs6000_emit_prologue): Likewise.
414 (rs6000_emit_epilogue): Likewise.
415 (rs6000_output_function_epilogue): Likewise.
416 (output_profile_hook): Likewise.
417 (output_function_profiler): Likewise.
418 (rs6000_trampoline_size): Likewise.
419 (rs6000_trampoline_init): Likewise.
420 (rs6000_elf_output_toc_section_asm_op): Likewise.
421 (rs6000_elf_encode_section_info): Likewise.
422 (rs6000_elf_reloc_rw_mask): Likewise.
423 (rs6000_elf_declare_function_name): Likewise.
424 (rs6000_function_arg_boundary): Treat ABI_ELFv2 the same as ABI_AIX,
425 except that rs6000_compat_align_parm is always assumed false.
426 (rs6000_gimplify_va_arg): Likewise.
427 (rs6000_call_aix): Update comment.
428 (rs6000_sibcall_aix): Likewise.
429 * config/rs6000/rs6000.md ("tls_gd_aix<TLSmode:tls_abi_suffix>"):
430 Treat ABI_ELFv2 the same as ABI_AIX.
431 ("*tls_gd_call_aix<TLSmode:tls_abi_suffix>"): Likewise.
432 ("tls_ld_aix<TLSmode:tls_abi_suffix>"): Likewise.
433 ("*tls_ld_call_aix<TLSmode:tls_abi_suffix>"): Likewise.
434 ("load_toc_aix_si"): Likewise.
435 ("load_toc_aix_di"): Likewise.
437 ("call_value"): Likewise.
438 ("*call_local_aix<mode>"): Likewise.
439 ("*call_value_local_aix<mode>"): Likewise.
440 ("*call_nonlocal_aix<mode>"): Likewise.
441 ("*call_value_nonlocal_aix<mode>"): Likewise.
442 ("*call_indirect_aix<mode>"): Likewise.
443 ("*call_value_indirect_aix<mode>"): Likewise.
444 ("sibcall"): Likewise.
445 ("sibcall_value"): Likewise.
446 ("*sibcall_aix<mode>"): Likewise.
447 ("*sibcall_value_aix<mode>"): Likewise.
448 * config/rs6000/predicates.md ("symbol_ref_operand"): Likewise.
449 ("current_file_function_operand"): Likewise.
451 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
453 * config/rs6000/rs6000.c (rs6000_arg_partial_bytes): Simplify logic
454 by making use of the fact that for vector / floating point arguments
455 passed both in VRs/FPRs and in the fixed parameter area, the partial
456 bytes mechanism is in fact not used.
458 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
460 * config/rs6000/rs6000.c (rs6000_psave_function_arg): New function.
461 (rs6000_finish_function_arg): Likewise.
462 (rs6000_function_arg): Use rs6000_psave_function_arg and
463 rs6000_finish_function_arg to handle both vector and floating
464 point arguments that are also passed in GPRs / the stack.
466 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
468 * config/rs6000/rs6000.c (USE_FP_FOR_ARG_P): Remove TYPE argument.
469 (USE_ALTIVEC_FOR_ARG_P): Likewise.
470 (rs6000_darwin64_record_arg_advance_recurse): Update uses.
471 (rs6000_function_arg_advance_1):Likewise.
472 (rs6000_darwin64_record_arg_recurse): Likewise.
473 (rs6000_function_arg): Likewise.
474 (rs6000_arg_partial_bytes): Likewise.
476 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
478 * config/rs6000/rs6000.c (rs6000_option_override_internal): Replace
479 "DEFAULT_ABI != ABI_AIX" test by testing for ABI_V4 or ABI_DARWIN.
480 (rs6000_savres_strategy): Likewise.
481 (rs6000_return_addr): Likewise.
482 (rs6000_emit_load_toc_table): Replace "DEFAULT_ABI != ABI_AIX" by
483 testing for ABI_V4 (since ABI_DARWIN is impossible here).
484 (rs6000_emit_prologue): Likewise.
485 (legitimate_lo_sum_address_p): Simplify DEFAULT_ABI test.
486 (rs6000_elf_declare_function_name): Remove duplicated test.
487 * config/rs6000/rs6000.md ("load_toc_v4_PIC_1"): Explicitly test
488 for ABI_V4 (instead of "DEFAULT_ABI != ABI_AIX" test).
489 ("load_toc_v4_PIC_1_normal"): Likewise.
490 ("load_toc_v4_PIC_1_476"): Likewise.
491 ("load_toc_v4_PIC_1b"): Likewise.
492 ("load_toc_v4_PIC_1b_normal"): Likewise.
493 ("load_toc_v4_PIC_1b_476"): Likewise.
494 ("load_toc_v4_PIC_2"): Likewise.
495 ("load_toc_v4_PIC_3b"): Likewise.
496 ("load_toc_v4_PIC_3c"): Likewise.
497 * config/rs6000/rs6000.h (RS6000_REG_SAVE): Simplify DEFAULT_ABI test.
498 (RS6000_SAVE_AREA): Likewise.
499 (FP_ARG_MAX_REG): Likewise.
500 (RETURN_ADDRESS_OFFSET): Likewise.
501 * config/rs6000/sysv.h (TARGET_TOC): Test for ABI_V4 instead
503 (SUBTARGET_OVERRIDE_OPTIONS): Likewise.
504 (MINIMAL_TOC_SECTION_ASM_OP): Likewise.
506 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
508 * config/rs6000/rs6000.c (rs6000_call_indirect_aix): Rename to ...
509 (rs6000_call_aix): ... this. Handle both direct and indirect calls.
510 Create call insn directly instead of via various gen_... routines.
511 Mention special registers used by the call in CALL_INSN_FUNCTION_USAGE.
512 (rs6000_sibcall_aix): New function.
513 * config/rs6000/rs6000.md (TOC_SAVE_OFFSET_32BIT): Remove.
514 (TOC_SAVE_OFFSET_64BIT): Likewise.
515 (AIX_FUNC_DESC_TOC_32BIT): Likewise.
516 (AIX_FUNC_DESC_TOC_64BIT): Likewise.
517 (AIX_FUNC_DESC_SC_32BIT): Likewise.
518 (AIX_FUNC_DESC_SC_64BIT): Likewise.
519 ("call" expander): Call rs6000_call_aix.
520 ("call_value" expander): Likewise.
521 ("call_indirect_aix<ptrsize>"): Replace this pattern ...
522 ("call_indirect_aix<ptrsize>_nor11"): ... and this pattern ...
523 ("*call_indirect_aix<mode>"): ... by this insn pattern.
524 ("call_value_indirect_aix<ptrsize>"): Replace this pattern ...
525 ("call_value_indirect_aix<ptrsize>_nor11"): ... and this pattern ...
526 ("*call_value_indirect_aix<mode>"): ... by this insn pattern.
527 ("*call_nonlocal_aix32", "*call_nonlocal_aix64"): Replace by ...
528 ("*call_nonlocal_aix<mode>"): ... this pattern.
529 ("*call_value_nonlocal_aix32", "*call_value_nonlocal_aix64"): Replace
530 ("*call_value_nonlocal_aix<mode>"): ... by this pattern.
531 ("*call_local_aix<mode>"): New insn pattern.
532 ("*call_value_local_aix<mode>"): Likewise.
533 ("sibcall" expander): Call rs6000_sibcall_aix.
534 ("sibcall_value" expander): Likewise. Move earlier in file.
535 ("*sibcall_nonlocal_aix<mode>"): Replace by ...
536 ("*sibcall_aix<mode>"): ... this pattern.
537 ("*sibcall_value_nonlocal_aix<mode>"): Replace by ...
538 ("*sibcall_value_aix<mode>"): ... this pattern.
539 * config/rs6000/rs6000-protos.h (rs6000_call_indirect_aix): Remove.
540 (rs6000_call_aix): Add prototype.
541 (rs6000_sibcall_aix): Likewise.
543 2013-11-14 Jakub Jelinek <jakub@redhat.com>
546 * asan.c (asan_emit_stack_protection): Ensure -fsection-anchors
547 isn't confused by the artificial decl.
549 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
551 * config/rs6000/rs6000.c (rs6000_emit_prologue): Do not place a
552 RTX_FRAME_RELATED_P marker on the UNSPEC_MOVESI_FROM_CR insn.
553 Instead, add USEs of all modified call-saved CR fields to the
554 insn storing the result to the stack slot, and provide an
555 appropriate REG_FRAME_RELATED_EXPR for that insn.
556 * config/rs6000/rs6000.md ("*crsave"): New insn pattern.
557 * config/rs6000/predicates.md ("crsave_operation"): New predicate.
559 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
560 Alan Modra <amodra@gmail.com>
562 * function.c (assign_parms): Use all.reg_parm_stack_space instead
563 of re-evaluating REG_PARM_STACK_SPACE target macro.
564 (locate_and_pad_parm): New parameter REG_PARM_STACK_SPACE. Use it
565 instead of evaluating target macro REG_PARM_STACK_SPACE every time.
566 (assign_parm_find_entry_rtl): Update call.
567 * calls.c (initialize_argument_information): Update call.
568 (emit_library_call_value_1): Likewise.
569 * expr.h (locate_and_pad_parm): Update prototype.
571 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
573 * calls.c (store_unaligned_arguments_into_pseudos): Skip PARALLEL
576 2013-11-14 DJ Delorie <dj@redhat.com>
578 * config/rx/rx.c (rx_mode_dependent_address_p): Allow offsets up
581 2013-11-14 Jeff Law <law@redhat.com>
583 * tree-ssa-threadedge.c (thread_through_normal_block): Only push the
584 EDGE_START_JUMP_THREAD marker if the jump threading path is empty.
586 2013-11-14 James Greenhalgh <james.greenhalgh@arm.com>
588 * doc/invoke.texi: Update documentation for AArch64's -mcpu
591 2013-11-14 James Greenhalgh <james.greenhalgh@arm.com>
593 * config/aarch64/aarch64-cores.def (example-1): Remove.
594 (example-2): Likewise.
595 * config/aarch64/aarch64-tune.md: Regenerate.
596 * config/aarch64/aarch64.md: Do not include "large.md" or "small.md".
597 (generic_sched): Remove "large", "small".
598 * config/aarch64/large.md: Delete.
599 * config/aarch64/small.md: Delete.
601 2013-11-14 James Greenhalgh <james.greenhalgh@arm.com>
603 * config/aarch64/aarch64-cores.def (cortex-a57): Tune for cortexa15.
604 * config/aarch64/aarch64-tune.md: Regenerate.
605 * config/aarch64/aarch64.md: Include cortex-a15 pipeline model.
606 (generic_sched): "no" if we are tuning for cortexa15.
607 * config/arm/cortex-a15.md: Include cortex-a15-neon.md by
610 2013-11-14 James Greenhalgh <james.greenhalgh@arm.com>
612 * config/aarch64/aarch64-arches.def (armv8-a): Tune for cortex-a53.
613 * config/aarch64/aarch64.md: Do not include aarch64-generic.md.
614 * config/aarch64/aarch64.c (aarch64_tune): Initialize to cortexa53.
615 (all_cores): Use cortexa53 when tuning for "generic".
616 (aarch64_override_options): Fix comment.
617 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Set to cortexa53.
618 * config/aarch64/aarch64-generic.md: Delete.
620 2013-11-14 James Greenhalgh <james.greenhalgh@arm.com>
622 * config/aarch64/aarch64.c (all_architectures): Remove "generic".
624 2013-11-14 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
626 * config/aarch64/aarch64.c: Include aarch-cost-tables.h.
627 (generic_rtx_cost_table): Remove.
628 (aarch64_rtx_costs): Use fields from cpu_cost_table.
629 * config/aarch64/aarch64-protos.h (tune_params): Use cpu_cost_table for
631 (cpu_rtx_cost_table): Remove.
633 2013-11-14 Julian Brown <julian@codesourcery.com>
634 Joey Ye <joey.ye@arm.com>
636 * config/arm/arm.c (arm_cortex_m_branch_cost): New.
638 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune, arm_9e_tune,
639 arm_v6t2_tune, arm_cortex_tune, arm_cortex_a15_tune,
640 arm_cortex_a5_tune, arm_v6m_tune): Add comments for Sched adj cost.
641 * config/arm/arm-cores.def (cortex-m4, cortex-m3): Use arm_v7m_tune.
643 2013-11-14 Kirill Yukhin <kirill.yukhin@intel.com>
646 * config/ia64/ia64.c (ia64_split_tmode_move): Relax `dead'
649 2013-11-14 Jakub Jelinek <jakub@redhat.com>
650 Uros Bizjak <ubizjak@gmail.com>
653 * config/i386/i386.md (*anddi_2): Only allow CCZmode if
654 operands[2] satisfies_constraint_Z that might have bit 31 set.
656 2013-11-13 Jeff Law <law@redhat.com>
658 PR tree-optimization/59102
659 * gimple-ssa-isolate-paths.c
660 (insert_trap_and_remove_trailing_statments): Ensure STMT is a
661 gimple assignment before looking at gimple_assign_lhs.
663 2013-11-13 Vladimir Makarov <vmakarov@redhat.com>
665 * ira.c: Add comment about threads at the top of file.
667 2013-11-13 Vladimir Makarov <vmakarov@redhat.com>
669 * ira-color.c (coalesce_allocnos): Don't allocate and free
672 2013-11-14 Tom de Vries <tom@codesourcery.com>
674 * tree-ssa-tail-merge.c (gimple_equal_p): Add test for structural
675 equality for GIMPLE_ASSIGN.
677 2013-11-14 Tom de Vries <tom@codesourcery.com>
679 * tree-ssa-tail-merge.c (gimple_operand_equal_value_p): Factor new
681 (gimple_equal_p): ... here.
683 2013-11-14 Tom de Vries <tom@codesourcery.com>
685 * trans-mem.c (is_tm_ending): New function.
686 * gimple.h (is_tm_ending): Declare.
687 * tree-ssa-tail-merge.c (gimple_equal_p): Remove test on
689 (find_duplicate): Use is_tm_ending instead of is_tm_ending_fndecl.
691 2013-11-14 Tom de Vries <tom@codesourcery.com>
693 * tree-ssa-tail-merge.c (gimple_equal_p): Remove equal variable.
695 2013-11-13 Andrew MacLeod <amacleod@redhat.com>
697 * gimple-walk.h: New File. Relocate prototypes from gimple.h.
698 (struct walk_stmt_info): Relocate here from gimple.h.
699 * gimple-iterator.h: New File. Relocate prototypes from gimple.h.
700 (struct gimple_stmt_iterator_d): Relocate here from gimple.h.
701 (gsi_start_1, gsi_none, gsi_start_bb, gsi_last_1, gsi_last_bb,
702 gsi_end_p, gsi_one_before_end_p, gsi_next, gsi_prev, gsi_stmt,
703 gsi_after_labels, gsi_next_nondebug, gsi_prev_nondebug,
704 gsi_start_nondebug_bb, gsi_start_nondebug_after_labels_bb,
705 gsi_last_nondebug_bb, gsi_bb, gsi_seq): Relocate here from gimple.h.
706 * gimple.h (struct gimple_stmt_iterator_d): Move to gimple-iterator.h.
707 (gsi_start_1, gsi_none, gsi_start_bb, gsi_last_1, gsi_last_bb,
708 gsi_end_p, gsi_one_before_end_p, gsi_next, gsi_prev, gsi_stmt,
709 gsi_after_labels, gsi_next_nondebug, gsi_prev_nondebug,
710 gsi_start_nondebug_bb, gsi_start_nondebug_after_labels_bb,
711 gsi_last_nondebug_bb, gsi_bb, gsi_seq): Move to gimple-iterator.h.
712 (struct walk_stmt_info): Move to gimple-walk.h.
713 (gimple_seq_set_location): Move to gimple.c
714 * gimple-walk.c: New File.
715 (walk_gimple_seq_mod, walk_gimple_seq, walk_gimple_asm, walk_gimple_op,
716 walk_gimple_stmt, get_base_loadstore, walk_stmt_load_store_addr_ops,
717 walk_stmt_load_store_ops): Relocate here from gimple.c.
718 * gimple-iterator.c: Include gimple-iterator.h.
719 * gimple.c (walk_gimple_seq_mod, walk_gimple_seq, walk_gimple_asm,
720 walk_gimple_op, walk_gimple_stmt, get_base_loadstore,
721 walk_stmt_load_store_addr_ops, walk_stmt_load_store_ops): Move to
723 (gimple_seq_set_location): Relocate from gimple.h.
724 * tree-phinodes.h (set_phi_nodes): Move to tree-phinodes.c.
725 * tree-phinodes.c (set_phi_nodes): Relocate from tree-phinodes.h.
726 * gengtype.c (open_base_files): Add gimple-iterator.h to include list.
727 * Makefile.in (OBJS): Add gimple-walk.o
728 * asan.c: Update Include list as required for gimple-iterator.h and
730 * cfgexpand.c: Likewise.
731 * cfgloop.c: Likewise.
732 * cfgloopmanip.c: Likewise.
733 * cgraph.c: Likewise.
734 * cgraphbuild.c: Likewise.
735 * cgraphunit.c: Likewise.
736 * gimple-fold.c: Likewise.
737 * gimple-low.c: Likewise.
738 * gimple-pretty-print.c: Likewise.
739 * gimple-ssa-isolate-paths.c: Likewise.
740 * gimple-ssa-strength-reduction.c: Likewise.
741 * gimple-streamer-in.c: Likewise.
742 * gimple-streamer-out.c: Likewise.
743 * gimplify.c: Likewise.
744 * graphite-blocking.c: Likewise.
745 * graphite-clast-to-gimple.c: Likewise.
746 * graphite-dependences.c: Likewise.
747 * graphite-interchange.c: Likewise.
748 * graphite-optimize-isl.c: Likewise.
749 * graphite-poly.c: Likewise.
750 * graphite-scop-detection.c: Likewise.
751 * graphite-sese-to-poly.c: Likewise.
752 * graphite.c: Likewise.
753 * ipa-inline-analysis.c: Likewise.
754 * ipa-profile.c: Likewise.
755 * ipa-prop.c: Likewise.
756 * ipa-pure-const.c: Likewise.
757 * ipa-split.c: Likewise.
758 * lto-streamer-in.c: Likewise.
759 * lto-streamer-out.c: Likewise.
760 * omp-low.c: Likewise.
761 * predict.c: Likewise.
762 * profile.c: Likewise.
764 * tracer.c: Likewise.
765 * trans-mem.c: Likewise.
766 * tree-call-cdce.c: Likewise.
767 * tree-cfg.c: Likewise.
768 * tree-cfgcleanup.c: Likewise.
769 * tree-complex.c: Likewise.
770 * tree-data-ref.c: Likewise.
771 * tree-dfa.c: Likewise.
772 * tree-eh.c: Likewise.
773 * tree-emutls.c: Likewise.
774 * tree-if-conv.c: Likewise.
775 * tree-inline.c: Likewise.
776 * tree-into-ssa.c: Likewise.
777 * tree-loop-distribution.c: Likewise.
778 * tree-nested.c: Likewise.
779 * tree-nrv.c: Likewise.
780 * tree-object-size.c: Likewise.
781 * tree-outof-ssa.c: Likewise.
782 * tree-parloops.c: Likewise.
783 * tree-predcom.c: Likewise.
784 * tree-profile.c: Likewise.
785 * tree-scalar-evolution.c: Likewise.
786 * tree-sra.c: Likewise.
787 * tree-ssa-ccp.c: Likewise.
788 * tree-ssa-coalesce.c: Likewise.
789 * tree-ssa-copy.c: Likewise.
790 * tree-ssa-copyrename.c: Likewise.
791 * tree-ssa-dce.c: Likewise.
792 * tree-ssa-dom.c: Likewise.
793 * tree-ssa-dse.c: Likewise.
794 * tree-ssa-forwprop.c: Likewise.
795 * tree-ssa-ifcombine.c: Likewise.
796 * tree-ssa-live.c: Likewise.
797 * tree-ssa-loop-ch.c: Likewise.
798 * tree-ssa-loop-im.c: Likewise.
799 * tree-ssa-loop-ivcanon.c: Likewise.
800 * tree-ssa-loop-ivopts.c: Likewise.
801 * tree-ssa-loop-manip.c: Likewise.
802 * tree-ssa-loop-niter.c: Likewise.
803 * tree-ssa-loop-prefetch.c: Likewise.
804 * tree-ssa-loop.c: Likewise.
805 * tree-ssa-math-opts.c: Likewise.
806 * tree-ssa-phiopt.c: Likewise.
807 * tree-ssa-phiprop.c: Likewise.
808 * tree-ssa-pre.c: Likewise.
809 * tree-ssa-propagate.c: Likewise.
810 * tree-ssa-reassoc.c: Likewise.
811 * tree-ssa-sink.c: Likewise.
812 * tree-ssa-strlen.c: Likewise.
813 * tree-ssa-structalias.c: Likewise.
814 * tree-ssa-tail-merge.c: Likewise.
815 * tree-ssa-ter.c: Likewise.
816 * tree-ssa-threadedge.c: Likewise.
817 * tree-ssa-threadupdate.c: Likewise.
818 * tree-ssa-uncprop.c: Likewise.
819 * tree-ssa-uninit.c: Likewise.
820 * tree-ssa.c: Likewise.
821 * tree-stdarg.c: Likewise.
822 * tree-switch-conversion.c: Likewise.
823 * tree-tailcall.c: Likewise.
824 * tree-vect-data-refs.c: Likewise.
825 * tree-vect-generic.c: Likewise.
826 * tree-vect-loop-manip.c: Likewise.
827 * tree-vect-loop.c: Likewise.
828 * tree-vect-patterns.c: Likewise.
829 * tree-vect-slp.c: Likewise.
830 * tree-vect-stmts.c: Likewise.
831 * tree-vectorizer.c: Likewise.
832 * tree-vrp.c: Likewise.
835 * value-prof.c: Likewise.
836 * vtable-verify.c: Likewise.
838 2013-11-13 Steven Bosscher <steven@gcc.gnu.org>
840 * gimple-ssa-isolate-paths.c (pass_isolate_erroneous_paths): Comment
843 2013-11-13 Jeff Law <law@redhat.com>
845 * PR middle-end/59119
846 * gimple-ssa-isolate-paths.c (find_implicit_erroneous_behaviour): New
847 function, extracted from gimple_ssa_isolate_erroneous_paths.
848 (find_explicit_erroneous_behaviour): Similarly.
849 (insert_trap_and_remove_trailing_statements): Remove statements
852 2013-11-13 Steven Bosscher <steven@gcc.gnu.org>
854 * cfgrtl.c (can_fallthru): Reorder code to move tablejump check up.
855 Make that check explicit. BB_HEAD cannot be NULL, remove check for it.
856 * haifa-sched.c (ready_remove_first_dispatch): Check INSN_P before
857 looking at INSN_CODE.
858 * reload1.c (delete_dead_insn) Do not expect JUMP_TABLE_DATA to be an
859 active_insn_p object, respect basic block boundaries.
860 * reorg.c (follow_jumps): Use invariant that JUMP_TABLE_DATA always
861 follows immediately after the jump table data label.
862 * config/nds32/nds32.c (nds32_output_casesi_pc_relative): Likewise.
863 * config/sh/sh.c (barrier_align): Likewise. Rearrange code such
864 that JUMP_TABLE_DATA is not expected to be an active_insn_p object.
866 2013-11-13 Teresa Johnson <tejohnson@google.com>
869 * predict.c (drop_profile): Error is currently too strict.
870 (handle_missing_profiles): Pass call_count to drop_profile.
872 2013-11-13 Teresa Johnson <tejohnson@google.com>
875 * ipa-inline.c (edge_badness): Fix overflow.
877 2013-11-13 Vladimir Makarov <vmakarov@redhat.com>
879 PR rtl-optimization/59036
880 * ira-color.c (struct allocno_color_data): Add new members
881 first_thread_allocno, next_thread_allocno, thread_freq.
882 (sorted_copies): New static var.
883 (allocnos_conflict_by_live_ranges_p, copy_freq_compare_func): Move up.
884 (allocno_thread_conflict_p, merge_threads)
885 (form_threads_from_copies, form_threads_from_bucket)
886 (form_threads_from_colorable_allocno, init_allocno_threads): New
888 (bucket_allocno_compare_func): Add comparison by thread frequency
890 (add_allocno_to_ordered_bucket): Rename to
891 add_allocno_to_ordered_colorable_bucket. Remove parameter.
892 (push_only_colorable): Call form_threads_from_bucket.
893 (color_pass): Call init_allocno_threads. Use
894 consideration_allocno_bitmap instead of coloring_allocno_bitmap
895 for nuillify allocno color data.
896 (ira_initiate_assign, ira_finish_assign): Allocate/free sorted_copies.
897 (coalesce_allocnos): Use static sorted copies.
899 2013-11-13 Jakub Jelinek <jakub@redhat.com>
901 * passes.c (execute_todo): Don't call do_per_function if
903 (execute_one_ipa_transform_pass, execute_one_pass): Don't call
904 execute_function_dump if dump_file is NULL.
906 2013-11-13 Martin Jambor <mjambor@suse.cz>
908 * cgraph.c (cgraph_get_create_node): Do what
909 cgraph_get_create_real_symbol_node used to do.
910 (cgraph_get_create_real_symbol_node): Removed. Changed all users to
911 call cgraph_get_create_node.
912 * cgraph.h (cgraph_get_create_real_symbol_node): Removed.
913 * lto-streamer-in.c (input_function): Call cgraph_get_node instead of
914 cgraph_get_create_node. Assert we get a node.
916 2013-11-13 Tejas Belagod <tejas.belagod@arm.com>
918 * config/aarch64/aarch64-simd.md (vec_extract): New.
920 2013-11-13 Tejas Belagod <tejas.belagod@arm.com>
922 * config/aarch64/aarch64-simd.md (vec_set<mode>): Add w -> w option to
925 2013-11-13 Eric Botcazou <ebotcazou@adacore.com>
927 * cfgexpand.c (expand_used_vars): Allocate space for partitions based
928 on PARM_DECLs or RESULT_DECLs only if they are ignored for debug info
929 or if optimization is enabled.
930 * tree-ssa-coalesce.c (coalesce_ssa_name): If optimization is disabled,
931 require that all the names based on a PARM_DECL or a RESULT_DECL that
932 isn't ignored for debug info be coalesced.
934 2013-11-13 Jan-Benedict Glaw <jbglaw@lug-owl.de>
936 * config/c6x/c6x.c: Include "gimple-expr.h".
938 2013-11-13 Richard Biener <rguenther@suse.de>
940 * gimple-streamer-out.c (output_gimple_stmt): Also wrap
941 decls in ADDR_EXPR operands inside a MEM_REF and optimize that.
942 * gimple-streamer-in.c (input_gimple_stmt): Remove now dead code
943 dealing with type mismatches inside component reference chains.
945 2013-11-13 Marc Glisse <marc.glisse@inria.fr>
947 PR tree-optimization/59077
948 * ipa-pure-const.c (better_state): Update *state.
950 2013-11-13 Christophe Lyon <christophe.lyon@linaro.org>
952 * config/aarch64/aarch64.h (FRAME_GROWS_DOWNWARD): Define to 1.
953 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
954 Update offset calculations.
956 2013-11-13 Eric Botcazou <ebotcazou@adacore.com>
959 * dwarf2out.c (add_byte_size_attribute): Also use int_size_in_bytes
960 for fields. Do not add the attribute if the size is negative.
962 2013-11-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
964 * config/arm/arm.c: Include aarch-cost-tables.h.
965 (generic_extra_costs): Move from here...
966 * config/arm/aarch-cost-tables.h: ... To here. New file.
968 2013-11-13 Alexander Ivchenko <alexander.ivchenko@intel.com>
969 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
970 Sergey Lega <sergey.s.lega@intel.com>
971 Anna Tikhonova <anna.tikhonova@intel.com>
972 Ilya Tocar <ilya.tocar@intel.com>
973 Andrey Turetskiy <andrey.turetskiy@intel.com>
974 Ilya Verbin <ilya.verbin@intel.com>
975 Kirill Yukhin <kirill.yukhin@intel.com>
976 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
978 * config/i386/i386.c (ix86_print_operand): Support z-masking.
979 * config/i386/predicate.md (const_0_to_4_operand): New.
980 (const_0_to_5_operand): Ditto.
981 * config/i386/sse.md (UNSPEC_COMPRESS): New.
982 (UNSPEC_COMPRESS_STORE): Ditto.
983 (UNSPEC_EXPAND): Ditto.
984 (UNSPEC_EMBEDDED_ROUNDING): Ditto.
985 (define_mode_attr ssescalarsize): Ditto.
986 (avx512f_load<mode>_mask): Ditto.
987 (avx512f_store<mode>_mask): Ditto.
988 (avx512f_storedqu<mode>_mask): Ditto.
989 (avx512f_vmcmp<mode>3_mask): Ditto.
990 (avx512f_fmadd_<mode>_mask): Ditto.
991 (avx512f_fmadd_<mode>_mask3): Ditto.
992 (avx512f_fmsub_<mode>_mask): Ditto.
993 (avx512f_fmsub_<mode>_mask3): Ditto.
994 (avx512f_fnmadd_<mode>_mask): Ditto.
995 (avx512f_fnmadd_<mode>_mask3): Ditto.
996 (avx512f_fnmsub_<mode>_mask): Ditto.
997 (avx512f_fnmsub_<mode>_mask3): Ditto.
998 (avx512f_fmaddsub_<mode>_mask): Ditto.
999 (avx512f_fmaddsub_<mode>_mask3): Ditto.
1000 (avx512f_fmsubadd_<mode>_mask): Ditto.
1001 (avx512f_fmsubadd_<mode>_mask3): Ditto.
1002 (vec_unpacku_float_lo_v16si): Ditto.
1003 (avx512f_vextract<shuffletype>32x4_mask): Ditto.
1004 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
1005 (avx512f_vextract<shuffletype>64x4_mask): Ditto.
1006 (vec_extract_lo_<mode>_maskm): Ditto.
1007 (vec_extract_hi_<mode>_maskm): Ditto.
1008 (avx512f_vternlog<mode>_mask): Ditto.
1009 (avx512f_shufps512_mask): Ditto.
1010 (avx512f_fixupimm<mode>_mask): Ditto.
1011 (avx512f_shufpd512_mask): Ditto.
1012 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
1013 (avx512f_<code>v8div16qi2_mask/trunc): Ditto.
1014 (*avx512f_<code>v8div16qi2_store_mask): Ditto.
1015 (ashr<mode>3<mask_name>): Ditto.
1016 (avx512f_vinsert<shuffletype>32x4_mask): Ditto.
1017 (avx512f_vinsert<shuffletype>64x4_mask): Ditto.
1018 (avx512f_shuf_<shuffletype>64x2_mask): Ditto.
1019 (avx512f_shuf_<shuffletype>32x4_mask): Ditto.
1020 (avx512f_pshufdv3_mask): Ditto.
1021 (avx512f_perm<mode>_mask): Ditto.
1022 (avx512f_vpermi2var<mode>3_mask): Ditto.
1023 (avx512f_vpermt2var<mode>3_mask): Ditto.
1024 (avx512f_compress<mode>_mask): Ditto.
1025 (avx512f_compressstore<mode>_mask): Ditto.
1026 (avx512f_expand<mode>_mask): Ditto.
1027 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>): Extend
1029 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
1030 (<plusminus_insn><mode>3<mask_name>): Ditto.
1031 (*<plusminus_insn><mode>3<mask_name>): Ditto.
1032 (mul<mode>3<mask_name>): Ditto.
1033 (*mul<mode>3<mask_name>): Ditto.
1034 (<sse>_div<mode>3<mask_name>): Ditto.
1035 (<mask_codefor>rcp14<mode><mask_name>): Ditto.
1036 (<sse>_sqrt<mode>2<mask_name>): Ditto.
1037 (<mask_codefor>rsqrt14<mode><mask_name>): Ditto.
1038 (<code><mode>3<mask_name>/smaxmin): Ditto.
1039 (*<code><mode>3_finite<mask_name>/smaxmin): Ditto.
1040 (*<code><mode>3<mask_name>/smaxmin): Ditto.
1041 (float<sseintvecmodelower><mode>2<mask_name>): Ditto.
1042 (ufloatv16siv16sf2<mask_name>): Ditto.
1043 (<mask_codefor>avx512f_fix_notruncv16sfv16si<mask_name>): Ditto.
1044 (<mask_codefor>avx512f_ufix_notruncv16sfv16si<mask_name>): Ditto.
1045 (<fixsuffix>fix_truncv16sfv16si2<mask_name>): Ditto.
1046 (float<si2dfmodelower><mode>2<mask_name>): Ditto.
1047 (ufloatv8siv8df<mask_name>): Ditto.
1048 (<mask_codefor>avx512f_cvtpd2dq512<mask_name>): Ditto.
1049 (avx512f_ufix_notruncv8dfv8si<mask_name>): Ditto.
1050 (<fixsuffix>fix_truncv8dfv8si2<mask_name>): Ditto.
1051 (<mask_codefor>avx512f_cvtpd2ps512<mask_name>): Ditto.
1052 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix><mask_name>): Ditto.
1053 (<mask_codefor>avx512f_unpckhps512<mask_name>): Ditto.
1054 (<mask_codefor>avx512f_unpcklps512<mask_name>): Ditto.
1055 (<mask_codefor>avx512f_movshdup512<mask_name>): Ditto.
1056 (<mask_codefor>avx512f_movsldup512<mask_name>): Ditto.
1057 (<mask_codefor>avx512f_vextract<shuffletype>32x4_1<mask_name>): Ditto.
1058 (vec_extract_lo_<mode><mask_name>): Ditto.
1059 (vec_extract_hi_<mode><mask_name>): Ditto.
1060 (<mask_codefor>avx512f_unpckhpd512<mask_name>): Ditto.
1061 (avx512f_movddup512<mask_name>): Ditto.
1062 (avx512f_unpcklpd512<mask_name>): Ditto.
1063 (*avx512f_unpcklpd512<mask_name>): Ditto.
1064 (*avx512f_vmscalef<mode>): Ditto.
1065 (avx512f_scalef<mode><mask_name>): Ditto.
1066 (avx512f_getexp<mode><mask_name>): Ditto.
1067 (<mask_codefor>avx512f_align<mode><mask_name>): Ditto.
1068 (avx512f_rndscale<mode><mask_name>): Ditto.
1069 (avx512f_shufps512_1<mask_name>): Ditto.
1070 (avx512f_shufpd512_1<mask_name>): Ditto.
1071 (<plusminus_insn><mode>3<mask_name>): Ditto.
1072 (*<plusminus_insn><mode>3<mask_name>): Ditto.
1073 (vec_widen_umult_even_v16si<mask_name>): Ditto.
1074 (*vec_widen_umult_even_v16si<mask_name>): Ditto.
1075 (vec_widen_smult_even_v16si<mask_name>): Ditto.
1076 (*vec_widen_smult_even_v16si<mask_name>): Ditto.
1077 (mul<mode>3<mask_name>): Ditto.
1078 (*<sse4_1_avx2>_mul<mode>3<mask_name>): Ditto.
1079 (<shift_insn><mode>3<mask_name>): Ditto.
1080 (avx512f_<rotate>v<mode><mask_name>/rotate): Ditto.
1081 (avx512f_<rotate><mode><mask_name>): Ditto.
1082 (<code><mode>3<mask_name>/maxmin): Ditto.
1083 (*avx2_<code><mode>3<mask_name>/maxmin): Ditto.
1084 (<sse2_avx2>_andnot<mode>3<mask_name>): Ditto.
1085 (*andnot<mode>3<mask_name>): Ditto.
1086 (<mask_codefor><code><mode>3<mask_name>/any_logic): Ditto.
1087 (<mask_codefor>avx512f_interleave_highv16si<mask_name>): Ditto.
1088 (<mask_codefor>avx512f_interleave_lowv16si<mask_name>): Ditto.
1089 (<mask_codefor>avx512f_vinsert<shuffletype>32x4_1<mask_name>): Ditto.
1090 (vec_set_lo_<mode><mask_name>): Ditto.
1091 (vec_set_hi_<mode><mask_name>): Ditto.
1092 (avx512f_shuf_<shuffletype>64x2_1<mask_name>): Ditto.
1093 (avx512f_shuf_<shuffletype>32x4_1<mask_name>): Ditto.
1094 (avx512f_pshufd_1<mask_name>): Ditto.
1095 (<mask_codefor>abs<mode>2<mask_name>): Ditto.
1096 (<mask_codefor>avx512f_<code>v16qiv16si2<mask_name>): Ditto.
1097 (avx512f_<code>v16hiv16si2<mask_name>/any_extend): Ditto.
1098 (avx512f_<code>v8qiv8di2<mask_name>/any_extend): Ditto.
1099 (avx512f_<code>v8hiv8di2<mask_name>/any_extend): Ditto.
1100 (avx512f_<code>v8siv8di2<mask_name>/any_extend): Ditto.
1101 (avx512er_exp2<mode><mask_name>): Ditto.
1102 (<mask_codefor>avx512er_rcp28<mode><mask_name>): Ditto.
1103 (<mask_codefor>avx512er_rsqrt28<mode><mask_name>): Ditto.
1104 (<avx2_avx512f>_permvar<mode><mask_name>): Ditto.
1105 (<avx2_avx512f>_perm<mode>_1<mask_name>): Ditto.
1106 (<mask_codefor>avx512f_vec_dup<mode><mask_name>): Ditto.
1107 (<mask_codefor>avx512f_broadcast<mode><mask_name>/V16FI): Ditto.
1108 (<mask_codefor>avx512f_broadcast<mode><mask_name>/V8FI): Ditto.
1109 (<mask_codefor>avx512f_vec_dup_gpr<mode><mask_name>): Ditto.
1110 (<mask_codefor>avx512f_vec_dup_mem<mode><mask_name>): Ditto.
1111 (<sse2_avx_avx512f>_vpermil<mode><mask_name>/VF2): Ditto.
1112 (<sse2_avx_avx512f>_vpermil<mode><mask_name>/VF1): Ditto.
1113 (*<sse2_avx_avx512f>_vpermilp<mode><mask_name>): Ditto.
1114 (<sse2_avx_avx512f>_vpermilvar<mode>3<mask_name>): Ditto.
1115 (<avx2_avx512f>_ashrv<mode><mask_name>): Ditto.
1116 (<avx2_avx512f>_<shift_insn>v<mode><mask_name>): Ditto.
1117 (<mask_codefor>avx512f_vcvtph2ps512<mask_name>): Ditto.
1118 (<mask_codefor>avx512f_vcvtps2ph512<mask_name>): Ditto.
1119 (avx512f_getmant<mode><mask_name>): Ditto.
1120 (clz<mode>2<mask_name>): Ditto.
1121 (<mask_codefor>conflict<mode><mask_name>): Ditto.
1122 (*srcp14<mode>): Remove visibility.
1123 (*rsqrt14<mode>): Ditto.
1124 (*fma_fmsub_<mode>): Ditto.
1125 (*fma_fnmadd_<mode>): Ditto.
1126 (*avx512f_rndscale<mode>): Ditto.
1127 * config/i386/subst.md: New file.
1129 2013-11-13 Joseph Myers <joseph@codesourcery.com>
1131 * doc/extend.texi (Statement Exprs, Typeof): Discuss __auto_type.
1132 * ginclude/stdatomic.h (kill_dependency, atomic_store_explicit)
1133 (atomic_load_explicit, atomic_exchange_explicit)
1134 (atomic_compare_exchange_strong_explicit)
1135 (atomic_compare_exchange_weak_explicit): Use __auto_type to
1136 declare variable initialized with PTR argument.
1138 2013-11-12 Jeff Law <law@redhat.com>
1140 * tree-ssa-threadedge.c (thread_around_empty_blocks): New argument
1141 backedge_seen_p. Set, use and pass it to children appropriately.
1142 (thread_through_normal_block): Similarly.
1143 (thread_across_edge): Similarly.
1145 * gimple-ssa-isolate-paths.c (check_loadstore): Mark discovered
1146 memory references as volatile.
1147 (insert_trap_and_remove_trailing_statements): Fix comment.
1149 2013-11-12 Vladimir Makarov <vmakarov@redhat.com>
1152 * ira-costs.c (record_operand_costs): Check operands number for
1155 2013-11-12 Michael Meissner <meissner@linux.vnet.ibm.com>
1158 * config/rs6000/rs6000.md (movdi_internal32): Eliminate
1159 constraints that would allow DImode into the traditional Altivec
1160 registers, but cause undesirable code generation when loading 0 as
1162 (movdi_internal64): Likewise.
1163 (cmp<mode>_fpr): Do not use %x for CR register output.
1164 (extendsfdf2_fpr): Fix constraints when -mallow-upper-df and
1165 -mallow-upper-sf debug switches are used.
1167 2013-11-12 Andrew MacLeod <amacleod@redhat.com>
1169 * gimple-expr.h (create_tmp_var_name, create_tmp_var_raw,
1170 create_tmp_var, create_tmp_reg, mark_addressable, is_gimple_reg_rhs):
1171 Relocate prototypes from gimple.h.
1172 * gimplify.h: New File. Relocate some prototypes from gimple.h here.
1173 (gimple_predicate, enum fallback, enum gimplify_status): Relocate
1175 * gimple.h: Move some prototypes to gimplify.h.
1176 (gimple_predicate, enum fallback, enum gimplify_status): Move to
1178 (gimple_do_not_emit_location_p, gimple_set_do_not_emit_location):
1179 Relocate from gimpify.c.
1180 * gimple-expr.c (remove_suffix, tmp_var_id_num, create_tmp_var_name,
1181 create_tmp_var_raw, create_tmp_var, create_tmp_reg, mark_addressable,
1182 is_gimple_reg_rhs) Relocate from gimplify.c.
1183 * gimplify.c (mark_addressable): Move to gimple-expr.c.
1184 (gimple_seq_add_stmt_without_update): Move to gimple.c.
1185 (remove_suffix, tmp_var_id_num, create_tmp_var_name,
1186 create_tmp_var_raw, create_tmp_var, create_tmp_reg,
1187 is_gimple_reg_rhs): Move to gimple-expr.c.
1188 (should_carry_location_p): Move to gimple.c.
1189 (gimple_do_not_emit_location_p, gimple_set_do_not_emit_location): Move
1191 (annotate_one_with_location, annotate_all_with_location_after,
1192 annotate_all_with_location): Move to gimple.c.
1193 (compare_case_labels, sort_case_labels,
1194 preprocess_case_label_vec_for_gimple): Move to gimple.c.
1195 (rhs_predicate_for): Make static.
1196 (gimplify_assign): Relocate from gimple.c.
1197 * gimple.c (gimplify_assign): Move to gimplify.c.
1198 (gimple_seq_add_stmt_without_update, should_carry_location_p,
1199 annotate_one_with_location, annotate_all_with_location_after,
1200 annotate_all_with_location, compare_case_labels, sort_case_labels,
1201 preprocess_case_label_vec_for_gimple): Relocate from gimplify.c.
1202 * tree.h (unshare_expr, unshare_expr_without_location,
1203 mark_addressable): Move prototypes to gimplify.h.
1204 * Makefile.in (GTFILES): gimple-expr.c now has the GTY tag for
1206 * asan.c: Include gimplify.h rather than gimple.h.
1207 * cfgloopmanip.c: Likewise.
1208 * cgraphunit.c: Likewise.
1209 * cilk-common.c: Likewise.
1210 * dwarf2out.c: Dont include gimple.h.
1211 * fold-const.c: Include gimplify.h rather than gimple.h.
1212 * function.c: Likewise.
1213 * gimple-fold.c: Likewise.
1214 * gimple-ssa-strength-reduction.c: Likewise.
1215 * graphite-clast-to-gimple.c: Likewise.
1216 * graphite-sese-to-poly.c: Likewise.
1217 * ipa-prop.c: Likewise.
1218 * ipa-split.c: Likewise.
1220 * langhooks.c: Dont include gimple.h.
1221 * loop-init.c: Include gimplify.h rather than gimple.h.
1222 * omp-low.c: Likewise.
1224 * stor-layout.c: Likewise.
1225 * targhooks.c: Likewise.
1226 * trans-mem.c: Likewise.
1227 * tree-affine.c: Likewise.
1228 * tree-cfg.c: Likewise.
1229 * tree-cfgcleanup.c: Likewise.
1230 * tree-complex.c: Likewise.
1231 * tree-if-conv.c: Likewise.
1232 * tree-inline.c: Likewise.
1233 * tree-iterator.c: Likewise.
1234 * tree-loop-distribution.c: Likewise.
1235 * tree-nested.c: Likewise.
1236 * tree-parloops.c: Likewise.
1237 * tree-predcom.c: Likewise.
1238 * tree-profile.c: Likewise.
1239 * tree-scalar-evolution.c: Likewise.
1240 * tree-sra.c: Likewise.
1241 * tree-ssa-address.c: Likewise.
1242 * tree-ssa-ccp.c: Likewise.
1243 * tree-ssa-dce.c: Likewise.
1244 * tree-ssa-forwprop.c: Likewise.
1245 * tree-ssa-ifcombine.c: Likewise.
1246 * tree-ssa-loop-im.c: Likewise.
1247 * tree-ssa-loop-ivopts.c: Likewise.
1248 * tree-ssa-loop-manip.c: Likewise.
1249 * tree-ssa-loop-niter.c: Likewise.
1250 * tree-ssa-loop-prefetch.c: Likewise.
1251 * tree-ssa-loop-unswitch.c: Likewise.
1252 * tree-ssa-math-opts.c: Likewise.
1253 * tree-ssa-phiopt.c: Likewise.
1254 * tree-ssa-phiprop.c: Likewise.
1255 * tree-ssa-pre.c: Likewise.
1256 * tree-ssa-propagate.c: Likewise.
1257 * tree-ssa-reassoc.c: Likewise.
1258 * tree-ssa-sccvn.c: Likewise.
1259 * tree-ssa-strlen.c: Likewise.
1260 * tree-ssa.c: Likewise.
1261 * tree-switch-conversio: Likewise.n.c
1262 * tree-tailcall.c: Likewise.
1263 * tree-vect-data-refs.c: Likewise.
1264 * tree-vect-generic.c: Likewise.
1265 * tree-vect-loop-manip.c: Likewise.
1266 * tree-vect-loop.c: Likewise.
1267 * tree-vect-patterns.c: Likewise.
1268 * tree-vect-stmts.c: Likewise.
1270 * value-prof.c: Likewise.
1271 * config/aarch64/aarch64.c: Include gimplify.h instead of gimple.h.
1272 * config/alpha/alpha.c: Likewise.
1273 * config/darwin.c: Likewise.
1274 * config/i386/i386.c: Likewise.
1275 * config/ia64/ia64.c: Likewise.
1276 * config/mep/mep.c: Likewise.
1277 * config/mips/mips.c: Likewise.
1278 * config/rs6000/rs6000.c: Likewise.
1279 * config/s390/s390.c: Likewise.
1280 * config/sh/sh.c: Likewise.
1281 * config/sparc/sparc.c: Likewise.
1282 * config/spu/spu.c: Likewise.
1283 * config/stormy16/stormy16.c: Likewise.
1284 * config/tilegx/tilegx.c: Likewise.
1285 * config/tilepro/tilepro.c: Likewise.
1286 * config/xtensa/xtensa.c: Likewise.
1288 2013-11-12 Adam Butcher <adam@jessamine.co.uk>
1290 * tree.c (grow_tree_vec_stat): New function ...
1291 * tree.h (grow_tree_vec_stat) (grow_tree_vec): ... and its declaration
1292 and macro front-end.
1294 2013-11-12 Marek Polacek <polacek@redhat.com>
1296 * final.c (update_alignments): Initialize label to NULL_RTX.
1298 2013-11-12 Jeff Law <law@redhat.com>
1300 * gimple-ssa-isolate-paths.c (check_loadstore): New function.
1301 (insert_trap_and_remove_trailing_statements): New argument OP which
1302 is the NULL pointer. Emit the trap after the load/store through
1303 the NULL pointer. Simplify the RHS of a store through a NULL pointer
1304 when trivial to do so.
1305 (isolate_path): Corresponding changes.
1306 (gimple_ssa_isolate_erroneous_path): Likewise.
1308 2013-11-12 Teresa Johnson <tejohnson@google.com>
1309 Jan Hubicka <jh@suse.cz>
1311 * predict.c (drop_profile): New function.
1312 (handle_missing_profiles): Ditto.
1313 (counts_to_freqs): Don't overwrite estimated frequencies
1314 when function has no profile counts.
1315 * predict.h (handle_missing_profiles): Declare.
1316 * tree-inline.c (freqs_to_counts): New function.
1317 (copy_cfg_body): Invoke freqs_to_counts as needed.
1318 * tree-profile.c (tree_profiling): Invoke handle_missing_profiles.
1320 2013-11-12 H.J. Lu <hongjiu.lu@intel.com>
1323 * config/i386/x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL):
1325 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Set for m_HASWELL.
1327 2013-11-12 H.J. Lu <hongjiu.lu@intel.com>
1330 * config/i386/i386.c (ix86_option_override_internal): Check
1331 X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL and
1332 X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL for
1333 MASK_AVX256_SPLIT_UNALIGNED_LOAD and
1334 MASK_AVX256_SPLIT_UNALIGNED_STORE.
1336 * config/i386/x86-tune.def (X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL):
1337 Clear m_COREI7_AVX and update comments.
1338 (X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL): Likewise.
1340 2013-11-12 Martin Jambor <mjambor@suse.cz>
1342 PR rtl-optimization/10474
1343 * ira.c (interesting_dest_for_shprep): New function.
1344 (split_live_ranges_for_shrink_wrap): Likewise.
1345 (find_moveable_pseudos): Move calculation of dominance info,
1346 df_analysios and the final anlyses to...
1347 (ira): ...here, call split_live_ranges_for_shrink_wrap.
1349 2013-11-12 Bin Cheng <bin.cheng@arm.com>
1351 * tree-ssa-loop-ivopts.c (force_expr_to_var_cost): Refactor the code.
1352 Handle type conversion.
1354 2013-11-11 Martin Liska <marxin.liska@gmail.com>
1355 Jan Hubicka <jh@suse.cz>
1357 * cgraph.c (dump_cgraph_node): Profile dump added.
1358 * cgraph.h (struct cgraph_node): New time profile variable added.
1359 * cgraphclones.c (cgraph_clone_node): Time profile is cloned.
1360 * gcov-io.h (gcov_type): New profiler type introduced.
1361 * ipa-profile.c (lto_output_node): Streaming for time profile added.
1362 (input_node): Time profiler is read from LTO stream.
1363 * predict.c (maybe_hot_count_p): Hot prediction changed.
1364 * profile.c (instrument_values): New case for time profiler added.
1365 (compute_value_histograms): Read of time profile.
1366 * tree-pretty-print.c (dump_function_header): Time profiler is dumped.
1367 * tree-profile.c (init_ic_make_global_vars): Time profiler
1369 (gimple_init_edge_profiler): TP function instrumentation.
1370 (gimple_gen_time_profiler): New.
1371 * value-prof.c (gimple_add_histogram_value): Support for time profiler
1373 (dump_histogram_value): TP type added to dumps.
1374 (visit_hist): More sensitive check that takes TP into account.
1375 (gimple_find_values_to_profile): TP instrumentation.
1376 * value-prof.h (hist_type): New histogram type added.
1377 (struct histogram_value_t): Pointer to struct function added.
1378 * libgcc/Makefile.in: New GCOV merge function for TP added.
1379 * libgcov.c: function_counter variable introduced.
1380 (_gcov_merge_time_profile): New.
1381 (_gcov_time_profiler): New.
1383 2013-11-11 Marc Glisse <marc.glisse@inria.fr>
1384 Jeff Law <law@redhat.com>
1386 * tree-ssa-alias.c (stmt_kills_ref_p_1): Use
1387 ao_ref_init_from_ptr_and_size for builtins.
1389 2013-11-11 Uros Bizjak <ubizjak@gmail.com>
1390 H.J. Lu <hongjiu.lu@intel.com>
1393 * config/i386/x86-tune.def
1394 (X86_TUNE_MISALIGNED_MOVE_STRING_PRO_EPILOGUES): Rename from
1395 TARGET_MISALIGNED_MOVE_STRING_PROLOGUES.
1396 * config/i386/i386.h
1397 (TARGET_MISALIGNED_MOVE_STRING_PRO_EPILOGUES): Rename from
1398 TARGET_MISALIGNED_MOVE_STRING_PROLOGUES_EPILOGUES. Update for renamed
1399 X86_TUNE_MISALIGNED_MOVE_STRING_PRO_EPILOGUES.
1400 * config/i386/i386.c (ix86_expand_set_or_movmem): Use
1401 TARGET_MISALIGNED_MOVE_STRING_PRO_EPILOGUES to calculate
1402 misaligned_prologue_used. Check that
1403 desired_aling <= epilogue_size_needed.
1405 2013-11-11 Cong Hou <congh@google.com>
1407 PR tree-optimization/59050
1408 * tree-vect-data-refs.c (comp_dr_addr_with_seg_len_pair): Bug fix.
1410 2013-11-11 Joern Rennecke <joern.rennecke@embecosm.com>
1413 * expmed.c (emit_store_flag): Fail for const-const comparison.
1415 2013-11-11 Tristan Gingold <gingold@adacore.com>
1416 Eric Botcazou <ebotcazou@adacore.com>
1418 * tree.h (CONSTRUCTOR_NO_CLEARING): Define.
1419 * tree-core.h (CONSTRUCTOR_NO_CLEARING): Document it.
1420 * tree.def (CONSTRUCTOR): Likewise.
1421 * doc/generic.texi (CONSTRUCTOR): Likewise. Update description.
1422 * gimplify.c (gimplify_init_constructor): Do not clear the object when
1423 the constructor is incomplete and CONSTRUCTOR_NO_CLEARING is set.
1425 2013-11-11 Basile Starynkevitch <basile@starynkevitch.net>
1427 * toplev.c (toplev_main): Move PLUGIN_FINISH invocation before
1430 2013-11-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1432 * config/arm/arm.c (arm_new_rtx_costs): Return after handling
1435 2013-11-11 Joern Rennecke <joern.rennecke@embecosm.com>
1437 * config/arc/arc.h (LOGICAL_OP_NON_SHORT_CIRCUIT): Define.
1439 2013-11-08 Jeff Law <law@redhat.com>
1441 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
1442 threading paths first, then perform PHI node checks if applicable.
1444 2013-11-10 Karlson2k <k2k@narod.ru>
1445 Kai Tietz <ktietz@redhat.com>
1448 * configure.ac: Adding for exported symbols check
1449 and for rdynamic-check executable-extension.
1450 * configure: Regenerated.
1452 2013-11-10 Uros Bizjak <ubizjak@gmail.com>
1454 * mode-switching.c (optimize_mode_switching): Mark block as
1455 nontransparent, if last_mode at block exit is different from no_mode.
1457 2013-11-09 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1459 * function.c (NAME__MAIN): Move to...
1460 * cfgexpand.c (NAME__MAIN): ...here.
1462 2013-11-09 Richard Sandiford <rdsandiford@googlemail.com>
1464 * target.def (can_use_doloop_p): New hook.
1465 * doc/tm.texi.in (TARGET_CAN_USE_DOLOOP_P): Add.
1466 * doc/tm.texi: Regenerate.
1467 * doc/md.texi (doloop_begin, doloop_end): Update documentation.
1468 * hooks.h (hook_bool_dint_dint_uint_true): Declare.
1469 * hooks.c (hook_bool_dint_dint_uint_true): New function.
1470 * targhooks.h (can_use_doloop_if_innermost): Declare.
1471 * targhooks.c (can_use_doloop_if_innermost): New function.
1472 * target.h: Include double-int.h.
1473 * loop-doloop.c (doloop_optimize): Call targetm.can_use_doloop_p.
1474 Remove iteration count, maximum iteration count, loop depth and
1475 enter-at-top inputs from doloop_begin and doloop_end.
1476 * config/arc/arc.md (doloop_begin, doloop_end): Update for new
1478 * config/arc/arc.c (arc_can_use_doloop_p): New function.
1479 (TARGET_CAN_USE_DOLOOP_P): Define.
1480 * config/arm/thumb2.md (doloop_end): Update for new interface.
1481 * config/arm/arm.c (TARGET_CAN_USE_DOLOOP_P): Define.
1482 * config/bfin/bfin.md (doloop_end): Update for new interface.
1483 * config/bfin/bfin.c (bfin_can_use_doloop_p): New function.
1484 (TARGET_CAN_USE_DOLOOP_P): Define.
1485 * config/c6x/c6x.md (doloop_end): Update for new interface.
1486 * config/ia64/ia64.md (doloop_end): Update for new interface.
1487 * config/ia64/ia64.c (TARGET_CAN_USE_DOLOOP_P): Define.
1488 * config/mep/mep.md (doloop_begin, doloop_end): Update for new
1490 * config/mep/mep.c (mep_emit_doloop): Likewise.
1491 (TARGET_CAN_USE_DOLOOP_P): Define.
1492 * config/rs6000/rs6000.md (doloop_end): Update for new interface.
1493 * config/rs6000/rs6000.c (TARGET_CAN_USE_DOLOOP_P): Define.
1494 * config/s390/s390.md (doloop_end): Update for new interface.
1495 * config/sh/sh.md (doloop_end): Likewise.
1496 * config/spu/spu.md (doloop_end): Likewise.
1497 * config/spu/spu.c (TARGET_CAN_USE_DOLOOP_P): Define.
1498 * config/tilegx/tilegx.md (doloop_end): Update for new interface.
1499 * config/tilegx/tilegx.c (TARGET_CAN_USE_DOLOOP_P): Define.
1500 * config/tilepro/tilepro.md (doloop_end): Update for new interface.
1501 * config/tilepro/tilepro.c (TARGET_CAN_USE_DOLOOP_P): Define.
1502 * config/v850/v850.md (doloop_begin, doloop_end): Update for new
1504 * config/v850/v850.c (TARGET_CAN_USE_DOLOOP_P): Define.
1506 2013-11-08 H.J. Lu <hongjiu.lu@intel.com>
1509 * doc/extend.texi: Move Cilk Plus Builtins node before Other
1512 2013-11-08 Andrew MacLeod <amacleod@redhat.com>
1513 Joseph Myers <joseph@codesourcery.com>
1515 * ginclude/stdatomic.h: New file.
1516 * Makefile.in (USER_H): Add stdatomic.h.
1518 2013-11-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1520 * config/arm/arm.c (arm_new_rtx_costs): Break after handling
1523 2013-11-08 Jeff Law <law@redhat.com>
1525 * tree-ssa-threadupdate.h (delete_thread_path): Declare.
1526 * tree-ssa-threadupdate.c (delete_thread_path): New function.
1527 (ssa_redirect_edges, thread_block_1): Use it.
1528 (thread_through_loop_header, mark_threaded_blocks): Likewise.
1529 (thread_through_all_blocks, register_jump_thread): Likewise.
1530 * tree-ssa-threadedge.c (thread_across_edge): Likewise.
1532 2013-11-08 James Greenhalgh <james.greenhalgh@arm.com>
1534 * config/arm/aarch-common.c
1535 (search_term): New typedef.
1536 (shift_rtx_costs): New array.
1537 (arm_rtx_shift_left_p): New.
1538 (arm_find_sub_rtx_with_search_term): Likewise.
1539 (arm_find_sub_rtx_with_code): Likewise.
1540 (arm_early_load_addr_dep): Add sanity checking.
1541 (arm_no_early_alu_shift_dep): Likewise.
1542 (arm_no_early_alu_shift_value_dep): Likewise.
1543 (arm_no_early_mul_dep): Likewise.
1544 (arm_no_early_store_addr_dep): Likewise.
1546 2013-11-08 Richard Biener <rguenther@suse.de>
1548 PR tree-optimization/59047
1549 * tree-predcom.c (ref_at_iteration): Handle bitfield accesses properly.
1551 2013-11-08 Ilya Enkovich <ilya.enkovich@intel.com>
1553 * common.opt (fcheck-pointer-bounds): Move to ...
1554 * c-family/c.opt: ... here.
1555 * langhooks-def.h (LANG_HOOKS_CHKP_SUPPORTED): Remove.
1556 (LANG_HOOKS_INITIALIZER): Remove LANG_HOOKS_CHKP_SUPPORTED.
1557 * langhooks.h (lang_hooks): Remove chkp_supported field.
1558 * toplev.c (process_options): Remove chkp_supported check.
1560 2013-11-08 Richard Biener <rguenther@suse.de>
1562 PR tree-optimization/59038
1563 PR tree-optimization/58955
1564 * tree-loop-distribution.c (pg_add_dependence_edges): Revert
1565 previous change. Handle known dependences correctly.
1567 2013-11-08 Tom de Vries <tom@codesourcery.com>
1569 * config/rs6000/t-xilinx: Remove duplicate contents.
1571 2013-11-07 Andrew MacLeod <amacleod@redhat.com>
1572 Joseph Myers <joseph@codesourcery.com>
1574 * tree-core.h (enum cv_qualifier): Add TYPE_QUAL_ATOMIC.
1575 (enum tree_index): Add TI_ATOMICQI_TYPE, TI_ATOMICHI_TYPE,
1576 TI_ATOMICSI_TYPE, TI_ATOMICDI_TYPE and TI_ATOMICTI_TYPE.
1577 (struct tree_base): Add atomic_flag field.
1578 * tree.h (TYPE_ATOMIC): New accessor macro.
1579 (TYPE_QUALS, TYPE_QUALS_NO_ADDR_SPACE): Add TYPE_QUAL_ATOMIC.
1580 (TYPE_QUALS_NO_ADDR_SPACE_NO_ATOMIC): New macro.
1581 (atomicQI_type_node, atomicHI_type_node, atomicSI_type_node)
1582 (atomicDI_type_node, atomicTI_type_node): New macros for type nodes.
1583 * tree.c (set_type_quals): Set TYPE_ATOMIC.
1584 (find_atomic_core_type): New function.
1585 (build_qualified_type): Adjust alignment for qualified types.
1586 (build_atomic_base): New function
1587 (build_common_tree_nodes): Build atomicQI_type_node,
1588 atomicHI_type_node, atomicSI_type_node, atomicDI_type_node and
1590 * print-tree.c (print_node): Print atomic qualifier.
1591 * tree-pretty-print.c (dump_generic_node): Print atomic type attribute.
1592 * target.def (atomic_assign_expand_fenv): New hook.
1593 * doc/tm.texi.in (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): New @hook.
1594 * doc/tm.texi: Regenerate.
1595 * targhooks.c (default_atomic_assign_expand_fenv): New function.
1596 * targhooks.h (default_atomic_assign_expand_fenv): Declare.
1597 * sync-builtins.def (__atomic_feraiseexcept): New built-in function.
1598 * config/i386/i386-builtin-types.def (VOID_FTYPE_PUSHORT): New
1600 * config/i386/i386.c (enum ix86_builtins): Add
1601 IX86_BUILTIN_FNSTENV, IX86_BUILTIN_FLDENV, IX86_BUILTIN_FNSTSW and
1602 IX86_BUILTIN_FNCLEX.
1603 (bdesc_special_args): Add __builtin_ia32_fnstenv,
1604 __builtin_ia32_fldenv, __builtin_ia32_fnstsw and __builtin_ia32_fnclex.
1605 (ix86_expand_builtin): Handle the new built-in functions.
1606 (ix86_atomic_assign_expand_fenv): New function.
1607 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): New macro.
1608 * config/i386/i386.md (UNSPECV_FNSTENV, UNSPECV_FLDENV)
1609 (UNSPECV_FNSTSW, UNSPECV_FNCLEX): New unspecs.
1610 (fnstenv, fldenv, fnstsw, fnclex): New insns.
1612 2013-11-07 Steve Ellcey <sellcey@mips.com>
1614 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add fp64 directory.
1615 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add -mfp64 flag.
1616 (MULTILIB_DIRNAMES): Add fp64 directory.
1617 (MULTILIB_EXCEPTIONS): Add new exclusions.
1619 2013-11-07 Aldy Hernandez <aldyh@redhat.com>
1621 * gimplify.c (gimple_regimplify_operands): Do not set
1623 * graphite-sese-to-poly.c (remove_simple_copy_phi): Same.
1624 (rewrite_close_phi_out_of_ssa): Same.
1625 (rewrite_phi_out_of_ssa): Same.
1626 (rewrite_degenerate_phi): Same.
1627 (handle_scalar_deps_crossing_scop_limits): Same.
1628 * tree-if-conv.c (predicate_scalar_phi): Same.
1629 * tree-parloops.c (create_loads_for_reductions): Same.
1630 (create_final_loads_for_reduction): Same.
1631 (create_loads_and_stores_for_name): Same.
1632 (transform_to_exit_first_loop): Same.
1633 (create_parallel_loop): Same.
1634 * tree-ssa-loop-im.c
1635 (move_computations_dom_walker::before_dom_children): Same.
1636 * tree-ssa-loop-manip.c (rewrite_phi_with_iv): Same.
1637 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children): Same.
1638 * tree-ssa-propagate.c (substitute_and_fold): Same.
1639 * tree-vect-loop.c (vect_finalize_reduction): Same.
1640 * tree-vect-stmts.c (vectorizable_call): Same.
1642 2013-11-07 Mike Stump <mikestump@comcast.net>
1644 * config/pdp11/pdp11.c: Include dbxout.h.
1645 * config/picochip/picochip.c: Likewise.
1647 2013-11-07 Cong Hou <congh@google.com>
1649 PR tree-optimization/56764
1650 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
1651 Combine alias checks if it is possible to amortize the runtime
1652 overhead. Return the number of alias checks after merging.
1653 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
1654 Use the function vect_create_cond_for_alias_checks () to check
1655 the number of alias checks.
1657 2013-11-07 Jeff Law <law@redhat.com>
1659 * varpool.c (ctor_for_folding): Fix typo in comment.
1661 2013-11-07 Joern Rennecke <joern.rennecke@embecosm.com>
1663 * config/arc/arc.c (arc_ifcvt): Use commutativity, e.g.:
1664 reg_a := reg_b + reg_a ==> reg_a := reg_a + reg_b
1666 2013-11-07 Jeff Law <law@redhat.com>
1668 * doc/invoke.texi (-fisolate-erroneous-paths): Document.
1670 * gimple-ssa-isolate-paths.c (gate_isolate_erroneous_paths):
1671 No longer check if we have __builtin_trap, assume it's available.
1673 2013-11-07 Diego Novillo <dnovillo@google.com>
1675 * attribs.c (lookup_scoped_attribute_spec): Make static.
1676 (get_attribute_namespace): Likewise.
1677 * builtins.c (more_const_call_expr_args_p): Move from tree.h.
1678 (validate_arglist): Move earlier in the file. Make static.
1679 (expand_stack_restore): Move from stmt.c
1680 (expand_stack_save): Move from stmt.c
1681 (rewrite_call_expr_array): Move earlier in the file.
1682 (rewrite_call_expr_valist): Likewise.
1683 * cfgexpand.c: Include hard-reg-set.h before tree.h
1686 (expand_asm_loc): Move from stmt.c.
1687 (n_occurrences): Move from stmt.c.
1688 (check_operand_nalternatives): Move from stmt.c.
1689 (tree_conflicts_with_clobbers_p): Move from stmt.c.
1690 (expand_asm_operands): Move from stmt.c
1691 (expand_asm_stmt): Move from stmt.c
1692 (expand_computed_goto): Move from stmt.c
1693 (expand_goto): Move from stmt.c
1694 (expand_null_return_1): Move from stmt.c
1695 (expand_null_return): Move from stmt.c
1696 (expand_value_return): Move from stmt.c
1697 (expand_return): Move from stmt.c
1698 (expand_main_function): Move from function.c
1699 (stack_protect_prologue): Move from function.c
1700 * cgraphclones.c (build_function_type_skip_args): Move from tree.c.
1701 (build_function_decl_skip_args): Move from tree.c.
1702 * explow.c (tree_expr_size): Move from tree.c.
1703 * expr.c (addr_expr_of_non_mem_decl_p): Remove.
1704 (fields_length): Move from tree.c.
1705 * fold-const.c (size_low_cst): Move from tree.c.
1706 (tree_expr_nonzero_warnv_p): Make static. Move earlier in the file.
1707 (tree_expr_nonzero_p): Make static. Move earlier in the file.
1708 (fold_build3_initializer_loc): Remove.
1709 (tree_invalid_nonnegative_warnv_p): Make static.
1710 * function.c (expand_main_function): Move to cfgexpand.c.
1711 (stack_protect_prologue): Move to cfgexpand.c.
1712 (set_insn_locations): Move earlier in the file.
1713 * gimple-fold.c: Include langhooks.h.
1714 (truth_type_for): Move from tree.c.
1715 * print-tree.c (print_vec_tree): Remove.
1716 * stmt.c (expand_computed_goto): Move to cfgexpand.c.
1717 (expand_goto): Move to cfgexpand.c.
1718 (n_occurrences): Move to cfgexpand.c.
1719 (expand_asm_loc): Move to cfgexpand.c
1720 (tree_conflicts_with_clobbers_p): Move to cfgexpand.c.
1721 (expand_asm_operands): Move to cfgexpand.c.
1722 (expand_asm_stmt): Move to cfgexpand.c.
1723 (check_operand_nalternatives): Move to cfgexpand.c
1724 (expand_null_return): Move to cfgexpand.c.
1725 (expand_value_return): Move to cfgexpand.c.
1726 (expand_null_return_1): Move to cfgexpand.c.
1727 (expand_return): Move to cfgexpand.c.
1728 (expand_stack_save): Move to builtins.c.
1729 (expand_stack_restore): Move to builtins.c
1730 * symtab.c: Include output.h.
1731 (decl_assembler_name_hash): Move from tree.c.
1732 (decl_assembler_name_equal): Move from tree.c.
1733 * trans-mem.c (is_tm_safe_or_pure): Move from tree.h.
1734 * tree-eh.c (in_array_bounds_p): Move from tree.c.
1735 (range_in_array_bounds_p): Move from tree.c.
1736 * tree-object-size.c (fini_object_sizes): Make static.
1737 * tree-ssa-dom.c (iterative_hash_exprs_commutative): Move from tree.h.
1738 * tree-vrp.c (ssa_name_nonnegative_p): Remove.
1739 * tree.c (decl_assembler_name_equal): Move to symtab.c.
1740 (tree_expr_size): Move to explow.c.
1741 (decl_assembler_name_hash): Move to symtab.c.
1742 (real_twop): Remove.
1743 (tree_expr_size): Move to explow.c.
1744 (stabilize_reference_1): Move earlier in the file. Make static.
1745 (omp_remove_redundant_declare_simd_attrs): Remove.
1746 (simple_cst_list_equal): Move earlier in the file. Make static.
1747 (size_low_cst): Move to fold-const.c.
1748 (build_type_no_quals): Remove.
1749 (build_function_type_skip_args): Move to cgraphclones.c.
1750 (build_function_decl_skip_args): Move to cgraphclones.c.
1751 (in_array_bounds_p): Move to tree-eh.c.
1752 (range_in_array_bounds_p): Move to tree-eh.c.
1753 (truth_type_for): Move to gimple-fold.c.
1754 (list_equal_p): Remove.
1755 * tree.h (decl_assembler_name_equal): Remove.
1756 (decl_assembler_name_hash): Remove.
1757 (truth_type_for): Remove.
1758 (build_type_no_quals): Remove.
1759 (build_function_decl_skip_args): Remove.
1760 (in_array_bounds_p): Remove.
1761 (range_in_array_bounds_p): Remove.
1762 (size_low_cst): Remove.
1763 (omp_remove_redundant_declare_simd_attrs): Remove.
1764 (tree_expr_size): Remove.
1765 (fields_length): Remove.
1766 (stabilize_reference_1): Remove.
1767 (expand_goto): Remove.
1768 (expand_stack_save): Remove.
1769 (expand_stack_restore): Remove.
1770 (expand_return): Remove.
1771 (fold_build3_initializer_loc): Remove.
1772 (tree_expr_nonzero_p): Remove.
1773 (tree_invalid_nonnegative_warnv_p): Remove.
1774 (tree_expr_nonzero_warnv_p): Remove.
1775 (fold_builtin_snprintf_chk): Remove.
1776 (validate_arglist): Remove.
1777 (iterative_hash_exprs_commutative): Move to tree-ssa-dom.c.
1778 (simple_cst_list_equal): Remove.
1779 (real_twop): Remove.
1780 (expand_main_function): Remove.
1781 (stack_protect_prologue): Remove.
1782 (print_vec_tree): Remove.
1783 (lookup_scoped_attribute_spec): Remove.
1784 (get_attribute_namespace): Remove.
1785 (expand_computed_goto): Remove.
1786 (expand_asm_stmt): Remove.
1787 (list_equal_p): Remove.
1788 (ssa_name_nonnegative_p): Remove.
1789 (fini_object_sizes): Remove.
1790 (addr_expr_of_non_mem_decl_p): Remove.
1791 (is_tm_safe_or_pure): Move to trans-mem.c.
1792 (more_const_call_expr_args_p): Remove.
1793 (save_vtable_map_decl): Remove.
1795 2013-11-07 Thomas Schwinge <thomas@codesourcery.com>
1797 * doc/sourcebuild.texi (Top Level) <lto-plugin>: GNU ld can use
1798 linker plugins, too.
1800 * config/arc/arc.h (LINK_COMMAND_SPEC): For -ftree-parallelize-loops=*,
1801 link to libgomp and its dependencies.
1802 * config/ia64/hpux.h (LIB_SPEC): Likewise.
1803 * config/pa/pa-hpux11.h (LIB_SPEC): Likewise.
1804 * config/pa/pa64-hpux.h (LIB_SPEC): Likewise.
1805 * gcc.c (GOMP_SELF_SPECS): Update comment about libgomp's dependencies.
1807 2013-11-07 Jakub Jelinek <jakub@redhat.com>
1809 * tree-ssa-loop-niter.c: Include tree-ssanames.h.
1810 (determine_value_range): Add loop argument. Use get_range_info to
1812 (bound_difference): Adjust caller.
1814 2013-11-07 Richard Biener <rguenther@suse.de>
1815 Jakub Jelinek <jakub@redhat.com>
1817 * tree-vrp.c (find_assert_locations): Pre-seed live bitmaps for loop
1818 latches from header PHI arguments from the latch edge.
1820 2013-11-07 Paolo Carlini <paolo.carlini@oracle.com>
1823 * varasm.c (output_constant): Handle NULLPTR_TYPE.
1825 2013-11-07 H.J. Lu <hongjiu.lu@intel.com>
1827 * config/i386/i386.c (ix86_expand_set_or_movmem): Don't set
1828 misaligned_prologue_used when it has been set.
1830 2013-11-07 Yury Gribov <y.gribov@samsung.com>
1831 Jakub Jelinek <jakub@redhat.com>
1834 * asan.c (get_mem_refs_of_builtin_call): Allow
1835 integer literals as addresses in instrumented builtins.
1837 2013-11-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1839 * config/aarch64/aarch64.c (aarch64_legitimize_reload_address):
1840 Explain why plus_constant is not used.
1842 2013-11-07 Richard Biener <rguenther@suse.de>
1844 * tree-ssa-ccp.c (canonicalize_float_value): Rename to ...
1845 (canonicalize_value): ... this. Also handle stripping of
1847 (get_value, set_lattice_value, get_value_for_expr): Adjust.
1848 * gimple-fold.c (canonicalize_constructor_val): Strip TREE_OVERFLOW.
1849 * tree-ssa-threadedge.c (set_ssa_name_value): Likewise.
1851 2013-11-07 Richard Biener <rguenther@suse.de>
1853 * tree-dfa.c (get_ref_base_and_extent): Fix casting.
1855 2013-11-07 H.J. Lu <hongjiu.lu@intel.com>
1858 * config/i386/i386.md (push peepholer/splitter): Use Pmode
1859 with stack_pointer_rtx.
1861 2013-11-07 Bin Cheng <bin.cheng@arm.com>
1863 * tree-ssa-loop-ivopts.c (get_shiftadd_cost): Check equality
1864 using operand_equal_p.
1866 2013-11-07 Bin Cheng <bin.cheng@arm.com>
1868 * tree-ssa-loop-ivopts.c (alloc_iv): Lower address expressions.
1869 * tree-affine.c (get_inner_reference_aff): Return base.
1870 * tree-affine.h (get_inner_reference_aff): Change prototype.
1872 2013-11-06 Tobias Burnus <burnus@net-b.de>
1874 * doc/invoke.texi (Wdate-time): Fix typo.
1876 2013-11-06 Oleg Endo <olegendo@gcc.gnu.org>
1878 * config/sh/sh.md (addsf3, divsf3, divsf3_i, rsqrtsf2, cmpgtdf_t,
1879 cmpeqdf_t, *ieee_ccmpeqdf_t, negdf2, sqrtdf2, absdf2): Use
1880 fp_arith_reg_operand instead of arith_reg_operand.
1882 2013-11-06 Oleg Endo <olegendo@gcc.gnu.org>
1884 * config/sh/sh.md (adddi3): Remove empty constraints.
1885 Remove can_create_pseudo_p and arith_reg_operand check.
1886 (adddi3_compact, subdi3_compact, *negdi2): Remove constraints.
1887 Split before reload.
1889 2013-11-06 Jeff Law <law@redhat.com>
1890 Tom Tromey <tromey@redhat.com>
1892 * gdbinit.in: Disable strict type checking.
1894 2013-11-06 Vladimir Makarov <vmakarov@redhat.com>
1896 * tree-pass.h (make_pass_live_range_shrinkage): New external.
1897 * timevar.def (TV_LIVE_RANGE_SHRINKAGE): New.
1898 * sched-rgn.c (gate_handle_live_range_shrinkage): New.
1899 (rest_of_handle_live_range_shrinkage): Ditto
1900 (class pass_live_range_shrinkage): Ditto.
1901 (pass_data_live_range_shrinkage): Ditto.
1902 (make_pass_live_range_shrinkage): Ditto.
1903 * sched-int.h (initialize_live_range_shrinkage): New prototype.
1904 (finish_live_range_shrinkage): Ditto.
1905 * sched-deps.c (create_insn_reg_set): Make void return value.
1906 * passes.def: Add pass_live_range_shrinkage.
1907 * ira.c (update_equiv_regs): Don't move if flag_live_range_shrinkage.
1908 * haifa-sched.c (live_range_shrinkage_p): New.
1909 (initialize_live_range_shrinkage, finish_live_range_shrinkage):
1911 (rank_for_schedule): Add code for pressure relief through live
1913 (schedule_insn): Print more debug info.
1914 (sched_init): Setup SCHED_PRESSURE_WEIGHTED for pressure relief
1915 through live range shrinkage.
1916 * doc/invoke.texi (-flive-range-shrinkage): New.
1917 * common.opt (flive-range-shrinkage): New.
1919 2013-11-06 Uros Bizjak <ubizjak@gmail.com>
1922 * config/i386/i386.c (ix86_avx_u128_mode_needed): Require
1923 AVX_U128_DIRTY mode for call_insn RTXes that use AVX256 registers.
1924 (ix86_avx_u128_mode_needed): Return AVX_U128_DIRTY mode for call_insn
1925 RTXes that return in AVX256 register.
1927 2013-11-06 Richard Biener <rguenther@suse.de>
1929 PR tree-optimization/58653
1930 * tree-predcom.c (ref_at_iteration): Rewrite to generate a MEM_REF.
1931 (prepare_initializers_chain): Adjust.
1933 2013-11-06 Andrew MacLeod <amacleod@redhat.com>
1935 * gimple.h (block_in_transaction): Move to basic-block.h and rename.
1936 (gimple_in_transaction): Use bb_in_transaction.
1937 * basic-block.h (bb_in_transaction): Relocate here and rename.
1938 * tree-ssa-loop-im.c (execute_sm): Use bb_in_transaction.
1940 2013-11-06 Richard Biener <rguenther@suse.de>
1942 * tree.c (drop_tree_overflow): New function.
1943 * tree.h (drop_tree_overflow): Declare.
1944 * gimplify.c (gimplify_expr): Drop TREE_OVERFLOW.
1945 * tree-vrp.c (range_int_cst_singleton_p): Use
1946 is_overflow_infinity instead of testing TREE_OVERFLOW.
1947 (extract_range_from_assert): Likewise.
1948 (zero_nonzero_bits_from_vr): Likewise.
1949 (extract_range_basic): Likewise.
1950 (register_new_assert_for): Use drop_tree_overflow.
1951 (vrp_visit_phi_node): Likewise.
1953 2013-11-06 Eric Botcazou <ebotcazou@adacore.com>
1955 * config/i386/i386.c (ix86_expand_prologue): Optimize stack
1956 checking for leaf functions without dynamic stack allocation.
1957 * config/ia64/ia64.c (ia64_emit_probe_stack_range): Adjust.
1958 (ia64_expand_prologue): Likewise.
1959 * config/mips/mips.c (mips_expand_prologue): Likewise.
1960 * config/rs6000/rs6000.c (rs6000_emit_prologue): Likewise.
1961 * config/sparc/sparc.c (sparc_expand_prologue): Likewise.
1962 (sparc_flat_expand_prologue): Likewise.
1964 2013-11-06 James Greenhalgh <james.greenhalgh@arm.com>
1966 * config/aarch64/arm_neon.h (__ST2_LANE_FUNC): Better model data size.
1967 (__ST3_LANE_FUNC): Likewise.
1968 (__ST4_LANE_FUNC): Likewise.
1970 2013-11-06 Nick Clifton <nickc@redhat.com>
1972 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define the
1973 name returned by msp430_mcu_name.
1974 (LIB_SPEC): If a -T option has not been specified then set a
1975 default, mcu-specific, linker script.
1976 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more mcu names.
1977 * config/msp430/msp430.c (msp430x_names): Likewise.
1978 Alpha sort the names for ease of comparison.
1979 (msp430_mcu_name): New function: Returns a string suitable for
1980 use as a C preprocessor symbol based upon the name of the MCU
1982 (msp430_option_override): Accept msp430x and msp430xv2 as generic
1984 * config/msp430/msp430-protos.h (msp430_mcu_name): Prototype.
1986 * gcc.c (do_spec_1): Do not insert a space after a %* substitution
1987 unless it is the last part of a spec substring.
1988 * doc/invoke.texi (Spec Files): Document space insertion
1991 2013-11-06 Christian Bruel <christian.bruel@st.com>
1993 * config/sh/sh-mem.cc (sh_expand_cmpnstr, sh_expand_cmpstr):
1994 Factorize probabilities, Use adjust_address instead of
1995 adjust_automodify_address when possible. Enable for optimize.
1996 (sh_expand_strlen): New function.
1997 * config/sh/sh-protos.h (sh_expand_strlen): Declare.
1998 * config/sh/sh.md (strlensi): New pattern.
1999 (UNSPEC_BUILTIN_STRLEN): Define.
2001 2013-11-06 Jakub Jelinek <jakub@redhat.com>
2004 * expr.c (get_bit_range): Handle *offset == NULL_TREE.
2005 (expand_assignment): If *bitpos is negative, set *offset
2006 and adjust *bitpos, so that it is not negative.
2008 2013-11-06 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
2010 * config/i386/bdver3.md : Added two additional decoder units
2011 to support issue rate of 4 and remodeled vector unit.
2012 * config/i386/i386.c (ix86_issue_rate): Issue rate for BD
2013 architectures is set to 4.
2014 * config/i386/i386.c (ia32_multipass_dfa_lookahead): DFA
2015 lookahead is set to 4 for BD architectures.
2017 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2019 * config/rs6000/rs6000.c (rs6000_option_override_internal):
2020 Remove restriction against use of VSX instructions when generating
2021 code for little endian mode.
2023 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2025 * config/rs6000/altivec.md (mulv4si3): Ensure we generate vmulouh
2026 for both big and little endian.
2027 (mulv8hi3): Swap input operands for merge high and merge low
2028 instructions for little endian.
2030 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2032 * config/rs6000/altivec.md (vec_widen_umult_even_v16qi): Change
2033 define_insn to define_expand that uses even patterns for big
2034 endian and odd patterns for little endian.
2035 (vec_widen_smult_even_v16qi): Likewise.
2036 (vec_widen_umult_even_v8hi): Likewise.
2037 (vec_widen_smult_even_v8hi): Likewise.
2038 (vec_widen_umult_odd_v16qi): Likewise.
2039 (vec_widen_smult_odd_v16qi): Likewise.
2040 (vec_widen_umult_odd_v8hi): Likewise.
2041 (vec_widen_smult_odd_v8hi): Likewise.
2042 (altivec_vmuleub): New define_insn.
2043 (altivec_vmuloub): Likewise.
2044 (altivec_vmulesb): Likewise.
2045 (altivec_vmulosb): Likewise.
2046 (altivec_vmuleuh): Likewise.
2047 (altivec_vmulouh): Likewise.
2048 (altivec_vmulesh): Likewise.
2049 (altivec_vmulosh): Likewise.
2051 2013-11-05 Mike Stump <mikestump@comcast.net>
2053 * Makefile.in (mostlyclean): Remove c-family objects.
2055 2013-11-05 Ian Lance Taylor <iant@google.com>
2057 * config/i386/sync.md (atomic_compare_and_swap<dwi>_doubleword):
2058 If possible, add .cfi directives to record change to bx.
2059 * config/i386/i386.c (ix86_emit_cfi): New function.
2060 * config/i386/i386-protos.h (ix86_emit_cfi): Declare.
2062 2013-11-05 Steven Bosscher <steven@gcc.gnu.org>
2064 * rtlanal.c (tablejump_p): Expect a JUMP_TABLE_DATA to always follow
2065 immediately after a label for a tablejump pattern.
2067 * config/arm/arm.c (is_jump_table): Remove.
2068 (create_fix_barrier): Use tablejump_p instead.
2069 (arm_reorg): Likewise.
2070 (thumb1_output_casesi): Expect JUMP_TABLE_DATA to always be NEXT_INSN.
2071 (thumb2_output_casesi): Likewise.
2072 * config/aarch64/aarch64.c (aarch64_output_casesi): Likewise.
2073 * config/sh/sh.md (casesi_worker_1, casesi_worker_2,
2074 casesi_shift_media, casesi_load_media): Likewise.
2075 * config/iq2000/iq2000.md: Likewise (in anonymous define_insn).
2076 * config/microblaze/microblaze.md: Likewise.
2078 2013-11-05 Tobias Burnus <burnus@net-b.de>
2080 * doc/invoke.texi (-Wdate-time): Document.
2082 2013-11-05 Richard Sandiford <rdsandiford@googlemail.com>
2084 * double-int.c (lshift_double, rshift_double): Remove
2085 SHIFT_COUNT_TRUNCATED handling.
2087 2013-11-05 Jeff Law <law@redhat.com>
2089 * Makefile.in (OBJS): Add gimple-ssa-isolate-paths.o
2090 * common.opt (-fisolate-erroneous-paths): Add option and documentation.
2091 * gimple-ssa-isolate-paths.c: New file.
2092 * gimple.c (check_loadstore): New function.
2093 (infer_nonnull_range): Moved into gimple.c from tree-vrp.c
2094 Verify OP is in the argument list and the argument corresponding
2095 to OP is a pointer type. Use operand_equal_p rather than
2096 pointer equality when testing if OP is on the nonnull list.
2097 Use check_loadstore rather than count_ptr_derefs. Handle
2098 GIMPLE_RETURN statements.
2099 * tree-vrp.c (infer_nonnull_range): Remove.
2100 * gimple.h (infer_nonnull_range): Declare.
2101 * opts.c (default_options_table): Add OPT_fisolate_erroneous_paths.
2102 * passes.def: Add pass_isolate_erroneous_paths.
2103 * timevar.def (TV_ISOLATE_ERRONEOUS_PATHS): New timevar.
2104 * tree-pass.h (make_pass_isolate_erroneous_paths): Declare.
2105 * tree-ssa.c (struct count_ptr_d): Remove.
2106 (count_ptr_derefs, count_uses_and_derefs): Remove.
2107 * tree-ssa.h (count_uses_and_derefs): Remove.
2109 2013-11-05 Jakub Jelinek <jakub@redhat.com>
2111 PR rtl-optimization/58997
2112 * loop-iv.c (iv_subreg): For IV_UNKNOWN_EXTEND, expect
2113 get_iv_value to be in iv->mode rather than iv->extend_mode.
2114 (iv_extend): Likewise. Otherwise, if iv->extend != extend,
2115 use lowpart_subreg on get_iv_value before calling simplify_gen_unary.
2116 * loop-unswitch.c (may_unswitch_on): Make sure op[i] is in the right
2119 2013-11-05 Andrew MacLeod <amacleod@redhat.com>
2121 * gimple.h: Move some prototypes to gimple-expr.h and add to include
2123 (extract_ops_from_tree, gimple_call_addr_fndecl, is_gimple_reg_type):
2124 Move to gimple-expr.h.
2125 * gimple-expr.h: New file. Relocate some prototypes from gimple.h.
2126 (types_compatible_p, is_gimple_reg_type, is_gimple_variable,
2127 is_gimple_id, virtual_operand_p, is_gimple_addressable,
2128 is_gimple_constant, extract_ops_from_tree, gimple_call_addr_fndecl):
2130 * gimple.c (extract_ops_from_tree_1, gimple_cond_get_ops_from_tree,
2131 gimple_set_body, gimple_body, gimple_has_body_p, is_gimple_lvalue,
2132 is_gimple_condexpr, is_gimple_addressable, is_gimple_constant,
2133 is_gimple_address, is_gimple_invariant_address,
2134 is_gimple_ip_invariant_address, is_gimple_min_invariant,
2135 is_gimple_ip_invariant, is_gimple_variable, is_gimple_id,
2136 virtual_operand_p, is_gimple_reg, is_gimple_val, is_gimple_asm_val,
2137 is_gimple_min_lval, is_gimple_call_addr, is_gimple_mem_ref_addr,
2138 gimple_decl_printable_name, useless_type_conversion_p,
2139 types_compatible_p, gimple_can_coalesce_p, copy_var_decl): Move to
2141 * gimple-expr.c: New File.
2142 (useless_type_conversion_p, gimple_set_body, gimple_body,
2143 gimple_has_body_p, gimple_decl_printable_name, copy_var_decl,
2144 gimple_can_coalesce_p, extract_ops_from_tree_1,
2145 gimple_cond_get_ops_from_tree, is_gimple_lvalue, is_gimple_condexpr,
2146 is_gimple_address, is_gimple_invariant_address,
2147 is_gimple_ip_invariant_address, is_gimple_min_invariant,
2148 is_gimple_ip_invariant, is_gimple_reg, is_gimple_val,
2149 is_gimple_asm_val, is_gimple_min_lval, is_gimple_call_addr,
2150 is_gimple_mem_ref_addr): Relocate here.
2151 * Makefile.in (OBJS): Add gimple-expr.o.
2153 2013-11-05 David Malcolm <dmalcolm@redhat.com>
2155 * gengtype-parse.c (struct_field_seq): Support empty structs.
2157 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
2159 * config/i386/t-rtems (MULTILIB_MATCHES): Fix option typos.
2161 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
2163 * config/i386/i386-c.c (ix86_target_macros): Define _SOFT_FLOAT
2165 * config/i386/rtemself.h (TARGET_OS_CPP_BUILTINS): Do not define
2167 (LONG_DOUBLE_TYPE_SIZE): New define.
2168 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Ditto.
2170 2013-11-05 Paolo Carlini <paolo.carlini@oracle.com>
2173 * doc/extend.texi [visibility ("visibility_type")]: Add example
2174 about visibility attribute on namespace declaration.
2176 2013-11-05 Richard Biener <rguenther@suse.de>
2179 * passes.def (all_passes): Start with pass_fixup_cfg again.
2181 2013-11-05 Richard Biener <rguenther@suse.de>
2183 PR tree-optimization/58955
2184 * tree-loop-distribution.c (pg_add_dependence_edges): Fix
2187 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2189 * config/rs6000/vector.md (vec_pack_sfix_trunc_v2df): Adjust for
2191 (vec_pack_ufix_trunc_v2df): Likewise.
2193 2013-11-05 H.J. Lu <hongjiu.lu@intel.com>
2196 * doc/md.texi (@code{movmem@var{m}}): Specify Pmode as mode of
2197 pattern, instead of word_mode.
2199 * expr.c (emit_block_move_via_movmem): Don't use mode wider than
2201 (set_storage_via_setmem): Likewise.
2203 2013-11-05 Andrew MacLeod <amacleod@redhat.com>
2205 * tree-outof-ssa.c (queue_phi_copy_p): Combine phi_ssa_name_p from
2206 gimple.h and the rest of the condition in eliminate_build.
2207 (eliminate_build): Call new routine.
2208 * gimple.h (phi_ssa_name_p): Delete.
2210 2013-11-05 Trevor Saunders <tsaunders@mozilla.com>
2212 * vec.c (vec_prefix::calculate_allocation): Don't try to handle the
2213 case of no prefix and reserving zero slots, because when that's the
2214 case we'll never get here.
2215 * vec.h (va_heap::reserve): Don't try and handle
2216 vec_prefix::calculate_allocation returning zero because that should
2219 2013-11-05 Richard Biener <rguenther@suse.de>
2222 * tree-dfa.c (get_ref_base_and_extent): Merge common code
2223 in MEM_REF and TARGET_MEM_REF handling. Make sure to
2224 process trailing array detection before diving into the
2225 view-converted object (and possibly apply some extra offset).
2227 2013-11-05 Joseph Myers <joseph@codesourcery.com>
2229 * config/i386/i386.c (ix86_float_exceptions_rounding_supported_p):
2231 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): Define.
2233 2013-11-05 Marc Glisse <marc.glisse@inria.fr>
2235 PR tree-optimization/58958
2236 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Use
2237 get_addr_base_and_unit_offset instead of get_ref_base_and_extent.
2239 2013-11-05 Marc Glisse <marc.glisse@inria.fr>
2241 * tree-ssa-alias.h (ranges_overlap_p): Handle negative offsets.
2242 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Likewise.
2244 2013-11-05 Jakub Jelinek <jakub@redhat.com>
2246 PR tree-optimization/58984
2247 * ipa-prop.c (ipa_load_from_parm_agg_1): Add SIZE_P argument,
2248 set *SIZE_P if non-NULL on success.
2249 (ipa_load_from_parm_agg, ipa_analyze_indirect_call_uses): Adjust
2251 (ipcp_transform_function): Likewise. Punt if size of access
2252 is different from TYPE_SIZE on v->value's type.
2254 2013-11-05 Tobias Burnus <burnus@net-b.de>
2256 * doc/invoke.texi (-fopenmp-simd): Document new option.
2257 * gimplify.c (gimplify_body): Accept -fopenmp-simd.
2258 * omp-low.c (execute_expand_omp, execute_lower_omp): Ditto.
2259 * tree.c (attribute_value_equal): Ditto.
2261 2013-11-04 Wei Mi <wmi@google.com>
2263 * sched-rgn.c (add_branch_dependences): Keep insns in
2264 a SCHED_GROUP at the end of BB to remain their location.
2266 2013-11-04 Wei Mi <wmi@google.com>
2268 * config/i386/i386.c (memory_address_length): Extract a part
2269 of code to rip_relative_addr_p.
2270 (rip_relative_addr_p): New Function.
2271 (ix86_macro_fusion_p): Ditto.
2272 (ix86_macro_fusion_pair_p): Ditto.
2273 * config/i386/i386.h: Add new tune features about macro-fusion.
2274 * config/i386/x86-tune.def (DEF_TUNE): Ditto.
2275 * doc/tm.texi: Generated.
2276 * doc/tm.texi.in: Ditto.
2277 * haifa-sched.c (try_group_insn): New Function.
2278 (group_insns_for_macro_fusion): Ditto.
2279 (sched_init): Call group_insns_for_macro_fusion.
2280 * target.def: Add two hooks: macro_fusion_p and
2281 macro_fusion_pair_p.
2283 2013-11-04 Kostya Serebryany <kcc@google.com>
2285 Update to match the changed asan API.
2286 * asan.c (asan_function_start): New function.
2287 (asan_emit_stack_protection): Update the string stored in the
2288 stack red zone to match new API. Store the PC of the current
2289 function in the red zone.
2290 (asan_global_struct): Update the __asan_global definition to match
2292 (asan_add_global): Ditto.
2293 * asan.h (asan_function_start): New prototype.
2294 * final.c (final_start_function): Call asan_function_start.
2295 * sanitizer.def (BUILT_IN_ASAN_INIT): Rename __asan_init_v1
2298 2013-11-04 Wei Mi <wmi@google.com>
2300 * config/i386/i386-c.c (ix86_target_macros_internal): Separate
2301 PROCESSOR_COREI7_AVX out from PROCESSOR_COREI7.
2302 * config/i386/i386.c (ix86_option_override_internal): Ditto.
2303 (ix86_issue_rate): Ditto.
2304 (ix86_adjust_cost): Ditto.
2305 (ia32_multipass_dfa_lookahead): Ditto.
2306 (ix86_sched_init_global): Ditto.
2307 (get_builtin_code_for_version): Ditto.
2308 * config/i386/i386.h (enum target_cpu_default): Ditto.
2309 (enum processor_type): Ditto.
2310 * config/i386/x86-tune.def (DEF_TUNE): Ditto.
2312 2013-11-04 Vladimir Makarov <vmakarov@redhat.com>
2314 PR rtl-optimization/58967
2315 * config/rs6000/rs6000.c (legitimate_lo_sum_address_p): Remove
2316 !lra_in_progress for mode sizes bigger word.
2318 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2320 * config/rs6000/altivec.md (vec_widen_umult_hi_v16qi): Swap
2321 arguments to merge instruction for little endian.
2322 (vec_widen_umult_lo_v16qi): Likewise.
2323 (vec_widen_smult_hi_v16qi): Likewise.
2324 (vec_widen_smult_lo_v16qi): Likewise.
2325 (vec_widen_umult_hi_v8hi): Likewise.
2326 (vec_widen_umult_lo_v8hi): Likewise.
2327 (vec_widen_smult_hi_v8hi): Likewise.
2328 (vec_widen_smult_lo_v8hi): Likewise.
2330 2013-11-04 Ian Lance Taylor <iant@google.com>
2332 * builtins.def (ATTR_NOTHROWCALL_LEAF_LIST): Define.
2333 * sync-builtins.def: Use ATTR_NOTHROWCALL_LEAF_LIST for all sync
2334 builtins that take pointers.
2335 * lto-opts.c (lto_write_options): Write -fnon-call-exceptions if set.
2336 * lto-wrapper.c (merge_and_complain): Collect OPT_fnon_call_exceptions.
2337 (run_gcc): Pass -fnon-call-exceptions.
2339 2013-11-04 Jakub Jelinek <jakub@redhat.com>
2341 * optabs.c (expand_vec_perm): Revert one incorrect line from
2344 PR tree-optimization/58978
2345 * tree-vrp.c (all_imm_uses_in_stmt_or_feed_cond): Don't modify
2346 use_stmt by single_imm_use directly. Only call single_imm_use
2349 2013-11-04 Vladimir Makarov <vmakarov@redhat.com>
2351 PR rtl-optimization/58968
2352 * lra-spills.c (return_regno_p): New function.
2353 (lra_final_code_change): Use it.
2355 2013-11-04 Joseph Myers <joseph@codesourcery.com>
2357 * doc/cpp.texi (__GCC_IEC_559, __GCC_IEC_559_COMPLEX): Document macros.
2358 * target.def (float_exceptions_rounding_supported_p): New hook.
2359 * targhooks.c (default_float_exceptions_rounding_supported_p): New
2361 * targhooks.h (default_float_exceptions_rounding_supported_p): Declare.
2362 * doc/tm.texi.in (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P):
2364 * doc/tm.texi: Regenerate.
2365 * config.gcc (powerpc*-*-linux*): Set extra_objs.
2366 * config/rs6000/rs6000-linux.c: New file.
2367 * config/rs6000/rs6000-protos.h
2368 (rs6000_linux_float_exceptions_rounding_supported_p): Declare.
2369 * config/rs6000/linux.h
2370 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): New macro.
2371 * config/rs6000/linux64.h
2372 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): Likewise.
2373 * config/rs6000/t-linux (rs6000-linux.o): New rule.
2374 * config/rs6000/t-linux64 (rs6000-linux.o): Likewise.
2376 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2378 * config/rs6000/vsx.md (*vsx_le_perm_store_<mode> for VSX_D):
2379 Replace the define_insn_and_split with a define_insn and two
2380 define_splits, with the split after reload re-permuting the source
2381 register to its original value.
2382 (*vsx_le_perm_store_<mode> for VSX_W): Likewise.
2383 (*vsx_le_perm_store_v8hi): Likewise.
2384 (*vsx_le_perm_store_v16qi): Likewise.
2386 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2388 * config/rs6000/vector.md (vec_pack_trunc_v2df): Adjust for
2391 2013-11-04 Jakub Jelinek <jakub@redhat.com>
2393 PR tree-optimization/58946
2394 * tree-ssa-reassoc.c (maybe_optimize_range_tests): Update all
2395 bbs with bbinfo[idx].op != NULL before all blocks with
2396 bbinfo[idx].op == NULL.
2398 2013-11-04 Richard Sandiford <rdsandiford@googlemail.com>
2400 * config/avr/avr-log.c (avr_double_int_pop_digit): Delete.
2401 (avr_dump_double_int_hex): Likewise.
2402 (avr_log_vadump): Remove %D and %X handling.
2403 * config/avr/avr.c (avr_double_int_push_digit): Delete.
2404 (avr_map_op_t): Change map from double_int to unsigned int.
2405 (avr_map_op): Update accordingly.
2406 (avr_map, avr_map_metric, avr_has_nibble_0xf, avr_map_decompose)
2407 (avr_move_bits, avr_out_insert_bits, avr_fold_builtin): Operate on
2408 unsigned ints rather than double_ints.
2410 2013-11-03 Marek Polacek <polacek@redhat.com>
2412 Implement -fsanitize=vla-bound.
2413 * opts.c (common_handle_option): Handle vla-bound.
2414 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE): Define.
2415 * flag-types.h (enum sanitize_code): Add SANITIZE_VLA.
2416 * asan.c (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR.
2418 2013-11-02 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
2420 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Adjust for
2423 2013-11-02 Uros Bizjak <ubizjak@gmail.com>
2425 * config/i386/constraints.md (Ts, Tv): New address constrains.
2426 * config/i386/i386.md (*lea<mode>, *<mode>_<bndcheck>): Use Ts
2427 constraint for address_no_seg_operand.
2428 * config/i386/sse.md (*avx512pf_gatherpf<mode>_mask)
2429 (*avx512pf_gatherpf<mode>, *avx512pf_scatterpf<mode>_mask)
2430 (*avx512pf_scatterpf<mode>, *avx2_gathersi<mode>)
2431 (*avx2_gathersi<mode>_2, *avx2_gatherdi<mode>, *avx2_gatherdi<mode>_2)
2432 (*avx2_gatherdi<mode>_3, *avx2_gatherdi<mode>_4)
2433 (*avx512f_gathersi<mode>, *avx512f_gathersi<mode>_2)
2434 (*avx512f_gatherdi<mode>, *avx512f_gatherdi<mode>_2)
2435 (*avx512f_scattersi<mode> *avx512f_scatterdi<mode>): Use Tv
2436 constraint for vsib_address_operand.
2438 2013-11-02 Steven Bosscher <steven@gcc.gnu.org>
2440 * gcse.c (pre_delete): Remove references to regmove from comments.
2441 * recog.c: (validate_replace_rtx_1): Likewise.
2442 * config/rl78/rl78.c: Likewise.
2443 * config/v850/v850.h: Likewise, and remove unused ENABLE_REGMOVE_PASS.
2444 * common/config/m32r/m32r-common.c: Don't manipulate OPT_fregmove.
2445 * common/config/mmix/mmix-common.c: Likewise.
2447 2013-11-01 Trevor Saunders <tsaunders@mozilla.com>
2449 * function.c (reorder_blocks): Convert block_stack to a stack_vec.
2450 * gimplify.c (gimplify_compound_lval): Likewise.
2451 * graphite-clast-to-gimple.c (gloog): Likewise.
2452 * graphite-dependences.c (loop_is_parallel_p): Likewise.
2453 * graphite-scop-detection.c (scopdet_basic_block_info): Likewise.
2454 (limit_scop); Likewise.
2455 (build_scops): Likewise.
2456 (dot_scop): Likewise.
2457 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
2458 (build_scop_drs): Likewise.
2459 (insert_stmts): Likewise.
2460 (insert_out_of_ssa_copy): Likewise.
2461 (remove_phi): Likewise.
2462 (rewrite_commutative_reductions_out_of_ssa_close_phi): Likewise.
2463 * hw-doloop.c (discover_loop): Likewise.
2464 * tree-call-cdce.c (shrink_wrap_one_built_in_call): Likewise.
2465 * tree-dfa.c (dump_enumerated_decls): Likewise.
2466 * tree-if-conv.c (if_convertable_loop_p): Likewise.
2467 * tree-inline.c (tree_function_versioning): Likewise.
2468 * tree-loop-distribution.c (build_rdg): Likewise.
2469 (rdg_flag_vertex_and_dependent): Likewise.
2470 (distribute_loop): Likewise.
2471 * tree-parloops.c (loop_parallel_p): Likewise.
2472 (eliminate_local_variables): Likewise.
2473 (separate_decls_in_region): Likewise.
2474 * tree-predcom.c (tree_predictive_commoning_loop): Likewise.
2475 * tree-ssa-phiopt.c (cond_if_else_store_replacement): Likewise.
2476 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
2477 * tree-vect-loop.c (vect_analyze_scaler_cycles_1): Likewise.
2478 * tree-vect-patterns.c (vect_pattern_recog): Likewise.
2479 * tree-vect-stmts.c (vect_mark_stmts_to_be_vectorized): Likewise.
2480 (vectorizable_condition): Likewise.
2482 2013-11-01 Uros Bizjak <ubizjak@gmail.com>
2484 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): Always define as 0/1.
2485 * configure: Regenerate.
2486 * config/i386/i386.md (*movdi_internal): Change
2487 HAVE_AS_IX86_INTERUNIT_MOVQ to runtime check.
2488 (*movdf_internal): Ditto.
2489 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
2490 * config/i386/sse.md (vec_concatv2di): Output interunit movq
2491 for HAVE_AS_IX86_INTERUNIT_MOVQ targets.
2493 2013-10-31 Robert Suchanek <Robert.Suchanek@imgtec.com>
2495 * lra-spills.c (assign_spill_hard_regs): Remove statement terminator
2498 2013-10-31 David Malcolm <dmalcolm@redhat.com>
2500 Automated part of renaming of symtab_node_base to symtab_node.
2502 Patch autogenerated by rename_symtab.py from
2503 https://github.com/davidmalcolm/gcc-refactoring-scripts
2504 revision 58bb219cc090b2f4516a9297d868c245495ee622
2505 with ChangeLog entry fixed up by hand.
2507 * cgraph.c (x_cgraph_nodes_queue): Rename symtab_node_base to
2509 (cgraph_node_for_asm): Likewise.
2510 * cgraph.h (symtab_node_base): Likewise.
2511 (cgraph_node): Likewise.
2512 (varpool_node): Likewise.
2513 (is_a_helper <cgraph_node>::test): Likewise.
2514 (is_a_helper <varpool_node>::test): Likewise.
2515 (symtab_nodes): Likewise.
2516 (symtab_register_node): Likewise.
2517 (symtab_unregister_node): Likewise.
2518 (symtab_remove_node): Likewise.
2519 (symtab_get_node): Likewise.
2520 (symtab_node_for_asm): Likewise.
2521 (symtab_node_asm_name): Likewise.
2522 (symtab_node_name): Likewise.
2523 (symtab_insert_node_to_hashtable): Likewise.
2524 (symtab_add_to_same_comdat_group): Likewise.
2525 (symtab_dissolve_same_comdat_group_list): Likewise.
2526 (dump_symtab_node): Likewise.
2527 (debug_symtab_node): Likewise.
2528 (dump_symtab_base): Likewise.
2529 (verify_symtab_node): Likewise.
2530 (verify_symtab_base): Likewise.
2531 (symtab_used_from_object_file_p): Likewise.
2532 (symtab_alias_ultimate_target): Likewise.
2533 (symtab_resolve_alias): Likewise.
2534 (fixup_same_cpp_alias_visibility): Likewise.
2535 (symtab_for_node_and_aliases): Likewise.
2536 (symtab_nonoverwritable_alias): Likewise.
2537 (availability symtab_node_availability): Likewise.
2538 (symtab_semantically_equivalent_p): Likewise.
2539 (fixup_same_cpp_alias_visibility): Likewise.
2540 (symtab_prevail_in_asm_name_hash): Likewise.
2542 (varpool): Likewise.
2543 (varpool_first_variable): Likewise.
2544 (varpool_next_variable): Likewise.
2545 (varpool_first_static_initializer): Likewise.
2546 (varpool_next_static_initializer): Likewise.
2547 (varpool_first_defined_variable): Likewise.
2548 (varpool_next_defined_variable): Likewise.
2549 (cgraph_first_defined_function): Likewise.
2550 (cgraph_next_defined_function): Likewise.
2551 (cgraph_first_function): Likewise.
2552 (cgraph_next_function): Likewise.
2553 (cgraph_first_function_with_gimple_body): Likewise.
2554 (cgraph_next_function_with_gimple_body): Likewise.
2555 (symtab_alias_target): Likewise.
2556 (symtab_real_symbol_p): Likewise.
2557 (symtab_can_be_discarded): Likewise.
2558 * cgraphbuild.c (mark_address): Likewise.
2559 (mark_load): Likewise.
2560 (mark_store): Likewise.
2561 * cgraphunit.c (decide_is_symbol_needed): Likewise.
2563 (enqueue_node): Likewise.
2564 (referred_to_p): Likewise.
2565 (cgraph_process_same_body_aliases): Likewise.
2566 (analyze_functions): Likewise.
2567 (handle_alias_pairs): Likewise.
2568 (output_weakrefs): Likewise.
2569 (compile): Likewise.
2570 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
2571 * ipa-inline-analysis.c (inline_write_summary): Likewise.
2572 * ipa-prop.c (remove_described_reference): Likewise.
2573 (try_decrement_rdesc_refcount): Likewise.
2574 (ipa_edge_duplication_hook): Likewise.
2575 * ipa-ref.c (ipa_record_reference): Likewise.
2576 (ipa_maybe_record_reference): Likewise.
2577 (ipa_clone_ref): Likewise.
2578 (ipa_clone_references): Likewise.
2579 (ipa_clone_referring): Likewise.
2580 (ipa_find_reference): Likewise.
2581 (ipa_remove_stmt_references): Likewise.
2582 (ipa_clear_stmts_in_references): Likewise.
2583 * ipa-ref.h (symtab_node_base): Likewise.
2584 (ipa_ref): Likewise.
2585 (ipa_record_reference): Likewise.
2586 (ipa_maybe_record_reference): Likewise.
2587 (ipa_clone_references): Likewise.
2588 (ipa_clone_referring): Likewise.
2589 (ipa_clone_ref): Likewise.
2590 (ipa_find_reference): Likewise.
2591 (ipa_remove_stmt_references): Likewise.
2592 (ipa_clear_stmts_in_references): Likewise.
2593 * ipa-reference.c (ipa_reference_write_optimization_summary):
2595 * ipa.c (enqueue_node): Likewise.
2596 (process_references): Likewise.
2597 (walk_polymorphic_call_targets): Likewise.
2598 (symtab_remove_unreachable_nodes): Likewise.
2599 (address_taken_from_non_vtable_p): Likewise.
2600 (comdat_can_be_unshared_p_1): Likewise.
2601 (comdat_can_be_unshared_p): Likewise.
2602 (can_replace_by_local_alias): Likewise.
2603 (function_and_variable_visibility): Likewise.
2604 * is-a.h: Likewise (within example in comment).
2605 * lto-cgraph.c (input_cgraph_opt_summary): Likewise.
2606 (lto_symtab_encoder_encode): Likewise.
2607 (lto_symtab_encoder_delete_node): Likewise.
2608 (lto_symtab_encoder_in_partition_p): Likewise.
2609 (lto_set_symtab_encoder_in_partition): Likewise.
2610 (output_refs): Likewise.
2611 (compute_ltrans_boundary): Likewise.
2612 (output_symtab): Likewise.
2613 (input_node): Likewise.
2614 (input_ref): Likewise.
2615 (input_edge): Likewise.
2616 (input_cgraph_1): Likewise.
2617 (input_refs): Likewise.
2618 (output_cgraph_opt_summary): Likewise.
2619 (input_node_opt_summary): Likewise.
2620 (input_cgraph_opt_section): Likewise.
2621 * lto-section-in.c (lto_free_function_in_decl_state_for_node):
2623 * lto-streamer-out.c (lto_output): Likewise.
2624 (output_symbol_p): Likewise.
2625 (produce_symtab): Likewise.
2626 * lto-streamer.h (lto_encoder_entry): Likewise.
2627 (lto_free_function_in_decl_state_for_node): Likewise.
2628 (lto_symtab_encoder_encode): Likewise.
2629 (lto_symtab_encoder_delete_node): Likewise.
2630 (lto_symtab_encoder_in_partition_p): Likewise.
2631 (lto_set_symtab_encoder_in_partition): Likewise.
2632 (lto_symtab_encoder_lookup): Likewise.
2633 (lsei_node): Likewise.
2634 (lto_symtab_encoder_deref): Likewise.
2635 * symtab.c (symtab_hash): Likewise.
2636 (assembler_name_hash): Likewise.
2637 (symtab_nodes): Likewise.
2638 (hash_node): Likewise.
2639 (eq_node): Likewise.
2640 (hash_node_by_assembler_name): Likewise.
2641 (eq_assembler_name): Likewise.
2642 (insert_to_assembler_name_hash): Likewise.
2643 (unlink_from_assembler_name_hash): Likewise.
2644 (symtab_prevail_in_asm_name_hash): Likewise.
2645 (symtab_register_node): Likewise.
2646 (symtab_insert_node_to_hashtable): Likewise.
2647 (symtab_unregister_node): Likewise.
2648 (symtab_get_node): Likewise.
2649 (symtab_remove_node): Likewise.
2650 (symtab_initialize_asm_name_hash): Likewise.
2651 (symtab_node_for_asm): Likewise.
2652 (symtab_add_to_same_comdat_group): Likewise.
2653 (symtab_dissolve_same_comdat_group_list): Likewise.
2654 (symtab_node_asm_name): Likewise.
2655 (symtab_node_name): Likewise.
2656 (dump_symtab_base): Likewise.
2657 (dump_symtab_node): Likewise.
2658 (dump_symtab): Likewise.
2659 (debug_symtab_node): Likewise.
2660 (verify_symtab_base): Likewise.
2661 (verify_symtab_node): Likewise.
2662 (verify_symtab): Likewise.
2663 (symtab_used_from_object_file_p): Likewise.
2664 (symtab_node_availability): Likewise.
2665 (symtab_alias_ultimate_target): Likewise.
2666 (fixup_same_cpp_alias_visibility): Likewise.
2667 (symtab_resolve_alias): Likewise.
2668 (symtab_for_node_and_aliases): Likewise.
2669 (symtab_for_node_and_aliases): Likewise.
2670 (symtab_nonoverwritable_alias_1): Likewise.
2671 (symtab_nonoverwritable_alias): Likewise.
2672 (symtab_semantically_equivalent_p): Likewise.
2673 * value-prof.c (init_node_map): Likewise.
2674 * varasm.c (find_decl): Likewise.
2675 * varpool.c (varpool_node_for_asm): Likewise.
2676 (varpool_remove_unreferenced_decls): Likewise.
2678 2013-10-31 David Malcolm <dmalcolm@redhat.com>
2680 Manual part of renaming of symtab_node_base to symtab_node.
2682 * ipa-ref.h (symtab_node): Remove typedef to pointer type, as it
2683 clashes with the preferred name for the base class.
2684 (const_symtab_node): Remove redundant typedef.
2686 2013-10-31 Jakub Jelinek <jakub@redhat.com>
2688 * optabs.c (expand_vec_perm): Avoid vector mode punning
2689 SUBREGs in SET_DEST.
2690 * expmed.c (store_bit_field_1): Likewise.
2691 * config/i386/sse.md (movdi_to_sse, vec_pack_sfix_trunc_v2df,
2692 vec_pack_sfix_v2df, vec_shl_<mode>, vec_shr_<mode>,
2693 vec_interleave_high<mode>, vec_interleave_low<mode>): Likewise.
2694 * config/i386/i386.c (ix86_expand_vector_move_misalign,
2695 ix86_expand_sse_movcc, ix86_expand_int_vcond, ix86_expand_vec_perm,
2696 ix86_expand_sse_unpack, ix86_expand_args_builtin,
2697 ix86_expand_vector_init_duplicate, ix86_expand_vector_set,
2698 emit_reduc_half, expand_vec_perm_blend, expand_vec_perm_pshufb,
2699 expand_vec_perm_interleave2, expand_vec_perm_pshufb2,
2700 expand_vec_perm_vpshufb2_vpermq,
2701 expand_vec_perm_vpshufb2_vpermq_even_odd, expand_vec_perm_even_odd_1,
2702 expand_vec_perm_broadcast_1, expand_vec_perm_vpshufb4_vpermq2,
2703 ix86_expand_sse2_mulv4si3, ix86_expand_pinsr): Likewise.
2704 (expand_vec_perm_palignr): Likewise. Modify a copy of *d rather
2707 2013-10-31 Uros Bizjak <ubizjak@gmail.com>
2709 * config/i386/i386.c (ix86_expand_sse2_abs): Rename function arguments.
2710 Use gcc_unreachable for unhandled modes. Do not check results of
2711 expand_simple_binop. If not expanded to target, move the result.
2713 2013-10-31 Chung-Ju Wu <jasonwucj@gmail.com>
2714 Shiva Chen <shiva0217@gmail.com>
2716 * config.gcc (nds32*-*-*): Add nds32 target.
2717 * config/nds32/nds32.c: New file.
2718 * config/nds32/nds32.h: New file.
2719 * config/nds32/nds32.md: New file.
2720 * config/nds32/constants.md: New file.
2721 * config/nds32/constraints.md: New file.
2722 * config/nds32/iterators.md: New file.
2723 * config/nds32/nds32-doubleword.md: New file.
2724 * config/nds32/nds32-intrinsic.md: New file.
2725 * config/nds32/nds32_intrinsic.h: New file.
2726 * config/nds32/nds32-modes.def: New file.
2727 * config/nds32/nds32-multiple.md: New file.
2728 * config/nds32/nds32.opt: New file.
2729 * config/nds32/nds32-opts.h: New file.
2730 * config/nds32/nds32-protos.h: New file.
2731 * config/nds32/nds32-peephole2.md: New file.
2732 * config/nds32/pipelines.md: New file.
2733 * config/nds32/predicates.md: New file.
2734 * config/nds32/t-mlibs: New file.
2735 * common/config/nds32: New directory and files.
2737 * doc/invoke.texi (NDS32 options): Document nds32 specific options.
2738 * doc/md.texi (NDS32 family): Document nds32 specific constraints.
2739 * doc/install.texi (Cross-Compiler-Specific Options): Document
2740 --with-nds32-lib for nds32 target.
2741 * doc/extend.texi (Function Attributes, Target Builtins): Document
2742 nds32 specific attributes.
2744 2013-10-31 Vladimir Makarov <vmakarov@redhat.com>
2746 * lra-constraints (process_alt_operands): Use the result
2747 elimination register for operand when matching constraints.
2749 2013-10-31 Jakub Jelinek <jakub@redhat.com>
2751 * tree-vrp.c (maybe_set_nonzero_bits): New function.
2752 (remove_range_assertions): Call it.
2754 * tree.c (tree_ctz): New function.
2755 * tree.h (tree_ctz): New prototype.
2756 * tree-ssanames.h (get_range_info, get_nonzero_bits): Change
2757 first argument from tree to const_tree.
2758 * tree-ssanames.c (get_range_info, get_nonzero_bits): Likewise.
2759 * tree-vectorizer.h (vect_generate_tmps_on_preheader): New prototype.
2760 * tree-vect-loop-manip.c (vect_generate_tmps_on_preheader): No longer
2762 * expr.c (highest_pow2_factor): Reimplemented using tree_ctz.
2763 * tree-vect-loop.c (vect_analyze_loop_operations,
2764 vect_transform_loop): Don't force scalar loop for bound just because
2765 number of iterations is unknown, only do it if it is not known to be
2766 a multiple of vectorization_factor.
2767 * builtins.c (get_object_alignment_2): Use tree_ctz on offset.
2769 * gimple-pretty-print.c (dump_ssaname_info): Print newline also
2770 in case of VR_VARYING. Print get_nonzero_bits if not all ones.
2771 * tree-ssanames.h (struct range_info_def): Add nonzero_bits field.
2772 (set_nonzero_bits, get_nonzero_bits): New prototypes.
2773 * tree-ssa-ccp.c (get_default_value): Use get_range_info to see if
2774 a default def isn't partially constant.
2775 (ccp_finalize): If after IPA, set_range_info if integral SSA_NAME
2776 is known to be partially zero.
2777 (evaluate_stmt): If we'd return otherwise VARYING, use get_range_info
2778 to see if a default def isn't partially constant.
2779 * tree-ssanames.c (set_range_info): Initialize nonzero_bits upon
2780 creation of a range, if VR_RANGE, try to improve nonzero_bits from
2782 (set_nonzero_bits, get_nonzero_bits): New functions.
2784 * tree-cfg.c (assert_unreachable_fallthru_edge_p): New function.
2785 * tree-cfg.h (assert_unreachable_fallthru_edge_p): New prototype.
2786 * tree-vrp.c (all_imm_uses_in_stmt_or_feed_cond): New function.
2787 (remove_range_assertions): If ASSERT_EXPR_VAR has no other immediate
2788 uses but in the condition and ASSERT_EXPR and the other successor of
2789 the predecessor bb is __builtin_unreachable (), set_range_info of the
2790 ASSERT_EXPR_VAR to the range info of the ASSERT_EXPR's lhs.
2792 2013-10-31 Martin Jambor <mjambor@suse.cz>
2794 PR rtl-optimization/58934
2796 2013-10-30 Martin Jambor <mjambor@suse.cz>
2797 PR rtl-optimization/10474
2798 * ira.c (find_moveable_pseudos): Do not calculate dominance info
2800 (interesting_dest_for_shprep): New function.
2801 (split_live_ranges_for_shrink_wrap): Likewise.
2802 (ira): Calculate dominance info and df analysis. Call
2803 split_live_ranges_for_shrink_wrap.
2805 2013-10-31 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
2806 Yury Gribov <y.gribov@samsung.com>
2809 * asan.c (asan_clear_shadow): Allocate a new vreg for temporary
2810 shadow pointer to avoid clobbering the main one.
2812 2013-10-31 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2814 * lower-subreg.c (resolve_simple_move): Copy REG_INC note.
2816 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
2819 * ira-color.c (update_costs_from_copies): Add new parameter. Use
2820 it for calling update_costs_from_allocno.
2821 (assign_hard_reg): Call restore_costs_from_copies only for
2822 !retry_p. Pass new argument to update_costs_from_copies.
2823 (color_pass): Pass new argument to update_costs_from_copies.
2824 (ira_mark_allocation_change): Ditto.
2826 2013-10-30 Sharad Singhai <singhai@google.com>
2829 * opts.c (common_handle_option): Remove deprecated option
2830 -ftree-vectorizer-verbose.
2831 * doc/invoke.texi (Debugging Options): Ditto.
2832 * opts-global.c (handle_common_deferred_options): Ditto.
2833 (dump_remap_tree_vectorizer_verbose): Delete.
2834 * common.opt: Set -ftree-vectorizer-verbose as an ignored option.
2836 2013-10-30 DJ Delorie <dj@redhat.com>
2838 * config/rx/rx.c (ADD_RX_BUILTIN0): New macro, used for builtins
2839 that take no arguments.
2841 2013-10-30 Joern Rennecke <joern.rennecke@embecosm.com>
2844 * reload1.c (update_eliminables_and_spill): New function, broken
2846 (reload): Use it. Check for frame size change after frame size
2847 alignment, and call update_eliminables_and_spill first if continue-ing.
2849 2013-10-30 Cong Hou <congh@google.com>
2852 * config/i386/i386-protos.h (ix86_expand_sse2_abs): New function.
2853 * config/i386/i386.c (ix86_expand_sse2_abs): New function.
2854 * config/i386/sse.md: Add SSE2 support to abs (8/16/32-bit-int).
2856 2013-10-18 Mikael Pettersson <mikpelinux@gmail.com>
2858 PR rtl-optimization/58369
2859 * reload1.c (compute_reload_subreg_offset): New function.
2860 (choose_reload_regs): Use it to pass endian-correct
2861 offset to subreg_regno_offset.
2863 2013-10-30 Tobias Burnus <burnus@net-b.de>
2866 * tree-cfg.c (replace_loop_annotate): Replace warning by
2869 2013-10-30 Jason Merrill <jason@redhat.com>
2871 * configure.ac (loose_warn): Add -Wno-format if
2872 --disable-build-format-warnings.
2874 2013-10-30 David Malcolm <dmalcolm@redhat.com>
2876 * cgraphunit.c (analyze_functions): Split symtab_node declarations
2877 onto multiple lines to make things easier for rename_symtab.py.
2879 * symtab.c (symtab_dissolve_same_comdat_group_list): Likewise.
2880 (symtab_semantically_equivalent_p): Likewise.
2882 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
2885 * lra.c (check_rtl): Remove address check before LRA work.
2887 2013-10-30 Marc Glisse <marc.glisse@inria.fr>
2889 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Look for a
2890 POINTER_PLUS_EXPR in the defining statement.
2892 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
2894 * regmove.c: Remove.
2895 * tree-pass.h (make_pass_regmove): Remove.
2896 * timevar.def (TV_REGMOVE): Remove.
2897 * passes.def (pass_regmove): Remove.
2898 * opts.c (default_options_table): Remove entry for regmove.
2899 * doc/passes.texi: Remove regmove pass description.
2900 * doc/invoke.texi (-foptimize-register-move, -fregmove): Remove
2902 (-fdump-rtl-regmove): Ditto.
2903 * common.opt (foptimize-register-move, fregmove): Ignore.
2904 * Makefile.in (OBJS): Remove regmove.o.
2905 * regmove.c: Remove.
2906 * ira-int.h (struct ira_allocno_pref, ira_pref_t): New structure
2908 (struct ira_allocno) New member allocno_prefs.
2909 (ALLOCNO_PREFS): New macro.
2910 (ira_prefs, ira_prefs_num): New external vars.
2911 (ira_setup_alts, ira_get_dup_out_num, ira_debug_pref): New prototypes.
2912 (ira_debug_prefs, ira_debug_allocno_prefs, ira_create_pref): Ditto.
2913 (ira_add_allocno_pref, ira_remove_pref, ira_remove_allocno_prefs):
2915 (ira_add_allocno_copy_to_list): Remove prototype.
2916 (ira_swap_allocno_copy_ends_if_necessary): Ditto.
2917 (ira_pref_iterator): New type.
2918 (ira_pref_iter_init, ira_pref_iter_cond): New functions.
2919 (FOR_EACH_PREF): New macro.
2920 * ira.c (commutative_constraint_p): Move from ira-conflicts.c.
2921 (ira_get_dup_out_num): Ditto. Rename from get_dup_num. Modify the
2923 (ira_setup_alts): New function.
2924 (decrease_live_ranges_number): New function.
2925 (ira): Call the above function.
2926 * ira-build.c (ira_prefs, ira_prefs_num): New global vars.
2927 (ira_create_allocno): Initialize allocno prefs.
2928 (pref_pool, pref_vec): New static vars.
2929 (initiate_prefs, find_allocno_pref, ira_create_pref): New
2931 (add_allocno_pref_to_list, ira_add_allocno_pref, print_pref): Ditto.
2932 (ira_debug_pref, print_prefs, ira_debug_prefs): Ditto.
2933 (print_allocno_prefs, ira_debug_allocno_prefs, finish_pref): Ditto.
2934 (ira_remove_pref, ira_remove_allocno_prefs, finish_prefs): Ditto.
2935 (ira_add_allocno_copy_to_list): Make static. Rename to
2936 add_allocno_copy_to_list.
2937 (ira_swap_allocno_copy_ends_if_necessary): Make static. Rename to
2938 swap_allocno_copy_ends_if_necessary.
2939 (remove_unnecessary_allocnos, remove_low_level_allocnos): Call
2940 ira_remove_allocno_prefs.
2941 (ira_flattening): Ditto.
2942 (ira_build): Call initiate_prefs, print_prefs.
2943 (ira_destroy): Call finish_prefs.
2944 * ira-color.c (struct update_cost_record): New.
2945 (struct allocno_color_data): Add new member update_cost_records.
2946 (update_cost_record_pool): New static var.
2947 (init_update_cost_records, get_update_cost_record): New functions.
2948 (free_update_cost_record_list, finish_update_cost_records): Ditto.
2949 (struct update_cost_queue_elem): Add member from.
2950 (initiate_cost_update): Call init_update_cost_records.
2951 (finish_cost_update): Call finish_update_cost_records.
2952 (queue_update_cost, get_next_update_cost): Add new param from.
2953 (Update_allocno_cost, update_costs_from_allocno): New functions.
2954 (update_costs_from_prefs): Ditto.
2955 (update_copy_costs): Rename to update_costs_from_copies.
2956 (restore_costs_from_copies): New function.
2957 (update_conflict_hard_regno_costs): Don't go back.
2958 (assign_hard_reg): Call restore_costs_from_copies. Add printing
2960 (pop_allocnos): Add priniting more debug info.
2961 (color_allocnos): Remove prefs for conflicting hard regs.
2962 Call update_costs_from_prefs.
2963 * ira-conflicts.c (commutative_constraint_p): Move to ira.c
2964 (get_dup_num): Rename, modify, and move to ira.c
2965 (process_regs_for_copy): Add prefs.
2966 (add_insn_allocno_copies): Put src as first arg of
2967 process_regs_for_copy. Remove dead code. Call ira_setup_alts.
2968 * ira-costs.c (record_reg_classes): Modify and move code into
2969 record_operands_costs.
2970 (find_costs_and_classes): Create prefs for the hard reg of small
2972 (process_bb_node_for_hard_reg_moves): Add prefs.
2974 2013-10-30 Richard Biener <rguenther@suse.de>
2977 * basic-block.h (pre_and_rev_post_order_compute_fn): New function.
2978 * cfganal.c (pre_and_rev_post_order_compute_fn): New function
2980 (pre_and_rev_post_order_compute): ... which now wraps it.
2981 * graph.c (draw_cfg_nodes_no_loops): Use
2982 pre_and_rev_post_order_compute_fn to avoid ICEing and dependence
2985 2013-10-30 Christian Bruel <christian.bruel@st.com>
2987 * config/sh/sh-mem.cc (sh_expand_cmpnstr): New function.
2988 (sh_expand_cmpstr): Handle known align and schedule improvements.
2989 * config/sh/sh-protos.h (sh_expand_cmpstrn): Declare.
2990 * config/sh/sh.md (cmpstrnsi): New pattern.
2992 2013-10-30 Martin Jambor <mjambor@suse.cz>
2994 PR rtl-optimization/10474
2995 * ira.c (find_moveable_pseudos): Do not calculate dominance info
2997 (interesting_dest_for_shprep): New function.
2998 (split_live_ranges_for_shrink_wrap): Likewise.
2999 (ira): Calculate dominance info and df analysis. Call
3000 split_live_ranges_for_shrink_wrap.
3002 2013-10-30 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3005 * config/arm/arm.c (arm_expand_epilogue_apcs_frame): Emit blockage.
3007 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
3009 * tree-core.h (tree_index): Add TI_POINTER_BOUNDS_TYPE.
3010 * tree.h (POINTER_BOUNDS_P): New.
3011 (BOUNDED_TYPE_P): New.
3013 (pointer_bounds_type_node): New.
3014 * tree.c (build_common_tree_nodes): Initialize
3015 pointer_bounds_type_node.
3016 * gimple.h (gimple_call_get_nobnd_arg_index): New.
3017 (gimple_call_num_nobnd_args): New.
3018 (gimple_call_nobnd_arg): New.
3019 (gimple_return_retbnd): New.
3020 (gimple_return_set_retbnd): New
3021 * gimple.c (gimple_build_return): Increase number of ops
3022 for return statement.
3023 (gimple_call_get_nobnd_arg_index): New.
3024 * gimple-pretty-print.c (dump_gimple_return): Print second op.
3026 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
3028 * ipa.c (cgraph_build_static_cdtor_1): Support contructors
3029 with "chkp ctor" and "bnd_legacy" attributes.
3030 * gimplify.c (gimplify_init_constructor): Avoid infinite
3031 loop during gimplification of bounds initializer.
3033 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
3035 * c-family/c-common.c (handle_bnd_variable_size_attribute): New.
3036 (handle_bnd_legacy): New.
3037 (c_common_attribute_table): Add bnd_variable_size and bnd_legacy.
3038 * doc/extend.texi: Document bnd_variable_size and bnd_legacy
3041 2013-10-29 Ilya Enkovich <ilya.enkovich@intel.com>
3043 * builtin-types.def (BT_FN_VOID_CONST_PTR): New.
3044 (BT_FN_PTR_CONST_PTR): New.
3045 (BT_FN_CONST_PTR_CONST_PTR): New.
3046 (BT_FN_PTR_CONST_PTR_SIZE): New.
3047 (BT_FN_PTR_CONST_PTR_CONST_PTR): New.
3048 (BT_FN_VOID_PTRPTR_CONST_PTR): New.
3049 (BT_FN_VOID_CONST_PTR_SIZE): New.
3050 (BT_FN_PTR_CONST_PTR_CONST_PTR_SIZE): New.
3051 * chkp-builtins.def: New.
3052 * builtins.def: include chkp-builtins.def.
3053 (DEF_CHKP_BUILTIN): New.
3054 * builtins.c (expand_builtin): Support BUILT_IN_CHKP_INIT_PTR_BOUNDS,
3055 BUILT_IN_CHKP_NULL_PTR_BOUNDS, BUILT_IN_CHKP_COPY_PTR_BOUNDS,
3056 BUILT_IN_CHKP_CHECK_PTR_LBOUNDS, BUILT_IN_CHKP_CHECK_PTR_UBOUNDS,
3057 BUILT_IN_CHKP_CHECK_PTR_BOUNDS, BUILT_IN_CHKP_SET_PTR_BOUNDS,
3058 BUILT_IN_CHKP_NARROW_PTR_BOUNDS, BUILT_IN_CHKP_STORE_PTR_BOUNDS,
3059 BUILT_IN_CHKP_GET_PTR_LBOUND, BUILT_IN_CHKP_GET_PTR_UBOUND,
3060 BUILT_IN_CHKP_BNDMK, BUILT_IN_CHKP_BNDSTX, BUILT_IN_CHKP_BNDCL,
3061 BUILT_IN_CHKP_BNDCU, BUILT_IN_CHKP_BNDLDX, BUILT_IN_CHKP_BNDRET,
3062 BUILT_IN_CHKP_INTERSECT, BUILT_IN_CHKP_ARG_BND, BUILT_IN_CHKP_NARROW,
3063 BUILT_IN_CHKP_EXTRACT_LOWER, BUILT_IN_CHKP_EXTRACT_UPPER.
3064 * common.opt (fcheck-pointer-bounds): New.
3065 * toplev.c (process_options): Check Pointer Bounds Checker is
3067 * doc/extend.texi: Document Pointer Bounds Checker built-in functions.
3069 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
3071 * target.def (builtin_chkp_function): New.
3072 (chkp_bound_type): New.
3073 (chkp_bound_mode): New.
3074 (fn_abi_va_list_bounds_size): New.
3075 (load_bounds_for_arg): New.
3076 (store_bounds_for_arg): New.
3077 * targhooks.h (default_load_bounds_for_arg): New.
3078 (default_store_bounds_for_arg): New.
3079 (default_fn_abi_va_list_bounds_size): New.
3080 (default_chkp_bound_type): New.
3081 (default_chkp_bound_mode): New.
3082 (default_builtin_chkp_function): New.
3083 * targhooks.c (default_load_bounds_for_arg): New.
3084 (default_store_bounds_for_arg): New.
3085 (default_fn_abi_va_list_bounds_size): New.
3086 (default_chkp_bound_type): New.
3087 (default_chkp_bound_mode); New.
3088 (default_builtin_chkp_function): New.
3089 * doc/tm.texi.in (TARGET_FN_ABI_VA_LIST_BOUNDS_SIZE): New.
3090 (TARGET_LOAD_BOUNDS_FOR_ARG): New.
3091 (TARGET_STORE_BOUNDS_FOR_ARG): New.
3092 (TARGET_BUILTIN_CHKP_FUNCTION): New.
3093 (TARGET_CHKP_BOUND_TYPE): New.
3094 (TARGET_CHKP_BOUND_MODE): New.
3095 * doc/tm.texi: Regenerated.
3096 * langhooks.h (lang_hooks): Add chkp_supported field.
3097 * langhooks-def.h (LANG_HOOKS_CHKP_SUPPORTED): New.
3098 (LANG_HOOKS_INITIALIZER); Add LANG_HOOKS_CHKP_SUPPORTED.
3100 2013-10-29 Andrew Pinski <apinski@cavium.com>
3102 * tree-ssa-ifcombine.c: Include rtl.h and tm_p.h.
3103 (ifcombine_ifandif): Handle cases where maybe_fold_and_comparisons
3104 fails, combining the branches anyways.
3105 (tree_ssa_ifcombine): Inverse the order of the basic block walk,
3106 increases the number of combinings.
3107 * gimple.h (gsi_start_nondebug_after_labels_bb): New function.
3109 2013-10-29 Mike Stump <mikestump@comcast.net>
3111 * machmode.def (PARTIAL_INT_MODE): Add precision and name.
3112 * genmodes.c (PARTIAL_INT_MODE): Add precision and name.
3113 (make_vector_mode): Increase namebuf to 16.
3114 (emit_insn_modes_h): When processing BImode, don't
3115 also match partial int modes.
3116 (emit_class_narrowest_mode): Likewise.
3118 * config/bfin/bfin-modes.def: Add precision to PDI.
3119 * config/m32c/m32c-modes.def: Add precision to PSI.
3120 * config/msp430/msp430-modes.def: Add precision to PSI.
3121 * config/rs6000/rs6000-modes.def: Add precision to PTI.
3122 * config/sh/sh-modes.def: Add precision to PSI and PDI.
3124 2013-10-29 Oleg Endo <olegendo@gcc.gnu.org>
3127 * config/sh/sh.md (*addc): Rename existing variations to ...
3128 (*addc_r_r_1, *addc_2r_1, *addc_r_1): ... these.
3129 (*addc_r_lsb, *addc_r_r_lsb, *addc_r_lsb_r, *addc_2r_lsb, *addc_r_msb,
3130 *addc_r_r_msb, *addc_2r_msb): New insn_and_split patterns.
3131 * config/sh/sh.c (addsubcosts): Handle some addc special cases.
3133 2013-10-29 Teresa Johnson <tejohnson@google.com>
3136 * tree-ssa-tail-merge.c (replace_block_by): Tolerate profile
3137 insanities when updating probabilities.
3139 2013-10-29 David Malcolm <dmalcolm@redhat.com>
3141 * gdbhooks.py (CGraphNodePrinter.to_string): Update gdb
3142 prettyprinter for cgraph_node to reflect the conversion of the
3143 symtable types to a C++ class hierarchy: it now *is* a
3144 symtab_node_base, rather than having one (named "symbol").
3146 2013-10-29 Balaji V. Iyer <balaji.v.iyer@intel.com>
3148 * builtins.c (is_builtin_name): Added a check for __cilkrts_detach and
3149 __cilkrts_pop_frame. If matched, then return true for built-in
3151 (expand_builtin): Added BUILT_IN_CILK_DETACH and
3152 BUILT_IN_CILK_POP_FRAME case.
3153 * langhooks-def.h (lhd_install_body_with_frame_cleanup): New prototype.
3154 (lhs_cilk_detect_spawn): Likewise.
3155 (LANG_HOOKS_DECLS): Added LANG_HOOKS_CILKPLUS.
3156 (LANG_HOOKS_CILKPLUS_DETECT_SPAWN_AND_UNWRAP): New #define.
3157 (LANG_HOOKS_CILKPLUS_FRAME_CLEANUP): Likewise.
3158 (LANG_HOOKS_CILKPLUS_GIMPLIFY_SPAWN): Likewise.
3159 (LANG_HOOKS_CILKPLUS): Likewise.
3160 * tree.h (CILK_SPAWN_FN): Likewise.
3161 * builtin.def (DEF_CILK_BUILTIN_STUB): Likewise.
3162 * Makefile.in (C_COMMON_OBJS): Added c-family/cilk.o.
3163 (OBJS): Added cilk-common.o.
3164 (BUILTINS_DEF): Added cilk-builtins.def.
3165 * langhooks.c (lhd_install_body_with_frame_cleanup): New function.
3166 (lhd_cilk_detect_spawn): Likewise.
3167 * langhooks.h (lang_hooks_for_cilkplus): New struct.
3168 (struct lang_hooks): Added new field called "cilkplus."
3169 * cilk-common.c: New file.
3171 * cilk-builtins.def: Likewise.
3172 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Added
3173 "__cilk" macro and set it to 200.
3174 * function.h (struct function::cilk_frame_decl): New field.
3175 (struct function::is_cilk_function): Likewise.
3176 (struct function::calls_cilk_spawn): Likewise.
3177 * gimplify.c (gimplify_call_expr): Added a check if the function call
3178 being gimplified is a spawn detach point. If so, then add pop_frame
3179 and detach function calls.
3180 (gimplify_expr): Added a CILK_SPAWN_STMT and CILK_SYNC_STMT case
3181 for gimplifying _Cilk_spawn and _Cilk_sync statements.
3182 (gimplify_return_expr): Added a check for _Cilk_spawn usage in
3183 function. If so, added a _Cilk_sync and gimplified it.
3184 (gimplify_modify_expr): Added a check for _Cilk_spawn in MODIFY and
3185 INIT_EXPRs. If so, then call gimplify_cilk_spawn.
3186 * ipa-inline-analysis (initialize_inline_failed): Prevent inlining of
3188 (can_inline_edge_p): Prevent inling of spawnee function.
3189 * ira.c (ira_setup_eliminable_regset): Force usage of frame pointer
3190 for functions that use Cilk keywords.
3191 * tree-inline.h (struct copy_body_data::remap_var_for_cilk): New field.
3192 * tree-pretty-print.c (dump_generic_node): Added CILK_SPAWN_STMT and
3193 CILK_SYNC_STMT cases.
3194 * tree.def (DEFTREECODE): Added CILK_SPAWN_STMT and CILK_SYNC_STMT
3196 * generic.texi (CILK_SPAWN_STMT): Added documentation for _Cilk_spawn.
3197 (CILK_SYNC_STMT): Added documentation for _Cilk_sync.
3198 * passes.texi (Cilk Keywords): New section that describes the compiler
3199 code changes for handling Cilk Keywords.
3201 2013-10-29 David Malcolm <dmalcolm@redhat.com>
3203 Patch autogenerated by refactor_symtab.py from
3204 https://github.com/davidmalcolm/gcc-refactoring-scripts
3205 revision 58bb219cc090b2f4516a9297d868c245495ee622
3207 * asan.c (asan_finish_file): Update for conversion of symtab types to
3208 a true class hierarchy.
3209 * cfgexpand.c (estimated_stack_frame_size): Likewise.
3210 * cgraph.c (cgraph_get_body): Likewise.
3211 (cgraph_get_create_real_symbol_node): Likewise.
3212 (verify_cgraph_node): Likewise.
3213 (verify_edge_corresponds_to_fndecl): Likewise.
3214 (verify_edge_count_and_frequency): Likewise.
3215 (cgraph_will_be_removed_from_program_if_no_direct_calls): Likewise.
3216 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
3217 (cgraph_can_remove_if_no_direct_calls_and_refs_p): Likewise.
3218 (cgraph_node_cannot_return): Likewise.
3219 (cgraph_set_pure_flag_1): Likewise.
3220 (cgraph_set_const_flag_1): Likewise.
3221 (cgraph_set_nothrow_flag_1): Likewise.
3222 (cgraph_make_node_local_1): Likewise.
3223 (cgraph_for_node_and_aliases): Likewise.
3224 (cgraph_for_node_thunks_and_aliases): Likewise.
3225 (cgraph_node_can_be_local_p): Likewise.
3226 (cgraph_node_cannot_be_local_p_1): Likewise.
3227 (cgraph_function_body_availability): Likewise.
3228 (dump_cgraph_node): Likewise.
3229 (cgraph_rtl_info): Likewise.
3230 (cgraph_mark_address_taken_node): Likewise.
3231 (cgraph_remove_node): Likewise.
3232 (cgraph_release_function_body): Likewise.
3233 (cgraph_update_edges_for_call_stmt_node): Likewise.
3234 (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
3235 (cgraph_make_edge_direct): Likewise.
3236 (cgraph_resolve_speculation): Likewise.
3237 (cgraph_speculative_call_info): Likewise.
3238 (cgraph_turn_edge_to_speculative): Likewise.
3239 (cgraph_create_edge_1): Likewise.
3240 (cgraph_set_call_stmt): Likewise.
3241 (cgraph_node_for_asm): Likewise.
3242 (cgraph_add_thunk): Likewise.
3243 (cgraph_same_body_alias): Likewise.
3244 (cgraph_create_function_alias): Likewise.
3245 (cgraph_create_node): Likewise.
3246 (cgraph_create_empty_node): Likewise.
3247 (record_function_versions): Likewise.
3248 (used_from_object_file_p): Likewise.
3249 * cgraph.h (symtab_can_be_discarded): Likewise.
3250 (symtab_real_symbol_p): Likewise.
3251 (cgraph_mark_force_output_node): Likewise.
3252 (cgraph_edge_recursive_p): Likewise.
3253 (symtab_alias_target): Likewise.
3254 (varpool_all_refs_explicit_p): Likewise.
3255 (varpool_can_remove_if_no_refs): Likewise.
3256 (cgraph_only_called_directly_or_aliased_p): Likewise.
3257 (cgraph_next_function_with_gimple_body): Likewise.
3258 (cgraph_first_function_with_gimple_body): Likewise.
3259 (cgraph_function_with_gimple_body_p): Likewise.
3260 (cgraph_next_function): Likewise.
3261 (cgraph_first_function): Likewise.
3262 (cgraph_next_defined_function): Likewise.
3263 (cgraph_first_defined_function): Likewise.
3264 (varpool_next_defined_variable): Likewise.
3265 (varpool_first_defined_variable): Likewise.
3266 (varpool_next_static_initializer): Likewise.
3267 (varpool_first_static_initializer): Likewise.
3268 (varpool_next_variable): Likewise.
3269 (varpool_first_variable): Likewise.
3270 (varpool_node_name): Likewise.
3271 (varpool): Likewise.
3273 (is_a_helper <varpool_node>::test): Likewise.
3274 (is_a_helper <cgraph_node>::test): Likewise.
3275 (varpool_variable_node): Likewise.
3276 (cgraph_function_or_thunk_node): Likewise.
3277 (varpool_alias_target): Likewise.
3278 (cgraph_alias_target): Likewise.
3279 (cgraph_node_name): Likewise.
3280 (varpool_node_asm_name): Likewise.
3281 (cgraph_node_asm_name): Likewise.
3282 * cgraphbuild.c (remove_cgraph_callee_edges): Likewise.
3283 (cgraph_rebuild_references): Likewise.
3284 (rebuild_cgraph_edges): Likewise.
3285 (record_eh_tables): Likewise.
3286 (build_cgraph_edges): Likewise.
3287 (mark_store): Likewise.
3288 (mark_load): Likewise.
3289 (mark_address): Likewise.
3290 (record_type_list): Likewise.
3291 (record_reference): Likewise.
3292 * cgraphclones.c (cgraph_materialize_all_clones): Likewise.
3293 (cgraph_materialize_clone): Likewise.
3294 (cgraph_function_versioning): Likewise.
3295 (cgraph_copy_node_for_versioning): Likewise.
3296 (update_call_expr): Likewise.
3297 (cgraph_find_replacement_node): Likewise.
3298 (cgraph_create_virtual_clone): Likewise.
3299 (cgraph_clone_node): Likewise.
3300 * cgraphunit.c (compile): Likewise.
3301 (output_weakrefs): Likewise.
3302 (output_in_order): Likewise.
3303 (expand_function): Likewise.
3304 (assemble_thunks_and_aliases): Likewise.
3305 (expand_thunk): Likewise.
3306 (mark_functions_to_output): Likewise.
3307 (handle_alias_pairs): Likewise.
3308 (analyze_functions): Likewise.
3309 (walk_polymorphic_call_targets): Likewise.
3310 (varpool_finalize_decl): Likewise.
3311 (process_function_and_variable_attributes): Likewise.
3312 (cgraph_process_same_body_aliases): Likewise.
3313 (analyze_function): Likewise.
3314 (cgraph_add_new_function): Likewise.
3315 (cgraph_finalize_function): Likewise.
3316 (referred_to_p): Likewise.
3317 (cgraph_reset_node): Likewise.
3318 (cgraph_process_new_functions): Likewise.
3319 (enqueue_node): Likewise.
3320 (decide_is_symbol_needed): Likewise.
3321 * coverage.c (coverage_compute_profile_id): Likewise.
3322 * dbxout.c (dbxout_expand_expr): Likewise.
3323 * dwarf2out.c (premark_types_used_by_global_vars_helper): Likewise.
3324 (reference_to_unused): Likewise.
3325 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
3326 * gimplify.c (unvisit_body): Likewise.
3327 (unshare_body): Likewise.
3328 * ipa-cp.c (ipcp_generate_summary): Likewise.
3329 (ipcp_decision_stage): Likewise.
3330 (identify_dead_nodes): Likewise.
3331 (decide_whether_version_node): Likewise.
3332 (decide_about_value): Likewise.
3333 (perhaps_add_new_callers): Likewise.
3334 (create_specialized_node): Likewise.
3335 (update_profiling_info): Likewise.
3336 (ipcp_propagate_stage): Likewise.
3337 (estimate_local_effects): Likewise.
3338 (good_cloning_opportunity_p): Likewise.
3339 (devirtualization_time_bonus): Likewise.
3340 (propagate_constants_accross_call): Likewise.
3341 (initialize_node_lattices): Likewise.
3342 (ipcp_cloning_candidate_p): Likewise.
3343 (determine_versionability): Likewise.
3344 (print_all_lattices): Likewise.
3345 (print_lattice): Likewise.
3346 (ipcp_discover_new_direct_edges): Likewise.
3347 * ipa-devirt.c (ipa_devirt): Likewise.
3348 (likely_target_p): Likewise.
3349 (update_type_inheritance_graph): Likewise.
3350 (possible_polymorphic_call_target_p): Likewise.
3351 (dump_possible_polymorphic_call_targets): Likewise.
3352 (devirt_variable_node_removal_hook): Likewise.
3353 (record_binfo): Likewise.
3354 (maybe_record_node): Likewise.
3355 (build_type_inheritance_graph): Likewise.
3356 * ipa-inline-analysis.c (inline_write_summary): Likewise.
3357 (inline_generate_summary): Likewise.
3358 (inline_analyze_function): Likewise.
3359 (do_estimate_growth): Likewise.
3360 (simple_edge_hints): Likewise.
3361 (estimate_node_size_and_time): Likewise.
3362 (estimate_edge_devirt_benefit): Likewise.
3363 (compute_inline_parameters): Likewise.
3364 (estimate_function_body_sizes): Likewise.
3365 (compute_bb_predicates): Likewise.
3366 (initialize_inline_failed): Likewise.
3367 (dump_inline_summary): Likewise.
3368 (dump_inline_edge_summary): Likewise.
3369 * ipa-inline-transform.c (inline_transform): Likewise.
3370 (preserve_function_body_p): Likewise.
3371 (save_inline_function_body): Likewise.
3372 (inline_call): Likewise.
3373 (clone_inlined_nodes): Likewise.
3374 (can_remove_node_now_p): Likewise.
3375 (can_remove_node_now_p_1): Likewise.
3376 * ipa-inline.c (early_inliner): Likewise.
3377 (early_inline_small_functions): Likewise.
3378 (inline_always_inline_functions): Likewise.
3379 (ipa_inline): Likewise.
3380 (flatten_function): Likewise.
3381 (inline_small_functions): Likewise.
3382 (speculation_useful_p): Likewise.
3383 (recursive_inlining): Likewise.
3384 (update_caller_keys): Likewise.
3385 (reset_edge_caches): Likewise.
3386 (update_edge_key): Likewise.
3387 (edge_badness): Likewise.
3388 (relative_time_benefit): Likewise.
3389 (want_inline_self_recursive_call_p): Likewise.
3390 (want_inline_small_function_p): Likewise.
3391 (want_early_inline_function_p): Likewise.
3392 (num_calls): Likewise.
3393 (can_early_inline_edge_p): Likewise.
3394 (can_inline_edge_p): Likewise.
3395 (report_inline_failed_reason): Likewise.
3396 * ipa-profile.c (ipa_profile): Likewise.
3397 (ipa_propagate_frequency): Likewise.
3398 (ipa_propagate_frequency_1): Likewise.
3399 (ipa_profile_generate_summary): Likewise.
3400 * ipa-prop.c (ipcp_transform_function): Likewise.
3401 (read_replacements_section): Likewise.
3402 (ipa_prop_read_section): Likewise.
3403 (ipa_modify_call_arguments): Likewise.
3404 (ipa_print_node_params): Likewise.
3405 (propagate_controlled_uses): Likewise.
3406 (update_indirect_edges_after_inlining): Likewise.
3407 (remove_described_reference): Likewise.
3408 (ipa_make_edge_direct_to_target): Likewise.
3409 (ipa_analyze_node): Likewise.
3410 (ipa_analyze_params_uses): Likewise.
3411 (ipa_compute_jump_functions): Likewise.
3412 (ipa_get_callee_param_type): Likewise.
3413 (ipa_print_node_jump_functions): Likewise.
3414 (ipa_initialize_node_params): Likewise.
3415 (ipa_populate_param_decls): Likewise.
3416 (ipa_func_spec_opts_forbid_analysis_p): Likewise.
3417 (write_agg_replacement_chain): Likewise.
3418 (ipa_write_node_info): Likewise.
3419 (ipa_edge_duplication_hook): Likewise.
3420 (try_decrement_rdesc_refcount): Likewise.
3421 * ipa-pure-const.c (propagate_nothrow): Likewise.
3422 (propagate_pure_const): Likewise.
3423 (pure_const_read_summary): Likewise.
3424 (pure_const_write_summary): Likewise.
3425 (analyze_function): Likewise.
3426 * ipa-ref-inline.h (ipa_ref_referred_ref_list): Likewise.
3427 (ipa_ref_referring_ref_list): Likewise.
3428 * ipa-ref.c (ipa_clear_stmts_in_references): Likewise.
3429 (ipa_remove_stmt_references): Likewise.
3430 (ipa_find_reference): Likewise.
3431 (ipa_dump_referring): Likewise.
3432 (ipa_dump_references): Likewise.
3433 (ipa_record_reference): Likewise.
3434 * ipa-reference.c (ipa_reference_read_optimization_summary): Likewise.
3435 (ipa_reference_write_optimization_summary): Likewise.
3436 (write_node_summary_p): Likewise.
3437 (propagate): Likewise.
3438 (read_write_all_from_decl): Likewise.
3439 (generate_summary): Likewise.
3440 (analyze_function): Likewise.
3441 (propagate_bits): Likewise.
3442 (ipa_reference_get_not_written_global): Likewise.
3443 (ipa_reference_get_not_read_global): Likewise.
3444 * ipa-split.c (execute_split_functions): Likewise.
3445 (split_function): Likewise.
3446 * ipa-utils.c (ipa_merge_profiles): Likewise.
3447 (dump_cgraph_node_set): Likewise.
3448 (ipa_reverse_postorder): Likewise.
3449 (ipa_edge_within_scc): Likewise.
3450 (ipa_get_nodes_in_cycle): Likewise.
3451 (ipa_free_postorder_info): Likewise.
3452 (ipa_reduced_postorder): Likewise.
3453 (searchc): Likewise.
3454 (recursive_call_p): Likewise.
3455 * ipa.c (ipa_cdtor_merge): Likewise.
3456 (record_cdtor_fn): Likewise.
3457 (function_and_variable_visibility): Likewise.
3458 (varpool_externally_visible_p): Likewise.
3459 (cgraph_externally_visible_p): Likewise.
3460 (comdat_can_be_unshared_p): Likewise.
3461 (comdat_can_be_unshared_p_1): Likewise.
3462 (address_taken_from_non_vtable_p): Likewise.
3463 (ipa_discover_readonly_nonaddressable_vars): Likewise.
3464 (symtab_remove_unreachable_nodes): Likewise.
3465 (walk_polymorphic_call_targets): Likewise.
3466 (process_references): Likewise.
3467 (enqueue_node): Likewise.
3468 (has_addr_references_p): Likewise.
3469 (cgraph_non_local_node_p_1): Likewise.
3470 * is-a.h (varpool_analyze_node): Likewise.
3471 * lto-cgraph.c (input_symtab): Likewise.
3472 (merge_profile_summaries): Likewise.
3473 (input_cgraph_1): Likewise.
3474 (input_edge): Likewise.
3475 (input_varpool_node): Likewise.
3476 (input_node): Likewise.
3477 (input_overwrite_node): Likewise.
3478 (compute_ltrans_boundary): Likewise.
3479 (output_refs): Likewise.
3480 (lto_output_varpool_node): Likewise.
3481 (lto_output_node): Likewise.
3482 (reachable_from_other_partition_p): Likewise.
3483 (referenced_from_other_partition_p): Likewise.
3484 (lto_output_edge): Likewise.
3485 (output_node_opt_summary): Likewise.
3486 (add_node_to): Likewise.
3487 (reachable_from_this_partition_p): Likewise.
3488 (lto_set_symtab_encoder_in_partition): Likewise.
3489 (lto_symtab_encoder_in_partition_p): Likewise.
3490 (lto_set_symtab_encoder_encode_initializer): Likewise.
3491 (lto_symtab_encoder_encode_initializer_p): Likewise.
3492 (lto_set_symtab_encoder_encode_body): Likewise.
3493 (lto_symtab_encoder_encode_body_p): Likewise.
3494 * lto-section-in.c (lto_free_function_in_decl_state_for_node):
3496 * lto-streamer-in.c (lto_read_body): Likewise.
3497 (fixup_call_stmt_edges): Likewise.
3498 (fixup_call_stmt_edges_1): Likewise.
3499 * lto-streamer-out.c (produce_symtab): Likewise.
3500 (output_symbol_p): Likewise.
3501 (write_symbol): Likewise.
3502 (lto_output): Likewise.
3503 (copy_function): Likewise.
3504 (output_function): Likewise.
3505 * passes.c (function_called_by_processed_nodes_p): Likewise.
3506 (ipa_write_optimization_summaries): Likewise.
3507 (ipa_write_summaries): Likewise.
3508 (do_per_function_toporder): Likewise.
3509 (do_per_function): Likewise.
3510 (dump_passes): Likewise.
3511 * symtab.c (symtab_semantically_equivalent_p): Likewise.
3512 (symtab_nonoverwritable_alias): Likewise.
3513 (symtab_nonoverwritable_alias_1): Likewise.
3514 (symtab_for_node_and_aliases): Likewise.
3515 (symtab_resolve_alias): Likewise.
3516 (fixup_same_cpp_alias_visibility): Likewise.
3517 (symtab_alias_ultimate_target): Likewise.
3518 (symtab_used_from_object_file_p): Likewise.
3519 (verify_symtab_base): Likewise.
3520 (dump_symtab_base): Likewise.
3521 (symtab_node_name): Likewise.
3522 (symtab_node_asm_name): Likewise.
3523 (symtab_dissolve_same_comdat_group_list): Likewise.
3524 (symtab_add_to_same_comdat_group): Likewise.
3525 (symtab_unregister_node): Likewise.
3526 (symtab_insert_node_to_hashtable): Likewise.
3527 (symtab_register_node): Likewise.
3528 (unlink_from_assembler_name_hash): Likewise.
3529 (insert_to_assembler_name_hash): Likewise.
3530 (eq_assembler_name): Likewise.
3531 (hash_node_by_assembler_name): Likewise.
3532 (eq_node): Likewise.
3533 (hash_node): Likewise.
3534 * toplev.c (wrapup_global_declaration_2): Likewise.
3535 * trans-mem.c (ipa_tm_execute): Likewise.
3536 (ipa_tm_transform_clone): Likewise.
3537 (ipa_tm_transform_transaction): Likewise.
3538 (ipa_tm_transform_calls_redirect): Likewise.
3539 (ipa_tm_insert_gettmclone_call): Likewise.
3540 (ipa_tm_insert_irr_call): Likewise.
3541 (ipa_tm_create_version): Likewise.
3542 (ipa_tm_create_version_alias): Likewise.
3543 (ipa_tm_mark_forced_by_abi_node): Likewise.
3544 (ipa_tm_mark_force_output_node): Likewise.
3545 (ipa_tm_diagnose_tm_safe): Likewise.
3546 (ipa_tm_mayenterirr_function): Likewise.
3547 (ipa_tm_scan_irr_function): Likewise.
3548 (ipa_tm_note_irrevocable): Likewise.
3549 (ipa_tm_scan_calls_clone): Likewise.
3550 (get_cg_data): Likewise.
3551 * tree-eh.c (tree_could_trap_p): Likewise.
3552 * tree-emutls.c (ipa_lower_emutls): Likewise.
3553 (create_emultls_var): Likewise.
3554 (lower_emutls_function_body): Likewise.
3555 (gen_emutls_addr): Likewise.
3556 (emutls_decl): Likewise.
3557 (new_emutls_decl): Likewise.
3558 * tree-inline.c (tree_function_versioning): Likewise.
3559 (optimize_inline_calls): Likewise.
3560 (expand_call_inline): Likewise.
3561 (estimate_num_insns): Likewise.
3562 (copy_bb): Likewise.
3563 (delete_unreachable_blocks_update_callgraph): Likewise.
3564 * tree-nested.c (gimplify_all_functions): Likewise.
3565 (create_nesting_tree): Likewise.
3566 (check_for_nested_with_variably_modified): Likewise.
3567 * tree-pretty-print.c (dump_function_header): Likewise.
3568 * tree-profile.c (tree_profiling): Likewise.
3569 * tree-sra.c (ipa_sra_preliminary_function_checks): Likewise.
3570 (modify_function): Likewise.
3571 (convert_callers): Likewise.
3572 (convert_callers_for_node): Likewise.
3573 * tree-ssa-structalias.c (ipa_pta_execute): Likewise.
3574 (associate_varinfo_to_alias): Likewise.
3575 (create_variable_info_for): Likewise.
3576 (get_constraint_for_ssa_var): Likewise.
3577 * tree-vectorizer.c (increase_alignment): Likewise.
3578 * tree.c (find_decls_types_in_var): Likewise.
3579 (find_decls_types_in_node): Likewise.
3580 (free_lang_data_in_decl): Likewise.
3581 * value-prof.c (gimple_ic_transform): Likewise.
3582 (gimple_ic): Likewise.
3583 (check_ic_target): Likewise.
3584 (init_node_map): Likewise.
3585 * varasm.c (decl_binds_to_current_def_p): Likewise.
3586 (default_binds_local_p_1): Likewise.
3587 (dump_tm_clone_pairs): Likewise.
3588 (assemble_alias): Likewise.
3589 (find_decl): Likewise.
3590 (mark_decl_referenced): Likewise.
3591 * varpool.c (varpool_for_node_and_aliases): Likewise.
3592 (varpool_extra_name_alias): Likewise.
3593 (varpool_create_variable_alias): Likewise.
3594 (add_new_static_var): Likewise.
3595 (varpool_finalize_named_section_flags): Likewise.
3596 (varpool_remove_unreferenced_decls): Likewise.
3597 (enqueue_node): Likewise.
3598 (varpool_assemble_decl): Likewise.
3599 (assemble_aliases): Likewise.
3600 (varpool_analyze_node): Likewise.
3601 (cgraph_variable_initializer_availability): Likewise.
3602 (varpool_add_new_variable): Likewise.
3603 (ctor_for_folding): Likewise.
3604 (dump_varpool_node): Likewise.
3605 (varpool_remove_initializer): Likewise.
3606 (varpool_remove_node): Likewise.
3607 (varpool_node_for_decl): Likewise.
3608 (varpool_create_empty_node): Likewise.
3609 * config/i386/i386.c (ix86_generate_version_dispatcher_body): Likewise.
3610 (ix86_get_function_versions_dispatcher): Likewise.
3612 2013-10-29 David Malcolm <dmalcolm@redhat.com>
3614 * cgraph.h (symtab_node_base): Convert to a class;
3615 add GTY((desc ("%h.type"), tag ("SYMTAB_SYMBOL"))), and take
3616 chain_next/prev from symtab_node_def.
3617 (cgraph_node): Inherit from symtab_node; add GTY option
3618 tag ("SYMTAB_FUNCTION").
3619 (varpool_node): Inherit from symtab_node; add GTY option
3620 tag ("SYMTAB_VARIABLE").
3621 (symtab_node_def): Remove.
3622 (is_a_helper <cgraph_node>::test (symtab_node_def *)): Convert to...
3623 (is_a_helper <cgraph_node>::test (symtab_node_base *)): ...this.
3624 (is_a_helper <varpool_node>::test (symtab_node_def *)): Convert to...
3625 (is_a_helper <varpool_node>::test (symtab_node_base *)): ...this.
3627 * ipa-ref.h (symtab_node_def): Drop.
3628 (symtab_node): Change underlying type from symtab_node_def to
3630 (const_symtab_node): Likwise.
3632 * is-a.h: Update examples in comment.
3634 * symtab.c (symtab_hash): Change symtab_node_def to symtab_node_base.
3635 (assembler_name_hash): Likewise.
3637 2013-10-29 Martin Liska <marxin.liska@gmail.com>
3639 * doc/tree-ssa.texi (gimple_phi_result): Document.
3640 (gimple_phi_num_args, gimple_phi_arg): Likewise.
3641 (gimple_phi_arg_edge, gimple_phi_arg_def): Likewise.
3642 (PHI_RESULT, PHI_NUM_ARGS): Remove.
3643 (PHI_ARG_ELT, PHI_ARG_EDGE, PHI_ARG_DEF): Likewise.
3645 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
3647 * expr.h: Revert change and include tree-core.h.
3648 * rtl.h: Revert change and don't include tree-core.h.
3650 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
3652 * config/darwin.c: Include gimple.h.
3653 * config/i386/winnt.c: Likewise.
3655 2013-10-29 Marc Glisse <marc.glisse@inria.fr>
3657 PR tree-optimization/19831
3658 * tree-ssa-alias.c (stmt_kills_ref_p_1): Handle BUILT_IN_FREE.
3660 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
3662 * tree-outof-ssa.h: Remove include files.
3663 * tree-outof-ssa.c: Add required include files from tree-outof-ssa.h.
3665 * tree-ssa-coalesce.c: Likewise.
3666 * cfgexpand.c: Likewise.
3667 * tree-ssa-ter.c: Likewise.
3668 * ipa-prop.h: Remove gimple.h and tree-core.h from include list.
3669 * lto-streamer.h: Likewise.
3670 * cgraphbuild.c: Add gimple.h to include list.
3671 * data-streamer-in.c: Likewise.
3672 * ipa-cp.c: Likewise.
3673 * tree-streamer.c: Likewise.
3674 * lto-compress.c: Likewise.
3675 * ipa-reference.c: Likewise.
3676 * data-streamer-out.c: Likewise.
3677 * lto-cgraph.c: Likewise.
3678 * cgraphclones.c: Likewise.
3679 * ipa-utils.c: Likewise.
3680 * data-streamer.c: Likewise.
3681 * ipa-split.c: Likewise.
3682 * lto-section-in.c: Likewise.
3683 * tree-streamer-out.c: Likewise.
3684 * ipa-prop.c: Likewise.
3685 * tree-streamer-in.c: Likewise.
3686 * symtab.c: Likewise.
3687 * opts-global.c: Likewise.
3688 * lto-opts.c: Likewise.
3689 * lto-section-out.c: Likewise.
3690 * lto-streamer.c: Likewise.
3691 * rtl.h: Add tree-core.h to include list.
3692 * expr.h: Remove tree-core.h from include list.
3693 * gimple.h: Likewise.
3694 * ipa-utils.h: Likewise.
3695 * streamer-hooks.h: Likewise.
3696 * streamer-hooks.c: Include input.h.
3698 2013-10-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3700 * config/arm/arm.c (cortexa7_extra_costs): New table.
3701 (arm_cortex_a7_tune): New.
3702 * config/arm/arm-cores.def: Use cortex_a7 tuning for cortex-a7.
3704 2013-10-29 Eric Botcazou <ebotcazou@adacore.com>
3706 * expr.c (expand_expr_real_1) <MEM_EXPR>: Eliminate small redundancy.
3708 2013-10-29 David Malcolm <dmalcolm@redhat.com>
3710 * doc/gty.texi ("Inheritance and GTY"): Make it clear that
3711 to use autogenerated markers for a class-hierarchy, every class
3712 must have a GTY marker.
3713 * gengtype.h (struct type): Add linked list of subclasses to
3714 the "s" member of the union.
3715 (add_subclass): New decl.
3716 * gengtype-state.c (read_state_struct_type): Set up subclass
3718 * gengtype.c (get_ultimate_base_class): New.
3719 (add_subclass): New.
3720 (new_structure): Set up subclass linked list.
3721 (set_gc_used_type): Propagate usage information to subclasses.
3722 (output_mangled_typename): Use get_ultimate_base_class.
3723 (walk_subclasses): Use the subclass linked list, avoiding an
3724 O(N^2) when writing out all types.
3725 (walk_type): Issue an error if the base class is missing a tag,
3726 rather than generating bogus C code. Add a gcc_unreachable
3727 default case, in case people omit tags from concrete subclasses,
3728 or get the values wrong.
3729 (write_func_for_structure): Issue an error for subclasses for
3730 which the base doesn't have a "desc", since otherwise the
3731 autogenerated routines for the base would silently fail to visit
3732 any subclass fields.
3733 (write_root): Use get_ultimate_base_class, tweaking constness of
3734 tp to match that function's signature.
3736 2013-10-29 David Malcolm <dmalcolm@redhat.com>
3738 * doc/gty.texi (GTY Options): Add note about inheritance to
3739 description of desc and tag.
3740 (Inheritance and GTY): New.
3742 2013-10-29 David Malcolm <dmalcolm@redhat.com>
3744 * gengtype-parse.c (opts_have): Drop "static" so that
3745 we can use this from gengtype.c.
3746 * gengtype.c (set_gc_used_type): Mark any base class as used;
3747 update field traversal to visit inherited fields.
3748 (output_mangled_typename): Convert references to classes within
3749 an inheritance hierarchy to reference the ultimate base class,
3750 since only it will have gt_ functions.
3751 (get_string_option): New.
3752 (walk_subclasses): New.
3753 (walk_type): Treat GTY structs that have a "desc" as being the
3754 root of an inheritance hierarchy. Generate a switch on it
3755 within the marking function which walks all subclasses, adding
3756 cases for them via walk_subclasses. For subclasses, visit all
3757 fields of the type (including inherited ones).
3758 (write_func_for_structure): Don't write fns for subclasses, only
3759 for the ultimate base class within an inheritance hierarchy.
3760 Subclasses-marking will be handled by the base class marking functions.
3761 (write_types): Likewise.
3762 (write_local_func_for_structure): Likewise.
3763 (USED_BY_TYPED_GC_P): Emit allocators for subclasses that have
3764 a "tag" option (and are thus concrete subclasses).
3765 (write_root): Use the marker function for the ultimate base class.
3766 * gengtype.h (FOR_ALL_INHERITED_FIELDS): New.
3767 (opts_have): Add declaration.
3769 2013-10-28 Vladimir Makarov <vmakarov@redhat.com>
3771 * lra-spills.c (lra_final_code_change): Remove useless move insns
3772 originated from moves of pseudos.
3774 2013-10-28 Jeff Law <law@redhat.com>
3776 * lower-subreg.c (resolve_simple_move): Fix comment typo.
3778 2013-10-28 Trevor Saunders <tsaunders@mozilla.com>
3780 * df-scan.c (df_collection_rec): Adjust.
3781 (copy_defs): New constant.
3782 (copy_uses): Likewise.
3783 (copy_eq_uses): Likewise.
3784 (copy_mw): Likewise.
3785 (copy_all): Likewise.
3786 (df_insn_rescan): Adjust.
3787 (df_notes_rescan): Likewise.
3788 (df_swap_refs): Likewise.
3789 (df_sort_and_compress_refs): Likewise.
3790 (df_sort_and_compress_mws): Likewise.
3791 (df_install_refs): Likewise.
3792 (df_install_mws): Likewise.
3793 (df_refs_add_to_chains): Add flags parameter controlling which vectors
3795 (df_bb_refs_record): Adjust.
3796 (df_record_entry_block_defs): Likewise.
3797 (df_record_exit_block_defs): Likewise.
3798 (df_refs_verify): Likewise.
3799 (df_mws_verify): Likewise.
3800 (df_insn_refs_verify): Likewise.
3801 (df_bb_verify): Likewise.
3802 * ipa-pure-const.c (finish_state): Remove.
3803 (propagate): Adjust.
3804 * tree-data-ref.c tree-ssa-alias.c tree-ssa-loop-ivcanon.c
3805 tree-ssa-threadedge.c tree-vect-loop-manip.c tree-vect-slp.c
3806 var-tracking.c: Adjust.
3807 * vec.c (stack_vecs): Remove.
3808 (register_stack_vec): Likewise.
3809 (stack_vec_register_index): Likewise.
3810 (unregister_stack_vec): Likewise.
3811 * vec.h (struct va_stack): Remove.
3812 (struct vec<T, A, vl_ptr>): Specialize as
3813 struct vec<T, va_heap, vl_ptr> instead since va_heap is the only
3814 allocation strategy compatable with the vl_ptr layout.
3815 (struct vec<T, va_gc, vl_ptr>): Remove because it now gets an empty
3816 specialization anyway.
3817 (class stack_vec): New class.
3818 (vec_stack_alloc): Remove.
3819 (vec<T, va_heap, vl_ptr>::using_auto_storage): New method.
3821 2013-10-28 Alexander Ivchenko <alexander.ivchenko@intel.com>
3822 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3823 Sergey Lega <sergey.s.lega@intel.com>
3824 Anna Tikhonova <anna.tikhonova@intel.com>
3825 Ilya Tocar <ilya.tocar@intel.com>
3826 Andrey Turetskiy <andrey.turetskiy@intel.com>
3827 Ilya Verbin <ilya.verbin@intel.com>
3828 Kirill Yukhin <kirill.yukhin@intel.com>
3829 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3831 * config/i386/i386.md (prefetch): Allow TARGET_AVX512PF.
3832 (*prefetch_avx512pf_<mode>): New.
3833 * config/i386/sse.md (avx512f_vmcmp<mode>3): Ditto.
3834 (avx512f_maskcmp<mode>3): Ditto.
3835 (vashrv16si3): Ditto.
3837 2013-10-28 Alexander Ivchenko <alexander.ivchenko@intel.com>
3838 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3839 Sergey Lega <sergey.s.lega@intel.com>
3840 Anna Tikhonova <anna.tikhonova@intel.com>
3841 Ilya Tocar <ilya.tocar@intel.com>
3842 Andrey Turetskiy <andrey.turetskiy@intel.com>
3843 Ilya Verbin <ilya.verbin@intel.com>
3844 Kirill Yukhin <kirill.yukhin@intel.com>
3845 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3847 * config/i386/i386.md (any_truncate): New.
3848 (trunsuffix): Ditto.
3849 * config/i386/predicates.md (const_8_to_9_operand): New.
3850 (const_10_to_11_operand): Ditto.
3851 (const_12_to_13_operand): Ditto.
3852 (const_14_to_15_operand): Ditto.
3853 (const_16_to_19_operand): Ditto.
3854 (const_20_to_23_operand): Ditto.
3855 (const_24_to_27_operand): Ditto.
3856 (const_28_to_31_operand): Ditto.
3857 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_FIX_NOTRUNC.
3858 (cvtusi2<ssescalarmodesuffix>32): New.
3859 (cvtusi2<ssescalarmodesuffix>64): Ditto.
3860 (ufloatv16siv16sf2): Ditto.
3861 (avx512f_fix_notruncv16sfv16si): Ditto.
3862 (avx512f_ufix_notruncv16sfv16si): Ditto.
3863 (avx512f_vcvtss2usi): Ditto.
3864 (avx512f_vcvtss2usiq): Ditto.
3865 (avx512f_vcvttss2usi): Ditto.
3866 (avx512f_vcvttss2usiq): Ditto.
3867 (avx512f_vcvtsd2usi): Ditto.
3868 (avx512f_vcvtsd2usiq): Ditto.
3869 (avx512f_vcvttsd2usi): Ditto.
3870 (avx512f_vcvttsd2usiq): Ditto.
3871 (ufloatv8siv8df): Ditto.
3872 (avx512f_cvtdq2pd512_2): Ditto.
3873 (avx512f_cvtpd2dq512): Ditto.
3874 (avx512f_ufix_notruncv8dfv8si): Ditto.
3875 (avx512f_cvtpd2ps512): Ditto.
3876 (vec_unpacks_lo_v16sf): Ditto.
3877 (vec_unpacks_hi_v16sf): Ditto.
3878 (vec_unpacks_float_hi_v16si): Ditto.
3879 (vec_unpacks_float_lo_v16si): Ditto.
3880 (avx512f_unpckhps512): Ditto.
3881 (avx512f_unpcklps512): Ditto.
3882 (avx512f_movshdup512): Ditto.
3883 (avx512f_movsldup512): Ditto.
3884 (vec_extract_lo_v32hi): Ditto.
3885 (vec_extract_hi_v32hi): Ditto.
3886 (vec_extract_lo_v64qi): Ditto.
3887 (vec_extract_hi_v64qi): Ditto.
3888 (avx512f_unpckhpd512): Ditto.
3889 (avx512f_movddup512): Ditto.
3890 (avx512f_unpcklpd512): Ditto.
3891 (*avx512f_unpcklpd512): Ditto.
3892 (avx512f_shufps512_1): Ditto.
3893 (avx512f_shufpd512_1): Ditto.
3894 (avx512f_interleave_highv8di): Ditto.
3895 (avx512f_interleave_lowv8di): Ditto.
3896 (PMOV_DST_MODE): Ditto.
3897 (pmov_src_mode): Ditto.
3898 (pmov_src_lower): Ditto.
3900 (*avx512f_<code><pmov_src_lower><mode>2): Ditto.
3901 (*avx512f_<code>v8div16qi2): Ditto.
3902 (*avx512f_<code>v8div16qi2_store): Ditto.
3903 (vec_widen_umult_even_v16si): Ditto.
3904 (*vec_widen_umult_even_v16si): Ditto.
3905 (vec_widen_smult_even_v16si): Ditto.
3906 (*vec_widen_smult_even_v16si): Ditto.
3907 (avx512f_interleave_highv16si): Ditto.
3908 (avx512f_interleave_lowv16si): Ditto.
3909 (avx512f_<code>v16qiv16si2): Ditto.
3910 (avx512f_<code>v16hiv16si2): Ditto.
3911 (avx512f_<code>v8qiv8di2): Ditto.
3912 (avx512f_<code>v8hiv8di2): Ditto.
3913 (avx512f_<code>v8siv8di2): Ditto.
3914 (avx512cd_maskb_vec_dupv8di): Ditto.
3915 (avx512cd_maskw_vec_dupv16si): Ditto.
3916 (avx512f_vcvtph2ps512): Ditto.
3917 (avx512f_vcvtps2ph512): Ditto.
3918 (VEC_EXTRACT_MODE): Extened with wider modes.
3919 (VEC_PERM_AVX2): Ditto.
3920 (VEC_PERM_CONST): Ditto.
3922 2013-10-28 Joern Rennecke <joern.rennecke@embecosm.com>
3924 * config/arc/arc.c (arc_ccfsm_post_advance):
3925 Add comment about TYPE_RETURN.
3927 2013-10-28 Bin Cheng <bin.cheng@arm.com>
3929 * tree-ssa-loop-ivopts.c (strip_offset_1): Change parameter type.
3930 Count DECL_FIELD_BIT_OFFSET for COMPONENT_REF.
3931 (strip_offset): Convert offset to unsigned number.
3933 2013-10-27 Tom de Vries <tom@codesourcery.com>
3935 * cfgexpand.c (gimple_expand_cfg): Remove test for parm_birth_insn.
3936 Don't commit insertions after NOTE_INSN_FUNCTION_BEG.
3938 2013-10-27 Oleg Endo <olegendo@gcc.gnu.org>
3940 * config/sh/sh.c (MSW, LSW): Move and rename macros to...
3941 * config/sh/sh.h (SH_REG_MSW_OFFSET, SH_REG_LSW_OFFSET): ... here.
3942 (TARGET_BIG_ENDIAN): New macro.
3943 * config/sh/sh.md: Use it instead of !TARGET_LITTLE_ENDIAN.
3944 Use SH_REG_MSW_OFFSET and SH_REG_LSW_OFFSET.
3945 * config/sh/sh.c: Likewise.
3946 * config/sh/sh.h: Likewise.
3948 2013-10-27 Hans-Peter Nilsson <hp@axis.com>
3950 * config/cris/cris.c (cris_emit_trap_for_misalignment): Replace the
3951 removed PRED_MUDFLAP with PRED_NORETURN. Correct file-path in comment.
3953 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
3955 * config/sh/sh.md (movmemsi): Remove empty constraints and predicates.
3957 (cmpstr_t, cmpstrsi): Fix formatting.
3959 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
3962 * config/sh/predicates.md (general_movdst_operand): Allow reg+reg
3963 addressing, do not use general_operand for memory operands.
3965 2013-10-26 Vladimir Makarov <vmakarov@redhat.com>
3968 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
3969 * lra-spills.c (lra_final_code_change): Remove useless move insns.
3971 2013-10-26 Jeff Law <law@redhat.com>
3973 * predict.c (PRED_MUDFLAP): Remove.
3974 * targhooks.c (build_va_arg_indirect_ref): Remove mudflap support.
3976 * Makefile.in (C_COMMON_OBJS): Remove tree-mudflap.
3977 (OBJS): Remove tree-nomudflap.o
3978 (GTFILES): Remove tree-mudflap.c
3979 * builtins.c (expand_builtin_alloc): Remove mudflap support.
3980 * gcc.c (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
3981 (mfwrap_spec, mflib_spec): Likewise.
3982 (cpp_unique_options, cc1_options, static_specs): Likewise.
3983 * gimplify (gimplify_vla_decl, build_va_arg_indirect_ref): Likewise.
3984 * passes.def: Likewise.
3985 * toplev.c (compile_file, process_options): Likewise.
3986 * tree-inline.c (copy_tree_r): Likewise.
3987 * tree-pass.,h (make_pass_mudflap_1, make_pass_mudflap_2): Likewise.
3988 * varasm.c (make_decl_rtl, make_decl_rtl_for_debug): Likewise.
3989 (build_constant_desc, output_constant_def_contents): Likewise.
3990 (categorize_decl_for_section): Likewise.
3991 * tree-mudflap.c: Removed.
3992 * tree-mudflap.h: Removed.
3993 * tree-nomudflap.c: Removed.
3994 * bfin/uclinux.h (MFWRAP_SPEC): Remove.
3995 * moxie/uclinux.h (MFWRAP_SPEC): Likewise.
3996 * rs6000/aix.h (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
3997 * config/sol2.h (MFLIB_SPEC): Likewise.
3998 * doc/install.texi: Remove mudflap references.
3999 * doc/passes.texi: Similarly.
4000 * doc/sourcebuild.texi: Similarly.
4001 * doc/invoke.texi: Remove mudlfap related options.
4003 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
4005 PR rtl-optimization/58759
4006 * lra-constraints.c (lra_constraints): Remove wrong condition to
4007 remove insn setting up an equivalent pseudo.
4009 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
4011 * config/rs6000/rs6000-protos.h
4012 (rs6000_secondary_memory_needed_mode): New prototype.
4013 * config/rs6000/rs6000.c: Include ira.h.
4014 (TARGET_LRA_P): Redefine.
4015 (rs6000_legitimate_offset_address_p): Call
4016 legitimate_constant_pool_address_p in strict mode for LRA.
4017 (rs6000_legitimate_address_p): Ditto.
4018 (legitimate_lo_sum_address_p): Add code for LRA. Use lra_in_progress.
4019 (rs6000_emit_move): Add LRA version of code to generate load/store
4021 (rs6000_secondary_memory_needed_mode): New.
4022 (rs6000_alloc_sdmode_stack_slot): Do nothing for LRA.
4023 (rs6000_secondary_reload_class): Return NO_REGS for LRA for
4024 constants, memory, and FP registers.
4025 (rs6000_lra_p): New.
4026 * config/rs6000/rs6000.h (SECONDARY_MEMORY_NEEDED_MODE): New macro.
4027 * config/rs6000/rs6000.opt (mlra): New option.
4028 * lra-spills.c (lra_final_code_change): Remove useless move insns.
4030 2013-10-25 Yufeng Zhang <yufeng.zhang@arm.com>
4032 * tree-ssa-math-opts.c (convert_plusminus_to_widen): Call
4033 has_single_use () and not do the conversion if has_single_use ()
4034 returns false for the multiplication result.
4036 2013-10-25 David Malcolm <dmalcolm@redhat.com>
4038 * tree.h (EXCEPTIONAL_CLASS_P): Rename parameter from "CODE"
4039 to "NODE", since this works on a "tree", not an
4041 (CONSTANT_CLASS_P): Likewise.
4044 (INDIRECT_REF_P): Likewise.
4045 (REFERENCE_CLASS_P): Likewise.
4046 (COMPARISON_CLASS_P): Likewise.
4047 (UNARY_CLASS_P): Likewise.
4048 (BINARY_CLASS_P): Likewise.
4049 (STATEMENT_CLASS_P): Likewise.
4050 (VL_EXP_CLASS_P): Likewise.
4051 (EXPRESSION_CLASS_P): Likewise.
4052 (IS_TYPE_OR_DECL_P): Likewise.
4054 2013-10-25 Marc Glisse <marc.glisse@inria.fr>
4056 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Look for an
4057 ADDR_EXPR in the defining statement.
4059 2013-10-25 Richard Biener <rguenther@suse.de>
4061 PR tree-optimization/58626
4062 * tree-loop-distribution.c (enum rdg_dep_type): Remove
4063 anti_dd, output_dd and input_dd.
4064 (struct rdg_edge): Remove level and relation members.
4065 (RDGE_LEVEL, RDGE_RELATION): Remove.
4066 (dot_rdg_1): Adjust.
4067 (create_rdg_edge_for_ddr): Remove.
4068 (create_rdg_edges_for_scalar): Adjust.
4069 (create_edge_for_control_dependence): Likewise.
4070 (create_rdg_edges): Split into ...
4071 (create_rdg_flow_edges): ... this
4072 (create_rdg_cd_edges): ... and this.
4074 (build_rdg): Likewise, do not compute data dependences or
4076 (pg_add_dependence_edges): New function.
4078 (distribute_loop): First apply all non-dependence based
4079 partition mergings. Then compute dependences between partitions
4080 and merge and order partitions according to them.
4082 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
4084 PR rtl-optimization/58831
4085 * alias.c (init_alias_analysis): At the beginning of each iteration,
4086 set the reg_seen[N] bit if static_reg_base_value[N] is non-null.
4088 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
4090 * recog.c (search_ofs): New static variable moved from...
4091 (peep2_find_free_register): ...here.
4092 (peephole2_optimize): Initialize it.
4094 2013-10-25 Tobias Burnus <burnus@net-b.de>
4096 * doc/invoke.texi (fopenmp): Change supported OpenMP version to 4.0.
4098 2013-10-25 Uros Bizjak <ubizjak@gmail.com>
4100 * config/i386/i386.h (TARGET_MPX): New define.
4101 (TARGET_MPX_P): Ditto.
4103 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
4105 * config/i386/constraints.md (B): New.
4108 * config/i386/i386-c.c (ix86_target_macros_internal): Add __MPX__.
4109 * config/i386/i386-modes.def (BND32): New.
4111 * config/i386/i386-protos.h (ix86_bnd_prefixed_insn_p): New.
4112 * config/i386/i386.c (isa_opts): Add mmpx.
4113 (regclass_map): Add bound registers.
4114 (dbx_register_map): Likewise.
4115 (dbx64_register_map): Likewise.
4116 (svr4_dbx_register_map): Likewise.
4118 (ix86_option_override_internal): Support MPX ISA.
4119 (ix86_conditional_register_usage): Support bound registers.
4120 (print_reg): Likewise.
4121 (ix86_code_end): Add MPX bnd prefix.
4122 (output_set_got): Likewise.
4123 (ix86_output_call_insn): Likewise.
4124 (ix86_print_operand): Add '!' (MPX bnd) print prefix support.
4125 (ix86_print_operand_punct_valid_p): Likewise.
4126 (ix86_print_operand_address): Support UNSPEC_BNDMK_ADDR and
4128 (ix86_class_likely_spilled_p): Add bound regs support.
4129 (ix86_hard_regno_mode_ok): Likewise.
4130 (x86_order_regs_for_local_alloc): Likewise.
4131 (ix86_bnd_prefixed_insn_p): New.
4132 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Fix to new value.
4133 (FIXED_REGISTERS): Add bound registers.
4134 (CALL_USED_REGISTERS): Likewise.
4135 (REG_ALLOC_ORDER): Likewise.
4136 (HARD_REGNO_NREGS): Likewise.
4138 (VALID_BND_REG_MODE): New.
4139 (FIRST_BND_REG): New.
4140 (LAST_BND_REG): New.
4141 (reg_class): Add BND_REGS.
4142 (REG_CLASS_NAMES): Likewise.
4143 (REG_CLASS_CONTENTS): Likewise.
4145 (ANY_BND_REG_P): New.
4147 (HI_REGISTER_NAMES): Add bound registers.
4148 * config/i386/i386.md (UNSPEC_BNDMK): New.
4149 (UNSPEC_BNDMK_ADDR): New.
4150 (UNSPEC_BNDSTX): New.
4151 (UNSPEC_BNDLDX): New.
4152 (UNSPEC_BNDLDX_ADDR): New.
4153 (UNSPEC_BNDCL): New.
4154 (UNSPEC_BNDCU): New.
4155 (UNSPEC_BNDCN): New.
4156 (UNSPEC_MPX_FENCE): New.
4159 (type): Add mpxmov, mpxmk, mpxchk, mpxld, mpxst.
4160 (length_immediate): Likewise.
4161 (prefix_0f): Likewise.
4163 (prefix_rep): Check for bnd prefix.
4164 (length_nobnd): New.
4165 (length): Use length_nobnd if specified.
4170 (*jcc_1): Add bnd prefix and rename length attr to length_nobnd.
4173 (simple_return_internal): Likewise.
4174 (simple_return_pop_internal): Likewise.
4175 (*indirect_jump): Add MPX bnd prefix.
4176 (*tablejump_1): Likewise.
4177 (simple_return_internal_long): Likewise.
4178 (simple_return_indirect_internal): Likewise.
4182 (*mov<mode>_internal_mpx): New.
4183 (<mode>_<bndcheck>): New.
4184 (*<mode>_<bndcheck>): New.
4189 * config/i386/predicates.md (lea_address_operand): Rename to...
4190 (address_no_seg_operand): ... this.
4191 (address_mpx_no_base_operand): New.
4192 (address_mpx_no_index_operand): New.
4193 (bnd_mem_operator): New.
4194 * config/i386/i386.opt (mmpx): New.
4195 * doc/invoke.texi: Add documentation for the flags -mmpx, -mno-mpx.
4196 * doc/rtl.texi Add documentation for BND32mode and BND64mode.
4198 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
4200 * mode-classes.def (MODE_POINTER_BOUNDS): New.
4201 * tree.def (POINTER_BOUNDS_TYPE): New.
4202 * genmodes.c (complete_mode): Support MODE_POINTER_BOUNDS.
4203 (POINTER_BOUNDS_MODE): New.
4204 (make_pointer_bounds_mode): New.
4205 * machmode.h (POINTER_BOUNDS_MODE_P): New.
4206 * stor-layout.c (int_mode_for_mode): Support MODE_POINTER_BOUNDS.
4207 (layout_type): Support POINTER_BOUNDS_TYPE.
4208 * tree-pretty-print.c (dump_generic_node): Support POINTER_BOUNDS_TYPE.
4209 * tree.c (build_int_cst_wide): Support POINTER_BOUNDS_TYPE.
4210 (type_contains_placeholder_1): Likewise.
4211 * tree.h (POINTER_BOUNDS_TYPE_P): New.
4212 * varasm.c (output_constant): Support POINTER_BOUNDS_TYPE.
4213 * doc/rtl.texi (MODE_POINTER_BOUNDS): New.
4215 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
4217 * expr.c (expand_expr_real_1): Use mode of memory reference rather than
4218 mode of address computation when calling memory_address_addr_space.
4220 2013-08-24 Richard Henderson <rth@twiddle.net>
4223 * optabs.c (maybe_emit_atomic_exchange): Use create_input_operand
4224 instead of create_convert_operand_to.
4225 (maybe_emit_sync_lock_test_and_set): Likewise.
4226 (expand_atomic_compare_and_swap): Likewise.
4227 (maybe_emit_compare_and_swap_exchange_loop): Don't convert_modes.
4229 2013-08-24 Sriraman Tallam <tmsriram@google.com>
4231 * cgraph.c (cgraph_fnver_htab): Move GTY((...)) to be before htab_t.
4232 Change param_is to use the struct and not the pointer to the struct.
4234 2013-10-24 Andrew MacLeod <amacleod@redhat.com>
4236 * builtins.c (dummy_object, gimplify_va_arg_expr): Move to gimplify.c.
4237 * gimplify.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
4238 Move to targhooks.c.
4239 (dummy_object, gimplify_va_arg_expr): Relocate from builtins.c.
4240 * targhooks.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
4241 Relocate from gimplify.c.
4242 * targhooks.h: Add 2 prototypes.
4243 * tree.h: Delete 2 prototypes.
4245 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
4247 * tree-ssa-loop-ivopts.c (multiplier_allowed_in_address_p ): Check both
4248 [reg+mult*reg] and [mult*reg] to determine if multiplier is allowed.
4250 2013-10-24 Cong Hou <congh@google.com>
4252 * convert.c (convert_to_real): Guard those unsafe math function
4253 convertions with flag_unsafe_math_optimizations. Handle sqrt()
4256 2013-10-24 Markus Trippelsdorf <markus@trippelsdorf.de>
4259 * cgraph.c (cgraph_create_edge_1): Add indirect_unknown_callee
4261 (cgraph_create_edge): Use the new argument.
4262 (cgraph_create_indirect_edge): Likewise.
4264 2013-10-24 Joern Rennecke <joern.rennecke@embecosm.com>
4266 * config/arc/arc.c (arc_ccfsm_post_advance): Also handle
4268 (arc_ifcvt) <case 1 and 2>: Check that arc_ccfsm_post_advance
4269 changes statep->state.
4271 2013-10-24 Tobias Burnus <burnus@net-b.de>
4274 * tree-cfg.c (replace_loop_annotate): New function.
4275 (execute_build_cfg): Call it.
4276 * gimplify.c (gimple_boolify, gimplify_expr): Handle ANNOTATE_EXPR.
4277 * internal-fn.c (expand_ANNOTATE): New function.
4278 * internal-fn.def (ANNOTATE): Define as new internal function.
4279 * tree-core.h (tree_node_kind): Add annot_expr_ivdep_kind.
4280 * tree-pretty-print.c (dump_generic_node): Handle ANNOTATE_EXPR.
4281 * tree.def (ANNOTATE_EXPR): New DEFTREECODE.
4282 * doc/extend.texi (Pragmas): Document #pragma ivdep.
4283 * doc/generic.texi (Expressions): Document ANNOTATE_EXPR.
4285 2013-10-17 Ian Bolton <ian.bolton@arm.com>
4286 Marcus Shawcroft <marcus.shawcroft@arm.com>
4288 * config/aarch64/aarch64.c (aarch64_preferred_reload_class):
4289 Special case reload SP+C into none GENERAL_REGS.
4291 2013-10-24 Michael Matz <matz@suse.de>
4293 * gengtype.c (is_file_equal): Check that files will be same length.
4295 2013-10-25 Christian Bruel <christian.bruel@st.com>
4297 * config.gcc (sh-*): Add sh-mem.o to extra_obj.
4298 * config/sh/t-sh (sh-mem.o): New rule.
4299 * config/sh/sh-mem.cc (expand_block_move): Moved here.
4300 (sh_expand_cmpstr): New function.
4301 * config/sh/sh.c (force_into, expand_block_move): Move to sh-mem.c.
4302 * config/sh/sh-protos.h (sh_expand_cmpstr): Declare.
4303 * config/sh/sh.md (cmpstrsi, cmpstr_t): New patterns.
4304 (rotlhi3_8): Rename.
4306 2013-10-24 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4308 * configure.ac (ZW_PROG_COMPILER_DEPENDENCIES): Use CXX instead of CC.
4309 * Makefile.in (CXXDEPMODE): Assign and change users.
4310 (CCDEPMODE): Delete.
4311 * configure: Regenerate.
4313 2013-10-23 David Malcolm <dmalcolm@redhat.com>
4315 * gengtype-parse.c (require_without_advance): New.
4316 (type): For GTY-marked types that are not GTY((user)), parse any
4317 base classes, requiring them to be single-inheritance, and not
4318 be templates. For non-GTY-marked types and GTY((user)),
4319 continue to skip over any C++ inheritance specification.
4320 * gengtype-state.c (state_writer::write_state_struct_type):
4321 Write base class of type (if any).
4322 (read_state_struct_type): Read base class of type (if any).
4323 * gengtype.c (new_structure): Add a "base_class" parameter.
4324 (create_optional_field_): Update for new parameter to new_structure.
4325 (adjust_field_rtx_def): Likewise.
4326 (adjust_field_tree_exp): Likewise.
4327 * gengtype.h (struct type): Add "base_class" field to the s
4329 (new_structure): Add "base" parameter.
4331 2013-10-23 Sriraman Tallam <tmsriram@google.com>
4334 * config/i386/i386.c (ix86_option_override_internal):
4335 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...)
4336 (ix86_valid_target_attribute_tree):
4337 Change TARGET_64BIT to TARGET_64BIT_P (opts->...)
4338 Change TARGET_SSE to TARGET_SSE_P (opts->...)
4340 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
4342 * tree-ssa-loop.h: Remove include files.
4343 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
4344 * cfgloopmanip.c: Move required includes from tree-ssa-loop.h.
4345 * graphite-clast-to-gimple.c: Likewise.
4346 * graphite-scop-detection.c: Likewise.
4347 * graphite-sese-to-poly.c: Likewise.
4348 * ipa-inline-analysis.c: Likewise.
4349 * ipa-pure-const.c: Likewise.
4350 * loop-init.c: Likewise.
4351 * passes.c: Likewise.
4352 * predict.c: Likewise.
4353 * tree-cfg.c: Likewise.
4354 * tree-cfgcleanup.c: Likewise.
4355 * tree-chrec.c: Likewise.
4356 * tree-data-ref.c: Likewise.
4357 * tree-loop-distribution.c: Likewise.
4358 * tree-parloops.c: Likewise.
4359 * tree-predcom.c: Likewise.
4360 * tree-scalar-evolution.c: Likewise.
4361 * tree-ssa-address.c: Likewise.
4362 * tree-ssa.c: Likewise.
4363 * tree-ssa-dce.c: Likewise.
4364 * tree-ssa-loop.c: Likewise.
4365 * tree-ssa-loop-im.c: Likewise.
4366 * tree-ssa-loop-ivcanon.c: Likewise.
4367 * tree-ssa-loop-ivopts.c: Likewise.
4368 * tree-ssa-loop-manip.c: Likewise.
4369 * tree-ssa-loop-niter.c: Likewise.
4370 * tree-ssa-loop-prefetch.c: Likewise.
4371 * tree-ssa-loop-unswitch.c: Likewise.
4372 * tree-ssa-reassoc.c: Likewise.
4373 * tree-vect-data-refs.c: Likewise.
4374 * tree-vect-loop.c: Likewise.
4375 * tree-vect-loop-manip.c: Likewise.
4376 * tree-vectorizer.c: Likewise.
4377 * tree-vect-stmts.c: Likewise.
4378 * tree-vrp.c: Likewise.
4380 2013-10-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4382 * config/rs6000/altivec.md (mulv8hi3): Adjust for little endian.
4384 2013-10-23 Jakub Jelinek <jakub@redhat.com>
4386 PR tree-optimization/58775
4387 PR tree-optimization/58791
4388 * tree-ssa-reassoc.c (reassoc_stmt_dominates_stmt_p): New function.
4389 (insert_stmt_after): Rewritten, don't move the stmt, but really
4391 (get_stmt_uid_with_default): Remove.
4392 (build_and_add_sum): Use insert_stmt_after and
4393 reassoc_stmt_dominates_stmt_p. Fix up uid if bb contains only labels.
4394 (update_range_test): Set uid on stmts added by
4395 force_gimple_operand_gsi. Don't immediately modify statements
4396 in inter-bb optimization, just update oe->op values.
4397 (optimize_range_tests): Return bool whether any changed have been made.
4398 (update_ops): New function.
4399 (struct inter_bb_range_test_entry): New type.
4400 (maybe_optimize_range_tests): Perform statement changes here.
4401 (not_dominated_by, appears_later_in_bb, get_def_stmt,
4402 ensure_ops_are_available): Remove.
4403 (find_insert_point): Rewritten.
4404 (rewrite_expr_tree): Remove MOVED argument, add CHANGED argument,
4405 return LHS of the (new resp. old) stmt. Don't call
4406 ensure_ops_are_available, don't reuse SSA_NAMEs, recurse first
4407 instead of last, move new stmt at the right place.
4408 (linearize_expr, repropagate_negates): Don't reuse SSA_NAMEs.
4409 (negate_value): Likewise. Set uids.
4410 (break_up_subtract_bb): Initialize uids.
4411 (reassociate_bb): Adjust rewrite_expr_tree caller.
4412 (do_reassoc): Don't call renumber_gimple_stmt_uids.
4414 2013-10-23 David Edelsohn <dje.gcc@gmail.com>
4417 * config/rs6000/rs6000.md (mulsi3_internal1 and splitter): Add
4418 TARGET_32BIT final condition.
4419 (mulsi3_internal2 and splitter): Same.
4421 2013-10-23 Jeff Law <law@redhat.com>
4423 * tree-ssa-threadedge.c (thread_across_edge): Do not allow threading
4424 through joiner blocks with abnormal outgoing edges.
4426 * tree-ssa-threadupdate.c (thread_block_1): Renamed from thread_block.
4427 Add parameter JOINERS, to allow/disallow threading through joiner
4429 (thread_block): New. Call thread_block_1.
4430 (mark_threaded_blocks): Remove code to filter out certain cases
4431 of threading through joiner blocks.
4432 (thread_through_all_blocks): Document how we can have a dangling
4433 edge AUX field and clear it.
4435 2013-10-23 Ian Lance Taylor <iant@google.com>
4437 * doc/invoke.texi (Option Summary): Remove -fno-default-inline.
4438 (C++ Dialect Options): Likewise.
4439 (Optimize Options): Likewise.
4441 2013-10-23 Tom de Vries <tom@codesourcery.com>
4443 PR tree-optimization/58805
4444 * tree-ssa-tail-merge.c (stmt_local_def): Add gimple_vdef check.
4446 2013-10-23 Jakub Jelinek <jakub@redhat.com>
4448 * tree-vect-patterns.c (vect_recog_divmod_pattern): Optimize
4449 sequence based on get_range_info returned range.
4451 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
4453 * tree-ssa.h: Remove all #include's
4454 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
4455 * alias.c: Move required includes from tree-ssa.h.
4457 * builtins.c: Likewise.
4458 * calls.c: Likewise.
4459 * cfgexpand.c: Likewise.
4460 * cfghooks.c: Likewise.
4461 * cfgloop.c: Likewise.
4462 * cfgloopmanip.c: Likewise.
4463 * cgraph.c: Likewise.
4464 * cgraphbuild.c: Likewise.
4465 * cgraphclones.c: Likewise.
4466 * cgraphunit.c: Likewise.
4468 * except.c: Likewise.
4470 * final.c: Likewise.
4471 * fold-const.c: Likewise.
4472 * ggc-page.c: Likewise.
4473 * gimple-builder.c: Likewise.
4474 * gimple-fold.c: Likewise.
4475 * gimple-iterator.c: Likewise.
4476 * gimple-low.c: Likewise.
4477 * gimple-pretty-print.c: Likewise.
4478 * gimple-ssa-strength-reduction.c: Likewise.
4479 * gimple-streamer-in.c: Likewise.
4480 * gimple-streamer-out.c: Likewise.
4481 * gimplify.c: Likewise.
4482 * graphite-blocking.c: Likewise.
4483 * graphite-clast-to-gimple.c: Likewise.
4484 * graphite-dependences.c: Likewise.
4485 * graphite-interchange.c: Likewise.
4486 * graphite-optimize-isl.c: Likewise.
4487 * graphite-poly.c: Likewise.
4488 * graphite-scop-detection.c: Likewise.
4489 * graphite-sese-to-poly.c: Likewise.
4490 * graphite.c: Likewise.
4491 * ipa-cp.c: Likewise.
4492 * ipa-inline-analysis.c: Likewise.
4493 * ipa-inline-transform.c: Likewise.
4494 * ipa-inline.c: Likewise.
4495 * ipa-prop.c: Likewise.
4496 * ipa-pure-const.c: Likewise.
4497 * ipa-reference.c: Likewise.
4498 * ipa-split.c: Likewise.
4499 * ipa-utils.c: Likewise.
4500 * loop-init.c: Likewise.
4501 * lto-cgraph.c: Likewise.
4502 * lto-section-in.c: Likewise.
4503 * lto-section-out.c: Likewise.
4504 * lto-streamer-in.c: Likewise.
4505 * lto-streamer-out.c: Likewise.
4506 * lto-streamer.c: Likewise.
4507 * omp-low.c: Likewise.
4508 * passes.c: Likewise.
4509 * predict.c: Likewise.
4510 * print-tree.c: Likewise.
4511 * profile.c: Likewise.
4513 * targhooks.c: Likewise.
4514 * tracer.c: Likewise.
4515 * trans-mem.c: Likewise.
4516 * tree-call-cdce.c: Likewise.
4517 * tree-cfg.c: Likewise.
4518 * tree-cfgcleanup.c: Likewise.
4519 * tree-chrec.c: Likewise.
4520 * tree-complex.c: Likewise.
4521 * tree-data-ref.c: Likewise.
4522 * tree-dfa.c: Likewise.
4523 * tree-eh.c: Likewise.
4524 * tree-emutls.c: Likewise.
4525 * tree-if-conv.c: Likewise.
4526 * tree-inline.c: Likewise.
4527 * tree-into-ssa.c: Likewise.
4528 * tree-loop-distribution.c: Likewise.
4529 * tree-mudflap.c: Likewise.
4530 * tree-nested.c: Likewise.
4531 * tree-nrv.c: Likewise.
4532 * tree-object-size.c: Likewise.
4533 * tree-outof-ssa.c: Likewise.
4534 * tree-parloops.c: Likewise.
4535 * tree-phinodes.c: Likewise.
4536 * tree-predcom.c: Likewise.
4537 * tree-pretty-print.c: Likewise.
4538 * tree-profile.c: Likewise.
4539 * tree-scalar-evolution.c: Likewise.
4540 * tree-sra.c: Likewise.
4541 * tree-ssa-address.c: Likewise.
4542 * tree-ssa-alias.c: Likewise.
4543 * tree-ssa-ccp.c: Likewise.
4544 * tree-ssa-coalesce.c: Likewise.
4545 * tree-ssa-copy.c: Likewise.
4546 * tree-ssa-copyrename.c: Likewise.
4547 * tree-ssa-dce.c: Likewise.
4548 * tree-ssa-dom.c: Likewise.
4549 * tree-ssa-dse.c: Likewise.
4550 * tree-ssa-forwprop.c: Likewise.
4551 * tree-ssa-ifcombine.c: Likewise.
4552 * tree-ssa-live.c: Likewise.
4553 * tree-ssa-loop-ch.c: Likewise.
4554 * tree-ssa-loop-im.c: Likewise.
4555 * tree-ssa-loop-ivcanon.c: Likewise.
4556 * tree-ssa-loop-ivopts.c: Likewise.
4557 * tree-ssa-loop-manip.c: Likewise.
4558 * tree-ssa-loop-niter.c: Likewise.
4559 * tree-ssa-loop-prefetch.c: Likewise.
4560 * tree-ssa-loop-unswitch.c: Likewise.
4561 * tree-ssa-loop.c: Likewise.
4562 * tree-ssa-math-opts.c: Likewise.
4563 * tree-ssa-operands.c: Likewise.
4564 * tree-ssa-phiopt.c: Likewise.
4565 * tree-ssa-phiprop.c: Likewise.
4566 * tree-ssa-pre.c: Likewise.
4567 * tree-ssa-propagate.c: Likewise.
4568 * tree-ssa-reassoc.c: Likewise.
4569 * tree-ssa-sccvn.c: Likewise.
4570 * tree-ssa-sink.c: Likewise.
4571 * tree-ssa-strlen.c: Likewise.
4572 * tree-ssa-structalias.c: Likewise.
4573 * tree-ssa-tail-merge.c: Likewise.
4574 * tree-ssa-ter.c: Likewise.
4575 * tree-ssa-threadedge.c: Likewise.
4576 * tree-ssa-threadupdate.c: Likewise.
4577 * tree-ssa-uncprop.c: Likewise.
4578 * tree-ssa-uninit.c: Likewise.
4579 * tree-ssa.c: Likewise.
4580 * tree-ssanames.c: Likewise.
4581 * tree-stdarg.c: Likewise.
4582 * tree-streamer-in.c: Likewise.
4583 * tree-switch-conversion.c: Likewise.
4584 * tree-tailcall.c: Likewise.
4585 * tree-vect-data-refs.c: Likewise.
4586 * tree-vect-generic.c: Likewise.
4587 * tree-vect-loop-manip.c: Likewise.
4588 * tree-vect-loop.c: Likewise.
4589 * tree-vect-patterns.c: Likewise.
4590 * tree-vect-slp.c: Likewise.
4591 * tree-vect-stmts.c: Likewise.
4592 * tree-vectorizer.c: Likewise.
4593 * tree-vrp.c: Likewise.
4596 * value-prof.c: Likewise.
4597 * var-tracking.c: Likewise.
4598 * varpool.c: Likewise.
4599 * vtable-verify.c: Likewise.
4601 2013-10-23 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4603 * config/tilegx/tilegx.c: Include "tree.h".
4605 2013-10-23 Jakub Jelinek <jakub@redhat.com>
4607 * gimple-pretty-print.c (dump_ssaname_info): Always print "# " before
4608 the info, not after it.
4609 (gump_gimple_phi): Add COMMENT argument, if true, print "# " after
4610 dump_ssaname_info call.
4611 (pp_gimple_stmt_1): Adjust caller.
4612 (dump_phi_nodes): Likewise. Don't print "# " here.
4614 2013-10-22 Jan Hubicka <jh@suse.cz>
4616 * i386.h (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES_EPILOGUES): New
4618 * x86-tune.def (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES): Define it.
4619 * i386.c (expand_small_movmem_or_setmem): New function.
4620 (expand_set_or_movmem_prologue_epilogue_by_misaligned_moves): New
4622 (alg_usable_p): Add support for value ranges; cleanup.
4623 (ix86_expand_set_or_movmem): Add support for misaligned moves.
4625 2013-10-22 Sterling Augustine <saugustine@google.com>
4627 * doc/invoke.texi: Document -ggnu-pubnames.
4628 * common.opt: Add new option -ggnu-pubnames and modify -gpubnames
4630 * dwarf2out.c: Include gdb/gdb-index.h.
4631 (DEBUG_PUBNAMES_SECTION, DEBUG_PUBTYPES_SECTION): Handle
4632 debug_generate_pub_sections.
4633 (is_java, output_pubtables, output_pubname): New functions.
4634 (include_pubname_in_output): Handle debug_generate_pub_sections at
4636 (size_of_pubnames): Use new local space_for_flags based on
4637 debug_generate_pub_sections.
4638 (output_pubnames): Unify pubnames and pubtypes output logic.
4639 Genericize comments. Call output_pubname.
4640 (dwarf2out_finish): Move logic to output_pubnames and call it.
4642 2013-10-22 Uros Bizjak <ubizjak@gmail.com>
4645 * config/i386/i386.c (put_condition_code) <case GTU, case LEU>:
4646 Remove CCCmode handling.
4647 <case LTU>: Return 'c' suffix for CCCmode.
4648 <case GEU>: Return 'nc' suffix for CCCmode.
4649 (ix86_cc_mode) <case GTU, case LEU>: Do not generate overflow checks.
4650 * config/i386/i386.md (*sub<mode>3_cconly_overflow): Remove.
4651 (*sub<mode>3_cc_overflow): Ditto.
4652 (*subsi3_zext_cc_overflow): Ditto.
4654 2013-10-22 Steve Ellcey <sellcey@mips.com>
4656 * config/mips/mips.c (mips_rtx_costs): Fix cost estimate for nor
4657 (AND (NOT OP1) (NOT OP2)).
4659 2013-10-22 Bill Schmidt <wschmidt@vnet.ibm.com>
4661 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
4662 meaning of merge-high and merge-low masks for little endian; avoid
4663 use of vector-pack masks for little endian for mismatched modes.
4665 2013-10-22 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4667 * config/tilepro/tilepro.c: Include "tree.h".
4669 2013-10-22 Andreas Schwab <schwab@suse.de>
4671 * config/m68k/m68k.c (notice_update_cc): Handle register conflict
4674 2013-10-22 Paolo Carlini <paolo.carlini@oracle.com>
4676 * doc/contrib.texi ([Fran@,{c}ois Dumont], [Tim Shen],
4677 [Ed Smith-Rowland]): New entries.
4678 ([Stephen M. Webb]): Update.
4680 2013-10-22 Andrew MacLeod <amacleod@redhat.com>
4682 * tree-ssa-ter.h: Remove duplicate copy of file contents.
4684 2013-10-21 Marek Polacek <polacek@redhat.com>
4687 * fold-const.c (fold_range_test): Return 0 if the type is not
4690 2013-10-21 Richard Sandiford <rdsandiford@googlemail.com>
4692 * system.h: Move hwint.h include further down.
4693 * hwint.h (sext_hwi, zext_hwi): Define unconditionally. Add
4694 gcc_checking_asserts.
4695 * hwint.c (sext_hwi, zext_hwi): Delete ENABLE_CHECKING versions.
4697 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
4699 Fix volatile issues in optimize_bit_field_compare.
4700 * fold-const.c (optimize_bit_field_compare): Bail out if
4701 lvolatilep or rvolatilep.
4703 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
4705 Fix DECL_BIT_FIELD depencency on flag_strict_volatile_bitfields
4706 and get_inner_reference returning different pmode for non-volatile
4707 bit-field members dependent on flag_strict_volatile_bitfields.
4708 * stor-layout.c (layout_decl): Remove special handling of
4709 flag_strict_volatile_bitfields.
4710 * expr.c (get_inner_reference): Don't use DECL_BIT_FIELD
4711 if flag_strict_volatile_bitfields > 0 and TREE_THIS_VOLATILE.
4713 2013-10-21 Paulo Matos <pmatos@broadcom.com>
4715 * ipa-inline.c (edge_badness): Cap edge->count at max_count for badness
4718 2013-10-21 Jeff Law <law@redhat.com>
4720 * tree-ssa-threadedge.c (thread_through_normal_block): New
4721 argument VISITED. Remove VISISTED as a local variable. When we
4722 have a threadable jump, verify the destination of the jump has not
4724 (thread_across_edge): Allocate VISITED bitmap once at function
4725 scope and use it throughout. Make sure to set appropriate bits in
4726 VISITED for E (start of jump thread path).
4727 * tree-ssa-threadupdate.c (mark_threaded_blocks): Reject threading
4728 through a joiner if any edge on the path has a recorded jump thread.
4730 2013-10-21 Ian Lance Taylor <iant@google.com>
4732 * doc/invoke.texi (Optimize Options): For -fno-toplevel-reorder,
4733 don't imply that attributes can solve all problems.
4734 (Directory Options): Fix typo.
4736 2013-10-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4738 * config/arm/arm.c (cortexa9_extra_costs): Update mult costs for
4739 extend and extend_add.
4741 2013-10-21 Richard Biener <rguenther@suse.de>
4743 PR tree-optimization/58794
4744 * fold-const.c (operand_equal_p): Compare FIELD_DECL operand
4745 of COMPONENT_REFs with OEP_CONSTANT_ADDRESS_OF left in place.
4747 2013-10-21 Richard Biener <rguenther@suse.de>
4750 * fold-const.c (fold_binary_loc): Fold ((T) (X /[ex] C)) * C
4751 to (T) X for sign-changing conversions (or no conversion).
4753 2013-10-20 Uros Bizjak <ubizjak@gmail.com>
4755 * config/i386/i386.md (kxnor<mode>): Add FLAGS_REG clobber.
4757 2013-10-20 Jan Hubicka <jh@suse.cz>
4759 * config/i386/i386-tune.def: Add comment; organize into categories
4761 2013-10-21 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
4763 * config/i386/i386.c (expand_set_or_movmem_via_loop): Add issetmem
4764 argument. Update function comment.
4765 (expand_set_or_movmem_via_rep): New function combining
4766 expand_movmem_via_rep_mov and expand_setmem_via_rep_stos.
4767 (expand_movmem_via_rep_mov): Remove.
4768 expand_setmem_via_rep_stos): Remove.
4769 (expand_movmem_epilogue): Update calls correspondingly.
4770 (expand_setmem_epilogue_via_loop): Likewise.
4772 (expand_setmem_epilogue): Add VEC_VALUE argument, refactor.
4773 (expand_set_or_movmem_prologue): New function combining
4774 expand_movmem_prologue and expand_setmem_prologue.
4775 (expand_movmem_prologue): Remove.
4776 (expand_setmem_prologue): Remove.
4777 (expand_set_or_movmem_constant_prologue): New function combining
4778 expand_constant_movmem_prologue and expand_constant_setmem_prologue.
4779 (expand_constant_movmem_prologue): Remove.
4780 (expand_constant_setmem_prologue): Remove.
4781 (promote_duplicated_reg): Allow vector-const0 value.
4782 (ix86_expand_set_or_movmem): New function combining ix86_expand_movmem
4783 and ix86_expand_setmem.
4784 (ix86_expand_movmem): Call ix86_expand_set_or_movmem.
4785 (ix86_expand_setmem): Call ix86_expand_set_or_movmem.
4787 2013-10-21 Diego Novillo <dnovillo@google.com>
4789 * asan.c: Include tree.h
4790 * bb-reorder.c: Likewise.
4791 * cfgcleanup.c: Likewise.
4792 * cfgloopmanip.c: Likewise.
4793 * data-streamer-in.c: Likewise.
4794 * data-streamer-out.c: Likewise.
4795 * data-streamer.c: Likewise.
4796 * dwarf2cfi.c: Likewise.
4797 * graphite-blocking.c: Likewise.
4798 * graphite-clast-to-gimple.c: Likewise.
4799 * graphite-dependences.c: Likewise.
4800 * graphite-interchange.c: Likewise.
4801 * graphite-optimize-isl.c: Likewise.
4802 * graphite-poly.c: Likewise.
4803 * graphite-scop-detection.c: Likewise.
4804 * graphite-sese-to-poly.c: Likewise.
4805 * graphite.c: Likewise.
4806 * ipa-devirt.c: Likewise.
4807 * ipa-profile.c: Likewise.
4810 * loop-init.c: Likewise.
4811 * loop-unroll.c: Likewise.
4812 * lower-subreg.c: Likewise.
4813 * lto/lto-object.c: Likewise.
4814 * recog.c: Likewise.
4815 * reginfo.c: Likewise.
4816 * tree-loop-distribution.c: Likewise.
4817 * tree-parloops.c: Likewise.
4818 * tree-ssa-strlen.c: Likewise.
4819 * tree-streamer.c: Likewise.
4820 * value-prof.c: Likewise.
4821 * target-globals.c: Likewise.
4822 * expr.h: Include tree-core.h instead of tree.h.
4823 * gimple.h: Likewise.
4824 * ipa-prop.h: Likewise.
4825 * ipa-utils.h: Likewise.
4826 * lto-streamer.h: Likewise.
4827 * streamer-hooks.h: Likewise.
4828 * ipa-reference.h: Include cgraph.h instead of tree.h.
4829 * cgraph.h: Include basic-block.h instead of tree.h.
4830 * tree-streamer.h: Do not include tree.h.
4831 * genattrtab.c (write_header): Generate inclusion of tree.h.
4832 * genautomata.c (main): Likewise.
4833 * genemit.c: Likewise.
4834 * genopinit.c: Likewise.
4835 * genoutput.c (output_prologue): Likewise.
4836 * genpeep.c: Likewise.
4838 2013-10-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4840 * config/rs6000/altivec.md (vec_unpacku_hi_v16qi): Adjust for
4842 (vec_unpacku_hi_v8hi): Likewise.
4843 (vec_unpacku_lo_v16qi): Likewise.
4844 (vec_unpacku_lo_v8hi): Likewise.
4846 2013-10-20 Jan Hubicka <jh@suse.cz>
4848 * config/i386/x86-tune.def (X86_TUNE_SLOW_IMUL_IMM32_MEM,
4849 X86_TUNE_SLOW_IMUL_IMM8): Keep enabled only for K8 and AMDFAM10.
4850 (X86_TUNE_USE_VECTOR_FP_CONVERTS): Disable for generic.
4852 2013-10-20 Richard Sandiford <rdsandiford@googlemail.com>
4854 * config/mips/mips.h (ISA_HAS_WSBH): Define.
4855 * config/mips/mips.md (UNSPEC_WSBH, UNSPEC_DSBH, UNSPEC_DSHD): New
4857 (bswaphi2, bswapsi2, bswapdi2, wsbh, dsbh, dshd): New patterns.
4859 2013-10-19 John David Anglin <danglin@gcc.gnu.org>
4862 * system.h: Undef m_slot.
4864 2013-10-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4866 * config/rs6000/rs6000.c (vspltis_constant): Make sure we check
4867 all elements for both endian flavors.
4869 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
4872 * config/i386/i386.c (ix86_function_value_regno): Add DX_REG,
4873 ST1_REG and XMM1_REG for 32bit and 64bit targets. Also add DI_REG
4874 and SI_REG for 64bit SYSV ABI targets.
4876 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
4878 * mode-switching.c (create_pre_exit): Rename maybe_builtin_apply
4879 to multi_reg_return. Clarify that we are skipping USEs of multiple
4880 return registers. Use bool type where appropriate.
4882 2013-10-18 Jan Hubicka <jh@suse.cz>
4884 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Disable accumulation
4886 * x86-tune.def (X86_TUNE_USE_LEAVE): Update comment.
4887 (X86_TUNE_PUSH_MEMORY): Likewise.
4888 (X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL,
4889 X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL): New.
4890 (X86_TUNE_ACCUMULATE_OUTGOING_ARGS, X86_TUNE_ALWAYS_FANCY_MATH_387):
4892 * i386.c (x86_accumulate_outgoing_args, x86_arch_always_fancy_math_387,
4893 x86_avx256_split_unaligned_load, x86_avx256_split_unaligned_store):
4895 (ix86_option_override_internal): Update to use tune features instead
4898 2013-10-18 Cong Hou <congh@google.com>
4900 PR tree-optimization/58508
4901 * tree-vect-loop-manip.c (vect_loop_versioning): Hoist loop invariant
4902 statement that contains data refs with zero-step.
4904 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
4906 * tree-ssa.h: Don't include gimple-low.h, tree-ssa-address.h,
4907 sbitmap.h, tree-ssa-threadedge.h, tree-ssa-dom.h and tree-cfgcleanup.h.
4908 * gimple-low.c (gimple_check_call_arg,
4909 gimple_check_call_matching_types): Move to cgraph.c.
4910 * gimple-low.h: Remove prototype.
4911 * cgraph.c: (gimple_check_call_arg, gimple_check_call_matching_types):
4912 Relocate from gimple-low.c.
4913 * cgraph.h: Add prototype. Don't include basic-block.h.
4914 * gimplify.c: Add gimple-low to include list.
4915 * omp-low.c: Add gimple-low and tree-cfgcleanup.h to include list.
4916 * tree-eh.c: Add gimple-low to include list.
4917 * tree-nested.c: Likewise.
4918 * cfgexpand.c: Add tree-ssa-address.h to include list.
4920 * gimple-fold.c: Likewise.
4921 * gimple-ssa-strength-reduction.c: Likewise.
4922 * trans-mem.c: Likewise.
4923 * tree-mudflap.c: Likewise.
4924 * tree-ssa-loop-ivopts.c: Likewise.
4925 * tree-ssa-dom.c: Include tree-ssa-threadedge.h and tree-ssa-dom.h.
4926 (degenerate_phi_result): Move to tree-phinodes.c.
4927 * tree-ssa-loop-ch.c: Include tree-ssa-threadedge.h.
4928 * tree-ssa-threadedge.c: Likewise.
4929 * tree-vrp.c: Likewise.
4930 * tree-phinodes.c (degenerate_phi_result): Relocate here.
4931 * tree-ssa-dom.h (degenerate_phi_result): Remove Prototype.
4932 * tree-phinodes.h (degenerate_phi_result): Add prototype.
4933 * tree-ssa-copy.c: Include tree-ssa-dom.h.
4934 * tree-ssa-forwprop.c: Likewise.
4935 * tree-cfgcleanup.c (execute_cleanup_cfg_post_optimizing,
4936 pass_data_cleanup_cfg_post_optimizing,
4937 make_pass_cleanup_cfg_post_optimizing): Relocate from tree-optimize.c.
4938 * tree-optimize.c: Delete File.
4939 * graphite.c: Include tree-cfgcleanup.h.
4940 * passes.c: Likewise.
4941 * tree-cfg.c: Likewise.
4942 * tree-profile.c: Likewise.
4943 * tree-ssa-dse.c: Likewise.
4944 * tree-ssa-loop-ivcanon.c: Likewise.
4945 * tree-switch-conversion.c: Don't include tree-ssa-operands.h.
4946 * tree-outof-ssa.c: Include sbitmap.h.
4947 * tree-ssa-live.c: Likewise.
4948 * tree-ssa-propagate.c: Likewise.
4949 * tree-ssa-structalias.c: Likewise.
4950 * tree-stdarg.c: Likewise.
4951 * Makefile.in (OBJS): Delete tree-optimize.o.
4952 * basic-block.h (gcov_type, gcov_type_unsigned): Move to coretypes.h.
4953 * coretypes.h (gcov_type, gcov_type_unsigned): Relocate here.
4954 * varasm.c: Include basic-block.h.
4955 * cfgloop.h: Include function.h instead of basic-block.h
4956 (bb_loop_depth): Move to cfgloop.c.
4957 * cfgloop.c (bb_loop_depth): Relocate from cfgloop.h.
4959 2013-10-18 Teresa Johnson <tejohnson@google.com>
4961 * predict.c (probably_never_executed): Compare frequency-based
4962 count to number of training runs.
4963 * params.def (UNLIKELY_BB_COUNT_FRACTION): New parameter.
4965 2013-10-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4967 * config/arm/arm.c (cortexa9_extra_costs): New table.
4968 (arm_cortex_a9_tune): Use cortexa9_extra_costs.
4970 2013-10-18 Jeff Law <law@redhat.com>
4972 * tree-ssa-threadupdate.c: Do not include "tm.h" or "tm_p.h".
4974 * tree-ssa-threadupdate.c: Include "dbgcnt.h".
4975 (register_jump_thread): Add "registered_jump_thread" debug
4977 * dbgcnt.def (registered_jump_thread): New debug counter.
4979 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
4981 * config/rs6000/rs6000.c: Include cgraph.h.
4983 2013-10-18 Teresa Johnson <tejohnson@google.com>
4985 * tree-ssa-tail-merge.c (replace_block_by): Update edge
4986 weights during merging.
4988 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
4990 * tree-cfg.h: Rename from tree-flow.h. Remove #includes.
4991 * tree-ssa.h: Relocate required #includes from tree-cfg.h.
4992 * tree-ssa-operands.h: Remove prototype.
4993 * tree-ssa-operands.c (virtual_operand_p): Move to gimple.c.
4994 * gimple.c (virtual_operand_p): Relocate from gimple.c.
4995 * gimple.h: Add prototype.
4996 * gimple-ssa.h: Include tree-ssa-operands.h.
4997 * tree-dump.c: Add tree-cfg.h to include list.
4998 * tree-ssa-alias.c: Add ipa-reference.h to include list.
4999 * config/alpha/alpha.c: Include gimple-ssa.h instead of tree-flow.h.
5000 * config/i386/i386.c: Don't include tree-flow.h.
5001 * config/rs6000/rs6000.c: Likewise.
5003 2013-10-18 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5005 * config/frv/frv.c (frv_init_cumulative_args): Fix wrong cast.
5007 2013-10-18 Richard Biener <rguenther@suse.de>
5009 * stor-layout.c (layout_type): Do not change TYPE_PRECISION
5010 or TYPE_UNSIGNED of integral types.
5011 (set_min_and_max_values_for_integral_type): Leave TYPE_MIN/MAX_VALUE
5012 NULL_TREE for zero-precision integral types.
5014 2013-10-18 James Greenhalgh <james.greenhalgh@arm.com>
5016 * config/aarch64/arm_neon.h
5017 (vcvt<ds>_n_<fsu><32,64>_<fsu><32,64>): Correct argument types.
5019 2013-10-17 Sriraman Tallam <tmsriram@google.com>
5022 * opth-gen.awk: Define target_flags_explicit.
5024 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
5026 * config/rs6000/rs6000.c (enum rs6000_reload_reg_type): Add new
5027 fields to the reg_addr array that describes the valid addressing
5028 mode for any register, general purpose registers, floating point
5029 registers, and Altivec registers.
5030 (FIRST_RELOAD_REG_CLASS): Likewise.
5031 (LAST_RELOAD_REG_CLASS): Likewise.
5032 (struct reload_reg_map_type): Likewise.
5033 (reload_reg_map_type): Likewise.
5034 (RELOAD_REG_VALID): Likewise.
5035 (RELOAD_REG_MULTIPLE): Likewise.
5036 (RELOAD_REG_INDEXED): Likewise.
5037 (RELOAD_REG_OFFSET): Likewise.
5038 (RELOAD_REG_PRE_INCDEC): Likewise.
5039 (RELOAD_REG_PRE_MODIFY): Likewise.
5040 (reg_addr): Likewise.
5041 (mode_supports_pre_incdec_p): New helper functions to say whether
5042 a given mode supports PRE_INC, PRE_DEC, and PRE_MODIFY.
5043 (mode_supports_pre_modify_p): Likewise.
5044 (rs6000_debug_vector_unit): Rearrange the -mdebug=reg output to
5045 print the valid address mode bits for each mode.
5046 (rs6000_debug_print_mode): Likewise.
5047 (rs6000_debug_reg_global): Likewise.
5048 (rs6000_setup_reg_addr_masks): New function to set up the address
5049 mask bits for each type.
5050 (rs6000_init_hard_regno_mode_ok): Use memset to clear arrays.
5051 Call rs6000_setup_reg_addr_masks to set up the address mask bits.
5052 (rs6000_legitimate_address_p): Use mode_supports_pre_incdec_p and
5053 mode_supports_pre_modify_p to determine if PRE_INC, PRE_DEC, and
5054 PRE_MODIFY are supported.
5055 (rs6000_output_move_128bit): Change to use {src,dest}_vmx_p for altivec
5056 registers, instead of {src,dest}_av_p.
5057 (rs6000_print_options_internal): Tweak the debug output slightly.
5059 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
5061 * config/i386/sse.md (*vec_widen_smult_even_v8si): Remove
5064 2013-10-17 Andrew MacLeod <amacleod@redhat.com>
5066 * tree-flow.h (struct omp_region): Move to omp-low.c.
5067 Remove omp_ prototypes and variables.
5068 * gimple.h (omp_reduction_init): Move prototype to omp-low.h.
5069 (copy_var_decl): Relocate prototype from tree-flow.h.
5070 * gimple.c (copy_var_decl): Relocate from omp-low.c.
5071 * tree.h: Move prototype to omp-low.h.
5072 * omp-low.h: New File. Relocate prototypes here.
5073 * omp-low.c (struct omp_region): Make local here.
5074 (root_omp_region): Make static.
5075 (copy_var_decl) Move to gimple.c.
5076 (new_omp_region): Make static.
5077 (make_gimple_omp_edges): New. Refactored from tree-cfg.c make_edges.
5078 * tree-cfg.c: Include omp-low.h.
5079 (make_edges): Factor out OMP specific bits to make_gimple_omp_edges.
5080 * gimplify.c: Include omp-low.h.
5081 * tree-parloops.c: Likewise.
5083 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
5085 * config/i386/i386.c (ix86_fixup_binary_operands): When both source
5086 operands are in memory, prefer to force non-matched operand 1 to
5089 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
5092 * config/rs6000/rs6000.c (rs6000_legitimate_address_p): Only
5093 restrict TImode addresses to single indirect registers if both
5094 -mquad-memory and -mvsx-timode are used.
5095 (rs6000_output_move_128bit): Use quad_load_store_p to determine if
5096 we should emit load/store quad. Remove using %y for quad memory
5099 * config/rs6000/rs6000.md (mov<mode>_ppc64, TI/PTImode): Add
5100 constraints to allow load/store quad on machines where TImode is
5101 not allowed in VSX registers. Use 'n' instead of 'F' constraint
5102 for TImode to load integer constants.
5104 2013-10-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5106 * config/aarch64/aarch64.c (aarch64_print_operand): Handle 'c'.
5108 2013-10-17 Marcus Shawcroft <marcus.shawcroft@arm.com>
5110 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Adjust
5111 handling of STACK_REG.
5113 2013-10-17 Richard Biener <rguenther@suse.de>
5115 PR tree-optimization/58143
5116 * tree-ssa-loop-im.c (arith_code_with_undefined_signed_overflow):
5118 (rewrite_to_defined_overflow): Likewise.
5119 (move_computations_dom_walker::before_dom): Rewrite stmts
5120 with undefined signed overflow that are not always executed
5121 into unsigned arithmetic.
5123 2013-10-16 Michael Meissner <meissner@linux.vnet.ibm.com>
5126 * config/rs6000/rs6000.opt (rs6000_isa_flags_explicit): Move the
5127 explicit isa flag to be an options variable, instead of using
5128 global_options_set. Remove define from rs6000.h.
5129 * config/rs6000/rs6000.h (rs6000_isa_flags_explicit): Likewise.
5131 * config/rs6000/rs6000.c (rs6000_option_override_internal):
5132 Initialize rs6000_isa_flags_explicit.
5133 (rs6000_function_specific_save): Add gcc_options* parameter, so
5134 that the powerpc builds after the 2013-10-15 changes.
5135 (rs6000_function_specific_restore): Likewise.
5137 2013-10-16 DJ Delorie <dj@redhat.com>
5139 * config/rl78/rl78.c (rl78_alloc_address_registers_macax): Verify
5140 op is a REG before checking REGNO.
5141 (rl78_alloc_physical_registers): Verify pattern is a SET before
5144 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5146 * config/rs6000/vector.md (vec_unpacks_hi_v4sf): Correct for
5148 (vec_unpacks_lo_v4sf): Likewise.
5149 (vec_unpacks_float_hi_v4si): Likewise.
5150 (vec_unpacks_float_lo_v4si): Likewise.
5151 (vec_unpacku_float_hi_v4si): Likewise.
5152 (vec_unpacku_float_lo_v4si): Likewise.
5154 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5156 * config/rs6000/vsx.md (vsx_concat_<mode>): Adjust output for LE.
5157 (vsx_concat_v2sf): Likewise.
5159 2013-10-16 James Greenhalgh <james.greenhalgh@arm.com>
5161 * config/aarch64/aarch64.md
5162 (*mov<mode>_aarch64): Fix output template for DUP (element) Scalar.
5164 2013-10-16 Andrew MacLeod <amacleod@redhat.com>
5166 PR tree-optimization/58697
5167 * cfgloop.c (get_estimated_loop_iterations_int): Rename from
5168 estimated_loop_iterations_int.
5169 (max_stmt_executions_int): Call get_max_loop_iterations_int.
5170 (get_max_loop_iterations_int): New. HWINT version of
5171 get_max_loop_iterations.
5172 * cfgloop.h: Add prototypes.
5173 * loop-iv.c (find_simple_exit): call get_estimated_loop_iterations_int.
5174 * loop-unroll.c (decide_peel_once_rolling): Call
5175 get_estimated_loop_iterations_int.
5176 * tree-ssa-loop-niter.c (estimated_loop_iterations_int): Add back.
5177 * tree-ssa-loop-niter.h: Tweak prototypes.
5179 2013-10-16 David Malcolm <dmalcolm@redhat.com>
5181 * gengtype-parse.c (struct_field_seq): Ignore access-control
5182 keywords ("public:" etc).
5184 2013-10-16 Marcus Shawcroft <marcus.shawcroft@arm.com>
5186 * config/aarch64/aarch64.c (aarch64_regno_regclass): Classify
5187 FRAME_POINTER_REGNUM and ARG_POINTER_REGNUM as POINTER_REGS.
5189 2013-10-16 Yvan Roux <yvan.roux@linaro.org>
5191 * config/arm/arm.opt (mlra): New option.
5192 * config/arm/arm.c (arm_lra_p): New function.
5193 (TARGET_LRA_P): Define.
5195 2013-10-16 Paulo Matos <pmatos@broadcom.com>
5197 * tree-core.h (tree_code_name): Remove.
5198 * tree.h (get_tree_code_name): New prototype.
5199 * tree.c (tree_code_name): Make static.
5200 (get_tree_code_name): New function.
5201 (dump_tree_statistics, tree_check_failed, tree_not_check_failed,
5202 tree_class_check_failed, tree_range_check_failed,
5203 tree_not_class_check_failed, omp_clause_check_failed,
5204 tree_contains_struct_check_failed, tree_operand_check_failed): Use new
5205 wrapper get_tree_code_name instead of calling tree_code_name directly.
5206 * tree-vrp.c (dump_asserts_for): Likewise.
5207 * tree-dump.c (dequeue_and_dump): Likewise.
5208 * tree-pretty-print.c (do_niy, dump_generic_node): Likewise.
5209 * tree-pretty-print.h (pp_unsupported_tree): Likewise.
5210 * lto-streamer-out.c (lto_write_tree, DFS_write_tree): Likewise.
5211 * tree-ssa-dom.c (print_expr_hash_elt): Likewise.
5212 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs,
5213 dump_ternary_rhs, dump_gimple_assign, dump_gimple_cond,
5214 dump_gimple_omp_for): Likewise.
5215 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Likewise.
5216 * tree-ssa-pre.c (print_pre_expr): Likewise.
5217 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Likewise.
5218 * print-tree.c (print_node_brief, print_node): Likewise.
5219 * gimple.c (gimple_check_failed): Likewise.
5220 * lto-streamer.c (lto_tag_name, print_lto_report): Likewise.
5221 * config/frv/frv.c (frv_init_cumulative_args): Likewise.
5222 * config/mep/mep.c (mep_validate_vliw): Likewise.
5223 * config/iq2000/iq2000.c (init_cumulative_args): Likewise.
5224 * config/rs6000/rs6000.c (init_cumulative_args): Likewise.
5226 2013-10-16 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
5228 * config/i386/i386.c (ix86_option_override_internal): Enable FMA4
5231 2013-10-16 Hans-Peter Nilsson <hp@axis.com>
5233 * config/cris/t-elfmulti (MULTILIB_OPTIONS, MULTILIB_DIRNAMES)
5234 (MULTILIB_MATCHES): Add multilib for -march=v8.
5236 2013-10-15 Sriraman Tallam <tmsriram@google.com>
5239 * optc-save-gen.awk: Add extra parameter to the save and restore
5241 * opth-gen.awk: Generate new TARGET_* macros to accept a parameter.
5242 * tree.c (build_optimization_node): New parameter. Add extra parameter
5243 to call to cl_optimization_save.
5244 (build_target_option_node): New parameter. Add extra parameter
5245 to call to cl_target_option_save.
5246 * tree.h (build_optimization_node): New parameter.
5247 (build_target_option_node): New parameter.
5248 * c-family/c-common.c (handle_optimize_attribute): Fix calls to
5249 build_optimization_node and build_target_option_node.
5250 * c-family/c-pragma.c (handle_pragma_optimize): Ditto.
5251 (handle_pragma_push_options): Ditto.
5252 * toplev.c (process_options): Ditto.
5253 * opts.c (init_options_struct): Check for opts_set non-null.
5254 * target.def (target_option.save): New parameter.
5255 (target_option.restore): New parameter.
5256 * tm.texi: Generate.
5257 * config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
5258 (ix86_pragma_target_parse): Ditto.
5259 * config/i386/i386-protos.h (ix86_valid_target_attribute_tree): New
5261 * config/rs6000/rs6000.c (rs6000_option_override_internal): Fix calls
5262 to build_optimization_node and build_target_option_node.
5263 (rs6000_valid_attribute_p): Ditto.
5264 (rs6000_pragma_target_parse): Ditto.
5265 * config/i386/i386.opt (x_ix86_target_flags_explicit): New TargetSave
5267 * config/i386/i386.h:
5268 TARGET_64BIT_P: New Macro
5269 TARGET_MMX_P: New Macro.
5270 TARGET_3DNOW_P: New Macro.
5271 TARGET_3DNOW_A_P: New Macro.
5272 TARGET_SSE_P: New Macro.
5273 TARGET_SSE2_P: New Macro.
5274 TARGET_SSE3_P: New Macro.
5275 TARGET_SSSE3_P: New Macro.
5276 TARGET_SSE4_1_P: New Macro.
5277 TARGET_SSE4_2_P: New Macro.
5278 TARGET_AVX_P: New Macro.
5279 TARGET_AVX2_P: New Macro.
5280 TARGET_AVX512F_P: New Macro.
5281 TARGET_AVX512PF_P: New Macro.
5282 TARGET_AVX512ER_P: New Macro.
5283 TARGET_AVX512CD_P: New Macro.
5284 TARGET_FMA_P: New Macro.
5285 TARGET_SSE4A_P: New Macro.
5286 TARGET_FMA4_P: New Macro.
5287 TARGET_XOP_P: New Macro.
5288 TARGET_LWP_P: New Macro.
5289 TARGET_ABM_P: New Macro.
5290 TARGET_BMI_P: New Macro.
5291 TARGET_BMI2_P: New Macro.
5292 TARGET_LZCNT_P: New Macro.
5293 TARGET_TBM_P: New Macro.
5294 TARGET_POPCNT_P: New Macro.
5295 TARGET_SAHF_P: New Macro.
5296 TARGET_MOVBE_P: New Macro.
5297 TARGET_CRC32_P: New Macro.
5298 TARGET_AES_P: New Macro.
5299 TARGET_PCLMUL_P: New Macro.
5300 TARGET_CMPXCHG16B_P: New Macro.
5301 TARGET_FSGSBASE_P: New Macro.
5302 TARGET_RDRND_P: New Macro.
5303 TARGET_F16C_P: New Macro.
5304 TARGET_RTM_P: New Macro.
5305 TARGET_HLE_P: New Macro.
5306 TARGET_RDSEED_P: New Macro.
5307 TARGET_PRFCHW_P: New Macro.
5308 TARGET_ADX_P: New Macro.
5309 TARGET_FXSR_P: New Macro.
5310 TARGET_XSAVE_P: New Macro.
5311 TARGET_XSAVEOPT_P: New Macro.
5312 TARGET_LP64_P: New Macro.
5313 TARGET_X32_P: New Macro.
5314 TARGET_FPMATH_DEFAULT_P: New Macro.
5315 TARGET_FLOAT_RETURNS_IN_80387_P: New Macro.
5316 * config/i386/i386.c (ix86_option_override_internal): New parameters.
5318 Change ix86_tune_string to access opts->x_ix86_tune_string.
5319 Change ix86_isa_flags to access opts->x_ix86_isa_flags.
5320 Change ix86_arch_string to access opts->x_ix86_arch_string.
5321 Change ix86_stringop_alg to access opts->x_ix86_stringop_alg.
5322 Change ix86_pmode to access opts->x_ix86_pmode.
5323 Change ix86_abi to access opts->x_ix86_abi.
5324 Change ix86_cmodel to access opts->x_ix86_cmodel.
5325 Change ix86_asm_dialect to access opts->x_ix86_asm_dialect.
5326 Change ix86_isa_flags_explicit to access
5327 opts->x_ix86_isa_flags_explicit.
5328 Change ix86_dump_tunes to access opts->x_ix86_dump_tunes.
5329 Change ix86_regparm to access opts->x_ix86_regparm.
5330 Change ix86_branch_cost to access opts->x_ix86_branch_cost.
5331 Change ix86_preferred_stack_boundary_arg to access
5332 opts->x_ix86_preferred_stack_boundary_arg.
5333 Change ix86_force_align_arg_pointer to access
5334 opts->x_ix86_force_align_arg_pointer.
5335 Change ix86_incoming_stack_boundar_arg to access
5336 opts->x_ix86_incoming_stack_boundar_arg.
5337 Change ix86_fpmath to access opts->x_ix86_fpmath.
5338 Change ix86_veclibabi_type to access opts->x_ix86_veclibabi_type.
5339 Change ix86_recip_name to access opts->x_ix86_recip_name.
5340 Change ix86_stack_protector_guard to access
5341 opts->x_ix86_stack_protector_guard.
5342 Change ix86_tune_memcpy_strategy to access
5343 opts->x_ix86_tune_memcpy_strategy.
5344 Change ix86_tune_memset_strategy to access
5345 opts->x_ix86_tune_memset_strategy.
5346 Change global_options to access opts.
5347 Change global_options_set to access opts_set.
5348 Change TARGET_64BIT to TARGET_64BIT_P (opts->...).
5349 Change TARGET_MMX to TARGET_MMX_P (opts->...).
5350 Change TARGET_3DNOW to TARGET_3DNOW_P (opts->...).
5351 Change TARGET_3DNOW_A to TARGET_3DNOW_A_P (opts->...).
5352 Change TARGET_SSE to TARGET_SSE_P (opts->...).
5353 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...).
5354 Change TARGET_SSE3 to TARGET_SSE3_P (opts->...).
5355 Change TARGET_SSSE3 to TARGET_SSSE3_P (opts->...).
5356 Change TARGET_SSE4_1 to TARGET_SSE4_1_P (opts->...).
5357 Change TARGET_SSE4_2 to TARGET_SSE4_2_P (opts->...).
5358 Change TARGET_AVX to TARGET_AVX_P (opts->...).
5359 Change TARGET_AVX2 to TARGET_AVX2_P (opts->...).
5360 Change TARGET_AVX512F to TARGET_AVX512F_P (opts->...).
5361 Change TARGET_AVX512PF to TARGET_AVX512PF_P (opts->...).
5362 Change TARGET_AVX512ER to TARGET_AVX512ER_P (opts->...).
5363 Change TARGET_AVX512CD to TARGET_AVX512CD_P (opts->...).
5364 Change TARGET_FMA to TARGET_FMA_P (opts->...).
5365 Change TARGET_SSE4A to TARGET_SSE4A_P (opts->...).
5366 Change TARGET_FMA4 to TARGET_FMA4_P (opts->...).
5367 Change TARGET_XOP to TARGET_XOP_P (opts->...).
5368 Change TARGET_LWP to TARGET_LWP_P (opts->...).
5369 Change TARGET_ABM to TARGET_ABM_P (opts->...).
5370 Change TARGET_BMI to TARGET_BMI_P (opts->...).
5371 Change TARGET_BMI2 to TARGET_BMI2_P (opts->...).
5372 Change TARGET_LZCNT to TARGET_LZCNT_P (opts->...).
5373 Change TARGET_TBM to TARGET_TBM_P (opts->...).
5374 Change TARGET_POPCNT to TARGET_POPCNT_P (opts->...).
5375 Change TARGET_SAHF to TARGET_SAHF_P (opts->...).
5376 Change TARGET_MOVBE to TARGET_MOVBE_P (opts->...).
5377 Change TARGET_CRC32 to TARGET_CRC32_P (opts->...).
5378 Change TARGET_AES to TARGET_AES_P (opts->...).
5379 Change TARGET_PCLMUL to TARGET_PCLMUL_P (opts->...).
5380 Change TARGET_CMPXCHG16B to TARGET_CMPXCHG16B_P (opts->...).
5381 Change TARGET_FSGSBASE to TARGET_FSGSBASE_P (opts->...).
5382 Change TARGET_RDRND to TARGET_RDRND_P (opts->...).
5383 Change TARGET_F16C to TARGET_F16C_P (opts->...).
5384 Change TARGET_RTM to TARGET_RTM_P (opts->...).
5385 Change TARGET_HLE to TARGET_HLE_P (opts->...).
5386 Change TARGET_RDSEED to TARGET_RDSEED_P (opts->...).
5387 Change TARGET_PRFCHW to TARGET_PRFCHW_P (opts->...).
5388 Change TARGET_ADX to TARGET_ADX_P (opts->...).
5389 Change TARGET_FXSR to TARGET_FXSR_P (opts->...).
5390 Change TARGET_XSAVE to TARGET_XSAVE_P (opts->...).
5391 Change TARGET_XSAVEOPT to TARGET_XSAVEOPT_P (opts->...).
5392 Change TARGET_LP64 to TARGET_LP64_P (opts->...).
5393 Change TARGET_X32 to TARGET_X32_P (opts->...).
5394 Change TARGET_FPMATH_DEFAULT to TARGET_FPMATH_DEFAULT_P (opts->...).
5395 Change TARGET_FLOAT_RETURNS_IN_80387 to
5396 TARGET_FLOAT_RETURNS_IN_80387_P (opts->...).
5397 (ix86_function_specific_save): New parameter. Use opts-> fields
5398 to replace global fields.
5399 (ix86_function_specific_restore): Ditto.
5400 (ix86_valid_target_attribute_inner_p): New parameters.
5402 Fix call to ix86_handle_option and set_option.
5403 (ix86_valid_target_attribute_tree): New parameters.
5404 Change global_options to access opts.
5405 Change global_options_set to access opts_set.
5406 Fix call to ix86_valid_target_attribute_inner_p.
5407 Change ix86_tune_string to access opts->x_ix86_tune_string.
5408 Change ix86_arch_string to access opts->x_ix86_arch_string.
5409 Change ix86_fpmath to access opts->x_ix86_fpmath
5410 Fix call to ix86_option_override_internal.
5411 Fix call to ix86_add_new_builtins.
5412 Fix calls to build_optimization_node and build_target_option_node.
5413 (ix86_valid_target_attribute_p): Remove access to global_options.
5414 Use new gcc_options structure func_options.
5415 Fix call to ix86_valid_target_attribute_tree.
5416 Fix call to build_optimization_node.
5417 (get_builtin_code_for_version): Fix call to
5418 ix86_valid_target_attribute_tree.
5420 2013-10-15 David Malcolm <dmalcolm@redhat.com>
5422 * Makefile.in (PICFLAG): New.
5423 (enable_host_shared): New.
5424 (INTERNAL_CFLAGS): Use PICFLAG.
5425 (LIBIBERTY): Use pic build of libiberty.a if configured with
5426 --enable-host-shared.
5427 * configure.ac: Add --enable-host-shared, setting up new
5429 * configure: Regenerate.
5430 * doc/install.texi (--enable-shared): Add note contrasting it with ...
5431 (--enable-host-shared): New option.
5433 2013-10-15 Richard Biener <rguenther@suse.de>
5435 * tree-tailcall.c (find_tail_calls): Don't use tail-call recursion
5436 for built-in functions.
5438 2013-10-15 Zhenqiang Chen <zhenqiang.chen@arm.com>
5440 * tree-ssa-reassoc.c: Include rtl.h and tm_p.h.
5441 (optimize_range_tests_1): New function,
5442 extracted from optimize_range_tests.
5443 (optimize_range_tests_xor): Similarly.
5444 (optimize_range_tests_diff): New function.
5445 (optimize_range_tests): Use optimize_range_tests_1.
5447 2013-10-15 Cong Hou <congh@google.com>
5449 * tree-vect-loop.c (vect_is_simple_reduction_1): Relax the
5450 requirement of the reduction pattern so that one operand of the
5451 reduction operation can come from outside of the loop.
5453 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
5455 * config/arm/neon-schedgen.ml: Remove.
5456 * config/arm/cortex-a9-neon.md: Remove comment regarding
5459 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
5461 * config/arm/types: Remove old neon types.
5463 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
5465 * config/arm/cortex-a7.md
5466 (cortex_a7_neon_type): New.
5467 (cortex_a7_neon_mul): Update for new types.
5468 (cortex_a7_neon_mla): Likewise.
5469 (cortex_a7_neon): Likewise.
5471 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
5473 * config/arm/cortex-a15-neon.md
5474 (cortex_a15_neon_type): New,
5476 (cortex_a15_neon_int_1): Remove.
5477 (cortex_a15_neon_int_2): Likewise.
5478 (cortex_a15_neon_int_3): Likewise.
5479 (cortex_a15_neon_int_4): Likewise.
5480 (cortex_a15_neon_int_5): Likewise.
5481 (cortex_a15_neon_vqneg_vqabs): Likewise.
5482 (cortex_a15_neon_vmov): Likewise.
5483 (cortex_a15_neon_vaba): Likewise.
5484 (cortex_a15_neon_vaba_qqq): Likewise.
5485 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
5486 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
5487 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
5489 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
5490 (cortex_a15_neon_mla_qqq_8_16): Likewise.
5491 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
5492 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
5493 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
5494 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
5495 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
5496 (cortex_a15_neon_shift_1): Likewise.
5497 (cortex_a15_neon_shift_2): Likewise.
5498 (cortex_a15_neon_shift_3): Likewise.
5499 (cortex_a15_neon_vshl_ddd): Likewise.
5500 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
5501 (cortex_a15_neon_vsra_vrsra): Likewise.
5502 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
5503 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
5504 (cortex_a15_neon_bp_3cycle): Likewise.
5505 (cortex_a15_neon_ldm_2): Likewise.
5506 (cortex_a15_neon_stm_2): Likewise.
5507 (cortex_a15_neon_mcr): Likewise.
5508 (cortex_a15_neon_mrc): Likewise.
5509 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
5510 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
5511 (cortex_a15_neon_fp_vmul_ddd): Likewise.
5512 (cortex_a15_neon_fp_vmul_qqd): Likewise.
5513 (cortex_a15_neon_fp_vmla_ddd): Likewise.
5514 (cortex_a15_neon_fp_vmla_qqq): Likewise.
5515 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
5516 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
5517 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
5518 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
5519 (cortex_a15_neon_bp_simple): Likewise.
5520 (cortex_a15_neon_bp_2cycle): Likewise.
5521 (cortex_a15_neon_bp_3cycle): Likewise.
5522 (cortex_a15_neon_vld1_1_2_regs): Likewise.
5523 (cortex_a15_neon_vld1_3_4_regs): Likewise.
5524 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
5525 (cortex_a15_neon_vld2_4_regs): Likewise.
5526 (cortex_a15_neon_vld3_vld4): Likewise.
5527 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
5528 (cortex_a15_neon_vst1_3_4_regs): Likewise.
5529 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
5530 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
5531 (cortex_a15_neon_vst3_vst4): Rename to...
5532 (cortex_a15_neon_vst4): This, update for new attributes.
5533 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
5534 (cortex_a15_neon_vld3_vld4_lane): Likewise.
5535 (cortex_a15_neon_vst1_vst2_lane): Likewise.
5536 (cortex_a15_neon_vst3_vst4_lane): Likewise.
5537 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
5538 (cortex_a15_neon_ldm_2): Likewise.
5539 (cortex_a15_neon_stm_2): Likewise.
5540 (cortex_a15_neon_mcr): Likewise.
5541 (cortex_a15_neon_mcr_2_mcrr): Likewise.
5542 (cortex_a15_neon_mrc): Likewise.
5543 (cortex_a15_neon_mrrc): Likewise.
5545 (cortex_a15_neon_abd): New.
5546 (cortex_a15_neon_abd_q): Likewise.
5547 (cortex_a15_neon_aba): Likewise.
5548 (cortex_a15_neon_aba_q): Likewise.
5549 (cortex_a15_neon_acc): Likewise.
5550 (cortex_a15_neon_acc_q): Likewise.
5551 (cortex_a15_neon_arith_basic): Likewise.
5552 (cortex_a15_neon_arith_complex): Likewise.
5553 (cortex_a15_neon_multiply): Likewise.
5554 (cortex_a15_neon_multiply_q): Likewise.
5555 (cortex_a15_neon_mla): Likewise.
5556 (cortex_a15_neon_mla_q): Likewise.
5557 (cortex_a15_neon_sat_mla_long): Likewise.
5558 (cortex_a15_neon_shift_acc): Likewise.
5559 (cortex_a15_neon_shift_imm_basic): Likewise.
5560 (cortex_a15_neon_shift_imm_complex): Likewise.
5561 (cortex_a15_neon_shift_reg_basic): Likewise.
5562 (cortex_a15_neon_shift_reg_basic_q): Likewise.
5563 (cortex_a15_neon_shift_reg_complex): Likewise.
5564 (cortex_a15_neon_shift_reg_complex_q): Likewise.
5565 (cortex_a15_neon_fp_negabs): Likewise
5566 (cortex_a15_neon_fp_arith): Likewise
5567 (cortex_a15_neon_fp_arith_q): Likewise
5568 (cortex_a15_neon_fp_cvt_int): Likewise
5569 (cortex_a15_neon_fp_cvt_int_q): Likewise
5570 (cortex_a15_neon_fp_cvt_16): Likewise
5571 (cortex_a15_neon_fp_mul): Likewise
5572 (cortex_a15_neon_fp_mul_q): Likewise
5573 (cortex_a15_neon_fp_mla): Likewise
5574 (cortex_a15_neon_fp_mla_q): Likewise
5575 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
5576 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
5577 (cortex_a15_neon_bitops): Likewise.
5578 (cortex_a15_neon_bitops_q): Likewise.
5579 (cortex_a15_neon_from_gp): Likewise.
5580 (cortex_a15_neon_from_gp_q): Likewise.
5581 (cortex_a15_neon_tbl3_tbl4): Likewise.
5582 (cortex_a15_neon_zip_q): Likewise.
5583 (cortex_a15_neon_to_gp): Likewise.
5584 (cortex_a15_neon_load_a): Likewise.
5585 (cortex_a15_neon_load_b): Likewise.
5586 (cortex_a15_neon_load_c): Likewise.
5587 (cortex_a15_neon_load_d): Likewise.
5588 (cortex_a15_neon_load_e): Likewise.
5589 (cortex_a15_neon_load_f): Likewise.
5590 (cortex_a15_neon_store_a): Likewise.
5591 (cortex_a15_neon_store_b): Likewise.
5592 (cortex_a15_neon_store_c): Likewise.
5593 (cortex_a15_neon_store_d): Likewise.
5594 (cortex_a15_neon_store_e): Likewise.
5595 (cortex_a15_neon_store_f): Likewise.
5596 (cortex_a15_neon_store_g): Likewise.
5597 (cortex_a15_neon_store_h): Likewise.
5598 (cortex_a15_vfp_to_from_gp): Likewise.
5600 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
5602 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
5604 (cortex_a9_neon_vshl_ddd): Remove.
5605 (cortex_a9_neon_vst3_vst4): Likewise.
5606 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
5608 (cortex_a9_neon_bit_ops_q): New.
5610 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
5611 (cortex_a9_neon_int_2): Likewise.
5612 (cortex_a9_neon_int_3): Likewise.
5613 (cortex_a9_neon_int_4): Likewise.
5614 (cortex_a9_neon_int_5): Likewise.
5615 (cortex_a9_neon_vqneg_vqabs): Likewise.
5616 (cortex_a9_neon_vmov): Likewise.
5617 (cortex_a9_neon_vaba): Likewise.
5618 (cortex_a9_neon_vaba_qqq): Likewise.
5619 (cortex_a9_neon_shift_1): Likewise.
5620 (cortex_a9_neon_shift_2): Likewise.
5621 (cortex_a9_neon_shift_3): Likewise.
5622 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
5623 (cortex_a9_neon_vsra_vrsra): Likewise.
5624 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
5625 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
5626 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
5628 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
5629 (cortex_a9_neon_mla_qqq_8_16): Likewise.
5630 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
5632 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
5633 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
5634 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
5635 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
5636 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
5637 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
5638 (cortex_a9_neon_fp_vsum): Likewise.
5639 (cortex_a9_neon_fp_vmul_ddd): Likewise.
5640 (cortex_a9_neon_fp_vmul_qqd): Likewise.
5641 (cortex_a9_neon_fp_vmla_ddd): Likewise.
5642 (cortex_a9_neon_fp_vmla_qqq): Likewise.
5643 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
5644 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
5645 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
5646 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
5647 (cortex_a9_neon_bp_simple): Likewise.
5648 (cortex_a9_neon_bp_2cycle): Likewise.
5649 (cortex_a9_neon_bp_3cycle): Likewise.
5650 (cortex_a9_neon_ldr): Likewise.
5651 (cortex_a9_neon_str): Likewise.
5652 (cortex_a9_neon_vld1_1_2_regs): Likewise.
5653 (cortex_a9_neon_vld1_3_4_regs): Likewise.
5654 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
5655 (cortex_a9_neon_vld2_4_regs): Likewise.
5656 (cortex_a9_neon_vld3_vld4): Likewise.
5657 (cortex_a9_neon_vld1_vld2_lane): Likewise.
5658 (cortex_a9_neon_vld3_vld4_lane): Likewise.
5659 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
5660 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
5661 (cortex_a9_neon_vst1_3_4_regs): Likewise.
5662 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
5663 (cortex_a9_neon_vst1_vst2_lane): Likewise.
5664 (cortex_a9_neon_vst3_vst4_lane): Likewise.
5665 (cortex_a9_neon_mcr): Likewise.
5666 (cortex_a9_neon_mcr_2_mcrr): Likewise.
5667 (cortex_a9_neon_mrc): Likewise.
5668 (cortex_a9_neon_mrrc): Likewise.
5670 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
5672 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
5674 (cortex_a8_neon_vshl_ddd): Remove.
5675 (cortex_a8_neon_vst3_vst4): Likewise.
5676 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
5678 (cortex_a8_neon_bit_ops_q): New.
5680 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
5681 (cortex_a8_neon_int_2): Likewise..
5682 (cortex_a8_neon_int_3): Likewise.
5683 (cortex_a8_neon_int_5): Likewise.
5684 (cortex_a8_neon_vqneg_vqabs): Likewise.
5685 (cortex_a8_neon_int_4): Likewise.
5686 (cortex_a8_neon_vaba): Likewise.
5687 (cortex_a8_neon_vaba_qqq): Likewise.
5688 (cortex_a8_neon_shift_1): Likewise.
5689 (cortex_a8_neon_shift_2): Likewise.
5690 (cortex_a8_neon_shift_3): Likewise.
5691 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
5692 (cortex_a8_neon_vsra_vrsra): Likewise.
5693 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
5694 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
5695 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
5697 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
5698 (cortex_a8_neon_mla_qqq_8_16): Likewise.
5699 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
5701 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
5702 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
5703 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
5704 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
5705 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
5706 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
5707 (cortex_a8_neon_fp_vsum): Likewise.
5708 (cortex_a8_neon_fp_vmul_ddd): Likewise.
5709 (cortex_a8_neon_fp_vmul_qqd): Likewise.
5710 (cortex_a8_neon_fp_vmla_ddd): Likewise.
5711 (cortex_a8_neon_fp_vmla_qqq): Likewise.
5712 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
5713 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
5714 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
5715 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
5716 (cortex_a8_neon_bp_simple): Likewise.
5717 (cortex_a8_neon_bp_2cycle): Likewise.
5718 (cortex_a8_neon_bp_3cycle): Likewise.
5719 (cortex_a8_neon_ldr): Likewise.
5720 (cortex_a8_neon_str): Likewise.
5721 (cortex_a8_neon_vld1_1_2_regs): Likewise.
5722 (cortex_a8_neon_vld1_3_4_regs): Likewise.
5723 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
5724 (cortex_a8_neon_vld2_4_regs): Likewise.
5725 (cortex_a8_neon_vld3_vld4): Likewise.
5726 (cortex_a8_neon_vld1_vld2_lane): Likewise.
5727 (cortex_a8_neon_vld3_vld4_lane): Likewise.
5728 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
5729 (cortex_a8_neon_vst1_3_4_regs): Likewise.
5730 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
5731 (cortex_a8_neon_vst1_vst2_lane): Likewise.
5732 (cortex_a8_neon_vst3_vst4_lane): Likewise.
5733 (cortex_a8_neon_mcr): Likewise.
5734 (cortex_a8_neon_mcr_2_mcrr): Likewise.
5735 (cortex_a8_neon_mrc): Likewise.
5736 (cortex_a8_neon_mrrc): Likewise.
5738 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
5740 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
5742 * config/aarch64/aarch64-simd.md (neon_type): Remove.
5743 (aarch64_simd_dup<mode>): Add "type" attribute.
5744 (aarch64_dup_lane<mode>): Likewise.
5745 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
5746 (*aarch64_simd_mov<mode>): Likewise.
5747 (aarch64_simd_mov_from_<mode>low): Likewise.
5748 (aarch64_simd_mov_from_<mode>high): Likewise.
5749 (orn<mode>3): Likewise.
5750 (bic<mode>3): Likewise.
5751 (add<mode>3): Likewise.
5752 (sub<mode>3): Likewise.
5753 (mul<mode>3): Likewise.
5754 (*aarch64_mul3_elt<mode>): Likewise.
5755 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
5756 (*aarch64_mul3_elt_to_128df): Likewise.
5757 (*aarch64_mul3_elt_to_64v2df): Likewise.
5758 (neg<mode>2): Likewise.
5759 (abs<mode>2): Likewise.
5760 (abd<mode>_3): Likewise.
5761 (aba<mode>_3): Likewise.
5762 (fabd<mode>_3): Likewise.
5763 (*fabd_scalar<mode>3): Likewise.
5764 (and<mode>3): Likewise.
5765 (ior<mode>3): Likewise.
5766 (xor<mode>3): Likewise.
5767 (one_cmpl<mode>2): Likewise.
5768 (aarch64_simd_vec_set<mode>): Likewise.
5769 (aarch64_simd_lshr<mode>): Likewise.
5770 (aarch64_simd_ashr<mode>): Likewise.
5771 (aarch64_simd_imm_shl<mode>): Likewise.
5772 (aarch64_simd_reg_sshl<mode): Likewise.
5773 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
5774 (aarch64_simd_reg_shl<mode>_signed): Likewise.
5775 (aarch64_simd_vec_setv2di): Likewise.
5776 (aarch64_simd_vec_set<mode>): Likewise.
5777 (aarch64_mla<mode>): Likewise.
5778 (*aarch64_mla_elt<mode>): Likewise.
5779 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
5780 (aarch64_mls<mode>): Likewise.
5781 (*aarch64_mls_elt<mode>): Likewise.
5782 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
5783 (<su><maxmin><mode>3): Likewise.
5784 (move_lo_quad_<mode>): Likewise.
5785 (aarch64_simd_move_hi_quad_<mode>): Likewise.
5786 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
5787 (vec_pack_trunc_<mode>): Likewise.
5788 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
5789 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
5790 (*aarch64_<su>mlal_lo<mode>): Likewise.
5791 (*aarch64_<su>mlal_hi<mode>): Likewise.
5792 (*aarch64_<su>mlsl_lo<mode>): Likewise.
5793 (*aarch64_<su>mlsl_hi<mode>): Likewise.
5794 (*aarch64_<su>mlal<mode>): Likewise.
5795 (*aarch64_<su>mlsl<mode>): Likewise.
5796 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
5797 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
5798 (add<mode>3): Likewise.
5799 (sub<mode>3): Likewise.
5800 (mul<mode>3): Likewise.
5801 (div<mode>3): Likewise.
5802 (neg<mode>2): Likewise.
5803 (abs<mode>2): Likewise.
5804 (fma<mode>4): Likewise.
5805 (*aarch64_fma4_elt<mode>): Likewise.
5806 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
5807 (*aarch64_fma4_elt_to_128df): Likewise.
5808 (*aarch64_fma4_elt_to_64v2df): Likewise.
5809 (fnma<mode>4): Likewise.
5810 (*aarch64_fnma4_elt<mode>): Likewise.
5811 (*aarch64_fnma4_elt_<vswap_width_name><mode>
5812 (*aarch64_fnma4_elt_to_128df): Likewise.
5813 (*aarch64_fnma4_elt_to_64v2df): Likewise.
5814 (<frint_pattern><mode>2): Likewise.
5815 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
5816 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
5817 (vec_unpacks_lo_v4sf): Likewise.
5818 (aarch64_float_extend_lo_v2df): Likewise.
5819 (vec_unpacks_hi_v4sf): Likewise.
5820 (aarch64_float_truncate_lo_v2sf): Likewise.
5821 (aarch64_float_truncate_hi_v4sf): Likewise.
5822 (aarch64_vmls<mode>): Likewise.
5823 (<su><maxmin><mode>3): Likewise.
5824 (<maxmin_uns><mode>3): Likewise.
5825 (reduc_<sur>plus_<mode>): Likewise.
5826 (reduc_<sur>plus_v2di): Likewise.
5827 (reduc_<sur>plus_v2si): Likewise.
5828 (reduc_<sur>plus_<mode>): Likewise.
5829 (aarch64_addpv4sf): Likewise.
5830 (clz<mode>2): Likewise.
5831 (reduc_<maxmin_uns>_<mode>): Likewise.
5832 (reduc_<maxmin_uns>_v2di): Likewise.
5833 (reduc_<maxmin_uns>_v2si): Likewise.
5834 (reduc_<maxmin_uns>_<mode>): Likewise.
5835 (reduc_<maxmin_uns>_v4sf): Likewise.
5836 (aarch64_simd_bsl<mode>_internal): Likewise.
5837 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
5838 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
5839 (aarch64_get_lane<mode>): Likewise.
5840 (*aarch64_combinez<mode>): Likewise.
5841 (aarch64_combine<mode>): Likewise.
5842 (aarch64_simd_combine<mode>): Likewise.
5843 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
5844 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
5845 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
5846 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
5847 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
5848 (aarch64_<sur>h<addsub><mode>): Likewise.
5849 (aarch64_<sur><addsub>hn<mode>): Likewise.
5850 (aarch64_<sur><addsub>hn2<mode>): Likewise.
5851 (aarch64_pmul<mode>): Likewise.
5852 (aarch64_<su_optab><optab><mode>): Likewise.
5853 (aarch64_<sur>qadd<mode>): Likewise.
5854 (aarch64_sqmovun<mode>): Likewise.
5855 (aarch64_<sur>qmovn<mode>): Likewise.
5856 (aarch64_s<optab><mode>): Likewise.
5857 (aarch64_sq<r>dmulh<mode>): Likewise.
5858 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
5859 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
5860 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
5861 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
5862 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
5863 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
5864 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
5865 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
5866 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
5867 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
5868 (aarch64_sqdmull<mode>): Likewise.
5869 (aarch64_sqdmull_lane<mode>_internal): Likewise.
5870 (aarch64_sqdmull_n<mode>): Likewise.
5871 (aarch64_sqdmull2<mode>_internal): Likewise.
5872 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
5873 (aarch64_sqdmull2_n<mode>_internal): Likewise.
5874 (aarch64_<sur>shl<mode>): Likewise.
5875 (aarch64_<sur>q<r>shl<mode>
5876 (aarch64_<sur>shll_n<mode>): Likewise.
5877 (aarch64_<sur>shll2_n<mode>): Likewise.
5878 (aarch64_<sur>shr_n<mode>): Likewise.
5879 (aarch64_<sur>sra_n<mode>): Likewise.
5880 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
5881 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
5882 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
5883 (aarch64_cm<optab><mode>): Likewise.
5884 (aarch64_cm<optab>di): Likewise.
5885 (aarch64_cm<optab><mode>): Likewise.
5886 (aarch64_cm<optab>di): Likewise.
5887 (aarch64_cmtst<mode>): Likewise.
5888 (aarch64_cmtstdi): Likewise.
5889 (aarch64_cm<optab><mode>): Likewise.
5890 (*aarch64_fac<optab><mode>): Likewise.
5891 (aarch64_addp<mode>): Likewise.
5892 (aarch64_addpdi): Likewise.
5893 (sqrt<mode>2): Likewise.
5894 (vec_load_lanesoi<mode>): Likewise.
5895 (vec_store_lanesoi<mode>): Likewise.
5896 (vec_load_lanesci<mode>): Likewise.
5897 (vec_store_lanesci<mode>): Likewise.
5898 (vec_load_lanesxi<mode>): Likewise.
5899 (vec_store_lanesxi<mode>): Likewise.
5900 (*aarch64_mov<mode>): Likewise.
5901 (aarch64_ld2<mode>_dreg): Likewise.
5902 (aarch64_ld2<mode>_dreg): Likewise.
5903 (aarch64_ld3<mode>_dreg): Likewise.
5904 (aarch64_ld3<mode>_dreg): Likewise.
5905 (aarch64_ld4<mode>_dreg): Likewise.
5906 (aarch64_ld4<mode>_dreg): Likewise.
5907 (aarch64_tbl1<mode>): Likewise.
5908 (aarch64_tbl2v16qi): Likewise.
5909 (aarch64_combinev16qi): Likewise.
5910 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
5911 (aarch64_st2<mode>_dreg): Likewise.
5912 (aarch64_st2<mode>_dreg): Likewise.
5913 (aarch64_st3<mode>_dreg): Likewise.
5914 (aarch64_st3<mode>_dreg): Likewise.
5915 (aarch64_st4<mode>_dreg): Likewise.
5916 (aarch64_st4<mode>_dreg): Likewise.
5917 (*aarch64_simd_ld1r<mode>): Likewise.
5918 (aarch64_frecpe<mode>): Likewise.
5919 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
5920 (aarch64_frecps<mode>): Likewise.
5922 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
5924 * config/arm/iterators.md (V_elem_ch): New.
5926 (VQH_type): Likewise.
5927 * config/arm/arm.md (is_neon_type): New.
5928 (conds): Use is_neon_type.
5929 (anddi3_insn): Update type attribute.
5930 (xordi3_insn): Likewise.
5931 (one_cmpldi2): Likewise.
5932 * config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
5933 * config/arm/neon.md (neon_mov): Update type attribute.
5934 (*movmisalign<mode>_neon_store): Likewise.
5935 (*movmisalign<mode>_neon_load): Likewise.
5936 (vec_set<mode>_internal): Likewise.
5937 (vec_set<mode>_internal): Likewise.
5938 (vec_setv2di_internal): Likewise.
5939 (vec_extract<mode>): Likewise.
5940 (vec_extract<mode>): Likewise.
5941 (vec_extractv2di): Likewise.
5942 (*add<mode>3_neon): Likewise.
5943 (adddi3_neon): Likewise.
5944 (*sub<mode>3_neon): Likewise.
5945 (subdi3_neon): Likewise.
5946 (fma<VCVTF:mode>4): Likewise.
5947 (fma<VCVTF:mode>4_intrinsic): Likewise.
5948 (*fmsub<VCVTF:mode>4): Likewise.
5949 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
5950 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
5951 (ior<mode>3): Likewise.
5952 (and<mode>3): Likewise.
5953 (orn<mode>3_neon): Likewise.
5954 (orndi3_neon): Likewise.
5955 (bic<mode>3_neon): Likewise.
5956 (bicdi3_neon): Likewise.
5957 (xor<mode>3): Likewise.
5958 (one_cmpl<mode>2): Likewise.
5959 (abs<mode>2): Likewise.
5960 (neg<mode>2): Likewise.
5961 (negdi2_neon): Likewise.
5962 (*umin<mode>3_neon): Likewise.
5963 (*umax<mode>3_neon): Likewise.
5964 (*smin<mode>3_neon): Likewise.
5965 (*smax<mode>3_neon): Likewise.
5966 (vashl<mode>3): Likewise.
5967 (vashr<mode>3_imm): Likewise.
5968 (vlshr<mode>3_imm): Likewise.
5969 (ashl<mode>3_signed): Likewise.
5970 (ashl<mode>3_unsigned): Likewise.
5971 (neon_load_count): Likewise.
5972 (ashldi3_neon_noclobber): Likewise.
5973 (ashldi3_neon): Likewise.
5974 (signed_shift_di3_neon): Likewise.
5975 (unsigned_shift_di3_neon): Likewise.
5976 (ashrdi3_neon_imm_noclobber): Likewise.
5977 (lshrdi3_neon_imm_noclobber): Likewise.
5978 (<shift>di3_neon): Likewise.
5979 (widen_ssum<mode>3): Likewise.
5980 (widen_usum<mode>3): Likewise.
5981 (quad_halves_<code>v4si): Likewise.
5982 (quad_halves_<code>v4sf): Likewise.
5983 (quad_halves_<code>v8hi): Likewise.
5984 (quad_halves_<code>v16qi): Likewise.
5985 (reduc_splus_v2di): Likewise.
5986 (neon_vpadd_internal<mode>): Likewise.
5987 (neon_vpsmin<mode>): Likewise.
5988 (neon_vpsmax<mode>): Likewise.
5989 (neon_vpumin<mode>): Likewise.
5990 (neon_vpumax<mode>): Likewise.
5991 (*ss_add<mode>_neon): Likewise.
5992 (*us_add<mode>_neon): Likewise.
5993 (*ss_sub<mode>_neon): Likewise.
5994 (*us_sub<mode>_neon): Likewise.
5995 (neon_vadd<mode>_unspec): Likewise.
5996 (neon_vaddl<mode>): Likewise.
5997 (neon_vaddw<mode>): Likewise.
5998 (neon_vhadd<mode>): Likewise.
5999 (neon_vqadd<mode>): Likewise.
6000 (neon_vaddhn<mode>): Likewise.
6001 (neon_vmul<mode>): Likewise.
6002 (neon_vfms<VCVTF:mode>): Likewise.
6003 (neon_vmlal<mode>): Likewise.
6004 (neon_vmls<mode>): Likewise.
6005 (neon_vmlsl<mode>): Likewise.
6006 (neon_vqdmulh<mode>): Likewise.
6007 (neon_vqdmlal<mode>): Likewise.
6008 (neon_vqdmlsl<mode>): Likewise.
6009 (neon_vmull<mode>): Likewise.
6010 (neon_vqdmull<mode>): Likewise.
6011 (neon_vsub<mode>_unspec): Likewise.
6012 (neon_vsubl<mode>): Likewise.
6013 (neon_vsubw<mode>): Likewise.
6014 (neon_vqsub<mode>): Likewise.
6015 (neon_vhsub<mode>): Likewise.
6016 (neon_vsubhn<mode>): Likewise.
6017 (neon_vceq<mode>): Likewise.
6018 (neon_vcge<mode>): Likewise.
6019 (neon_vcgeu<mode>): Likewise.
6020 (neon_vcgt<mode>): Likewise.
6021 (neon_vcgtu<mode>): Likewise.
6022 (neon_vcle<mode>): Likewise.
6023 (neon_vclt<mode>): Likewise.
6024 (neon_vcage<mode>): Likewise.
6025 (neon_vcagt<mode>): Likewise.
6026 (neon_vtst<mode>): Likewise.
6027 (neon_vabd<mode>): Likewise.
6028 (neon_vabdl<mode>): Likewise.
6029 (neon_vaba<mode>): Likewise.
6030 (neon_vabal<mode>): Likewise.
6031 (neon_vmax<mode>): Likewise.
6032 (neon_vmin<mode>): Likewise.
6033 (neon_vpaddl<mode>): Likewise.
6034 (neon_vpadal<mode>): Likewise.
6035 (neon_vpmax<mode>): Likewise.
6036 (neon_vpmin<mode>): Likewise.
6037 (neon_vrecps<mode>): Likewise.
6038 (neon_vrsqrts<mode>): Likewise.
6039 (neon_vqabs<mode>): Likewise.
6040 (neon_vqneg<mode>): Likewise.
6041 (neon_vcls<mode>): Likewise.
6042 (clz<mode>2): Likewise.
6043 (popcount<mode>2): Likewise.
6044 (neon_vrecpe<mode>): Likewise.
6045 (neon_vrsqrte<mode>): Likewise.
6046 (neon_vget_lane<mode>_sext_internal): Likewise.
6047 (neon_vget_lane<mode>_zext_internal): Likewise.
6048 (neon_vdup_n<mode>): Likewise.
6049 (neon_vdup_n<mode>): Likewise.
6050 (neon_vdup_nv2di): Likewise.
6051 (neon_vdup_lane<mode>_interal): Likewise.
6052 (*neon_vswp<mode>): Likewise.
6053 (neon_vcombine<mode>): Likewise.
6054 (float<mode><V_cvtto>2): Likewise.
6055 (floatuns<mode><V_cvtto>2): Likewise.
6056 (fix_trunc<mode><V_cvtto>2): Likewise.
6057 (fixuns_trunc<mode><V_cvtto>2
6058 (neon_vcvt<mode>): Likewise.
6059 (neon_vcvt<mode>): Likewise.
6060 (neon_vcvtv4sfv4hf): Likewise.
6061 (neon_vcvtv4hfv4sf): Likewise.
6062 (neon_vcvt_n<mode>): Likewise.
6063 (neon_vcvt_n<mode>): Likewise.
6064 (neon_vmovn<mode>): Likewise.
6065 (neon_vqmovn<mode>): Likewise.
6066 (neon_vqmovun<mode>): Likewise.
6067 (neon_vmovl<mode>): Likewise.
6068 (neon_vmul_lane<mode>): Likewise.
6069 (neon_vmul_lane<mode>): Likewise.
6070 (neon_vmull_lane<mode>): Likewise.
6071 (neon_vqdmull_lane<mode>): Likewise.
6072 (neon_vqdmulh_lane<mode>): Likewise.
6073 (neon_vqdmulh_lane<mode>): Likewise.
6074 (neon_vmla_lane<mode>): Likewise.
6075 (neon_vmla_lane<mode>): Likewise.
6076 (neon_vmlal_lane<mode>): Likewise.
6077 (neon_vqdmlal_lane<mode>): Likewise.
6078 (neon_vmls_lane<mode>): Likewise.
6079 (neon_vmls_lane<mode>): Likewise.
6080 (neon_vmlsl_lane<mode>): Likewise.
6081 (neon_vqdmlsl_lane<mode>): Likewise.
6082 (neon_vext<mode>): Likewise.
6083 (neon_vrev64<mode>): Likewise.
6084 (neon_vrev32<mode>): Likewise.
6085 (neon_vrev16<mode>): Likewise.
6086 (neon_vbsl<mode>_internal): Likewise.
6087 (neon_vshl<mode>): Likewise.
6088 (neon_vqshl<mode>): Likewise.
6089 (neon_vshr_n<mode>): Likewise.
6090 (neon_vshrn_n<mode>): Likewise.
6091 (neon_vqshrn_n<mode>): Likewise.
6092 (neon_vqshrun_n<mode>): Likewise.
6093 (neon_vshl_n<mode>): Likewise.
6094 (neon_vqshl_n<mode>): Likewise.
6095 (neon_vqshlu_n<mode>): Likewise.
6096 (neon_vshll_n<mode>): Likewise.
6097 (neon_vsra_n<mode>): Likewise.
6098 (neon_vsri_n<mode>): Likewise.
6099 (neon_vsli_n<mode>): Likewise.
6100 (neon_vtbl1v8qi): Likewise.
6101 (neon_vtbl2v8qi): Likewise.
6102 (neon_vtbl3v8qi): Likewise.
6103 (neon_vtbl4v8qi): Likewise.
6104 (neon_vtbl1v16qi): Likewise.
6105 (neon_vtbl2v16qi): Likewise.
6106 (neon_vcombinev16qi): Likewise.
6107 (neon_vtbx1v8qi): Likewise.
6108 (neon_vtbx2v8qi): Likewise.
6109 (neon_vtbx3v8qi): Likewise.
6110 (neon_vtbx4v8qi): Likewise.
6111 (*neon_vtrn<mode>_insn): Likewise.
6112 (*neon_vzip<mode>_insn): Likewise.
6113 (*neon_vuzp<mode>_insn): Likewise.
6114 (neon_vld1<mode>): Likewise.
6115 (neon_vld1_lane<mode>): Likewise.
6116 (neon_vld1_lane<mode>): Likewise.
6117 (neon_vld1_dup<mode>): Likewise.
6118 (neon_vld1_dup<mode>): Likewise.
6119 (neon_vld1_dupv2di): Likewise.
6120 (neon_vst1<mode>): Likewise.
6121 (neon_vst1_lane<mode>): Likewise.
6122 (neon_vst1_lane<mode>): Likewise.
6123 (neon_vld2<mode>): Likewise.
6124 (neon_vld2<mode>): Likewise.
6125 (neon_vld2_lane<mode>): Likewise.
6126 (neon_vld2_lane<mode>): Likewise.
6127 (neon_vld2_dup<mode>): Likewise.
6128 (neon_vst2<mode>): Likewise.
6129 (neon_vst2<mode>): Likewise.
6130 (neon_vst2_lane<mode>): Likewise.
6131 (neon_vst2_lane<mode>): Likewise.
6132 (neon_vld3<mode>): Likewise.
6133 (neon_vld3qa<mode>): Likewise.
6134 (neon_vld3qb<mode>): Likewise.
6135 (neon_vld3_lane<mode>): Likewise.
6136 (neon_vld3_lane<mode>): Likewise.
6137 (neon_vld3_dup<mode>): Likewise.
6138 (neon_vst3<mode>): Likewise.
6139 (neon_vst3qa<mode>): Likewise.
6140 (neon_vst3qb<mode>): Likewise.
6141 (neon_vst3_lane<mode>): Likewise.
6142 (neon_vst3_lane<mode>): Likewise.
6143 (neon_vld4<mode>): Likewise.
6144 (neon_vld4qa<mode>): Likewise.
6145 (neon_vld4qb<mode>): Likewise.
6146 (neon_vld4_lane<mode>): Likewise.
6147 (neon_vld4_lane<mode>): Likewise.
6148 (neon_vld4_dup<mode>): Likewise.
6149 (neon_vst4<mode>): Likewise.
6150 (neon_vst4qa<mode>): Likewise.
6151 (neon_vst4qb<mode>): Likewise.
6152 (neon_vst4_lane<mode>): Likewise.
6153 (neon_vst4_lane<mode>): Likewise.
6154 (neon_vec_unpack<US>_lo_<mode>): Likewise.
6155 (neon_vec_unpack<US>_hi_<mode>): Likewise.
6156 (neon_vec_<US>mult_lo_<mode>): Likewise.
6157 (neon_vec_<US>mult_hi_<mode>): Likewise.
6158 (neon_vec_<US>shiftl_<mode>): Likewise.
6159 (neon_unpack<US>_<mode>): Likewise.
6160 (neon_vec_<US>mult_<mode>): Likewise.
6161 (vec_pack_trunc_<mode>): Likewise.
6162 (neon_vec_pack_trunc_<mode>): Likewise.
6163 (neon_vabd<mode>_2): Likewise.
6164 (neon_vabd<mode>_3): Likewise.
6166 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
6168 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
6169 (load_pair): Update type attribute.
6170 (store_pair): Update type attribute.
6171 * config/aarch64/iterators.md (q): New.
6173 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
6175 * config/arm/types.md: Add new types for Neon insns.
6177 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
6178 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6179 Sergey Lega <sergey.s.lega@intel.com>
6180 Anna Tikhonova <anna.tikhonova@intel.com>
6181 Ilya Tocar <ilya.tocar@intel.com>
6182 Andrey Turetskiy <andrey.turetskiy@intel.com>
6183 Ilya Verbin <ilya.verbin@intel.com>
6184 Kirill Yukhin <kirill.yukhin@intel.com>
6185 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6187 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
6188 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
6189 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
6191 (srcp14<mode>): Ditto.
6192 (rsqrt14<mode>): Ditto.
6193 (rsqrt14<mode>): Ditto.
6194 (avx512f_vmscalef<mode>): Ditto.
6195 (avx512f_scalef<mode>): Ditto.
6196 (avx512f_getexp<mode>): Ditto.
6197 (avx512f_sgetexp<mode>): Ditto.
6198 (avx512f_fixupimm<mode>): Ditto.
6199 (avx512f_sfixupimm<mode>): Ditto.
6200 (avx512f_rndscale<mode>): Ditto.
6201 (*avx512er_exp2<mode>): Ditto.
6202 (*avx512er_rcp28<mode>): Ditto.
6203 (avx512er_rsqrt28<mode>): Ditto.
6204 (avx512f_getmant<mode>): Ditto.
6205 (avx512f_getmant<mode>): Ditto.
6206 (avx512f_rndscale<mode>): Fix formatting.
6208 2013-10-15 Martin Jambor <mjambor@suse.cz>
6210 * ipa-utils.h (ipa_edge_within_scc): Declare.
6211 * ipa-cp.c (edge_within_scc): Moved...
6212 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
6214 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
6215 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6216 Sergey Lega <sergey.s.lega@intel.com>
6217 Anna Tikhonova <anna.tikhonova@intel.com>
6218 Ilya Tocar <ilya.tocar@intel.com>
6219 Andrey Turetskiy <andrey.turetskiy@intel.com>
6220 Ilya Verbin <ilya.verbin@intel.com>
6221 Kirill Yukhin <kirill.yukhin@intel.com>
6222 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6224 * config/i386/predicates.md (const_8_to_15_operand): New.
6225 (const_16_to_31_operand): Ditto.
6226 * config/i386/sse.md (V8FI): New.
6228 (reduc_splus_v8df): Ditto.
6229 (reduc_splus_v16sf): Ditto.
6230 (avx512f_vextract<shuffletype>32x4_1): Ditto.
6231 (vec_extract_hi_<mode>): Ditto.
6232 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
6233 (vec_set_lo_<mode>): Ditto.
6234 (vec_set_hi_<mode>): Ditto.
6235 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
6236 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
6237 (avx512f_pshufd_1): Ditto.
6238 (avx512f_broadcast<mode>): Ditto.
6239 (avx512f_broadcast<mode>): Ditto.
6240 (define_split): Split vec_extract_lo into move.
6241 (ssequartermode): Ditto.
6242 (ssedoublemode): Extened with wider modes.
6243 (vec_extract_lo_<mode>): Ditto.
6245 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
6246 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6247 Sergey Lega <sergey.s.lega@intel.com>
6248 Anna Tikhonova <anna.tikhonova@intel.com>
6249 Ilya Tocar <ilya.tocar@intel.com>
6250 Andrey Turetskiy <andrey.turetskiy@intel.com>
6251 Ilya Verbin <ilya.verbin@intel.com>
6252 Kirill Yukhin <kirill.yukhin@intel.com>
6253 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6255 * config/i386/predicates.md (register_or_constm1_operand): New.
6256 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
6257 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
6258 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
6259 UNSPEC_SCATTER_PREFETCH
6261 (avx512f_ucmp<mode>3): Ditto.
6262 (avx512f_vternlog<mode>): Ditto.
6263 (avx512f_align<mode>): Ditto.
6264 (<shift_insn><mode>3): Ditto.
6265 (avx512f_<rotate>v<mode>): Ditto.
6266 (avx512f_<rotate><mode>): Ditto.
6267 (avx512f_eq<mode>3): Ditto.
6268 (avx512f_eq<mode>3_1): Ditto.
6269 (avx512f_gt<mode>3): Ditto.
6270 (avx512f_testm<mode>3): Ditto.
6271 (avx512f_testnm<mode>3): Ditto.
6272 (avx512pf_gatherpf<mode>): Ditto.
6273 (*avx512pf_gatherpf<mode>_mask): Ditto.
6274 (*avx512pf_gatherpf<mode>): Ditto.
6275 (avx512pf_scatterpf<mode>): Ditto.
6276 (*avx512pf_scatterpf<mode>_mask): Ditto.
6277 (*avx512pf_scatterpf<mode>): Ditto.
6278 (avx512f_vec_dup_gpr<mode>): Ditto.
6279 (clz<mode>2): Ditto.
6280 (conflict<mode>): Ditto.
6281 (REDUC_SMINMAX_MODE): Extened with wider modes.
6282 (reduc_<code>_<mode>): Ditto.
6283 (vlshr<mode>3): Ditto.
6284 (vashl<mode>3): Ditto.
6286 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
6287 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6288 Sergey Lega <sergey.s.lega@intel.com>
6289 Anna Tikhonova <anna.tikhonova@intel.com>
6290 Ilya Tocar <ilya.tocar@intel.com>
6291 Andrey Turetskiy <andrey.turetskiy@intel.com>
6292 Ilya Verbin <ilya.verbin@intel.com>
6293 Kirill Yukhin <kirill.yukhin@intel.com>
6294 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6296 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
6299 (avx512fmaskmode): Ditto.
6300 (bcstscalarsuff): Ditto.
6301 (avx512f_blendm<mode>): Ditto.
6302 (cmp_imm_predicate): Ditto.
6303 (avx512f_cmp<mode>3): Ditto.
6304 (avx512f_vec_dup<mode>): Ditto.
6305 (avx512f_vec_dup_mem<mode>): Ditto.
6306 (avx512f_vpermi2var<mode>3): Ditto.
6307 (avx512f_vpermt2var<mode>3): Ditto.
6308 (vec_init<mode>): Ditto.
6309 (avx512f_gathersi<mode>): Ditto.
6310 (*avx512f_gathersi<mode>): Ditto.
6311 (*avx512f_gathersi<mode>_2): Ditto.
6312 (avx512f_gatherdi<mode>): Ditto.
6313 (*avx512f_gatherdi<mode>): Ditto.
6314 (*avx512f_gatherdi<mode>_2): Ditto.
6315 (avx512f_scattersi<mode>): Ditto.
6316 (*avx512f_scattersi<mode>): Ditto.
6317 (avx512f_scatterdi<mode>): Ditto.
6318 (*avx512f_scatterdi<mode>): Ditto.
6319 (sseintprefix): Extened with wider modes.
6320 (VEC_GATHER_IDXSI): Ditto.
6321 (VEC_GATHER_IDXDI): Ditto.
6322 (VEC_GATHER_SRCDI): Ditto.
6324 2013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
6325 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6327 * config/arm/t-aprofile: New file.
6328 * config.gcc: Handle --with-multilib-list option.
6330 2013-10-15 Bernd Schmidt <bernds@codesourcery.com>
6332 * reload1.c (reloads_unique_chain_p): Ensure that r1 is
6335 2013-10-15 Richard Biener <rguenther@suse.de>
6337 * tree-loop-distribution.c (build_empty_rdg): Inline into
6339 (rdg_flag_vertex): Inline into single user.
6340 (rdg_flag_vertex_and_dependent): Likewise.
6341 (build_rdg_partition_for_vertex): Remove processed bitmap.
6342 (rdg_build_partitions): Simplify.
6344 2013-10-15 Richard Biener <rguenther@suse.de>
6346 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
6347 Restructure forwarding through conversions and copies to
6348 avoid performing copy-propagation the wrong way. Adjust
6349 recursion invocations.
6350 (forward_propagate_addr_expr): Add argument stating if we
6351 are recursing from a single-use.
6352 (ssa_forward_propagate_and_combine): Adjust.
6354 2013-10-14 David Malcolm <dmalcolm@redhat.com>
6356 * dumpfile.h (gcc::dump_manager): New class, to hold state
6357 relating to dumpfile management.
6358 (get_dump_file_name): Remove in favor of method of dump_manager.
6359 (dump_initialized_p): Likewise.
6360 (dump_start): Likewise.
6361 (dump_finish): Likewise.
6362 (dump_switch_p): Likewise.
6363 (dump_register): Likewise.
6364 (get_dump_file_info): Likewise.
6365 * context.c (gcc::context::context): Construct the dump_manager
6367 * context.h (gcc::context::get_dumps): New.
6368 (gcc::context::m_dumps): New.
6369 * coverage.c (coverage_init): Port to dump_manager API.
6370 * dumpfile.c (extra_dump_files): Convert to field of gcc::dump_manager.
6371 (extra_dump_files_in_use): Likewise.
6372 (extra_dump_files_alloced): Likewise.
6373 (gcc::dump_manager::dump_manager): New.
6374 (dump_register): Convert to...
6375 (gcc::dump_manager::dump_register): ...method, replacing
6376 function-static next_dump with m_next_dump field.
6377 (get_dump_file_info): Convert to...
6378 (gcc::dump_manager::get_dump_file_info): ...method.
6379 (get_dump_file_name): Convert to...
6380 (gcc::dump_manager::get_dump_file_name): ...method.
6381 (dump_start): Convert to...
6382 (gcc::dump_manager::dump_start): ...method.
6383 (dump_finish): Convert to...
6384 (gcc::dump_manager::dump_finish): ...method.
6385 (dump_begin): Replace body with...
6386 (gcc::dump_manager::dump_begin): ...new method.
6387 (dump_phase_enabled_p): Convert to...
6388 (gcc::dump_manager::dump_phase_enabled_p): ...method.
6389 (dump_phase_enabled_p): Convert to...
6390 (gcc::dump_manager::dump_phase_enabled_p): ...method.
6391 (dump_initialized_p): Convert to...
6392 (gcc::dump_manager::dump_initialized_p): ...method.
6393 (dump_flag_name): Replace body with...
6394 (gcc::dump_manager::dump_flag_name): ...new method.
6395 (dump_enable_all): Convert to...
6396 (gcc::dump_manager::dump_enable_all): ...new method.
6397 (opt_info_enable_passes): Convert to...
6398 (gcc::dump_manager::opt_info_enable_passes): ...new method.
6399 (dump_switch_p_1): Convert to...
6400 (gcc::dump_manager::dump_switch_p_1): ...new method.
6401 (dump_switch_p): Convert to...
6402 (gcc::dump_manager::dump_switch_p): ...new method.
6403 (opt_info_switch_p): Port to dump_manager API.
6404 (enable_rtl_dump_file): Likewise.
6405 * opts-global.c (handle_common_deferred_options): Port to new
6407 * passes.c (pass_manager::finish_optimization_passes): Likewise.
6408 (pass_manager::register_one_dump_file): Likewise.
6409 (pass_manager::register_pass): Likewise.
6410 (pass_init_dump_file): Likewise.
6411 (pass_fini_dump_file): Likewise.
6412 * statistics.c (statistics_early_init): Likewise.
6414 2013-10-14 Richard Biener <rguenther@suse.de>
6416 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
6417 iterative_hash_canonical_type, gimple_canonical_type_hash,
6418 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
6419 gimple_register_canonical_type, print_gimple_types_stats,
6420 free_gimple_type_tables): Move to lto/lto.c
6421 (gt-gimple.h): Do not include.
6422 * gimple.h (gimple_register_canonical_type,
6423 print_gimple_types_stats, free_gimple_type_tables): Remove.
6424 * Makefile.in (GTFILES): Remove gimple.c.
6426 2013-10-14 Travis Snoozy <quandary@remstate.com>
6429 * config/msp430/msp430.c (msp430_option_override): Correct thinko
6430 scanning for msp430x targets.
6432 2013-10-14 Eric Botcazou <ebotcazou@adacore.com>
6435 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
6436 (registers_ok_for_ldd_peep): Move around.
6437 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
6438 * config/sparc/sparc.md (widening peepholes): Use it.
6440 2013-10-14 Richard Biener <rguenther@suse.de>
6444 * gimple.c (iterative_hash_canonical_type): Make sure to
6445 record the hash into the correct hashtable slot.
6447 2013-10-13 Eric Botcazou <ebotcazou@adacore.com>
6449 PR rtl-optimization/58662
6450 * combine.c (try_combine): Take into account death nodes on I2 when
6451 splitting a PARALLEL of two independent SETs. Fix dump message.
6453 2013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
6456 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
6457 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
6458 * config/sh/t-sh (sh_treg_combine.o): New entry.
6459 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
6460 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
6461 (register_sh_passes): New function. Register sh_treg_combine pass.
6462 (sh_option_override): Invoke it.
6463 (sh_canonicalize_comparison): Handle op0_preserve_value.
6464 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
6465 opportunities. Canonicalize branch condition.
6466 (nott): Allow only if pseudos can be created for non-SH2A.
6468 2013-10-12 H.J. Lu <hongjiu.lu@intel.com>
6471 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
6472 (ix86_expand_movmem): Replace copy_addr_to_reg with
6473 ix86_copy_addr_to_reg.
6474 (ix86_expand_setmem): Likewise.
6476 2013-10-12 Alexander Monakov <amonakov@ispras.ru>
6478 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
6479 provided by ix86_fp_compare_mode instead of CCFPUmode.
6481 2013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
6483 * config/aarch64/arm_neon.h
6484 (vtbx<1,3>_<psu>8): Fix register constriants.
6486 2013-10-11 Jeff Law <law@redhat.com>
6488 PR tree-optimization/58640
6489 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
6490 threading paths that cross over two loop entry points.
6492 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6494 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
6495 handle vector float as well.
6496 (*vsx_le_perm_load_v4si): Likewise.
6497 (*vsx_le_perm_store_v2di): Likewise.
6498 (*vsx_le_perm_store_v4si): Likewise.
6500 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6502 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
6503 directly to circumvent subtract from splat{31} workaround.
6504 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
6506 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
6507 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
6508 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
6509 (altivec_vperm_<mode>): Convert to define_insn_and_split to
6510 separate big and little endian logic.
6511 (*altivec_vperm_<mode>_internal): New define_insn.
6512 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
6513 separate big and little endian logic.
6514 (*altivec_vperm_<mode>_uns_internal): New define_insn.
6515 (vec_permv16qi): Add little endian logic.
6517 2013-10-11 Marc Glisse <marc.glisse@inria.fr>
6519 * doc/extend.texi (returns_nonnull): Remove arguments.
6521 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
6522 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6523 Sergey Lega <sergey.s.lega@intel.com>
6524 Anna Tikhonova <anna.tikhonova@intel.com>
6525 Ilya Tocar <ilya.tocar@intel.com>
6526 Andrey Turetskiy <andrey.turetskiy@intel.com>
6527 Ilya Verbin <ilya.verbin@intel.com>
6528 Kirill Yukhin <kirill.yukhin@intel.com>
6529 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6531 * config/i386/sse.md (VI48F_256_512): New.
6532 (avx2_permvar<mode>): Change to ...
6533 (<avx2_avx512f>_permvar<mode>): This.
6535 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
6536 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6537 Sergey Lega <sergey.s.lega@intel.com>
6538 Anna Tikhonova <anna.tikhonova@intel.com>
6539 Ilya Tocar <ilya.tocar@intel.com>
6540 Andrey Turetskiy <andrey.turetskiy@intel.com>
6541 Ilya Verbin <ilya.verbin@intel.com>
6542 Kirill Yukhin <kirill.yukhin@intel.com>
6543 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6545 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
6546 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
6547 * config/i386/sse.md (VI4_AVX): New.
6548 (sf2simodelower): Ditto.
6549 (sse2_cvtps2dq): Change to ...
6550 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
6552 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
6553 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6554 Sergey Lega <sergey.s.lega@intel.com>
6555 Anna Tikhonova <anna.tikhonova@intel.com>
6556 Ilya Tocar <ilya.tocar@intel.com>
6557 Andrey Turetskiy <andrey.turetskiy@intel.com>
6558 Ilya Verbin <ilya.verbin@intel.com>
6559 Kirill Yukhin <kirill.yukhin@intel.com>
6560 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6562 * config/i386/sse.md (V_512): New.
6564 (vcond<V_512:mode><VF_512:mode>): Ditto.
6565 (vcond<V_512:mode><VI_512:mode>): Ditto.
6566 (vcondu<V_512:mode><VI_512:mode>): Ditto.
6568 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
6569 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6570 Sergey Lega <sergey.s.lega@intel.com>
6571 Anna Tikhonova <anna.tikhonova@intel.com>
6572 Ilya Tocar <ilya.tocar@intel.com>
6573 Andrey Turetskiy <andrey.turetskiy@intel.com>
6574 Ilya Verbin <ilya.verbin@intel.com>
6575 Kirill Yukhin <kirill.yukhin@intel.com>
6576 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6578 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
6579 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
6581 (fma<mode>4): Removed condition.
6582 (fms<mode>4): Ditto.
6583 (fnma<mode>4): Ditto.
6584 (fnms<mode>4): Ditto.
6585 (fma4i_fmadd_<mode>): Ditto.
6586 (*fma_fmadd_<mode>): Ditto.
6587 (*fma_fmsub_<mode>): Ditto.
6588 (*fma_fnmadd_<mode>): Ditto.
6589 (*fma_fnmsub_<mode>): Ditto.
6590 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
6591 (*fma_fmaddsub_<mode>): Ditto.
6592 (*fma_fmsubadd_<mode>): Ditto.
6593 (*fmai_fmadd_<mode>): Ditto.
6594 (*fmai_fmsub_<mode>): Ditto.
6595 (*fmai_fnmadd_<mode>): Ditto.
6596 (*fmai_fnmsub_<mode>): Ditto.
6598 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
6599 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6600 Sergey Lega <sergey.s.lega@intel.com>
6601 Anna Tikhonova <anna.tikhonova@intel.com>
6602 Ilya Tocar <ilya.tocar@intel.com>
6603 Andrey Turetskiy <andrey.turetskiy@intel.com>
6604 Ilya Verbin <ilya.verbin@intel.com>
6605 Kirill Yukhin <kirill.yukhin@intel.com>
6606 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6608 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
6609 (VI124_256): Changed to ...
6610 (VI124_256_48_512): This.
6611 (ssepackmode): Extended with wider modes.
6612 (<code><mode>3): Changed iterator.
6613 (*avx2_<code><mode>3): Ditto.
6614 (vec_pack_trunc_<mode>): Ditto.
6616 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
6617 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6618 Sergey Lega <sergey.s.lega@intel.com>
6619 Anna Tikhonova <anna.tikhonova@intel.com>
6620 Ilya Tocar <ilya.tocar@intel.com>
6621 Andrey Turetskiy <andrey.turetskiy@intel.com>
6622 Ilya Verbin <ilya.verbin@intel.com>
6623 Kirill Yukhin <kirill.yukhin@intel.com>
6624 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6626 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
6627 (VI8F_256_512): Ditto.
6628 (abs<mode>2): Changed iterator.
6629 (avx2_perm<mode>): Changed to ...
6630 (<avx2_avx512f>_perm<mode>): This.
6631 (avx2_perm<mode>_1): Changed to ...
6632 (<avx2_avx512f>_perm<mode>_1): This.
6634 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
6635 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6636 Sergey Lega <sergey.s.lega@intel.com>
6637 Anna Tikhonova <anna.tikhonova@intel.com>
6638 Ilya Tocar <ilya.tocar@intel.com>
6639 Andrey Turetskiy <andrey.turetskiy@intel.com>
6640 Ilya Verbin <ilya.verbin@intel.com>
6641 Kirill Yukhin <kirill.yukhin@intel.com>
6642 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6644 * config/i386/sse.md (VI48_AVX512F): New.
6645 (VI48_AVX2): Changed to ...
6646 (VI48_AVX2_48_AVX512F): This.
6647 (avx2_ashrv<mode>): Changed to ...
6648 (<avx2_avx512f>_ashrv<mode>): This.
6649 (avx2_<shift_insn>v<mode>): Changed to ...
6650 (<avx2_avx512f>_<shift_insn>v<mode>): This.
6652 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
6653 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6654 Sergey Lega <sergey.s.lega@intel.com>
6655 Anna Tikhonova <anna.tikhonova@intel.com>
6656 Ilya Tocar <ilya.tocar@intel.com>
6657 Andrey Turetskiy <andrey.turetskiy@intel.com>
6658 Ilya Verbin <ilya.verbin@intel.com>
6659 Kirill Yukhin <kirill.yukhin@intel.com>
6660 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6662 * config/i386/sse.md (VI4_AVX512F): New.
6663 (VI8_AVX2_AVX512F): Ditto.
6664 (mul<mode>3): Extended with wider modes.
6665 (*<sse4_1_avx2>_mul<mode>3): Ditto.
6666 (mul<mode>3): Ditto.
6667 (vec_widen_<s>mult_odd_<mode>): Ditto.
6669 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
6670 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6671 Sergey Lega <sergey.s.lega@intel.com>
6672 Anna Tikhonova <anna.tikhonova@intel.com>
6673 Ilya Tocar <ilya.tocar@intel.com>
6674 Andrey Turetskiy <andrey.turetskiy@intel.com>
6675 Ilya Verbin <ilya.verbin@intel.com>
6676 Kirill Yukhin <kirill.yukhin@intel.com>
6677 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6679 * config/i386/sse.md (VI2_AVX512F): New.
6680 (VI124_AVX512F): Ditto.
6681 (sseunpackmode): Extended with wider modes.
6682 (sseunpackfltmode): Ditto.
6683 (vec_unpacks_float_hi_<mode>): Ditto.
6684 (vec_unpacks_float_lo_<mode>): Ditto.
6685 (vec_unpacku_float_hi_<mode>): Ditto.
6686 (vec_unpacku_float_lo_<mode>): Ditto.
6687 (vec_unpacks_lo_<mode>): Ditto.
6688 (vec_unpacks_hi_<mode>): Ditto.
6689 (vec_unpacku_lo_<mode>): Ditto.
6690 (vec_unpacku_hi_<mode>): Ditto.
6692 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
6693 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6694 Sergey Lega <sergey.s.lega@intel.com>
6695 Anna Tikhonova <anna.tikhonova@intel.com>
6696 Ilya Tocar <ilya.tocar@intel.com>
6697 Andrey Turetskiy <andrey.turetskiy@intel.com>
6698 Ilya Verbin <ilya.verbin@intel.com>
6699 Kirill Yukhin <kirill.yukhin@intel.com>
6700 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6702 * config/i386/i386.md (multdiv): New.
6703 (multdiv_mnemonic): Ditto.
6704 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
6705 (<sse>_vm<multdiv_mnemonic><mode>3): This.
6706 (<sse>_vmdiv<mode>3): Removed.
6708 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
6709 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6710 Sergey Lega <sergey.s.lega@intel.com>
6711 Anna Tikhonova <anna.tikhonova@intel.com>
6712 Ilya Tocar <ilya.tocar@intel.com>
6713 Andrey Turetskiy <andrey.turetskiy@intel.com>
6714 Ilya Verbin <ilya.verbin@intel.com>
6715 Kirill Yukhin <kirill.yukhin@intel.com>
6716 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6718 * config/i386/sse.md (V): Extended with wider modes.
6720 (ssehalfvecmode): Ditto.
6722 (ssepackfltmode): Ditto.
6723 (avx_vec_concat<mode>): Ditto.
6724 (V_256_512): New iterator.
6725 (VF2_512_256): Ditto.
6726 (si2dfmode): New attribute.
6727 (si2dfmodelower): Ditto.
6729 (concat_tg_mode): Ditto.
6730 (floatv4siv4df2): Changed to ...
6731 (float<si2dfmodelower><mode>2): This.
6732 (avx_cvtps2pd256): Changed to ...
6733 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
6734 (vec_pack_trunc_v4df): Changed to ...
6735 (vec_pack_trunc_<mode>): This.
6736 (avx_vpermil<mode>): Changed to ...
6737 (<sse2_avx_avx512f>_vpermil<mode>): This.
6738 (<fixsuffix>fix_truncv8dfv8si2): New.
6739 (vec_pack_sfix_trunc_v8df): Ditto.
6740 (avx512f_rndscale<mode>): Ditto.
6741 (avx512f_roundpd512): Ditto.
6742 (vec_pack_ufix_trunc_<mode>): Updated iterator.
6744 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
6745 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6746 Sergey Lega <sergey.s.lega@intel.com>
6747 Anna Tikhonova <anna.tikhonova@intel.com>
6748 Ilya Tocar <ilya.tocar@intel.com>
6749 Andrey Turetskiy <andrey.turetskiy@intel.com>
6750 Ilya Verbin <ilya.verbin@intel.com>
6751 Kirill Yukhin <kirill.yukhin@intel.com>
6752 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6754 * config/i386/i386.md (any_fix): New iterator.
6755 (fixsuffix): New attribute.
6756 * config/i386/sse.md (VF1): Extened with wider modes.
6760 (sseintvecmodelower): Ditto.
6761 (ssescalarmode): Ditto.
6762 (ssescalarnum): Ditto.
6764 (ssexmmmode): Ditto.
6765 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
6766 (<sse>_rcp<mode>2): Change iterator.
6767 (rsqrt<mode>2): Ditto.
6768 (<sse>_rsqrt<mode>2): Ditto.
6769 (avx2_vec_dup<mode>): Ditto.
6770 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
6771 (round<mode>2_sfix): Ditto.
6772 (avx2_pbroadcast<mode>): Ditto.
6773 (*andnot<mode>3): Handle XI mode.
6774 (*<code><mode>3): Ditto.
6775 (AVXTOSSEMODE): Removed.
6776 (avx_vpermil<mode>): Changed to ...
6777 (<sse2_avx_avx512f>_vpermil<mode>): This.
6779 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
6780 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6781 Sergey Lega <sergey.s.lega@intel.com>
6782 Anna Tikhonova <anna.tikhonova@intel.com>
6783 Ilya Tocar <ilya.tocar@intel.com>
6784 Andrey Turetskiy <andrey.turetskiy@intel.com>
6785 Ilya Verbin <ilya.verbin@intel.com>
6786 Kirill Yukhin <kirill.yukhin@intel.com>
6787 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6789 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
6790 (<sse>_comi): Ditto.
6791 (<sse>_ucomi): Ditto.
6792 (sse_cvtss2siq_2): Ditto.
6793 (sse2_cvtsd2si): Ditto.
6794 (sse2_cvtsd2siq): Ditto.
6795 (sse2_cvttsd2si): Ditto.
6796 (sse2_cvttsd2siq): Ditto.
6797 (<shift_insn><mode>3): Ditto.
6798 (sse2_cvtsi2sdq): Update constraint and prefix.
6799 (sse_cvtsi2ss): Update prefix.
6800 (sse_cvtsi2ssq): Ditto.
6802 2013-10-11 Jakub Jelinek <jakub@redhat.com>
6804 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
6805 ignore internal calls.
6807 2013-10-11 Richard Biener <rguenther@suse.de>
6809 * tree-pretty-print.c (dump_generic_node): Allow to dump both (D)
6810 and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with (OVF) if
6811 TREE_OVERFLOW is set.
6813 2013-10-11 Thomas Schwinge <thomas@codesourcery.com>
6815 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
6817 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
6818 * gimple.h: Likewise.
6820 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
6822 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
6825 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
6827 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
6828 * configure: Regenerate.
6830 2013-10-11 Jakub Jelinek <jakub@redhat.com>
6832 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
6833 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
6834 formatting fixes, use pp_colon instead of pp_character (..., ':'),
6835 similarly pp_right_paren.
6836 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
6837 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
6838 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
6839 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
6841 (omp_declare_simd_clauses_equal,
6842 omp_remove_redundant_declare_simd_attrs): New functions.
6843 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
6844 (walk_tree_1): Handle new OpenMP 4.0 clauses.
6845 * tree.h (OMP_LOOP_CHECK): Define.
6846 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
6847 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
6848 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
6849 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
6850 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
6851 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
6852 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
6853 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
6854 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
6855 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
6856 OMP_CLAUSE_SIMDLEN_EXPR): Define.
6857 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
6858 (omp_remove_redundant_declare_simd_attrs): New prototype.
6859 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
6860 GIMPLE_OMP_TEAMS): New codes.
6861 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
6862 * omp-low.c (struct omp_context): Add cancel_label and cancellable
6864 (target_nesting_level): New variable.
6865 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
6866 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
6867 for collapse > 1 static schedule unless ordered.
6868 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
6869 (determine_parallel_type): Adjust get_ws_args_for caller.
6870 (install_var_field): Handle mask & 4 for double indirection.
6871 (scan_sharing_clauses): Ignore shared clause on teams construct.
6872 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
6873 (create_omp_child_function): If inside target or declare target
6874 constructs, set "omp declare target" attribute on the child function.
6875 (find_combined_for): New function.
6876 (scan_omp_parallel): Handle combined loops.
6877 (scan_omp_target, scan_omp_teams): New functions.
6878 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
6879 restrictions and set ctx->cancellable for cancellable constructs.
6880 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
6881 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
6882 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
6883 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
6884 (omp_clause_aligned_alignment): New function.
6885 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR on decls.
6886 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
6887 on teams constructs. Handle user defined reductions and new
6889 (lower_reduction_clauses): Don't set placeholder to address of ref
6890 if it has already the right type.
6891 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
6892 (expand_parallel_call): Use the new non-_start suffixed builtins,
6893 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
6894 and GOMP_parallel_end after the call.
6895 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
6896 (expand_omp_for_init_counts): Handle combined loops.
6897 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
6899 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
6900 end of cancellable loops.
6901 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
6902 Likewise. Handle collapse > 1 loops.
6903 (expand_omp_simd): Handle combined loops.
6904 (expand_omp_for): Add inner_stmt argument, adjust callers of
6905 expand_omp_for* functions, use expand_omp_for_static*chunk even
6906 for collapse > 1 unless ordered.
6907 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
6908 of cancellable sections.
6909 (expand_omp_single): Remove need_barrier variable, just rely on
6910 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
6911 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
6912 (expand_omp_atomic_load, expand_omp_atomic_store,
6913 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
6914 (expand_omp_target): New function.
6915 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
6916 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
6917 (build_omp_regions_1): Immediately close region for
6918 GF_OMP_TARGET_KIND_UPDATE.
6919 (maybe_add_implicit_barrier_cancel): New function.
6920 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
6922 (lower_omp_single): Likewise. Add clobber after the barrier.
6923 (lower_omp_taskgroup): New function.
6924 (lower_omp_for): Handle combined loops. Adjust
6925 lower_rec_input_clauses caller. Handle cancellation.
6926 (lower_depend_clauses): New function.
6927 (lower_omp_taskreg): Lower depend clauses. Adjust
6928 lower_rec_input_clauses caller. Add clobber after the call. Handle
6930 (lower_omp_target, lower_omp_teams): New functions.
6931 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
6932 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
6933 and GOMP_cancellation_point calls.
6934 (lower_omp): Fold stmts inside of target region.
6935 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
6936 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
6937 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
6938 (BT_FN_VOID_OMPFN_PTR_UINT,
6939 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
6940 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
6941 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
6942 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
6943 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
6944 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
6945 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
6946 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
6947 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
6948 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
6949 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
6950 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
6951 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
6952 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
6953 BUILT_IN_GOMP_PARALLEL_END.
6954 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
6955 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
6956 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
6957 GF_OMP_FOR_KIND_DISTRIBUTE.
6958 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
6959 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
6960 (dump_gimple_omp_return): Print lhs if it has any.
6961 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
6962 gimple_omp_atomic_seq_cst_p.
6963 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
6964 and GIMPLE_OMP_TEAMS.
6965 * langhooks.c (lhd_omp_mappable_type): New function.
6966 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
6967 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
6969 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
6970 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
6971 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and ORT_TARGET.
6972 (struct gimplify_omp_ctx): Add combined_loop field.
6973 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
6974 on stmts inside of target region.
6975 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
6976 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
6977 ORT_TARGET and ORT_TARGET_DATA.
6978 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
6980 (omp_notice_threadprivate_variable): Complain about threadprivate
6981 variables in target region.
6982 (omp_notice_variable): Complain about vars with non-mappable type
6983 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
6984 (omp_check_private): Ignore ORT_TARGET* regions.
6985 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
6986 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
6987 (find_combined_omp_for): New function.
6988 (gimplify_omp_for): Handle gimplification of combined loops.
6989 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
6991 (gimplify_omp_target_update): New function.
6992 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
6993 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
6994 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
6995 (gimplify_body): If fndecl has "omp declare target" attribute, add
6996 implicit ORT_TARGET context around it.
6997 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
6998 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
6999 * tree-nested.c (convert_nonlocal_reference_stmt,
7000 convert_local_reference_stmt, convert_gimple_call): Handle
7001 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
7002 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
7003 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
7004 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
7005 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
7006 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
7007 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
7008 BUILT_IN_GOMP_LOOP_END_CANCEL,
7009 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
7010 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
7011 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
7012 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
7013 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
7014 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
7015 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
7016 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
7017 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
7018 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
7019 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
7020 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
7021 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
7022 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
7023 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
7024 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
7025 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
7026 gimple_build_omp_teams): New functions.
7027 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
7028 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
7029 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
7030 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
7031 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
7032 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
7033 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
7034 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
7035 GF_OMP_ATOMIC_SEQ_CST): New.
7036 (gimple_build_omp_taskgroup, gimple_build_omp_target,
7037 gimple_build_omp_teams): New prototypes.
7038 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
7039 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
7040 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
7041 GIMPLE_OMP_SINGLE as end of range.
7042 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
7043 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
7044 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
7045 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
7046 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
7047 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
7048 gimple_omp_target_kind, gimple_omp_target_set_kind,
7049 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
7050 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
7051 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
7052 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
7053 gimple_omp_teams_set_clauses): New inlines.
7054 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
7055 and GIMPLE_OMP_TASKGROUP.
7056 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
7057 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
7058 enum omp_clause_proc_bind_kind): New.
7059 (union omp_clause_subcode): Add depend_kind, map_kind and
7060 proc_bind_kind fields.
7061 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
7062 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
7063 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
7064 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
7065 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
7067 2013-10-10 Teresa Johnson <tejohnson@google.com>
7069 * predict.c (tree_estimate_probability): Add new parameter
7070 for estimate_bb_frequencies.
7071 (estimate_bb_frequencies): Add new parameter to force estimation.
7072 (rebuild_frequencies): When max frequency in function is small,
7073 recompute counts from frequencies.
7074 * predict.h (estimate_bb_frequencies): New parameter.
7076 2013-10-10 David Malcolm <dmalcolm@redhat.com>
7078 * ipa-inline.c (ipa_inline): Fix leak of "order" when
7079 optimizations are disabled.
7081 2013-10-10 David Malcolm <dmalcolm@redhat.com>
7083 * coverage.c (coverage_finish): Fix leak of da_file_name.
7085 2013-10-10 Jan Hubicka <jh@suse.cz>
7087 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
7088 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
7091 2013-10-10 Jakub Jelinek <jakub@redhat.com>
7094 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
7095 if any labels are in FALLTHRU_BB, use a special label emitted
7096 immediately after the asm goto insn rather than label_rtx
7098 (expand_asm_stmt): Adjust caller.
7099 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
7100 edge if the last insn in predecessor is a jump with single successor,
7101 but it isn't simplejump_p.
7103 2013-10-10 Richard Biener <rguenther@suse.de>
7105 PR tree-optimization/58656
7106 * tree-ssa-pre.c (phi_translate): Do not cache failed translations.
7108 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
7110 * gimplify.c: Include expr.h and tm_p.h for targets with special
7111 va-arg padding requirements.
7113 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
7115 * tree-flow.h: Move some prototypes to gimple.h.
7116 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
7117 * gimple.h: Relocate some prototypes from tree-flow.h
7118 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
7120 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
7121 (build_va_arg_indirect_ref): Relocate and make static.
7122 (std_gimplify_va_arg_expr): Relocate here.
7123 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
7124 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
7126 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7128 * doc/md.texi: Document the mnemonic attribute.
7130 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7133 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
7134 (set_attr_alternative x ...) when searching for user defined
7137 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
7139 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
7141 2013-10-09 Easwaran Raman <eraman@google.com>
7143 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
7144 * cfgexpand.c (defer_stack_allocation): ...use here
7145 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
7147 2013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
7149 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
7150 (operand_equal_for_value_replacement): New function, extracted from
7151 value_replacement and enhanced to catch more cases.
7152 (value_replacement): Use operand_equal_for_value_replacement.
7154 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
7156 * loop-doloop.c (doloop_modify, doloop_optimize): Use
7157 get_max_loop_iterations.
7159 2013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7161 * config/arm/aarch-common.c (arm_early_load_addr_dep):
7162 Place comment above function.
7164 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
7166 * tree-flow.h: Remove all remaining prototypes, enums and structs that
7167 are not related to tree-cfg.c.
7168 * tree-ssa-address.h: New file. Relocate prototypes.
7169 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
7170 (addr_for_mem_ref): New. Combine call to get_address_description and
7171 return addr_for_mem_ref.
7172 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
7173 * tree-ssa-live.h: Adjust prototypes.
7174 * passes.c: Include tree-ssa-live.h.
7175 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
7176 * graphite.c (graphite_transform_loops): Make static.
7177 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
7178 make_pass_graphite, pass_data_graphite_transforms,
7179 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
7180 * ipa-pure-const.c (warn_function_noreturn): Make static.
7181 (execute_warn_function_noreturn, gate_warn_function_noreturn,
7182 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
7183 Relocate from tree-cfg.c
7184 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
7186 (execute_warn_function_noreturn, gate_warn_function_noreturn,
7187 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
7188 Move to ipa-pure-const.c.
7189 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
7190 Relocate from tree-optimize.c.
7191 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
7192 make_pass_fixup_cfg): Move to tree-cfg.c.
7193 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
7194 Relocate some prototypes.
7195 * tree-data-ref.h (tree_check_data_deps) Add prototype.
7196 * tree-dump.c (dump_function_to_file): Remove prototype.
7197 Add tree-flow.h to the include file.
7198 * tree-dump.h: Remove prototype.
7199 * tree-parloops.h: New File. Add prototypes.
7200 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
7201 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
7202 from tree-ssa-loop.c.
7203 * tree-predcom.c (run_tree_predictive_commoning,
7204 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
7205 Relocate here from tree-ssa-loop.c.
7206 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
7207 ssa_name_values.release ().
7208 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
7209 (ssa_name_values): Relocate from tree-flow.h.
7210 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
7211 * tree-ssa-loop.c (run_tree_predictive_commoning,
7212 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
7213 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
7214 make_pass_graphite, pass_data_graphite_transforms,
7215 make_pass_graphite_transforms, gate_tree_parallelize_loops,
7216 tree_parallelize_loops, pass_data_parallelize_loops,
7217 make_pass_parallelize_loops): Move to other files.
7218 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
7220 * tree.h: Remove prototypes from tree-address.c.
7222 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
7224 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
7225 (struct int_tree_map): Move to tree-hasher.h
7226 (SCALE, LABEL, PERCENT): Move to gimple.h
7227 * tree-flow-inline.h: Delete. Move functions to other files.
7228 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
7229 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
7230 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
7231 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
7232 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
7233 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
7234 * tree-hasher.h: Don't include tree-flow.h.
7235 (struct int_tree_map): Relocate from tree-flow.h.
7236 * tree-sra.c (contains_view_convert_expr_p): Relocate from
7237 tree-flow-inline.h and make static.
7238 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
7240 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
7241 tree-flow-inline.h and make static.
7242 * tree.h (is_global_var, may_be_aliased): Relocate from
7244 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
7245 * value-prof.c: No longer include tree-flow-inline.h.
7246 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
7248 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
7250 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
7251 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
7252 (enum move_pos): Move to tree-ssa-loop-im.h
7253 * cfgloop.h: Move some prototypes.
7254 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
7255 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
7256 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
7257 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
7258 (loop_containing_stmt): Relocate from tree-flow-inline.h.
7259 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
7260 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
7261 (enum move_pos): Relocate here.
7262 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
7264 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
7265 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
7266 make_pass_lim): Relocate here from tree-ssa-loop.c.
7267 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
7269 (loop_edge_to_cancel, unloop_loops): Make static.
7270 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
7271 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
7272 (tree_complete_unroll, gate_tree_complete_unroll,
7273 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
7274 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
7275 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
7276 * tree-ssa-loop-ivopts.c: Remove local prototypes.
7277 (stmt_invariant_in_loop_p): Remove unused function.
7278 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
7279 * tree-ssa-loop-manip.h: New file. Add prototypes.
7280 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
7281 (gcov_type_to_double_int): Move to cfgloop.h.
7282 (double_int_cmp, bound_index,
7283 estimate_numbers_of_iterations_loop): Make static.
7284 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
7285 (max_loop_iterations): Factor out get_max_loop_iterations.
7286 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
7288 * tree-ssa-loop-niter.h: New file. Add prototypes.
7289 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
7290 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
7291 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
7292 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
7293 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
7294 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
7295 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
7296 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
7297 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
7298 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
7299 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
7300 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
7301 pass_data_complete_unroll, make_pass_complete_unroll,
7302 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
7303 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
7304 tree-ssa-loop-ivcanon.c.
7305 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
7306 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
7307 tree-ssa-loop-prefetch.c.
7308 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
7310 (get_lsm_tmp_name): Relocate and add suffix parameter.
7311 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
7312 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
7313 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
7314 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
7315 (get_estimated_loop_iterations): Factor out accessor from
7316 estimated_loop_iterations in tree-ssa-loop-niter.c.
7317 (get_max_loop_iterations): Factor out accessor from
7318 _max_loop_iterations in tree-ssa-niter.c.
7319 * loop-unroll.c (decide_unroll_constant_iterations,
7320 decide_unroll_runtime_iterations, decide_peel_simple,
7321 decide_unroll_stupid): Use new get_* accessors.
7323 2013-10-09 Marc Glisse <marc.glisse@inria.fr>
7325 PR tree-optimization/20318
7326 * doc/extend.texi (returns_nonnull): New function attribute.
7327 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
7329 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
7330 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
7332 2013-10-09 Eric Botcazou <ebotcazou@adacore.com>
7335 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
7336 false if both components are bitfields.
7338 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
7340 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
7341 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
7342 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
7343 * config/aarch64/aarch64.h
7344 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
7345 * config/aarch64/aarch64-simd-builtins.def
7346 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
7348 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
7350 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
7352 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
7354 * config/aarch64/arm_neon.h (vdiv_f64): Added.
7356 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
7358 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
7359 (vneg_f64): New intrinsic.
7360 (vneg_s8): Asm replaced with C.
7361 (vneg_s16): Likewise.
7362 (vneg_s32): Likewise.
7363 (vneg_s64): New intrinsic.
7364 (vnegq_f32): Asm replaced with C.
7365 (vnegq_f64): Likewise.
7366 (vnegq_s8): Likewise.
7367 (vnegq_s16): Likewise.
7368 (vnegq_s32): Likewise.
7369 (vnegq_s64): Likewise.
7371 2013-10-09 Renlin Li <Renlin.Li@arm.com>
7373 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
7375 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7377 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
7378 packed stack special handling.
7379 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
7380 back to fixed stack slots for FPRs saved due to stdarg.
7382 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7384 * config/s390/s390.c (s390_frame_info): Restructure function.
7386 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7388 * config/s390/s390.c (struct s390_frame_layout): New field
7390 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
7391 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
7392 regs_ever_clobbered to char*.
7393 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
7394 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
7395 (s390_register_info): Enable FPR save slots. Move/Copy some
7396 functionality into ...
7397 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
7398 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
7400 (s390_frame_info): Do gpr slot allocation here now. stdarg does
7401 not imply a stack frame.
7402 (s390_init_frame_layout): Remove variable clobbered_regs.
7403 (s390_update_register_info): Remove function.
7404 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
7405 cannot be used for register renaming.
7406 (s390_hard_regno_scratch_ok): New function.
7407 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
7408 (s390_initial_elimination_offset): Change offset calculation of
7409 the return address pointer.
7410 (save_gprs): Deal with only r6 being saved from the call-saved regs.
7411 (restore_gprs): Set frame related flag.
7412 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
7413 (s390_emit_prologue): Call s390_register_info instead of
7414 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
7415 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
7416 (s390_optimize_prologue): Call s390_optimize_register_info.
7417 Try to remove also FPR slot save/restore INSNs. Remove frame
7418 related flags from restore INSNs.
7420 2013-10-08 DJ Delorie <dj@redhat.com>
7422 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
7425 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
7426 avoid conflict with the MI use of %c.
7427 * config/rl78/rl78-real.md: change %c to %C throughout.
7428 * config/rl78/rl78-virt.md: Likewise.
7430 2013-10-08 Jan Hubicka <jh@suse.cz>
7432 * config/i386/i386.c (ix86_option_override_internal): Switch
7433 to SSE math for -ffast-math when target ISA supports SSE2.
7435 2013-10-08 Andrew MacLeod <amacleod@redhat.com>
7437 * tree-flow.h: Remove some prototypes.
7438 * tree.h: Remove some protypes, add a couple.
7439 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
7440 using_eh_for_cleanups_p): Add interface routines for front ends.
7441 * tree-eh.h: New file. Add protoptyes.
7442 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
7443 (add_stmt_to_eh_lp_fn): Make static.
7444 (lower_try_finally): Use new using_eh_for_cleanups_p.
7445 * emit-rtl.c: Include tree-eh.h.
7446 * gimple.h: Include tree-eh.h.
7448 2013-10-08 Marc Glisse <marc.glisse@inria.fr>
7450 PR tree-optimization/58480
7451 * tree-vrp.c (infer_nonnull_range): New function.
7452 (infer_value_range): Call infer_nonnull_range.
7454 2013-10-08 Dehao Chen <dehao@google.com>
7456 PR tree-optimization/58619
7457 * tree-inline.c (copy_phis_for_bb): Combine location data
7460 2013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
7463 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
7464 RTX_FRAME_RELATED_P on INSN.
7466 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7468 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
7469 (altivec_expand_vec_perm_const): Call it.
7471 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7473 * config/rs6000/vector.md (mov<mode>): Emit permuted move
7474 sequences for LE VSX loads and stores at expand time.
7475 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
7477 * config/rs6000/rs6000.c (rs6000_const_vec): New.
7478 (rs6000_gen_le_vsx_permute): New.
7479 (rs6000_gen_le_vsx_load): New.
7480 (rs6000_gen_le_vsx_store): New.
7481 (rs6000_gen_le_vsx_move): New.
7482 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
7483 (*vsx_le_perm_load_v4si): New.
7484 (*vsx_le_perm_load_v8hi): New.
7485 (*vsx_le_perm_load_v16qi): New.
7486 (*vsx_le_perm_store_v2di): New.
7487 (*vsx_le_perm_store_v4si): New.
7488 (*vsx_le_perm_store_v8hi): New.
7489 (*vsx_le_perm_store_v16qi): New.
7490 (*vsx_xxpermdi2_le_<mode>): New.
7491 (*vsx_xxpermdi4_le_<mode>): New.
7492 (*vsx_xxpermdi8_le_V8HI): New.
7493 (*vsx_xxpermdi16_le_V16QI): New.
7494 (*vsx_lxvd2x2_le_<mode>): New.
7495 (*vsx_lxvd2x4_le_<mode>): New.
7496 (*vsx_lxvd2x8_le_V8HI): New.
7497 (*vsx_lxvd2x16_le_V16QI): New.
7498 (*vsx_stxvd2x2_le_<mode>): New.
7499 (*vsx_stxvd2x4_le_<mode>): New.
7500 (*vsx_stxvd2x8_le_V8HI): New.
7501 (*vsx_stxvd2x16_le_V16QI): New.
7503 2013-10-07 Renlin Li <Renlin.Li@arm.com>
7505 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
7507 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7509 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
7510 loop to work also for 31bit ABI.
7511 Save the stack pointer for frame_size > 0.
7513 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7515 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
7516 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
7517 constraint letters from expanders.
7518 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
7519 retry count to general_operand.
7520 ("tabort"): Give operand 0 a mode.
7521 ("tabort_1"): Add mode and constraint letter for operand 0.
7522 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
7524 2013-10-04 Jeff Law <law@redhat.com>
7526 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
7528 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
7530 (thread_across_edge): Here. Call it.
7532 2013-10-04 Cary Coutant <ccoutant@google.com>
7534 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
7535 discarding a location list expression (or a piece of one).
7537 2013-10-03 Jan Hubicka <jh@suse.cz>
7539 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
7540 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
7541 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
7543 2013-10-03 Jan Hubicka <jh@suse.cz>
7545 * config/i386/i386.c (ix86_option_override_internal): Do not enable
7546 accumulate-outgoing-args when producing unwind info.
7548 2013-10-03 Wei Mi <wmi@google.com>
7550 * lra-constraints.c (insert_move_for_subreg): New function
7551 extracted from simplify_operand_subreg.
7552 (simplify_operand_subreg): Add reload for paradoxical subreg.
7554 2013-10-03 Rong Xu <xur@google.com>
7556 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
7557 the candidate of builtin_expect such that we should fix the
7558 size/time estimation.
7559 (estimate_function_body_sizes): Do the acutally size/time fix-up
7562 2013-10-03 Rong Xu <xur@google.com>
7564 * predict.c (tree_predict_by_opcode): Get the probability
7565 for builtin_expect from param builtin_expect_probability.
7566 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
7567 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
7568 * doc/invoke.texi: Add documentation for builtin-expect-probability.
7570 2013-10-03 Marc Glisse <marc.glisse@inria.fr>
7573 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
7574 * calls.c (alloca_call_p): Use get_callee_fndecl.
7575 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
7576 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
7578 (vrp_visit_stmt): Remove duplicated code.
7580 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
7582 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
7583 ceildf2, btruncdf2, instead of vsx_* name.
7585 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
7586 iterators to only do V2DF and V4SF here. Move the DF code to
7587 rs6000.md where it is combined with SF mode. Replace <VSv> with
7588 just 'v' since only vector operations are handled with these insns
7589 after moving the DF support to rs6000.md.
7590 (vsx_sub<mode>3): Likewise.
7591 (vsx_mul<mode>3): Likewise.
7592 (vsx_div<mode>3): Likewise.
7593 (vsx_fre<mode>2): Likewise.
7594 (vsx_neg<mode>2): Likewise.
7595 (vsx_abs<mode>2): Likewise.
7596 (vsx_nabs<mode>2): Likewise.
7597 (vsx_smax<mode>3): Likewise.
7598 (vsx_smin<mode>3): Likewise.
7599 (vsx_sqrt<mode>2): Likewise.
7600 (vsx_rsqrte<mode>2): Likewise.
7601 (vsx_fms<mode>4): Likewise.
7602 (vsx_nfma<mode>4): Likewise.
7603 (vsx_copysign<mode>3): Likewise.
7604 (vsx_btrunc<mode>2): Likewise.
7605 (vsx_floor<mode>2): Likewise.
7606 (vsx_ceil<mode>2): Likewise.
7607 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
7608 (vsx_sminsf3): Likewise.
7609 (vsx_fmadf4): Likewise.
7610 (vsx_fmsdf4): Likewise.
7611 (vsx_nfmadf4): Likewise.
7612 (vsx_nfmsdf4): Likewise.
7613 (vsx_cmpdf_internal1): Likewise.
7615 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
7616 simpler to select whether a target has SPE or traditional floating
7617 point support in iterators.
7618 (TARGET_DF_SPE): Likewise.
7619 (TARGET_SF_FPR): Likewise.
7620 (TARGET_DF_FPR): Likewise.
7621 (TARGET_SF_INSN): Macros to say whether floating point support
7622 exists for a given operation for expanders.
7623 (TARGET_DF_INSN): Likewise.
7625 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
7626 combining of SF/DF mode operations, using both traditional and VSX
7634 (abs<mode>2): Combine SF/DF modes using traditional floating point
7635 instructions. Add support for using the upper DF registers with
7636 VSX support, and SF registers with power8-vector support. Update
7637 expanders for operations supported by both the SPE and traditional
7638 floating point units.
7639 (abs<mode>2_fpr): Likewise.
7640 (nabs<mode>2): Likewise.
7641 (nabs<mode>2_fpr): Likewise.
7642 (neg<mode>2): Likewise.
7643 (neg<mode>2_fpr): Likewise.
7644 (add<mode>3): Likewise.
7645 (add<mode>3_fpr): Likewise.
7646 (sub<mode>3): Likewise.
7647 (sub<mode>3_fpr): Likewise.
7648 (mul<mode>3): Likewise.
7649 (mul<mode>3_fpr): Likewise.
7650 (div<mode>3): Likewise.
7651 (div<mode>3_fpr): Likewise.
7652 (sqrt<mode>3): Likewise.
7653 (sqrt<mode>3_fpr): Likewise.
7654 (fre<Fs>): Likewise.
7655 (rsqrt<mode>2): Likewise.
7656 (cmp<mode>_fpr): Likewise.
7657 (smax<mode>3): Likewise.
7658 (smin<mode>3): Likewise.
7659 (smax<mode>3_vsx): Likewise.
7660 (smin<mode>3_vsx): Likewise.
7661 (negsf2): Delete SF operations that are merged with DF.
7668 (fmasf4_fpr): Likewise.
7669 (fmssf4_fpr): Likewise.
7670 (nfmasf4_fpr): Likewise.
7671 (nfmssf4_fpr): Likewise.
7672 (sqrtsf2): Likewise.
7673 (rsqrtsf_internal1): Likewise.
7674 (smaxsf3): Likewise.
7675 (sminsf3): Likewise.
7676 (cmpsf_internal1): Likewise.
7677 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
7678 (negdf2): Delete DF operations that are merged with SF.
7680 (nabsdf2): Likewise.
7686 (rsqrtdf_internal1): Likewise.
7687 (fmadf4_fpr): Likewise.
7688 (fmsdf4_fpr): Likewise.
7689 (nfmadf4_fpr): Likewise.
7690 (nfmsdf4_fpr): Likewise.
7691 (sqrtdf2): Likewise.
7692 (smaxdf3): Likewise.
7693 (smindf3): Likewise.
7694 (cmpdf_internal1): Likewise.
7695 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
7696 (btrunc<mode>2): Delete separate expander, and combine with the
7697 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
7698 (btrunc<mode>2_fpr): Likewise.
7699 (ceil<mode>2): Likewise.
7700 (ceil<mode>2_fpr): Likewise.
7701 (floor<mode>2): Likewise.
7702 (floor<mode>2_fpr): Likewise.
7703 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
7704 Add support for using the upper registers with VSX and
7705 power8-vector. Move insns to be closer to the define_expands. On
7706 VSX systems, prefer the traditional form of FMA over the VSX
7707 version, since the traditional form allows the target not to
7708 overlap with the inputs.
7709 (fms<mode>4_fpr): Likewise.
7710 (nfma<mode>4_fpr): Likewise.
7711 (nfms<mode>4_fpr): Likewise.
7713 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7714 Richard Earnshaw <richard.earnshaw@arm.com>
7716 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
7717 (struct mult_cost_table): Likewise.
7718 (struct mem_cost_table): Likewise.
7719 (struct fp_cost_table): Likewise.
7720 (struct vector_cost_table): Likewise.
7721 (cpu_cost_table): Likewise.
7722 * config/arm/arm.opt (mold-rts-costs): New option.
7723 (mnew-generic-costs): Likewise.
7724 * config/arm/arm.c (generic_extra_costs): New table.
7725 (cortexa15_extra_costs): Likewise.
7726 (arm_slowmul_tune): Use NULL as new costs.
7727 (arm_fastmul_tune): Likewise.
7728 (arm_strongarm_tune): Likewise.
7729 (arm_xscale_tune): Likewise.
7730 (arm_9e_tune): Likewise.
7731 (arm_v6t2_tune): Likewise.
7732 (arm_cortex_a5_tune): Likewise.
7733 (arm_cortex_a9_tune): Likewise.
7734 (arm_v6m_tune): Likewise.
7735 (arm_fa726te_tune): Likewise.
7736 (arm_cortex_a15_tune): Use cortex15_extra_costs.
7737 (arm_cortex_tune): Use generict_extra_costs.
7738 (shifter_op_p): New function.
7739 (arm_unspec_cost): Likewise.
7740 (LIBCALL_COST): Define.
7741 (arm_new_rtx_costs): New function.
7742 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
7743 table is available. Use old costs otherwise unless mnew-generic-costs
7745 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
7746 (cpu_cost_table): Declare.
7748 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
7751 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
7752 (*subs_mul_imm_<mode>)
7753 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
7754 (*sub_<shift>_<mode>)
7755 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
7756 Remove k constraint.
7758 2013-10-03 Ian Bolton <ian.bolton@arm.com>
7760 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
7762 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
7764 2013-10-02 Teresa Johnson <tejohnson@google.com>
7766 * predict.c (probably_never_executed): New function.
7767 (probably_never_executed_bb_p): Invoke probably_never_executed.
7768 (probably_never_executed_edge_p): Ditto.
7769 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
7770 Treat profile insanities conservatively.
7772 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
7774 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
7776 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
7778 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
7779 use it. Use smaller increase for scratch. Don't increase reject
7780 for early clobber scratch.
7781 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
7782 setting eliminated regs except setting fp from hfp.
7783 (lra_eliminate): Check lra_insn_recog_data on NULL.
7785 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
7788 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
7789 setting -mvsx-timode by default until the underlying problem is fixed.
7790 (RS6000_CPU, power7 defaults): Likewise.
7792 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
7794 * config/x-linux (host-linux.o): Remove header dependencies.
7795 Use $(COMPILE) and $(POSTCOMPILE).
7796 * config/t-linux-android (linux-android.o): Ditto.
7798 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
7800 * Makefile.in (expmed.o-warn): Remove.
7802 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
7804 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
7805 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
7807 2013-10-02 Teresa Johnson <tejohnson@google.com>
7809 * dojump.c (do_jump_1): Divide probability between
7810 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
7812 2013-10-02 Tom Tromey <tromey@redhat.com>
7814 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
7816 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
7818 * tree-flow.h: Remove some prototypes.
7819 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
7820 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
7821 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
7822 mark_virtual_phi_result_for_renaming): Relocate here.
7823 * tree-into-ssa.h: Add prototypes.
7824 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
7825 single_pred_before_succ_order.
7826 (blocks_in_phiopt_order): Rename and move to cfganal.c.
7827 (nonfreeing_call_p) Move to gimple.c.
7828 * cfganal.c (single_pred_before_succ_order): Move and renamed from
7830 * basic-block.h (single_pred_before_succ_order): Add prototype.
7831 * gimple.c (nonfreeing_call_p): Relocate here.
7832 * gimple.h: Add prototype.
7833 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
7834 * tree-ssa-dom.h: New file. Relocate prototypes here.
7835 * tree-ssa.h: Include tree-ssa-dom.h.
7837 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
7839 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
7840 Use $(COMPILE) and $(POSTCOMPILE).
7842 * config/alpha/x-alpha (driver-alpha.o): Ditto.
7844 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
7846 * tree-flow.h: Remove some prototypes.
7847 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
7848 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
7849 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
7850 propagate_tree_value*): Move from here to...
7851 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
7852 propagate_tree_value*): Relocate here.
7853 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
7854 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
7855 * gimple-fold.c: Remove gimple-fold.h from include list.
7856 * tree-vrp.c: Remove gimple-fold.h from include list.
7857 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
7858 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
7859 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
7860 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
7861 * sese.c: Add tree-ssa-propagate.h to include list.
7863 2013-10-02 Richard Biener <rguenther@suse.de>
7865 * tree-loop-distribution.c: Include tree-vectorizer.h for
7867 (enum partition_kind): Remove PKIND_REDUCTION.
7868 (struct partition_s): Remove has_writes member, add reduction_p member.
7869 (partition_alloc): Adjust.
7870 (partition_builtin_p): Likewise.
7871 (partition_has_writes): Remove.
7872 (partition_reduction_p): New function.
7873 (partition_merge_into): Likewise.
7874 (generate_code_for_partition): Commonize builtin partition
7876 (rdg_cannot_recompute_vertex_p): Remove.
7877 (already_processed_vertex_p): Likewise.
7878 (rdg_flag_vertex): Do not set has_writes.
7879 (classify_partition): Adjust.
7880 (rdg_build_partitions): Do not set has_writes, treat all
7881 partitions as useful.
7882 (distribute_loop): Record number of library calls generated. Adjust.
7883 (tree_loop_distribution): Report number of loops and library
7884 calls generated as opt-info.
7886 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
7888 * tree-flow.h: Include new .h files. Move prototypes.
7889 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
7890 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
7891 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
7892 * tree-pretty-print.h: Add prototypes from tree-flow.h.
7893 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
7894 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
7895 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
7896 * tree.h (get_ref_base_and_extent): Move prototype out.
7897 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
7899 * gimple-low.h: New File. Add prototypes from tree-flow.h.
7900 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
7901 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
7902 * tree-scalar-evolution.c: Include tree.h.
7903 * sese.c: Include tree.h.
7904 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
7905 * dwarf2out.c: Include tree-dfa.h.
7906 * tree-chrec.c: Include tree.h.
7907 * tree-data-ref.c: Include tree.h.
7909 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
7911 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
7914 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7916 * config/t-sol2 (sol2-c.o): Remove header dependencies.
7917 Use $(COMPILE) and $(POSTCOMPILE).
7918 (sol2-cxx.o): Likewise.
7919 (sol2-stubs.o): Likewise.
7921 * config/x-solaris (host-solaris.o): Likewise.
7923 * config/sparc/t-sparc (sparc.o): Remove.
7924 (sparc-c.o): Remove header dependencies.
7925 Use $(COMPILE) and $(POSTCOMPILE).
7926 * config/sparc/x-sparc: Likewise.
7928 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
7930 * config/arc/arc-opts.h: Add 2013 to Copyright years.
7931 * config/arc/arc700.md: Likewise.
7932 * config/arc/arc-modes.def: Likewise.
7933 * config/arc/arc-simd.h: Likewise.
7934 * config/arc/t-arc-uClibc: Likewise.
7935 * config/arc/t-arc-newlib: Likewise.
7937 2013-10-02 Renlin Li <renlin.li@arm.com>
7939 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
7941 (aarch64_expand_epilogue): Likewise.
7943 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7944 Yufeng Zhang <yufeng.zhang@arm.com>
7946 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
7948 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
7949 'base_in' represent a conversion and legal_cast_p_1 holds; set
7950 'base_in' with the returned value from get_unwidened.
7952 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7954 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
7955 plus_constant is not used.
7957 2013-10-01 Wei Mi <wmi@google.com>
7959 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
7960 * config/i386/i386.md: Add define_peephole2 to
7961 break partial reg stall for cvtss2sd/cvtsd2ss.
7963 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
7965 * config/arc/arc.c (pass_arc_ifcvt::clone):
7966 Update for ctxt_ -> m_ctxt change.
7968 2013-10-01 Jeff Law <law@redhat.com>
7970 * tree-ssa-threadupdate.c (struct redirection_data): Delete
7971 outgoing_edge and intermediate_edge fields. Instead store the path.
7972 (redirection_data::hash): Hash on the last edge's destination index.
7973 (redirection_data::equal): Check the entire thread path.
7974 (lookup_redirectio_data): Corresponding changes.
7975 (create_edge_and_update_destination_phis): Likewise.
7976 (thread_single_edge): Likewise.
7978 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
7979 Diego Novillo <dnovillo@google.com>
7981 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
7982 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
7983 (vld32wh_insn, vld32wl_insn): Delete commented-out old
7984 versions of these patterns.
7986 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
7987 (__builtin_arc_aligned): Likewise.
7989 * config/arc/arc.md: Expand adc_0 comment stating the intended
7990 purpose and why it isn't ready.
7991 Replace commented out call_value_via_label_mixed with a
7992 plain comment about bl_s.
7994 * config/arc/arc.c (stdio.h): Don't include directly.
7995 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
7996 Remove if (1) condition.
7997 (arc_encode_section_info): Fix comment.
7999 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
8001 * config/arc/arc.c (arc_conditional_register_usage):
8002 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
8003 Also set reg_alloc_order for DMA config regs.
8005 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
8006 Jeremy Bennett <jeremy.bennett@embecosm.com>
8008 * doc/install.texi (--with-cpu): Mention ARC.
8009 (arc-*-elf32): New paragraph.
8010 (arc-linux-uclibc): Likewise.
8011 * doc/md.texi (Machine Constraints): Add ARC part.
8012 * doc/invoke.texi: (menu): Add ARC Options.
8013 (Machine Dependent Options) <ARC Options>: Add synopsis.
8014 (node ARC Options): Add.
8015 * doc/extend.texi (long_call / short_call attribute): Add ARC.
8016 (ARC Built-in Functions): New section defining
8017 generic ARC built-in functions.
8018 (ARC SIMD Built-in Functions): New section defining SIMD specific
8020 (Declaring Attributes of Functions): Extended
8021 description of short_call and long_call attributes for ARC and
8022 added index entries.
8024 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
8025 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
8026 Joern Rennecke <joern.rennecke@embecosm.com>
8027 Muhammad Khurram Riaz <khurram.riaz@arc.com>
8028 Brendan Kehoe <brendan@zen.org>
8029 Michael Eager <eager@eagercon.com>
8030 Simon Cook <simon.cook@embecosm.com>
8031 Jeremy Bennett <jeremy.bennett@embecosm.com>
8033 * config/arc, common/config/arc: New directories.
8035 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
8036 Brendan Kehoe <brendan@zen.org>
8037 Simon Cook <simon.cook@embecosm.com>
8039 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
8041 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
8043 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
8044 * tree-ssa-coalesce.h: New. Move prototype to here.
8045 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
8046 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
8047 (gimple_can_coalesce_p): Move to...
8048 * gimple.c (gimple_can_coalesce_p): Here.
8050 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
8052 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
8053 (dump_decl_set): Move to gimple.c.
8054 * gimple.h: Don't include tree-ssa-operands.h.
8055 (dump_decl_set): Add prototype.
8056 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
8057 Move to gimple-ssa.h.
8058 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
8059 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
8060 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
8061 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
8062 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
8064 (dump_decl_set): Relocate here.
8065 * gimple-ssa.h: New file.
8066 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
8067 Relocate from gimple.h.
8068 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
8069 * tree-ssa-operands.c (swap_ssa_operands): Rename from
8070 swap_tree_operands and remove non-ssa path.
8071 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
8072 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
8074 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
8075 vect_is_simple_reduction_1): Use swap_ssa_operands.
8076 * tree-flow.h: Move various prototypes to tree-phinodes.h.
8077 (enum need_phi_state): Move to tree-into-ssa.c.
8078 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
8079 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
8080 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
8081 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
8082 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
8083 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
8084 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
8085 num_imm_uses): Move to ssa-iterators.h.
8086 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
8087 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
8089 (op_iter_done, op_iter_next_def, op_iter_next_tree,
8090 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
8091 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
8092 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
8093 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
8094 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
8095 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
8096 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
8097 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
8098 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
8099 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
8100 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
8101 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
8102 (set_phi_nodes): Move to tree-phinodes.h.
8103 * tree-ssa-operands.h (enum ssa_op_iter_type,
8104 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
8105 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
8106 (dump_decl_set): Remove prototype.
8107 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
8108 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
8109 (set_phi_nodes): Relocate from tree-flow-inline.h.
8110 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
8112 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
8113 include list. Temporarily add gimple.h to include list.
8114 * ssa-iterators.h: New file.
8115 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
8116 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
8117 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
8118 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
8119 Relocate from tree-ssa-operands.h.
8120 (delink_imm_use, link_imm_use_to_list, link_imm_use,
8121 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
8122 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
8123 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
8124 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
8125 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
8126 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
8127 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
8128 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
8129 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
8130 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
8131 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
8132 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
8133 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
8134 Relocate from tree-flow-inline.h.
8135 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
8137 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
8140 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
8141 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
8142 Insert '\t' to output template.
8143 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
8144 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
8145 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
8146 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
8148 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
8150 * doc/install.texi (Host/target specific installation notes for GCC):
8151 Put @anchor before @heading.
8152 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
8153 Use @email for email addresses.
8155 2013-10-01 Jeff Law <law@redhat.com>
8157 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
8158 a vec. Only delete the path if we create one without successfully
8159 registering a jump thread.
8160 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
8162 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
8163 (paths): New vector of jump threading paths.
8164 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
8165 (THREAD_PATH): New accessor macro for the entire thread path.
8166 (lookup_redirection_data): Get intermediate and final outgoing edge
8167 from the thread path.
8168 (create_edge_and_update_destination_phis): Copy the threading path.
8169 (ssa_fix_duplicate_block_edges): Get edges and block types from the
8170 jump threading path.
8171 (ssa_redirect_edges): Get edges and block types from the jump threading
8172 path. Free the path vector.
8173 (thread_block): Get edges from the jump threading path. Look at the
8174 entire path to see if we thread to a loop exit. If we cancel a jump
8175 thread request, then free the path vector.
8176 (thread_single_edge): Get edges and block types from the jump threading
8177 path. Free the path vector.
8178 (thread_through_loop_header): Get edges and block types from the jump
8179 threading path. Free the path vector.
8180 (mark_threaded_blocks): Iterate over the vector of paths and store
8181 the path on the appropriate edge. Get edges and block types from the
8182 jump threading path.
8183 (mark_threaded_blocks): Get edges and block types from the jump
8184 threading path. Free the path vector.
8185 (thread_through_all_blocks): Use the vector of paths rather than
8186 a vector of 3-edge sets.
8187 (register_jump_thread): Accept pointer to a path vector rather
8188 than the path vector itself. Store the path vector for later use.
8191 2013-10-01 Jakub Jelinek <jakub@redhat.com>
8192 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8195 * config/s390/s390.c (s390_split_branches): Modify check for table
8197 (s390_chunkify_start): Rearrange table jump insn check in order to
8198 deal with compare and branch insns correctly.
8200 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
8204 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
8205 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
8206 define_insn_and_split.
8207 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
8208 (shiftsi3_compare): New pattern.
8210 * config/arm/unspecs.md (UNSPEC_RRX): New.
8212 2013-10-01 Alan Modra <amodra@gmail.com>
8214 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
8215 casing inout operands.
8217 2013-10-01 Richard Biener <rguenther@suse.de>
8219 PR tree-optimization/58553
8220 * tree-loop-distribution.c (struct partition_s): Add niter member.
8221 (classify_partition): Populate niter member for the partition
8222 and properly identify whether the relevant store happens before
8223 or after the loop exit.
8224 (generate_memset_builtin): Use niter member from the partition.
8225 (generate_memcpy_builtin): Likewise.
8227 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
8229 * vec.h (vec_prefix, vec): Prefix member names with "m_".
8230 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
8232 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
8234 * basic-block.h (edge_list): Prefix member names with "m_".
8235 * context.h (context): Likewise.
8236 * domwalk.h (dom_walker): Likewise.
8237 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
8238 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
8239 * hash-table.h (hash_table): Likewise.
8240 * machmode.h (bit_field_mode_iterator): Likewise.
8241 * pass_manager.h (pass_list): Likewise.
8242 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
8243 * tree-pass.h (pass_data): Likewise.
8244 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
8245 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
8246 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
8247 * asan.c (pass_data_asan): Update accordingly.
8248 * cfganal.c (control_dependences::find_control_dependence): Likewise.
8249 (control_dependences::control_dependences): Likewise.
8250 (control_dependences::~control_dependences): Likewise.
8251 (control_dependences::~control_dependences): Likewise.
8252 (control_dependences::get_edges_dependent_on): Likewise.
8253 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
8254 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
8255 * context.c (gcc::context::context): Likewise.
8256 * cprop.c (pass_rtl_cprop::clone): Likewise.
8257 * domwalk.c (dom_walker::walk): Likewise.
8258 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
8259 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
8260 * mode-switching.c (pass_mode_switching::clone): Likewise.
8261 * passes.c (opt_pass::opt_pass): Likewise.
8262 (pass_manager::pass_manager): Likewise.
8263 * predict.c (pass_strip_predict_hints::clone): Likewise.
8264 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
8265 (pass_split_all_insns::clone): Likewise.
8266 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
8268 (bit_field_mode_iterator::next_mode): Likewise.
8269 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
8270 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
8271 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
8272 * tree-complex.c (pass_lower_complex::clone): Likewise.
8273 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
8274 * tree-object-size.c (pass_object_sizes::clone): Likewise.
8275 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
8276 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
8277 (pass_fold_builtins::clone): Likewise.
8278 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
8279 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
8280 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
8281 (pass_cd_dce::clone): Likewise.
8282 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
8283 (pass_phi_only_cprop::clone): Likewise.
8284 * tree-ssa-dse.c (pass_dse::clone): Likewise.
8285 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
8286 * tree-ssa-loop.c (pass_lim::clone): Likewise.
8287 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
8288 * tree-ssa-pre.c (pass_fre::clone): Likewise.
8289 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
8290 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
8291 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
8292 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
8293 * tree-vrp.c (pass_vrp::clone): Likewise.
8294 * tsan.c (pass_tsan::clone): Likewise.
8296 2013-09-30 Jakub Jelinek <jakub@redhat.com>
8299 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
8300 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
8303 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
8304 optimization, punt if sign_bit_p looked through any zero extension.
8306 2013-09-30 Teresa Johnson <tejohnson@google.com>
8308 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
8309 Update redirected out edge count in joiner case.
8310 (ssa_redirect_edges): Common the joiner and non-joiner cases
8311 so that joiner case gets profile updates.
8313 2013-09-30 Richard Biener <rguenther@suse.de>
8315 PR tree-optimization/58554
8316 * tree-loop-distribution.c (classify_partition): Require
8317 unconditionally executed stores for memcpy and memset recognition.
8318 (tree_loop_distribution): Calculate dominance info.
8320 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
8322 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
8323 (NO_PROFILE_COUNTERS): Likewise.
8324 (PROFILE_HOOK): Likewise.
8325 (FUNCTION_PROFILER): Likewise.
8326 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
8328 2013-09-30 Iain Sandoe <iain@codesourcery.com>
8330 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
8331 calls and defines in TARGET_MACHO conditional.
8332 (load_macho_picbase_di): Likewise.
8333 (reload_macho_picbase): Likewise.
8334 (reload_macho_picbase_si): Likewise.
8335 (reload_macho_picbase_di): Likewise.
8336 (nonlocal_goto_receiver): Likewise.
8338 2013-09-30 Nick Clifton <nickc@redhat.com>
8340 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
8341 that use the MSP430X ISA.
8342 (msp430_option_override): Scan -mmcu command line option for any
8343 MCU name that supports the MSP430X ISA.
8344 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
8345 -mmcu options which enable the MSP430X ISA.
8347 2013-09-30 Richard Biener <rguenther@suse.de>
8350 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
8351 before looking for setjmp-like calls.
8353 2013-09-29 Iain Sandoe <iain@codesourcery.com>
8356 * config/darwin-protos.h (machopic_get_function_picbase): New.
8357 * config/darwin.c (machopic_get_function_picbase): New.
8358 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
8359 label for a new func. (load_macho_picbase_di): Likewise.
8360 (reload_macho_picbase): New expand.
8361 (reload_macho_picbase_si): New insn.
8362 (reload_macho_picbase_di): New insn.
8363 (nonlocal_goto_receiver): New define and split.
8364 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
8365 (unspecv enum): Add UNSPECV_NLGR.
8367 2013-09-29 Iain Sandoe <iain@codesourcery.com>
8369 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
8370 that altivec registers are correctly sized on Darwin.
8372 2013-09-29 Iain Sandoe <iain@codesourcery.com>
8374 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
8375 darwin-driver.o): Use COMPILE and POSTCOMPILE.
8376 * config/x-darwin (host-darwin.o): Likewise.
8377 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
8378 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
8379 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
8381 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
8383 * doc/invoke.texi: Fix usage of @tie{} command.
8385 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
8387 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
8389 2013-09-28 Jan Hubicka <jh@suse.cz>
8391 * config/i386/x86-tune.def: Add documentation for each of the options;
8394 2013-09-28 Jan Hubicka <jh@suse.cz>
8396 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
8398 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
8399 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
8400 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
8401 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
8402 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
8404 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
8406 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
8407 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
8408 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
8409 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
8410 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
8411 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
8412 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
8413 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
8414 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
8415 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
8416 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
8417 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
8418 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
8419 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
8420 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
8421 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
8422 gimple.h, godump.c, graphite-clast-to-gimple.c,
8423 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
8424 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
8425 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
8426 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
8427 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
8428 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
8429 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
8430 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
8431 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
8432 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
8433 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
8434 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
8435 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
8436 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
8437 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
8438 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
8439 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
8440 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
8441 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
8442 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
8443 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
8444 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
8445 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
8446 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
8447 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
8448 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
8449 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
8450 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
8451 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
8452 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
8453 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
8454 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
8455 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
8456 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
8457 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
8458 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
8459 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
8460 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
8461 whitespace before "(".
8463 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
8465 * expr.h (extract_bit_field): Remove packedp parameter.
8466 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
8467 from forward declaration.
8468 (store_split_bit_field): Remove packedp arg from calls to
8469 extract_fixed_bit_field.
8470 (extract_bit_field_1): Remove packedp parameter and packedp
8471 argument from recursive calls and calls to extract_fixed_bit_field.
8472 (extract_bit_field): Remove packedp parameter and corresponding
8473 arg to extract_bit_field_1.
8474 (extract_fixed_bit_field): Remove packedp parameter. Remove code
8476 (extract_split_bit_field): Remove packedp arg from call to
8477 extract_fixed_bit_field.
8478 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
8479 (copy_blkmode_from_reg): Likewise.
8480 (copy_blkmode_to_reg): Likewise.
8481 (read_complex_part): Likewise.
8482 (store_field): Likewise.
8483 (expand_expr_real_1): Likewise.
8484 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
8485 to extract_bit_field.
8486 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
8487 call to extract_bit_field.
8488 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
8489 call to extract_bit_field.
8490 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
8491 and special packedp behavior from -fstrict-volatile-bitfields
8494 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8496 * lra-eliminations.c (init_elim_table): Guard value_p.
8498 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
8500 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
8501 DFmode, DImode, and SFmode in the upper VSX registers based on the
8502 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
8503 if -mpower8-vector. Combine -mvsx-timode handling with the rest
8504 of the VSX register handling.
8506 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
8508 (zero_extendsidi2_lfiwzx): Add support for loading into the
8509 Altivec registers with -mpower8-vector. Use wu/wv constraints to
8510 only do VSX memory options on Altivec registers.
8511 (extendsidi2_lfiwax): Likewise.
8512 (extendsfdf2_fpr): Likewise.
8513 (mov<mode>_hardfloat, SF/SD modes): Likewise.
8514 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
8515 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
8516 (movdi_internal64): Likewise.
8518 2013-09-27 Xinliang David Li <davidxl@google.com>
8520 * opts.c (finish_options): Adjust parameters
8521 according to vect cost model.
8522 (common_handle_option): Set dynamic vect cost
8524 targhooks.c (default_add_stmt_cost): Compute stmt cost
8526 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
8527 Use helper function.
8528 * tree-vectorizer.h (unlimited_cost_model): New function.
8529 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
8530 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
8532 (vect_enhance_data_refs_alignment): Ditto.
8533 * flag-types.h: New enum.
8534 * common/config/i386/i386-common.c (ix86_option_init_struct):
8535 No need to initialize vect_cost_model flag.
8536 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
8539 2013-09-27 Diego Novillo <dnovillo@google.com>
8541 * gimple.h (enum ssa_mode): Remove.
8543 2013-09-27 Paulo Matos <pmatos@broadcom.com>
8545 * cfgloop.h (number_of_loops): Fix typo in check for null.
8547 2013-09-27 Jakub Jelinek <jakub@redhat.com>
8550 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
8551 are children of outermost saved_cfun's loop, and set it up to
8552 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
8553 if loop != loop0 and SESE region contains bbs that belong to loop0.
8555 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
8557 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
8558 (binary_scale_code_p, get_base_term, get_index_term): New functions.
8559 (set_address_segment, set_address_base, set_address_index)
8560 (set_address_disp): Accept the argument unconditionally.
8561 (baseness): Remove must_be_base_p and must_be_index_p checks.
8562 (decompose_normal_address): Classify as much as possible in the
8565 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
8567 * cse.c (count_reg_usage): Handle INT_LIST.
8568 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
8569 * reginfo.c (reg_scan_mark_refs): Likewise.
8570 * reload1.c (eliminate_regs_1): Likewise.
8572 2013-09-27 Iain Sandoe <iain@codesourcery.com>
8575 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
8578 2013-09-27 Richard Biener <rguenther@suse.de>
8580 PR tree-optimization/58459
8581 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
8582 restriction not propagating into loops.
8584 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
8586 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
8587 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
8588 * doc/tree-ssa.texi (Walking use-def chains): Delete.
8590 2013-09-26 Richard Biener <rguenther@suse.de>
8592 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
8594 2013-09-26 Richard Biener <rguenther@suse.de>
8596 * alias.h (component_uses_parent_alias_set): Rename to ...
8597 (component_uses_parent_alias_set_from): ... this.
8598 * alias.c (component_uses_parent_alias_set): Rename to ...
8599 (component_uses_parent_alias_set_from): ... this and return
8601 (reference_alias_ptr_type_1): Use the result from
8602 component_uses_parent_alias_set_from instead of stripping
8603 components one at a time.
8604 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
8606 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
8608 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
8609 Move prototypes to...
8610 * tree-ssa-ter.h: New File. Move prototypes here.
8611 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
8612 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
8613 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
8614 from is_replaceable_p.
8615 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
8616 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
8617 (process_replaceable): Use ter_is_replaceable_p.
8618 (find_replaceable_in_bb): Use ter_is_replaceable_p.
8619 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
8620 newly refactored ssa_is_replaceable_p.
8621 * cfgexpand.c: Include tree-outof-ssa.h.
8622 * ssaexpand.h: Delete.
8624 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
8626 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
8627 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
8629 (create_gimple_tmp): Delete.
8630 (get_expr_type, build_assign, build_type_cast): Move to...
8631 * gimple-builder.c: New File.
8632 (get_expr_type): Relocate from gimple.c.
8633 (build_assign, build_type_cast): Change to only create ssanames.
8634 * gimple.h: Move prototypes to...
8635 * gimple-builder.h: New File. Here.
8636 * tree-ssa.h: And here.
8637 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
8638 count_uses_and_derefs): Relocate from gimple.c.
8639 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
8640 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
8641 * tree-ssa-math-opts (execute_cse_reciprocals): Use
8642 gimple_replace_ssa_lhs.
8643 * asan.c: Include gimple-builder.h.
8644 * Makefile.in: Add gimple-builder.o.
8646 2013-09-26 Richard Biener <rguenther@suse.de>
8648 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
8649 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
8650 (loe_visit_block): Use gcc_checking_assert.
8651 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
8652 gimple_assign_ssa_name_copy_p.
8653 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
8656 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
8658 * config/rs6000/t-rs6000 (rs6000.o): Remove.
8659 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
8661 2013-09-26 Richard Biener <rguenther@suse.de>
8663 PR tree-optimization/58539
8664 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
8665 the fact that debug statements are not taking part in loop-closed
8668 2013-09-26 Nick Clifton <nickc@redhat.com>
8670 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
8671 time warning message.
8672 (msp430_print_operand_raw): Delete unused letter parameter.
8673 (TARGET_PRINT_OPERAND_ADDRESS): Define.
8674 (msp430_print_operand_address): New function.
8675 (msp430_print_operand): Move address printing code from here to
8677 * config/msp430/msp430.md (movsipsi2): Add comment in generated
8679 (zero_extendpsisi2): Likewise.
8680 (extendpsisi2): New pattern.
8681 (andneghi3): New pattern.
8683 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
8685 * config/aarch64/aarch64.opt (mlra): New option.
8686 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
8687 (TARGET_LRA_P): Define.
8689 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
8691 * expr.c (expand_assignment): Remove obsolete comment.
8693 2013-09-25 Jeff Law <law@redhat.com>
8695 * tree-flow.h (thread_through_all_blocks): Prototype moved into
8696 tree-ssa-threadupdate.h.
8697 (register_jump_thread): Similarly.
8698 * tree-ssa-threadupdate.h: New header file.
8699 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
8700 * tree-vrp.c: Likewise.
8701 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
8702 (thread_around_empty_blocks): Change type of path vector argument to
8703 an edge,type pair from just an edge. Initialize both elements when
8704 appending to a jump threading path. Tweak references to elements
8706 (thread_across_edge): Similarly. Release memory for the elements
8708 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
8709 (dump_jump_thread_path): New function broken out from
8710 register_jump_thread.
8711 (register_jump_thread): Use dump_jump_thread_path. Change type of
8712 path vector entries. Search the path for NULL edges and dump
8713 the path if one is found. Tweak the conversion of path to 3-edge
8714 form to use the block copy type information embedded in the path.
8716 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
8718 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
8720 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
8721 Vladimir Makarov <vmakarov@redhat.com>
8723 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
8724 from the least significant bit.
8725 (strip_address_mutations): Add bitfield operations handling.
8726 (must_be_index_p): Add shifting and rotate operations handling.
8727 (set_address_base): Use must_be_base_p predicate.
8728 (set_address_index): Use must_be_index_p predicate.
8730 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
8731 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8732 Sergey Lega <sergey.s.lega@intel.com>
8733 Anna Tikhonova <anna.tikhonova@intel.com>
8734 Ilya Tocar <ilya.tocar@intel.com>
8735 Andrey Turetskiy <andrey.turetskiy@intel.com>
8736 Ilya Verbin <ilya.verbin@intel.com>
8737 Kirill Yukhin <kirill.yukhin@intel.com>
8738 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8740 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
8742 (ix86_expand_vector_move_misalign): Support new unaligned load and
8743 stores and use new names.
8744 (CODE_FOR_sse2_storedqu): Rename to ...
8745 (CODE_FOR_sse2_storedquv16qi): ... this.
8746 (CODE_FOR_sse2_loaddqu): Rename to ...
8747 (CODE_FOR_sse2_loaddquv16qi): ... this.
8748 (CODE_FOR_avx_loaddqu256): Rename to ...
8749 (CODE_FOR_avx_loaddquv32qi): ... this.
8750 (CODE_FOR_avx_storedqu256): Rename to ...
8751 (CODE_FOR_avx_storedquv32qi): ... this.
8752 * config/i386/i386.md (fpint_logic): New.
8753 * config/i386/sse.md (VMOVE): Extend for AVX512.
8757 (VI_UNALIGNED_LOADSTORE): Ditto.
8758 (sse2_avx_avx512f): Ditto.
8759 (sse2_avx2): Extend for AVX512.
8760 (sse4_1_avx2): Ditto.
8761 (avx2_avx512f): New.
8762 (sse): Extend for AVX512.
8765 (avxsizesuffix): Ditto.
8766 (sseintvecmode): Ditto.
8768 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
8769 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
8770 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
8771 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
8772 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
8773 (<sse2_avx_avx512f>_storedqu<mode): ... this.
8774 (<sse>_movnt<mode>): Replace constraint "x" with "v".
8775 (STORENT_MODE): Extend for AVX512.
8776 (*absneg<mode>2): Replace constraint "x" with "v".
8777 (*mul<mode>3): Ditto.
8778 (*ieee_smin<mode>3): Ditto.
8779 (*ieee_smax<mode>3): Ditto.
8780 (avx_cmp<mode>3): Replace VF with VF_128_256.
8781 (*<sse>_maskcmp<mode>3_comm): Ditto.
8782 (<sse>_maskcmp<mode>3): Ditto.
8783 (<sse>_andnot<mode>3): Extend for AVX512.
8784 (<code><mode>3, anylogic): Replace VF with VF_128_256.
8785 (<code><mode>3, fpint_logic): New.
8786 (*<code><mode>3): Extend for AVX512.
8787 (avx512flogicsuff): New.
8788 (avx512f_<logic><mode>): Ditto.
8789 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
8791 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
8792 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
8793 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
8794 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
8795 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
8796 (xop_vpermil2<mode>3): Ditto.
8797 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
8798 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
8799 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
8800 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
8802 2013-09-25 Tom Tromey <tromey@redhat.com>
8804 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
8805 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
8806 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
8807 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
8808 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
8809 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
8810 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
8811 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
8812 (GRAPHITE_HTAB_H): Remove.
8814 2013-09-25 Tom Tromey <tromey@redhat.com>
8816 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
8818 2013-09-25 Tom Tromey <tromey@redhat.com>
8820 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
8822 2013-09-25 Tom Tromey <tromey@redhat.com>
8824 * config/i386/t-i386 (i386.o): Remove.
8825 (i386-c.o): Use COMPILE and POSTCOMPILE.
8827 2013-09-25 Tom Tromey <tromey@redhat.com>
8829 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
8831 2013-09-25 Tom Tromey <tromey@redhat.com>
8833 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
8834 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
8835 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
8836 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
8837 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
8838 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
8839 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
8840 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
8841 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
8842 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
8843 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
8844 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
8845 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
8846 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
8847 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
8848 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
8849 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
8850 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
8851 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
8852 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
8853 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
8854 (tree-ssa-pre.o, tree-ssa-sccvn.o)
8855 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
8856 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
8857 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
8858 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
8859 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
8860 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
8861 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
8862 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
8863 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
8864 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
8865 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
8866 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
8867 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
8868 (graphite-clast-to-gimple.o, graphite-dependences.o)
8869 (graphite-interchange.o, graphite-poly.o)
8870 (graphite-scop-detection.o, graphite-sese-to-poly.o)
8871 (graphite-optimize-isl.o, tree-vect-loop.o)
8872 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
8873 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
8874 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
8875 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
8876 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
8877 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
8878 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
8879 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
8880 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
8881 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
8882 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
8883 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
8884 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
8885 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
8886 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
8887 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
8888 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
8889 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
8890 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
8891 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
8892 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
8893 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
8894 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
8895 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
8896 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
8897 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
8898 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
8899 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
8900 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
8901 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
8902 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
8903 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
8904 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
8905 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
8906 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
8907 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
8908 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
8909 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
8910 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
8911 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
8912 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
8913 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
8914 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
8915 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
8916 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
8917 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
8918 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
8919 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
8920 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
8921 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
8922 (gcov-dump.o): Remove.
8923 (default-c.o): Use COMPILE and POSTCOMPILE.
8924 (CFLAGS-gcc.o): New variable.
8925 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
8927 2013-09-25 Tom Tromey <tromey@redhat.com>
8929 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
8930 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
8931 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
8932 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
8933 (c-family/c-pragma.o, c-family/c-pretty-print.o)
8934 (c-family/c-semantics.o, c-family/c-ada-spec.o)
8935 (c-family/array-notation-common.o, c-family/stub-objc.o)
8936 (c-family/c-ubsan.o): Remove.
8938 2013-09-25 Tom Tromey <tromey@redhat.com>
8940 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
8942 2013-09-25 Tom Tromey <tromey@redhat.com>
8944 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
8945 to add -DENABLE_SHARED_LIBGCC.
8946 (gcc.o): Don't use subshell.
8948 2013-09-25 Tom Tromey <tromey@redhat.com>
8950 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
8951 * configure.ac: Don't invoke AM_PROG_CC_C_O.
8952 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
8953 * configure, config.in: Rebuild.
8955 2013-09-25 Tom Tromey <tromey@redhat.com>
8957 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
8958 (COMPILE, POSTCOMPILE): New variables.
8959 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
8960 (DEPFILES): New variable.
8961 Include ".Po" files.
8962 * configure.ac: Add checks for dependency checking.
8963 * configure, aclocal.m4: Regenerate.
8965 2013-09-25 Tom Tromey <tromey@redhat.com>
8967 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
8968 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
8970 2013-09-25 Tom Tromey <tromey@redhat.com>
8972 * Makefile.in (generated_files): Add options.h,
8973 target-hooks-def.h, insn-opinit.h,
8974 common/common-target-hooks-def.h, pass-instances.def,
8975 c-family/c-target-hooks-def.h.
8977 2013-09-25 Jeff Law <law@redhat.com>
8979 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
8980 than foo[foo.length () - 1] to access last member in a vec.
8981 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
8983 2013-09-25 Richard Biener <rguenther@suse.de>
8986 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
8988 2013-09-25 Jan Hubicka <jh@suse.cz>
8990 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
8993 2013-09-25 Marek Polacek <polacek@redhat.com>
8996 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
8997 when determining the type name.
8999 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
9001 * config/sh/sh.md: Fix formatting.
9003 2013-09-24 Xinliang David Li <davidxl@google.com>
9005 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
9006 max peel iterations parameter.
9007 * param.def: New parameter.
9008 * doc/invoke.texi: Document New parameter.
9010 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
9012 * gimple-pretty-print.c: Various whitespace tweaks.
9013 * tree-core.h: Likewise.
9014 * tree-pretty-print.c: Likewise.
9015 * tree-ssa-alias.c: Likewise.
9016 * tree-ssa-copy.c: Likewise.
9017 * tree-ssanames.c: Likewise.
9018 * tree-ssanames.h: Likewise.
9019 * tree-vrp.c: Likewise.
9021 2013-09-24 Alan Modra <amodra@gmail.com>
9025 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
9026 for output operands that disallow regs. Don't use EXPAND_WRITE on
9029 2013-09-24 Richard Biener <rguenther@suse.de>
9032 * tree.c (reference_alias_ptr_type): Move ...
9033 * alias.c (reference_alias_ptr_type): ... here and implement
9034 in terms of the new reference_alias_ptr_type_1.
9035 (ref_all_alias_ptr_type_p): New helper.
9036 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
9037 use ref_all_alias_ptr_type_p.
9038 (get_deref_alias_set): Add flag_strict_aliasing check here.
9039 (reference_alias_ptr_type_1): New function, split out from ...
9040 (get_alias_set): ... here.
9041 (alias_ptr_types_compatible_p): New function.
9042 * alias.h (reference_alias_ptr_type): Declare.
9043 (alias_ptr_types_compatible_p): Likewise.
9044 * tree.h (reference_alias_ptr_type): Remove.
9045 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
9046 to compare MEM_REF alias types.
9048 2013-09-24 Richard Biener <rguenther@suse.de>
9050 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
9052 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
9054 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
9055 reload helper function arrays into a single array reg_addr.
9056 (reload_fpr_gpr): Likewise.
9057 (reload_gpr_vsx): Likewise.
9058 (reload_vsx_gpr): Likewise.
9059 (struct rs6000_reg_addr): Likewise.
9060 (reg_addr): Likewise.
9061 (rs6000_debug_reg_global): Change rs6000_vector_reload,
9062 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
9063 (rs6000_init_hard_regno_mode_ok): Likewise.
9064 (rs6000_secondary_reload_direct_move): Likewise.
9065 (rs6000_secondary_reload): Likewise.
9067 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
9068 constraints: wu, ww, and wy. Repurpose wv constraint added during
9069 power8 changes. Put wg constraint in alphabetical order.
9071 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
9072 for future work to add ISA 2.07 VSX single precision support.
9073 (-mvsx-scalar-double): Change default from -1 to 1, update
9074 documentation comment.
9075 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
9076 (-mupper-regs-df): New debug switch to control whether DF values
9077 can go in the traditional Altivec registers.
9078 (-mupper-regs-sf): New debug switch to control whether SF values
9079 can go in the traditional Altivec registers.
9081 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
9083 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
9084 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
9085 Add new constraints, wu/ww/wy. Repurpose wv constraint.
9086 (rs6000_debug_legitimate_address_p): Print if we are running
9087 before, during, or after reload.
9088 (rs6000_secondary_reload): Add a comment.
9089 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
9091 * config/rs6000/constraints.md (wa constraint): Sort w<x>
9092 constraints. Update documentation string.
9093 (wd constraint): Likewise.
9094 (wf constraint): Likewise.
9095 (wg constraint): Likewise.
9096 (wn constraint): Likewise.
9097 (ws constraint): Likewise.
9098 (wt constraint): Likewise.
9099 (wx constraint): Likewise.
9100 (wz constraint): Likewise.
9101 (wu constraint): New constraint for ISA 2.07 SFmode scalar
9103 (ww constraint): Likewise.
9104 (wy constraint): Likewise.
9105 (wv constraint): Repurpose ISA 2.07 constraint that we did not use
9106 in the previous submissions.
9107 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
9109 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
9111 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
9112 (REG_BR_PROB): Say that the probability is stored in an int_list.
9113 * reg-notes.def: Update commentary to mention INT_LIST.
9114 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
9115 (INT_LIST): New rtx.
9116 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
9117 * rtlanal.c (int_reg_note_p): New function.
9118 (alloc_reg_note): Assert that the note does not have an int argument.
9119 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
9120 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
9121 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
9122 rather than an INSN_LIST. Handle INT_LIST.
9123 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
9124 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
9125 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
9126 Manipulate them as ints rather than rtxes.
9127 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
9128 * regmove.c (copy_src_to_dest): Likewise.
9129 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
9131 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
9132 into the cases that need it.
9133 * config/arm/arm.c (arm_unwind_emit): Likewise.
9135 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
9136 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
9137 * loop-doloop.c (add_test, doloop_modify): Likewise.
9138 * loop-unswitch.c (compare_and_jump_seq): Likewise.
9139 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
9140 * predict.c (combine_predictions_for_insn): Likewise.
9141 * print-rtl.c (print_rtx): Handle INT_LIST.
9142 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
9143 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
9144 * config/arm/arm.c (emit_unlikely_jump): Likewise.
9145 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
9146 (ix86_split_fp_branch, predict_jump): Likewise.
9147 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
9148 * config/sh/sh.c (expand_cbranchsi4): Likewise.
9149 * config/spu/spu.c (ea_load_store_inline): Likewise.
9151 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
9152 value of a REG_BR_PROB note.
9153 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
9154 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
9155 * emit-rtl.c (try_split): Likewise.
9156 * predict.c (br_prob_note_reliable_p): Likewise.
9157 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
9158 * reorg.c (mostly_true_jump): Likewise.
9159 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
9160 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
9161 * config/i386/i386.c (ix86_print_operand): Likewise.
9162 * config/ia64/ia64.c (ia64_print_operand): Likewise.
9163 * config/mmix/mmix.c (mmix_print_operand): Likewise.
9164 * config/rs6000/rs6000.c (output_cbranch): Likewise.
9165 * config/s390/s390.c (s390_expand_tbegin): Likewise.
9166 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
9167 * config/sparc/sparc.c (output_cbranch): Likewise.
9168 * config/spu/spu.c (get_branch_target): Likewise.
9169 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
9170 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
9172 2013-09-23 Jan Hubicka <jh@suse.cz>
9174 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
9176 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
9177 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
9179 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
9180 (gimple_fold_call): Dump inconsistent devirtualizations; add
9181 sanity check for type based devirtualizations.
9182 * ipa-prop.c: Include ipa-utils.h
9183 (ipa_intraprocedural_devirtualization): Add sanity check.
9184 (try_make_edge_direct_virtual_call): Likewise.
9186 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
9188 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
9189 assignment statements.
9191 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
9193 * gimple-pretty-print.c (dump_ssaname_info): New function.
9194 (dump_gimple_phi): Call it.
9195 (pp_gimple_stmt_1): Likewise.
9196 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
9197 (range_info_def): Declare.
9198 * tree-pretty-print.c (pp_double_int): New function.
9199 (dump_generic_node): Call it.
9200 * tree-pretty-print.h (pp_double_int): Declare.
9201 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
9202 * tree-ssanames.h (range_info_def): New structure.
9203 (value_range_type): Move definition here.
9204 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
9206 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
9208 (set_range_info): New function.
9209 (get_range_info): Likewise.
9210 (duplicate_ssa_name_range_info): Likewise.
9211 (duplicate_ssa_name_fn): Check pointer type and call
9212 duplicate_ssa_name_range_info.
9213 * tree-ssa-copy.c (fini_copy_prop): Likewise.
9214 * tree-vrp.c (value_range_type): Remove definition, now in
9216 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
9217 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
9218 (SSA_NAME_RANGE_INFO): New macro.
9220 2013-09-23 Richard Biener <rguenther@suse.de>
9222 PR tree-optimization/58464
9223 * tree-ssa-pre.c (phi_trans_lookup): Remove.
9224 (phi_trans_add): Change to add conditionally on being not present.
9225 (phi_translate_1): Remove recursion detection here.
9226 (phi_translate): Pre-seed the cache with NULL to catch
9227 recursion here in a more generic way.
9228 (bitmap_find_leader): Adjust comment.
9229 (get_representative_for): Dump value-numbers.
9230 (create_expression_by_pieces): Likewise.
9231 (insert_into_preds_of_block): Likewise.
9233 2013-09-23 Christian Bruel <christian.bruel@st.com>
9236 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
9237 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
9239 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
9242 2013-09-20 Renlin Li <renlin.li@arm.com>
9244 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
9246 (aarch64_expand_epilogue): Likewise.
9247 (aarch64_legitimize_reload_address): Likewise.
9249 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
9251 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
9252 NULL_TREE before pushing them onto the vector. Likewise for labels.
9254 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
9256 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
9257 comparison operators when -fno-trapping-math is in effect.
9258 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
9259 comparison operators in TFmode and assert that unsupported operators
9261 (ia64_print_operand): Likewise.
9263 2013-09-21 Jan Hubicka <jh@suse.cz>
9265 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
9266 (sse_typeless_stores): Enable for core
9267 (sse_load0_by_pxor): Likewise.
9268 (four_jump_limit): Disable for core.
9269 (pad_returns): Likewise.
9270 (avoid_vector_decode): Likewise.
9271 (fuse_cmp_and_branch): Enable for cores.
9272 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
9274 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
9277 * config/pa/pa.c (pa_option_override): Disable auto increment and
9278 decrement instructions until reload is completed.
9280 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
9281 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
9282 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
9284 2013-09-20 DJ Delorie <dj@redhat.com>
9285 Nick Clifton <nickc@redhat.com>
9287 * config/rl78/rl78.c: Various whitespace and comment tweaks.
9288 (need_to_save): Save bank 0 on interrupts.
9289 (characterize_address): Strip far address wrappers.
9290 (rl78_as_legitimate_address): Likewise.
9291 (transcode_memory_rtx): Likewise.
9292 (rl78_peep_movhi_p): Disable this peephole after devirt.
9293 (rl78_propogate_register_origins): Forget all origins when a
9295 * config/rl78/rl78-virt.md: Various whitespace tweaks.
9296 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
9298 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
9299 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
9300 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
9302 * config/rl78/constraints.md: Various whitespace and paren tweaks.
9304 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
9306 * config/pa/pa.md: In "scc" insn patterns, change output template to
9307 handle const0_rtx in reg_or_0_operand operands.
9309 2013-09-20 Martin Husemann <martin@NetBSD.org>
9312 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
9313 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
9315 2013-09-20 Richard Biener <rguenther@suse.de>
9318 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
9319 remembering SSA name version and block index.
9320 (new_scev_info_str): Adjust.
9321 (hash_scev_info): Likewise. Also hash the block index.
9322 (eq_scev_info): Adjust.
9323 (find_var_scev_info): Likewise.
9324 (struct instantiate_cache_entry): Remove.
9325 (struct instantiate_cache_type): Use a htab to map name, block
9327 (instantiate_cache_type::~instantiate_cache_type): Adjust.
9328 (get_instantiated_value_entry): Likewise.
9329 (hash_idx_scev_info, eq_idx_scev_info): New functions.
9330 (instantiate_scev_name): Adjust.
9332 2013-09-20 Jeff Law <law@redhat.com>
9334 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
9336 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
9338 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
9339 Call aarch64_simd_expand_args to update op[argc].
9341 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
9343 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
9346 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
9348 * gengtype.c (file_rules): Added rule for *.cc files.
9349 (get_output_file_with_visibility): Give fatal message when no
9352 2013-09-20 Renlin Li <renlin.li@arm.com>
9354 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
9355 (aarch64_expand_epilogue): Likewise.
9356 (aarch64_legitimize_reload_address): Likewise.
9358 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
9361 * expr.c (expand_assignment): Remove misalignp code path.
9363 2013-09-20 Marek Polacek <polacek@redhat.com>
9366 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
9367 TYPE_PRECISION. Add asserts.
9369 2013-09-20 Richard Biener <rguenther@suse.de>
9371 PR tree-optimization/58453
9372 * tree-loop-distribution.c (distribute_loop): Apply the cost
9373 model for -ftree-loop-distribute-patterns, too.
9375 2013-09-20 Richard Biener <rguenther@suse.de>
9378 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
9379 make type comparison less strict.
9381 2013-09-20 Alan Modra <amodra@gmail.com>
9383 * configure: Regenerate.
9384 * aclocal.m4: Regenerate.
9386 2013-09-20 Marek Polacek <polacek@redhat.com>
9389 * doc/extend.texi: Document that attribute used is meant to be used
9390 on variables with static storage duration.
9392 2013-09-19 Jakub Jelinek <jakub@redhat.com>
9394 PR tree-optimization/58472
9395 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
9396 simd_lane_access set inv_p = false.
9397 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
9398 the simduid magic VAR_DECL.
9400 2013-09-19 Jan Hubicka <jh@suse.cz>
9402 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
9404 2013-09-17 Jeff Law <law@redhat.com>
9406 * tree-ssa-dom.c (record_temporary_equivalences): New function
9407 split out of dom_opt_dom_walker::after_dom_children.
9408 (dom_opt_dom_walker::thread_across_edge): Move common code
9409 in here from dom_opt_dom_walker::after_dom_children.
9410 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
9412 2013-09-19 Jan Hubicka <jh@suse.cz>
9414 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
9415 (TARGET_GENERIC): Use PROCESOR_GENERIC
9416 (enum processor_type): Unify generic32 and 64.
9417 * i386.md (cpu): Likewise.
9418 * x86-tune.def (use_leave): Enable for generic32.
9419 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
9420 * athlon.md: Change generic64 to generic in all occurences.
9421 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
9422 (ix86_target_macros_internal): Likewise.
9423 * driver-i386.c (host_detect_local_cpu): Likewise.
9424 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
9426 (generic_memcpy, generic_memset, generic_cost): This one.
9427 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
9428 (m_GENERIC32, m_GENERIC64): Remove.
9429 (m_GENERIC): Turn into one flag.
9430 (processor_target): Unify generic tunnings.
9431 (ix86_option_override_internal): Replace generic32/64 by generic.
9432 (ix86_issue_rate): Likewise.
9433 (ix86_adjust_cost): Likewise.
9435 2013-09-19 Jan Hubicka <jh@suse.cz>
9437 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
9438 of speculative flag.
9440 2013-09-19 Jakub Jelinek <jakub@redhat.com>
9442 * omp-low.c (expand_omp_sections): Always pass len - 1 to
9443 GOMP_sections_start, even if !exit_reachable.
9445 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
9447 * lra-constraints.c (need_for_all_save_p): Use macro
9448 HARD_REGNO_CALL_PART_CLOBBERED.
9449 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
9450 set up pseudo conflict hard regs.
9452 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
9455 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
9458 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
9460 PR rtl-optimization/58438
9461 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
9462 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
9463 from previous subpasses.
9465 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
9467 * arm.c (arm_get_frame_offsets): Validate architecture supports
9468 LDRD/STRD before accepting the tuning preference.
9469 (arm_expand_prologue): Likewise.
9470 (arm_expand_epilogue): Likewise.
9472 2013-09-18 Richard Biener <rguenther@suse.de>
9474 PR tree-optimization/58417
9475 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
9476 have chrecs with symbols defined in the loop as operands.
9477 (chrec_fold_multiply): Likewise.
9478 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
9479 parameters before folding binary operations.
9480 (struct instantiate_cache_entry_hasher): Remove.
9481 (struct instantiate_cache_type): Use a pointer-map.
9482 (instantiate_cache_type::instantiate_cache_type): New function.
9483 (instantiate_cache_type::get): Likewise.
9484 (instantiate_cache_type::set): Likewise.
9485 (instantiate_cache_type::~instantiate_cache_type): Adjust.
9486 (get_instantiated_value_entry): Likewise.
9487 (global_cache): New global.
9488 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
9489 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
9490 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
9491 (instantiate_scev_name): Adjust.
9492 (instantiate_scev): Construct global instead of local cache.
9493 (resolve_mixers): Likewise.
9495 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
9496 Paolo Carlini <paolo.carlini@oracle.com>
9499 * doc/implement-cxx.texi: Fix references to the C++ standards.
9501 2013-09-18 Jakub Jelinek <jakub@redhat.com>
9503 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
9504 * tree-vect-data-refs.c (vect_analyze_data_refs): For
9505 simd_lane_access drs, update also DR_ALIGNED_TO.
9507 2013-09-18 Marek Polacek <polacek@redhat.com>
9510 * doc/extend.texi: Document no_sanitize_undefined attribute.
9511 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
9512 no_sanitize_undefined attribute.
9514 2013-09-18 Nick Clifton <nickc@redhat.com>
9516 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
9517 (ASM_DECLARE_FUNCTION_NAME): Define.
9519 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
9521 * compare-elim.c (find_comparison_dom_walker): New class
9522 (find_comparisons_in_bb): Rename to
9523 find_comparison_dom_walker::before_dom_children
9524 (find_comparisons): Adjust
9525 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
9527 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
9528 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
9529 (init_walk_dominator_tree): Remove declaration.
9530 (fini_walk_dominator_tree): Remove declaration.
9531 * fwprop.c (single_def_use_dom_walker): New class
9532 (single_def_use_enter_block): Convert to
9533 single_def_use_dom_walker::before_dom_children.
9534 (single_def_use_leave_block): Convert to
9535 single_def_use_dom_walker::after_dom_children.
9536 (build_single_def_use_links): Adjust.
9537 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
9539 (find_candidates_in_block): Convert to
9540 find_candidates_dom_walker::before_dom_children.
9541 (execute_strength_reduction): Adjust.
9542 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
9543 (sese_dom_walker): New class.
9544 (sese_dom_walker::sese_dom_walker): New constructor.
9545 (sese_dom_walker::~sese_dom_walker): New destructor.
9546 (build_sese_conditions_before): Convert to
9547 sese_dom_walker::before_dom_children.
9548 (build_sese_conditions_after): Convert to
9549 sese_dom_walker::after_dom_children.
9550 (build_poly_scop): Adjust
9551 * tree-into-ssa.c (rewrite_dom_walker): New class
9552 (rewrite_enter_block): Convert to
9553 rewrite_dom_walker::before_dom_children.
9554 (rewrite_leave_block): Convert to
9555 rewrite_dom_walker::after_dom_children.
9556 (rewrite_update_dom_walker): New class.
9557 (rewrite_update_enter_block): Convert to
9558 rewrite_update_dom_walker::before_dom_children.
9559 (rewrite_update_leave_block): Convert to
9560 rewrite_update_dom_walker::after_dom_children.
9561 (rewrite_blocks, rewrite_into_ssa): Adjust.
9562 (mark_def_dom_walker): New class.
9563 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
9564 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
9565 (mark_def_sites_blocks): Convert to
9566 mark_def_dom_walker::before_dom_children.
9567 (mark_def_site_blocks): Remove.
9568 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
9569 (tree_ssa_dominator_optimize): Adjust.
9570 (dom_thread_across_edge): Convert to method
9571 dom_opt_dom_walker::thread_across_edge.
9572 (dom_opt_enter_block): Convert to member function
9573 dom_opt_dom_walker::before_dom_children.
9574 (dom_opt_leave_block): Convert to member function
9575 dom_opt_dom_walker::after_dom_children.
9576 * tree-ssa-dse.c (dse_dom_walker): New class.
9577 (dse_enter_block): Convert to member function
9578 dse_dom_walker::before_dom_children.
9579 (tree_ssa_dse): Adjust.
9580 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
9581 (determine_invariantness_stmt): Convert to method
9582 invariantness_dom_walker::before_dom_children.
9583 (determine_invariantness): Remove
9584 (move_computations_dom_walker): New class.
9585 (move_computations_stmt): Convert to method
9586 move_computations_dom_walker::before_dom_children.
9587 (move_computations, tree_ssa_lim): Adjust.
9588 * tree-ssa-phiopt.c (nontrapping_dom_walker): New class.
9589 (nt_init_block): Convert to method
9590 notrappping_dom_walker::before_dom_children.
9591 (nt_fini_block): Convert to method
9592 method nontrapping_dom_walker::after_dom_children.
9593 (get_non_trapping): Adjust.
9594 * tree-ssa-pre.c (eliminate_dom_walker): New class.
9595 (eliminate_bb): Convert to method
9596 eliminate_dom_walker::before_dom_children.
9597 (eliminate_leave_block): Convert to method
9598 eliminate_dom_walker::after_dom_children.
9599 (eliminate): Adjust.
9600 * tree-ssa-strlen.c (strlen_dom_walker): New class.
9601 (strlen_enter_block): Convert to method
9602 strlen_dom_walker::before_dom_children.
9603 (strlen_leave_block): Convert to method
9604 method strlen_dom_walker::after_dom_children.
9605 (tree_ssa_strlen): Adjust.
9606 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
9607 (tree_ssa_uncprop): Adjust.
9608 (uncprop_leave_block): Convert to method
9609 uncprop_dom_walker::after_dom_children.
9610 (uncprop_leave_block): Convert to method
9611 uncprop_dom_walker::before_dom_children.
9613 2013-09-18 Bin Cheng <bin.cheng@arm.com>
9615 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
9616 branch in same basic block. Check both src and dest of the move insn.
9618 2013-09-17 Nick Clifton <nickc@redhat.com>
9620 * config/rl78/rl78-real.md (bf): New pattern.
9622 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
9623 (rl78_print_operand): Do not put a # before a %B.
9624 * config/rl78/rl78.opt: Tweak doc strings.
9626 2013-09-17 DJ Delorie <dj@redhat.com>
9628 * config/rl78/constraints.md (Wcv): Allow up to $r31.
9629 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
9630 (rl78_option_override): Likewise, if -mallregs.
9631 (is_virtual_register): Likewise.
9632 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
9633 (REGNO_OK_FOR_BASE_P): Likewise.
9634 * config/rl78/rl78.opt (-mallregs): New.
9636 2013-09-17 Nick Clifton <nickc@redhat.com>
9638 * config/rl78/rl78.c (need_to_save): Change return type to bool.
9639 For interrupt functions: save all call clobbered registers if the
9640 interrupt handler is not a leaf function.
9641 (rl78_expand_prologue): Always recompute the frame information.
9642 For interrupt functions: only select bank 0 if one of the bank 0
9643 registers is going to be psuhed.
9645 2013-09-17 DJ Delorie <dj@redhat.com>
9647 * config/rl78/constraints.md: For each W* constraint, rename to C*
9648 and create a W* constraint that checks for an optional ES: prefix
9650 * config/rl78/rl78.md (UNS_ES_ADDR): New.
9651 (es_addr): New. Used to wrap far addresses.
9652 * config/rl78/rl78-protos.h (rl78_es_addr): New.
9653 (rl78_es_base): New.
9654 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
9655 wrapped far addresses.
9656 (rl78_print_operand_1): Unwrap far addresses before processing.
9657 (rl78_lo16): Wrap far addresses in unspecs.
9658 (rl78_es_addr): New.
9659 (rl78_es_base): New.
9660 (insn_ok_now): Check for not-yet-wrapped far addresses.
9661 (transcode_memory_rtx): Properly re-wrap far addresses.
9663 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
9665 * config/sparc/t-rtems: Add leon3 multilibs.
9667 2013-09-17 Cong Hou <congh@google.com>
9669 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
9670 when checking the dot production pattern. The type of rhs operand
9671 of multiply is now checked correctly.
9673 2013-09-17 Jeff Law <law@redhat.com>
9675 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
9676 edge implied equivalences into successor phis.
9677 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
9678 here from tree-ssa-threadedge.c.
9679 (mark_threaded_blocks): When threading through a joiner, if both
9680 successors of the joiner's clone reach the same block, verify the
9681 PHI arguments are equal. If not, cancel the jump threading request.
9682 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
9683 tree-ssa-threadupdate.c
9684 (thread_across_edge): Don't check PHI argument equality when
9685 threading through joiner block here.
9687 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
9689 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
9690 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
9691 (warn_uninit, warn_uninitialized_vars,
9692 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
9693 Move to tree-ssa-uninit.c.
9694 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
9695 (has_undefined_value_p): New. Pass dependant parts of
9696 ssa_undefined_value_p.
9697 (uninit_undefined_value_p): Use has_undefined_value_p.
9698 (warn_uninit, warn_uninitialized_vars,
9699 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
9700 Move from tree-ssa.c.
9701 * tree-ssa.h: Adjust prototypes.
9703 2013-09-17 Jan Hubicka <jh@suse.cz>
9706 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
9707 * ipa-inline.c (can_inline_edge_p): Do not downgrade
9708 FUNCTION_NOT_OPTIMIZED.
9709 * ipa-inline-analysis.c (compute_inline_parameters): Function
9710 not optimized is not inlinable unless it is alwaysinline.
9711 (inline_analyze_function): Force calls in not optimized
9712 function not inlinable.
9714 2013-09-17 Jan Hubicka <jh@suse.cz>
9717 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
9719 * ipa.c (function_and_variable_visibility): Likewise.
9720 * ipa-profile.c (ipa_profile): Likewise.
9722 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
9725 * cgraph.c (cgraph_function_body_availability): Check for ifunc
9726 attribute, and don't inline the resolver in this case.
9728 2013-09-17 Teresa Johnson <tejohnson@google.com>
9730 * coverage.c (get_coverage_counts): Add missing newline.
9732 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9734 PR tree-optimization/58088
9735 * fold-const.c (mask_with_trailing_zeros): New function.
9736 (fold_binary_loc): Make sure we don't recurse infinitely
9737 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
9738 Use mask_with_trailing_zeros where appropriate.
9740 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
9742 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
9743 of 'prev' var to get better distance estimation.
9745 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
9747 * tree-inline.h (struct copy_body_data): Add transform_parameter.
9748 * tree-inline.c (is_parameter_of): New predicate.
9749 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
9750 a parameter has been remapped.
9751 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
9752 (optimize_inline_calls): Initialize transform_parameter.
9753 (copy_gimple_seq_and_replace_locals): Likewise.
9754 (tree_function_versioning): Likewise.
9755 (maybe_inline_call_in_expr): Likewise.
9757 2013-09-17 Nick Clifton <nickc@redhat.com>
9759 * config/msp430/msp430-protos.h: Add prototypes for new functions.
9760 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
9762 (is_attr_func): New function.
9763 (msp430_is_interrupt_func): New function.
9764 (is_naked_func): New function.
9765 (is_reentrant_func): New function.
9766 (is_critical_func): New function.
9767 (msp430_start_function): Add annotations for function attributes.
9768 (msp430_attr): New function.
9769 (msp430_attribute_table): New.
9770 (msp430_function_section): New function.
9771 (TARGET_ASM_FUNCTION_SECTION): Define.
9772 (msp430_builtin): New enum.
9773 (msp430_init_builtins): New function.
9774 (msp430_builtin_devl): New function.
9775 (msp430_expand_builtin): New function.
9776 (TARGET_INIT_BUILTINS): Define.
9777 (TARGET_EXPAND_BUILTINS): Define.
9778 (TARGET_BUILTIN_DECL): Define.
9779 (msp430_expand_prologue): Add support for naked, interrupt,
9780 critical and reentrant functions.
9781 (msp430_expand_epilogue): Likewise.
9782 (msp430_print_operand): Handle 'O' character.
9783 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
9785 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
9786 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
9787 (pushm): Use a 'n' rather than an 'i' constraint.
9788 (msp_return): Add generation of the interrupt return instruction.
9789 (disable_interrupts): New pattern.
9790 (enable_interrupts): New pattern.
9791 (push_intr_state): New pattern.
9792 (pop_intr_state): New pattern.
9793 (bic_SR): New pattern.
9794 (bis_SR): New pattern.
9795 * doc/extend.texi: Document MSP430 function attributes and builtin
9798 2013-09-17 Richard Biener <rguenther@suse.de>
9800 PR tree-optimization/58432
9801 * tree-loop-distribution.c (tree_loop_distribution): Also
9802 scan PHIs for outside loop uses and seed a partition from them.
9804 2013-09-17 Bin Cheng <bin.cheng@arm.com>
9806 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
9807 (restructure_reference): Call backtrace_base_for_ref.
9809 2013-09-17 Alan Modra <amodra@gmail.com>
9812 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
9815 2013-09-16 DJ Delorie <dj@redhat.com>
9817 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
9818 vregs location for RL78/G10.
9819 (rl78_expand_prologue): Avoid SEL on G10.
9820 (rl78_expand_epilogue): Likewise.
9821 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
9822 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
9823 __RL78_G10__ when appropriate.
9824 (ASM_SPEC): Pass -mg10 along to the assembler.
9825 * config/rl78/rl78.md (sel_rb): Disable for G10.
9826 * config/rl78/rl78.opt: Add -mg10 option.
9827 * config/rl78/t-rl78: Add -mg10 multilib.
9829 2013-09-16 Xinliang David Li <davidxl@google.com>
9831 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
9832 * omp-low.c (omp_max_vf): Ditto.
9833 (expand_omp_simd): Ditto.
9834 * tree-vectorizer.c (vectorize_loops): Ditto.
9835 (gate_vect_slp): Ditto.
9836 (gate_increase_alignment): Ditto.
9837 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
9838 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
9839 (gate_tree_vectorize): Name change.
9840 (tree_vectorize): Ditto.
9841 (pass_vectorize::gate): Call new function.
9842 (pass_vectorize::execute): Ditto.
9843 * opts.c: O3 default setting change.
9844 (finish_options): Check new flag.
9845 * doc/invoke.texi: Document new flags.
9846 * common.opt: New flags.
9848 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
9850 * doc/tm.texi.in (Cond Exec Macros): Remove node.
9851 (Condition Code): Don't reference it.
9852 * doc/tm.texi: Regenerate.
9854 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
9857 * lra-constraints.c (undo_optional_reloads): Consider all optional
9858 reload even if it did not get a hard reg.
9860 2013-09-16 Teresa Johnson <tejohnson@google.com>
9862 * dumpfile.c (dump_loc): Remove newline emission.
9863 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
9864 emission to dump_printf_loc calls where missing.
9865 (vect_mark_for_runtime_alias_test): Ditto.
9866 (vect_analyze_data_ref_dependence): Ditto.
9867 (vect_analyze_data_ref_dependences): Ditto.
9868 (vect_slp_analyze_data_ref_dependence): Ditto.
9869 (vect_slp_analyze_data_ref_dependences): Ditto.
9870 (vect_compute_data_ref_alignment): Ditto.
9871 (vect_update_misalignment_for_peel): Ditto.
9872 (vect_verify_datarefs_alignment): Ditto.
9873 (vector_alignment_reachable_p): Ditto.
9874 (vect_get_data_access_cost): Ditto.
9875 (vect_enhance_data_refs_alignment): Ditto.
9876 (vect_find_same_alignment_drs): Ditto.
9877 (vect_analyze_data_refs_alignment): Ditto.
9878 (vect_analyze_group_access): Ditto.
9879 (vect_analyze_data_ref_access): Ditto.
9880 (vect_analyze_data_ref_accesses): Ditto.
9881 (vect_prune_runtime_alias_test_list): Ditto.
9882 (vect_analyze_data_refs): Ditto.
9883 (vect_create_addr_base_for_vector_ref): Ditto.
9884 (vect_create_data_ref_ptr): Ditto.
9885 (vect_grouped_store_supported): Ditto.
9886 (vect_grouped_load_supported): Ditto.
9887 * value-prof.c (check_counter): Ditto.
9888 (check_ic_target): Ditto.
9889 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
9890 (vect_recog_widen_mult_pattern): Ditto.
9891 (vect_recog_widen_sum_pattern): Ditto.
9892 (vect_recog_over_widening_pattern): Ditto.
9893 (vect_recog_widen_shift_pattern): Ditto.
9894 (vect_recog_rotate_pattern): Ditto.
9895 (vect_recog_vector_vector_shift_pattern): Ditto.
9896 (vect_recog_divmod_pattern): Ditto.
9897 (vect_recog_mixed_size_cond_pattern): Ditto.
9898 (vect_recog_bool_pattern): Ditto.
9899 (vect_pattern_recog_1): Ditto.
9900 (vect_pattern_recog): Ditto.
9901 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
9902 (vect_is_simple_iv_evolution): Ditto.
9903 (vect_analyze_scalar_cycles_1): Ditto.
9904 (vect_get_loop_niters): Ditto.
9905 (vect_analyze_loop_1): Ditto.
9906 (vect_analyze_loop_form): Ditto.
9907 (vect_analyze_loop_operations): Ditto.
9908 (vect_analyze_loop_2): Ditto.
9909 (vect_analyze_loop): Ditto.
9910 (report_vect_op): Ditto.
9911 (vect_is_slp_reduction): Ditto.
9912 (vect_is_simple_reduction_1): Ditto.
9913 (vect_get_known_peeling_cost): Ditto.
9914 (vect_estimate_min_profitable_iters): Ditto.
9915 (vect_model_reduction_cost): Ditto.
9916 (vect_model_induction_cost): Ditto.
9917 (get_initial_def_for_induction): Ditto.
9918 (vect_create_epilog_for_reduction): Ditto.
9919 (vectorizable_reduction): Ditto.
9920 (vectorizable_induction): Ditto.
9921 (vectorizable_live_operation): Ditto.
9922 (vect_loop_kill_debug_uses): Ditto.
9923 (vect_transform_loop): Ditto.
9924 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
9925 (vect_stmt_relevant_p): Ditto.
9926 (process_use): Ditto.
9927 (vect_mark_stmts_to_be_vectorized): Ditto.
9928 (vect_model_simple_cost): Ditto.
9929 (vect_model_promotion_demotion_cost): Ditto.
9930 (vect_model_store_cost): Ditto.
9931 (vect_get_store_cost): Ditto.
9932 (vect_model_load_cost): Ditto.
9933 (vect_get_load_cost): Ditto.
9934 (vect_init_vector_1): Ditto.
9935 (vect_get_vec_def_for_operand): Ditto.
9936 (vect_finish_stmt_generation): Ditto.
9937 (vectorizable_call): Ditto.
9938 (vectorizable_conversion): Ditto.
9939 (vectorizable_assignment): Ditto.
9940 (vectorizable_shift): Ditto.
9941 (vectorizable_operation): Ditto.
9942 (vectorizable_store): Ditto.
9943 (vectorizable_load): Ditto.
9944 (vectorizable_condition): Ditto.
9945 (vect_analyze_stmt): Ditto.
9946 (vect_transform_stmt): Ditto.
9947 (vect_is_simple_use): Ditto.
9948 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
9949 (vect_can_advance_ivs_p): Ditto.
9950 (vect_update_ivs_after_vectorizer): Ditto.
9951 (vect_do_peeling_for_loop_bound): Ditto.
9952 (vect_gen_niters_for_prolog_loop): Ditto.
9953 (vect_update_inits_of_drs): Ditto.
9954 (vect_create_cond_for_alias_checks): Ditto.
9955 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
9956 (vect_build_slp_tree_1): Ditto.
9957 (vect_supported_load_permutation_p): Ditto.
9958 (vect_analyze_slp_instance): Ditto.
9959 (vect_analyze_slp): Ditto.
9960 (vect_make_slp_decision): Ditto.
9961 (vect_detect_hybrid_slp): Ditto.
9962 (vect_bb_vectorization_profitable_p): Ditto.
9963 (vect_slp_analyze_bb_1): Ditto.
9964 (vect_update_slp_costs_according_to_vf): Ditto.
9965 (vect_get_mask_element): Ditto.
9966 (vect_transform_slp_perm_load): Ditto.
9967 (vect_schedule_slp_instance): Ditto.
9968 (vect_schedule_slp): Ditto.
9969 (vect_slp_transform_bb): Ditto.
9970 * profile.c (read_profile_edge_counts): Ditto.
9971 (compute_branch_probabilities): Ditto.
9972 * coverage.c (get_coverage_counts): Ditto.
9974 2013-09-16 Diego Novillo <dnovillo@google.com>
9976 * tree-core.h: Add missing comment lines from refactoring of tree.h.
9978 2013-09-16 Jan Hubicka <jh@suse.cz>
9980 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
9981 abstract functions; for static functions check the presence of body.
9983 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
9985 * config/aarch64/aarch64-simd-builtins.def (fma): New.
9986 * config/aarch64/aarch64-simd.md
9987 (aarch64_mla_elt<mode>): New.
9988 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
9989 (aarch64_mls_elt<mode>): Likewise.
9990 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
9991 (aarch64_fma4_elt<mode>): Likewise.
9992 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
9993 (aarch64_fma4_elt_to_128v2df): Likewise.
9994 (aarch64_fma4_elt_to_64df): Likewise.
9995 (fnma<mode>4): Likewise.
9996 (aarch64_fnma4_elt<mode>): Likewise.
9997 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
9998 (aarch64_fnma4_elt_to_128v2df): Likewise.
9999 (aarch64_fnma4_elt_to_64df): Likewise.
10000 * config/aarch64/iterators.md (VDQSF): New.
10001 * config/aarch64/arm_neon.h
10002 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
10003 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
10005 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
10007 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
10008 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
10009 (aarch64_mul3_elt_to_128df): Likewise.
10010 (aarch64_mul3_elt_to_64v2df): Likewise.
10011 * config/aarch64/iterators.md (VEL): Also handle DFmode.
10013 (VMUL_CHANGE_NLANES) Likewise.
10016 * config/aarch64/arm_neon.h
10017 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
10019 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
10021 * config/aarch64/arm_neon.h
10022 (vcvtx_high_f32_f64): Fix parameters.
10024 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
10025 Uros Bizjak <ubizjak@gmail.com>
10027 * config/alpha.c: Include tree-ssanames.h.
10029 2013-09-16 Richard Biener <rguenther@suse.de>
10031 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
10032 (dot_rdg_1): Handle control_dd.
10033 (create_edge_for_control_dependence): New function.
10034 (create_rdg_edges): Add control dependences if asked for.
10035 (build_rdg): Likewise.
10036 (generate_loops_for_partition): If there are not necessary
10037 control stmts remove all their dependencies.
10038 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
10039 (distribute_loop): Pass on control dependences.
10040 (tree_loop_distribution): Compute control dependences and remove
10041 restriction on number of loop nodes.
10043 2013-09-16 Jakub Jelinek <jakub@redhat.com>
10045 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
10046 for internal calls.
10048 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
10050 * cse.c (try_const_anchors): Punt on CC modes.
10052 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
10054 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
10056 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
10059 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
10060 calls to word_mode.
10062 2013-09-14 Iain Sandoe <iain@codesourcery.com>
10065 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
10067 (darwin_objc1_section): Likewise.
10068 (darwin_file_end): Emit Image Info section when required.
10070 2013-09-14 Jan Hubicka <jh@suse.cz>
10072 * tree-into-ssa.c (gate_into_ssa): New.
10073 (pass_data_build_ssa): Use it.
10074 * cgraph.h (expand_thunk): Update prototype.
10075 * cgraphunit.c (analyze_function): Expand thunks early.
10076 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
10077 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
10078 set lowered flag; do not add new function.
10079 (assemble_thunks_and_aliases): Update.
10080 * tree-ssa.c (gate_init_datastructures): New gate.
10081 (pass_data_init_datastructures): Use it.
10083 2013-09-14 Iain Sandoe <iain@codesourcery.com>
10086 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
10087 xmm register set described in the psABI.
10089 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
10091 * dwarf2out.c (should_emit_struct_debug): Add check
10092 for type_decl variable is not NULL.
10094 2013-09-13 Jacek Caban <jacek@codeweavers.com>
10096 * config.gcc: Use new winnt-c.c target hooks
10097 * config/t-winnt: New file
10098 * config/winnt-c.c: New file
10099 * doc/tm.texi.in: Document new hook
10100 * doc/tm.texi: Regenerated
10102 2013-09-13 Jan Hubicka <jh@suse.cz>
10104 PR middle-end/58094
10105 * ipa-inline.c (check_callers): New function.
10106 (check_caller_edge): Remove.
10107 (want_inline_function_to_all_callers_p): Also permit alises that are
10109 (inline_to_all_callers): Terminate the walk when devirtualization
10110 introduce new calls.
10112 2013-09-13 Jan Hubicka <jh@suse.cz>
10114 * ipa-inline-analysis.c (struct growth_data): Add node.
10115 (do_estimate_growth_1): Fix detection of recursion.
10117 2013-09-13 Jakub Jelinek <jakub@redhat.com>
10119 PR tree-optimization/58392
10120 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
10121 to avoid shadowing of outer loop variable. If
10122 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
10123 replace_by_duplicate_decl simduid of loops that have it set and
10124 set dest_cfun->has_simduid_loops and/or
10125 dest_cfun->has_force_vect_loops.
10126 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
10127 instead of maybe_lookup_decl.
10128 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
10129 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
10130 call. Copy over force_vect and copy and remap simduid. Set
10131 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
10132 (copy_cfg_body): Remove blocks_to_copy argument. Use
10133 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
10134 caller. Don't set cfun->has_simduid_loops and/or
10135 cfun->has_force_vect_loops here.
10136 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
10138 (expand_call_inline, tree_function_versioning): Adjust copy_body
10141 2013-09-13 Martin Jambor <mjambor@suse.cz>
10144 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
10145 the assert if the edge was a speculative one.
10147 2013-09-13 Richard Biener <rguenther@suse.de>
10149 * tree-data-ref.h (known_dependences_p): Move here ...
10150 * tree-loop-distribution.c (known_dependences_p): ... from here.
10151 (dump_rdg_component, debug_rdg_component): Remove.
10152 (dump_rdg): Adjust.
10153 (generate_loops_for_partition): Use gimple_uid instead of
10154 relying on matching stmt visit order.
10155 (rdg_build_partitions): Take starting stmt vector.
10156 (ldist_gen): Merge into ...
10157 (distribute_loop): ... this function. Do not compute starting
10159 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
10161 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10163 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
10164 Set type attribute correctly. Set predicable_short_it attribute.
10165 (cmpsi_shiftsi): Remove %? from output template.
10167 2013-09-13 Richard Biener <rguenther@suse.de>
10169 * tree-loop-distribution.c (struct rdg_component,
10170 rdg_defs_used_in_other_loops_p, free_rdg_components,
10171 rdg_build_components): Remove.
10172 (stmts_from_loop): Do not record virtual PHIs.
10173 (generate_loops_for_partition): Skip virtual PHIs.
10174 (build_rdg_partition_for_component): Rename to ...
10175 (build_rdg_partition_for_vertex): ... this and adjust.
10176 (rdg_build_partitions): Take a vector of starting vertices
10177 instead of components. Remove unnecessary leftover handling.
10178 (ldist_gen): Do not build components or record other stores.
10179 (distribute_loop): Do not distribute loops containing stmts
10182 2013-09-13 Christian Bruel <christian.bruel@st.com>
10185 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
10187 2013-09-13 Kai Tietz <ktietz@redhat.com>
10189 * config.gcc: Separate cases for mingw and cygwin targets,
10190 and add 64-bit cygwin target case.
10192 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
10193 dll-export inline-functions.
10194 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
10196 2013-09-13 Jeff Law <law@redhat.com>
10198 PR middle-end/58387
10200 2013-09-06 Jeff Law <law@redhat.com>
10202 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
10203 edge implied equivalences into successor phis.
10205 2013-09-12 DJ Delorie <dj@redhat.com>
10207 * config/rl78/rl78-virt.md: Change from | to \; for asm line
10210 2013-09-12 Brooks Moses <bmoses@google.com>
10213 * Makefile.in: Do not install driver binaries in $(target)/bin.
10215 2013-09-12 DJ Delorie <dj@redhat.com>
10217 * config/rl78/rl78.opt (mrelax): New.
10218 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
10219 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
10221 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
10223 (rl78_expand_epilogue): Likewise.
10225 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
10227 PR middle-end/58335
10228 * lra-eliminations.c (remove_reg_equal_offset_note): New.
10229 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
10230 pointer elimination with using remove_reg_equal_offset_note.
10232 2013-09-12 DJ Delorie <dj@redhat.com>
10234 * config/msp430/: New port.
10235 * config.gcc (msp430): Added.
10236 * doc/invoke.texi: Document MSP430 options.
10237 * doc/install.texi: Document msp430-elf
10238 * doc/md.texi: Document msp430-elf
10239 * doc/contrib.texi: Document msp430-elf
10241 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
10244 2013-09-12 Martin Jambor <mjambor@suse.cz>
10247 * ipa-prop.c (remove_described_reference): Give up if the edge in the
10248 reference descriptor is NULL.
10249 (ipa_edge_removal_hook): If owning a reference descriptor, set its
10252 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
10254 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
10255 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
10256 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
10257 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
10258 make_temp_ssa_name): move to tree-ssanames.h
10259 * tree-ssa-alias.h: Move prototype.
10260 * tree-ssa.h: Include tree-ssanames.h.
10261 * tree-ssanames.c (FREE_SSANAMES): Move to here.
10262 * tree-ssanames.h: New. Move items from tree-flow*.h
10263 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
10265 2013-09-12 Richard Biener <rguenther@suse.de>
10267 PR tree-optimization/58404
10268 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
10269 propagate non-invariant addresses into dereferences wrapped
10270 in component references.
10272 2013-09-12 Richard Biener <rguenther@suse.de>
10274 PR tree-optimization/58402
10275 * passes.def: Move pass_late_warn_uninitialized later.
10277 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
10279 * tree-ssa.h: New. Move content from tree-flow.h and
10280 tree-flow-inline.h.
10281 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
10282 Move prototypes belonging to tree-ssa.c.
10283 * tree-flow-inline.h (redirect_edge_var_map_def,
10284 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
10286 * gimple.h: Adjust prototypes.
10287 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
10289 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
10290 * tree.h: Move prototype to tree-ssa.h.
10291 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
10292 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
10293 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
10294 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
10295 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
10296 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
10297 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
10298 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
10299 graphite-blocking.c, graphite-clast-to-gimple.c,
10300 graphite-dependences.c, graphite-interchange.c,
10301 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
10302 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
10303 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
10304 ipa-reference.c, ipa-split.c, ipa-utils.c,
10305 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
10306 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
10307 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
10308 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
10309 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
10310 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
10311 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
10312 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
10313 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
10314 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
10315 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
10316 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
10317 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
10318 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
10319 value-prof.c, var-tracking.c,
10320 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
10322 2013-09-12 Richard Biener <rguenther@suse.de>
10324 PR tree-optimization/58396
10325 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
10326 (build_rdg): Take a loop-nest parameter, fix memleaks.
10327 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
10329 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
10331 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
10334 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
10336 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
10337 The FMA instruction names should have a 'v' prefix.
10339 2013-09-12 Richard Biener <rguenther@suse.de>
10341 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
10342 (dot_rdg): Use popen instead of system in optional code.
10343 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
10344 (already_processed_vertex_p): Adjust.
10345 (has_anti_or_output_dependence, predecessor_has_mem_write,
10346 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
10347 rdg_flag_uses): Remove.
10348 (rdg_flag_vertex): Simplify.
10349 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
10351 (build_rdg_partition_for_component): Process the first vertex
10352 of a component only.
10353 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
10355 2013-09-12 Alan Modra <amodra@gmail.com>
10357 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
10359 2013-09-11 DJ Delorie <dj@redhat.com>
10360 Nick Clifton <nickc@redhat.com>
10362 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
10363 (rl78_stack_based_mem): New.
10364 * config/rl78/constraints.md (Iv08): New.
10373 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
10375 (movsi): Change from expand to insn-and-split.
10376 (ashrsi3): Clobber AX.
10380 * config/rl78/rl78.md (CC_REG): Fix.
10381 (addsi3): Allow memory and immediate operands.
10382 (addsi3_internal): Split into...
10383 (addsi3_internal_virt): ...new, and ...
10384 (addsi3_internal_real): ...new.
10386 (subsi3_internal_virt): New.
10387 (subsi3_internal_real): New.
10388 (mulsi3): Add memory operand.
10389 (mulsi3_rl78): Likewise.
10390 (mulsi3_g13): Likewise.
10391 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
10392 (cbranchqi4_real): Add more constraint options.
10393 (cbranchhi4_real): Expand pattern.
10394 (cbranchhi4_real_signed): New.
10395 (cbranchhi4_real_inverted): New.
10396 (cbranchsi4_real_lt): New.
10397 (cbranchsi4_real_ge): New.
10398 (cbranchsi4_real_signed): New.
10399 (cbranchsi4_real): New.
10401 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
10403 (lshrsi3_virt): Likewise.
10404 (ashlsi3_virt): Likewise.
10405 (cbranchqi4_virt_signed): New.
10406 (cbranchhi4_virt_signed): New.
10407 (cbranchsi4_virt): New.
10408 * config/rl78/rl78.c: Whitespace fixes throughout.
10409 (move_elim_pass): New.
10410 (pass_data_rl78_move_elim): New.
10411 (pass_rl78_move_elim): New.
10412 (make_pass_rl78_move_elim): New.
10413 (rl78_devirt_info): Run devirt earlier.
10414 (rl78_move_elim_info): New.
10415 (rl78_asm_file_start): Register it.
10416 (rl78_split_movsi): New.
10417 (rl78_as_legitimate_address): Allow virtual base registers when
10419 (rl78_addr_space_convert): Remove spurious debug stuff.
10420 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
10421 (rl78_print_operand): More cases for not printing '#'.
10422 (rl78_expand_compare): Remove most of the logic.
10423 (content_memory): New.
10424 (clear_content_memory): New.
10425 (get_content_index): New.
10426 (get_content_name): New.
10427 (display_content_memory): New.
10428 (update_content): New.
10429 (record_content): New.
10430 (already_contains): New.
10431 (insn_ok_now): Re-recog insns with virtual registers.
10432 (add_postponed_content_update): New.
10433 (process_postponed_content_update): New.
10434 (gen_and_emit_move): New.
10435 (transcode_memory_rtx): Record new location content.
10436 Use gen_and_emit_move.
10437 (force_into_acc): New.
10438 (move_to_acc): Use gen_and_emit_move.
10439 (move_from_acc): Likewise.
10440 (move_acc_to_reg): Likewise.
10441 (move_to_x): Likewise.
10442 (move_to_hl): Likewise.
10443 (move_to_de): Likewise.
10444 (rl78_alloc_physical_registers_op1): Record location content.
10445 (has_constraint): New.
10446 (rl78_alloc_physical_registers_op2): Record location content.
10447 Optimize use of HL.
10448 (rl78_alloc_physical_registers_ro1): Likewise.
10449 (rl78_alloc_physical_registers_cmp): Likewise.
10450 (rl78_alloc_physical_registers_umul): Likewise.
10451 (rl78_alloc_address_registers_macax): New.
10452 (rl78_alloc_physical_registers): Initialize and set location
10453 content memory as needed.
10454 (rl78_reorg): Make sure split2 is called.
10455 (rl78_rtx_costs): New.
10457 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
10459 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
10460 for (not (neg ...)) and (neg (not ...)) cases.
10462 2013-09-11 Richard Biener <rguenther@suse.de>
10464 PR middle-end/58377
10465 * passes.def: Split critical edges before late uninit warning passes.
10466 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
10468 2013-09-11 Jakub Jelinek <jakub@redhat.com>
10470 PR tree-optimization/58385
10471 * fold-const.c (build_range_check): If both low and high are NULL,
10472 use omit_one_operand_loc to preserve exp side-effects.
10474 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10476 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
10478 2013-09-11 Richard Biener <rguenther@suse.de>
10480 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
10481 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
10482 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
10483 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
10484 stmts_from_loop, known_dependences_p, build_empty_rdg,
10485 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
10486 * tree-loop-distribution.c: ... here.
10487 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
10488 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
10489 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
10490 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
10491 * tree-loop-distribution.c: ... here.
10492 * tree-loop-distribution.c: Include gimple-pretty-print.h.
10493 (struct partition_s): Add loops member.
10494 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
10495 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
10496 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
10498 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
10499 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
10500 Sergey Lega <sergey.s.lega@intel.com>
10501 Anna Tikhonova <anna.tikhonova@intel.com>
10502 Ilya Tocar <ilya.tocar@intel.com>
10503 Andrey Turetskiy <andrey.turetskiy@intel.com>
10504 Ilya Verbin <ilya.verbin@intel.com>
10505 Kirill Yukhin <kirill.yukhin@intel.com>
10506 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
10508 * config/i386/constraints.md (k): New.
10510 * config/i386/i386.c (const regclass_map): Add new mask registers.
10511 (dbx_register_map): Ditto.
10512 (dbx64_register_map): Ditto.
10513 (svr4_dbx_register_map): Ditto.
10514 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
10516 (ix86_preferred_reload_class): Disable constants for mask registers.
10517 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
10518 (ix86_hard_regno_mode_ok): Support new mask registers.
10519 (x86_order_regs_for_local_alloc): Ditto.
10520 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
10521 (FIXED_REGISTERS): Add new mask registers.
10522 (CALL_USED_REGISTERS): Ditto.
10523 (REG_ALLOC_ORDER): Ditto.
10524 (VALID_MASK_REG_MODE): New.
10525 (FIRST_MASK_REG): Ditto.
10526 (LAST_MASK_REG): Ditto.
10527 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
10528 (MAYBE_MASK_CLASS_P): New.
10529 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
10530 (REG_CLASS_CONTENTS): Ditto.
10531 (MASK_REGNO_P): New.
10532 (ANY_MASK_REG_P): Ditto.
10533 (HI_REGISTER_NAMES): Add new mask registers.
10534 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
10535 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
10537 (attribute "type"): Add mskmov, msklog.
10538 (attribute "length_immediate"): Support them.
10539 (attribute "memory"): Ditto.
10540 (attribute "prefix_0f"): Ditto.
10541 (*movhi_internal): Support new mask registers.
10542 (*movqi_internal): Ditto.
10543 (define_split): Split out clobber pattern is a logic
10544 insn on mask registers.
10545 (*k<logic><mode>): New.
10546 (*andhi_1): Extend to support mask regs.
10547 (*andqi_1): Extend to support mask regs.
10548 (kandn<mode>): New.
10549 (define_split): Split and-not to and and not if operands
10551 (*<code><mode>_1): Separate HI mode to new pattern...
10552 (*<code>hi_1): This.
10553 (*<code>qi_1): Extend to support mask regs.
10554 (kxnor<mode>): New.
10555 (kortestzhi): Ditto.
10556 (kortestchi): Ditto.
10558 (*one_cmpl<mode>2_1): Remove HImode and handle it...
10559 (*one_cmplhi2_1): ...Here, now with mask registers support.
10560 (*one_cmplqi2_1): Support new mask registers.
10561 (HI/QImode arithmetics splitter): Don't split if mask registers
10563 (HI/QImode not splitter): Ditto.
10564 * config/i386/predicated.md (mask_reg_operand): New.
10565 (general_reg_operand): Ditto.
10567 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
10569 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
10570 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
10572 2013-09-10 Jeff Law <law@redhat.com>
10574 PR tree-optimization/58380
10575 * tree-ssa-threadupdate.c (thread_block): Recognize another case
10576 of threading through a buried loop header.
10578 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
10579 return value for single successor case.
10581 2013-09-10 Jan Hubicka <jh@suse.cz>
10583 * ipa-devirt.c (ipa_devirt): Enable with LTO.
10585 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
10588 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
10589 support conditional execution.
10590 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
10592 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
10594 * lra.c (lra): Clear lra_optional_reload_pseudos before every
10596 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
10597 Check destination too to check move insn.
10598 (undo_optional_reloads): Add check that the original peudo did not
10599 changed its allocation and the optional reload was inherited on last
10600 inheritance pass. Break loop after deciding to keep optional reload.
10601 (lra_undo_inheritance): Add check that inherited pseudo still in
10604 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
10606 * config/aarch64/aarch64.md (generic_sched): New.
10607 * config/aarch64/aarch64-generic.md (load): Make conditional
10608 on generic_sched attribute.
10609 (nonload): Likewise.
10611 2013-09-10 Jan Hubicka <jh@suse.cz>
10613 * lto-cgraph.c: Include ipa-utils.h.
10614 (compute_ltrans_boundary): Also add possible targets into the boundary.
10616 2013-09-10 Jan Hubicka <jh@suse.cz>
10618 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
10619 VAR_DECL of vtable rather than full expression.
10621 2013-09-10 Jan Hubicka <jh@suse.cz>
10622 Paolo Carlini <paolo.carlini@oracle.com>
10624 * cgraphunit.c (analyze_functions): Save input_location, set it
10625 to UNKNOWN_LOCATION and restore it at the end.
10627 2013-09-10 Martin Jambor <mjambor@suse.cz>
10629 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
10630 represented by a thunk.
10632 2013-09-10 Jeff Law <law@redhat.com>
10634 PR tree-optimization/58343
10635 * tree-ssa-threadupdate.c (thread_block): Identify and disable
10636 jump threading requests through loop headers buried in the middle
10637 of a jump threading path.
10639 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
10640 in return value/type.
10642 2013-09-10 Jakub Jelinek <jakub@redhat.com>
10644 PR rtl-optimization/58365
10645 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
10646 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
10649 2013-09-10 Richard Biener <rguenther@suse.de>
10651 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
10652 * tree-data-ref.c (create_rdg_vertices): Collect all data
10653 references, signal failure to the caller, use data-ref API.
10654 (build_rdg): Compute data references only once. Maintain lifetime
10655 of data references and data dependences from within RDG.
10656 (free_rdg): Free dependence relations.
10657 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
10658 inventing extra dependences.
10659 (distribute_loop): Update for RDG API changes.
10661 2013-09-10 Kai Tietz <ktietz@redhat.com>
10663 * doc/invoke.texi (fms-extensions): Document changed
10664 behavior for ms-abi targets.
10665 * config/i386/i386.c (ix86_option_override_internal):
10666 Set default value of option -fms-extension for ms-abi targets.
10668 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
10670 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
10672 2013-09-10 Alan Modra <amodra@gmail.com>
10675 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
10677 2013-09-10 Alan Modra <amodra@gmail.com>
10679 * config/rs6000/predicates.md (add_cint_operand): New.
10680 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
10681 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
10682 using add_cint_operand.
10683 (largetoc_high_plus_aix): Likewise.
10685 2013-09-09 Jakub Jelinek <jakub@redhat.com>
10687 PR tree-optimization/58364
10688 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
10689 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
10690 the current range can't be an unconditional true or false.
10692 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
10694 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
10696 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
10698 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
10700 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
10703 * doc/invoke.texi (-Wdelete-incomplete): Document it.
10705 2013-09-09 Ian Bolton <ian.bolton@arm.com>
10707 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
10708 NO_REGS for immediate that can't be moved directly into FP_REGS.
10710 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10712 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
10713 comparison with negated operand.
10714 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
10717 2013-09-09 Richard Biener <rguenther@suse.de>
10719 PR middle-end/58326
10720 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
10721 of a subloop record all its block as affecting loop-closed SSA form.
10723 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
10725 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
10726 of an rtx/bitpos pair.
10727 (store_fixed_bit_field): Update accordingly.
10729 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
10731 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
10733 * builtins.c (expand_errno_check): Likewise.
10734 * dwarf2cfi.c (init_return_column_size): Likewise.
10735 * except.c (sjlj_mark_call_sites): Likewise.
10736 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
10737 * lra-constraints.c (emit_inc): Likewise.
10738 * ree.c (combine_set_extension): Likewise.
10739 * regmove.c (fixup_match_2): Likewise.
10740 * reload1.c (inc_for_reload): Likewise.
10742 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
10744 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
10745 (force_to_mode, simplify_shift_const_1, simplify_comparison):
10746 Use gen_int_mode with the mode of the associated simplify_* call.
10747 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
10748 * expmed.c (expand_shift_1): Likewise.
10749 * function.c (instantiate_virtual_regs_in_insn): Likewise.
10750 * loop-iv.c (iv_number_of_iterations): Likewise.
10751 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
10752 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
10754 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
10756 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
10757 of the associated expand_* call.
10758 (asan_emit_stack_protection): Likewise.
10759 * builtins.c (round_trampoline_addr): Likewise.
10760 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
10761 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
10762 (emit_store_flag): Likewise.
10763 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
10764 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
10766 * function.c (instantiate_virtual_regs_in_insn): Likewise.
10767 * ifcvt.c (noce_try_store_flag_constants): Likewise.
10768 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
10769 * modulo-sched.c (generate_prolog_epilog): Likewise.
10770 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
10771 (expand_ctz, expand_ffs, expand_unop): Likewise.
10773 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
10775 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
10776 of the associated gen_rtx_* call.
10777 * caller-save.c (init_caller_save): Likewise.
10778 * combine.c (find_split_point, make_extraction): Likewise.
10779 (make_compound_operation): Likewise.
10780 * dwarf2out.c (mem_loc_descriptor): Likewise.
10781 * explow.c (plus_constant, probe_stack_range): Likewise.
10782 * expmed.c (expand_mult_const): Likewise.
10783 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
10784 * reload1.c (init_reload): Likewise.
10785 * valtrack.c (cleanup_auto_inc_dec): Likewise.
10786 * var-tracking.c (adjust_mems): Likewise.
10787 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
10788 rather than gen_rtx_fmt_ee.
10790 2013-09-09 Jan Hubicka <jh@suse.cz>
10792 PR middle-end/58294
10793 * value-prof.c (gimple_ic): Copy also abnormal edges.
10795 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
10797 * asan.c (asan_shadow_cst): Use gen_int_mode.
10799 2013-09-08 Jan Hubicka <jh@suse.cz>
10801 * ipa-profile.c: Add toplevel comment.
10802 (ipa_propagate_frequency_1): Be more conservative when profile is read.
10803 (contains_hot_call_p): New function.
10804 (ipa_propagate_frequency): Set frequencies based on counts when
10806 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
10807 profile; do not tamper with profile after inlining if it is read.
10809 2013-09-08 Jan Hubicka <jh@suse.cz>
10811 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
10814 2013-09-08 Jan Hubicka <jh@suse.cz>
10816 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
10817 summary generation.
10819 2013-09-08 Jeff Law <law@redhat.com>
10822 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
10825 2013-09-08 Andi Kleen <ak@linux.intel.com>
10827 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
10829 2013-09-08 Jan Hubicka <jh@suse.cz>
10831 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
10833 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
10835 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
10836 for non-debug insns.
10837 * lra.c (new_insn_reg): Take the containing insn as a parameter.
10838 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
10839 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
10842 2013-09-08 Jan Hubicka <jh@suse.cz>
10844 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
10845 targets and devirtualize to BUILT_IN_UNREACHABLE.
10846 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
10847 * ipa.c (walk_polymorphic_call_targets): New function.
10848 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
10849 functions; use the new timevar.
10850 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
10851 was removed from the program.
10852 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
10853 not consider it in the walk when its vtable is dead.
10854 (possible_polymorphic_call_targets_1): Pass anonymous flag to
10856 (devirt_variable_node_removal_hook): New function.
10857 (possible_polymorphic_call_targets): Also register
10858 devirt_variable_node_removal_hook.
10859 (ipa_devirt): Do not do non-speculative devirtualization.
10860 (gate_ipa_devirt): One execute if devirtualizing speculatively.
10862 2013-09-08 Jan Hubicka <jh@suse.cz>
10864 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
10865 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
10866 varpool_remove_variable_insertion_hook): Declare.
10867 * varpool.c (varpool_node_hook_list): New structure.
10868 (first_varpool_node_removal_hook,
10869 first_varpool_variable_insertion_hook): New variables.
10870 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
10871 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
10872 varpool_remove_variable_insertion_hook,
10873 varpool_call_variable_insertion_hooks): New functions.
10874 (varpool_remove_node): Use it.
10876 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
10879 * diagnostic.c (diagnostic_build_prefix): When s.file is
10880 "<built-in>" don't output line and column numbers.
10882 2013-09-06 Jan Hubicka <jh@suse.cz>
10884 * cgraphunit.c (expand_thunk): Get body before touching arguments.
10885 * lto-streamer-out.c: Stream thunks, too.
10886 * lto-streamer-in.c (input_function): Pop cfun here
10887 (lto_read_body): Instead of here.
10889 2013-09-06 Caroline Tice <cmtice@google.com>
10891 * doc/install.texi: Add documentation for the --enable-vtable-verify
10892 and the --disable-libvtv configure options.
10894 2013-09-06 Jeff Law <law@redhat.com>
10896 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
10897 edge implied equivalences into successor phis.
10899 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
10901 * resource.c (mark_referenced_resources): Handle COND_EXEC.
10903 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
10905 * resource.c (mark_target_live_regs): Compute resources taking
10906 into account if a call is predicated or not.
10908 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
10910 * toplev.c (output_stack_usage): Be prepared for suffixes created by
10911 the compiler in the function names.
10913 2013-09-06 Jan Hubicka <jh@suse.cz>
10915 PR middle-end/58094
10916 * ipa-inline.c (has_caller_p): New function.
10917 (want_inline_function_to_all_callers_p): Use it.
10918 (sum_callers, inline_to_all_callers): Break out from ...
10919 (ipa_inline): ... here.
10921 2013-09-06 Jan Hubicka <jh@suse.cz>
10923 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
10924 only when AVX is enabled.
10926 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
10928 * config/aarch64/aarch64.md
10929 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
10930 is fpsimd_<load/store>2.
10931 (load_pair<mode>): Likewise.
10932 (store_pair<mode>): Likewise.
10934 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
10936 * config/arm/types.md (type): Add "mrs" type.
10937 * config/aarch64/aarch64.md
10938 (aarch64_load_tp_hard): Make type "mrs".
10939 * config/arm/arm.md
10940 (load_tp_hard): Make type "mrs".
10941 * config/arm/cortex-a15.md: Update with new attributes.
10942 * config/arm/cortex-a5.md: Update with new attributes.
10943 * config/arm/cortex-a53.md: Update with new attributes.
10944 * config/arm/cortex-a7.md: Update with new attributes.
10945 * config/arm/cortex-a8.md: Update with new attributes.
10946 * config/arm/cortex-a9.md: Update with new attributes.
10947 * config/arm/cortex-m4.md: Update with new attributes.
10948 * config/arm/cortex-r4.md: Update with new attributes.
10949 * config/arm/fa526.md: Update with new attributes.
10950 * config/arm/fa606te.md: Update with new attributes.
10951 * config/arm/fa626te.md: Update with new attributes.
10952 * config/arm/fa726te.md: Update with new attributes.
10954 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
10956 * config/aarch64/aarch64.md
10957 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
10958 (*movtf_aarch64): Likewise.
10959 * config/arm/arm.md
10960 (thumb1_movdi_insn): Use "multiple" for type where more than one
10961 instruction is used for a move.
10962 (*arm32_movhf): Likewise.
10963 (*thumb_movdf_insn): Likewise.
10965 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
10967 * config/arm/types.md (type): Rename fcpys to fmov.
10968 * config/arm/vfp.md
10969 (*arm_movsi_vfp): Rename type fcpys as fmov.
10970 (*thumb2_movsi_vfp): Likewise
10971 (*movhf_vfp_neon): Likewise
10972 (*movhf_vfp): Likewise
10973 (*movsf_vfp): Likewise
10974 (*thumb2_movsf_vfp): Likewise
10975 (*movsfcc_vfp): Likewise
10976 (*thumb2_movsfcc_vfp): Likewise
10977 * config/aarch64/aarch64-simd.md
10978 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
10979 * config/aarch64/aarch64.md
10980 (*movsi_aarch64): Replace type mov_reg with fmovs.
10981 (*movdi_aarch64): Likewise
10982 (*movsf_aarch64): Likewise
10983 (*movdf_aarch64): Likewise
10985 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
10986 * config/arm/iwmmxt.md
10987 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
10988 * config/arm/arm1020e.md: Update with new attributes.
10989 * config/arm/cortex-a15-neon.md: Update with new attributes.
10990 * config/arm/cortex-a5.md: Update with new attributes.
10991 * config/arm/cortex-a53.md: Update with new attributes.
10992 * config/arm/cortex-a7.md: Update with new attributes.
10993 * config/arm/cortex-a8-neon.md: Update with new attributes.
10994 * config/arm/cortex-a9.md: Update with new attributes.
10995 * config/arm/cortex-m4-fpu.md: Update with new attributes.
10996 * config/arm/cortex-r4f.md: Update with new attributes.
10997 * config/arm/marvell-pj4.md: Update with new attributes.
10998 * config/arm/vfp11.md: Update with new attributes.
11000 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
11002 * config/aarch64/aarch64.md
11003 (*madd<mode>): Fix type attribute.
11004 (*maddsi_uxtw): Likewise.
11005 (*msub<mode>): Likewise.
11006 (*msubsi_uxtw): Likewise.
11007 (<su_optab>maddsidi4): Likewise.
11008 (<su_optab>msubsidi4): Likewise.
11010 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
11012 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
11013 * config/arm/arm.md (core_cycles): Remove fdiv.
11014 * config/arm/vfp.md:
11015 (*sqrtsf2_vfp): Update for attribute changes.
11016 (*sqrtdf2_vfp): Likewise.
11017 * config/aarch64/aarch64.md:
11018 (sqrt<mode>2): Update for attribute changes.
11019 * config/arm/arm1020e.md: Update with new attributes.
11020 * config/arm/cortex-a15-neon.md: Update with new attributes.
11021 * config/arm/cortex-a5.md: Update with new attributes.
11022 * config/arm/cortex-a53.md: Update with new attributes.
11023 * config/arm/cortex-a7.md: Update with new attributes.
11024 * config/arm/cortex-a8-neon.md: Update with new attributes.
11025 * config/arm/cortex-a9.md: Update with new attributes.
11026 * config/arm/cortex-m4-fpu.md: Update with new attributes.
11027 * config/arm/cortex-r4f.md: Update with new attributes.
11028 * config/arm/marvell-pj4.md: Update with new attributes.
11029 * config/arm/vfp11.md: Update with new attributes.
11031 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
11033 * config/arm/types.md
11034 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
11035 * config/aarch64/aarch64.md
11036 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
11038 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
11039 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
11040 (float<GPI:mode><GPF:mode>2): Likewise.
11041 * config/arm/vfp.md
11042 (*truncsisf2_vfp): Update with new attributes.
11043 (*truncsidf2_vfp): Likewise.
11044 (fixuns_truncsfsi2): Likewise.
11045 (fixuns_truncdfsi2): Likewise.
11046 (*floatsisf2_vfp): Likewise.
11047 (*floatsidf2_vfp): Likewise.
11048 (floatunssisf2): Likewise.
11049 (floatunssidf2): Likewise.
11050 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
11051 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
11052 * config/arm/arm1020e.md: Update with new attributes.
11053 * config/arm/cortex-a15-neon.md: Update with new attributes.
11054 * config/arm/cortex-a5.md: Update with new attributes.
11055 * config/arm/cortex-a53.md: Update with new attributes.
11056 * config/arm/cortex-a7.md: Update with new attributes.
11057 * config/arm/cortex-a8-neon.md: Update with new attributes.
11058 * config/arm/cortex-a9.md: Update with new attributes.
11059 * config/arm/cortex-m4-fpu.md: Update with new attributes.
11060 * config/arm/cortex-r4f.md: Update with new attributes.
11061 * config/arm/marvell-pj4.md: Update with new attributes.
11062 * config/arm/vfp11.md: Update with new attributes.
11064 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
11066 * config/aarch64/arm_neon.h
11067 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
11068 (vqtbx<1,2,3,4><q>_s8): Likewise.
11070 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
11072 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
11073 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
11074 (add<mode>3): Add type attribute.
11075 (add<mode>3): Likewise.
11076 (usadd<mode>3): Likewise.
11077 (ssadd<mode>3): Likewise.
11078 (sub<mode>3): Likewise.
11079 (sub<mode>3): Likewise.
11080 (ussub<mode>3): Likewise.
11081 (sssub<mode>3): Likewise.
11082 (ssmulsa3): Likewise.
11083 (usmulusa3): Likewise.
11084 (arm_usatsihi): Likewise.
11085 * config/arm/vfp.md
11086 (*movdi_vfp): Add types for all instructions.
11087 (*movdi_vfp_cortexa8): Likewise.
11088 (*movhf_vfp_neon): Likewise.
11089 (*movhf_vfp): Likewise.
11090 (*movdf_vfp): Likewise.
11091 (*thumb2_movdf_vfp): Likewise.
11092 (*thumb2_movdfcc_vfp): Likewise.
11093 * config/arm/arm.md: Add type attribute to all insn patterns.
11094 (*thumb1_adddi3): Add type attribute.
11095 (*arm_adddi3): Likewise.
11096 (*adddi_sesidi_di): Likewise.
11097 (*adddi_zesidi_di): Likewise.
11098 (*thumb1_addsi3): Likewise.
11099 (addsi3_compare0): Likewise.
11100 (*addsi3_compare0_scratch): Likewise.
11101 (*compare_negsi_si): Likewise.
11102 (cmpsi2_addneg): Likewise.
11103 (*addsi3_carryin_<optab>): Likewise.
11104 (*addsi3_carryin_alt2_<optab>): Likewise.
11105 (*addsi3_carryin_clobercc_<optab>): Likewise.
11106 (*subsi3_carryin): Likewise.
11107 (*subsi3_carryin_const): Likewise.
11108 (*subsi3_carryin_compare): Likewise.
11109 (*subsi3_carryin_compare_const): Likewise.
11110 (*arm_subdi3): Likewise.
11111 (*thumb_subdi3): Likewise.
11112 (*subdi_di_zesidi): Likewise.
11113 (*subdi_di_sesidi): Likewise.
11114 (*subdi_zesidi_di): Likewise.
11115 (*subdi_sesidi_di): Likewise.
11116 (*subdi_zesidi_ze): Likewise.
11117 (thumb1_subsi3_insn): Likewise.
11118 (*arm_subsi3_insn): Likewise.
11119 (*anddi3_insn): Likewise.
11120 (*anddi_zesidi_di): Likewise.
11121 (*anddi_sesdi_di): Likewise.
11122 (*ne_zeroextracts): Likewise.
11123 (*ne_zeroextracts): Likewise.
11124 (*ite_ne_zeroextr): Likewise.
11125 (*ite_ne_zeroextr): Likewise.
11126 (*anddi_notdi_di): Likewise.
11127 (*anddi_notzesidi): Likewise.
11128 (*anddi_notsesidi): Likewise.
11129 (andsi_notsi_si): Likewise.
11130 (thumb1_bicsi3): Likewise.
11131 (*iordi3_insn): Likewise.
11132 (*iordi_zesidi_di): Likewise.
11133 (*iordi_sesidi_di): Likewise.
11134 (*thumb1_iorsi3_insn): Likewise.
11135 (*xordi3_insn): Likewise.
11136 (*xordi_zesidi_di): Likewise.
11137 (*xordi_sesidi_di): Likewise.
11138 (*arm_xorsi3): Likewise.
11139 (*andsi_iorsi3_no): Likewise.
11140 (*smax_0): Likewise.
11141 (*smax_m1): Likewise.
11142 (*arm_smax_insn): Likewise.
11143 (*smin_0): Likewise.
11144 (*arm_smin_insn): Likewise.
11145 (*arm_umaxsi3): Likewise.
11146 (*arm_uminsi3): Likewise.
11147 (*minmax_arithsi): Likewise.
11148 (*minmax_arithsi_): Likewise.
11149 (*satsi_<SAT:code>): Likewise.
11150 (arm_ashldi3_1bit): Likewise.
11151 (arm_ashrdi3_1bit): Likewise.
11152 (arm_lshrdi3_1bit): Likewise.
11153 (*arm_negdi2): Likewise.
11154 (*thumb1_negdi2): Likewise.
11155 (*arm_negsi2): Likewise.
11156 (*thumb1_negsi2): Likewise.
11157 (*negdi_extendsid): Likewise.
11158 (*negdi_zero_extend): Likewise.
11159 (*arm_abssi2): Likewise.
11160 (*thumb1_abssi2): Likewise.
11161 (*arm_neg_abssi2): Likewise.
11162 (*thumb1_neg_abss): Likewise.
11163 (one_cmpldi2): Likewise.
11164 (extend<mode>di2): Likewise.
11165 (*compareqi_eq0): Likewise.
11166 (*arm_extendhisi2addsi): Likewise.
11167 (*arm_movdi): Likewise.
11168 (*thumb1_movdi_insn): Likewise.
11169 (*arm_movt): Likewise.
11170 (*thumb1_movsi_insn): Likewise.
11171 (pic_add_dot_plus_four): Likewise.
11172 (pic_add_dot_plus_eight): Likewise.
11173 (tls_load_dot_plus_eight): Likewise.
11174 (*thumb1_movhi_insn): Likewise.
11175 (*thumb1_movsf_insn): Likewise.
11176 (*movdf_soft_insn): Likewise.
11177 (*thumb_movdf_insn): Likewise.
11178 (cbranchsi4_insn): Likewise.
11179 (cbranchsi4_scratch): Likewise.
11180 (*negated_cbranchsi4): Likewise.
11181 (*tbit_cbranch): Likewise.
11182 (*tlobits_cbranch): Likewise.
11183 (*tstsi3_cbranch): Likewise.
11184 (*cbranchne_decr1): Likewise.
11185 (*addsi3_cbranch): Likewise.
11186 (*addsi3_cbranch_scratch): Likewise.
11187 (*arm_cmpdi_insn): Likewise.
11188 (*arm_cmpdi_unsig): Likewise.
11189 (*arm_cmpdi_zero): Likewise.
11190 (*thumb_cmpdi_zero): Likewise.
11191 (*deleted_compare): Likewise.
11192 (*mov_scc): Likewise.
11193 (*mov_negscc): Likewise.
11194 (*mov_notscc): Likewise.
11195 (*cstoresi_eq0_thumb1_insn): Likewise.
11196 (cstoresi_nltu_thumb1): Likewise.
11197 (cstoresi_ltu_thu): Likewise.
11198 (thumb1_addsi3_addgeu): Likewise.
11199 (*arm_jump): Likewise.
11200 (*thumb_jump): Likewise.
11201 (*check_arch2): Likewise.
11202 (arm_casesi_internal): Likewise.
11203 (thumb1_casesi_dispatch): Likewise.
11204 (*arm_indirect_jump): Likewise.
11205 (*thumb1_indirect_jump): Likewise.
11207 (*and_scc): Likewise.
11208 (*ior_scc): Likewise.
11209 (*compare_scc): Likewise.
11210 (*cond_move): Likewise.
11211 (*cond_arith): Likewise.
11212 (*cond_sub): Likewise.
11213 (*cmp_ite0): Likewise.
11214 (*cmp_ite1): Likewise.
11215 (*cmp_and): Likewise.
11216 (*cmp_ior): Likewise.
11217 (*ior_scc_scc): Likewise.
11218 (*ior_scc_scc_cmp): Likewise.
11219 (*and_scc_scc): Likewise.
11220 (*and_scc_scc_cmp): Likewise.
11221 (*and_scc_scc_nod): Likewise.
11222 (*negscc): Likewise.
11223 (movcond_addsi): Likewise.
11224 (movcond): Likewise.
11225 (*ifcompare_plus_move): Likewise.
11226 (*if_plus_move): Likewise.
11227 (*ifcompare_move_plus): Likewise.
11228 (*if_move_plus): Likewise.
11229 (*ifcompare_arith_arith): Likewise.
11230 (*if_arith_arith): Likewise.
11231 (*ifcompare_arith_move): Likewise.
11232 (*if_arith_move): Likewise.
11233 (*ifcompare_move_arith): Likewise.
11234 (*if_move_arith): Likewise.
11235 (*ifcompare_move_not): Likewise.
11236 (*if_move_not): Likewise.
11237 (*ifcompare_not_move): Likewise.
11238 (*if_not_move): Likewise.
11239 (*ifcompare_shift_move): Likewise.
11240 (*if_shift_move): Likewise.
11241 (*ifcompare_move_shift): Likewise.
11242 (*if_move_shift): Likewise.
11243 (*ifcompare_shift_shift): Likewise.
11244 (*ifcompare_not_arith): Likewise.
11245 (*ifcompare_arith_not): Likewise.
11246 (*if_arith_not): Likewise.
11247 (*ifcompare_neg_move): Likewise.
11248 (*if_neg_move): Likewise.
11249 (*ifcompare_move_neg): Likewise.
11250 (*if_move_neg): Likewise.
11251 (prologue_thumb1_interwork): Likewise.
11252 (*cond_move_not): Likewise.
11253 (*sign_extract_onebit): Likewise.
11254 (*not_signextract_onebit): Likewise.
11255 (stack_tie): Likewise.
11256 (align_4): Likewise.
11257 (align_8): Likewise.
11258 (consttable_end): Likewise.
11259 (consttable_1): Likewise.
11260 (consttable_2): Likewise.
11261 (consttable_4): Likewise.
11262 (consttable_8): Likewise.
11263 (consttable_16): Likewise.
11264 (*thumb1_tablejump): Likewise.
11265 (prefetch): Likewise.
11266 (force_register_use): Likewise.
11267 (thumb_eh_return): Likewise.
11268 (load_tp_hard): Likewise.
11269 (load_tp_soft): Likewise.
11270 (tlscall): Likewise.
11271 (*arm_movtas_ze): Likewise.
11272 (*arm_rev): Likewise.
11273 (*arm_revsh): Likewise.
11274 (*arm_rev16): Likewise.
11275 * config/arm/thumb2.md
11276 (*thumb2_smaxsi3): Likewise.
11277 (*thumb2_sminsi3): Likewise.
11278 (*thumb32_umaxsi3): Likewise.
11279 (*thumb2_uminsi3): Likewise.
11280 (*thumb2_negdi2): Likewise.
11281 (*thumb2_abssi2): Likewise.
11282 (*thumb2_neg_abss): Likewise.
11283 (*thumb2_movsi_insn): Likewise.
11284 (tls_load_dot_plus_four): Likewise.
11285 (*thumb2_movhi_insn): Likewise.
11286 (*thumb2_mov_scc): Likewise.
11287 (*thumb2_mov_negs): Likewise.
11288 (*thumb2_mov_negs): Likewise.
11289 (*thumb2_mov_nots): Likewise.
11290 (*thumb2_mov_nots): Likewise.
11291 (*thumb2_movsicc_): Likewise.
11292 (*thumb2_movsfcc_soft_insn): Likewise.
11293 (*thumb2_indirect_jump): Likewise.
11294 (*thumb2_and_scc): Likewise.
11295 (*thumb2_ior_scc): Likewise.
11296 (*thumb2_ior_scc_strict_it): Likewise.
11297 (*thumb2_cond_move): Likewise.
11298 (*thumb2_cond_arith): Likewise.
11299 (*thumb2_cond_ari): Likewise.
11300 (*thumb2_cond_sub): Likewise.
11301 (*thumb2_negscc): Likewise.
11302 (*thumb2_movcond): Likewise.
11303 (thumb2_casesi_internal): Likewise.
11304 (thumb2_casesi_internal_pic): Likewise.
11305 (*thumb2_alusi3_short): Likewise.
11306 (*thumb2_mov<mode>_shortim): Likewise.
11307 (*thumb2_addsi_short): Likewise.
11308 (*thumb2_subsi_short): Likewise.
11309 (thumb2_addsi3_compare0): Likewise.
11310 (*thumb2_cbz): Likewise.
11311 (*thumb2_cbnz): Likewise.
11312 (*thumb2_one_cmplsi2_short): Likewise.
11313 (*thumb2_negsi2_short): Likewise.
11314 (*orsi_notsi_si): Likewise.
11315 * config/arm/arm1020e.md: Update with new attributes.
11316 * config/arm/arm1026ejs.md: Update with new attributes.
11317 * config/arm/arm1136jfs.md: Update with new attributes.
11318 * config/arm/arm926ejs.md: Update with new attributes.
11319 * config/arm/cortex-a15.md: Update with new attributes.
11320 * config/arm/cortex-a5.md: Update with new attributes.
11321 * config/arm/cortex-a53.md: Update with new attributes.
11322 * config/arm/cortex-a7.md: Update with new attributes.
11323 * config/arm/cortex-a8.md: Update with new attributes.
11324 * config/arm/cortex-a9.md: Update with new attributes.
11325 * config/arm/cortex-m4.md: Update with new attributes.
11326 * config/arm/cortex-r4.md: Update with new attributes.
11327 * config/arm/fa526.md: Update with new attributes.
11328 * config/arm/fa606te.md: Update with new attributes.
11329 * config/arm/fa626te.md: Update with new attributes.
11330 * config/arm/fa726te.md: Update with new attributes.
11332 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
11334 * config/aarch64/aarch64-simd.md
11335 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
11336 <vwx> iterator to ensure correct register choice.
11337 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
11338 (aarch64_sqdmull_n<mode>): Likewise.
11339 (aarch64_sqdmull2_n<mode>_internal): Likewise.
11340 * config/aarch64/arm_neon.h
11341 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
11342 (vml<as><q>_n_<su>16): Likewise.
11343 (vml<as>l_high_lane<q>_<su>16): Likewise.
11344 (vml<as>l_high_n_<su>16): Likewise.
11345 (vml<as>l_lane<q>_<su>16): Likewise.
11346 (vml<as>l_n_<su>16): Likewise.
11347 (vmul<q>_lane<q>_<su>16): Likewise.
11348 (vmul<q>_n_<su>16): Likewise.
11349 (vmull_lane<q>_<su>16): Likewise.
11350 (vmull_n_<su>16): Likewise.
11351 (vmull_high_lane<q>_<su>16): Likewise.
11352 (vmull_high_n_<su>16): Likewise.
11353 (vqrdmulh<q>_n_s16): Likewise.
11355 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
11357 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
11358 have the correct lane parameter.
11360 2013-09-06 Richard Biener <rguenther@suse.de>
11362 * cfganal.c (control_dependences::~control_dependences):
11363 Properly free all of the vector.
11365 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
11368 * config/i386/i386.c (ix86_conditional_register_usage):
11369 Proper initialize extended SSE registers.
11371 2013-09-06 Jan Hubicka <jh@suse.cz>
11373 PR tree-optimization/58311
11374 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
11376 2013-09-06 Jan Hubicka <jh@suse.cz>
11378 * Makefile.in (tree-sra.o): Update dependencies.
11379 * tree-sra.c: Include ipa-utils.h
11380 (scan_function): Use recursive_call_p.
11381 (has_caller_p): New function.
11382 (cgraph_for_node_and_aliases): Count also callers of aliases.
11384 2013-09-06 Jan Hubicka <jh@suse.cz>
11386 PR middle-end/58094
11387 * cgraph.h (symtab_semantically_equivalent_p): Declare.
11388 * tree-tailcall.c: Include ipa-utils.h.
11389 (find_tail_calls): Use it.
11390 * ipa-pure-const.c (check_call): Likewise.
11391 * ipa-utils.c (recursive_call_p): New function.
11392 * ipa-utils.h (recursive_call_p): Dclare.
11393 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
11394 (symtab_semantically_equivalent_p): New function.
11395 * Makefile.in (tree-tailcall.o): Update dependencies.
11397 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
11399 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
11400 non-inlinable part.
11402 2013-09-06 Richard Biener <rguenther@suse.de>
11404 * lto-streamer.h (lto_global_var_decls): Remove.
11405 * Makefile.in (OBJS): Remove lto-symtab.o.
11406 (lto-symtab.o): Remove.
11407 (GTFILES): Remove lto-symtab.c
11408 * lto-symtab.c: Move to lto/
11410 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11412 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
11413 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
11414 (UNSPEC_FPINT_RINT): New constant definitions.
11415 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
11416 definition with 2 attributes.
11417 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
11418 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
11421 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11423 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
11424 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
11425 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
11426 attribute to "z196_alone".
11427 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
11430 2013-09-06 Richard Biener <rguenther@suse.de>
11432 * basic-block.h (class control_dependences): New.
11433 * tree-ssa-dce.c (control_dependence_map): Remove.
11435 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
11436 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
11437 find_pdom, find_control_dependence, find_all_control_dependences):
11439 (mark_control_dependent_edges_necessary,
11440 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
11441 tree_dce_done, perform_tree_ssa_dce): Adjust.
11442 * cfganal.c (set_control_dependence_map_bit,
11443 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
11444 find_all_control_dependences): Move from tree-ssa-dce.c and
11445 implement as methods of control_dependences class.
11446 (control_dependences::control_dependences): New.
11447 (control_dependences::~control_dependences): Likewise.
11448 (control_dependences::get_edges_dependent_on): Likewise.
11449 (control_dependences::get_edge): Likewise.
11451 2013-09-04 Jan Hubicka <jh@suse.cz>
11453 * tree.c (types_same_for_odr): Drop overactive check.
11454 * ipa-devirt.c (hash_type_name): Likewise.
11456 2013-09-04 Jan Hubicka <jh@suse.cz>
11458 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
11459 (analyze_functions): ... here.
11461 2013-09-04 Jan Hubicka <jh@suse.cz>
11463 PR middle-end/58201
11464 * cgraphunit.c (analyze_functions): Clear AUX fields
11465 after processing; initialize assembler name has.
11467 2013-09-05 Jeff Law <law@redhat.com>
11469 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
11470 from thread_around_empty_block. Record threading path into PATH.
11471 Recurse if threading through the initial block is successful.
11472 (thread_across_edge): Corresponding changes to slightly simplify.
11474 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
11476 * config/aarch64/aarch64.md
11477 (type): Remove frecpe, frecps, frecpx.
11478 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
11479 fix to be a TARGET_SIMD instruction.
11480 (aarch64_frecps): Remove.
11481 * config/aarch64/aarch64-simd.md
11482 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
11483 (aarch64_frecps<mode>): Handle all float/vector of float modes.
11485 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
11486 Sofiane Naci <sofiane.naci@arm.com>
11488 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
11489 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
11490 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
11491 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
11492 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
11493 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
11494 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
11495 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
11496 into "clz, "rbit". Rename "shift" to "shift_imm".
11497 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
11498 changes. Update for attribute changes all occurrences of arlo_* and
11500 * config/arm/arm-fixed.md: Update for attribute changes
11501 all occurrences of arlo_* types.
11502 * config/arm/thumb2.md: Update for attribute changes all occurrences
11504 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
11505 (cortexa7_older_only): Likewise.
11506 (cortexa7_younger): Likewise.
11507 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
11508 (1020alu_shift_op): Likewise.
11509 (1020alu_shift_reg_op): Likewise.
11510 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
11511 (alu_shift_op): Likewise.
11512 (alu_shift_reg_op): Likewise.
11513 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
11514 (11_alu_shift_op): Likewise.
11515 (11_alu_shift_reg_op): Likewise.
11516 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
11517 (9_alu_shift_reg_op): Likewise.
11518 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
11520 (cortex_a15_alu_shift): Likewise.
11521 (cortex_a15_alu_shift_reg): Likewise.
11522 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
11524 (cortex_a5_alu_shift): Likewise.
11525 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
11527 (cortex_a53_alu_shift): Likewise.
11528 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
11530 (cortex_a7_alu_reg): Likewise.
11531 (cortex_a7_alu_shift): Likewise.
11532 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
11534 (cortex_a8_alu_shift): Likewise.
11535 (cortex_a8_alu_shift_reg): Likewise.
11536 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
11537 (cortex_a9_dp_shift): Likewise.
11538 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
11540 * config/arm/cortex-r4.md
11541 (cortex_r4_alu): Update for attribute changes.
11542 (cortex_r4_mov): Likewise.
11543 (cortex_r4_alu_shift_reg): Likewise.
11544 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
11545 (526_alu_shift_op): Likewise.
11546 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
11547 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
11548 (626te_alu_shift_op): Likewise.
11549 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
11550 (726te_alu_shift_op): Likewise.
11551 (726te_alu_shift_reg_op): Likewise.
11552 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
11553 (mp626_alu_shift_op): Likewise.
11554 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
11555 (pj4_alu_conds): Likewise.
11556 (pj4_shift): Likewise.
11557 (pj4_shift_conds): Likewise.
11558 (pj4_alu_shift): Likewise.
11559 (pj4_alu_shift_conds): Likewise.
11560 * config/aarch64/aarch64.md: Update for attribute change
11561 all occurrences of arlo_* and shift* types.
11563 2013-09-05 Mike Stump <mikestump@comcast.net>
11565 * tree.h: Move documentation for tree_function_decl to tree-core.h
11566 with the declaration.
11568 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
11571 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
11572 looking for widest mode.
11574 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
11576 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
11578 2013-09-05 Richard Biener <rguenther@suse.de>
11580 PR tree-optimization/58137
11581 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
11582 Do not create vectors of pointers.
11583 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
11584 types for the components of the vector initializer.
11585 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
11586 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
11588 2013-09-05 Martin Jambor <mjambor@suse.cz>
11590 * ipa-prop.c (remove_described_reference): Accept missing references,
11591 return false if that hppens, otherwise return true.
11592 (cgraph_node_for_jfunc): New function.
11593 (try_decrement_rdesc_refcount): Likewise.
11594 (try_make_edge_direct_simple_call): Use them.
11595 (ipa_edge_removal_hook): Remove references from rdescs.
11596 (ipa_edge_duplication_hook): Clone rdescs and their references
11597 when the new edge has the same caller as the old one.
11598 * cgraph.c (cgraph_resolve_speculation): Remove speculative
11599 reference before removing any edges.
11601 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
11603 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
11605 * thumb2.md (thumb2_storewb_parisi): New pattern.
11607 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
11609 * config/aarch64/aarch64-option-extensions.def: Add
11610 AARCH64_OPT_EXTENSION of 'crc'.
11611 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
11612 (AARCH64_ISA_CRC): Ditto.
11613 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
11614 description of the CRC extension.
11616 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
11618 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
11619 * config/rs6000/linux.h: Ditto.
11620 * alpha/linux.h: Ditto.
11621 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
11622 no_c99_libc_has_function.
11623 * config/c6x/uclinux-elf.h: Ditto.
11624 * config/lm32/uclinux-elf.h: Ditto.
11625 * config/m68k/uclinux.h: Ditto.
11626 * config/moxie/uclinux.h: Ditto.
11627 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
11628 (crisv32-*-linux*, cris-*-linux*): Ditto.
11629 * config/bfin/bfin.c: Include "tm_p.h".
11631 2013-09-05 Richard Biener <rguenther@suse.de>
11633 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
11634 check for a definition without a basic-block.
11636 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
11637 Sofiane Naci <sofiane.naci@arm.com>
11639 * config/aarch64/aarch64.md
11640 (*movti_aarch64): Rename r_2_f and f_2_r.
11641 (*movsf_aarch64): Likewise.
11642 (*movdf_aarch64): Likewise.
11643 (*movtf_aarch64): Likewise.
11644 (aarch64_movdi_<mode>low): Likewise.
11645 (aarch64_movdi_<mode>high): Likewise.
11646 (aarch64_mov<mode>high_di): Likewise.
11647 (aarch64_mov<mode>low_di): Likewise.
11648 (aarch64_movtilow_tilow): Likewise.
11649 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
11650 values to config/arm/types.md
11651 (attribute "conds"): Update for attribute change.
11652 (anddi3_insn): Likewise.
11653 (iordi3_insn): Likewise.
11654 (xordi3_insn): Likewise.
11655 (one_cmpldi2): Likewise.
11656 * config/arm/types.md (type): Add Neon types.
11657 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
11658 use "type" attribute.
11659 (movmisalign<mode>_neon_store): Likewise.
11660 (movmisalign<mode>_neon_load): Likewise.
11661 (vec_set<mode>_internal): Likewise.
11662 (vec_setv2di_internal): Likewise.
11663 (vec_extract<mode>): Likewise.
11664 (vec_extractv2di): Likewise.
11665 (add<mode>3_neon): Likewise.
11666 (adddi3_neon): Likewise.
11667 (sub<mode>3_neon): Likewise.
11668 (subdi3_neon): Likewise.
11669 (mul<mode>3_neon): Likewise.
11670 (mul<mode>3add<mode>_neon): Likewise.
11671 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
11672 (fma<VCVTF:mode>4)): Likewise.
11673 (fma<VCVTF:mode>4_intrinsic): Likewise.
11674 (fmsub<VCVTF:mode>4)): Likewise.
11675 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
11676 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
11677 (ior<mode>3): Likewise.
11678 (and<mode>3): Likewise.
11679 (anddi3_neon): Likewise.
11680 (orn<mode>3_neon): Likewise.
11681 (orndi3_neon): Likewise.
11682 (bic<mode>3_neon): Likewise.
11683 (bicdi3_neon): Likewise.
11684 (xor<mode>3): Likewise.
11685 (one_cmpl<mode>2): Likewise.
11686 (abs<mode>2): Likewise.
11687 (neg<mode>2): Likewise.
11688 (umin<mode>3_neon): Likewise.
11689 (umax<mode>3_neon): Likewise.
11690 (smin<mode>3_neon): Likewise.
11691 (smax<mode>3_neon): Likewise.
11692 (vashl<mode>3): Likewise.
11693 (vashr<mode>3_imm): Likewise.
11694 (vlshr<mode>3_imm): Likewise.
11695 (ashl<mode>3_signed): Likewise.
11696 (ashl<mode>3_unsigned): Likewise.
11697 (neon_load_count): Likewise.
11698 (ashldi3_neon_noclobber): Likewise.
11699 (signed_shift_di3_neon): Likewise.
11700 (unsigned_shift_di3_neon): Likewise.
11701 (ashrdi3_neon_imm_noclobber): Likewise.
11702 (lshrdi3_neon_imm_noclobber): Likewise.
11703 (widen_ssum<mode>3): Likewise.
11704 (widen_usum<mode>3): Likewise.
11705 (quad_halves_<code>v4si): Likewise.
11706 (quad_halves_<code>v4sf): Likewise.
11707 (quad_halves_<code>v8hi): Likewise.
11708 (quad_halves_<code>v16qi): Likewise.
11709 (reduc_splus_v2di): Likewise.
11710 (neon_vpadd_internal<mode>): Likewise.
11711 (neon_vpsmin<mode>): Likewise.
11712 (neon_vpsmax<mode>): Likewise.
11713 (neon_vpumin<mode>): Likewise.
11714 (neon_vpumax<mode>): Likewise.
11715 (ss_add<mode>_neon): Likewise.
11716 (us_add<mode>_neon): Likewise.
11717 (ss_sub<mode>_neon): Likewise.
11718 (us_sub<mode>_neon): Likewise.
11719 (neon_vadd<mode>_unspec): Likewise.
11720 (neon_vaddl<mode>): Likewise.
11721 (neon_vaddw<mode>): Likewise.
11722 (neon_vhadd<mode>): Likewise.
11723 (neon_vqadd<mode>): Likewise.
11724 (neon_vaddhn<mode>): Likewise.
11725 (neon_vmul<mode>): Likewise.
11726 (neon_vmla<mode>): Likewise.
11727 (neon_vmlal<mode>): Likewise.
11728 (neon_vmls<mode>): Likewise.
11729 (neon_vmlsl<mode>): Likewise.
11730 (neon_vqdmulh<mode>): Likewise.
11731 (neon_vqdmlal<mode>): Likewise.
11732 (neon_vqdmlsl<mode>): Likewise.
11733 (neon_vmull<mode>): Likewise.
11734 (neon_vqdmull<mode>): Likewise.
11735 (neon_vsub<mode>_unspec): Likewise.
11736 (neon_vsubl<mode>): Likewise.
11737 (neon_vsubw<mode>): Likewise.
11738 (neon_vqsub<mode>): Likewise.
11739 (neon_vhsub<mode>): Likewise.
11740 (neon_vsubhn<mode>): Likewise.
11741 (neon_vceq<mode>): Likewise.
11742 (neon_vcge<mode>): Likewise.
11743 (neon_vcgeu<mode>): Likewise.
11744 (neon_vcgt<mode>): Likewise.
11745 (neon_vcgtu<mode>): Likewise.
11746 (neon_vcle<mode>): Likewise.
11747 (neon_vclt<mode>): Likewise.
11748 (neon_vcage<mode>): Likewise.
11749 (neon_vcagt<mode>): Likewise.
11750 (neon_vtst<mode>): Likewise.
11751 (neon_vabd<mode>): Likewise.
11752 (neon_vabdl<mode>): Likewise.
11753 (neon_vaba<mode>): Likewise.
11754 (neon_vabal<mode>): Likewise.
11755 (neon_vmax<mode>): Likewise.
11756 (neon_vmin<mode>): Likewise.
11757 (neon_vpaddl<mode>): Likewise.
11758 (neon_vpadal<mode>): Likewise.
11759 (neon_vpmax<mode>): Likewise.
11760 (neon_vpmin<mode>): Likewise.
11761 (neon_vrecps<mode>): Likewise.
11762 (neon_vrsqrts<mode>): Likewise.
11763 (neon_vqabs<mode>): Likewise.
11764 (neon_vqneg<mode>): Likewise.
11765 (neon_vcls<mode>): Likewise.
11766 (clz<mode>2): Likewise.
11767 (popcount<mode>2): Likewise.
11768 (neon_vrecpe): Likewise.
11769 (neon_vrsqrte): Likewise.
11770 (neon_vget_lane<mode>_sext_internal): Likewise.
11771 (neon_vget_lane<mode>_zext_internal): Likewise.
11772 (neon_vdup_n<mode>): Likewise.
11773 (neon_vdup_nv2di): Likewise.
11774 (neon_vdpu_lane<mode>_internal): Likewise.
11775 (neon_vswp<mode>): Likewise.
11776 (float<mode><V_cvtto>2): Likewise.
11777 (floatuns<mode><V_cvtto>2): Likewise.
11778 (fix_trunc<mode><V_cvtto>)2): Likewise
11779 (fixuns_trunc<mode><V_cvtto)2): Likewise.
11780 (neon_vcvt<mode>): Likewise.
11781 (neon_vcvtv4sfv4hf): Likewise.
11782 (neon_vcvtv4hfv4sf): Likewise.
11783 (neon_vcvt_n<mode>): Likewise.
11784 (neon_vmovn<mode>): Likewise.
11785 (neon_vqmovn<mode>): Likewise.
11786 (neon_vqmovun<mode>): Likewise.
11787 (neon_vmovl<mode>): Likewise.
11788 (neon_vmul_lane<mode>): Likewise.
11789 (neon_vmull_lane<mode>): Likewise.
11790 (neon_vqdmull_lane<mode>): Likewise.
11791 (neon_vqdmulh_lane<mode>): Likewise.
11792 (neon_vmla_lane<mode>): Likewise.
11793 (neon_vmlal_lane<mode>): Likewise.
11794 (neon_vqdmlal_lane<mode>): Likewise.
11795 (neon_vmls_lane<mode>): Likewise.
11796 (neon_vmlsl_lane<mode>): Likewise.
11797 (neon_vqdmlsl_lane<mode>): Likewise.
11798 (neon_vext<mode>): Likewise.
11799 (neon_vrev64<mode>): Likewise.
11800 (neon_vrev32<mode>): Likewise.
11801 (neon_vrev16<mode>): Likewise.
11802 (neon_vbsl<mode>_internal): Likewise.
11803 (neon_vshl<mode>): Likewise.
11804 (neon_vqshl<mode>): Likewise.
11805 (neon_vshr_n<mode>): Likewise.
11806 (neon_vshrn_n<mode>): Likewise.
11807 (neon_vqshrn_n<mode>): Likewise.
11808 (neon_vqshrun_n<mode>): Likewise.
11809 (neon_vshl_n<mode>): Likewise.
11810 (neon_vqshl_n<mode>): Likewise.
11811 (neon_vqshlu_n<mode>): Likewise.
11812 (neon_vshll_n<mode>): Likewise.
11813 (neon_vsra_n<mode>): Likewise.
11814 (neon_vsri_n<mode>): Likewise.
11815 (neon_vsli_n<mode>): Likewise.
11816 (neon_vtbl1v8qi): Likewise.
11817 (neon_vtbl2v8qi): Likewise.
11818 (neon_vtbl3v8qi): Likewise.
11819 (neon_vtbl4v8qi): Likewise.
11820 (neon_vtbx1v8qi): Likewise.
11821 (neon_vtbx2v8qi): Likewise.
11822 (neon_vtbx3v8qi): Likewise.
11823 (neon_vtbx4v8qi): Likewise.
11824 (neon_vtrn<mode>_internal): Likewise.
11825 (neon_vzip<mode>_internal): Likewise.
11826 (neon_vuzp<mode>_internal): Likewise.
11827 (neon_vld1<mode>): Likewise.
11828 (neon_vld1_lane<mode>): Likewise.
11829 (neon_vld1_dup<mode>): Likewise.
11830 (neon_vld1_dupv2di): Likewise.
11831 (neon_vst1<mode>): Likewise.
11832 (neon_vst1_lane<mode>): Likewise.
11833 (neon_vld2<mode>): Likewise.
11834 (neon_vld2_lane<mode>): Likewise.
11835 (neon_vld2_dup<mode>): Likewise.
11836 (neon_vst2<mode>): Likewise.
11837 (neon_vst2_lane<mode>): Likewise.
11838 (neon_vld3<mode>): Likewise.
11839 (neon_vld3qa<mode>): Likewise.
11840 (neon_vld3qb<mode>): Likewise.
11841 (neon_vld3_lane<mode>): Likewise.
11842 (neon_vld3_dup<mode>): Likewise.
11843 (neon_vst3<mode>): Likewise.
11844 (neon_vst3qa<mode>): Likewise.
11845 (neon_vst3qb<mode>): Likewise.
11846 (neon_vst3_lane<mode>): Likewise.
11847 (neon_vld4<mode>): Likewise.
11848 (neon_vld4qa<mode>): Likewise.
11849 (neon_vld4qb<mode>): Likewise.
11850 (neon_vld4_lane<mode>): Likewise.
11851 (neon_vld4_dup<mode>): Likewise.
11852 (neon_vst4<mode>): Likewise.
11853 (neon_vst4qa<mode>): Likewise.
11854 (neon_vst4qb<mode>): Likewise.
11855 (neon_vst4_lane<mode>): Likewise.
11856 (neon_vec_unpack<US>_lo_<mode>): Likewise.
11857 (neon_vec_unpack<US>_hi_<mode>): Likewise.
11858 (neon_vec_<US>mult_lo_<mode>): Likewise.
11859 (neon_vec_<US>mult_hi_<mode>): Likewise.
11860 (neon_vec_<US>shiftl_<mode>): Likewise.
11861 (neon_unpack<US>_<mode>): Likewise.
11862 (neon_vec_<US>mult_<mode>): Likewise.
11863 (vec_pack_trunc_<mode>): Likewise.
11864 (neon_vec_pack_trunk_<mode>): Likewise.
11865 (neon_vabd<mode>_2): Likewise.
11866 (neon_vabd<mode>_3): Likewise.
11867 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
11868 (thumb2_movsi_vfp): Likewise.
11869 (movdi_vfp): Likewise.
11870 (movdi_vfp_cortexa8): Likewise.
11871 (movhf_vfp_neon): Likewise.
11872 (movhf_vfp): Likewiwse.
11873 (movsf_vfp): Likewiwse.
11874 (thumb2_movsf_vfp): Likewiwse.
11875 (movdf_vfp): Likewise.
11876 (thumb2_movdf_vfp): Likewise.
11877 (movsfcc_vfp): Likewise.
11878 (thumb2_movsfcc_vfp): Likewise.
11879 (movdfcc_vfp): Likewise.
11880 (thumb2_movdfcc_vfp): Likewise.
11881 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
11882 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
11883 (v10_v2c): Likewise.
11884 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
11886 (cortex_a15_neon_int_2): Likewise.
11887 (cortex_a15_neon_int_3): Likewise.
11888 (cortex_a15_neon_int_4): Likewise.
11889 (cortex_a15_neon_int_5): Likewise.
11890 (cortex_a15_neon_vqneg_vqabs): Likewise.
11891 (cortex_a15_neon_vmov): Likewise.
11892 (cortex_a15_neon_vaba): Likewise.
11893 (cortex_a15_neon_vaba_qqq): Likewise.
11894 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
11895 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
11896 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
11897 scalar_64_32_long_scalar): Likewise.
11898 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
11899 (cortex_a15_neon_mla_qqq_8_16): Likewise.
11900 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
11901 lotype_qdd_64_32_long): Likewise.
11902 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
11903 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
11904 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
11905 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
11906 (cortex_a15_neon_shift_1): Likewise.
11907 (cortex_a15_neon_shift_2): Likewise.
11908 (cortex_a15_neon_shift_3): Likewise.
11909 (cortex_a15_neon_vshl_ddd): Likewise.
11910 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
11911 (cortex_a15_neon_vsra_vrsra): Likewise.
11912 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
11913 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
11914 (cortex_a15_neon_fp_vmul_ddd): Likewise.
11915 (cortex_a15_neon_fp_vmul_qqd): Likewise.
11916 (cortex_a15_neon_fp_vmla_ddd): Likewise.
11917 (cortex_a15_neon_fp_vmla_qqq): Likewise.
11918 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
11919 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
11920 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
11921 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
11922 (cortex_a15_neon_bp_simple): Likewise.
11923 (cortex_a15_neon_bp_2cycle): Likewise.
11924 (cortex_a15_neon_bp_3cycle): Likewise.
11925 (cortex_a15_neon_vld1_1_2_regs): Likewise.
11926 (cortex_a15_neon_vld1_3_4_regs): Likewise.
11927 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
11928 (cortex_a15_neon_vld2_4_regs): Likewise.
11929 (cortex_a15_neon_vld3_vld4): Likewise.
11930 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
11931 (cortex_a15_neon_vst1_3_4_regs): Likewise.
11932 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
11933 (cortex_a15_neon_vst3_vst4): Likewise.
11934 (cortex_a15_neon_vld1_vld2_lane): Likewise.
11935 (cortex_a15_neon_vld3_vld4_lane" 10
11936 (cortex_a15_neon_vst1_vst2_lane): Likewise.
11937 (cortex_a15_neon_vst3_vst4_lane): Likewise.
11938 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
11939 (cortex_a15_neon_ldm_2): Likewise.0
11940 (cortex_a15_neon_stm_2): Likewise.
11941 (cortex_a15_neon_mcr): Likewise.
11942 (cortex_a15_neon_mcr_2_mcrr): Likewise.
11943 (cortex_a15_neon_mrc): Likewise.
11944 (cortex_a15_neon_mrrc): Likewise.
11945 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
11947 (cortex_a15_alu_shift): Likewise.
11948 (cortex_a15_alu_shift_reg): Likewise.
11949 (cortex_a15_mult32): Likewise.
11950 (cortex_a15_mult64): Likewise.
11951 (cortex_a15_block): Likewise.
11952 (cortex_a15_branch): Likewise.
11953 (cortex_a15_load1): Likewise.
11954 (cortex_a15_load3): Likewise.
11955 (cortex_a15_store1): Likewise.
11956 (cortex_a15_store3): Likewise.
11957 (cortex_a15_call): Likewise.
11958 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
11959 (cortex_a5_f2r): Likewise.
11960 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
11962 (cortex_a53_f2r): Likewise.
11963 * config/arm/cortex-a7.md
11964 (cortex_a7_branch): Update for attribute change.
11965 (cortex_a7_call): Likewise.
11966 (cortex_a7_alu_imm): Likewise.
11967 (cortex_a7_alu_reg): Likewise.
11968 (cortex_a7_alu_shift): Likewise.
11969 (cortex_a7_mul): Likewise.
11970 (cortex_a7_load1): Likewise.
11971 (cortex_a7_store1): Likewise.
11972 (cortex_a7_load2): Likewise.
11973 (cortex_a7_store2): Likewise.
11974 (cortex_a7_load3): Likewise.
11975 (cortex_a7_store3): Likewise.
11976 (cortex_a7_load4): Likewise.
11977 (cortex_a7_store4): Likewise.
11978 (cortex_a7_fpalu): Likewise.
11979 (cortex_a7_fconst): Likewise.
11980 (cortex_a7_fpmuls): Likewise.
11981 (cortex_a7_neon_mul): Likewise.
11982 (cortex_a7_fpmacs): Likewise.
11983 (cortex_a7_neon_mla: Likewise.
11984 (cortex_a7_fpmuld: Likewise.
11985 (cortex_a7_fpmacd: Likewise.
11986 (cortex_a7_fpfmad: Likewise.
11987 (cortex_a7_fdivs: Likewise.
11988 (cortex_a7_fdivd: Likewise.
11989 (cortex_a7_r2f: Likewise.
11990 (cortex_a7_f2r: Likewise.
11991 (cortex_a7_f_flags: Likewise.
11992 (cortex_a7_f_loads: Likewise.
11993 (cortex_a7_f_loadd: Likewise.
11994 (cortex_a7_f_stores: Likewise.
11995 (cortex_a7_f_stored: Likewise.
11996 (cortex_a7_neon): Likewise.
11997 * config/arm/cortex-a8-neon.md
11998 (cortex_a8_neon_mrc): Update for attribute change.
11999 (cortex_a8_neon_mrrc): Likewise.
12000 (cortex_a8_neon_int_1): Likewise.
12001 (cortex_a8_neon_int_2): Likewise.
12002 (cortex_a8_neon_int_3): Likewise.
12003 (cortex_a8_neon_int_4): Likewise.
12004 (cortex_a8_neon_int_5): Likewise.
12005 (cortex_a8_neon_vqneg_vqabs): Likewise.
12006 (cortex_a8_neon_vmov): Likewise.
12007 (cortex_a8_neon_vaba): Likewise.
12008 (cortex_a8_neon_vaba_qqq): Likewise.
12009 (cortex_a8_neon_vsma): Likewise.
12010 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
12011 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
12012 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
12013 long_scalar): Likewise.
12014 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
12015 (cortex_a8_neon_mla_qqq_8_16): Likewise.
12016 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
12017 long_scalar_qdd_64_32_long): Likewise.
12018 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
12019 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
12020 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
12021 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
12022 (cortex_a8_neon_shift_1): Likewise.
12023 (cortex_a8_neon_shift_2): Likewise.
12024 (cortex_a8_neon_shift_3): Likewise.
12025 (cortex_a8_neon_vshl_ddd): Likewise.
12026 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
12027 (cortex_a8_neon_vsra_vrsra): Likewise.
12028 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
12029 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
12030 (cortex_a8_neon_fp_vsum): Likewise.
12031 (cortex_a8_neon_fp_vmul_ddd): Likewise.
12032 (cortex_a8_neon_fp_vmul_qqd): Likewise.
12033 (cortex_a8_neon_fp_vmla_ddd): Likewise.
12034 (cortex_a8_neon_fp_vmla_qqq): Likewise.
12035 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
12036 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
12037 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
12038 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
12039 (cortex_a8_neon_bp_simple): Likewise.
12040 (cortex_a8_neon_bp_2cycle): Likewise.
12041 (cortex_a8_neon_bp_3cycle): Likewise.
12042 (cortex_a8_neon_ldr): Likewise.
12043 (cortex_a8_neon_str): Likewise.
12044 (cortex_a8_neon_vld1_1_2_regs): Likewise.
12045 (cortex_a8_neon_vld1_3_4_regs): Likewise.
12046 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
12047 (cortex_a8_neon_vld2_4_regs): Likewise.
12048 (cortex_a8_neon_vld3_vld4): Likewise.
12049 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
12050 (cortex_a8_neon_vst1_3_4_regs): Likewise.
12051 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
12052 (cortex_a8_neon_vst3_vst4): Likewise.
12053 (cortex_a8_neon_vld1_vld2_lane): Likewise.
12054 (cortex_a8_neon_vld3_vld4_lane): Likewise.
12055 (cortex_a8_neon_vst1_vst2_lane): Likewise.
12056 (cortex_a8_neon_vst3_vst4_lane): Likewise.
12057 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
12058 (cortex_a8_neon_mcr): Likewise.
12059 (cortex_a8_neon_mcr_2_mcrr): Likewise.
12060 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
12061 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
12063 (ca9_neon_mrrc): Likewise.
12064 (cortex_a9_neon_int_1): Likewise.
12065 (cortex_a9_neon_int_2): Likewise.
12066 (cortex_a9_neon_int_3): Likewise.
12067 (cortex_a9_neon_int_4): Likewise.
12068 (cortex_a9_neon_int_5): Likewise.
12069 (cortex_a9_neon_vqneg_vqabs): Likewise.
12070 (cortex_a9_neon_vmov): Likewise.
12071 (cortex_a9_neon_vaba): Likewise.
12072 (cortex_a9_neon_vaba_qqq): Likewise.
12073 (cortex_a9_neon_vsma): Likewise.
12074 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
12075 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
12076 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
12077 long_scalar): Likewise.
12078 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
12079 (cortex_a9_neon_mla_qqq_8_16): Likewise.
12080 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
12081 long_scalar_qdd_64_32_long): Likewise.
12082 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
12083 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
12084 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
12085 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
12086 (cortex_a9_neon_shift_1): Likewise.
12087 (cortex_a9_neon_shift_2): Likewise.
12088 (cortex_a9_neon_shift_3): Likewise.
12089 (cortex_a9_neon_vshl_ddd): Likewise.
12090 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
12091 (cortex_a9_neon_vsra_vrsra): Likewise.
12092 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
12093 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
12094 (cortex_a9_neon_fp_vsum): Likewise.
12095 (cortex_a9_neon_fp_vmul_ddd): Likewise.
12096 (cortex_a9_neon_fp_vmul_qqd): Likewise.
12097 (cortex_a9_neon_fp_vmla_ddd): Likewise.
12098 (cortex_a9_neon_fp_vmla_qqq): Likewise.
12099 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
12100 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
12101 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
12102 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
12103 (cortex_a9_neon_bp_simple): Likewise.
12104 (cortex_a9_neon_bp_2cycle): Likewise.
12105 (cortex_a9_neon_bp_3cycle): Likewise.
12106 (cortex_a9_neon_ldr): Likewise.
12107 (cortex_a9_neon_str): Likewise.
12108 (cortex_a9_neon_vld1_1_2_regs): Likewise.
12109 (cortex_a9_neon_vld1_3_4_regs): Likewise.
12110 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
12111 (cortex_a9_neon_vld2_4_regs): Likewise.
12112 (cortex_a9_neon_vld3_vld4): Likewise.
12113 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
12114 (cortex_a9_neon_vst1_3_4_regs): Likewise.
12115 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
12116 (cortex_a9_neon_vst3_vst4): Likewise.
12117 (cortex_a9_neon_vld1_vld2_lane): Likewise.
12118 (cortex_a9_neon_vld3_vld4_lane): Likewise.
12119 (cortex_a9_neon_vst1_vst2_lane): Likewise.
12120 (cortex_a9_neon_vst3_vst4_lane): Likewise.
12121 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
12122 (cortex_a9_neon_mcr): Likewise.
12123 (cortex_a9_neon_mcr_2_mcrr): Likewise.
12124 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
12125 (cortex_a9_fps): Likewise.
12126 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
12128 (cortex_m4_fmuls): Likewise.
12129 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
12131 (cortex_r4_mrc): Likewise.
12132 * config/arm/iterators.md: Update comment referring to neon_type.
12133 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
12134 (iwmmxt_movsi_insn): Likewise.
12135 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
12137 (pj4_core_to_vfp): Likewise.
12138 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
12140 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
12141 (vfp_fstore): Likewise.
12142 * doc/md.texi: Change references to neon_type to refer to type.
12144 2013-09-04 Dodji Seketeli <dodji@redhat.com>
12146 * tree.h (DECL_BUILT_IN): Fix typo in comment.
12148 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
12150 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
12151 lglobl if not weak.
12153 2013-09-04 Easwaran Raman <eraman@google.com>
12155 PR middle-end/57370
12156 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
12157 (build_and_add_sum): Use it.
12158 (appears_later_in_bb): Simplify code.
12160 2013-09-04 Teresa Johnson <tejohnson@google.com>
12162 * dumpfile.c (dump_finish): Don't close stderr/stdout.
12164 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
12166 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
12168 2013-09-04 Jan Hubicka <jh@suse.cz>
12170 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
12171 * ipa-devirt.c: Include diganostic.h
12172 (odr_type_d): Add types and types_set.
12173 (hash_type_name): Work for types with vtables during LTO.
12174 (odr_hasher::remove): Fix comment; destroy types_set.
12175 (add_type_duplicate): New function,
12176 (get_odr_type): Use it.
12177 (dump_type_inheritance_graph): Dump type duplicates.
12178 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
12180 * tree.c (types_same_for_odr): Give exact answers on types with
12183 2013-09-04 Dodji Seketeli <dodji@redhat.com>
12185 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
12186 explaining their differences.
12188 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
12190 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
12192 2013-09-03 Jeff Law <law@redhat.com>
12194 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
12195 when not threading through a joiner block. Pass joiner/no joiner
12196 state to register_jump_thread.
12197 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
12198 state from argument rather than implying on path length.
12199 Dump the entire jump thread path into debugging dump.
12200 * tree-flow.h (register_jump_thread): Update prototype.
12202 2013-08-29 Xinliang David Li <davidxl@google.com>
12204 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
12205 Remove a trivial gcc_assert.
12207 2013-08-29 Xinliang David Li <davidxl@google.com>
12209 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
12210 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
12211 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
12212 Delay base decl alignment adjustment.
12213 * tree-vectorizer.c (vect_destroy_datarefs): New function.
12214 * tree-vectorizer.h: New data structure.
12215 (set_dr_misalignment): New function.
12216 (dr_misalignment): Ditto.
12217 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
12218 (vectorizable_load): Ditto.
12219 (ensure_base_align): New function.
12220 (vectorize_loops): Add dbg_cnt support.
12221 (execute_vect_slp): Ditto.
12222 * dbgcnt.def: New debug counter.
12223 * Makefile: New dependency.
12225 2013-09-03 Meador Inge <meadori@codesourcery.com>
12229 2013-08-30 Meador Inge <meadori@codesourcery.com>
12231 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
12233 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
12235 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
12236 function descriptor.
12238 2013-09-03 Richard Biener <rguenther@suse.de>
12240 * tree-affine.c (add_elt_to_tree): Fix association issue,
12241 avoid useless converts and make sure to always return a
12242 properly typed result.
12244 2013-09-03 Richard Biener <rguenther@suse.de>
12246 PR middle-end/57656
12247 * fold-const.c (negate_expr_p): Fix division case.
12248 (negate_expr): Likewise.
12250 2013-09-03 Richard Biener <rguenther@suse.de>
12253 * tree-streamer-out.c: Include tm.h.
12254 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
12256 2013-09-03 Jan Hubicka <jh@suse.cz>
12258 * tree-profile.c (tree_profiling): Cleanup CFG when done.
12260 2013-09-03 Alan Modra <amodra@gmail.com>
12262 * config.gcc (powerpc*-*-linux*): Add support for little-endian
12263 multilibs to big-endian target and vice versa.
12264 * config/rs6000/t-linux64: Use := assignment on all vars.
12265 (MULTILIB_EXTRA_OPTS): Remove fPIC.
12266 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
12267 * config/rs6000/t-linux64le: New file.
12268 * config/rs6000/t-linux64bele: New file.
12269 * config/rs6000/t-linux64lebe: New file.
12271 2013-09-02 Jan Hubicka <jh@suse.cz>
12273 * ipa-inline-transform.c (inline_transform): Do not
12274 optimize_inline_calls when not optimizing.
12276 2013-09-02 Jan Hubicka <jh@suse.cz>
12278 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
12279 duplicated nodes for assembler names.
12280 * symtab.c (symtab_unregister_node): Do not attempt to unlink
12281 hard registers from assembler name hash.
12283 2013-09-02 Jan Hubicka <jh@suse.cz>
12285 * ipa-split.c (execute_split_functions): Split externally visible
12286 functions called once.
12288 2013-09-02 Martin Jambor <mjambor@suse.cz>
12291 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
12292 linked list. When finding the correct duplicate, also consider also
12293 the caller in additon to its inlined_to node.
12295 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
12297 * config/aarch64/aarch64-simd-builtins.def
12298 (dup_lane_scalar): Remove.
12299 * config/aarch64/aarch64-simd.md
12300 (aarch64_simd_dup): Add 'w->w' alternative.
12301 (aarch64_dup_lane<mode>): Allow for VALL.
12302 (aarch64_dup_lane_scalar<mode>): Remove.
12303 (aarch64_dup_lane_<vswap_width_name><mode>): New.
12304 (aarch64_get_lane_signed<mode>): Add w->w altenative.
12305 (aarch64_get_lane_unsigned<mode>): Likewise.
12306 (aarch64_get_lane<mode>): Likewise.
12307 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
12308 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
12309 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
12310 (VCON): Change container of V2SF.
12311 (vswap_width_name): Likewise.
12312 * config/aarch64/arm_neon.h
12313 (__aarch64_vdup_lane_any): New.
12314 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
12315 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
12316 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
12318 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
12320 PR middle-end/56382
12321 * expr.c (emit_move_complex): Do not move complex FP values as parts if
12322 the source or the destination is a single hard register.
12324 2013-09-02 Richard Biener <rguenther@suse.de>
12326 PR middle-end/57511
12327 * tree-scalar-evolution.c (instantiate_scev_name): Allow
12330 2013-09-02 Richard Biener <rguenther@suse.de>
12332 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
12333 arithmetic to sizetype.
12335 2013-09-02 Bin Cheng <bin.cheng@arm.com>
12337 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
12338 Find auto-increment use both before and after candidate.
12340 2013-09-02 Marek Polacek <polacek@redhat.com>
12342 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
12344 2013-09-01 Jan Hubicka <jh@suse.cz>
12346 * Makefile.in: Add ipa-profile.o
12347 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
12348 * cgraph.c (struct cgraph_propagate_frequency_data,
12349 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
12350 ipa-profile.c; replace cgraph_ by ipa_ prefix.
12351 * cgraph.h (cgraph_propagate_frequency): Remove.
12352 * ipa-inline-analysis.c: Include ipa-utils.h;
12353 drop duplicated cfgloop.h.
12354 (inline_update_callee_summaries): Update.
12355 * ipa-profile.c: New file.
12356 * ipa-utils.h (ipa_propagate_frequency): Declare.
12357 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
12358 data-streamer.h, value-prof.h.
12359 (symtab_remove_unreachable_nodes): Update profile.
12360 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
12361 account_time_size, cmp_counts, dump_histogram,
12362 ipa_profile_generate_summary, ipa_profile_write_summary,
12363 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
12364 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
12365 Move to ipa-profile.c.
12367 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
12369 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
12371 2013-09-01 Jan Hubicka <jh@suse.cz>
12373 * common.opt (fdevirtualize-speculatively): New function.
12374 * invoke.texi (fdevirtualize-speculatively): Document.
12375 * ipa-devirt.c: Include ipa-inline.h
12376 (likely_target_p): New function.
12377 (ipa_devirt): New function.
12378 (gate_ipa_devirt): New function.
12379 (pass_data_ipa_devirt): New static var.
12380 (pass_ipa_devirt): Likewise.
12381 (make_pass_ipa_devirt): New function.
12382 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
12383 (common_handle_option): Disable devirtualization when
12384 value range profiling is available.
12385 * passes.def (pass_ipa_devirt): Add.
12386 * timever.def (TV_IPA_DEVIRT): New timevar.
12387 * tree-pass.h (make_pass_ipa_devirt):
12389 2013-09-01 Iain Sandoe <iain@codesourcery.com>
12391 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
12392 include sanitize(undefined).
12394 2013-08-31 Diego Novillo <dnovillo@google.com>
12396 * Makefile.in (TREE_CORE_H): Define.
12398 (GTFILES): Add tree-core.h.
12399 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
12401 * tree-core.h: New file.
12402 Move all data structures, enum, typedefs, global
12403 declarations and constants from ...
12404 * tree.h: ... here.
12406 2013-08-31 Jan Hubicka <jh@suse.cz>
12408 * bulitins.c (expand_builtin): Do not early exit for gcov
12409 instrumented functions.
12411 2013-08-31 Marek Polacek <polacek@redhat.com>
12413 * ubsan.c: Include tm_p.h.
12415 2013-08-31 Jan Hubicka <jh@suse.cz>
12417 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
12420 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
12421 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
12422 * tree-cfg.c (verify_gimple_label): ... here.
12423 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
12424 (ipa_merge_profiles): New function.
12425 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
12426 (lto_input_function_body): Likewise.
12427 * ipa-utils.h (ipa_merge_profiles): Declare.
12428 * lto-streamer.h (lto_input_function_body): Update prototype.
12429 (emit_label_in_global_context_p): Remove.
12430 * lto-symtab.c: Include ipa-utils.h
12431 (lto_cgraph_replace_node): Use ipa_merge_profiles.
12433 2013-08-31 Jan Hubicka <jh@suse.cz>
12435 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
12437 2013-08-31 Jan Hubicka <jh@suse.cz>
12439 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
12441 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
12443 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
12444 "cmp" RTX before signed_comparison_operator check to account
12445 for "code" changes.
12447 2013-08-30 Jan Hubicka <jh@suse.cz>
12449 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
12450 (detect_type_change_1): Rename to ...
12451 (detect_type_change): ... this one; early return on non-polymorphic
12453 (detect_type_change_ssa): Add comp_type parameter; update
12454 use of detect_type_change.
12455 (compute_complex_assign_jump_func): Add param_type parameter;
12456 update use of detect_type_change_ssa.
12457 (compute_complex_ancestor_jump_func): Likewise.
12458 (ipa_get_callee_param_type): New function.
12459 (ipa_compute_jump_functions_for_edge): Compute parameter type;
12460 update calls to the jump function computation functions.
12462 2013-08-30 Teresa Johnson <tejohnson@google.com>
12463 Steven Bosscher <steven@gcc.gnu.org>
12465 * cfgrtl.c (fixup_new_cold_bb): New routine.
12466 (commit_edge_insertions): Invoke fixup_partitions.
12467 (find_partition_fixes): New routine.
12468 (fixup_partitions): Ditto.
12469 (verify_hot_cold_block_grouping): Update comments.
12470 (rtl_verify_edges): Invoke find_partition_fixes.
12471 (rtl_verify_bb_pointers): Update comments.
12472 (rtl_verify_bb_layout): Ditto.
12473 * basic-block.h (probably_never_executed_edge_p): Declare.
12474 (fixup_partitions): Ditto.
12475 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
12476 * bb-reorder.c (sanitize_hot_paths): New function.
12477 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
12478 sanitize_hot_paths.
12479 * predict.c (probably_never_executed_edge_p): New routine.
12480 * cfg.c (check_bb_profile): Add partition insanity warnings.
12482 2013-08-30 Meador Inge <meadori@codesourcery.com>
12484 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
12486 2013-08-30 Marek Polacek <polacek@redhat.com>
12488 * Makefile.in (ubsan.o): Add.
12489 (c-family/c-ubsan.o): Add.
12490 (builtins.o): Add ubsan.h dependency.
12491 * ubsan.h: New file.
12492 * ubsan.c: New file.
12493 * common.opt: Add -fsanitize=undefined option.
12494 (flag_sanitize): Add variable.
12495 (fsanitize=): Add option. Add Driver.
12496 (fsanitize=thread): Remove option.
12497 (fsanitize=address): Likewise.
12498 (static-libubsan): New option.
12499 * doc/invoke.texi: Document the new flag and -static-libubsan.
12500 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
12501 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
12502 * builtin-attrs.def (ATTR_COLD): Define.
12503 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
12504 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
12505 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
12506 * flag-types.h (sanitize_code): New enum.
12507 * opts.c (common_handle_option): Parse command line arguments
12508 of -fsanitize=. Add -fsanitize=unreachable option.
12509 * varasm.c (get_variable_section): Adjust.
12510 (assemble_noswitch_variable): Likewise.
12511 (assemble_variable): Likewise.
12512 (output_constant_def_contents): Likewise.
12513 (categorize_decl_for_section): Likewise.
12514 (place_block_symbol): Likewise.
12515 (output_object_block): Likewise.
12516 * builtins.def: Likewise.
12517 * toplev.c (compile_file): Likewise.
12518 (process_options): Likewise.
12519 * cppbuiltin.c: Likewise.
12520 * tsan.c (tsan_pass): Likewise.
12521 (tsan_gate): Likewise.
12522 (tsan_gate_O0): Likewise.
12523 * cfgexpand.c (partition_stack_vars): Likewise.
12524 (expand_stack_vars): Likewise.
12525 (defer_stack_allocation): Likewise.
12526 (expand_used_vars): Likewise.
12527 * cfgcleanup.c (old_insns_match_p): Likewise.
12528 * asan.c (asan_finish_file): Likewise.
12529 (asan_instrument): Likewise.
12530 (gate_asan): Likewise.
12531 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
12532 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
12533 (asan_global_struct): Use pointer_sized_int_node instead
12534 calling build_nonstandard_integer_type.
12535 (initialize_sanitizer_builtins): Likewise.
12536 (asan_finish_file): Likewise.
12537 * gcc.c: Document %{%:function(args):X}.
12538 (static_spec_functions): Add sanitize.
12539 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
12540 store funcval != NULL there.
12541 (do_spec_1): Adjust handle_spec_function caller.
12542 (handle_braces): Allow %:function(args) as condition.
12543 (sanitize_spec_function): New function.
12544 (ADD_STATIC_LIBUBSAN_LIBS): Define.
12545 (LIBUBSAN_SPEC): Likewise.
12546 (LIBUBSAN_EARLY_SPEC): Likewise.
12547 (SANITIZER_SPEC): Handle libubsan.
12548 (SANITIZER_EARLY_SPEC): Likewise.
12549 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
12550 instead of fsanitize=address.
12551 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
12552 instead of fsanitize=address*.
12553 * builtins.c: Include ubsan.h.
12554 (fold_builtin_0): Instrument __builtin_unreachable.
12555 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
12556 instead of flag_asan.
12557 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
12558 (pointer_sized_int_node): Define.
12559 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
12561 2013-08-30 Mike Stump <mikestump@comcast.net>
12563 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
12566 2013-08-29 Jan Hubicka <jh@suse.cz>
12568 * cgraph.c (cgraph_function_body_availability): Handle weakref
12570 * passes.def: Remove pass_fixup_cfg.
12571 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
12572 track when we need to remove functions.
12573 (gate_ipa_inline): Execute inlining always; add comment why.
12574 (pass_data_ipa_inline): Remove TODO_remove_functions.
12575 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
12576 do not produce summaries.
12577 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
12578 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
12579 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
12580 constant pool and vtable.
12582 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
12584 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
12585 New arm_neon.h's internal macros to specify 64-bit constants.
12586 Avoid using stdint.h's macros.
12588 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
12590 * recog.c (verify_changes): Verify that changes[i].old is non-zero
12591 before applying REG_P.
12593 2013-08-30 Jakub Jelinek <jakub@redhat.com>
12595 PR tree-optimization/58277
12596 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
12597 after seeing too many stmts with vdef in between dombb and current
12598 bb, invalidate everything.
12600 2013-08-30 Richard Biener <rguenther@suse.de>
12602 * fold-const.c (fold_single_bit_test): Fix overflow test.
12604 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
12606 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
12607 and which can live in a register, always retrieve the value on entry.
12608 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
12609 passed by invisible reference specially.
12610 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
12611 (vt_add_function_parameter): Correctly deal with a parameter passed by
12612 invisible reference.
12614 2013-08-30 Jan Hubicka <jh@suse.cz>
12616 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
12618 2013-08-30 Richard Biener <rguenther@suse.de>
12620 PR tree-optimization/58228
12621 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
12622 allow invariant loads in nested loop vectorization.
12624 2013-08-30 Richard Biener <rguenther@suse.de>
12626 PR tree-optimization/58223
12627 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
12628 (has_anti_or_output_dependence): ... this and adjust to also
12629 look for output dependences.
12630 (mark_nodes_having_upstream_mem_writes): Adjust.
12631 (rdg_flag_uses): Likewise.
12633 2013-08-30 Richard Biener <rguenther@suse.de>
12635 PR tree-optimization/58010
12636 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
12637 assert that we have a loop-closed PHI.
12639 2013-08-29 Jan Hubicka <jh@suse.cz>
12641 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
12642 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
12643 * lto-section-in.c (lto_free_function_in_decl_state): New function.
12644 (lto_free_function_in_decl_state_for_node): New function.
12646 2013-08-29 Xinliang David Li <davidxl@google.com>
12648 * loop-unroll.c (report_unroll_peel): Minor message change.
12649 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
12650 Emit alignment peeling message with default -fopt-info.
12651 (vect_loop_versioning): Emit loop version info message.
12652 * tree-vectorizer.c (vectorize_loops): Minor message change.
12653 (execute_vect_slp): Ditto.
12655 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
12657 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
12658 of the clone from the DECL_NAME of the original function.
12660 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
12662 * passes.c (register_pass): Add overload.
12663 * tree-pass.h (register_pass): Forward declare it. Add comment.
12665 2013-08-29 Jan Hubicka <jh@suse.cz>
12667 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
12668 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
12669 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
12670 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
12671 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
12672 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
12673 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
12674 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
12676 2013-08-29 Teresa Johnson <tejohnson@google.com>
12678 * dumpfile.c (dump_loc): Output column number.
12679 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
12680 * doc/invoke.texi: Document optall -fopt-info flag.
12681 * profile.c (read_profile_edge_counts): Use new dump framework.
12682 (compute_branch_probabilities): Ditto.
12683 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
12684 when pass not in any opt group.
12685 * pass_manager.h (pass_manager::get_pass_profile): New method.
12686 * value-prof.c (check_counter): Use new dump framework.
12687 (check_ic_target): Ditto.
12688 * coverage.c (get_coverage_counts): Ditto.
12689 (coverage_init): Setup new dump framework.
12691 2013-08-29 Richard Biener <rguenther@suse.de>
12693 PR tree-optimization/58246
12694 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
12695 handle the dominance check inside a basic-block.
12697 2013-08-29 Richard Biener <rguenther@suse.de>
12699 PR middle-end/57287
12700 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
12701 of default defs that appear in abnormal PHI nodes.
12703 2013-08-29 Richard Biener <rguenther@suse.de>
12705 PR tree-optimization/57685
12706 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
12707 single-use operands to avoid exponential complexity.
12709 2013-08-28 Dehao Chen <dehao@google.com>
12711 * ipa-inline.c (edge_badness): Fix integer underflow.
12713 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
12715 * gtm-builtins.def (_ITM_free): Declare leaf.
12717 2013-08-28 Jakub Jelinek <jakub@redhat.com>
12720 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
12721 (*tls_local_dynamic_base_64_largepic): Likewise.
12722 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
12723 Remove predicate from call operand.
12724 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
12725 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
12727 2013-08-28 Jeff Law <law@redhat.com>
12729 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
12730 checks for the number of predecessors and successors allowed.
12731 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
12732 which require copying a joiner block if there is a request which
12733 is a subpath that requires no joiner block copying.
12735 2013-08-28 Jan Hubicka <jh@suse.cz>
12737 * lto-streamer-out.c (DFS_write_tree_body): Drop
12738 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
12739 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
12740 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
12741 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
12742 Do not read DECL_ERROR_ISSUED.
12743 (unpack_ts_decl_with_vis_value_fields): Do not read
12745 (lto_input_ts_binfo_tree_pointers): Do not read
12746 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
12747 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
12748 write DECL_ERROR_ISSUED..
12749 (pack_ts_decl_with_vis_value_fields): Do not write
12751 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
12752 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
12753 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
12754 * tree.h (tree_decl_common): Update comment.
12755 (DECL_ERROR_ISSUED): Remove.
12757 2013-08-28 Jakub Jelinek <jakub@redhat.com>
12759 PR middle-end/58257
12760 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
12762 2013-08-28 Jan Hubicka <jh@suse.cz>
12764 * builtins.def (free): Declare leaf.
12766 2013-08-27 David Malcolm <dmalcolm@redhat.com>
12768 * gdbhooks.py: New.
12769 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
12770 * configure: Regenerate.
12772 2013-08-27 Martin Jambor <mjambor@suse.cz>
12774 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
12775 (ipa_ancestor_jf_data): Likewise.
12776 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
12777 (ipa_get_jf_pass_through_type_preserved): New function.
12778 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
12779 (ipa_get_jf_ancestor_type_preserved): New function.
12780 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
12781 (ipa_get_jf_ancestor_result): Likewise.
12782 (propagate_vals_accross_pass_through): Use
12783 ipa_get_jf_pass_through_result to do all the value mappings.
12784 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
12785 type_preserved flag.
12786 (ipa_set_jf_cst_copy): New function.
12787 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
12788 (ipa_set_jf_arith_pass_through): Likewise.
12789 (ipa_set_ancestor_jf): Likewise.
12790 (compute_complex_assign_jump_func): Set type_preserved instead of
12792 (ipa_compute_jump_functions_for_edge): Likewise.
12793 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
12794 (update_jump_functions_after_inlining): Update type_preserved.
12795 Explicitely create jump functions when combining one with pass_through.
12796 (ipa_write_jump_function): Stream the type_preserved flags.
12797 (ipa_read_jump_function): Likewise.
12799 2013-08-27 Jakub Jelinek <jakub@redhat.com>
12800 Aldy Hernandez <aldyh@redhat.com>
12802 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
12803 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
12804 * function.h (struct function): Add has_force_vect_loops and
12806 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
12807 * gimple.c (gimple_build_omp_critical): Add KIND argument and
12809 * gimple.def: Update CLAUSES comments.
12810 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
12811 (gimple_build_omp_for): Add argument to prototype.
12812 (gimple_omp_for_kind): New.
12813 (gimple_omp_for_set_kind): New.
12814 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
12815 GOVD_DATA_SHARE_CLASS.
12816 (enum omp_region_type): Add ORT_SIMD.
12817 (gimple_add_tmp_var): Handle ORT_SIMD.
12818 (gimplify_var_or_parm_decl): Same.
12819 (is_gimple_stmt): Same.
12820 (omp_firstprivatize_variable): Same.
12821 (omp_add_variable): Only use splay_tree_insert if lookup failed.
12822 (omp_notice_variable): Handle ORT_SIMD.
12823 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
12824 (omp_check_private): Handle ORT_SIMD.
12825 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
12826 OMP_CLAUSE_SAFELEN.
12827 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
12828 Handle OMP_CLAUSE_LASTPRIVATE.
12829 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
12830 OMP_CLAUSE_SAFELEN.
12831 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
12832 (gimplify_expr): Handle OMP_SIMD.
12833 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
12834 (expand_GOMP_SIMD_VF): New.
12835 (expand_GOMP_SIMD_LAST_LANE): New.
12836 * internal-fn.def (GOMP_SIMD_LANE): New.
12837 (GOMP_SIMD_VF): New.
12838 (GOMP_SIMD_LAST_LANE): New.
12839 * omp-low.c: Include target.h.
12840 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
12841 OMP_CLAUSE_SAFELEN.
12842 (check_omp_nesting_restrictions): Same.
12844 (lower_rec_simd_input_clauses): New.
12845 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
12847 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
12848 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
12849 (expand_omp_build_assign): New.
12850 (expand_omp_for_init_counts): New.
12851 (expand_omp_for_init_vars): New.
12852 (extract_omp_for_update_vars): New.
12853 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
12854 and rewrite accordingly.
12855 (expand_omp_simd): New.
12856 (expand_omp_for): Use expand_omp_simd.
12857 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
12858 (lower_omp_for): Do not lower the body.
12859 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
12860 in their own loops.
12861 * tree-flow.h (find_omp_clause): Remove prototype.
12862 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
12863 forcing vectorization of the loop, or if flag_tree_vectorize.
12864 (gate_tree_if_conversion): Similarly.
12865 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
12866 gimple_build_omp_for.
12867 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
12868 * tree-parloops (create_parallel_loop): Pass kind argument to
12869 gimple_build_omp_for.
12870 * tree-pretty-print.c (dump_omp_clause): Add cases for
12871 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
12872 OMP_CLAUSE__SIMDUID_.
12873 (dump_generic_node): Handle OMP_SIMD.
12874 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
12875 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
12876 unroll OMP_SIMD loops here.
12877 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
12878 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
12880 (vect_analyze_data_refs): Handle simd loops.
12881 * tree-vect-loop.c (vectorizable_live_operation): Handle
12883 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
12884 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
12885 (vectorizable_load): Same.
12886 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
12887 (struct simduid_to_vf): New.
12888 (simduid_to_vf::hash): New.
12889 (simduid_to-vf::equal): New.
12890 (struct simd_array_to_simduid): New.
12891 (simd_array_to_simduid::hash): New.
12892 (simd_array_to_simduid::equal): New.
12893 (adjust_simduid_builtins): New.
12894 (struct note_simd_array_uses_struct): New.
12895 (note_simd_array_uses_cb): New.
12896 (note_simd_array_uses): New.
12897 (vectorize_loops): Handle simd hints and adjust simd builtins
12899 * tree-vectorizer.h (struct _stmt_vec_info): Add
12900 simd_lane_access_p field.
12901 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
12902 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
12903 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
12904 (omp_clause_code_name): Same.
12905 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
12906 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
12907 * tree.def (OMP_SIMD): New entry.
12908 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
12909 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
12910 (OMP_CLAUSE_DECL): Adjust range for new clauses.
12911 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
12912 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
12913 (OMP_CLAUSE_LINEAR_STEP): New.
12914 (OMP_CLAUSE_SAFELEN_EXPR): New.
12915 (OMP_CLAUSE__SIMDUID__DECL): New.
12916 (find_omp_clause): New prototype.
12918 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
12920 * config/i386/driver-i386.c (host_detect_local_cpu): Update
12921 Haswell processor detection.
12923 2013-08-27 Christian Widmer <shadow@umbrox.de>
12926 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
12927 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
12928 AVX2 capable processors.
12930 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
12932 * config/aarch64/arm_neon.h: Replace all inline asm implementations
12933 of vget_low_* with implementations in terms of other intrinsics.
12935 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
12937 PR middle-end/57219
12938 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
12939 values to -1, 0 and 1.
12941 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
12943 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
12944 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
12945 (<optab><mode>3_insn): Remove.
12946 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
12947 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
12948 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
12949 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
12950 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
12951 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
12952 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
12953 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
12954 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
12955 (ror<mode>3_insn): Likewise.
12956 * config/aarch64/predicates.md (aarch64_simd_register): New.
12958 2013-08-27 Richard Biener <rguenther@suse.de>
12960 PR tree-optimization/57521
12961 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
12962 one edge is non-critical.
12963 (find_phi_replacement_condition): Make sure to use a non-critical
12964 edge. Cleanup and remove old bug workarounds.
12965 (bb_postdominates_preds): Remove.
12966 (if_convertible_loop_p_1): Do not compute post-dominators.
12967 (combine_blocks): Do not free post-dominators.
12968 (main_tree_if_conversion): Likewise.
12969 (pass_data_if_conversion): Add TODO_verify_ssa.
12971 2013-08-27 DJ Delorie <dj@redhat.com>
12973 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
12975 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
12977 * function.c (assign_parm_find_data_types): Set passed_mode and
12978 nominal_mode to the TYPE_MODE of nominal_type for the built
12979 pointer type in case of the struct-pass-by-reference.
12981 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
12983 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
12984 (UINT16_TYPE): Change default to "unsigned int".
12986 * config/avr/avr.opt (mfract-convert-truncate): New option.
12987 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
12988 is set, round negative fractional integers according to n1169
12989 when converting to integer types.
12991 2013-08-26 Jan Hubicka <jh@suse.cz>
12993 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
12994 methods can not be called indirectly when their address is not taken.
12996 2013-08-26 Jan Hubicka <jh@suse.cz>
12998 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
13001 2013-08-26 Jan Hubicka <jh@suse.cz>
13003 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
13006 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
13008 * reload.c (find_valid_class): Allow classes that do not include
13009 FIRST_PSEUDO_REGISTER - 1.
13011 2013-08-26 Jan Hubicka <jh@suse.cz>
13013 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
13014 fix edge count/frequency when speculation failed; fix type check
13015 for the direct call.
13017 2013-08-26 Jan Hubicka <jh@suse.cz>
13019 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
13021 2013-08-26 Jan Hubicka <jh@suse.cz>
13023 * ipa-inline-transform.c (inline_transform): Be ready for basic block
13024 to be changed by edge redirection.
13026 2013-08-26 Jan Hubicka <jh@suse.cz>
13028 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
13029 formating; add sanity check.
13030 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
13031 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
13032 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
13033 (resolve_noninline_speculation): Update callee keys, too.
13035 2013-08-26 Jan Hubicka <jh@suse.cz>
13037 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
13038 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
13040 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
13042 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
13045 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
13047 * config/i386/i386.c (ix86_debug_options): Remove prototype.
13048 (x86_64_elf_select_section): Ditto.
13049 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
13051 (ix86_pass_by_reference): Ditto.
13052 (output_set_got): Ditto.
13053 (ix86_unary_operator_ok): Ditto.
13054 (ix86_expand_builtin): Ditto.
13056 2013-08-23 Jan Hubicka <jh@suse.cz>
13058 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
13060 2013-08-23 Jan Hubicka <jh@suse.cz>
13062 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
13063 (tree_decl_with_vis): Add FINAL field.
13065 2013-08-23 Jeff Law <law@redhat.com>
13067 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
13068 the debugging dump when the expression is fully redundant.
13070 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
13072 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
13073 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
13074 * pretty-print.c (pp_formatted_text_data): Likewise.
13075 (pp_write_text_to_stream): Likewise.
13076 (pp_write_text_as_dot_label_to_stream): Likewise.
13077 (pp_append_r): Likewise.
13078 (pp_format): Likewise.
13079 (pp_flush): Likewise.
13080 (pp_clear_output_area): Likewise.
13081 (pp_append_text): Likewise.
13082 (pp_formatted_text): Likewise.
13083 (pp_remaining_character_count_for_line): Likewise.
13084 (pp_newline): Likewise.
13085 (pp_character): Likewise.
13086 (output_buffer::~output_buffer): Define.
13087 (pretty_printer::~pretty_printer): Destruct output buffer.
13088 * pretty-print.h (output_buffer::~output_buffer): Declare.
13089 (pretty_printer::~pretty_printer): Declare virtual.
13091 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
13094 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
13095 HOST_WIDE_INT_M1U): New macros.
13096 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
13097 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
13098 unsigned -1 for lshift.
13099 * cse.c (cse_insn): Likewise.
13100 * double-int.c (rshift_double, lshift_double): Likewise.
13101 * builtins.c (fold_builtin_bitop): Likewise.
13102 * combine.c (force_to_mode): Likewise.
13103 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
13104 * simplify-rtx.c (simplify_const_unary_operation,
13105 simplify_const_binary_operation): Likewise.
13106 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
13107 check_va_list_escapes): Likewise.
13108 * rtlanal.c (nonzero_bits1): Likewise.
13109 * expmed.c (expand_smod_pow2): Likewise.
13110 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
13112 2013-08-23 Jan Hubicka <jh@suse.cz>
13114 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
13115 as having address taken.
13117 2013-08-23 Jan Hubicka <jh@suse.cz>
13119 * ipa-utils.h (method_class_type): Declare.
13120 * ipa-devirt.c (method_class_type): Export.
13122 * cgraphunit.c (analyze_functions): Do basic devirtualization;
13123 do not walk base classes of anonymous types.
13125 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
13127 PR rtl-optimization/58220
13128 PR regression/58221
13129 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
13130 handle SEQUENCE insns properly.
13132 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
13134 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
13136 (pp_newline_and_indent): Likewise.
13137 (pp_separate_with): Likewise.
13138 * pretty-print.c (pp_newline_and_flush): Define.
13139 (pp_newline_and_indent): Likewise.
13140 (pp_separate_with): Likewise.
13142 2013-08-23 Jakub Jelinek <jakub@redhat.com>
13145 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
13146 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
13148 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
13150 * config/i386/predicates.md (ext_sse_reg_operand): New.
13151 * config/i386/i386.md (*movti_internal): Use
13152 predicate to determine if EVEX is needed.
13153 (*movsi_internal): Ditto.
13154 (*movdf_internal): Ditto.
13155 (*movsf_internal): Ditto.
13156 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
13158 2013-08-23 Jakub Jelinek <jakub@redhat.com>
13160 PR tree-optimization/58209
13161 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
13162 (find_tail_calls): Give up for pointer result types if m is non-NULL.
13163 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
13164 emit POINTER_PLUS_EXPR.
13165 (create_tailcall_accumulator): For pointer result type accumulate in
13168 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
13170 * configure.ac: Add backslashes missing from the last change.
13171 * configure: Regenerate.
13173 2013-08-22 Jan Hubicka <jh@susue.cz>
13175 * ipa.c (function_and_variable_visibility): First remember function
13176 was global and then make it local.
13178 2013-08-22 Julian Brown <julian@codesourcery.com>
13180 * configure.ac: Add aarch64 to list of arches which use "nop" in
13182 * configure: Regenerate.
13184 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13186 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
13187 gnu_libc_has_function.
13188 * config/s390/tpf.h: Likewise.
13190 2013-08-22 Jan Hubicka <jh@susue.cz>
13192 * timevar.c (validate_phases): Add cast.
13194 2013-08-22 Jan Hubicka <jh@susue.cz>
13196 * timevar.c (validate_phases): Use size_t for memory.
13197 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
13199 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
13201 * pretty-print.h (output_buffer::output_buffer): Declare.
13202 (pretty_printer::pretty_printer): Likewise.
13203 (pp_construct): Remove.
13204 * pretty-print.c (output_buffer::output_buffer): Define.
13205 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
13206 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
13207 (print_gimple_expr): Likewise.
13208 (print_gimple_seq): Likewise.
13209 (gimple_dump_bb): Likewise.
13210 * sched-vis.c (dump_value_slim): Likewise.
13211 (dump_insn_slim): Likewise.
13212 (dump_rtl_slim): Likewise.
13213 (str_pattern_slim): Likewise.
13214 * tree-mudflap.c (mf_varname_tree): Likewise.
13215 * graph.c (print_graph_cfg): Likewise.
13216 (start_graph_dump): Likewise.
13217 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
13219 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
13220 pretty printer initialization.
13221 * coretypes.h (diagnostic_context): Remove superflous type alias
13223 (pretty_printer): Likewise. Declare directly as a class.
13224 (pretty_print_info): Remove declaration as class.
13225 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
13226 and pp_clear_output_area.
13227 (asan_add_global): Likewise.
13229 2013-08-22 Jan Hubicka <jh@suse.cz>
13231 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
13232 * ipa-utils.h (update_type_inheritance_graph): Declare.
13233 (possible_polymorphic_call_target_p): Declare.
13234 (possible_polymorphic_call_target_p): New.
13235 * ipa-devirt.c: Update toplevel comments.
13236 (cached_polymorphic_call_targets): Move up.
13237 (odr_type_d): Move ID down.
13238 (polymorphic_type_binfo_p): Update comment.
13239 (odr_hasher::remove): Likewise;
13240 (get_odr_type): Set anonymous_namespace.
13241 (dump_odr_type): Dump it.
13242 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
13243 (maybe_record_node): Record node in cached_polymorphic_call_targets.
13244 (record_binfo): Add comment.
13245 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
13247 (devirt_node_removal_hook): Do not iCE when cache is freed.
13248 (possible_polymorphic_call_target_p): New predicate.
13249 (update_type_inheritance_graph): New function.
13251 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
13252 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
13253 Sergey Lega <sergey.s.lega@intel.com>
13254 Anna Tikhonova <anna.tikhonova@intel.com>
13255 Ilya Tocar <ilya.tocar@intel.com>
13256 Andrey Turetskiy <andrey.turetskiy@intel.com>
13257 Ilya Verbin <ilya.verbin@intel.com>
13258 Kirill Yukhin <kirill.yukhin@intel.com>
13259 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
13261 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
13262 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
13263 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
13264 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
13265 (OPTION_MASK_ISA_AVX2_UNSET): Update.
13266 (OPTION_MASK_ISA_AVX512F_UNSET): New.
13267 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
13268 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
13269 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
13270 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
13271 OPT_mavx512pf, OPT_mavx512er cases.
13272 * config/i386/constraints.md (v): New constraint.
13273 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
13274 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
13275 (bit_AVX512CD): New.
13276 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
13277 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
13278 * config/i386/i386-c.c (ix86_target_macros_internal):
13279 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
13281 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
13282 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
13283 * config/i386/i386.c (regclass_map, dbx_register_map)
13284 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
13285 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
13286 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
13287 -mavx512pf options.
13288 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
13289 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
13290 -mavx512pf options. Fix formatting.
13291 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
13292 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
13293 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
13294 -mavx512cd, -mavx512pf options.
13295 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
13296 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
13297 (ix86_preferred_output_reload_class): Replace SSE_REGS with
13299 (ix86_hard_regno_mode_ok): Support 512-bit registers.
13300 (ix86_set_reg_reg_cost): Ditto.
13301 (x86_order_regs_for_local_alloc): Ditto.
13302 (MAX_VECT_LEN): Extend to 64-byte.
13303 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
13304 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
13305 (TARGET_AVX512ER, TARGET_AVX512CD): New.
13306 (BIGGEST_ALIGNMENT): Extend to 512-bits.
13307 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
13308 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
13309 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
13310 (SSE_REG_MODE_P): Support new modes.
13311 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
13312 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
13313 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
13314 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
13315 (REG_CLASS_CONTENTS): Add new registers.
13316 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
13317 (EXT_REX_SSE_REGNO_P): New.
13318 (HI_REGISTER_NAMES): Add new registers.
13319 * config/i386/i386.md: Define constants for new registers.
13320 (mode): Add new 512-bit modes.
13321 (prefix): Support evex prefix.
13322 (isa): Support avx512f, noavx512f, fma_avx512f.
13323 (ssemodesuffix): Add new 512-bit modes.
13325 (*movxi_internal_avx512f): Ditto.
13326 (*movdi_internal): Replace constraint "x" with the new constraint "v".
13328 (*movsi_internal): Likewise.
13329 (*movdf_internal): Likewise.
13330 (*movsf_internal): Likewise.
13331 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
13332 (<code><mode>3): Likewise.
13333 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
13335 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
13336 with the new constraint "v".
13337 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
13339 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
13340 with the new constraint "v".
13341 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
13342 (<sse2>_storedqu<avxsizesuffix>): Likewise.
13343 (*<plusminus_insn><mode>3): Likewise.
13344 (<sse>_vm<plusminus_insn><mode>3): Likewise.
13345 (*mul<mode>3): Likewise.
13346 (<sse>_vmmul<mode>3): Likewise.
13347 (<sse>_div<mode>3): Likewise.
13348 (<sse>_vmdiv<mode>3): Likewise.
13349 (<sse>_sqrt<mode>2): Likewise.
13350 (<sse>_vmsqrt<mode>2): Likewise.
13351 (*<code><mode>3_finite): Likewise.
13352 (*<code><mode>3) <smaxmin>: Likewise.
13353 (<sse>_vm<code><mode>3): Likewise.
13354 (*<code><mode>3) <any_logic>: Likewise.
13355 (*fma_fmadd_<mode>): Likewise.
13356 (*fma_fmsub_<mode>): Likewise.
13357 (*fma_fnmadd_<mode>): Likewise.
13358 (*fma_fnmsub_<mode>): Likewise.
13359 (*fma_fmaddsub_<mode>): Likewise.
13360 (*fma_fmsubadd_<mode>): Likewise.
13361 (*fmai_fmadd_<mode>): Likewise.
13362 (*fmai_fmsub_<mode>): Likewise.
13363 (*fmai_fnmadd_<mode>): Likewise.
13364 (*fmai_fnmsub_<mode>): Likewise.
13365 (sse_cvtsi2ss): Likewise.
13366 (sse_cvtsi2ssq): Likewise.
13367 (sse_cvtss2si): Likewise.
13368 (sse_cvtss2si_2): Likewise.
13369 (sse_cvtss2siq): Likewise.
13370 (sse_cvtss2siq_2): Likewise.
13371 (sse_cvttss2si): Likewise.
13372 (sse_cvtss2siq_2): Likewise.
13373 (float<sseintvecmodelower><mode>2): Likewise.
13374 (sse2_cvtsd2si_2): Likewise.
13375 (sse2_cvtsd2siq_2): Likewise.
13376 (*<plusminus_insn><mode>3): Likewise.
13377 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
13378 (*<sse4_1_avx2>_mul<mode>3): Likewise.
13379 (ashr<mode>3): Likewise.
13380 (<shift_insn><mode>3): Likewise.
13381 (avx2_<code><mode>3): Likewise.
13382 (*avx2_<code><mode>3): Likewise.
13383 (*andnot<mode>3): Likewise.
13384 (*<code><mode>3) <any_logic>: Likewise.
13385 (abs<mode>2): Likewise.
13386 (avx2_permvar<mode>): Likewise.
13387 (avx2_perm<mode>_1): Likewise.
13388 (*avx_vpermilp<mode>): Likewise.
13389 (avx_vpermilvar<mode>3): Likewise.
13390 (avx2_ashrv<mode>): Likewise.
13391 (avx2_<shift_insn>v<mode>): Likewise.
13392 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
13394 * doc/rtl.texi: Document XImode.
13396 2013-08-21 Jeff Law <law@redhat.com>
13398 * tree-flow.h (register_jump_thread): Pass vector of edges
13399 instead of each important edge.
13400 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
13401 thread path into a vector and pass that to register_jump_thread.
13402 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
13403 passed in edge vector to the current 3-edge form.
13406 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
13408 * dce.c (fini_dce): Call df_analyze again just in case
13409 delete_unmarked_insns removed anything.
13411 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
13413 * reload.h (struct reg_equivs): Rename to ..
13414 (struct reg_equivs_s): .. this.
13416 2013-08-20 Martin Liska <marxin.liska@gmail.com>
13418 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
13420 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
13422 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
13424 2013-08-21 Jeff Law <law@redhat.com>
13426 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
13427 simplify assignments too. If the RHS collapses to a singleton
13428 range, then return the value for the range.
13430 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
13432 * config/i386/sse.md (V16): Rename to...
13434 (mov<mode>): Update iterator name.
13435 (*mov<mode>_internal): Ditto.
13436 (push<mode>1): Ditto.
13437 (movmisalign<mode>): Ditto.
13439 2013-08-20 Jan Hubicka <jh@suse.cz>
13442 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
13443 entry for direct edges.
13444 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
13446 2013-08-20 David Malcolm <dmalcolm@redhat.com>
13448 Revert my last two changes, r201865 and r201864:
13451 2013-08-20 David Malcolm <dmalcolm@redhat.com>
13453 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
13454 instances can own GC refs.
13456 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
13457 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
13458 (gcc::context::gt_pch_nx): Likewise.
13459 (gcc::context::gt_pch_nx): Likewise.
13460 * ggc.h (gt_ggc_mx <T>): New.
13461 (gt_pch_nx_with_op <T>): New.
13462 (gt_pch_nx <T>): New.
13463 * passes.c (opt_pass::gt_ggc_mx): New.
13464 (opt_pass::gt_pch_nx): New.
13465 (opt_pass::gt_pch_nx_with_op): New.
13466 (pass_manager::gt_ggc_mx): New.
13467 (pass_manager::gt_pch_nx): New.
13468 (pass_manager::gt_pch_nx_with_op): New.
13469 (pass_manager::operator new): Use
13470 ggc_internal_cleared_alloc_stat rather than xcalloc.
13471 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
13472 (pass_manager::gt_ggc_mx): New.
13473 (pass_manager::gt_pch_nx): New.
13474 (pass_manager::gt_pch_nx_with_op): New.
13475 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
13476 (opt_pass::operator new): New.
13477 (opt_pass::gt_ggc_mx): New.
13478 (opt_pass::gt_pch_nx): New.
13479 (opt_pass::gt_pch_nx_with_op): New.
13482 2013-08-20 David Malcolm <dmalcolm@redhat.com>
13484 * Makefile.in (GTFILES): Add context.h.
13485 * context.c (gcc::context::operator new): New.
13486 (gcc::context::gt_ggc_mx): New.
13487 (gcc::context::gt_pch_nx): New.
13488 (gcc::context::gt_pch_nx): New.
13489 * context.h (gcc::context): Add GTY((user)) marking.
13490 (gcc::context::operator new): New.
13491 (gcc::context::gt_ggc_mx): New.
13492 (gcc::context::gt_pch_nx): New.
13493 (gcc::context::gt_pch_nx): New.
13494 (g): Add GTY marking.
13495 (gt_ggc_mx (gcc::context *)): New.
13496 (gt_pch_nx (gcc::context *)): New.
13497 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
13498 void *cookie)): New.
13499 * gengtype.c (open_base_files) <ifiles>: Add context.h.
13501 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
13503 * dce.c (fini_dce): Call df_analyze again just in case
13504 delete_unmarked_insns removed anything.
13506 2013-08-20 Teresa Johnson <tejohnson@google.com>
13508 PR rtl-optimizations/57451
13509 * final.c (reemit_insn_block_notes): Prevent lexical blocks
13510 from crossing split section boundaries.
13512 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
13514 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
13515 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
13516 with MULTLIB_DEFAULTS.
13518 2013-08-20 Nick Clifton <nickc@redhat.com>
13520 * target.def (narrow_volatile_bitfield): Note that the default
13521 value is false, not !TARGET_STRICT_ALIGN.
13522 * doc/tm.texi: Regenerate.
13524 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
13526 Fix LIB_SPEC for systems without libpthread.
13528 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
13529 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
13531 * config/i386/linux-common.h: Likewise.
13532 * config/mips/linux-common.h: Likewise.
13534 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
13536 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
13539 2013-08-20 David Malcolm <dmalcolm@redhat.com>
13541 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
13542 instances can own GC refs.
13544 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
13545 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
13546 (gcc::context::gt_pch_nx): Likewise.
13547 (gcc::context::gt_pch_nx): Likewise.
13548 * ggc.h (gt_ggc_mx <T>): New.
13549 (gt_pch_nx_with_op <T>): New.
13550 (gt_pch_nx <T>): New.
13551 * passes.c (opt_pass::gt_ggc_mx): New.
13552 (opt_pass::gt_pch_nx): New.
13553 (opt_pass::gt_pch_nx_with_op): New.
13554 (pass_manager::gt_ggc_mx): New.
13555 (pass_manager::gt_pch_nx): New.
13556 (pass_manager::gt_pch_nx_with_op): New.
13557 (pass_manager::operator new): Use
13558 ggc_internal_cleared_alloc_stat rather than xcalloc.
13559 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
13560 (pass_manager::gt_ggc_mx): New.
13561 (pass_manager::gt_pch_nx): New.
13562 (pass_manager::gt_pch_nx_with_op): New.
13563 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
13564 (opt_pass::operator new): New.
13565 (opt_pass::gt_ggc_mx): New.
13566 (opt_pass::gt_pch_nx): New.
13567 (opt_pass::gt_pch_nx_with_op): New.
13569 2013-08-20 David Malcolm <dmalcolm@redhat.com>
13571 * Makefile.in (GTFILES): Add context.h.
13572 * context.c (gcc::context::operator new): New.
13573 (gcc::context::gt_ggc_mx): New.
13574 (gcc::context::gt_pch_nx): New.
13575 (gcc::context::gt_pch_nx): New.
13576 * context.h (gcc::context): Add GTY((user)) marking.
13577 (gcc::context::operator new): New.
13578 (gcc::context::gt_ggc_mx): New.
13579 (gcc::context::gt_pch_nx): New.
13580 (gcc::context::gt_pch_nx): New.
13581 (g): Add GTY marking.
13582 (gt_ggc_mx (gcc::context *)): New.
13583 (gt_pch_nx (gcc::context *)): New.
13584 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
13585 void *cookie)): New.
13586 * gengtype.c (open_base_files) <ifiles>: Add context.h.
13588 2013-08-20 Alan Modra <amodra@gmail.com>
13591 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
13592 (rs6000_emit_epilogue): Likewise.
13594 2013-08-19 Dehao Chen <dehao@google.com>
13596 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
13598 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
13599 Jakub Jelinek <jakub@redhat.com>
13601 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
13602 (BUILT_IN_FABSD64): Likewise.
13603 (BUILT_IN_FABSD128): Likewise.
13604 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
13605 (fold_builtin_1): Likewise.
13606 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
13607 destination and source operands.
13608 (*abstd2_fpr): Likewise.
13609 (*nabstd2_fpr): Likewise.
13611 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
13613 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
13616 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
13618 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
13621 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
13623 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
13624 * builtins.c (default_libc_has_function): New.
13625 (gnu_libc_has_function): Ditto.
13626 (no_c99_libc_has_function): Ditto.
13627 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
13628 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
13629 (fold_builtin_sincos): Likewise.
13630 (fold_builtin_cexp): Likewise.
13631 * builtins.def (DEF_C94_BUILTIN): Likewise.
13632 (DEF_C99_BUILTIN): Likewise.
13633 (DEF_C99_C90RES_BUILTIN): Likewise.
13634 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
13635 definitions to using this define.
13636 * config/darwin-protos.h (darwin_libc_has_function): New.
13637 * config/darwin.c (darwin_libc_has_function): Ditto.
13638 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
13639 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
13640 * config/darwin.h: Ditto.
13641 * config/elfos.h: Ditto.
13642 * config/freebsd.h: Ditto.
13643 * config/i386/cygming.h: Ditto.
13644 * config/i386/djgpp.h: Ditto.
13645 * config/i386/i386-interix.h: Ditto.
13646 * config/microblaze/microblaze.h: Ditto.
13647 * config/mmix/mmix.h: Ditto.
13648 * config/gnu-user.h: Ditto.
13649 * config/ia64/hpux.h: Ditto.
13650 * config/pa/pa-hpux.h: Ditto.
13651 * config/pdp11/pdp11.h: Ditto.
13652 * config/picochip/picochip.h: Ditto.
13653 * config/linux.h: Ditto.
13654 * config/netbsd.h: Ditto.
13655 * config/openbsd.h: Ditto.
13656 * config/rs6000/aix43.h: Ditto.
13657 * config/rs6000/aix51.h: Ditto.
13658 * config/rs6000/aix52.h: Ditto.
13659 * config/rs6000/aix53.h: Ditto.
13660 * config/rs6000/aix61.h: Ditto.
13661 * config/rs6000/darwin.h: Ditto.
13662 * config/rs6000/linux.h: Ditto.
13663 * config/rs6000/linux64.h: Ditto.
13664 * config/s390/tpf.h: Ditto.
13665 * config/sol2-10.h: Ditto.
13666 * config/sol2.h: Ditto.
13667 * config/vms/vms.h: Ditto.
13668 * config/vxworks.h: Ditto.
13669 * config/linux-android.c (linux_android_libc_has_function):
13670 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
13671 * config/linux-protos.h (linux_android_libc_has_function):
13673 * config/i386/i386.c (ix86_libc_has_function): New.
13674 * config/i386/i386-protos.h
13675 (ix86_libc_has_function): New declaration.
13676 * config/i386/i386.md
13677 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
13678 ("isinf<mode>2): Likewise.
13679 * convert.c (convert_to_integer): Using new target hook
13680 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
13681 TARGET_C99_FUNCTIONS.
13682 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
13683 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
13684 * coretypes.h (function_class): New enum for different
13685 classes of functions.
13686 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
13687 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
13688 (TARGET_HAS_SINCOS): Likewise.
13689 (TARGET_LIBC_HAS_FUNCTION): New.
13690 * doc/tm.texi: Regenerated.
13691 * targhooks.h (default_libc_has_function): New declaration.
13692 (no_c99_libc_has_function): Ditto.
13693 (gnu_libc_has_function): Ditto.
13694 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
13695 and TARGET_HAS_SINCOS.
13697 2013-08-18 Jan Hubicka <jh@suse.cz>
13699 * Makeifle-in (ipa-devirt.o): New.
13700 (GTFILES): Add ipa-utils.h and ipa-devirt.c
13701 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
13702 (analyze_functions): Look into possible targets of polymorphic call.
13703 * dumpfile.c (dump_files): Add type-inheritance dump.
13704 * dumpfile.h (TDI_inheritance): New.
13705 * ipa-devirt.c: New file.
13706 * ipa-utils.h (odr_type_d): Forward declare.
13707 (odr_type): New type.
13708 (build_type_inheritance_graph): Declare.
13709 (possible_polymorphic_call_targets): Declare and introduce inline
13710 variant when only edge is pased.
13711 (dump_possible_polymorphic_call_targets): Likewise.
13712 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
13713 * tree.c (type_in_anonymous_namespace_p): Break out from ...
13714 (types_same_for_odr): ... here.
13715 * tree.h (type_in_anonymous_namespace_p): Declare.
13717 2013-08-18 Jakub Jelinek <jakub@redhat.com>
13719 PR tree-optimization/58006
13720 * tree-parloops.c (take_address_of): Don't ICE if get_name
13722 (eliminate_local_variables_stmt): Remove clobber stmts.
13724 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
13726 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
13727 error message is issued for an alias to undefined symbol.
13729 2013-08-18 Jan Hubicka <jh@suse.cz>
13731 * cgraph.c (cgraph_create_indirect_edge): Discover
13732 polymorphic calls and record basic info into indirect_info.
13733 * gimple-fold.c (gimple_fold_call): When doing BINFO based
13734 devirtualization, ignore objc function calls.
13735 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
13736 call with no parm index info.
13737 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
13738 * tree.c (virtual_method_call_p): New function.
13739 * tree.h (virtual_method_call_p): Declare.
13741 2013-08-16 Jan Hubicka <jh@suse.cz>
13743 PR middle-end/58179
13744 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
13746 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
13748 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
13751 2013-08-16 David Malcolm <dmalcolm@redhat.com>
13753 * gengtype.c (type_for_name): Add special-case support for
13754 locating types within the "gcc::" namespace.
13755 (open_base_files): Emit a "using namespace gcc" directive.
13757 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
13760 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
13761 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
13763 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
13764 array instead of each individual operand as a separate argument.
13765 (emit_fusion_gpr_load): Likewise.
13766 (expand_fusion_gpr_load): Add new function declaration.
13768 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
13769 signature to have the operands passed as an array, instead of as
13770 separate arguments. Allow ZERO_EXTEND to be in the memory
13771 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
13772 depend on the register live/dead flags when peepholes are run.
13773 (expand_fusion_gpr_load): New function to be called from the
13774 peephole2 pass, to change the register that addis sets to be the
13776 (emit_fusion_gpr_load): Change the calling signature to have the
13777 operands passed as an array, instead of as separate arguments.
13778 Allow ZERO_EXTEND to be in the memory address, and also
13779 SIGN_EXTEND if -mpower8-fusion-sign.
13781 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
13782 unspec enumeration.
13783 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
13784 adjust the register addis loads up in the peephole2 pass. Do not
13785 depend on the register live/dead state when the peephole pass is done.
13787 2013-08-16 David Malcolm <dmalcolm@redhat.com>
13789 * gengtype.c (create_user_defined_type): Ensure that the kind
13790 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
13791 declaration is seen before the GTY((user)) marking.
13793 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
13796 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
13798 2013-08-16 Jan Hubicka <jh@suse.cz>
13800 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
13801 arugment expected_type.
13802 (gimple_fold_call): Use it.
13803 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
13804 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
13805 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
13806 (try_make_edge_direct_virtual_call): Likewise.
13807 * tree.c (obj_type_ref_class): New.
13808 * tree.h (obj_type_ref_class): Use it.
13810 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
13812 * sched-vis.c (rtl_slim_pp_initialized): Remove.
13813 (rtl_slim_pp): Likewise.
13814 (init_rtl_slim_pretty_print): Likewise.
13815 (dump_value_slim): Don't call it. Use local pretty printer.
13816 (dump_insn_slim): Likewise.
13817 (dump_rtl_slim): Likewise.
13818 (str_pattern_slim): Likewise.
13819 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
13822 2013-08-16 Jakub Jelinek <jakub@redhat.com>
13824 PR tree-optimization/58164
13825 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
13826 walk gimple_goto_dest of GIMPLE_GOTO.
13828 PR tree-optimization/58165
13829 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
13830 bi_call must be the last stmt in a bb, don't split_block, instead
13831 use fallthru edge from it and give up if there is none.
13832 Release conds vector when returning early.
13834 2013-08-14 Xinliang David Li <davidxl@google.com>
13836 * config/i386/i386.c (ix86_option_override_internal):
13837 Remove unused variable and field.
13839 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13842 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
13843 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
13844 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
13845 and Linux, correct BLKmode alignment when 128-bit alignment is
13846 required and compatibility flag is not set.
13847 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
13848 for zero-size arguments when compatibility flag is not set.
13850 2013-08-14 Jakub Jelinek <jakub@redhat.com>
13852 PR tree-optimization/58145
13853 * tree-sra.c (build_ref_for_offset): If prev_base has
13854 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
13856 2013-08-14 Xinliang David Li <davidxl@google.com>
13858 * config/i386/i386.c (ix86_option_override_internal):
13859 Fix uninitialized variable error.
13861 2013-08-14 Xinliang David Li <davidxl@google.com>
13863 * config/i386/i386.opt: Define two new options.
13864 * config/i386/x86-tune.def: Add arch selector field in macros.
13865 * config/i386/i386.h: Adjust macro definition.
13866 * config/i386/i386.c (ix86_option_override_internal):
13868 (parse_mtune_ctrl_str): New function.
13869 (set_ix86_tune_features): New function.
13870 (ix86_function_specific_restore): Call the new helper function.
13872 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
13874 PR rtl-optimization/57662
13875 * sel-sched.c (code_motion_process_successors): When the current insn
13876 is removed after the recursive traversal, break from the loop.
13877 Add comments and debug printouts.
13879 2013-08-14 Jakub Jelinek <jakub@redhat.com>
13880 Alexandre Oliva <aoliva@redhat.com>
13883 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
13884 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
13885 there also UNSPEC_PLTOFF.
13887 2013-08-14 Marek Polacek <polacek@redhat.com>
13889 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
13890 NUM_CONDITIONS bit positions.
13892 2013-08-13 Cary Coutant <ccoutant@google.com>
13894 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
13895 (attr_checksum): Hash vector contents instead of pointer.
13896 (attr_checksum_ordered): Likewise.
13898 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
13900 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
13901 when Pmode != word_mode. Add length_address attribute.
13902 (sse3_monitor_<mode>): Merge from sse3_monitor and
13903 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
13904 Pmode != word_mode. Update insn length attribute.
13905 * config/i386/i386.c (ix86_option_override_internal): Update
13906 ix86_gen_monitor selection for merged sse3_monitor insn.
13908 2013-08-13 Julian Brown <julian@codesourcery.com>
13910 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
13911 perform invalid legitimization on greater-than-word-size modes for
13912 TARGET_E500_DOUBLE.
13914 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
13916 * ira.c (setup_class_translate_array): Use aclass instead of cl
13917 for classes not fully covered by allocno classes.
13919 2013-08-13 Jakub Jelinek <jakub@redhat.com>
13921 PR tree-optimization/57661
13922 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
13923 * tree-inline.c (tree_function_versioning): Initialize it.
13924 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
13925 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
13926 that is not being copied.
13929 * asan.c (instrument_strlen_call): Fix typo in comment.
13930 Use char * type even for the lhs of POINTER_PLUS_EXPR.
13932 2013-08-13 Steve Ellcey <sellcey@mips.com>
13934 * config/mips/mips.md (prefetch): Use lw instead of ld on
13935 loongson in 32bit mode.
13937 2013-08-13 Nick Clifton <nickc@redhat.com>
13939 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
13941 2013-08-13 Jan Hubicka <jh@suse.cz>
13943 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
13944 introduced edge; fix typo in sanity check.
13945 (cgraph_resolve_speculation): Export; improve diagnostic.
13946 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
13947 speculation at type mismatch.
13948 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
13949 (cgraph_resolve_speculation): Declare.
13950 (symtab_can_be_discarded): New function.
13951 * value-prof.c (gimple_ic_transform): Remove actual transform code.
13952 * ipa-inline-transform.c (speculation_removed): New global var.
13953 (clone_inlined_nodes): See if speculation can be removed.
13954 (inline_call): If speculations was removed, we growths may not match.
13955 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
13956 (speculation_useful_p): New function.
13957 (resolve_noninline_speculation): New function.
13958 (inline_small_functions): Resolve useless speculations.
13959 * ipa-inline.h (speculation_useful_p): Declare
13960 * ipa.c (can_replace_by_local_alias): Simplify.
13961 (ipa_profile): Produce speculative calls in non-lto, too;
13962 add simple cost model; produce local aliases.
13964 2013-08-13 David Malcolm <dmalcolm@redhat.com>
13966 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
13969 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
13971 * config/i386/i386.c (ix86_function_versions): Use error + inform.
13973 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
13975 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
13976 iterator instead of X87MODEF.
13978 2013-08-12 Perez Read <netfirewall@gmail.com>
13981 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
13982 operand 0 for intel asm alternative.
13983 (*movabs<mode>_2): Ditto for operand 1.
13985 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
13987 * config/aarch64/arm_none.h
13988 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
13990 2013-08-12 Nick Clifton <nickc@redhat.com>
13992 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
13994 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
13996 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
13997 expand for QI/HImode operand to produce more effictive code for
13998 unsigned char(short) --> float(double) conversion.
14000 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
14002 * doc/invoke.texi: Mention that -ftls-model does not force the final
14005 2013-08-12 Marek Polacek <polacek@redhat.com>
14006 Marc Glisse <marc.glisse@inria.fr>
14008 PR tree-optimization/57980
14009 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
14010 when creating -1 constant.
14012 2013-08-10 Jan Hubicka <jh@suse.cz>
14014 Workaround binutils PR14342.
14015 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
14016 (gimple_init_edge_profiler): Likewise.
14017 (gimple_gen_ic_func_profiler): Likewise.
14019 2013-08-09 Jan Hubicka <jh@suse.cz>
14021 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
14023 2013-08-09 Xinliang David Li <davidxl@google.com>
14025 * config/i386/stringop.def: New file.
14026 * config/i386/stringop.opt: New file.
14027 * config/i386/i386-opts.h: Include stringopt.def.
14028 * config/i386/i386.opt: Include stringopt.opt.
14029 * config/i386/i386.c (ix86_option_override_internal):
14030 Override default size based stringop inline strategies with options.
14031 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
14034 2013-08-09 Jan Hubicka <jh@suse.cz>
14036 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
14038 2013-08-09 Jan Hubicka <jh@suse.cz>
14040 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
14042 (dump_cgraph_node): Dump profile-id.
14043 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
14044 and common_target_probability.
14045 * lto-cgraph.c (lto_output_edge): Stream common targets.
14046 (lto_output_node): Stream profile ids.
14047 (input_node): Stream profile ids.
14048 (input_edge): Stream common targets.
14049 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
14050 * ipa.c: Include value-prof.h
14051 (ipa_profile_generate_summary): Turn indirect call statement histograms
14052 into common targets.
14053 (ipa_profile): Turn common targets into speculative edges.
14055 2013-08-09 Jan Hubicka <jh@suse.cz>
14057 * cgraph.h (cgraph_node): Add profile_id.
14058 * value-prof.c (cgraph_node_map): Turn into pointer_map.
14059 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
14060 (del_node_map): Update.
14061 (find_func_by_funcdef_no): Replace by ...
14062 (find_func_by_profile_id): ... this one.
14063 (gimple_ic_transform): Do not remove useful histograms when
14064 speculation is not done; dump info when indirect call removal
14066 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
14067 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
14068 (__gcov_indirect_call_profiler_v2): .. this one.
14069 * profile.h (init_node_map): Update.
14070 * coverage.c (coverage_compute_profile_id): New function.
14071 * coverage.h (coverage_compute_profile_id): Declare.
14072 * tree-profile.c (init_ic_make_global_vars): Make
14073 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
14074 (gimple_init_edge_profiler): Update prototype of
14075 __gcov_indirect_call_profiler.
14076 (gimple_gen_ic_func_profiler): Simplify.
14077 (tree_profiling): Use init_node_map
14079 2013-08-09 Jan Hubicka <jh@suse.cz>
14081 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
14082 non-speculative refs.
14083 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
14084 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
14085 (cgraph_set_call_stmt): Likewise.
14086 (cgraph_create_edge_1): Fix release checking compilatoin;
14087 clear lto_stmt_uid.
14088 (cgraph_free_edge): Free indirect info.
14089 (cgraph_turn_edge_to_speculative): New function.
14090 (cgraph_speculative_call_info): New function.
14091 (cgraph_make_edge_direct): Return direct edge; handle speculation.
14092 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
14093 (dump_cgraph_node): Dump speculation.
14094 (verify_edge_count_and_frequency): Accept speculative edges.
14095 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
14096 (verify_cgraph_node): Handle speculation.
14097 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
14098 (cgraph_set_call_stmt): Update prototype.
14099 (cgraph_make_edge_direct): Update prototype.
14100 (cgraph_speculative_call_info): Declare.
14101 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
14102 to change; update call of ipa_find_references.
14103 * ipa-ref.c (ipa_record_reference): Fix return value; clear
14104 lto_stmt_uid and speculative flags.
14105 (ipa_dump_references): Dump speculation.
14106 (ipa_clone_references): Clone speculative flag.
14107 (ipa_clone_referring): Likewise.
14108 (ipa_clone_ref): New function.
14109 (ipa_find_reference): Look into lto_stmt_uids
14110 (ipa_clear_stmts_in_references): Do not clear speculative calls.
14111 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
14112 (ipa_find_reference): Update declaration.
14113 (ipa_clone_ref): Declare.
14114 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
14115 stream speculative flag.
14116 (lto_output_ref): Stream statements uids and speculation.
14117 (input_ref): Likewise.
14118 (input_edge): Stream speuclation.
14119 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
14120 (cgraph_set_call_stmt_including_clones): Handle speculation.
14121 * ipa-inline.c (heap_edge_removal_hook): New function.
14122 (inline_small_functions): Register it.
14123 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
14124 also initialize refs.
14125 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
14127 (try_make_edge_direct_simple_call): Likewise.
14128 (try_make_edge_direct_simple_call): Likewise.
14129 (update_indirect_edges_after_inlining): Likewise.
14130 (remove_described_reference): Look proper lto_stmt_uid.
14131 (propagate_controlled_uses): Likewise.
14132 (propagate_controlled_uses): Liekwise.
14133 * tree-inline.c (copy_bb): Copy speculative edges.
14134 (redirect_all_calls): New function.
14135 (copy_cfg_body): Do redirection after loop info is updated.
14136 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
14138 2013-08-09 Jan Hubicka <jh@suse.cz>
14140 * lto-streamer-out.c (output_function): Renumber PHIs.
14141 * lto-streamer-in.c (input_function): Likewise.
14143 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
14145 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
14146 (get_lane_unsigned): Likewise.
14147 (dup_lane_scalar): Likewise.
14148 (get_lane): enable for VALL.
14149 * config/aarch64/aarch64-simd.md
14150 (aarch64_dup_lane_scalar<mode>): Remove.
14151 (aarch64_get_lane_signed<mode>): Likewise.
14152 (aarch64_get_lane_unsigned<mode>): Likewise.
14153 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
14154 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
14155 (aarch64_get_lane<mode>): Enable for all vector modes.
14156 (aarch64_get_lanedi): Remove misleading constraints.
14157 * config/aarch64/arm_neon.h
14158 (__aarch64_vget_lane_any): Define.
14159 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
14160 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
14161 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
14162 * config/aarch64/iterators.md (VDQQH): New.
14163 (VDQQHS): Likewise.
14164 (vwcore): Likewise.
14166 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
14168 * configure.ac: Add GAS check for LEON instructions on SPARC.
14169 * configure: Regenerate.
14170 * config.in: Likewise.
14171 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
14173 * config/sparc/sparc.opt (LEON, LEON3): New masks.
14174 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
14176 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
14177 (AS_LEON_FLAG): New macro.
14178 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
14179 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
14180 Deal with LEON and LEON3 for the memory model.
14181 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
14182 (atomic_compare_and_swap<mode>_1): Likewise.
14183 (*atomic_compare_and_swap<mode>_1): Likewise.
14185 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
14187 * config/arm/neon.md (vcond): Fix floating-point vector
14188 comparisons against 0.
14190 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
14192 * lra-constraints.c (emit_spill_move): Remove assert.
14193 (process_alt_operands): Add more debugging
14194 output. Increase reject for spilling into memory. Decrease
14195 reject for reloading scratch.
14196 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
14198 2013-08-08 Steve Ellcey <sellcey@mips.com>
14200 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
14201 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
14202 micromips incompatible. Add nan2008.
14203 (MULTILIB_DIRNAMES): Add nan2008.
14204 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
14205 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
14206 and micromips incompatible. Add nan2008.
14207 (MULTILIB_DIRNAMES): Add nan2008.
14208 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
14210 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
14212 PR rtl-optimization/58079
14213 * combine.c (combine_simplify_rtx): Avoid using SUBST if
14214 simplify_comparison has widened a comparison with an integer.
14216 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14218 * config/arm/neon.md (movmisalign<mode>): Disable when we
14219 don't allow unaligned accesses.
14220 (*movmisalign<mode>_neon_store): Likewise.
14221 (*movmisalign<mode>_neon_load): Likewise.
14222 (*movmisalign<mode>_neon_store): Likewise.
14223 (*movmisalign<mode>_neon_load): Likewise.
14225 2013-08-08 Jan Hubicka <jh@suse.cz>
14227 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
14228 (make_pass_rebuild_cgraph_edges): Also clear references.
14229 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
14230 * ipa-inline-transform.c (inline_transform): Remove all references
14232 * cgraphunit.c (expand_function): Remove all references after
14234 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
14235 (ipa_find_reference): Rewrite to iterator.
14236 (remove_stmt_references): Likewise.
14237 (ipa_clear_stmts_in_references): New function.
14238 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
14239 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
14241 * ipa-split.c (split_function): Remove references in split function.
14243 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
14246 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
14247 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
14249 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
14252 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
14253 suggested mode for the assignment isn't compatible with the
14254 registers required.
14256 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
14259 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
14261 2013-08-07 Xinliang David Li <davidxl@google.com>
14263 * config/i386/i386.opt: New option -mtune-ctrl=.
14264 * config/i386/x86-tune.def: New file.
14265 * config/i386/i386.h: include x86-tune.def.
14266 * config/i386/i386.c (ix86_option_override_internal):
14267 Parsing -mtune-ctrl= option and set tune features.
14269 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
14272 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
14273 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
14276 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
14278 * rtl.h (update_alignments): Declare.
14279 * final.c (grow_label_align): New function extracted from...
14280 (shorten_branches): ...here. Call it.
14281 (update_alignments): New function.
14282 * reorg.c (sibling_labels): New variable.
14283 (get_label_before): Add SIBLING parameter. If it is non-zero, push
14284 the new label along with it onto the sibling_labels vector.
14285 (fill_simple_delay_slots): Adjust call to get_label_before.
14286 (fill_slots_from_thread): Likewise.
14287 (relax_delay_slots): Likewise.
14288 (make_return_insns): Likewise.
14289 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
14291 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
14293 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
14294 document its semantics.
14295 (diagnostic_report_diagnostic): Adjust accordingly.
14297 2013-08-07 David Malcolm <dmalcolm@redhat.com>
14299 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
14300 (sparc_option_override): ...and port to new C++ pass API.
14301 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
14303 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
14305 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
14307 2013-08-06 Caroline Tice <cmtice@google.com>
14309 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
14310 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
14311 * tree-pass.h: Add pass_vtable_verify.
14312 * varasm.c (assemble_variable): Add code to properly set the comdat
14313 section and name for the .vtable_map_vars section.
14314 (assemble_vtyv_preinit_initializer): New function.
14315 (default_sectin_type_flags): Make sure .vtable_map_vars section has
14317 * output.h: Add function decl for assemble_vtv_preinit_initializer.
14318 * vtable-verify.c: New file.
14319 * vtable-verify.h: New file.
14320 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
14321 initialiation levels.
14322 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
14323 * passes.def: Insert pass_vtable_verify.
14324 * aclocal.m4: Reorder includes.
14325 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
14326 -fvtv-counts options.
14327 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
14328 as appropriate, if -fvtable-verify=... is used.
14329 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
14330 -fvtable-verify=... is used.
14331 * Makefile.in (OBJS): Add vtable-verify.o to list.
14332 (vtable-verify.o): Add new build rule.
14333 (GTFILES): Add vtable-verify.c to list.
14334 * common.opt (fvtable-verify=): New flag.
14335 (vtv_priority): Values for fvtable-verify= flag.
14336 (fvtv-counts): New flag.
14337 (fvtv-debug): New flag.
14338 * tree.h (save_vtable_map_decl): New extern function decl.
14340 2013-08-07 David Malcolm <dmalcolm@redhat.com>
14342 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
14343 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
14344 (pass_data_rl78_devirt): ...new pass_data instance and...
14345 (make_pass_rl78_devirt): ...new function.
14346 (rl78_asm_file_start): Port pass registration to new C++ API.
14348 2013-08-07 David Malcolm <dmalcolm@redhat.com>
14350 * coretypes.h (rtl_opt_pass): Add.
14351 (gcc::context): Add.
14352 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
14353 (epiphany_init): Port to new C++ pass API.
14354 (epiphany_optimize_mode_switching): Likewise.
14355 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
14356 (pass_manager::get_pass_mode_switching): New.
14357 (pass_manager::get_pass_peephole2): New.
14358 * mode-switching.c (pass_mode_switching): Add clone method.
14359 * recog.c (pass_peephole2): Add clone method.
14360 (pass_split_all_insns): Add clone method.
14362 2013-08-06 David Malcolm <dmalcolm@redhat.com>
14364 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
14365 (mips_option_override): ...here, porting to new C++ API for passes.
14367 2013-08-06 Jan Hubicka <jh@suse.cz>
14369 * cgraph.c (cgraph_get_body): New function based on lto.c
14371 * cgraph.h (cgraph_get_body): Declare.
14372 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
14374 * cgraphunit.c (expand_function): Get body prior expanding.
14375 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
14376 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
14378 * passes.c (do_per_function_toporder): Get body.
14379 * tree-inline.c (expand_call_inline): Get body prior inlining it.
14380 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
14382 2013-08-06 Martin Jambor <mjambor@suse.cz>
14385 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
14386 is not re-finalized. Rename second parameter to no_collect.
14388 2013-08-06 Martin Jambor <mjambor@suse.cz>
14390 PR middle-end/58041
14391 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
14392 MEM_REF has proper alignment information.
14394 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
14397 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
14399 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
14400 rtx (*) (rtx, ...) with insn_gen_fn.
14401 * genoutput.c (output_insn_data): Cast gen_? function pointers to
14402 insn_gen_fn::stored_funcptr. Add initializer braces.
14404 2013-08-05 David Malcolm <dmalcolm@redhat.com>
14406 Rewrite how instances of passes are cloned to remove assumptions
14407 about their sizes (thus allowing pass subclasses to have
14408 additional data fields, albeit non-GC-managed ones at this point).
14410 * passes.c (make_pass_instance): Now that passes have clone
14411 methods, rewrite this function to eliminate XNEW and memcpy
14412 calls that used hardcoded sizes. Since this function no longer
14413 creates pass instances, rename it to...
14414 (add_pass_instance): ...this. Document the old way that passes were
14415 numbered and flagged, and rework this function to continue using it.
14416 (next_pass_1): Add an initial_pass argument for use by
14418 (position_pass): When adding multiple instances of a pass, use
14419 the pass's clone method, rather than relying on the XNEW/memcpy
14420 within the former make_pass_instance (now add_pass_instance).
14421 (pass_manager::pass_manager): When invoking next_pass_1, also supply
14422 the initial instance of the current pass within the pass manager.
14424 2013-08-05 David Malcolm <dmalcolm@redhat.com>
14426 This is the automated part of the conversion of passes from C
14427 structs to C++ classes.
14429 Patch autogenerated by refactor_passes.py from
14430 https://github.com/davidmalcolm/gcc-refactoring-scripts
14431 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
14433 * asan.c (pass_asan): Convert from a global struct to a subclass of
14434 gimple_opt_pass along with...
14435 (pass_data_asan): ...new pass_data instance and...
14436 (make_pass_asan): ...new function.
14437 (pass_asan_O0): Convert from a global struct to a subclass of
14438 gimple_opt_pass along with...
14439 (pass_data_asan_O0): ...new pass_data instance and...
14440 (make_pass_asan_O0): ...new function.
14441 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
14442 subclass of rtl_opt_pass along with...
14443 (pass_data_inc_dec): ...new pass_data instance and...
14444 (make_pass_inc_dec): ...new function.
14445 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
14446 a subclass of rtl_opt_pass along with...
14447 (pass_data_reorder_blocks): ...new pass_data instance and...
14448 (make_pass_reorder_blocks): ...new function.
14449 (pass_duplicate_computed_gotos): Convert from a global struct to a
14450 subclass of rtl_opt_pass along with...
14451 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
14452 (make_pass_duplicate_computed_gotos): ...new function.
14453 (pass_partition_blocks): Convert from a global struct to a subclass of
14454 rtl_opt_pass along with...
14455 (pass_data_partition_blocks): ...new pass_data instance and...
14456 (make_pass_partition_blocks): ...new function.
14457 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
14458 struct to a subclass of rtl_opt_pass along with...
14459 (pass_data_branch_target_load_optimize1): ...new pass_data instance
14461 (make_pass_branch_target_load_optimize1): ...new function.
14462 (pass_branch_target_load_optimize2): Convert from a global struct to a
14463 subclass of rtl_opt_pass along with...
14464 (pass_data_branch_target_load_optimize2): ...new pass_data instance
14466 (make_pass_branch_target_load_optimize2): ...new function.
14467 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
14468 of rtl_opt_pass along with...
14469 (pass_data_jump): ...new pass_data instance and...
14470 (make_pass_jump): ...new function.
14471 (pass_jump2): Convert from a global struct to a subclass of
14472 rtl_opt_pass along with...
14473 (pass_data_jump2): ...new pass_data instance and...
14474 (make_pass_jump2): ...new function.
14475 * cfgexpand.c (pass_expand): Convert from a global struct to a
14476 subclass of rtl_opt_pass along with...
14477 (pass_data_expand): ...new pass_data instance and...
14478 (make_pass_expand): ...new function.
14479 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
14480 of rtl_opt_pass along with...
14481 (pass_data_free_cfg): ...new pass_data instance and...
14482 (make_pass_free_cfg): ...new function.
14483 (pass_into_cfg_layout_mode): Convert from a global struct to a
14484 subclass of rtl_opt_pass along with...
14485 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
14486 (make_pass_into_cfg_layout_mode): ...new function.
14487 (pass_outof_cfg_layout_mode): Convert from a global struct to a
14488 subclass of rtl_opt_pass along with...
14489 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
14490 (make_pass_outof_cfg_layout_mode): ...new function.
14491 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
14492 struct to a subclass of gimple_opt_pass along with...
14493 (pass_data_build_cgraph_edges): ...new pass_data instance and...
14494 (make_pass_build_cgraph_edges): ...new function.
14495 (pass_rebuild_cgraph_edges): Convert from a global struct to a
14496 subclass of gimple_opt_pass along with...
14497 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
14498 (make_pass_rebuild_cgraph_edges): ...new function.
14499 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
14500 subclass of gimple_opt_pass along with...
14501 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
14503 (make_pass_remove_cgraph_callee_edges): ...new function.
14504 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
14505 struct to a subclass of rtl_opt_pass along with...
14506 (pass_data_stack_adjustments): ...new pass_data instance and...
14507 (make_pass_stack_adjustments): ...new function.
14508 * combine.c (pass_combine): Convert from a global struct to a subclass
14509 of rtl_opt_pass along with...
14510 (pass_data_combine): ...new pass_data instance and...
14511 (make_pass_combine): ...new function.
14512 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
14513 global struct to a subclass of rtl_opt_pass along with...
14514 (pass_data_compare_elim_after_reload): ...new pass_data instance
14516 (make_pass_compare_elim_after_reload): ...new function.
14517 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
14518 of rtl_opt_pass along with...
14519 (pass_data_rtl_cprop): ...new pass_data instance and...
14520 (make_pass_rtl_cprop): ...new function.
14521 * cse.c (pass_cse): Convert from a global struct to a subclass of
14522 rtl_opt_pass along with...
14523 (pass_data_cse): ...new pass_data instance and...
14524 (make_pass_cse): ...new function.
14525 (pass_cse2): Convert from a global struct to a subclass of
14526 rtl_opt_pass along with...
14527 (pass_data_cse2): ...new pass_data instance and...
14528 (make_pass_cse2): ...new function.
14529 (pass_cse_after_global_opts): Convert from a global struct to a
14530 subclass of rtl_opt_pass along with...
14531 (pass_data_cse_after_global_opts): ...new pass_data instance and...
14532 (make_pass_cse_after_global_opts): ...new function.
14533 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
14534 of rtl_opt_pass along with...
14535 (pass_data_ud_rtl_dce): ...new pass_data instance and...
14536 (make_pass_ud_rtl_dce): ...new function.
14537 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
14538 rtl_opt_pass along with...
14539 (pass_data_fast_rtl_dce): ...new pass_data instance and...
14540 (make_pass_fast_rtl_dce): ...new function.
14541 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
14542 a subclass of rtl_opt_pass along with...
14543 (pass_data_df_initialize_opt): ...new pass_data instance and...
14544 (make_pass_df_initialize_opt): ...new function.
14545 (pass_df_initialize_no_opt): Convert from a global struct to a
14546 subclass of rtl_opt_pass along with...
14547 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
14548 (make_pass_df_initialize_no_opt): ...new function.
14549 (pass_df_finish): Convert from a global struct to a subclass of
14550 rtl_opt_pass along with...
14551 (pass_data_df_finish): ...new pass_data instance and...
14552 (make_pass_df_finish): ...new function.
14553 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
14554 rtl_opt_pass along with...
14555 (pass_data_rtl_dse1): ...new pass_data instance and...
14556 (make_pass_rtl_dse1): ...new function.
14557 (pass_rtl_dse2): Convert from a global struct to a subclass of
14558 rtl_opt_pass along with...
14559 (pass_data_rtl_dse2): ...new pass_data instance and...
14560 (make_pass_rtl_dse2): ...new function.
14561 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
14562 subclass of rtl_opt_pass along with...
14563 (pass_data_dwarf2_frame): ...new pass_data instance and...
14564 (make_pass_dwarf2_frame): ...new function.
14565 * except.c (pass_set_nothrow_function_flags): Convert from a global
14566 struct to a subclass of rtl_opt_pass along with...
14567 (pass_data_set_nothrow_function_flags): ...new pass_data instance
14569 (make_pass_set_nothrow_function_flags): ...new function.
14570 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
14571 subclass of rtl_opt_pass along with...
14572 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
14574 (make_pass_convert_to_eh_region_ranges): ...new function.
14575 * final.c (pass_compute_alignments): Convert from a global struct to a
14576 subclass of rtl_opt_pass along with...
14577 (pass_data_compute_alignments): ...new pass_data instance and...
14578 (make_pass_compute_alignments): ...new function.
14579 (pass_final): Convert from a global struct to a subclass of
14580 rtl_opt_pass along with...
14581 (pass_data_final): ...new pass_data instance and...
14582 (make_pass_final): ...new function.
14583 (pass_shorten_branches): Convert from a global struct to a subclass of
14584 rtl_opt_pass along with...
14585 (pass_data_shorten_branches): ...new pass_data instance and...
14586 (make_pass_shorten_branches): ...new function.
14587 (pass_clean_state): Convert from a global struct to a subclass of
14588 rtl_opt_pass along with...
14589 (pass_data_clean_state): ...new pass_data instance and...
14590 (make_pass_clean_state): ...new function.
14591 * function.c (pass_instantiate_virtual_regs): Convert from a global
14592 struct to a subclass of rtl_opt_pass along with...
14593 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
14594 (make_pass_instantiate_virtual_regs): ...new function.
14595 (pass_leaf_regs): Convert from a global struct to a subclass of
14596 rtl_opt_pass along with...
14597 (pass_data_leaf_regs): ...new pass_data instance and...
14598 (make_pass_leaf_regs): ...new function.
14599 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
14600 subclass of rtl_opt_pass along with...
14601 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
14603 (make_pass_thread_prologue_and_epilogue): ...new function.
14604 (pass_match_asm_constraints): Convert from a global struct to a
14605 subclass of rtl_opt_pass along with...
14606 (pass_data_match_asm_constraints): ...new pass_data instance and...
14607 (make_pass_match_asm_constraints): ...new function.
14608 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
14609 subclass of rtl_opt_pass along with...
14610 (pass_data_rtl_fwprop): ...new pass_data instance and...
14611 (make_pass_rtl_fwprop): ...new function.
14612 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
14613 rtl_opt_pass along with...
14614 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
14615 (make_pass_rtl_fwprop_addr): ...new function.
14616 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
14617 rtl_opt_pass along with...
14618 (pass_data_rtl_pre): ...new pass_data instance and...
14619 (make_pass_rtl_pre): ...new function.
14620 (pass_rtl_hoist): Convert from a global struct to a subclass of
14621 rtl_opt_pass along with...
14622 (pass_data_rtl_hoist): ...new pass_data instance and...
14623 (make_pass_rtl_hoist): ...new function.
14624 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
14625 subclass of gimple_opt_pass along with...
14626 (pass_data_lower_cf): ...new pass_data instance and...
14627 (make_pass_lower_cf): ...new function.
14628 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
14629 from a global struct to a subclass of gimple_opt_pass along with...
14630 (pass_data_strength_reduction): ...new pass_data instance and...
14631 (make_pass_strength_reduction): ...new function.
14632 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
14633 of rtl_opt_pass along with...
14634 (pass_data_rtl_ifcvt): ...new pass_data instance and...
14635 (make_pass_rtl_ifcvt): ...new function.
14636 (pass_if_after_combine): Convert from a global struct to a subclass of
14637 rtl_opt_pass along with...
14638 (pass_data_if_after_combine): ...new pass_data instance and...
14639 (make_pass_if_after_combine): ...new function.
14640 (pass_if_after_reload): Convert from a global struct to a subclass of
14641 rtl_opt_pass along with...
14642 (pass_data_if_after_reload): ...new pass_data instance and...
14643 (make_pass_if_after_reload): ...new function.
14644 * init-regs.c (pass_initialize_regs): Convert from a global struct to
14645 a subclass of rtl_opt_pass along with...
14646 (pass_data_initialize_regs): ...new pass_data instance and...
14647 (make_pass_initialize_regs): ...new function.
14648 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
14649 of ipa_opt_pass_d along with...
14650 (pass_data_ipa_cp): ...new pass_data instance and...
14651 (make_pass_ipa_cp): ...new function.
14652 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
14653 global struct to a subclass of gimple_opt_pass along with...
14654 (pass_data_inline_parameters): ...new pass_data instance and...
14655 (make_pass_inline_parameters): ...new function.
14656 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
14657 subclass of gimple_opt_pass along with...
14658 (pass_data_early_inline): ...new pass_data instance and...
14659 (make_pass_early_inline): ...new function.
14660 (pass_ipa_inline): Convert from a global struct to a subclass of
14661 ipa_opt_pass_d along with...
14662 (pass_data_ipa_inline): ...new pass_data instance and...
14663 (make_pass_ipa_inline): ...new function.
14664 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
14665 struct to a subclass of gimple_opt_pass along with...
14666 (pass_data_local_pure_const): ...new pass_data instance and...
14667 (make_pass_local_pure_const): ...new function.
14668 (pass_ipa_pure_const): Convert from a global struct to a subclass of
14669 ipa_opt_pass_d along with...
14670 (pass_data_ipa_pure_const): ...new pass_data instance and...
14671 (make_pass_ipa_pure_const): ...new function.
14672 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
14673 to a subclass of ipa_opt_pass_d along with...
14674 (pass_data_ipa_reference): ...new pass_data instance and...
14675 (make_pass_ipa_reference): ...new function.
14676 * ipa-split.c (pass_split_functions): Convert from a global struct to
14677 a subclass of gimple_opt_pass along with...
14678 (pass_data_split_functions): ...new pass_data instance and...
14679 (make_pass_split_functions): ...new function.
14680 (pass_feedback_split_functions): Convert from a global struct to a
14681 subclass of gimple_opt_pass along with...
14682 (pass_data_feedback_split_functions): ...new pass_data instance and...
14683 (make_pass_feedback_split_functions): ...new function.
14684 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
14685 global struct to a subclass of simple_ipa_opt_pass along with...
14686 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
14688 (make_pass_ipa_function_and_variable_visibility): ...new function.
14689 (pass_ipa_free_inline_summary): Convert from a global struct to a
14690 subclass of simple_ipa_opt_pass along with...
14691 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
14692 (make_pass_ipa_free_inline_summary): ...new function.
14693 (pass_ipa_whole_program_visibility): Convert from a global struct to a
14694 subclass of ipa_opt_pass_d along with...
14695 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
14697 (make_pass_ipa_whole_program_visibility): ...new function.
14698 (pass_ipa_profile): Convert from a global struct to a subclass of
14699 ipa_opt_pass_d along with...
14700 (pass_data_ipa_profile): ...new pass_data instance and...
14701 (make_pass_ipa_profile): ...new function.
14702 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
14703 ipa_opt_pass_d along with...
14704 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
14705 (make_pass_ipa_cdtor_merge): ...new function.
14706 * ira.c (pass_ira): Convert from a global struct to a subclass of
14707 rtl_opt_pass along with...
14708 (pass_data_ira): ...new pass_data instance and...
14709 (make_pass_ira): ...new function.
14710 (pass_reload): Convert from a global struct to a subclass of
14711 rtl_opt_pass along with...
14712 (pass_data_reload): ...new pass_data instance and...
14713 (make_pass_reload): ...new function.
14714 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
14715 subclass of rtl_opt_pass along with...
14716 (pass_data_cleanup_barriers): ...new pass_data instance and...
14717 (make_pass_cleanup_barriers): ...new function.
14718 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
14719 of rtl_opt_pass along with...
14720 (pass_data_loop2): ...new pass_data instance and...
14721 (make_pass_loop2): ...new function.
14722 (pass_rtl_loop_init): Convert from a global struct to a subclass of
14723 rtl_opt_pass along with...
14724 (pass_data_rtl_loop_init): ...new pass_data instance and...
14725 (make_pass_rtl_loop_init): ...new function.
14726 (pass_rtl_loop_done): Convert from a global struct to a subclass of
14727 rtl_opt_pass along with...
14728 (pass_data_rtl_loop_done): ...new pass_data instance and...
14729 (make_pass_rtl_loop_done): ...new function.
14730 (pass_rtl_move_loop_invariants): Convert from a global struct to a
14731 subclass of rtl_opt_pass along with...
14732 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
14733 (make_pass_rtl_move_loop_invariants): ...new function.
14734 (pass_rtl_unswitch): Convert from a global struct to a subclass of
14735 rtl_opt_pass along with...
14736 (pass_data_rtl_unswitch): ...new pass_data instance and...
14737 (make_pass_rtl_unswitch): ...new function.
14738 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
14739 subclass of rtl_opt_pass along with...
14740 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
14742 (make_pass_rtl_unroll_and_peel_loops): ...new function.
14743 (pass_rtl_doloop): Convert from a global struct to a subclass of
14744 rtl_opt_pass along with...
14745 (pass_data_rtl_doloop): ...new pass_data instance and...
14746 (make_pass_rtl_doloop): ...new function.
14747 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
14748 a subclass of rtl_opt_pass along with...
14749 (pass_data_lower_subreg): ...new pass_data instance and...
14750 (make_pass_lower_subreg): ...new function.
14751 (pass_lower_subreg2): Convert from a global struct to a subclass of
14752 rtl_opt_pass along with...
14753 (pass_data_lower_subreg2): ...new pass_data instance and...
14754 (make_pass_lower_subreg2): ...new function.
14755 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
14756 struct to a subclass of ipa_opt_pass_d along with...
14757 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
14758 (make_pass_ipa_lto_gimple_out): ...new function.
14759 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
14760 of ipa_opt_pass_d along with...
14761 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
14762 (make_pass_ipa_lto_finish_out): ...new function.
14763 * mode-switching.c (pass_mode_switching): Convert from a global struct
14764 to a subclass of rtl_opt_pass along with...
14765 (pass_data_mode_switching): ...new pass_data instance and...
14766 (make_pass_mode_switching): ...new function.
14767 * modulo-sched.c (pass_sms): Convert from a global struct to a
14768 subclass of rtl_opt_pass along with...
14769 (pass_data_sms): ...new pass_data instance and...
14770 (make_pass_sms): ...new function.
14771 * omp-low.c (pass_expand_omp): Convert from a global struct to a
14772 subclass of gimple_opt_pass along with...
14773 (pass_data_expand_omp): ...new pass_data instance and...
14774 (make_pass_expand_omp): ...new function.
14775 (pass_lower_omp): Convert from a global struct to a subclass of
14776 gimple_opt_pass along with...
14777 (pass_data_lower_omp): ...new pass_data instance and...
14778 (make_pass_lower_omp): ...new function.
14779 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
14780 of gimple_opt_pass along with...
14781 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
14782 (make_pass_diagnose_omp_blocks): ...new function.
14783 * passes.c (pass_early_local_passes): Convert from a global struct to
14784 a subclass of simple_ipa_opt_pass along with...
14785 (pass_data_early_local_passes): ...new pass_data instance and...
14786 (make_pass_early_local_passes): ...new function.
14787 (pass_all_early_optimizations): Convert from a global struct to a
14788 subclass of gimple_opt_pass along with...
14789 (pass_data_all_early_optimizations): ...new pass_data instance and...
14790 (make_pass_all_early_optimizations): ...new function.
14791 (pass_all_optimizations): Convert from a global struct to a subclass
14792 of gimple_opt_pass along with...
14793 (pass_data_all_optimizations): ...new pass_data instance and...
14794 (make_pass_all_optimizations): ...new function.
14795 (pass_all_optimizations_g): Convert from a global struct to a subclass
14796 of gimple_opt_pass along with...
14797 (pass_data_all_optimizations_g): ...new pass_data instance and...
14798 (make_pass_all_optimizations_g): ...new function.
14799 (pass_rest_of_compilation): Convert from a global struct to a subclass
14800 of rtl_opt_pass along with...
14801 (pass_data_rest_of_compilation): ...new pass_data instance and...
14802 (make_pass_rest_of_compilation): ...new function.
14803 (pass_postreload): Convert from a global struct to a subclass of
14804 rtl_opt_pass along with...
14805 (pass_data_postreload): ...new pass_data instance and...
14806 (make_pass_postreload): ...new function.
14807 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
14808 subclass of rtl_opt_pass along with...
14809 (pass_data_gcse2): ...new pass_data instance and...
14810 (make_pass_gcse2): ...new function.
14811 * postreload.c (pass_postreload_cse): Convert from a global struct to
14812 a subclass of rtl_opt_pass along with...
14813 (pass_data_postreload_cse): ...new pass_data instance and...
14814 (make_pass_postreload_cse): ...new function.
14815 * predict.c (pass_profile): Convert from a global struct to a subclass
14816 of gimple_opt_pass along with...
14817 (pass_data_profile): ...new pass_data instance and...
14818 (make_pass_profile): ...new function.
14819 (pass_strip_predict_hints): Convert from a global struct to a subclass
14820 of gimple_opt_pass along with...
14821 (pass_data_strip_predict_hints): ...new pass_data instance and...
14822 (make_pass_strip_predict_hints): ...new function.
14823 * recog.c (pass_peephole2): Convert from a global struct to a subclass
14824 of rtl_opt_pass along with...
14825 (pass_data_peephole2): ...new pass_data instance and...
14826 (make_pass_peephole2): ...new function.
14827 (pass_split_all_insns): Convert from a global struct to a subclass of
14828 rtl_opt_pass along with...
14829 (pass_data_split_all_insns): ...new pass_data instance and...
14830 (make_pass_split_all_insns): ...new function.
14831 (pass_split_after_reload): Convert from a global struct to a subclass
14832 of rtl_opt_pass along with...
14833 (pass_data_split_after_reload): ...new pass_data instance and...
14834 (make_pass_split_after_reload): ...new function.
14835 (pass_split_before_regstack): Convert from a global struct to a
14836 subclass of rtl_opt_pass along with...
14837 (pass_data_split_before_regstack): ...new pass_data instance and...
14838 (make_pass_split_before_regstack): ...new function.
14839 (pass_split_before_sched2): Convert from a global struct to a subclass
14840 of rtl_opt_pass along with...
14841 (pass_data_split_before_sched2): ...new pass_data instance and...
14842 (make_pass_split_before_sched2): ...new function.
14843 (pass_split_for_shorten_branches): Convert from a global struct to a
14844 subclass of rtl_opt_pass along with...
14845 (pass_data_split_for_shorten_branches): ...new pass_data instance
14847 (make_pass_split_for_shorten_branches): ...new function.
14848 * ree.c (pass_ree): Convert from a global struct to a subclass of
14849 rtl_opt_pass along with...
14850 (pass_data_ree): ...new pass_data instance and...
14851 (make_pass_ree): ...new function.
14852 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
14853 subclass of rtl_opt_pass along with...
14854 (pass_data_stack_regs): ...new pass_data instance and...
14855 (make_pass_stack_regs): ...new function.
14856 (pass_stack_regs_run): Convert from a global struct to a subclass of
14857 rtl_opt_pass along with...
14858 (pass_data_stack_regs_run): ...new pass_data instance and...
14859 (make_pass_stack_regs_run): ...new function.
14860 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
14861 subclass of rtl_opt_pass along with...
14862 (pass_data_cprop_hardreg): ...new pass_data instance and...
14863 (make_pass_cprop_hardreg): ...new function.
14864 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
14865 subclass of rtl_opt_pass along with...
14866 (pass_data_reginfo_init): ...new pass_data instance and...
14867 (make_pass_reginfo_init): ...new function.
14868 * regmove.c (pass_regmove): Convert from a global struct to a subclass
14869 of rtl_opt_pass along with...
14870 (pass_data_regmove): ...new pass_data instance and...
14871 (make_pass_regmove): ...new function.
14872 * regrename.c (pass_regrename): Convert from a global struct to a
14873 subclass of rtl_opt_pass along with...
14874 (pass_data_regrename): ...new pass_data instance and...
14875 (make_pass_regrename): ...new function.
14876 * reorg.c (pass_delay_slots): Convert from a global struct to a
14877 subclass of rtl_opt_pass along with...
14878 (pass_data_delay_slots): ...new pass_data instance and...
14879 (make_pass_delay_slots): ...new function.
14880 (pass_machine_reorg): Convert from a global struct to a subclass of
14881 rtl_opt_pass along with...
14882 (pass_data_machine_reorg): ...new pass_data instance and...
14883 (make_pass_machine_reorg): ...new function.
14884 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
14885 of rtl_opt_pass along with...
14886 (pass_data_sched): ...new pass_data instance and...
14887 (make_pass_sched): ...new function.
14888 (pass_sched2): Convert from a global struct to a subclass of
14889 rtl_opt_pass along with...
14890 (pass_data_sched2): ...new pass_data instance and...
14891 (make_pass_sched2): ...new function.
14892 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
14893 to a subclass of rtl_opt_pass along with...
14894 (pass_data_stack_ptr_mod): ...new pass_data instance and...
14895 (make_pass_stack_ptr_mod): ...new function.
14896 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
14897 to a subclass of rtl_opt_pass along with...
14898 (pass_data_rtl_store_motion): ...new pass_data instance and...
14899 (make_pass_rtl_store_motion): ...new function.
14900 * tracer.c (pass_tracer): Convert from a global struct to a subclass
14901 of gimple_opt_pass along with...
14902 (pass_data_tracer): ...new pass_data instance and...
14903 (make_pass_tracer): ...new function.
14904 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
14905 to a subclass of gimple_opt_pass along with...
14906 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
14907 (make_pass_diagnose_tm_blocks): ...new function.
14908 (pass_lower_tm): Convert from a global struct to a subclass of
14909 gimple_opt_pass along with...
14910 (pass_data_lower_tm): ...new pass_data instance and...
14911 (make_pass_lower_tm): ...new function.
14912 (pass_tm_init): Convert from a global struct to a subclass of
14913 gimple_opt_pass along with...
14914 (pass_data_tm_init): ...new pass_data instance and...
14915 (make_pass_tm_init): ...new function.
14916 (pass_tm_mark): Convert from a global struct to a subclass of
14917 gimple_opt_pass along with...
14918 (pass_data_tm_mark): ...new pass_data instance and...
14919 (make_pass_tm_mark): ...new function.
14920 (pass_tm_edges): Convert from a global struct to a subclass of
14921 gimple_opt_pass along with...
14922 (pass_data_tm_edges): ...new pass_data instance and...
14923 (make_pass_tm_edges): ...new function.
14924 (pass_tm_memopt): Convert from a global struct to a subclass of
14925 gimple_opt_pass along with...
14926 (pass_data_tm_memopt): ...new pass_data instance and...
14927 (make_pass_tm_memopt): ...new function.
14928 (pass_ipa_tm): Convert from a global struct to a subclass of
14929 simple_ipa_opt_pass along with...
14930 (pass_data_ipa_tm): ...new pass_data instance and...
14931 (make_pass_ipa_tm): ...new function.
14932 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
14933 subclass of gimple_opt_pass along with...
14934 (pass_data_call_cdce): ...new pass_data instance and...
14935 (make_pass_call_cdce): ...new function.
14936 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
14937 subclass of gimple_opt_pass along with...
14938 (pass_data_build_cfg): ...new pass_data instance and...
14939 (make_pass_build_cfg): ...new function.
14940 (pass_split_crit_edges): Convert from a global struct to a subclass of
14941 gimple_opt_pass along with...
14942 (pass_data_split_crit_edges): ...new pass_data instance and...
14943 (make_pass_split_crit_edges): ...new function.
14944 (pass_warn_function_return): Convert from a global struct to a
14945 subclass of gimple_opt_pass along with...
14946 (pass_data_warn_function_return): ...new pass_data instance and...
14947 (make_pass_warn_function_return): ...new function.
14948 (pass_warn_function_noreturn): Convert from a global struct to a
14949 subclass of gimple_opt_pass along with...
14950 (pass_data_warn_function_noreturn): ...new pass_data instance and...
14951 (make_pass_warn_function_noreturn): ...new function.
14952 (pass_warn_unused_result): Convert from a global struct to a subclass
14953 of gimple_opt_pass along with...
14954 (pass_data_warn_unused_result): ...new pass_data instance and...
14955 (make_pass_warn_unused_result): ...new function.
14956 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
14957 a subclass of gimple_opt_pass along with...
14958 (pass_data_merge_phi): ...new pass_data instance and...
14959 (make_pass_merge_phi): ...new function.
14960 * tree-complex.c (pass_lower_complex): Convert from a global struct to
14961 a subclass of gimple_opt_pass along with...
14962 (pass_data_lower_complex): ...new pass_data instance and...
14963 (make_pass_lower_complex): ...new function.
14964 (pass_lower_complex_O0): Convert from a global struct to a subclass of
14965 gimple_opt_pass along with...
14966 (pass_data_lower_complex_O0): ...new pass_data instance and...
14967 (make_pass_lower_complex_O0): ...new function.
14968 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
14969 subclass of gimple_opt_pass along with...
14970 (pass_data_lower_eh): ...new pass_data instance and...
14971 (make_pass_lower_eh): ...new function.
14972 (pass_refactor_eh): Convert from a global struct to a subclass of
14973 gimple_opt_pass along with...
14974 (pass_data_refactor_eh): ...new pass_data instance and...
14975 (make_pass_refactor_eh): ...new function.
14976 (pass_lower_resx): Convert from a global struct to a subclass of
14977 gimple_opt_pass along with...
14978 (pass_data_lower_resx): ...new pass_data instance and...
14979 (make_pass_lower_resx): ...new function.
14980 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
14981 of gimple_opt_pass along with...
14982 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
14983 (make_pass_lower_eh_dispatch): ...new function.
14984 (pass_cleanup_eh): Convert from a global struct to a subclass of
14985 gimple_opt_pass along with...
14986 (pass_data_cleanup_eh): ...new pass_data instance and...
14987 (make_pass_cleanup_eh): ...new function.
14988 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
14989 to a subclass of simple_ipa_opt_pass along with...
14990 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
14991 (make_pass_ipa_lower_emutls): ...new function.
14992 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
14993 a subclass of gimple_opt_pass along with...
14994 (pass_data_if_conversion): ...new pass_data instance and...
14995 (make_pass_if_conversion): ...new function.
14996 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
14997 subclass of gimple_opt_pass along with...
14998 (pass_data_build_ssa): ...new pass_data instance and...
14999 (make_pass_build_ssa): ...new function.
15000 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
15001 global struct to a subclass of gimple_opt_pass along with...
15002 (pass_data_loop_distribution): ...new pass_data instance and...
15003 (make_pass_loop_distribution): ...new function.
15004 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
15005 subclass of gimple_opt_pass along with...
15006 (pass_data_mudflap_1): ...new pass_data instance and...
15007 (make_pass_mudflap_1): ...new function.
15008 (pass_mudflap_2): Convert from a global struct to a subclass of
15009 gimple_opt_pass along with...
15010 (pass_data_mudflap_2): ...new pass_data instance and...
15011 (make_pass_mudflap_2): ...new function.
15012 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
15013 subclass of gimple_opt_pass along with...
15014 (pass_data_mudflap_1): ...new pass_data instance and...
15015 (make_pass_mudflap_1): ...new function.
15016 (pass_mudflap_2): Convert from a global struct to a subclass of
15017 gimple_opt_pass along with...
15018 (pass_data_mudflap_2): ...new pass_data instance and...
15019 (make_pass_mudflap_2): ...new function.
15020 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
15021 gimple_opt_pass along with...
15022 (pass_data_nrv): ...new pass_data instance and...
15023 (make_pass_nrv): ...new function.
15024 (pass_return_slot): Convert from a global struct to a subclass of
15025 gimple_opt_pass along with...
15026 (pass_data_return_slot): ...new pass_data instance and...
15027 (make_pass_return_slot): ...new function.
15028 * tree-object-size.c (pass_object_sizes): Convert from a global struct
15029 to a subclass of gimple_opt_pass along with...
15030 (pass_data_object_sizes): ...new pass_data instance and...
15031 (make_pass_object_sizes): ...new function.
15032 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
15033 global struct to a subclass of gimple_opt_pass along with...
15034 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
15036 (make_pass_cleanup_cfg_post_optimizing): ...new function.
15037 (pass_fixup_cfg): Convert from a global struct to a subclass of
15038 gimple_opt_pass along with...
15039 (pass_data_fixup_cfg): ...new pass_data instance and...
15040 (make_pass_fixup_cfg): ...new function.
15041 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
15042 (make_pass_mudflap_1): ...new function.
15043 (pass_mudflap_2): Replace declaration with that of...
15044 (make_pass_mudflap_2): ...new function.
15045 (pass_asan): Replace declaration with that of...
15046 (make_pass_asan): ...new function.
15047 (pass_asan_O0): Replace declaration with that of...
15048 (make_pass_asan_O0): ...new function.
15049 (pass_tsan): Replace declaration with that of...
15050 (make_pass_tsan): ...new function.
15051 (pass_tsan_O0): Replace declaration with that of...
15052 (make_pass_tsan_O0): ...new function.
15053 (pass_lower_cf): Replace declaration with that of...
15054 (make_pass_lower_cf): ...new function.
15055 (pass_refactor_eh): Replace declaration with that of...
15056 (make_pass_refactor_eh): ...new function.
15057 (pass_lower_eh): Replace declaration with that of...
15058 (make_pass_lower_eh): ...new function.
15059 (pass_lower_eh_dispatch): Replace declaration with that of...
15060 (make_pass_lower_eh_dispatch): ...new function.
15061 (pass_lower_resx): Replace declaration with that of...
15062 (make_pass_lower_resx): ...new function.
15063 (pass_build_cfg): Replace declaration with that of...
15064 (make_pass_build_cfg): ...new function.
15065 (pass_early_tree_profile): Replace declaration with that of...
15066 (make_pass_early_tree_profile): ...new function.
15067 (pass_cleanup_eh): Replace declaration with that of...
15068 (make_pass_cleanup_eh): ...new function.
15069 (pass_sra): Replace declaration with that of...
15070 (make_pass_sra): ...new function.
15071 (pass_sra_early): Replace declaration with that of...
15072 (make_pass_sra_early): ...new function.
15073 (pass_early_ipa_sra): Replace declaration with that of...
15074 (make_pass_early_ipa_sra): ...new function.
15075 (pass_tail_recursion): Replace declaration with that of...
15076 (make_pass_tail_recursion): ...new function.
15077 (pass_tail_calls): Replace declaration with that of...
15078 (make_pass_tail_calls): ...new function.
15079 (pass_tree_loop): Replace declaration with that of...
15080 (make_pass_tree_loop): ...new function.
15081 (pass_tree_loop_init): Replace declaration with that of...
15082 (make_pass_tree_loop_init): ...new function.
15083 (pass_lim): Replace declaration with that of...
15084 (make_pass_lim): ...new function.
15085 (pass_tree_unswitch): Replace declaration with that of...
15086 (make_pass_tree_unswitch): ...new function.
15087 (pass_predcom): Replace declaration with that of...
15088 (make_pass_predcom): ...new function.
15089 (pass_iv_canon): Replace declaration with that of...
15090 (make_pass_iv_canon): ...new function.
15091 (pass_scev_cprop): Replace declaration with that of...
15092 (make_pass_scev_cprop): ...new function.
15093 (pass_empty_loop): Replace declaration with that of...
15094 (make_pass_empty_loop): ...new function.
15095 (pass_record_bounds): Replace declaration with that of...
15096 (make_pass_record_bounds): ...new function.
15097 (pass_graphite): Replace declaration with that of...
15098 (make_pass_graphite): ...new function.
15099 (pass_graphite_transforms): Replace declaration with that of...
15100 (make_pass_graphite_transforms): ...new function.
15101 (pass_if_conversion): Replace declaration with that of...
15102 (make_pass_if_conversion): ...new function.
15103 (pass_loop_distribution): Replace declaration with that of...
15104 (make_pass_loop_distribution): ...new function.
15105 (pass_vectorize): Replace declaration with that of...
15106 (make_pass_vectorize): ...new function.
15107 (pass_slp_vectorize): Replace declaration with that of...
15108 (make_pass_slp_vectorize): ...new function.
15109 (pass_complete_unroll): Replace declaration with that of...
15110 (make_pass_complete_unroll): ...new function.
15111 (pass_complete_unrolli): Replace declaration with that of...
15112 (make_pass_complete_unrolli): ...new function.
15113 (pass_parallelize_loops): Replace declaration with that of...
15114 (make_pass_parallelize_loops): ...new function.
15115 (pass_loop_prefetch): Replace declaration with that of...
15116 (make_pass_loop_prefetch): ...new function.
15117 (pass_iv_optimize): Replace declaration with that of...
15118 (make_pass_iv_optimize): ...new function.
15119 (pass_tree_loop_done): Replace declaration with that of...
15120 (make_pass_tree_loop_done): ...new function.
15121 (pass_ch): Replace declaration with that of...
15122 (make_pass_ch): ...new function.
15123 (pass_ccp): Replace declaration with that of...
15124 (make_pass_ccp): ...new function.
15125 (pass_phi_only_cprop): Replace declaration with that of...
15126 (make_pass_phi_only_cprop): ...new function.
15127 (pass_build_ssa): Replace declaration with that of...
15128 (make_pass_build_ssa): ...new function.
15129 (pass_build_alias): Replace declaration with that of...
15130 (make_pass_build_alias): ...new function.
15131 (pass_build_ealias): Replace declaration with that of...
15132 (make_pass_build_ealias): ...new function.
15133 (pass_dominator): Replace declaration with that of...
15134 (make_pass_dominator): ...new function.
15135 (pass_dce): Replace declaration with that of...
15136 (make_pass_dce): ...new function.
15137 (pass_dce_loop): Replace declaration with that of...
15138 (make_pass_dce_loop): ...new function.
15139 (pass_cd_dce): Replace declaration with that of...
15140 (make_pass_cd_dce): ...new function.
15141 (pass_call_cdce): Replace declaration with that of...
15142 (make_pass_call_cdce): ...new function.
15143 (pass_merge_phi): Replace declaration with that of...
15144 (make_pass_merge_phi): ...new function.
15145 (pass_split_crit_edges): Replace declaration with that of...
15146 (make_pass_split_crit_edges): ...new function.
15147 (pass_pre): Replace declaration with that of...
15148 (make_pass_pre): ...new function.
15149 (pass_profile): Replace declaration with that of...
15150 (make_pass_profile): ...new function.
15151 (pass_strip_predict_hints): Replace declaration with that of...
15152 (make_pass_strip_predict_hints): ...new function.
15153 (pass_lower_complex_O0): Replace declaration with that of...
15154 (make_pass_lower_complex_O0): ...new function.
15155 (pass_lower_complex): Replace declaration with that of...
15156 (make_pass_lower_complex): ...new function.
15157 (pass_lower_vector): Replace declaration with that of...
15158 (make_pass_lower_vector): ...new function.
15159 (pass_lower_vector_ssa): Replace declaration with that of...
15160 (make_pass_lower_vector_ssa): ...new function.
15161 (pass_lower_omp): Replace declaration with that of...
15162 (make_pass_lower_omp): ...new function.
15163 (pass_diagnose_omp_blocks): Replace declaration with that of...
15164 (make_pass_diagnose_omp_blocks): ...new function.
15165 (pass_expand_omp): Replace declaration with that of...
15166 (make_pass_expand_omp): ...new function.
15167 (pass_expand_omp_ssa): Replace declaration with that of...
15168 (make_pass_expand_omp_ssa): ...new function.
15169 (pass_object_sizes): Replace declaration with that of...
15170 (make_pass_object_sizes): ...new function.
15171 (pass_strlen): Replace declaration with that of...
15172 (make_pass_strlen): ...new function.
15173 (pass_fold_builtins): Replace declaration with that of...
15174 (make_pass_fold_builtins): ...new function.
15175 (pass_stdarg): Replace declaration with that of...
15176 (make_pass_stdarg): ...new function.
15177 (pass_early_warn_uninitialized): Replace declaration with that of...
15178 (make_pass_early_warn_uninitialized): ...new function.
15179 (pass_late_warn_uninitialized): Replace declaration with that of...
15180 (make_pass_late_warn_uninitialized): ...new function.
15181 (pass_cse_reciprocals): Replace declaration with that of...
15182 (make_pass_cse_reciprocals): ...new function.
15183 (pass_cse_sincos): Replace declaration with that of...
15184 (make_pass_cse_sincos): ...new function.
15185 (pass_optimize_bswap): Replace declaration with that of...
15186 (make_pass_optimize_bswap): ...new function.
15187 (pass_optimize_widening_mul): Replace declaration with that of...
15188 (make_pass_optimize_widening_mul): ...new function.
15189 (pass_warn_function_return): Replace declaration with that of...
15190 (make_pass_warn_function_return): ...new function.
15191 (pass_warn_function_noreturn): Replace declaration with that of...
15192 (make_pass_warn_function_noreturn): ...new function.
15193 (pass_cselim): Replace declaration with that of...
15194 (make_pass_cselim): ...new function.
15195 (pass_phiopt): Replace declaration with that of...
15196 (make_pass_phiopt): ...new function.
15197 (pass_forwprop): Replace declaration with that of...
15198 (make_pass_forwprop): ...new function.
15199 (pass_phiprop): Replace declaration with that of...
15200 (make_pass_phiprop): ...new function.
15201 (pass_tree_ifcombine): Replace declaration with that of...
15202 (make_pass_tree_ifcombine): ...new function.
15203 (pass_dse): Replace declaration with that of...
15204 (make_pass_dse): ...new function.
15205 (pass_nrv): Replace declaration with that of...
15206 (make_pass_nrv): ...new function.
15207 (pass_rename_ssa_copies): Replace declaration with that of...
15208 (make_pass_rename_ssa_copies): ...new function.
15209 (pass_sink_code): Replace declaration with that of...
15210 (make_pass_sink_code): ...new function.
15211 (pass_fre): Replace declaration with that of...
15212 (make_pass_fre): ...new function.
15213 (pass_check_data_deps): Replace declaration with that of...
15214 (make_pass_check_data_deps): ...new function.
15215 (pass_copy_prop): Replace declaration with that of...
15216 (make_pass_copy_prop): ...new function.
15217 (pass_vrp): Replace declaration with that of...
15218 (make_pass_vrp): ...new function.
15219 (pass_uncprop): Replace declaration with that of...
15220 (make_pass_uncprop): ...new function.
15221 (pass_return_slot): Replace declaration with that of...
15222 (make_pass_return_slot): ...new function.
15223 (pass_reassoc): Replace declaration with that of...
15224 (make_pass_reassoc): ...new function.
15225 (pass_rebuild_cgraph_edges): Replace declaration with that of...
15226 (make_pass_rebuild_cgraph_edges): ...new function.
15227 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
15228 (make_pass_remove_cgraph_callee_edges): ...new function.
15229 (pass_build_cgraph_edges): Replace declaration with that of...
15230 (make_pass_build_cgraph_edges): ...new function.
15231 (pass_local_pure_const): Replace declaration with that of...
15232 (make_pass_local_pure_const): ...new function.
15233 (pass_tracer): Replace declaration with that of...
15234 (make_pass_tracer): ...new function.
15235 (pass_warn_unused_result): Replace declaration with that of...
15236 (make_pass_warn_unused_result): ...new function.
15237 (pass_diagnose_tm_blocks): Replace declaration with that of...
15238 (make_pass_diagnose_tm_blocks): ...new function.
15239 (pass_lower_tm): Replace declaration with that of...
15240 (make_pass_lower_tm): ...new function.
15241 (pass_tm_init): Replace declaration with that of...
15242 (make_pass_tm_init): ...new function.
15243 (pass_tm_mark): Replace declaration with that of...
15244 (make_pass_tm_mark): ...new function.
15245 (pass_tm_memopt): Replace declaration with that of...
15246 (make_pass_tm_memopt): ...new function.
15247 (pass_tm_edges): Replace declaration with that of...
15248 (make_pass_tm_edges): ...new function.
15249 (pass_split_functions): Replace declaration with that of...
15250 (make_pass_split_functions): ...new function.
15251 (pass_feedback_split_functions): Replace declaration with that of...
15252 (make_pass_feedback_split_functions): ...new function.
15253 (pass_strength_reduction): Replace declaration with that of...
15254 (make_pass_strength_reduction): ...new function.
15255 (pass_ipa_lower_emutls): Replace declaration with that of...
15256 (make_pass_ipa_lower_emutls): ...new function.
15257 (pass_ipa_function_and_variable_visibility): Replace declaration with
15259 (make_pass_ipa_function_and_variable_visibility): ...new function.
15260 (pass_ipa_tree_profile): Replace declaration with that of...
15261 (make_pass_ipa_tree_profile): ...new function.
15262 (pass_early_local_passes): Replace declaration with that of...
15263 (make_pass_early_local_passes): ...new function.
15264 (pass_ipa_whole_program_visibility): Replace declaration with that
15266 (make_pass_ipa_whole_program_visibility): ...new function.
15267 (pass_ipa_lto_gimple_out): Replace declaration with that of...
15268 (make_pass_ipa_lto_gimple_out): ...new function.
15269 (pass_ipa_increase_alignment): Replace declaration with that of...
15270 (make_pass_ipa_increase_alignment): ...new function.
15271 (pass_ipa_inline): Replace declaration with that of...
15272 (make_pass_ipa_inline): ...new function.
15273 (pass_ipa_free_lang_data): Replace declaration with that of...
15274 (make_pass_ipa_free_lang_data): ...new function.
15275 (pass_ipa_free_inline_summary): Replace declaration with that of...
15276 (make_pass_ipa_free_inline_summary): ...new function.
15277 (pass_ipa_cp): Replace declaration with that of...
15278 (make_pass_ipa_cp): ...new function.
15279 (pass_ipa_reference): Replace declaration with that of...
15280 (make_pass_ipa_reference): ...new function.
15281 (pass_ipa_pure_const): Replace declaration with that of...
15282 (make_pass_ipa_pure_const): ...new function.
15283 (pass_ipa_pta): Replace declaration with that of...
15284 (make_pass_ipa_pta): ...new function.
15285 (pass_ipa_lto_finish_out): Replace declaration with that of...
15286 (make_pass_ipa_lto_finish_out): ...new function.
15287 (pass_ipa_tm): Replace declaration with that of...
15288 (make_pass_ipa_tm): ...new function.
15289 (pass_ipa_profile): Replace declaration with that of...
15290 (make_pass_ipa_profile): ...new function.
15291 (pass_ipa_cdtor_merge): Replace declaration with that of...
15292 (make_pass_ipa_cdtor_merge): ...new function.
15293 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
15295 (make_pass_cleanup_cfg_post_optimizing): ...new function.
15296 (pass_init_datastructures): Replace declaration with that of...
15297 (make_pass_init_datastructures): ...new function.
15298 (pass_fixup_cfg): Replace declaration with that of...
15299 (make_pass_fixup_cfg): ...new function.
15300 (pass_expand): Replace declaration with that of...
15301 (make_pass_expand): ...new function.
15302 (pass_instantiate_virtual_regs): Replace declaration with that of...
15303 (make_pass_instantiate_virtual_regs): ...new function.
15304 (pass_rtl_fwprop): Replace declaration with that of...
15305 (make_pass_rtl_fwprop): ...new function.
15306 (pass_rtl_fwprop_addr): Replace declaration with that of...
15307 (make_pass_rtl_fwprop_addr): ...new function.
15308 (pass_jump): Replace declaration with that of...
15309 (make_pass_jump): ...new function.
15310 (pass_jump2): Replace declaration with that of...
15311 (make_pass_jump2): ...new function.
15312 (pass_lower_subreg): Replace declaration with that of...
15313 (make_pass_lower_subreg): ...new function.
15314 (pass_cse): Replace declaration with that of...
15315 (make_pass_cse): ...new function.
15316 (pass_fast_rtl_dce): Replace declaration with that of...
15317 (make_pass_fast_rtl_dce): ...new function.
15318 (pass_ud_rtl_dce): Replace declaration with that of...
15319 (make_pass_ud_rtl_dce): ...new function.
15320 (pass_rtl_dce): Replace declaration with that of...
15321 (make_pass_rtl_dce): ...new function.
15322 (pass_rtl_dse1): Replace declaration with that of...
15323 (make_pass_rtl_dse1): ...new function.
15324 (pass_rtl_dse2): Replace declaration with that of...
15325 (make_pass_rtl_dse2): ...new function.
15326 (pass_rtl_dse3): Replace declaration with that of...
15327 (make_pass_rtl_dse3): ...new function.
15328 (pass_rtl_cprop): Replace declaration with that of...
15329 (make_pass_rtl_cprop): ...new function.
15330 (pass_rtl_pre): Replace declaration with that of...
15331 (make_pass_rtl_pre): ...new function.
15332 (pass_rtl_hoist): Replace declaration with that of...
15333 (make_pass_rtl_hoist): ...new function.
15334 (pass_rtl_store_motion): Replace declaration with that of...
15335 (make_pass_rtl_store_motion): ...new function.
15336 (pass_cse_after_global_opts): Replace declaration with that of...
15337 (make_pass_cse_after_global_opts): ...new function.
15338 (pass_rtl_ifcvt): Replace declaration with that of...
15339 (make_pass_rtl_ifcvt): ...new function.
15340 (pass_into_cfg_layout_mode): Replace declaration with that of...
15341 (make_pass_into_cfg_layout_mode): ...new function.
15342 (pass_outof_cfg_layout_mode): Replace declaration with that of...
15343 (make_pass_outof_cfg_layout_mode): ...new function.
15344 (pass_loop2): Replace declaration with that of...
15345 (make_pass_loop2): ...new function.
15346 (pass_rtl_loop_init): Replace declaration with that of...
15347 (make_pass_rtl_loop_init): ...new function.
15348 (pass_rtl_move_loop_invariants): Replace declaration with that of...
15349 (make_pass_rtl_move_loop_invariants): ...new function.
15350 (pass_rtl_unswitch): Replace declaration with that of...
15351 (make_pass_rtl_unswitch): ...new function.
15352 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
15353 (make_pass_rtl_unroll_and_peel_loops): ...new function.
15354 (pass_rtl_doloop): Replace declaration with that of...
15355 (make_pass_rtl_doloop): ...new function.
15356 (pass_rtl_loop_done): Replace declaration with that of...
15357 (make_pass_rtl_loop_done): ...new function.
15358 (pass_web): Replace declaration with that of...
15359 (make_pass_web): ...new function.
15360 (pass_cse2): Replace declaration with that of...
15361 (make_pass_cse2): ...new function.
15362 (pass_df_initialize_opt): Replace declaration with that of...
15363 (make_pass_df_initialize_opt): ...new function.
15364 (pass_df_initialize_no_opt): Replace declaration with that of...
15365 (make_pass_df_initialize_no_opt): ...new function.
15366 (pass_reginfo_init): Replace declaration with that of...
15367 (make_pass_reginfo_init): ...new function.
15368 (pass_inc_dec): Replace declaration with that of...
15369 (make_pass_inc_dec): ...new function.
15370 (pass_stack_ptr_mod): Replace declaration with that of...
15371 (make_pass_stack_ptr_mod): ...new function.
15372 (pass_initialize_regs): Replace declaration with that of...
15373 (make_pass_initialize_regs): ...new function.
15374 (pass_combine): Replace declaration with that of...
15375 (make_pass_combine): ...new function.
15376 (pass_if_after_combine): Replace declaration with that of...
15377 (make_pass_if_after_combine): ...new function.
15378 (pass_ree): Replace declaration with that of...
15379 (make_pass_ree): ...new function.
15380 (pass_partition_blocks): Replace declaration with that of...
15381 (make_pass_partition_blocks): ...new function.
15382 (pass_match_asm_constraints): Replace declaration with that of...
15383 (make_pass_match_asm_constraints): ...new function.
15384 (pass_regmove): Replace declaration with that of...
15385 (make_pass_regmove): ...new function.
15386 (pass_split_all_insns): Replace declaration with that of...
15387 (make_pass_split_all_insns): ...new function.
15388 (pass_fast_rtl_byte_dce): Replace declaration with that of...
15389 (make_pass_fast_rtl_byte_dce): ...new function.
15390 (pass_lower_subreg2): Replace declaration with that of...
15391 (make_pass_lower_subreg2): ...new function.
15392 (pass_mode_switching): Replace declaration with that of...
15393 (make_pass_mode_switching): ...new function.
15394 (pass_sms): Replace declaration with that of...
15395 (make_pass_sms): ...new function.
15396 (pass_sched): Replace declaration with that of...
15397 (make_pass_sched): ...new function.
15398 (pass_ira): Replace declaration with that of...
15399 (make_pass_ira): ...new function.
15400 (pass_reload): Replace declaration with that of...
15401 (make_pass_reload): ...new function.
15402 (pass_clean_state): Replace declaration with that of...
15403 (make_pass_clean_state): ...new function.
15404 (pass_branch_prob): Replace declaration with that of...
15405 (make_pass_branch_prob): ...new function.
15406 (pass_value_profile_transformations): Replace declaration with that
15408 (make_pass_value_profile_transformations): ...new function.
15409 (pass_postreload_cse): Replace declaration with that of...
15410 (make_pass_postreload_cse): ...new function.
15411 (pass_gcse2): Replace declaration with that of...
15412 (make_pass_gcse2): ...new function.
15413 (pass_split_after_reload): Replace declaration with that of...
15414 (make_pass_split_after_reload): ...new function.
15415 (pass_branch_target_load_optimize1): Replace declaration with that
15417 (make_pass_branch_target_load_optimize1): ...new function.
15418 (pass_thread_prologue_and_epilogue): Replace declaration with that
15420 (make_pass_thread_prologue_and_epilogue): ...new function.
15421 (pass_stack_adjustments): Replace declaration with that of...
15422 (make_pass_stack_adjustments): ...new function.
15423 (pass_peephole2): Replace declaration with that of...
15424 (make_pass_peephole2): ...new function.
15425 (pass_if_after_reload): Replace declaration with that of...
15426 (make_pass_if_after_reload): ...new function.
15427 (pass_regrename): Replace declaration with that of...
15428 (make_pass_regrename): ...new function.
15429 (pass_cprop_hardreg): Replace declaration with that of...
15430 (make_pass_cprop_hardreg): ...new function.
15431 (pass_reorder_blocks): Replace declaration with that of...
15432 (make_pass_reorder_blocks): ...new function.
15433 (pass_branch_target_load_optimize2): Replace declaration with that
15435 (make_pass_branch_target_load_optimize2): ...new function.
15436 (pass_leaf_regs): Replace declaration with that of...
15437 (make_pass_leaf_regs): ...new function.
15438 (pass_split_before_sched2): Replace declaration with that of...
15439 (make_pass_split_before_sched2): ...new function.
15440 (pass_compare_elim_after_reload): Replace declaration with that of...
15441 (make_pass_compare_elim_after_reload): ...new function.
15442 (pass_sched2): Replace declaration with that of...
15443 (make_pass_sched2): ...new function.
15444 (pass_stack_regs): Replace declaration with that of...
15445 (make_pass_stack_regs): ...new function.
15446 (pass_stack_regs_run): Replace declaration with that of...
15447 (make_pass_stack_regs_run): ...new function.
15448 (pass_df_finish): Replace declaration with that of...
15449 (make_pass_df_finish): ...new function.
15450 (pass_compute_alignments): Replace declaration with that of...
15451 (make_pass_compute_alignments): ...new function.
15452 (pass_duplicate_computed_gotos): Replace declaration with that of...
15453 (make_pass_duplicate_computed_gotos): ...new function.
15454 (pass_variable_tracking): Replace declaration with that of...
15455 (make_pass_variable_tracking): ...new function.
15456 (pass_free_cfg): Replace declaration with that of...
15457 (make_pass_free_cfg): ...new function.
15458 (pass_machine_reorg): Replace declaration with that of...
15459 (make_pass_machine_reorg): ...new function.
15460 (pass_cleanup_barriers): Replace declaration with that of...
15461 (make_pass_cleanup_barriers): ...new function.
15462 (pass_delay_slots): Replace declaration with that of...
15463 (make_pass_delay_slots): ...new function.
15464 (pass_split_for_shorten_branches): Replace declaration with that of...
15465 (make_pass_split_for_shorten_branches): ...new function.
15466 (pass_split_before_regstack): Replace declaration with that of...
15467 (make_pass_split_before_regstack): ...new function.
15468 (pass_convert_to_eh_region_ranges): Replace declaration with that
15470 (make_pass_convert_to_eh_region_ranges): ...new function.
15471 (pass_shorten_branches): Replace declaration with that of...
15472 (make_pass_shorten_branches): ...new function.
15473 (pass_set_nothrow_function_flags): Replace declaration with that of...
15474 (make_pass_set_nothrow_function_flags): ...new function.
15475 (pass_dwarf2_frame): Replace declaration with that of...
15476 (make_pass_dwarf2_frame): ...new function.
15477 (pass_final): Replace declaration with that of...
15478 (make_pass_final): ...new function.
15479 (pass_rtl_seqabstr): Replace declaration with that of...
15480 (make_pass_rtl_seqabstr): ...new function.
15481 (pass_release_ssa_names): Replace declaration with that of...
15482 (make_pass_release_ssa_names): ...new function.
15483 (pass_early_inline): Replace declaration with that of...
15484 (make_pass_early_inline): ...new function.
15485 (pass_inline_parameters): Replace declaration with that of...
15486 (make_pass_inline_parameters): ...new function.
15487 (pass_update_address_taken): Replace declaration with that of...
15488 (make_pass_update_address_taken): ...new function.
15489 (pass_convert_switch): Replace declaration with that of...
15490 (make_pass_convert_switch): ...new function.
15491 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
15492 to a subclass of simple_ipa_opt_pass along with...
15493 (pass_data_ipa_tree_profile): ...new pass_data instance and...
15494 (make_pass_ipa_tree_profile): ...new function.
15495 * tree-sra.c (pass_sra_early): Convert from a global struct to a
15496 subclass of gimple_opt_pass along with...
15497 (pass_data_sra_early): ...new pass_data instance and...
15498 (make_pass_sra_early): ...new function.
15499 (pass_sra): Convert from a global struct to a subclass of
15500 gimple_opt_pass along with...
15501 (pass_data_sra): ...new pass_data instance and...
15502 (make_pass_sra): ...new function.
15503 (pass_early_ipa_sra): Convert from a global struct to a subclass of
15504 gimple_opt_pass along with...
15505 (pass_data_early_ipa_sra): ...new pass_data instance and...
15506 (make_pass_early_ipa_sra): ...new function.
15507 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
15508 subclass of gimple_opt_pass along with...
15509 (pass_data_ccp): ...new pass_data instance and...
15510 (make_pass_ccp): ...new function.
15511 (pass_fold_builtins): Convert from a global struct to a subclass of
15512 gimple_opt_pass along with...
15513 (pass_data_fold_builtins): ...new pass_data instance and...
15514 (make_pass_fold_builtins): ...new function.
15515 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
15516 subclass of gimple_opt_pass along with...
15517 (pass_data_copy_prop): ...new pass_data instance and...
15518 (make_pass_copy_prop): ...new function.
15519 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
15520 global struct to a subclass of gimple_opt_pass along with...
15521 (pass_data_rename_ssa_copies): ...new pass_data instance and...
15522 (make_pass_rename_ssa_copies): ...new function.
15523 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
15524 subclass of gimple_opt_pass along with...
15525 (pass_data_dce): ...new pass_data instance and...
15526 (make_pass_dce): ...new function.
15527 (pass_dce_loop): Convert from a global struct to a subclass of
15528 gimple_opt_pass along with...
15529 (pass_data_dce_loop): ...new pass_data instance and...
15530 (make_pass_dce_loop): ...new function.
15531 (pass_cd_dce): Convert from a global struct to a subclass of
15532 gimple_opt_pass along with...
15533 (pass_data_cd_dce): ...new pass_data instance and...
15534 (make_pass_cd_dce): ...new function.
15535 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
15536 subclass of gimple_opt_pass along with...
15537 (pass_data_dominator): ...new pass_data instance and...
15538 (make_pass_dominator): ...new function.
15539 (pass_phi_only_cprop): Convert from a global struct to a subclass of
15540 gimple_opt_pass along with...
15541 (pass_data_phi_only_cprop): ...new pass_data instance and...
15542 (make_pass_phi_only_cprop): ...new function.
15543 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
15544 subclass of gimple_opt_pass along with...
15545 (pass_data_dse): ...new pass_data instance and...
15546 (make_pass_dse): ...new function.
15547 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
15548 a subclass of gimple_opt_pass along with...
15549 (pass_data_forwprop): ...new pass_data instance and...
15550 (make_pass_forwprop): ...new function.
15551 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
15552 struct to a subclass of gimple_opt_pass along with...
15553 (pass_data_tree_ifcombine): ...new pass_data instance and...
15554 (make_pass_tree_ifcombine): ...new function.
15555 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
15556 subclass of gimple_opt_pass along with...
15557 (pass_data_ch): ...new pass_data instance and...
15558 (make_pass_ch): ...new function.
15559 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
15560 subclass of gimple_opt_pass along with...
15561 (pass_data_tree_loop): ...new pass_data instance and...
15562 (make_pass_tree_loop): ...new function.
15563 (pass_tree_loop_init): Convert from a global struct to a subclass of
15564 gimple_opt_pass along with...
15565 (pass_data_tree_loop_init): ...new pass_data instance and...
15566 (make_pass_tree_loop_init): ...new function.
15567 (pass_lim): Convert from a global struct to a subclass of
15568 gimple_opt_pass along with...
15569 (pass_data_lim): ...new pass_data instance and...
15570 (make_pass_lim): ...new function.
15571 (pass_tree_unswitch): Convert from a global struct to a subclass of
15572 gimple_opt_pass along with...
15573 (pass_data_tree_unswitch): ...new pass_data instance and...
15574 (make_pass_tree_unswitch): ...new function.
15575 (pass_predcom): Convert from a global struct to a subclass of
15576 gimple_opt_pass along with...
15577 (pass_data_predcom): ...new pass_data instance and...
15578 (make_pass_predcom): ...new function.
15579 (pass_vectorize): Convert from a global struct to a subclass of
15580 gimple_opt_pass along with...
15581 (pass_data_vectorize): ...new pass_data instance and...
15582 (make_pass_vectorize): ...new function.
15583 (pass_graphite): Convert from a global struct to a subclass of
15584 gimple_opt_pass along with...
15585 (pass_data_graphite): ...new pass_data instance and...
15586 (make_pass_graphite): ...new function.
15587 (pass_graphite_transforms): Convert from a global struct to a subclass
15588 of gimple_opt_pass along with...
15589 (pass_data_graphite_transforms): ...new pass_data instance and...
15590 (make_pass_graphite_transforms): ...new function.
15591 (pass_check_data_deps): Convert from a global struct to a subclass of
15592 gimple_opt_pass along with...
15593 (pass_data_check_data_deps): ...new pass_data instance and...
15594 (make_pass_check_data_deps): ...new function.
15595 (pass_iv_canon): Convert from a global struct to a subclass of
15596 gimple_opt_pass along with...
15597 (pass_data_iv_canon): ...new pass_data instance and...
15598 (make_pass_iv_canon): ...new function.
15599 (pass_scev_cprop): Convert from a global struct to a subclass of
15600 gimple_opt_pass along with...
15601 (pass_data_scev_cprop): ...new pass_data instance and...
15602 (make_pass_scev_cprop): ...new function.
15603 (pass_record_bounds): Convert from a global struct to a subclass of
15604 gimple_opt_pass along with...
15605 (pass_data_record_bounds): ...new pass_data instance and...
15606 (make_pass_record_bounds): ...new function.
15607 (pass_complete_unroll): Convert from a global struct to a subclass of
15608 gimple_opt_pass along with...
15609 (pass_data_complete_unroll): ...new pass_data instance and...
15610 (make_pass_complete_unroll): ...new function.
15611 (pass_complete_unrolli): Convert from a global struct to a subclass of
15612 gimple_opt_pass along with...
15613 (pass_data_complete_unrolli): ...new pass_data instance and...
15614 (make_pass_complete_unrolli): ...new function.
15615 (pass_parallelize_loops): Convert from a global struct to a subclass
15616 of gimple_opt_pass along with...
15617 (pass_data_parallelize_loops): ...new pass_data instance and...
15618 (make_pass_parallelize_loops): ...new function.
15619 (pass_loop_prefetch): Convert from a global struct to a subclass of
15620 gimple_opt_pass along with...
15621 (pass_data_loop_prefetch): ...new pass_data instance and...
15622 (make_pass_loop_prefetch): ...new function.
15623 (pass_iv_optimize): Convert from a global struct to a subclass of
15624 gimple_opt_pass along with...
15625 (pass_data_iv_optimize): ...new pass_data instance and...
15626 (make_pass_iv_optimize): ...new function.
15627 (pass_tree_loop_done): Convert from a global struct to a subclass of
15628 gimple_opt_pass along with...
15629 (pass_data_tree_loop_done): ...new pass_data instance and...
15630 (make_pass_tree_loop_done): ...new function.
15631 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
15632 struct to a subclass of gimple_opt_pass along with...
15633 (pass_data_cse_reciprocals): ...new pass_data instance and...
15634 (make_pass_cse_reciprocals): ...new function.
15635 (pass_cse_sincos): Convert from a global struct to a subclass of
15636 gimple_opt_pass along with...
15637 (pass_data_cse_sincos): ...new pass_data instance and...
15638 (make_pass_cse_sincos): ...new function.
15639 (pass_optimize_bswap): Convert from a global struct to a subclass of
15640 gimple_opt_pass along with...
15641 (pass_data_optimize_bswap): ...new pass_data instance and...
15642 (make_pass_optimize_bswap): ...new function.
15643 (pass_optimize_widening_mul): Convert from a global struct to a
15644 subclass of gimple_opt_pass along with...
15645 (pass_data_optimize_widening_mul): ...new pass_data instance and...
15646 (make_pass_optimize_widening_mul): ...new function.
15647 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
15648 subclass of gimple_opt_pass along with...
15649 (pass_data_phiopt): ...new pass_data instance and...
15650 (make_pass_phiopt): ...new function.
15651 (pass_cselim): Convert from a global struct to a subclass of
15652 gimple_opt_pass along with...
15653 (pass_data_cselim): ...new pass_data instance and...
15654 (make_pass_cselim): ...new function.
15655 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
15656 subclass of gimple_opt_pass along with...
15657 (pass_data_phiprop): ...new pass_data instance and...
15658 (make_pass_phiprop): ...new function.
15659 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
15660 subclass of gimple_opt_pass along with...
15661 (pass_data_pre): ...new pass_data instance and...
15662 (make_pass_pre): ...new function.
15663 (pass_fre): Convert from a global struct to a subclass of
15664 gimple_opt_pass along with...
15665 (pass_data_fre): ...new pass_data instance and...
15666 (make_pass_fre): ...new function.
15667 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
15668 subclass of gimple_opt_pass along with...
15669 (pass_data_reassoc): ...new pass_data instance and...
15670 (make_pass_reassoc): ...new function.
15671 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
15672 subclass of gimple_opt_pass along with...
15673 (pass_data_sink_code): ...new pass_data instance and...
15674 (make_pass_sink_code): ...new function.
15675 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
15676 subclass of gimple_opt_pass along with...
15677 (pass_data_strlen): ...new pass_data instance and...
15678 (make_pass_strlen): ...new function.
15679 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
15680 struct to a subclass of gimple_opt_pass along with...
15681 (pass_data_build_alias): ...new pass_data instance and...
15682 (make_pass_build_alias): ...new function.
15683 (pass_build_ealias): Convert from a global struct to a subclass of
15684 gimple_opt_pass along with...
15685 (pass_data_build_ealias): ...new pass_data instance and...
15686 (make_pass_build_ealias): ...new function.
15687 (pass_ipa_pta): Convert from a global struct to a subclass of
15688 simple_ipa_opt_pass along with...
15689 (pass_data_ipa_pta): ...new pass_data instance and...
15690 (make_pass_ipa_pta): ...new function.
15691 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
15692 subclass of gimple_opt_pass along with...
15693 (pass_data_uncprop): ...new pass_data instance and...
15694 (make_pass_uncprop): ...new function.
15695 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
15696 global struct to a subclass of gimple_opt_pass along with...
15697 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
15698 (make_pass_late_warn_uninitialized): ...new function.
15699 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
15700 to a subclass of gimple_opt_pass along with...
15701 (pass_data_init_datastructures): ...new pass_data instance and...
15702 (make_pass_init_datastructures): ...new function.
15703 (pass_early_warn_uninitialized): Convert from a global struct to a
15704 subclass of gimple_opt_pass along with...
15705 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
15706 (make_pass_early_warn_uninitialized): ...new function.
15707 (pass_update_address_taken): Convert from a global struct to a
15708 subclass of gimple_opt_pass along with...
15709 (pass_data_update_address_taken): ...new pass_data instance and...
15710 (make_pass_update_address_taken): ...new function.
15711 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
15712 struct to a subclass of gimple_opt_pass along with...
15713 (pass_data_release_ssa_names): ...new pass_data instance and...
15714 (make_pass_release_ssa_names): ...new function.
15715 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
15716 subclass of gimple_opt_pass along with...
15717 (pass_data_stdarg): ...new pass_data instance and...
15718 (make_pass_stdarg): ...new function.
15719 * tree-switch-conversion.c (pass_convert_switch): Convert from a
15720 global struct to a subclass of gimple_opt_pass along with...
15721 (pass_data_convert_switch): ...new pass_data instance and...
15722 (make_pass_convert_switch): ...new function.
15723 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
15724 to a subclass of gimple_opt_pass along with...
15725 (pass_data_tail_recursion): ...new pass_data instance and...
15726 (make_pass_tail_recursion): ...new function.
15727 (pass_tail_calls): Convert from a global struct to a subclass of
15728 gimple_opt_pass along with...
15729 (pass_data_tail_calls): ...new pass_data instance and...
15730 (make_pass_tail_calls): ...new function.
15731 * tree-vect-generic.c (pass_lower_vector): Convert from a global
15732 struct to a subclass of gimple_opt_pass along with...
15733 (pass_data_lower_vector): ...new pass_data instance and...
15734 (make_pass_lower_vector): ...new function.
15735 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
15736 gimple_opt_pass along with...
15737 (pass_data_lower_vector_ssa): ...new pass_data instance and...
15738 (make_pass_lower_vector_ssa): ...new function.
15739 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
15740 to a subclass of gimple_opt_pass along with...
15741 (pass_data_slp_vectorize): ...new pass_data instance and...
15742 (make_pass_slp_vectorize): ...new function.
15743 (pass_ipa_increase_alignment): Convert from a global struct to a
15744 subclass of simple_ipa_opt_pass along with...
15745 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
15746 (make_pass_ipa_increase_alignment): ...new function.
15747 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
15748 gimple_opt_pass along with...
15749 (pass_data_vrp): ...new pass_data instance and...
15750 (make_pass_vrp): ...new function.
15751 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
15752 subclass of simple_ipa_opt_pass along with...
15753 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
15754 (make_pass_ipa_free_lang_data): ...new function.
15755 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
15756 gimple_opt_pass along with...
15757 (pass_data_tsan): ...new pass_data instance and...
15758 (make_pass_tsan): ...new function.
15759 (pass_tsan_O0): Convert from a global struct to a subclass of
15760 gimple_opt_pass along with...
15761 (pass_data_tsan_O0): ...new pass_data instance and...
15762 (make_pass_tsan_O0): ...new function.
15763 * var-tracking.c (pass_variable_tracking): Convert from a global
15764 struct to a subclass of rtl_opt_pass along with...
15765 (pass_data_variable_tracking): ...new pass_data instance and...
15766 (make_pass_variable_tracking): ...new function.
15767 * web.c (pass_web): Convert from a global struct to a subclass of
15768 rtl_opt_pass along with...
15769 (pass_data_web): ...new pass_data instance and...
15770 (make_pass_web): ...new function.
15771 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
15772 declaration with that of...
15773 (make_pass_mode_switch_use): ...new function.
15774 (pass_resolve_sw_modes): Replace declaration with that of...
15775 (make_pass_resolve_sw_modes): ...new function.
15776 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
15777 from a global struct to a subclass of rtl_opt_pass along with...
15778 (pass_data_mode_switch_use): ...new pass_data instance and...
15779 (make_pass_mode_switch_use): ...new function.
15780 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
15781 from a global struct to a subclass of rtl_opt_pass along with...
15782 (pass_data_resolve_sw_modes): ...new pass_data instance and...
15783 (make_pass_resolve_sw_modes): ...new function.
15784 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
15785 struct to a subclass of rtl_opt_pass along with...
15786 (pass_data_insert_vzeroupper): ...new pass_data instance and...
15787 (make_pass_insert_vzeroupper): ...new function.
15788 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
15789 global struct to a subclass of rtl_opt_pass along with...
15790 (pass_data_work_around_errata): ...new pass_data instance and...
15791 (make_pass_work_around_errata): ...new function.
15792 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
15793 struct to a subclass of rtl_opt_pass along with...
15794 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
15795 (make_pass_mips_machine_reorg2): ...new function.
15797 2013-08-05 David Malcolm <dmalcolm@redhat.com>
15799 * passes.c (pass_manager::operator new): New.
15801 2013-08-05 David Malcolm <dmalcolm@redhat.com>
15803 Handwritten part of conversion of passes to C++ classes.
15805 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
15806 (toplev.o): Add dep on PASS_MANAGER_H.
15807 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
15808 of early local pases to reflect this moving from a global to a
15809 member of gcc::pass_manager.
15810 (cgraph_add_new_function): Likewise.
15811 * lto-cgraph.c (lto_output_node): Update for conversion of
15812 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
15813 * passes.c (opt_pass::clone): New.
15814 (opt_pass::gate): New.
15815 (opt_pass::execute): New.
15816 (opt_pass::opt_pass): New.
15817 (pass_manager::execute_early_local_passes): New.
15818 (pass_manager::execute_pass_mode_switching): new.
15819 (finish_optimization_passes): Convert to...
15820 (pass_manager::finish_optimization_passes): ...this.
15821 (finish_optimization_passes): Update for conversion of passes to
15823 (register_dump_files_1): Use has_gate since we cannot portably
15824 check a vtable entry against NULL.
15825 (dump_one_pass): Likewise.
15826 (ipa_write_summaries_2): Likewise.
15827 (ipa_write_optimization_summaries_1): Likewise.
15828 (ipa_read_summaries_1): Likewise.
15829 (ipa_read_optimization_summaries_1): Likewise.
15830 (execute_ipa_stmt_fixups): Likewise.
15831 (pass_manager::pass_manager): Rewrite pass-creation, invoking
15832 pass-creation functions rather than wiring up globals, and
15833 storing the results in fields of pass_manager generated using
15834 pass-instances.def.
15835 (pass_manager::dump_profile_report): Update for conversion of
15836 passes to C++ classes.
15837 (pass_manager::execute_ipa_summary_passes): Likewise.
15838 (execute_one_ipa_transform_pass): Likewise.
15839 (execute_one_pass): Use has_gate and has_execute since we cannot
15840 portably check a vtable entry against NULL.
15841 * pass_manager.h (pass_manager::finish_optimization_passes): New.
15842 (pass_manager): Use pass-instances.def to add fields for the
15843 various pass instances.
15844 * toplev.c (finalize): Update for move of
15845 finish_optimization_passes to a method of gcc::pass_manager.
15846 * toplev.h (finish_optimization_passes): Move to method of class
15848 * tree-pass.h (struct pass_data): New.
15849 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
15850 (opt_pass::gate): Convert to virtual function.
15851 (opt_pass::~opt_pass): New.
15852 (opt_pass::clone): New.
15853 (opt_pass::execute): Convert to virtual function.
15854 (opt_pass::opt_pass): New.
15855 (opt_pass::ctxt_): new.
15856 (gimple_opt_pass): Convert to subclass of opt_pass.
15857 (gimple_opt_pass::gimple_opt_pass): New.
15858 (rtl_opt_pass): Convert to subclass of opt_pass.
15859 (rtl_opt_pass::rtl_opt_pass): New.
15860 (ipa_opt_pass_d): Convert to subclass of opt_pass.
15861 (ipa_opt_pass_d::ipa_opt_pass_d): New.
15862 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
15863 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
15864 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
15865 invocation of pass_mode_switching to reflect this moving from a
15866 global to a member of gcc::pass_manager.
15867 (ix86_option_override): Rework how pass_insert_vzeroupper is
15868 added to the pass_manager to reflect autogenerated changes.
15869 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
15871 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
15873 PR rtl-optimization/57708
15874 * recog.c (peep2_find_free_register): Validate all regs in a
15877 2013-08-05 Jan Hubicka <jh@suse.cz>
15880 * cgraph.c (verify_cgraph_node): Accept local flags from other
15882 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
15883 (function_and_variable_visibility): Likewise.
15884 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
15886 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
15888 * graph.c (init_graph_slim_pretty_print): Remove.
15889 (print_graph_cfg): Do not call it. Use local pretty printer.
15890 (start_graph_dump): Likewise.
15892 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
15894 * gimple-pretty-print.c (buffer): Remove.
15895 (initialized): Likewise.
15896 (maybe_init_pretty_print): Likewise.
15897 (print_gimple_stmt): Do not call it. Use non-static local
15898 pretty_printer variable.
15899 (print_gimple_expr): Likewise.
15900 (print_gimple_seq): Likewise.
15901 (gimple_dump_bb): Likewise.
15903 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
15905 * asan.c (asan_pp): Remove.
15906 (asan_pp_initialized): Likewise.
15907 (asan_pp_initialize): Likewise.
15908 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
15909 (asan_emit_stack_protection): Tidy. Use local pretty printer.
15910 (asan_add_global): Likewise.
15912 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
15914 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
15915 * diagnostic.h (diagnostic_flush_buffer): Adjust.
15916 * pretty-print.c (pp_formatted_text_data): Likewise.
15917 (pp_indent): Rename from pp_base_indent.
15918 (pp_format): Rename from pp_base_format.
15919 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
15920 (pp_format_verbatim): Rename from pp_base_format_verbatim.
15921 (pp_flush): Rename from pp_base_flush.
15922 (pp_set_line_maximum_length): Rename from
15923 pp_base_set_line_maximum_length.
15924 (pp_clear_output_area): Rename from pp_base_clear_output_area.
15925 (pp_set_prefix): Rename from pp_base_set_prefix.
15926 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
15927 (pp_emit_prefix): Rename from pp_base_emit_prefix.
15928 (pp_append_text): Rename from pp_base_append_text.
15929 (pp_formatted_text): Rename from pp_base_formatted_text.
15930 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
15931 (pp_remaining_character_count_for_line): Rename from
15932 pp_base_remaining_character_count_for_line.
15933 (pp_newline): Rename from pp_base_newline.
15934 (pp_character): Rename from pp_base_character.
15935 (pp_string): Rename from pp_base_string.
15936 (pp_maybe_space): Rename from pp_base_maybe_space.
15937 * asan.c (asan_pp_string): Adjust.
15938 (asan_emit_stack_protection): Likewise.
15939 (asan_add_global): Likewise.
15940 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
15941 * tree-mudflap.c (mf_varname_tree): Likewise.
15942 * tree-pretty-print.c (pp_tree_identifier): Rename from
15943 pp_base_tree_identifier.
15944 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
15945 Declare as function.
15947 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
15949 * pretty-print.h (pp_bar_bar): New.
15950 (pp_ampersand_ampersand): Likewise.
15951 (pp_less_equal): Likewise.
15952 (pp_greater_equal): Likewise.
15953 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
15954 printer functions instead of pp_string or operators and punctuators.
15955 (dump_gimple_call): Likewise.
15956 (dump_gimple_omp_for): Likewise.
15957 (dump_gimple_transaction): Likewise.
15958 (dump_gimple_phi): Likewise.
15959 (pp_gimple_stmt_1): Likewise.
15960 * sched-vis.c (print_insn): Likewise.
15961 * tree-mudflap.c (mf_varname_tree): Likewise.
15962 * tree-pretty-print.c (dump_block_node): Likewise.
15963 (dump_generic_node): Likewise.
15965 2013-08-02 Jan Hubicka <jh@suse.cz>
15967 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
15969 * lto-streamer-out.c (tree_is_indexable): Results decls and
15970 parm decls are not indexable.
15971 (DFS_write_tree_body): Do not follow args and results.
15972 (hash_tree): Likewise.
15973 (output_functions): Rearrange so struct function is needed
15974 only when real body is output; be able to also ouptut abstract
15975 functions; output DECL_ARGUMENTS and DECL_RESULT.
15976 (lto_output): When not in WPA, ale store abstract functions.
15977 (write_symbol): Do not care about RESULT_DECL.
15978 (output_symbol_p): Handle correctly sbtract decls.
15979 * lto-streamer-in.c (input_function): Rearrange so struct
15980 function can be NULL at entry; allow streaming of
15981 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
15982 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
15983 sanity check during LTO.
15984 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
15985 RESULT_DECl and DECL_ARGUMENTS.
15986 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
15989 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
15991 * pretty-print.h (pp_underscore): New.
15993 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
15994 printer functions instead of pp_character.
15995 (dump_binary_rhs): Likewise.
15996 (dump_ternary_rhs): Likewise.
15997 (dump_gimple_call_args): Likewise.
15998 (pp_points_to_solution): Likewise.
15999 (dump_gimple_call): Likewise.
16000 (dump_gimple_switch): Likewise.
16001 (dump_gimple_cond): Likewise.
16002 (dump_gimple_bind): Likewise.
16003 (dump_gimple_try): Likewise.
16004 (dump_gimple_omp_for): Likewise.
16005 (dump_gimple_omp_continue): Likewise.
16006 (dump_gimple_omp_single): Likewise.
16007 (dump_gimple_omp_sections): Likewise.
16008 (dump_gimple_omp_block): Likewise.
16009 (dump_gimple_omp_critical): Likewise.
16010 (dump_gimple_transaction): Likewise.
16011 (dump_gimple_asm): Likewise.
16012 (dump_gimple_phi): Likewise.
16013 (dump_gimple_omp_parallel): Likewise.
16014 (dump_gimple_omp_task): Likewise.
16015 (dump_gimple_omp_atomic_load): Likewise.
16016 (dump_gimple_omp_atomic_store): Likewise.
16017 (dump_gimple_mem_ops): Likewise.
16018 (pp_gimple_stmt_1): Likewise.
16019 (pp_cfg_jump): Likewise.
16020 (dump_implicit_edges): Likewise.
16021 (gimple_dump_bb_for_graph): Likewise.
16022 * graph.c (draw_cfg_node): Likewise.
16023 * langhooks.c (lhd_print_error_function): Likewise.
16024 * sched-vis.c (print_exp): Likewise.
16025 (print_value): Likewise.
16026 (print_pattern): Likewise.
16027 (print_insn): Likewise.
16028 (rtl_dump_bb_for_graph): Likewise.
16029 * tree-pretty-print.c (dump_function_declaration): Likewise.
16030 (dump_array_domain): Likewise.
16031 (dump_omp_clause): Likewise.
16032 (dump_location): Likewise.
16033 (dump_generic_node): Likewise.
16034 (print_struct_decl): Likewise.
16035 * diagnostic.c (diagnostic_show_locus): Use pp_space.
16037 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
16039 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
16040 candidate table when replacing a candidate statement.
16041 (replace_rhs_if_not_dup): Likewise.
16042 (replace_one_candidate): Likewise.
16044 2013-08-02 Jan Hubicka <jh@suse.cz>
16045 Martin Liska <marxin.liska@gmail.com>
16047 * cgraphunit.c (add_new_function): Fix logic when adding from
16049 (assemble_thunk): Rename to ...
16050 (expand_thunk); .. this one; export; get it working with
16051 general functions; make produced gimple valid.
16052 * cgraph.h (expand_thunk): Declare.
16054 2013-08-02 Jan Hubicka <jh@suse.cz>
16056 * ipa-cp.c (gather_context_independent_values): Use
16057 ipa_get_param_move_cost.
16058 (get_replacement_map): Remove PARAM; move parameter folding
16060 (create_specialized_node): Update.
16061 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
16062 assert that we have gimple body; update move_cost.
16063 (count_formal_params): Assert that we have gimple body.
16064 (ipa_dump_param): New function.
16065 (ipa_alloc_node_params): Break out from ...
16066 (ipa_initialize_node_params): ... here.
16067 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
16068 (ipa_write_node_info): Stream move costs.
16069 (ipa_read_node_info): Read move costs.
16070 (ipa_update_after_lto_read): Do not recompute node params.
16071 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
16072 (ipa_get_param): Check we are not in WPA.
16073 (ipa_get_param_move_cost): New.
16074 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
16075 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
16076 parm numbers to be present.
16078 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
16080 PR rtl-optimization/58048
16081 * lra-constraints.c (process_alt_operands): Don't check asm
16082 operand on register.
16084 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
16086 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
16087 the implied StoreLoad barrier for atomic operations if before.
16089 2013-08-02 Jan Hubicka <jh@suse.cz>
16090 Martin Liska <marxin.liska@gmail.com>
16092 * cgraph.c (cgraph_function_body_availability): Do not check
16094 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
16095 symtab_node_availability): Declare.
16096 * ipa.c (can_replace_by_local_alias): New.
16097 (function_and_variable_visibility): Use it.
16098 * symtab.c (symtab_for_node_and_aliases,
16099 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
16101 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
16103 PR rtl-optimization/57963
16104 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
16105 (lra_constraints): Use them.
16107 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
16109 * config/arm/types.md (define_attr "type"): Add "load_acq"
16111 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
16113 (cortex_a53_store1): Likewise.
16115 2013-08-01 Jan Hubicka <jh@suse.cz>
16117 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
16118 partitions are not needed.
16120 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
16122 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
16123 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
16124 MAYBE_NON_Q_CLASS_P where appropriate.
16126 2013-08-01 Jan Hubicka <jh@suse.cz>
16128 * cgraph.h (release_function_body): Declare.
16129 * tree.c (free_lang_data_in_decl): Free, parameters and return values
16130 of unused delcarations.
16132 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16134 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
16135 RTL form when subtracting a constant.
16137 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16139 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
16140 Generate canonical plus rtx with negated immediate instead of minus
16142 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
16144 2013-08-01 Jan Hubicka <jh@suse.cz>
16146 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
16147 (cgraph_release_function_body): Likewise.
16148 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
16149 * cgraph.h (cgrpah_node): Rename abstract_and_needed
16150 to used_as_abstract_origin.
16151 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
16152 symbols used as abstract origins.
16153 * cgraphunit.c (analyze_functions): Update.
16154 * ipa.c (symtab_remove_unreachable_nodes): Recompute
16155 used_as_abstract_origin.
16156 * tree-inline.c (tree_function_versioning): Update
16157 used_as_abstract_origin; be ready for DECL_RESULT and
16158 DECL_ARGUMENTS to be NULL.
16160 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
16161 for abstract functions.
16162 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
16165 2013-08-01 Jan Hubicka <jh@suse.cz>
16167 * profile.c (compute_value_histograms): Fix thinko.
16169 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
16171 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
16172 aarch-common-protos.h to extra_headers.
16173 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
16174 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
16175 * config/aarch64/t-aarch64 (aarch-common.o): Define.
16177 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
16179 * config/aarch64/aarch64.md (define_attr "type"): Delete.
16180 Include "../arm/types.md". Define "type" attribute for all patterns.
16181 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
16184 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
16186 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
16187 to support power8 load fusion.
16188 (fusion_gpr_mem_load): Likewise.
16190 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
16192 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
16193 declarations for power8 load fusion.
16194 (emit_fusion_gpr_load): Likewise.
16196 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
16197 tuning for power8, turn on fusion mode by default. Turn on sign
16198 extending fusion mode if normal fusion mode is on, and we are at
16200 (fusion_gpr_load_p): New function, return true if we can fuse an
16201 addis instruction with a dependent load to a GPR.
16202 (emit_fusion_gpr_load): Emit the instructions for power8 load
16205 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
16206 (VSX load fusion peepholes): New peepholes to fuse together an
16207 addi instruction with a VSX load instruction.
16209 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
16210 peepholes to fuse an addis instruction with a load to a GPR base
16211 register. If we are supporting sign extending fusions, convert
16212 sign extending loads to zero extending loads and add an explicit
16215 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
16217 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
16218 aarch-common-protos.h to extra_headers.
16219 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
16220 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
16221 (arm_early_store_addr_dep): Likewise.
16222 (arm_no_early_alu_shift_dep): Likewise.
16223 (arm_no_early_alu_shift_value_dep): Likewise.
16224 (arm_no_early_mul_dep): Likewise.
16225 (arm_no_early_store_addr_dep): Likewise.
16226 (arm_mac_accumulator_is_mul_result): Likewise.
16227 (arm_mac_accumulator_is_result): Likewise.
16228 * config/arm/aarch-common.c: ... here. New file.
16229 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
16231 (arm_early_store_addr_dep): Likewise.
16232 (arm_no_early_alu_shift_dep): Likewise.
16233 (arm_no_early_alu_shift_value_dep): Likewise.
16234 (arm_no_early_mul_dep): Likewise.
16235 (arm_no_early_store_addr_dep): Likewise.
16236 (arm_mac_accumulator_is_mul_result): Likewise.
16237 (arm_mac_accumulator_is_result): Likewise.
16238 * config/arm/aarch-common-protos.h: ... here. New file.
16239 * config/arm/t-arm (aarch-common.o): Define.
16241 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
16243 * config/arm/arm.md: Include new file "types.md".
16244 (define_attr "type"): Move from here to ...
16245 (define_attr "mul32"): Likewise.
16246 (define_attr "mul64"): Likewise.
16247 * config/arm/types.md: ... here. New file.
16249 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
16251 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
16252 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
16254 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
16256 * gen-pass-instances.awk: Fix offset of substr().
16258 2013-07-31 David Malcolm <dmalcolm@redhat.com>
16260 * Makefile.in (pass-instances.def): New.
16261 (passes.o): Replace dependency on passes.def with one on
16264 * gen-pass-instances.awk: New.
16266 * passes.c (pass_manager::pass_manager): Use pass-instances.def
16267 rather than passes.def, updating local definition of NEXT_PASS
16268 macro to add an extra NUM parameter (currently unused).
16270 2013-07-30 David Malcolm <dmalcolm@redhat.com>
16272 * Makefile.in (PASS_MANAGER_H): New.
16273 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
16274 (passes.o): Likewise.
16275 (statistics.o): Likewise.
16276 (cgraphunit.o): Likewise.
16277 (context.o): Depend on PASS_MANAGER_H.
16279 * pass_manager.h: New.
16281 * cgraphunit.c (cgraph_add_new_function): Update for moves
16282 of globals to fields of pass_manager.
16283 (analyze_function): Likewise.
16284 (expand_function): Likewise.
16285 (ipa_passes): Likewise.
16286 (compile): Likewise.
16288 * context.c (context::context): New.
16289 * context.h (context::context): New.
16290 (context::get_passes): New.
16291 (context::passes_): New.
16293 * lto-cgraph.c (input_node): Update for moves of globals to
16294 fields of pass_manager.
16296 * passes.c (all_passes): Remove, in favor of a field of the
16297 same name within the new class pass_manager.
16298 (all_small_ipa_passes): Likewise.
16299 (all_lowering_passes): Likewise.
16300 (all_regular_ipa_passes): Likewise.
16301 (all_late_ipa_passes): Likewise.
16302 (all_lto_gen_passes): Likewise.
16303 (passes_by_id): Likewise.
16304 (passes_by_id_size): Likewise.
16305 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
16306 the new class pass_manager.
16307 (set_pass_for_id): Convert to...
16308 (pass_manager::set_pass_for_id): ...method.
16309 (get_pass_for_id): Convert to...
16310 (pass_manager::get_pass_for_id): ...method.
16311 (register_one_dump_file): Move body of implementation into...
16312 (pass_manager::register_one_dump_file): ...here.
16313 (register_dump_files_1): Convert to...
16314 (pass_manager::register_dump_files_1): ...method.
16315 (register_dump_files): Convert to...
16316 (pass_manager::register_dump_files): ...method.
16317 (create_pass_tab): Update for moves of globals to fields of
16319 (dump_passes): Move body of implementation into...
16320 (pass_manager::dump_passes): ...here.
16321 (register_pass): Move body of implementation into...
16322 (pass_manager::register_pass): ...here.
16323 (init_optimization_passes): Convert into...
16324 (pass_manager::pass_manager): ...constructor for new
16325 pass_manager class, and initialize the pass_lists array.
16326 (check_profile_consistency): Update for moves of globals to
16327 fields of pass_manager.
16328 (dump_profile_report): Move body of implementation into...
16329 (pass_manager::dump_profile_report): ...here.
16330 (ipa_write_summaries_1): Update for moves of pass lists from
16331 being globals to fields of pass_manager.
16332 (ipa_write_optimization_summaries): Likewise.
16333 (ipa_read_summaries): Likewise.
16334 (ipa_read_optimization_summaries): Likewise.
16335 (execute_all_ipa_stmt_fixups): Likewise.
16337 * statistics.c (statistics_fini): Update for moves of globals to
16338 fields of pass_manager.
16340 * toplev.c (general_init): Replace call to
16341 init_optimization_passes with construction of the pass_manager
16344 * tree-pass.h (all_passes): Remove, in favor of a field of the
16345 same name within the new class pass_manager.
16346 (all_small_ipa_passes): Likewise.
16347 (all_lowering_passes): Likewise.
16348 (all_regular_ipa_passes): Likewise.
16349 (all_lto_gen_passes): Likewise.
16350 (all_late_ipa_passes): Likewise.
16351 (passes_by_id): Likewise.
16352 (passes_by_id_size): Likewise.
16353 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
16354 the new class pass_manager.
16355 (get_pass_for_id): Remove.
16357 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
16359 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
16362 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
16364 * arm.md (mulhi3): New expand pattern.
16366 2013-07-30 Jan Hubicka <jh@suse.cz>
16367 Martin Liska <marxin.liska@gmail.com>
16369 * profile.c (compute_value_histograms): Do not ICE when
16370 there is mismatch only on some counters.
16372 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
16374 PR rtl-optimization/57637
16375 * function.c (move_insn_for_shrink_wrap): Also check the
16376 GEN set of the LIVE problem for the liveness analysis
16377 if it exists, otherwise give up.
16379 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
16381 PR tree-optimization/57993
16382 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
16383 replaced statement in the candidate table.
16384 (phi_add_costs): Return infinite cost when the hidden basis does
16385 not dominate all phis on which the candidate is dependent.
16386 (replace_one_candidate): Record replaced statement in the
16389 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
16391 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
16392 (ashlv2si3): New expander.
16393 (*ashlv2si3_i): New define_insn_and_split.
16394 * predicates.md (float_operation): Allow patterns with three
16395 basic sub-patterns.
16397 PR rtl-optimization/58021
16398 * mode-switching.c (create_pre_exit): Always split off preceding
16399 insns if we are not at the basic block head.
16401 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
16403 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
16404 (UCLIBC_DYNAMIC_LINKER): New macro.
16405 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
16407 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
16408 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
16410 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
16411 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
16412 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
16413 for SF and DF modes. Use ieee_quad_format for TF mode.
16414 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
16415 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
16416 (mips_option_override): Handle `-mnan=legacy'.
16417 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
16418 `-mabs=2008' and `-mnan=2008'.
16419 (OPTION_DEFAULT_SPECS): Add "nan" default.
16420 (ASM_SPEC): Handle `-mnan='.
16421 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
16422 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
16423 comment accordingly.
16424 (neg<mode>2): Likewise.
16425 * config/mips/mips.opt (mabs, mnan): New options.
16426 * doc/install.texi (Configuration): Document `--with-nan=' option.
16427 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
16429 (MIPS Options): Document them.
16430 * config.gcc <mips*-*-*>: Handle `--with-nan='.
16431 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
16432 * configure: Regenerate.
16433 * config.in: Regenerate.
16435 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
16437 * config/i386/i386.md (float post-reload splitters): Do not check
16438 for subregs of SSE registers.
16440 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
16441 H.J. Lu <hongjiu.lu@intel.com>
16445 * config/i386/i386.md (post-reload splitter
16446 to avoid partial SSE reg dependency stalls): New pattern.
16448 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
16450 * config/s390/s390.md ("movcc"): Swap load and store instructions.
16452 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
16454 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
16455 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
16457 2013-07-26 Cary Coutant <ccoutant@google.com>
16459 * dwarf2out.c (die_checksum_ordered): Don't include template
16460 instantiations in signature.
16461 (is_template_parameter): New function.
16462 (is_template_instantiation): New function.
16463 (generate_skeleton_bottom_up): Don't include template instantiations
16465 (generate_skeleton): Likewise.
16466 (break_out_comdat_types): Move recursive call to break out nested
16468 (prune_unused_types_mark_generic_parms_dies): Call
16469 is_template_parameter.
16471 2013-07-26 Ian Bolton <ian.bolton@arm.com>
16473 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
16474 uses vector registers.
16475 * config/aarch64/iterators.md: Add attributes rtn and vas.
16477 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16478 Richard Earnshaw <richard.earnshaw@arm.com>
16480 * combine.c (simplify_comparison): Re-canonicalize operands
16482 * config/arm/arm.md (movcond_addsi): New splitter.
16484 2013-07-25 Sterling Augustine <saugustine@google.com>
16486 * dwarf2out.c (size_of_pubnames): Move code to...
16487 (include_pubname_in_output): ...here. New.
16488 (want_pubnames): Rearrange.
16489 (output_pubnames): Call include_pubname_in_output. Move assertion.
16491 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
16493 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
16495 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
16498 * doc/extend.texi: Remove obsolete builtins. Fix
16499 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
16501 2013-07-25 Jan Hubicka <jh@suse.cz>
16503 * cgraph.c (release_function_body): Break out from ...
16504 (cgraph_release_function_body): ... this one; also release DECL_RESULT
16505 and DECL_ARGUMENTS.
16506 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
16507 old_tree in the map.
16508 (create_specialized_node): Update.
16509 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
16511 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
16512 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
16513 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
16514 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
16517 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16519 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
16520 addsi3_carryin_alt2_<optab>): Correct output template.
16522 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16524 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
16525 Adjust for arm_restrict_it.
16526 Remove trailing whitespace.
16528 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
16530 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
16531 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
16533 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
16535 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
16537 PR rtl-optimization/57960
16538 * lra-constraints.c (process_alt_operands): Use the right mode
16539 when checking strict_low.
16541 2013-07-25 Jan Hubicka <jh@suse.cz>
16543 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
16544 * cgraph.c (cgraph_remove_node): Do not release function body
16545 when in cgraph streaming.
16546 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
16547 in other partitions are not considered reachable; fix handling of
16550 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16552 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
16554 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16559 * config/arm/arm.md ("*sibcall_insn): Replace use of
16560 Ss with US. Adjust output for v5 and v4t.
16561 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
16563 * config/arm/constraints.md ("Ss"): Rename to US.
16565 2013-07-25 Terry Guo <terry.guo@arm.com>
16567 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
16568 shift_add/shift_sub0/shift_sub1 RTXs.
16570 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
16571 Anton Blanchard <anton@au1.ibm.com>
16573 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
16574 (altivec_vpks<VI_char>ss): Likewise.
16575 (altivec_vpks<VI_char>us): Likewise.
16576 (altivec_vpku<VI_char>us): Likewise.
16577 (altivec_vpku<VI_char>um): Likewise.
16579 2013-07-24 David Malcolm <dmalcolm@redhat.com>
16581 Introduce context class.
16583 * Makefile.in (CONTEXT_H): New.
16584 (OBJS): Add context.o.
16585 (toplev.o): Add CONTEXT_H to dependencies.
16588 * toplev.c (general_init): Create the singleton gcc::context instance.
16594 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
16596 PR rtl-optimization/57968
16597 * mode-switching.c (create_pre_exit): Allow instructions that
16598 don't set a return register to need a non-exit mode.
16600 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
16601 Anton Blanchard <anton@au1.ibm.com>
16603 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
16604 operands to vperm for little endian.
16605 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
16606 of lvsl to create the control mask for a vperm for little endian.
16608 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16609 Anton Blanchard <anton@au1.ibm.com>
16611 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
16612 two operands for little-endian.
16614 2013-07-23 Steve Ellcey <sellcey@mips.com>
16616 * config/mips/mips.c (mips_case_values_threshold): New.
16617 (TARGET_CASE_VALUES_THRESHOLD): Define.
16619 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16620 Anton Blanchard <anton@au1.ibm.com>
16622 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
16623 selection of field for vector splat in little endian mode.
16625 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
16627 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
16628 expanders to rs6000.md.
16629 (ior<mode>3): Likewise.
16630 (and<mode>3): Likewise.
16631 (one_cmpl<mode>2): Likewise.
16632 (nor<mode>3): Likewise.
16633 (andc<mode>3): Likewise.
16634 (eqv<mode>3): Likewise.
16635 (nand<mode>3): Likewise.
16636 (orc<mode>3): Likewise.
16638 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
16641 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
16642 to split multi-word logical operations.
16643 (rs6000_split_logical_di): Likewise.
16644 (rs6000_split_logical): Likewise.
16646 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
16647 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
16648 and allow TImode operations in 32-bit.
16649 (vsx_and<mode>3_64bit): Likewise.
16650 (vsx_ior<mode>3_32bit): Likewise.
16651 (vsx_ior<mode>3_64bit): Likewise.
16652 (vsx_xor<mode>3_32bit): Likewise.
16653 (vsx_xor<mode>3_64bit): Likewise.
16654 (vsx_one_cmpl<mode>2_32bit): Likewise.
16655 (vsx_one_cmpl<mode>2_64bit): Likewise.
16656 (vsx_nor<mode>3_32bit): Likewise.
16657 (vsx_nor<mode>3_64bit): Likewise.
16658 (vsx_andc<mode>3_32bit): Likewise.
16659 (vsx_andc<mode>3_64bit): Likewise.
16660 (vsx_eqv<mode>3_32bit): Likewise.
16661 (vsx_eqv<mode>3_64bit): Likewise.
16662 (vsx_nand<mode>3_32bit): Likewise.
16663 (vsx_nand<mode>3_64bit): Likewise.
16664 (vsx_orc<mode>3_32bit): Likewise.
16665 (vsx_orc<mode>3_64bit): Likewise.
16667 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
16668 logical types in GPRs.
16670 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
16671 logical insns to rs6000.md, and allow TImode operations in
16673 (altivec_ior<mode>3): Likewise.
16674 (altivec_xor<mode>3): Likewise.
16675 (altivec_one_cmpl<mode>2): Likewise.
16676 (altivec_nor<mode>3): Likewise.
16677 (altivec_andc<mode>3): Likewise.
16679 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
16680 attributes for moving the 128-bit logical operations into
16682 (BOOL_REGS_OUTPUT): Likewise.
16683 (BOOL_REGS_OP1): Likewise.
16684 (BOOL_REGS_OP2): Likewise.
16685 (BOOL_REGS_UNARY): Likewise.
16686 (BOOL_REGS_AND_CR0): Likewise.
16687 (one_cmpl<mode>2): Add support for DI logical operations on
16688 32-bit, splitting the operations to 32-bit.
16689 (anddi3): Likewise.
16690 (iordi3): Likewise.
16691 (xordi3): Likewise.
16692 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
16693 changes to combine the 32/64-bit code, allow logical operations on
16694 TI mode in 32-bit, and to use similar match_operator patterns like
16695 scalar mode uses. Combine the Altivec and VSX code for logical
16696 operations, and move it here.
16697 (ior<mode>3, 128-bit types): Likewise.
16698 (xor<mode>3, 128-bit types): Likewise.
16699 (one_cmpl<mode>3, 128-bit types): Likewise.
16700 (nor<mode>3, 128-bit types): Likewise.
16701 (andc<mode>3, 128-bit types): Likewise.
16702 (eqv<mode>3, 128-bit types): Likewise.
16703 (nand<mode>3, 128-bit types): Likewise.
16704 (orc<mode>3, 128-bit types): Likewise.
16705 (and<mode>3_internal): Likewise.
16706 (bool<mode>3_internal): Likewise.
16707 (boolc<mode>3_internal1): Likewise.
16708 (boolc<mode>3_internal2): Likewise.
16709 (boolcc<mode>3_internal1): Likewise.
16710 (boolcc<mode>3_internal2): Likewise.
16711 (eqv<mode>3_internal1): Likewise.
16712 (eqv<mode>3_internal2): Likewise.
16713 (one_cmpl1<mode>3_internal): Likewise.
16715 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
16717 * config/microblaze/microblaze.c (microblaze_expand_prologue):
16718 Rename flag_stack_usage to flag_stack_usage_info.
16720 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
16722 * config/microblaze/sync.md: New file.
16723 * config/microblaze/microblaze.md: Include sync.md
16724 * config/microblaze/microblaze.c: Add print_operand 'y'.
16725 * config/microblaze/constraints.md: Add memory_contraint
16726 'Q' which is a single register.
16728 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
16730 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
16732 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
16734 * reload.c (find_reloads): Exit loop once we find this operand
16735 cannot be reloaded somehow for this alternative.
16737 * reload.c (find_reloads): Exit loop once we find a hard register.
16739 * rtlanal.c (computed_jump_p): Exit loop once we find label
16742 * i386.c (ix86_pad_returns): Exit loop after setting replace.
16744 * cfgloopmanip.c (remove_path): Exit loop after setting
16747 * gensupport.c (subst_dup): Avoid loop if code is not
16748 MATCH_DUP nor MATCH_OP_DUP.
16750 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
16752 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
16754 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
16756 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
16757 true for SP_REGNUM if mode == ptr_mode.
16758 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
16759 with value R0_REGNUM + 31.
16761 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
16763 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
16764 pad pointer-typed argument downward.
16766 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
16768 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
16769 and __ILP32__ when the ILP32 model is in use.
16771 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
16773 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
16774 (aarch64_load_symref_appropriately): In the case of
16775 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
16776 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
16777 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
16778 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
16779 if 'mode' doesn't equal to 'ptr_mode'.
16780 (aarch64_output_mi_thunk): Add an assertion on the alignment of
16781 'vcall_offset'; change to call aarch64_emit_move differently depending
16782 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
16783 to calculate the upper bound of 'vcall_offset'.
16784 (aarch64_cannot_force_const_mem): Change to also return true if
16786 (aarch64_legitimize_reload_address): In the case of large
16787 displacements, add new local variable 'xmode' and an assertion
16788 based on it; change to use 'xmode' to generate the new rtx and
16790 (aarch64_asm_trampoline_template): Change to generate the template
16791 differently depending on TARGET_ILP32 or not; change to use
16792 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
16793 (aarch64_trampoline_size): Removed.
16794 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
16795 and replace immediate literals with it. Change to use 'ptr_mode'
16796 instead of 'DImode' and call convert_memory_address if the mode
16797 of 'fnaddr' doesn't equal to 'ptr_mode'.
16798 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
16800 (aarch64_elf_asm_destructor): Likewise.
16801 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
16802 on TARGET_ILP32 instead of aarch64_trampoline_size.
16803 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
16804 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
16805 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
16806 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
16807 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
16808 (storewb_pair<GPI:mode>_<P:mode>): ... this.
16809 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
16810 depending on the value of 'mode'.
16811 (add_losym_<mode>): New.
16812 (ldr_got_small_<mode>): New, based on ldr_got_small.
16813 (ldr_got_small): Remove.
16814 (ldr_got_small_sidi): New.
16815 * config/aarch64/iterators.md (P): New.
16816 (PTR): Change to 'ptr_mode' in the condition.
16818 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
16820 * config.gcc (aarch64*-*-*): Support --with-abi.
16821 (aarch64*-*-elf): Support --with-multilib-list.
16822 (aarch64*-*-linux*): Likewise.
16823 (supported_defaults): Add abi to aarch64*-*-*.
16824 * configure.ac: Mention AArch64 for --with-multilib-list.
16825 * configure: Re-generated.
16826 * config/aarch64/biarchilp32.h: New file.
16827 * config/aarch64/biarchlp64.h: New file.
16828 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
16830 (MULTILIB_DEFAULTS): Ditto.
16831 (DRIVER_SELF_SPECS): Ditto.
16832 (ASM_SPEC): Update to also substitute -mabi.
16833 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
16834 file whose name depends on -mabi= and -mbig-endian.
16835 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
16837 (POINTER_SIZE): New define.
16838 (POINTERS_EXTEND_UNSIGNED): Ditto.
16839 (enum aarch64_abi_type): New enumeration tag.
16840 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
16841 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
16842 (TARGET_ILP32): New define.
16843 * config/aarch64/aarch64.opt (mabi): New.
16844 (aarch64_abi): New.
16845 (ilp32, lp64): New values for -mabi.
16846 * config/aarch64/t-aarch64 (comma): New define.
16847 (MULTILIB_OPTIONS): Ditto.
16848 (MULTILIB_DIRNAMES): Ditto.
16849 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
16850 * doc/invoke.texi: Document -mabi for AArch64.
16852 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
16854 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
16856 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
16857 Anton Blanchard <anton@au1.ibm.com>
16859 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
16860 endianness when selecting field to splat.
16862 2013-07-22 Eric Christopher <echristo@gmail.com>
16864 * dwarf2out.c (die_odr_checksum): New function to use
16865 CHECKSUM_ macros and ULEB128 for DIE tag.
16866 (generate_type_signature): Use.
16868 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
16870 * config.gcc (sparc*-*-*): Accept leon3 processor.
16871 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
16872 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
16873 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
16874 * config/sparc/sparc.opt (enum processor_type): Add leon3.
16875 (mfix-ut699): Adjust comment.
16876 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
16877 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
16878 (CPP_CPU_SPEC): Likewise.
16879 (ASM_CPU_SPEC): Likewise.
16880 * config/sparc/sparc.c (leon3_cost): New constant.
16881 (sparc_option_override): Add leon3 support.
16882 (mem_ref): New function.
16883 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
16884 (sparc_do_work_around_errata): Look into the instruction in the delay
16885 slot and adjust accordingly. Add fix for the data cache nullify issues
16886 of the UT699. Change insertion position for the NOP.
16887 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
16888 (leon3_load): New reservation.
16889 (leon_store): Bump latency to 2.
16890 (grfpu): New automaton.
16891 (grfpu_alu): New unit.
16892 (grfpu_ds): Likewise.
16893 (leon_fp_alu): Adjust.
16894 (leon_fp_mult): Delete.
16895 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
16896 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
16897 * config/sparc/sparc.md (cpu): Add leon3.
16898 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
16899 (swapsi): Likewise.
16900 (atomic_test_and_set): Likewise.
16901 (ldstub): Likewise.
16903 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
16905 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
16906 default for R5900 targets.
16907 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
16908 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
16909 * config/mips/mips.c (mips_option_override): Report an error for
16910 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
16911 for -march=r5900 -mhard-float.
16913 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
16915 * df-problems.c (can_move_insns_across): Exit loop once we
16916 find a non-fixed, non-global register.
16918 * ipa-pure-const.c (propagate_nothrow): Exit loop after
16921 * omega.c (omega_eliminate_red): Break after setting red_found.
16922 (omega_problem_has_red_equations): Similarly after setting found.
16923 (omega_query_variable): Similarly after setting coupled.
16925 2013-07-22 Marek Polacek <polacek@redhat.com>
16927 * gimplify.c: Don't include gimple.h twice.
16929 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16931 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
16932 instead of TARGET_THUMB1.
16933 (Pz): New constraint.
16934 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
16936 (compare_negsi_si): Likewise.
16937 (compare_addsi2_op0): Likewise.
16938 (compare_addsi2_op1): Likewise.
16939 (addsi3_carryin_<optab>): Likewise.
16940 (addsi3_carryin_alt2_<optab>): Likewise.
16941 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
16942 for arm_restrict_it.
16943 (subsi3_carryin): Likewise.
16944 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
16945 (minmax_arithsi): Disable for arm_restrict_it.
16946 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
16947 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
16948 (satsi_<SAT:code>_shift): Likewise.
16949 (arm_shiftsi3): Add alternative for 16-bit encoding.
16950 (arm32_movhf): Disable for arm_restrict_it.
16951 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
16952 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
16954 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
16956 * config/arm/arm.md (attribute "insn"): Delete.
16957 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
16958 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
16959 (not_shiftsi): Update for attribute change.
16960 (not_shiftsi_compare0): Likewise.
16961 (not_shiftsi_compare0_scratch): Likewise.
16962 (arm_one_cmplsi2): Likewise.
16963 (thumb1_one_cmplsi2): Likewise.
16964 (notsi_compare0): Likewise.
16965 (notsi_compare0_scratch): Likewise.
16966 (thumb1_movdi_insn): Likewise.
16967 (arm_movsi_insn): Likewise.
16968 (movhi_insn_arch4): Likewise.
16969 (movhi_bytes): Likewise.
16970 (arm_movqi_insn): Likewise.
16971 (thumb1_movqi_insn): Likewise.
16972 (arm32_movhf): Likewise.
16973 (thumb1_movhf): Likewise.
16974 (arm_movsf_soft_insn): Likewise.
16975 (thumb1_movsf_insn): Likewise.
16976 (thumb_movdf_insn): Likewise.
16977 (movsicc_insn): Likewise.
16978 (movsfcc_soft_insn): Likewise.
16979 (and_scc): Likewise.
16980 (cond_move): Likewise.
16981 (if_move_not): Likewise.
16982 (if_not_move): Likewise.
16983 (if_shift_move): Likewise.
16984 (if_move_shift): Likewise.
16985 (if_shift_shift): Likewise.
16986 (if_not_arith): Likewise.
16987 (if_arith_not): Likewise.
16988 (cond_move_not): Likewise.
16989 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
16990 (neon_mov<mode>): Likewise.
16991 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
16992 (thumb2_movsi_vfp): Likewise.
16993 (movsf_vfp): Likewise.
16994 (thumb2_movsf_vfp): Likewise.
16995 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
16997 (cortexa7_older_only): Likewise.
16998 (cortexa7_younger): Likewise.
16999 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
17000 (1020alu_shift_op): Likewise.
17001 (1020alu_shift_reg_op): Likewise.
17002 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
17003 (alu_shift_op): Likewise.
17004 (alu_shift_reg_op): Likewise.
17005 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
17006 (11_alu_shift_op): Likewise.
17007 (11_alu_shift_reg_op): Likewise.
17008 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
17009 (9_alu_shift_reg_op): Likewise.
17010 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
17012 (cortex_a15_alu_shift): Likewise.
17013 (cortex_a15_alu_shift_reg): Likewise.
17014 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
17015 (cortex_a5_alu_shift): Likewise.
17016 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
17018 (cortex_a53_alu_shift): Likewise.
17019 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
17021 (cortex_a7_alu_reg): Likewise.
17022 (cortex_a7_alu_shift): Likewise.
17023 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
17024 (cortex_a8_alu_shift): Likewise.
17025 (cortex_a8_alu_shift_reg): Likewise.
17026 (cortex_a8_mov): Likewise.
17027 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
17028 (cortex_a9_dp_shift): Likewise.
17029 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
17030 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
17031 (cortex_r4_mov): Likewise.
17032 (cortex_r4_alu_shift): Likewise.
17033 (cortex_r4_alu_shift_reg): Likewise.
17034 * config/arm/fa526.md (526_alu_op): Update for attribute change.
17035 (526_alu_shift_op): Likewise.
17036 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
17037 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
17038 (626te_alu_shift_op): Likewise.
17039 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
17040 (726te_alu_op): Likewise.
17041 (726te_alu_shift_op): Likewise.
17042 (726te_alu_shift_reg_op): Likewise.
17043 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
17044 (mp626_alu_shift_op): Likewise.
17045 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
17046 (pj4_alu_e1_conds): Likewise.
17047 (pj4_alu): Likewise.
17048 (pj4_alu_conds): Likewise.
17049 (pj4_shift): Likewise.
17050 (pj4_shift_conds): Likewise.
17051 (pj4_alu_shift): Likewise.
17052 (pj4_alu_shift_conds): Likewise.
17054 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17056 * config/arm/predicates.md (shiftable_operator_strict_it):
17058 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
17059 Disable cond_exec version for arm_restrict_it.
17060 (thumb2_smaxsi3): Convert to generate cond_exec.
17061 (thumb2_sminsi3): Likewise.
17062 (thumb32_umaxsi3): Likewise.
17063 (thumb2_uminsi3): Likewise.
17064 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
17065 (thumb2_neg_abssi2): Likewise.
17066 (thumb2_mov_scc): Add alternative for 16-bit encoding.
17067 (thumb2_movsicc_insn): Adjust alternatives.
17068 (thumb2_mov_negscc): Disable for arm_restrict_it.
17069 (thumb2_mov_negscc_strict_it): New pattern.
17070 (thumb2_mov_notscc_strict_it): New pattern.
17071 (thumb2_mov_notscc): Disable for arm_restrict_it.
17072 (thumb2_ior_scc): Likewise.
17073 (thumb2_ior_scc_strict_it): New pattern.
17074 (thumb2_cond_move): Adjust for arm_restrict_it.
17075 (thumb2_cond_arith): Disable for arm_restrict_it.
17076 (thumb2_cond_arith_strict_it): New pattern.
17077 (thumb2_cond_sub): Adjust for arm_restrict_it.
17078 (thumb2_movcond): Likewise.
17079 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
17080 (thumb2_zero_extendhisi2_v6): Likewise.
17081 (thumb2_zero_extendqisi2_v6): Likewise.
17082 (orsi_notsi_si): Likewise.
17083 (orsi_not_shiftsi_si): Likewise.
17085 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
17087 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
17088 instruction sequence is 1 byte shorter.
17090 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
17092 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
17093 it is not needed after split.
17095 2013-07-20 Iain Sandoe <iain@codesourcery.com>
17098 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
17099 second label for nonlocal goto receivers. Don't output pic base labels
17100 unless we're producing PIC; mark that action unreachable().
17101 (ix86_save_reg): If the function contains a nonlocal label, save the
17103 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
17104 * config/darwin.c (emitted_pic_label_num): New GTY.
17105 (update_pic_label_number_if_needed): New.
17106 (machopic_output_function_base_name): Adjust for nonlocal receiver
17108 (machopic_should_output_picbase_label): New.
17109 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
17110 (nonlocal_goto_receiver): New insn and split.
17112 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
17114 * config/aarch64/aarch64-builtins.c
17115 (aarch64_fold_builtin): Fold abs in all modes.
17116 * config/aarch64/aarch64-simd-builtins.def
17117 (abs): Enable for all modes.
17118 * config/aarch64/arm_neon.h
17119 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
17120 (vabs_f64): Add missing intrinsic.
17122 2013-07-19 Ian Bolton <ian.bolton@arm.com>
17124 * config/aarch64/arm_neon.h (vabs_s64): New function
17126 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
17129 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
17130 * config/avr/avr.md (adjust_len): Add `round'.
17131 * config/avr/avr-protos.h (avr_out_round): New prototype.
17132 (avr_out_plus): Add `out_label' argument.
17133 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
17134 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
17135 Handle the case where `insn' is just a pattern.
17136 (avr_out_bitop): Handle the case where `insn' is just a pattern.
17137 (avr_out_round): New function.
17138 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
17140 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
17142 * config/microblaze/microblaze.c (microblaze_expand_prologue):
17143 Add check for flag_stack_usage to handle -fstack-usage support
17145 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
17147 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
17148 interaction for new Power8 flags and VSX.
17150 2013-07-18 Sriraman Tallam <tmsriram@google.com>
17152 PR middle-end/57698
17153 * tree-inline.c (expand_call_inline): Emit errors during
17154 early_inlining only if optimization is not turned on.
17156 2013-07-18 David Malcolm <dmalcolm@redhat.com>
17160 * passes.c (init_optimization_passes): Move the construction of
17161 the pass hierarchy into a new passes.def file.
17163 * Makefile.in (passes.o): Add dependency on passes.def.
17165 2013-07-18 David Malcolm <dmalcolm@redhat.com>
17167 * passes.c (init_optimization_passes): Introduce macros for
17168 constructing the tree of passes (INSERT_PASSES_AFTER,
17169 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
17170 TERMINATE_PASS_LIST).
17172 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
17173 Wei Mi <wmi@google.com>
17175 PR rtl-optimization/57878
17176 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
17178 (reload_pseudo_compare_func): Check nregs first for reload
17181 2013-07-18 David Malcolm <dmalcolm@redhat.com>
17183 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
17185 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
17187 * read-rtl.c (validate_const_int): Once an invalid character is
17188 seen, quit the loop.
17190 * gengtype.c (write_roots): Similarly once we find the "deletable"
17191 or "if_marked" option.
17193 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
17195 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
17196 "xtab" and "sat". Move value "clz" from here to ...
17197 (attriubte "type"): ... here.
17198 (satsi_<SAT:code>): Delete "insn" attribute.
17199 (satsi_<SAT:code>_shift): Likewise.
17200 (arm_zero_extendqisi2addsi): Likewise.
17201 (arm_extendqisi2addsi): Likewise.
17202 (clzsi2): Update for attribute changes.
17203 (rbitsi2): Likewise.
17204 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
17206 (arm_usatsihi): Likewise.
17207 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
17209 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
17211 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
17212 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
17213 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
17214 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
17215 in alphabetical order.
17216 (attribute "core_cycles"): Update for attribute changes.
17217 (arm_addsi3): Likewise.
17218 (addsi3_compare0): Likewise.
17219 (addsi3_compare0_scratch): Likewise.
17220 (addsi3_compare_op1): Likewise.
17221 (addsi3_compare_op2): Likewise.
17222 (compare_addsi2_op0): Likewise.
17223 (compare_addsi2_op1): Likewise.
17224 (addsi3_carryin_shift_<optab>): Likewise.
17225 (subsi3_carryin_shift): Likewise.
17226 (rsbsi3_carryin_shift): Likewise.
17227 (arm_subsi3_insn): Likewise.
17228 (subsi3_compare0): Likewise.
17229 (subsi3_compare): Likewise.
17230 (arm_andsi3_insn): Likewise.
17231 (thumb1_andsi3_insn): Likewise.
17232 (andsi3_compare0): Likewise.
17233 (andsi3_compare0_scratch): Likewise.
17234 (zeroextractsi_compare0_scratch
17235 (andsi_not_shiftsi_si): Likewise.
17236 (iorsi3_insn): Likewise.
17237 (iorsi3_compare0): Likewise.
17238 (iorsi3_compare0_scratch): Likewise.
17239 (arm_xorsi3): Likewise.
17240 (thumb1_xorsi3_insn): Likewise.
17241 (xorsi3_compare0): Likewise.
17242 (xorsi3_compare0_scratch): Likewise.
17243 (satsi_<SAT:code>_shift): Likewise.
17245 (arm_shiftsi3): Likewise.
17246 (shiftsi3_compare0): Likewise.
17247 (not_shiftsi): Likewise.
17248 (not_shiftsi_compare0): Likewise.
17249 (not_shiftsi_compare0_scratch): Likewise.
17250 (arm_one_cmplsi2): Likewise.
17251 (thumb_one_complsi2): Likewise.
17252 (notsi_compare0): Likewise.
17253 (notsi_compare0_scratch): Likewise.
17254 (thumb1_zero_extendhisi2): Likewise.
17255 (arm_zero_extendhisi2): Likewise.
17256 (arm_zero_extendhisi2_v6): Likewise.
17257 (arm_zero_extendhisi2addsi): Likewise.
17258 (thumb1_zero_extendqisi2): Likewise.
17259 (thumb1_zero_extendqisi2_v6): Likewise.
17260 (arm_zero_extendqisi2): Likewise.
17261 (arm_zero_extendqisi2_v6): Likewise.
17262 (arm_zero_extendqisi2addsi): Likewise.
17263 (thumb1_extendhisi2): Likewise.
17264 (arm_extendhisi2): Likewise.
17265 (arm_extendhisi2_v6): Likewise.
17266 (arm_extendqisi): Likewise.
17267 (arm_extendqisi_v6): Likewise.
17268 (arm_extendqisi2addsi): Likewise.
17269 (thumb1_extendqisi2): Likewise.
17270 (thumb1_movdi_insn): Likewise.
17271 (arm_movsi_insn): Likewise.
17272 (movsi_compare0): Likewise.
17273 (movhi_insn_arch4): Likewise.
17274 (movhi_bytes): Likewise.
17275 (arm_movqi_insn): Likewise.
17276 (thumb1_movqi_insn): Likewise.
17277 (arm32_movhf): Likewise.
17278 (thumb1_movhf): Likewise.
17279 (arm_movsf_soft_insn): Likewise.
17280 (thumb1_movsf_insn): Likewise.
17281 (movdf_soft_insn): Likewise.
17282 (thumb_movdf_insn): Likewise.
17283 (arm_cmpsi_insn): Likewise.
17284 (cmpsi_shiftsi): Likewise.
17285 (cmpsi_shiftsi_swp): Likewise.
17286 (arm_cmpsi_negshiftsi_si): Likewise.
17287 (movsicc_insn): Likewise.
17288 (movsfcc_soft_insn): Likewise.
17289 (arith_shiftsi): Likewise.
17290 (arith_shiftsi_compare0
17291 (arith_shiftsi_compare0_scratch
17292 (sub_shiftsi): Likewise.
17293 (sub_shiftsi_compare0
17294 (sub_shiftsi_compare0_scratch
17295 (and_scc): Likewise.
17296 (cond_move): Likewise.
17297 (if_plus_move): Likewise.
17298 (if_move_plus): Likewise.
17299 (if_move_not): Likewise.
17300 (if_not_move): Likewise.
17301 (if_shift_move): Likewise.
17302 (if_move_shift): Likewise.
17303 (if_shift_shift): Likewise.
17304 (if_not_arith): Likewise.
17305 (if_arith_not): Likewise.
17306 (cond_move_not): Likewise.
17307 (thumb1_ashlsi3): Set type attribute.
17308 (thumb1_ashrsi3): Likewise.
17309 (thumb1_lshrsi3): Likewise.
17310 (thumb1_rotrsi3): Likewise.
17311 (shiftsi3_compare0_scratch): Likewise.
17312 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
17313 (neon_mov<mode>): Likewise.
17314 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
17316 (thumb2_movsi_insn): Likewise.
17317 (thumb2_cmpsi_neg_shiftsi): Likewise.
17318 (thumb2_extendqisi_v6): Likewise.
17319 (thumb2_zero_extendhisi2_v6): Likewise.
17320 (thumb2_zero_extendqisi2_v6): Likewise.
17321 (thumb2_shiftsi3_short): Likewise.
17322 (thumb2_addsi3_compare0_scratch): Likewise.
17323 (orsi_not_shiftsi_si): Likewise.
17324 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
17325 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
17327 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
17328 (1020alu_shift_op): Likewise.
17329 (1020alu_shift_reg_op): Likewise.
17330 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
17331 (alu_shift_op): Likewise.
17332 (alu_shift_reg_op): Likewise.
17333 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
17334 (11_alu_shift_op): Likewise.
17335 (11_alu_shift_reg_op): Likewise.
17336 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
17337 (9_alu_shift_reg_op): Likewise.
17338 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
17340 (cortex_a15_alu_shift): Likewise.
17341 (cortex_a15_alu_shift_reg): Likewise.
17342 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
17344 (cortex_a5_alu_shift): Likewise.
17345 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
17347 (cortex_a53_alu_shift): Likewise.
17348 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
17350 (cortex_a7_alu_reg): Likewise.
17351 (cortex_a7_alu_shift): Likewise.
17352 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
17354 (cortex_a8_alu_shift): Likewise.
17355 (cortex_a8_alu_shift_reg): Likewise.
17356 (cortex_a8_mov): Likewise.
17357 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
17358 (cortex_a9_dp_shift): Likewise.
17359 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
17361 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
17363 (cortex_r4_mov): Likewise.
17364 (cortex_r4_alu_shift): Likewise.
17365 (cortex_r4_alu_shift_reg): Likewise.
17366 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
17367 (526_alu_shift_op): Likewise.
17368 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
17369 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
17370 (626te_alu_shift_op): Likewise.
17371 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
17372 (726te_alu_op): Likewise.
17373 (726te_alu_shift_op): Likewise.
17374 (726te_alu_shift_reg_op): Likewise.
17375 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
17376 (mp626_alu_shift_op): Likewise.
17377 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
17378 (pj4_alu_e1_conds): Likewise.
17379 (pj4_alu): Likewise.
17380 (pj4_alu_conds): Likewise.
17381 (pj4_shift): Likewise.
17382 (pj4_shift_conds): Likewise.
17383 (pj4_alu_shift): Likewise.
17384 (pj4_alu_shift_conds): Likewise.
17385 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
17387 (cortexa7_older_only): Likewise.
17388 (cortexa7_younger): Likewise.
17390 2013-07-18 David Malcolm <dmalcolm@redhat.com>
17392 * ipa-pure-const.c (generate_summary): Rename to...
17393 (pure_const_generate_summary): ... this.
17395 2013-07-17 Iain Sandoe <iain@codesourcery.com>
17397 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
17399 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
17402 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
17405 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17407 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
17408 enabled without -march=zEC12.
17409 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
17412 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
17414 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
17415 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
17416 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
17417 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
17418 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
17419 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
17420 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
17422 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
17423 and ISA_HAS_NMADD3_NMSUB3.
17424 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
17425 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
17426 (nmsub4<mode>, nmsub3<mode>): Likewise.
17427 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
17429 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
17431 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
17432 TARGET_MIPS5400 checking.
17434 2013-07-16 Jakub Jelinek <jakub@redhat.com>
17435 Peter Bergner <bergner@vnet.ibm.com>
17437 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
17438 registers in the comment.
17439 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
17440 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
17441 rather than FIRST_PSEUDO_REGISTERS.
17443 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
17445 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
17446 enable extra ISA flags with TARGET_HTM.
17448 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
17450 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
17453 2013-07-15 Cong Hou <congh@google.com>
17455 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
17456 in compare function for sorting.
17458 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
17460 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
17461 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
17462 * config/rs6000/rs6000.opt: Add -mhtm option.
17463 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
17464 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
17465 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
17466 __HTM__ if the HTM instructions are available.
17467 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
17468 htm_spr_reg_operand): New define_predicates.
17469 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
17470 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
17472 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
17473 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
17474 HTM builtin functions.
17475 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
17476 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
17477 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
17478 (rs6000_builtin_mask_calculate): Likewise.
17479 (rs6000_option_override_internal): Likewise.
17480 (bdesc_htm): Add new HTM builtin support.
17481 (htm_spr_num): New function.
17482 (htm_spr_regno): Likewise.
17483 (rs6000_htm_spr_icode): Likewise.
17484 (htm_expand_builtin): Likewise.
17485 (htm_init_builtins): Likewise.
17486 (rs6000_expand_builtin): Add support for HTM builtin functions.
17487 (rs6000_init_builtins): Likewise.
17488 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
17490 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
17491 (TARGET_HTM, MASK_HTM): Define macros.
17492 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
17493 (FIXED_REGISTERS): Likewise.
17494 (CALL_USED_REGISTERS): Likewise.
17495 (CALL_REALLY_USED_REGISTERS): Likewise.
17496 (REG_ALLOC_ORDER): Likewise.
17497 (enum reg_class): Likewise.
17498 (REG_CLASS_NAMES): Likewise.
17499 (REG_CLASS_CONTENTS): Likewise.
17500 (REGISTER_NAMES): Likewise.
17501 (ADDITIONAL_REGISTER_NAMES): Likewise.
17502 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
17503 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
17504 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
17505 * config/rs6000/htm.md: New file.
17506 * config/rs6000/htmintrin.h: New file.
17507 * config/rs6000/htmxlintrin.h: New file.
17509 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
17511 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
17512 Define SYMBOL_TINY_GOT, update comment.
17513 * config/aarch64/aarch64.c
17514 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
17515 (aarch64_expand_mov_immediate): Likewise.
17516 (aarch64_print_operand): Likewise.
17517 (aarch64_classify_symbol): Likewise.
17518 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
17519 (ldr_got_tiny): Define.
17521 2013-07-13 Tobias Grosser <tobias@grosser.es>
17523 PR tree-optimization/54094
17524 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
17525 scheduling dimension for the parallelism check from the polyhedral
17526 information in the AST.
17527 * graphite-dependences.c (carries_deps): Do not assume the schedule is
17530 2013-07-13 Jason Merrill <jason@redhat.com>
17532 * print-tree.c (debug_vec_tree): Use debug_raw.
17533 (debug_raw (vec<tree, va_gc> &)): New.
17534 (debug_raw (vec<tree, va_gc> *)): New.
17535 * tree.h: Declare them.
17537 2013-07-13 Bin Cheng <bin.cheng@arm.com>
17539 * ifcvt.c (ifcvt_after_combine): New static variable.
17540 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
17542 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
17543 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
17544 rest_of_handle_if_after_reload): Pass new argument for if_convert.
17546 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
17548 * config/mips/mips.c (mips_expand_call): Remove empty statement.
17550 2013-07-12 Michael Matz <matz@suse.de>
17552 PR middle-end/55771
17553 * convert.c (convert_to_real): Reject non-float inner types.
17555 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
17557 * config/aarch64/aarch64-protos.h
17558 (aarch64_simd_immediate_valid_for_move): Remove.
17559 * config/aarch64/aarch64.c (simd_immediate_info): New member.
17560 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
17562 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
17564 2013-07-11 Steve Ellcey <sellcey@mips.com>
17566 * config/mips/mips.c (mips_conditional_register_usage): Do not
17567 use t[0-7] registers in MIPS16 mode when optimizing for size.
17569 2013-07-11 Sriraman Tallam <tmsriram@google.com>
17571 * config/i386/i386.c (dispatch_function_versions): Fix array
17572 indexing of function_version_info to match actual_versions.
17574 2013-07-11 Teresa Johnson <tejohnson@google.com>
17576 * vec.h (struct va_gc): Move release out-of-line.
17577 (va_gc::release): Call ggc_free on released vec.
17579 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
17581 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
17582 Require GOT register as additional operand in UNSPEC.
17583 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
17584 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
17585 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
17586 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
17587 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
17588 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
17589 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
17591 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
17594 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
17595 name seen by assembler/linker rather if available.
17597 2013-07-11 Andreas Schwab <schwab@suse.de>
17599 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
17601 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
17603 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
17605 2013-07-10 Joseph Myers <joseph@codesourcery.com>
17607 * doc/tm.texi.in: Move hook documentation to ....
17608 * target.def: ... here.
17610 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
17611 text on @hook line.
17612 * doc/tm.texi: Regenerate.
17614 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
17617 * doc/invoke.texi: Document -Wconditionally-supported.
17619 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
17622 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
17625 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
17628 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
17629 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
17630 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
17631 Remove duplicate devices.
17632 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
17633 * config/avr/t-multilib: Regenerate.
17634 * config/avr/avr-tables.opt: Regenerate.
17635 * doc/avr-mmcu.texi: Regenerate.
17637 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
17640 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
17642 2013-07-10 Graham Stott <graham.stott@btinternet.com>
17644 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
17645 the cost of MULT when optimizing for size.
17647 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
17649 * config/cr16/cr16-protos.h: Don't include target.h.
17651 2013-07-09 Joseph Myers <joseph@codesourcery.com>
17653 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
17654 adjust register size for TDmode and TFmode for VSX registers.
17656 2013-07-08 Kai Tietz <ktietz@redhat.com>
17659 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
17660 hook_bool_const_tree_true.
17662 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17664 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
17665 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
17666 * config/s390/s390.md: Define FPR*_REGNUM constants.
17667 Fix FPR2_REGNUM constant (18 -> 17).
17668 ("*trunc<BFP:mode><DFP_ALL:mode>2")
17669 ("*trunc<DFP_ALL:mode><BFP:mode>2")
17670 ("trunc<BFP:mode><DFP_ALL:mode>2")
17671 ("trunc<DFP_ALL:mode><BFP:mode>2")
17672 ("*extend<BFP:mode><DFP_ALL:mode>2")
17673 ("*extend<DFP_ALL:mode><BFP:mode>2")
17674 ("extend<BFP:mode><DFP_ALL:mode>2")
17675 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
17678 2013-07-08 Graham Stott <graham.stott@btinternet.com>
17680 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
17682 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17684 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
17685 and cfun_fpr_bit_p to cfun_fpr_save_p.
17686 (s390_frame_area, s390_register_info, s390_frame_info)
17687 (s390_emit_prologue, s390_emit_epilogue)
17688 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
17690 * config/s390/s390.h: Define *_REGNUM macros for floating point
17693 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
17695 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
17697 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
17699 PR rtl-optimization/57786
17700 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
17701 and break out of the loop when it is set to false.
17703 2013-07-08 Jakub Jelinek <jakub@redhat.com>
17706 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
17707 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
17708 (const_int 63)) 0)).
17709 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
17710 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
17711 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
17713 PR rtl-optimization/57829
17714 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
17715 mask bits outside of mode are just sign-extension from mode to HWI.
17717 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
17719 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
17720 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
17721 adjust_address instead of change_address to keep info about alignment.
17722 (emit_strmov): Remove.
17723 (emit_memmov): New function.
17724 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
17725 (expand_movmem_epilogue): Likewise and return updated rtx for
17727 (expand_constant_movmem_prologue): Likewise and return updated rtx for
17728 destination and source.
17729 (decide_alignment): Refactor, handle vector_loop.
17730 (ix86_expand_movmem): Likewise.
17731 (ix86_expand_setmem): Likewise.
17732 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
17734 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
17736 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
17737 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
17739 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
17741 * config/i386/sse.md (sse_movlhps): Change alternative 3
17742 of operand 2 to "m".
17744 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
17747 * config/i386/sse.md (iptr): New mode attribute.
17748 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
17749 (<sse>_vm<plusminus_insn><mode>3): Ditto.
17750 (<sse>_vmmul<mode>3): Ditto.
17751 (<sse>_vmdiv<mode>3): Ditto.
17752 (sse_vmrcpv4sf2): Ditto.
17753 (<sse>_vmsqrt<mode>2): Ditto.
17754 (sse_vmrsqrtv4sf2): Ditto.
17755 (<sse>_vm<code><mode>3): Ditto.
17756 (avx_vmcmp<mode>3): Ditto.
17757 (<sse>_vmmaskcmp<mode>3): Ditto.
17758 (<sse>_comi): Ditto.
17759 (<sse>_ucomi): Ditto.
17760 (*xop_vmfrcz_<mode>): Ditto.
17761 (*fmai_fmadd_<mode>): Ditto.
17762 (*fmai_fmsub_<mode>): Ditto.
17763 (*fmai_fnmadd_<mode>): Ditto.
17764 (*fmai_fnmsub_<mode>): Ditto.
17765 (*fma4i_vmfmadd_<mode>): Ditto.
17766 (*fma4i_vmfmsub_<mode>): Ditto.
17767 (*fma4i_vmfnmadd_<mode>): Ditto.
17768 (*fma4i_vmfnmsub_<mode>): Ditto.
17769 (*xop_vmfrcz_<mode>): Ditto.
17770 (sse_cvtps2pi): Ditto.
17771 (sse_cvttps2pi): Ditto.
17772 (sse_cvtss2si): Ditto.
17773 (sse_cvtss2si_2): Ditto.
17774 (sse_cvtss2siq_2): Ditto.
17775 (sse_cvttss2si): Ditto.
17776 (sse_cvttss2siq): Ditto.
17777 (sse_cvtsd2si): Ditto.
17778 (sse_cvtsd2si_2): Ditto.
17779 (sse_cvtsd2siq_2): Ditto.
17780 (sse_cvttsd2si): Ditto.
17781 (sse_cvttsd2siq): Ditto.
17782 (sse_cvtsd2ss): Ditto.
17783 (sse_cvtss2sd): Ditto.
17784 (avx2_pbroadcast<mode>): Ditto.
17785 (avx2_pbroadcast<mode>_1): Ditto.
17786 (*avx_vperm_broadcast_v4sf): Ditto.
17788 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
17789 (sse_movlhps): Ditto.
17790 (sse_storehps): Ditto.
17791 (sse_loadhps): Ditto.
17792 (sse_storelps): Ditto.
17793 (sse_loadlps): Ditto.
17794 (*vec_concatv4sf): Ditto.
17795 (*vec_interleave_highv2df): Ditto.
17796 (*vec_interleave_lowv2df): Ditto.
17797 (*vec_extractv2df_1_sse): Ditto.
17798 (*vec_extractv2df_0_sse): Ditto.
17799 (sse2_storelpd): Ditto.
17800 (sse2_loadlpd): Ditto.
17801 (sse2_movsd): Ditto.
17802 (*vec_concatv4si): Ditto.
17803 (vec_concatv2di): Ditto.
17805 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
17806 for Intel asm dialect.
17807 (mmx_punpcklwd): Ditto.
17808 (mmx_punpckldq): Ditto.
17810 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
17811 for intel assembler dialect.
17813 2013-07-06 Jakub Jelinek <jakub@redhat.com>
17816 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
17817 for BUILT_IN_C{LZ,LRSB}*.
17818 * tree.h (CASE_INT_FN): Add FN##IMAX case.
17819 * tree-vrp.c (extract_range_basic): Handle
17820 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
17821 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
17822 fall thru to code calling set_value*.
17823 * builtins.c (expand_builtin): Remove *IMAX cases.
17824 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
17825 if width is bigger than 2*HWI.
17827 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
17829 PR rtl-optimization/55342
17830 * lra-int.h (lra_subreg_reload_pseudos): New.
17831 * lra.c: Add undoing optional reloads to the block diagram.
17832 (lra_subreg_reload_pseudos): New.
17833 (lra_optional_reload_pseudos): Change comments.
17834 (lra): Init and clear lra_subreg_reload_pseudos. Clear
17835 lra_optional_reload_pseudos after undo transformations.
17836 * lra-assigns.c (pseudo_prefix_title): New.
17837 (lra_setup_reg_renumber): Use it.
17838 (spill_for): Ditto. Check subreg reload pseudos too.
17839 (assign_by_spills): Consider subreg reload pseudos too.
17840 * lra-constraints.c (simplify_operand_subreg): Use
17841 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
17842 (curr_insn_transform): Recognize and do optional reloads.
17843 (undo_optional_reloads): New.
17844 (lra_undo_inheritance): Call undo_optional_reloads.
17846 2013-07-05 Thomas Quinot <quinot@adacore.com>
17848 * tree-complex.c (expand_complex_operations_1): Fix typo.
17850 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
17852 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
17853 (tune_params): New member 'const vec_costs'.
17854 * config/aarch64/aarch64.c (generic_vector_cost): New.
17855 (generic_tunings): New member 'generic_vector_cost'.
17856 (aarch64_builtin_vectorization_cost): New.
17857 (aarch64_add_stmt_cost): New.
17858 (TARGET_VECTORIZE_ADD_STMT_COST): New.
17859 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
17861 2013-07-03 Jakub Jelinek <jakub@redhat.com>
17864 * config/i386/predicates.md (vsib_address_operand): Disallow
17865 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
17867 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
17869 PR middle-end/55030
17870 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
17871 expand_builtin_setjmp_receiver.
17872 (expand_label): Adjust, call expand_builtin_setjmp_receiver
17873 with NULL for the label parameter.
17874 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
17875 the frame-pointer. Adjust comments.
17876 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
17877 only if LABEL is non-NULL.
17879 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
17881 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
17882 (ARM_ABI_AAPCS64): Ditto.
17884 (ARM_DEFAULT_ABI): Ditto.
17886 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
17888 * config/aarch64/aarch64-builtins.c
17889 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
17890 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
17892 * config/aarch64/aarch64-simd.md
17893 (aarch64_ld1<VALL:mode>): New.
17894 (aarch64_st1<VALL:mode>): Likewise.
17895 * config/aarch64/arm_neon.h
17896 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
17898 2013-07-02 Sriraman Tallam <tmsriram@google.com>
17900 * config/i386/i386.c (gate_insert_vzeroupper): Check if
17902 (ix86_option_override_internal):Turn on all -mavx target flags by
17903 default as they are dependent on AVX anyway.
17905 2013-07-02 Cary Coutant <ccoutant@google.com>
17907 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
17908 deterministic hash.
17909 (loc_checksum_ordered): Likewise.
17910 (hash_loc_operands): Remove inline keyword.
17912 2013-07-02 Jakub Jelinek <jakub@redhat.com>
17914 PR tree-optimization/57741
17915 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
17916 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
17917 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
17918 Allow REAL_CST step_exprs if flag_associative_math.
17919 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
17921 2013-07-02 Ian Bolton <ian.bolton@arm.com>
17923 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
17925 2013-07-02 Ian Bolton <ian.bolton@arm.com>
17927 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
17929 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17931 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
17933 (iorsi3_insn): Likewise.
17934 (arm_xorsi3): Likewise.
17936 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
17938 * arm.md (attribute "wtype"): Delete. Move attribute values from here
17940 (attribute "type"): ... here, and prefix with "wmmx_".
17941 (attribute "core_cycles"): Update for attribute changes.
17942 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
17943 (tbcstv4hi): Likewise.
17944 (tbcstv2si): Likewise.
17945 (iwmmxt_iordi3): Likewise.
17946 (iwmmxt_xordi3): Likewise.
17947 (iwmmxt_anddi3): Likewise.
17948 (iwmmxt_nanddi3): Likewise.
17949 (iwmmxt_arm_movdi): Likewise.
17950 (iwmmxt_movsi_insn): Likewise.
17951 (mov<mode>_internal): Likewise.
17952 (and<mode>3_iwmmxt): Likewise.
17953 (ior<mode>3_iwmmxt): Likewise.
17954 (xor<mode>3_iwmmxt): Likewise.
17955 (add<mode>3_iwmmxt): Likewise.
17956 (ssaddv8qi3): Likewise.
17957 (ssaddv4hi3): Likewise.
17958 (ssaddv2si3): Likewise.
17959 (usaddv8qi3): Likewise.
17960 (usaddv4hi3): Likewise.
17961 (usaddv2si3): Likewise.
17962 (sub<mode>3_iwmmxt): Likewise.
17963 (sssubv8qi3): Likewise.
17964 (sssubv4hi3): Likewise.
17965 (sssubv2si3): Likewise.
17966 (ussubv8qi3): Likewise.
17967 (ussubv4hi3): Likewise.
17968 (ussubv2si3): Likewise.
17969 (mulv4hi3_iwmmxt): Likewise.
17970 (smulv4hi3_highpart): Likewise.
17971 (umulv4hi3_highpart): Likewise.
17972 (iwmmxt_wmacs): Likewise.
17973 (iwmmxt_wmacsz): Likewise.
17974 (iwmmxt_wmacu): Likewise.
17975 (iwmmxt_wmacuz): Likewise.
17976 (iwmmxt_clrdi): Likewise.
17977 (iwmmxt_clrv8qi): Likewise.
17978 (iwmmxt_clr4hi): Likewise.
17979 (iwmmxt_clr2si): Likewise.
17980 (iwmmxt_uavgrndv8qi3): Likewise.
17981 (iwmmxt_uavgrndv4hi3): Likewise.
17982 (iwmmxt_uavgv8qi3): Likewise.
17983 (iwmmxt_uavgv4hi3): Likewise.
17984 (iwmmxt_tinsrb): Likewise.
17985 (iwmmxt_tinsrh): Likewise.
17986 (iwmmxt_tinsrw): Likewise.
17987 (iwmmxt_textrmub): Likewise.
17988 (iwmmxt_textrmsb): Likewise.
17989 (iwmmxt_textrmuh): Likewise.
17990 (iwmmxt_textrmsh): Likewise.
17991 (iwmmxt_textrmw): Likewise.
17992 (iwmxxt_wshufh): Likewise.
17993 (eqv8qi3): Likewise.
17994 (eqv4hi3): Likewise.
17995 (eqv2si3): Likewise.
17996 (gtuv8qi3): Likewise.
17997 (gtuv4hi3): Likewise.
17998 (gtuv2si3): Likewise.
17999 (gtv8qi3): Likewise.
18000 (gtv4hi3): Likewise.
18001 (gtv2si3): Likewise.
18002 (smax<mode>3_iwmmxt): Likewise.
18003 (umax<mode>3_iwmmxt): Likewise.
18004 (smin<mode>3_iwmmxt): Likewise.
18005 (umin<mode>3_iwmmxt): Likewise.
18006 (iwmmxt_wpackhss): Likewise.
18007 (iwmmxt_wpackwss): Likewise.
18008 (iwmmxt_wpackdss): Likewise.
18009 (iwmmxt_wpackhus): Likewise.
18010 (iwmmxt_wpackwus): Likewise.
18011 (iwmmxt_wpackdus): Likewise.
18012 (iwmmxt_wunpckihb): Likewise.
18013 (iwmmxt_wunpckihh): Likewise.
18014 (iwmmxt_wunpckihw): Likewise.
18015 (iwmmxt_wunpckilb): Likewise.
18016 (iwmmxt_wunpckilh): Likewise.
18017 (iwmmxt_wunpckilw): Likewise.
18018 (iwmmxt_wunpckehub): Likewise.
18019 (iwmmxt_wunpckehuh): Likewise.
18020 (iwmmxt_wunpckehuw): Likewise.
18021 (iwmmxt_wunpckehsb): Likewise.
18022 (iwmmxt_wunpckehsh): Likewise.
18023 (iwmmxt_wunpckehsw): Likewise.
18024 (iwmmxt_wunpckelub): Likewise.
18025 (iwmmxt_wunpckeluh): Likewise.
18026 (iwmmxt_wunpckeluw): Likewise.
18027 (iwmmxt_wunpckelsb): Likewise.
18028 (iwmmxt_wunpckelsh): Likewise.
18029 (iwmmxt_wunpckelsw): Likewise.
18030 (ror<mode>3): Likewise.
18031 (ashr<mode>3_iwmmxt): Likewise.
18032 (lshr<mode>3_iwmmxt): Likewise.
18033 (ashl<mode>3_iwmmxt): Likewise.
18034 (ror<mode>3_di): Likewise.
18035 (ashr<mode>3_di): Likewise.
18036 (lshr<mode>3_di): Likewise.
18037 (ashl<mode>3_di): Likewise.
18038 (iwmmxt_wmadds): Likewise.
18039 (iwmmxt_wmaddu): Likewise.
18040 (iwmmxt_tmia): Likewise.
18041 (iwmmxt_tmiaph): Likewise.
18042 (iwmmxt_tmiabb): Likewise.
18043 (iwmmxt_tmiatb): Likewise.
18044 (iwmmxt_tmiabt): Likewise.
18045 (iwmmxt_tmiatt): Likewise.
18046 (iwmmxt_tmovmskb): Likewise.
18047 (iwmmxt_tmovmskh): Likewise.
18048 (iwmmxt_tmovmskw): Likewise.
18049 (iwmmxt_waccb): Likewise.
18050 (iwmmxt_wacch): Likewise.
18051 (iwmmxt_waccw): Likewise.
18052 (iwmmxt_waligni): Likewise.
18053 (iwmmxt_walignr): Likewise.
18054 (iwmmxt_walignr0): Likewise.
18055 (iwmmxt_walignr1): Likewise.
18056 (iwmmxt_walignr2): Likewise.
18057 (iwmmxt_walignr3): Likewise.
18058 (iwmmxt_wsadb): Likewise.
18059 (iwmmxt_wsadh): Likewise.
18060 (iwmmxt_wsadbz): Likewise.
18061 (iwmmxt_wsadhz): Likewise.
18062 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
18063 (iwmmxt_wabsdiffb): Likewise.
18064 (iwmmxt_wabsdiffh): Likewise.
18065 (iwmmxt_wabsdiffw): Likewise.
18066 (iwmmxt_waddsubhx): Likewise
18067 (iwmmxt_wsubaddhx): Likewise.
18068 (addc<mode>3): Likewise.
18069 (iwmmxt_avg4): Likewise.
18070 (iwmmxt_avg4r): Likewise.
18071 (iwmmxt_wmaddsx): Likewise.
18072 (iwmmxt_wmaddux): Likewise.
18073 (iwmmxt_wmaddsn): Likewise.
18074 (iwmmxt_wmaddun): Likewise.
18075 (iwmmxt_wmulwsm): Likewise.
18076 (iwmmxt_wmulwum): Likewise.
18077 (iwmmxt_wmulsmr): Likewise.
18078 (iwmmxt_wmulumr): Likewise.
18079 (iwmmxt_wmulwsmr): Likewise.
18080 (iwmmxt_wmulwumr): Likewise.
18081 (iwmmxt_wmulwl): Likewise.
18082 (iwmmxt_wqmulm): Likewise.
18083 (iwmmxt_wqmulwm): Likewise.
18084 (iwmmxt_wqmulmr): Likewise.
18085 (iwmmxt_wqmulwmr): Likewise.
18086 (iwmmxt_waddbhusm): Likewise.
18087 (iwmmxt_waddbhusl): Likewise.
18088 (iwmmxt_wqmiabb): Likewise.
18089 (iwmmxt_wqmiabt): Likewise.
18090 (iwmmxt_wqmiatb): Likewise.
18091 (iwmmxt_wqmiatt): Likewise.
18092 (iwmmxt_wqmiabbn): Likewise.
18093 (iwmmxt_wqmiabtn): Likewise.
18094 (iwmmxt_wqmiatbn): Likewise.
18095 (iwmmxt_wqmiattn): Likewise.
18096 (iwmmxt_wmiabb): Likewise.
18097 (iwmmxt_wmiabt): Likewise.
18098 (iwmmxt_wmiatb): Likewise.
18099 (iwmmxt_wmiatt): Likewise.
18100 (iwmmxt_wmiabbn): Likewise.
18101 (iwmmxt_wmiabtn): Likewise.
18102 (iwmmxt_wmiatbn): Likewise.
18103 (iwmmxt_wmiattn): Likewise.
18104 (iwmmxt_wmiawbb): Likewise.
18105 (iwmmxt_wmiawbt): Likewise.
18106 (iwmmxt_wmiawtb): Likewise.
18107 (iwmmxt_wmiawtt): Likewise.
18108 (iwmmxt_wmiawbbn): Likewise.
18109 (iwmmxt_wmiawbtn): Likewise.
18110 (iwmmxt_wmiawtbn): Likewise.
18111 (iwmmxt_wmiawttn): Likewise.
18112 (iwmmxt_wmerge): Likewise.
18113 (iwmmxt_tandc<mode>3): Likewise.
18114 (iwmmxt_torc<mode>3): Likewise.
18115 (iwmmxt_torvsc<mode>3): Likewise.
18116 (iwmmxt_textrc<mode>3): Likewise.
18117 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
18118 (wmmxt_pack): Likewise.
18119 (wmmxt_mult_c1): Likewise.
18120 (wmmxt_mult_c2): Likewise.
18121 (wmmxt_alu_c1): Likewise.
18122 (wmmxt_alu_c2): Likewise.
18123 (wmmxt_alu_c3): Likewise.
18124 (wmmxt_transfer_c1): Likewise.
18125 (wmmxt_transfer_c2): Likewise.
18126 (wmmxt_transfer_c3): Likewise.
18127 (marvell_f_iwmmxt_wstr): Likewise.
18128 (marvell_f_iwmmxt_wldr): Likewise.
18130 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
18132 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
18134 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
18137 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
18138 * lra-constraints.c (need_for_split_p): Check call used hard regs
18139 living through calls.
18141 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
18142 call used regs for call insn.
18144 2013-06-28 Jakub Jelinek <jakub@redhat.com>
18147 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
18148 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
18150 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
18152 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
18154 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
18156 * lra-constraints.c (need_for_split_p): Check call used hard regs
18157 living through calls.
18159 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
18162 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
18163 to tie with any other modes. Eliminate Altivec vector mode tests,
18164 since these are a subset of ALTIVEC or VSX vector modes. Simplify
18165 code, to return 0 if testing MODE2 for a condition, if we've
18166 already tested MODE1 for the same condition.
18168 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
18170 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
18173 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
18175 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
18176 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
18178 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
18180 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
18182 (aarch64_symbolic_constant_p): Remove.
18183 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
18184 static. Fix line length and white space.
18185 (aarch64_symbolic_constant_p): Remove.
18186 * config/aarch64/predicates.md (aarch64_valid_symref):
18187 Use aarch64_classify_symbol_expression.
18189 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18191 * config/arm/constraints.md (Ts): New constraint.
18192 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
18194 (compare_scc): Use "Ts" constraint for operand 0.
18195 (ior_scc_scc): Likewise.
18196 (and_scc_scc): Likewise.
18197 (and_scc_scc_nodom): Likewise.
18198 (ior_scc_scc_cmp): Likewise for operand 7.
18199 (and_scc_scc_cmp): Likewise.
18200 * config/arm/thumb2.md (thumb2_movsi_insn):
18201 Add alternatives for 16-bit encodings.
18202 (thumb2_movhi_insn): Likewise.
18203 (thumb2_movsicc_insn): Likewise.
18204 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
18205 (thumb2_negscc): Use "Ts" constraint.
18206 Move mvn instruction outside cond_exec block.
18207 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
18208 for 16-bit encodings.
18210 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18212 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
18214 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
18215 (mulsi3subsi): Likewise.
18216 (mulsidi3adddi): Likewise.
18217 (mulsidi3_v6): Likewise.
18218 (umulsidi3_v6): Likewise.
18219 (umulsidi3adddi_v6): Likewise.
18220 (smulsi3_highpart_v6): Likewise.
18221 (umulsi3_highpart_v6): Likewise.
18222 (mulhisi3tb): Likewise.
18223 (mulhisi3bt): Likewise.
18224 (mulhisi3tt): Likewise.
18225 (maddhisi4): Likewise.
18226 (maddhisi4tb): Likewise.
18227 (maddhisi4tt): Likewise.
18228 (maddhidi4): Likewise.
18229 (maddhidi4tb): Likewise.
18230 (maddhidi4tt): Likewise.
18231 (zeroextractsi_compare0_scratch): Likewise.
18232 (insv_zero): Likewise.
18233 (insv_t2): Likewise.
18234 (anddi_notzesidi_di): Likewise.
18235 (anddi_notsesidi_di): Likewise.
18236 (andsi_notsi_si): Likewise.
18237 (iordi_zesidi_di): Likewise.
18238 (xordi_zesidi_di): Likewise.
18239 (andsi_iorsi3_notsi): Likewise.
18240 (smax_0): Likewise.
18241 (smax_m1): Likewise.
18242 (smin_0): Likewise.
18243 (not_shiftsi): Likewise.
18244 (unaligned_loadsi): Likewise.
18245 (unaligned_loadhis): Likewise.
18246 (unaligned_loadhiu): Likewise.
18247 (unaligned_storesi): Likewise.
18248 (unaligned_storehi): Likewise.
18249 (extv_reg): Likewise.
18250 (extzv_t2): Likewise.
18251 (divsi3): Likewise.
18252 (udivsi3): Likewise.
18253 (arm_zero_extendhisi2addsi): Likewise.
18254 (arm_zero_extendqisi2addsi): Likewise.
18255 (compareqi_eq0): Likewise.
18256 (arm_extendhisi2_v6): Likewise.
18257 (arm_extendqisi2addsi): Likewise.
18258 (arm_movt): Likewise.
18259 (thumb2_ldrd): Likewise.
18260 (thumb2_ldrd_base): Likewise.
18261 (thumb2_ldrd_base_neg): Likewise.
18262 (thumb2_strd): Likewise.
18263 (thumb2_strd_base): Likewise.
18264 (thumb2_strd_base_neg): Likewise.
18265 (arm_negsi2): Add alternative for 16-bit encoding.
18266 (arm_one_cmplsi2): Likewise.
18268 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18270 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
18271 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
18272 (movdfcc): Likewise.
18273 * config/arm/vfp.md (*thumb2_movsf_vfp):
18274 Disable predication for arm_restrict_it.
18275 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
18276 (*thumb2_movdfcc_vfp): Likewise.
18277 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
18278 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
18279 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
18280 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
18281 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
18282 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
18283 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
18284 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
18285 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
18286 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
18287 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
18288 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
18289 Disable predication for arm_restrict_it.
18291 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
18293 * config/i386/bmiintrin.h (_bextr_u32): New.
18294 (_bextr_u64): Ditto.
18296 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
18298 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
18299 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
18300 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
18301 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
18302 * config/mips/n32-elf.h: ...this new file.
18304 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
18307 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
18308 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
18310 2013-06-27 Catherine Moore <clm@codesourcery.com>
18312 * config/mips/mips-tables.opt: Regenerate.
18313 * config/mips/mips-cpus.def: Add m14ke and m14kec.
18314 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
18315 * doc/invoke.texi: Add -m14kc.
18317 2013-06-27 Jakub Jelinek <jakub@redhat.com>
18320 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
18321 constraints of operand 1 and 2.
18324 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
18325 to match RTL canonicalization. Swap predicates and
18326 constraints of operand 1 and 2.
18328 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
18330 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
18331 Process OP_INOUT regs for splitting too.
18333 2013-06-27 Jakub Jelinek <jakub@redhat.com>
18335 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
18336 decl before the loop, initialize to NULL.
18337 (vectorizable_load): Initialize ptr_incr to NULL.
18339 2013-06-27 Martin Jambor <mjambor@suse.cz>
18342 * ipa-ref.h (ipa_maybe_record_reference): Declare.
18343 * ipa-ref.c (ipa_maybe_record_reference): New function.
18344 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
18345 * ipa-cp.c (create_specialized_node): Record potential references from
18347 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
18349 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
18351 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
18352 parameter 'mode' of type 'enum machine_mode mode'; change to pass
18353 'mode' to force_reg.
18354 (aarch64_add_offset): Update calls to aarch64_force_temporary.
18355 (aarch64_expand_mov_immediate): Likewise.
18357 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
18359 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
18360 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
18362 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18364 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
18365 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
18366 (struct machine_function): Add tbegin_p.
18367 (s390_canonicalize_comparison): Fold CC mode compares to
18368 conditional jump if possible.
18369 (s390_emit_jump): Return the emitted jump.
18370 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
18371 Handle CCRAWmode compares.
18372 (s390_option_override): Default to -mhtm if available.
18373 (s390_reg_clobbered_rtx): Handle floating point regs as well.
18374 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
18375 FPRs instead of df_regs_ever_live_p.
18376 (s390_optimize_nonescaping_tx): New function.
18377 (s390_init_frame_layout): Extend clobbered_regs array to cover
18379 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
18380 (s390_expand_tbegin): New function.
18381 (enum s390_builtin): New enum definition.
18382 (code_for_builtin): New array definition.
18383 (s390_init_builtins): New function.
18384 (s390_expand_builtin): New function.
18385 (TARGET_INIT_BUILTINS): Define.
18386 (TARGET_EXPAND_BUILTIN): Define.
18387 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
18388 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
18389 (s390_alc_comparison): Likewise.
18390 * config/s390/s390-modes.def: Add CCRAWmode.
18391 * config/s390/s390.h (processor_flags): Add PF_TX.
18392 (TARGET_CPU_HTM): Define macro.
18393 (TARGET_HTM): Define macro.
18394 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
18395 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
18396 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
18397 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
18398 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
18399 (TBEGIN_MASK, TBEGINC_MASK): New constants.
18400 ("*cc_to_int"): Move up.
18401 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
18402 constants other than 0.
18403 ("*ccraw_to_int"): New insn and splitter definition.
18404 ("tbegin", "tbegin_nofloat", "tbegin_retry")
18405 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
18406 ("tx_assist"): New expander.
18407 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
18408 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
18409 * config/s390/s390.opt: Add -mhtm option.
18410 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
18411 * config/s390/htmxlintrin.h: New file.
18412 * config/s390/htmintrin.h: New file.
18413 * config/s390/s390intrin.h: New file.
18414 * doc/extend.texi: Document htm builtins.
18415 * config.gcc: Add the new header files to extra_headers.
18417 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
18419 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
18420 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
18422 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
18423 Pat Haugen <pthaugen@us.ibm.com>
18424 Peter Bergner <bergner@vnet.ibm.com>
18426 * config/rs6000/power8.md: New.
18427 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
18428 setting for power8 entry.
18429 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
18430 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
18431 test for Power4/Power5 only.
18432 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
18434 (force_new_group): Adjust comment.
18435 * config/rs6000/rs6000.md: Include power8.md.
18437 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
18439 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
18440 * config/arm/arm-protos.h (arm_max_conditional_execute): New
18442 (tune_params): Update comment.
18443 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
18444 (arm_max_conditional_execute): New function.
18445 (thumb2_final_prescan_insn): Use max_insn_skipped and
18446 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
18448 2013-06-25 Jakub Jelinek <jakub@redhat.com>
18450 PR tree-optimization/57705
18451 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
18452 SSA_NAME step, provided that it is not defined inside the loop.
18453 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
18454 (get_initial_def_for_induction): Handle SSA_NAME IV step.
18456 2013-06-25 Martin Jambor <mjambor@suse.cz>
18458 PR middle-end/57670
18459 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
18460 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
18462 (ipa_note_param_call): Initialize member_ptr flag.
18463 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
18464 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
18465 (ipa_write_indirect_edge_info): Stream member_ptr flag.
18466 (ipa_read_indirect_edge_info): Likewise.
18468 2013-06-25 Richard Biener <rguenther@suse.de>
18470 PR middle-end/56977
18471 * passes.c (init_optimization_passes): Move pass_fold_builtins
18472 and pass_dce earlier with -Og.
18474 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
18476 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
18477 <BIT_FIELD_REF>: Remove trailing TAB.
18478 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
18481 2013-06-24 Martin Jambor <mjambor@suse.cz>
18483 PR tree-optimization/57358
18484 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
18485 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
18486 (ipa_analyze_params_uses): Generate pessimistic info when true.
18488 2013-06-24 Martin Jambor <mjambor@suse.cz>
18490 PR tree-optimization/57539
18491 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
18492 global.inlined_to of the new node to it. All callers changed.
18493 * ipa-inline-transform.c (clone_inlined_nodes): New variable
18494 inlining_into, pass it to cgraph_clone_node.
18495 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
18496 ipa_free_edge_args_substructures.
18497 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
18498 rdesc linked list. Do not assert rdesc edges have inlined caller.
18499 Assert we have found an rdesc in the rdesc list.
18501 2013-06-24 Richard Biener <rguenther@suse.de>
18503 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
18504 (pointer_set_lookup): Declare.
18505 (class pointer_map): New template class implementing a
18506 generic pointer to T map.
18507 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
18508 pointer_map<T>::contains, pointer_map<T>::insert,
18509 pointer_map<T>::traverse): New functions.
18510 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
18511 (pointer_set_lookup): New function.
18512 (pointer_set_contains): Use pointer_set_lookup.
18513 (pointer_set_insert): Likewise.
18514 (insert_aux): Remove.
18515 (struct pointer_map_t): Embed a pointer_set_t.
18516 (pointer_map_create): Adjust.
18517 (pointer_map_destroy): Likewise.
18518 (pointer_map_contains): Likewise.
18519 (pointer_map_insert): Likewise.
18520 (pointer_map_traverse): Likewise.
18521 * tree-streamer.h (struct streamer_tree_cache_d): Use a
18522 pointer_map<unsigned> instead of a pointer_map_t.
18523 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
18524 (streamer_tree_cache_lookup): Likewise.
18525 (streamer_tree_cache_create): Likewise.
18526 (streamer_tree_cache_delete): Likewise.
18527 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
18528 pointer_map<unsigned> instead of a pointer_map_t.
18529 (lto_init_tree_ref_encoder): Adjust.
18530 (lto_destroy_tree_ref_encoder): Likewise.
18531 * lto-section-out.c (lto_output_decl_index): Likewise.
18532 (lto_record_function_out_decl_state): Likewise.
18533 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
18535 2013-06-24 Richard Biener <rguenther@suse.de>
18537 PR tree-optimization/57488
18538 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
18540 2013-06-24 Alan Modra <amodra@gmail.com>
18542 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
18543 (gen_easy_altivec_constant): Likewise.
18544 * config/rs6000/predicates.md (easy_vector_constant_add_self,
18545 easy_vector_constant_msb): Likewise.
18547 2013-06-23 Jakub Jelinek <jakub@redhat.com>
18550 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
18551 add missing return true.
18553 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
18556 * config/sh/predicates.md (general_extend_operand): Invoke
18557 general_movsrc_operand for memory operands.
18558 (general_movsrc_operand): Allow reg+reg addressing, do not use
18559 general_operand for memory operands.
18561 2013-06-23 Sriraman Tallam <tmsriram@google.com>
18563 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
18564 when current target options does not apply.
18565 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
18566 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
18567 * config/i386/bmiintrin.h: Pass appropriate target
18568 attributes to header.
18569 * config/i386/mmintrin.h: Ditto.
18570 * config/i386/nmmintrin.h: Ditto.
18571 * config/i386/avx2intrin.h: Ditto.
18572 * config/i386/fxsrintrin.h: Ditto.
18573 * config/i386/tbmintrin.h: Ditto.
18574 * config/i386/xsaveintrin.h: Ditto.
18575 * config/i386/f16cintrin.h: Ditto.
18576 * config/i386/xtestintrin.h: Ditto.
18577 * config/i386/xsaveoptintrin.h: Ditto.
18578 * config/i386/bmi2intrin.h: Ditto.
18579 * config/i386/lzcntintrin.h: Ditto.
18580 * config/i386/smmintrin.h: Ditto.
18581 * config/i386/wmmintrin.h: Ditto.
18582 * config/i386/x86intrin.h: Remove all header include guards.
18583 * config/i386/prfchwintrin.h: Ditto.
18584 * config/i386/pmmintrin.h: Ditto.
18585 * config/i386/tmmintrin.h: Ditto.
18586 * config/i386/xmmintrin.h: Ditto.
18587 * config/i386/popcntintrin.h: Ditto.
18588 * config/i386/rdseedintrin.h: Ditto.
18589 * config/i386/ammintrin.h: Ditto.
18590 * config/i386/emmintrin.h: Ditto.
18591 * config/i386/immintrin.h: Remove all header include guards.
18592 * config/i386/fma4intrin.h: Ditto.
18593 * config/i386/lwpintrin.h: Ditto.
18594 * config/i386/xopintrin.h: Ditto.
18595 * config/i386/ia32intrin.h: Ditto.
18596 * config/i386/avxintrin.h: Ditto.
18597 * config/i386/rtmintrin.h: Ditto.
18598 * config/i386/fmaintrin.h: Ditto.
18599 * config/i386/mm3dnow.h: Ditto.
18601 2013-06-22 Sriraman Tallam <tmsriram@google.com>
18603 * common/config/i386/i386-common.c: Handle LZCNT.
18605 2013-06-22 Andi Kleen <ak@linux.intel.com>
18607 * doc/extend.texi: Use __atomic_store_n instead of
18608 __atomic_store in HLE example.
18610 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
18612 * config/sh/sh.c: Remove <cstdlib> workaround.
18614 2013-06-21 Andi Kleen <ak@linux.intel.com>
18616 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
18618 2013-06-21 Andi Kleen <ak@linux.intel.com>
18620 * doc/extend.texi: Document that __atomic_clear and
18621 __atomic_test_and_set should only be used with bool.
18623 2013-06-20 Jan Hubicka <jh@suse.cz>
18625 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
18626 types_same_for_odr.
18627 * tree.c (decls_same_for_odr): New function.
18628 (same_for_edr): New function.
18629 (types_same_for_odr): New function.
18630 (get_binfo_at_offset): Use it.
18631 * tree.h (types_same_for_odr): Declare.
18633 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
18634 Jason Merrill <jason@redhat.com>
18636 * system.h: Include <cstdlib> as well as <stdlib.h>.
18638 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
18641 * config/i386/i386.c (construct_container): Report error if
18642 long double is used with disabled x87 float returns.
18644 2013-06-20 Jan Hubicka <jh@suse.cz>
18646 * lto-cgraph.c (input_symtab): Do not set cgraph state.
18648 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
18650 PR rtl-optimization/57425
18651 PR rtl-optimization/57569
18652 * alias.c (write_dependence_p): Remove parameters mem_mode and
18653 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
18654 Changed all callers.
18655 (canon_anti_dependence): Get comments and semantics in sync.
18656 Add parameter mem_canonicalized. Changed all callers.
18657 * rtl.h (canon_anti_dependence): Update prototype.
18659 2013-06-20 Richard Biener <rguenther@suse.de>
18661 * data-streamer-in.c (streamer_read_uhwi): Optimize single
18662 byte case, inline streamer_read_uchar and defer section
18665 2013-06-20 Richard Biener <rguenther@suse.de>
18667 PR tree-optimization/57584
18668 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
18669 SSA names into the expanded expression that take part in
18670 abnormal coalescing.
18672 2013-06-19 Sharad Singhai <singhai@google.com>
18674 * gcov.c (print_usage): Handle new option.
18675 (process_args): Ditto.
18676 (get_gcov_intermediate_filename): New function.
18677 (output_intermediate_file): New function.
18678 (output_gcov_file): New function
18679 (generate_results): Handle new option.
18680 (release_function): Relase demangled name.
18681 (read_graph_file): Handle demangled name.
18682 (output_lines): Ditto.
18683 * doc/gcov.texi: Document gcov intermediate format.
18685 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
18688 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
18689 (lra_emit_add): Use the functions. Add comment about Y as an
18692 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
18695 * collect2.c (collect_atexit): New.
18696 (collect_exit): Delete.
18697 (main): Register collect_atexit with atexit.
18698 (collect_wait): Change collect_exit to exit.
18700 * collect2.h (collect_exit): Delete.
18701 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
18703 2013-06-19 Wei Mi <wmi@google.com>
18705 PR rtl-optimization/57518
18706 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
18707 if regno is used in paradoxical subreg.
18708 (update_equiv_regs): Check pdx_subregs[regno] before
18709 set a reg to be equivalent with a mem.
18711 2013-06-19 Matthias Klose <doko@ubuntu.com>
18714 * file-find.h (find_a_file): Add a mode parameter.
18715 * file-find.c (find_a_file): Likewise.
18716 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
18717 with X_OK for the executables.
18718 * collect2.c (main): Call find_a_file with X_OK.
18720 2013-06-19 Steve Ellcey <sellcey@mips.com>
18723 * config/mips/mips.md (casesi_internal_mips16_<mode>):
18724 Use NEXT_INSN instead of next_real_insn.
18726 2013-06-19 Jan Hubicka <jh@suse.cz>
18728 * cgraph.h (const_value_known_p): Replace by ...
18729 (ctor_for_folding): .. this one.
18730 * cgraphunit.c (process_function_and_variable_attributes): Use it.
18731 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
18732 * expr.c (expand_expr_real_1): Likewise.
18733 (string_constant): Likewise.
18734 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
18735 * ipa.c (process_references): Likewise.
18736 (symtab_remove_unreachable_nodes): Likewise.
18737 * ipa-inline-analysis.c (param_change_prob): Likewise.
18738 * gimple-fold.c (canonicalize_constructor_val): Likewise.
18739 (get_base_constructor): Likwise.
18740 * varpool.c (varpool_remove_node): Likewise.
18741 (varpool_remove_initializer): LIkewise.
18742 (dump_varpool_node): LIkwise.
18743 (const_value_known_p): Rewrite to ...
18744 (ctor_for_folding): ... this one.
18746 2013-06-19 Jakub Jelinek <jakub@redhat.com>
18749 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
18750 PERSONALITY in $PATH derived prefixes.
18752 2013-06-19 Jeff Law <law@redhat.com>
18754 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
18757 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
18758 (simplify_bitwise_binary): Use it to simpify certain binary ops on
18761 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
18763 * config/arm/vfp.md: Move VFP instruction classification documentation
18765 * config/arm/arm.md: ... here. Update instruction classification
18768 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
18770 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
18771 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
18772 pattern. Use more efficient sequences on ARMv5 and Thumb2.
18774 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
18777 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
18778 with NEXT_INSN. Use tablejump_p to check for jump table data
18781 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
18784 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
18785 that now in C++ the value is correct per the C++ standards.
18787 2013-06-19 Richard Biener <rguenther@suse.de>
18789 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
18790 for global context.
18792 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18795 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18798 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
18799 with next_active_insn.
18801 2013-06-18 Sriraman Tallam <tmsriram@google.com>
18803 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
18804 functions are inlined during failures to flag an error.
18805 * tree-inline.c (expand_call_inline): Allow the error to be flagged
18806 in early inline pass.
18808 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
18810 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
18813 2013-06-18 Julian Brown <julian@codesourcery.com>
18815 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
18816 Permit virtual register pre-reload if !strict.
18817 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
18819 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
18821 * config/arm/neon.md (movmisalign<mode>): Use
18822 neon_perm_struct_or_reg_operand instead of
18823 neon_struct_or_register_operand.
18824 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
18825 neon_permissive_struct_operand instead of neon_struct_operand.
18826 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
18827 neon_vector_mem_operand.
18828 * config/arm/predicates.md (neon_struct_operand): Adjust call to
18829 neon_vector_mem_operand.
18830 (neon_permissive_struct_operand): New.
18831 (neon_struct_or_register_operand): Rename to...
18832 (neon_perm_struct_or_reg_operand): This. Adjust call to
18833 neon_vector_mem_operand.
18835 2013-06-18 Richard Biener <rguenther@suse.de>
18837 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
18838 * lto-streamer.h: Include pointer-set.h.
18839 (struct lto_decl_slot): Remove.
18840 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
18841 Remove next_index entry.
18842 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
18843 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
18844 (lto_init_tree_ref_encoder): Adjust.
18845 (lto_destroy_tree_ref_encoder): Likewise.
18846 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
18847 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
18848 (lto_output_decl_index): Adjust.
18849 (lto_new_out_decl_state): Likewise.
18850 (lto_record_function_out_decl_state): Likewise.
18851 * lto-streamer-out.c (copy_function): Likewise.
18853 2013-06-18 Richard Biener <rguenther@suse.de>
18855 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
18856 * cgraphunit.c: Include cfgloop.h.
18857 (init_lowered_empty_function): Initialize the loop tree.
18858 (assemble_thunk): Insert new BBs into loops.
18860 2013-06-18 Richard Biener <rguenther@suse.de>
18862 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
18863 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
18864 the map from cache entry to cache index optional.
18865 (streamer_tree_cache_replace_tree): Adjust accordingly.
18866 (streamer_tree_cache_append): Likewise.
18867 (streamer_tree_cache_delete): Likewise.
18868 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
18869 streamer cache map from cache entry to cache index.
18870 * lto-streamer-out.c (create_output_block): Adjust.
18872 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
18874 * config/arm/arm.md (attribute "insn"): Move multiplication and
18875 division attributes to...
18876 (attribute "type"): ... here. Remove mult.
18877 (attribute "mul32"): New attribute.
18878 (attribute "mul64"): Add umaal.
18879 (*arm_mulsi3): Update attributes.
18880 (*arm_mulsi3_v6): Likewise.
18881 (*thumb_mulsi3): Likewise.
18882 (*thumb_mulsi3_v6): Likewise.
18883 (*mulsi3_compare0): Likewise.
18884 (*mulsi3_compare0_v6): Likewise.
18885 (*mulsi_compare0_scratch): Likewise.
18886 (*mulsi_compare0_scratch_v6): Likewise.
18887 (*mulsi3addsi): Likewise.
18888 (*mulsi3addsi_v6): Likewise.
18889 (*mulsi3addsi_compare0): Likewise.
18890 (*mulsi3addsi_compare0_v6): Likewise.
18891 (*mulsi3addsi_compare0_scratch): Likewise.
18892 (*mulsi3addsi_compare0_scratch_v6): Likewise.
18893 (*mulsi3subsi): Likewise.
18894 (*mulsidi3adddi): Likewise.
18895 (*mulsi3addsi_v6): Likewise.
18896 (*mulsidi3adddi_v6): Likewise.
18897 (*mulsidi3_nov6): Likewise.
18898 (*mulsidi3_v6): Likewise.
18899 (*umulsidi3_nov6): Likewise.
18900 (*umulsidi3_v6): Likewise.
18901 (*umulsidi3adddi): Likewise.
18902 (*umulsidi3adddi_v6): Likewise.
18903 (*smulsi3_highpart_nov6): Likewise.
18904 (*smulsi3_highpart_v6): Likewise.
18905 (*umulsi3_highpart_nov6): Likewise.
18906 (*umulsi3_highpart_v6): Likewise.
18907 (mulhisi3): Likewise.
18908 (*mulhisi3tb): Likewise.
18909 (*mulhisi3bt): Likewise.
18910 (*mulhisi3tt): Likewise.
18911 (maddhisi4): Likewise.
18912 (*maddhisi4tb): Likewise.
18913 (*maddhisi4tt): Likewise.
18914 (maddhidi4): Likewise.
18915 (*maddhidi4tb): Likewise.
18916 (*maddhidi4tt): Likewise.
18917 (divsi3): Likewise.
18918 (udivsi3): Likewise.
18919 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
18920 (thumb2_mulsi_short_compare0): Likewise.
18921 (thumb2_mulsi_short_compare0_scratch): Likewise.
18922 * config/arm/arm1020e.md (1020mult1): Update attribute change.
18923 (1020mult2): Likewise.
18924 (1020mult3): Likewise.
18925 (1020mult4): Likewise.
18926 (1020mult5): Likewise.
18927 (1020mult6): Likewise.
18928 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
18930 (cortex_a15_mult64): Likewise.
18931 (cortex_a15_sdiv): Likewise.
18932 (cortex_a15_udiv): Likewise.
18933 * config/arm/arm1026ejs.md (mult1): Update attribute change.
18939 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
18940 (pj4_ir_div): Likewise.
18941 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
18942 (11_mult2): Likewise.
18943 (11_mult3): Likewise.
18944 (11_mult4): Likewise.
18945 (11_mult5): Likewise.
18946 (11_mult6): Likewise.
18947 (11_mult7): Likewise.
18948 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
18949 (cortex_a8_mla): Likewise.
18950 (cortex_a8_mull): Likewise.
18951 (cortex_a8_smulwy): Likewise.
18952 (cortex_a8_smlald): Likewise.
18953 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
18954 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
18955 (cortex_r4_mul_3): Likewise.
18956 (cortex_r4_mla_4): Likewise.
18957 (cortex_r4_mla_3): Likewise.
18958 (cortex_r4_smlald): Likewise.
18959 (cortex_r4_mull): Likewise.
18960 (cortex_r4_sdiv): Likewise.
18961 (cortex_r4_udiv): Likewise.
18962 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
18963 (cortex_a7_idiv): Likewise.
18964 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
18965 (9_mult2): Likewise.
18966 (9_mult3): Likewise.
18967 (9_mult4): Likewise.
18968 (9_mult5): Likewise.
18969 (9_mult6): Likewise.
18970 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
18971 (cortex_a53_sdiv): Likewise.
18972 (cortex_a53_udiv): Likewise.
18973 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
18974 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
18975 (mp626_mult2): Likewise.
18976 (mp626_mult3): Likewise.
18977 (mp626_mult4): Likewise.
18978 * config/arm/fa526.md (526_mult1): Update attribute change.
18979 (526_mult2): Likewise.
18980 * config/arm/arm-generic.md (mult): Update attribute change.
18981 (mult_ldsched_strongarm): Likewise.
18982 (mult_ldsched): Likewise.
18983 (multi_cycle): Likewise.
18984 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
18985 * config/arm/fa606te.md (606te_mult1): Update attribute change.
18986 (606te_mult2): Likewise.
18987 (606te_mult3): Likewise.
18988 (606te_mult4): Likewise.
18989 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
18990 (cortex_a9_mac16): Likewise.
18991 (cortex_a9_multiply): Likewise.
18992 (cortex_a9_mac): Likewise.
18993 (cortex_a9_multiply_long): Likewise.
18994 * config/arm/fa626te.md (626te_mult1): Update attribute change.
18995 (626te_mult2): Likewise.
18996 (626te_mult3): Likewise.
18997 (626te_mult4): Likewise.
18999 2013-06-18 Richard Biener <rguenther@suse.de>
19002 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
19004 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19007 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
19008 with next_active_insn.
19010 2013-06-18 Alan Modra <amodra@gmail.com>
19012 * config/rs6000/rs6000.h (enum data_align): New.
19013 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
19014 (DATA_ABI_ALIGNMENT): Define.
19015 (CONSTANT_ALIGNMENT): Correct comment.
19016 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
19017 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
19019 2013-06-17 David Malcolm <dmalcolm@redhat.com>
19021 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
19022 ATTRIBUTE_UNUSED marking.
19024 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
19026 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
19027 alternative and update.
19028 (aarch64_dup_lanedi): Delete.
19029 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
19030 * config/aarch64/aarch64-simd-builtins.def: Update.
19032 2013-06-17 Richard Biener <rguenther@suse.de>
19034 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
19035 (lto_input_scc): Declare.
19036 (lto_input_tree_1): Likewise.
19037 (struct lto_stats_d): Add num_tree_bodies_output and
19038 num_pickle_refs_output.
19039 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
19040 (lto_read_tree_1): Split out from ...
19041 (lto_read_tree): ... this.
19042 (lto_input_scc): New function.
19043 (lto_input_tree_1): Split out from ...
19044 (lto_input_tree): ... this. Handle LTO_tree_scc.
19045 (lto_data_in_create): Create the streamer cache without hashes.
19046 * lto-streamer-out.c (create_output_block): Create the streamer
19047 cache with hashes when not doing WPA.
19048 (lto_write_tree_1): Split out from ...
19049 (lto_write_tree): ... this.
19050 (get_symbol_initial_value): New function.
19051 (lto_output_tree_1): Split out from ...
19052 (lto_output_tree): ... this. Write trees as series of SCCs
19053 using a DFS walk via DFS_write_tree.
19054 (struct sccs, struct scc_entry): New types.
19055 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
19056 (DFS_write_tree_body): New function.
19057 (DFS_write_tree): Likewise.
19058 (hash_tree): Likewise.
19059 (scc_entry_compare): Likewise.
19060 (hash_scc): Likewise.
19061 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
19062 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
19063 TREE_CHAIN as regular reference.
19064 (streamer_read_integer_cst): Remove.
19065 (streamer_get_pickled_tree): Adjust.
19066 * tree-streamer-out.c (streamer_write_chain): Disable streaming
19067 of DECL_EXTERNALs in BLOCK_VARS for now.
19068 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
19070 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
19071 Add hash value argument and record that if hashes are recorded
19073 (streamer_tree_cache_insert_1): Adjust.
19074 (streamer_tree_cache_insert): Likewise.
19075 (streamer_tree_cache_insert_at): Rename to ...
19076 (streamer_tree_cache_replace_tree): ... this and adjust.
19077 (streamer_tree_cache_append): Adjust.
19078 (record_common_node): Likewise.
19079 (streamer_tree_cache_create): Add argument whether to
19080 record hash values together with trees.
19081 (streamer_tree_cache_delete): Adjust.
19082 * tree-streamer.h (struct streamer_tree_cache_d): Add
19084 (streamer_read_integer_cst): Remove.
19085 (streamer_tree_cache_insert): Adjust.
19086 (streamer_tree_cache_append): Likewise.
19087 (streamer_tree_cache_insert_at): Rename to ...
19088 (streamer_tree_cache_replace_tree): ... this and adjust.
19089 (streamer_tree_cache_create): Add argument whether to record hashes.
19090 (streamer_tree_cache_get): Rename to ...
19091 (streamer_tree_cache_get_tree): ... this.
19092 (streamer_tree_cache_get_hash): New function.
19093 * tree.c (cache_integer_cst): New function.
19094 * tree.h (cache_integer_cst): Declare.
19095 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
19096 * lto-symtab.c (lto_varpool_replace_node): Only release
19097 DECL_INITIAL of non-prevailing decls.
19098 * varpool.c (varpool_remove_initializer): Do not release
19099 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
19101 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
19103 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
19104 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
19105 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
19106 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
19107 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
19108 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
19109 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
19110 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
19111 instead of TARGET_64BIT.
19112 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
19113 Require ISA_HAS_<D>DIV.
19115 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
19117 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
19118 (mips*-*-linux*): Move default with_llsc setting to where other
19120 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
19122 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
19123 Likewise. Remove default with_tune setting. Move default float
19124 setting to its own block. Handle with_llsc in the same block as above.
19126 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
19128 PR rtl-optimization/57425
19129 PR rtl-optimization/57569
19130 * alias.c (write_dependence_p): Add new parameters mem_mode,
19131 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
19132 Changed all callers.
19133 (canon_anti_dependence): New function.
19134 * cse.c (check_dependence): Use canon_anti_dependence.
19135 * cselib.c (cselib_invalidate_mem): Likewise.
19136 * rtl.h (canon_anti_dependence): Declare.
19138 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
19140 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
19141 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
19142 ".set mips3" for 64-bit targets.
19144 2013-06-15 Dehao Chen <dehao@google.com>
19146 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
19147 * gimple-low.c (gimple_check_call_matching_types): Likewise.
19148 (gimple_check_call_args): Likewise.
19149 * value-prof.c (check_ic_target): Likewise.
19150 * ipa-inline.c (early_inliner): Likewise.
19151 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
19152 * cgraph.c (cgraph_create_edge_1): Likewise.
19153 (cgraph_make_edge_direct): Likewise.
19155 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
19158 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
19159 rs6000_output_move_128bit to handle emitting quad memory
19160 operations. Set attribute length to 8 bytes.
19162 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
19164 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
19166 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
19167 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
19168 (aarch64_<su>mlsl<mode>): Likewise.
19170 2013-06-14 Mike Stump <mikestump@comcast.net>
19172 * Makefile.in (TARGET_H): Add insn-codes.h.
19174 2013-06-14 Alan Modra <amodra@gmail.com>
19176 PR middle-end/57134
19177 PR middle-end/57586
19178 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
19179 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
19180 bitfield expansion when EXPAND_MEMORY.
19181 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
19183 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
19185 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
19186 test for clearing quad memory on 32-bit later.
19188 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
19190 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
19191 (fold_negate_expr): Likewise.
19192 (fold_real_zero_addition_p): Handle vectors.
19193 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
19195 2013-06-14 Alan Modra <amodra@gmail.com>
19197 * varasm.c (force_const_mem): Revert 2013-06-07 change.
19199 2013-06-13 Jan Hubicka <jh@suse.cz>
19201 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
19202 Local comdats are not externally visible.
19203 * symtab.c (dump_symtab_base): Dump externally visible.
19204 (verify_symtab_base): Verify back links in the symtab hash.
19206 2013-06-13 Bin Cheng <bin.cheng@arm.com>
19208 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
19209 CONVERT_EXPR as equal nodes.
19211 2013-06-13 Bin Cheng <bin.cheng@arm.com>
19213 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
19215 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
19217 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
19218 Generalize to complex and vector.
19219 * tree.c (build_all_ones_cst): New function.
19220 * tree.h (build_all_ones_cst): Declare it.
19222 2013-06-13 Alan Modra <amodra@gmail.com>
19224 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
19225 * config/rs6000/rs6000.md (signbittf2): New insn.
19226 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
19227 (abstf2_internal, cmptf_internal2): Likewise.
19228 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
19230 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
19231 Pat Haugen <pthaugen@us.ibm.com>
19232 Peter Bergner <bergner@vnet.ibm.com>
19234 * config/rs6000/rs6000.c (emit_load_locked): Add support for
19235 power8 byte, half-word, and quad-word atomic instructions.
19236 (emit_store_conditional): Likewise.
19237 (rs6000_expand_atomic_compare_and_swap): Likewise.
19238 (rs6000_expand_atomic_op): Likewise.
19240 * config/rs6000/sync.md (larx): Add new modes for power8.
19242 (AINT): New mode iterator to include TImode as well as normal
19243 integer modes on power8.
19244 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
19245 that VSX registers are not considered. Use AINT mode iterator
19246 instead of INT1 to allow inclusion of quad word atomic operations
19248 (load_locked<mode>): Likewise.
19249 (store_conditional<mode>): Likewise.
19250 (atomic_compare_and_swap<mode>): Likewise.
19251 (atomic_exchange<mode>): Likewise.
19252 (atomic_nand<mode>): Likewise.
19253 (atomic_fetch_<fetchop_name><mode>): Likewise.
19254 (atomic_nand_fetch<mode>): Likewise.
19255 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
19257 (ATOMIC): On power8, add QImode, HImode modes.
19258 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
19259 modes that promote to SImode.
19260 (load_lockedti): Convert TImode arguments to PTImode, so that we
19261 get a guaranteed even/odd register pair.
19262 (load_lockedpti): Likewise.
19263 (store_conditionalti): Likewise.
19264 (store_conditionalpti): Likewise.
19266 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
19267 atomic load/store instructions.
19270 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
19272 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
19274 (insn_count): New attribute, with most cases extracted from...
19275 (length): ...here. Redefine most cases in terms of insn_count.
19276 (single_insn): Delete.
19277 (can_delay): Use insn_count to check for single instructions.
19278 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
19279 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
19280 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
19281 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
19282 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
19283 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
19284 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
19285 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
19286 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
19287 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
19288 rather than "length".
19289 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
19290 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
19291 Use "insn_count" rather than "length".
19292 * config/mips/mips-dsp.md
19293 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
19294 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
19297 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
19299 PR tree-optimization/57361
19300 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
19302 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
19304 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
19306 (aarch64_simd_combine<mode>): New instruction expansion.
19307 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
19308 function prototype.
19309 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
19310 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
19312 2013-06-12 Jan Hubicka <jh@suse.cz>
19314 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
19315 decl has when in streaming stage.
19316 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
19317 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
19319 2013-06-12 Roland Stigge <stigge@antcom.de>
19322 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
19324 2013-06-12 Jakub Jelinek <jakub@redhat.com>
19326 PR tree-optimization/57537
19327 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
19328 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
19330 2013-06-12 Richard Biener <rguenther@suse.de>
19332 * data-streamer.h (streamer_write_char_stream): CSE
19333 obs->current_pointer.
19334 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
19335 streamer_write_char_stream manually and optimize the resulting loop.
19336 (streamer_write_hwi_stream): Likewise.
19338 2013-06-12 Jan Hubicka <jh@suse.cz>
19340 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
19341 * cgraph.h (varpool_create_empty_node): Declare.
19342 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
19344 * symtab.c (symtab_unregister_node): Be lax about missin entries
19346 (symtab_get_node): Update comment.
19347 * varpool.c (varpool_create_empty_node): Break out from ...
19348 (varpool_node_for_decl): ... here.
19349 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
19351 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
19353 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
19354 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
19355 part. Use straight-line flow at the end.
19356 <COMPONENT_REF>: Remove superfluous else.
19357 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
19359 2013-06-12 Jakub Jelinek <jakub@redhat.com>
19362 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
19363 target hook even for !TREE_PUBLIC decls. If no resolution info
19364 is available, return false for common and external decls.
19366 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
19368 * config/rl78/constraints.md (U): New constraint.
19369 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
19372 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
19375 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
19376 to allow returning address to AT_PLATFORM name.
19378 2013-06-11 Jan Hubicka <jh@suse.cz>
19380 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
19381 * cgraph.h (symtab_node_base): Add weakref flag.
19382 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
19383 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
19384 (output_weakrefs): Use weakref flag.
19385 * fold-const.c (simple_operand_p): Handle WEAK.
19386 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
19387 * ipa.c (varpool_externally_visible_p): Drop weakref.
19388 (function_and_variable_visibility): Update comment; fix weakref
19389 sanity checks; do not clear DECL_WEAK on them.
19390 * lto-cgraph.c (lto_output_node): update.
19391 (lto_output_varpool_node): Update.
19392 (input_overwrite_node): Update.
19393 (input_node): Update.
19394 (input_varpool_node): Update.
19395 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
19396 (lto_symtab_merge_symbols): Add sanity check.
19397 (lto_symtab_prevailing_decl): Do not special case weakrefs.
19398 * passes.c (rest_of_decl_compilation): Set static flag, too.
19399 * symtab.c (dump_symtab_base): Dump weakref.
19400 (verify_symtab_base): Sanity check weakrefs.
19401 (symtab_make_decl_local): Remove duplicated code.
19402 (symtab_alias_ultimate_target): Simplify.
19403 * varpool.c (varpool_create_variable_alias): Set weakref flag.
19405 2013-06-11 Tom de Vries <tom@codesourcery.com>
19407 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
19408 sequence_vect == NULL.
19410 2013-06-11 DJ Delorie <dj@redhat.com>
19412 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
19413 (rl78_unwind_word_mode): New.
19415 2013-06-11 David Malcolm <dmalcolm@redhat.com>
19417 * final.c (debug_prefix_maps): Make static.
19419 2013-06-11 David Malcolm <dmalcolm@redhat.com>
19421 * function.c (initial_trampoline): Remove stray copy.
19423 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
19425 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
19427 2013-06-11 Martin Jambor <mjambor@suse.cz>
19429 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
19430 within bounds at the beginning of the function.
19432 2013-06-11 Alan Modra <amodra@gmail.com>
19434 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
19436 (get_named_section): Don't NULL !DECL_P decl.
19438 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
19440 * doc/invoke.texi (core-avx2): Document.
19442 (atom): Updated with MOVBE.
19444 2013-06-11 Richard Biener <rguenther@suse.de>
19446 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
19448 2013-06-11 Anton Blanchard <anton@samba.org>
19450 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
19451 correct shift value in little-endian mode.
19453 2013-06-11 Jakub Jelinek <jakub@redhat.com>
19456 * varasm.c (get_variable_align): Move #endif to the right place.
19458 2013-06-10 Cary Coutant <ccoutant@google.com>
19460 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
19461 for hash so that hash table traversal order is deterministic.
19463 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
19464 Pat Haugen <pthaugen@us.ibm.com>
19465 Peter Bergner <bergner@vnet.ibm.com>
19467 * config/rs6000/vector.md (GPR move splitter): Do not split moves
19468 of vectors in GPRS if they are direct moves or quad word load or
19471 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
19473 (direct_move_p): Likewise.
19474 (quad_load_store_p): Likewise.
19476 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
19477 classes into bins based on the physical register type.
19478 (reg_class_to_reg_type): Likewise.
19479 (IS_STD_REG_TYPE): Likewise.
19480 (IS_FP_VECT_REG_TYPE): Likewise.
19481 (reload_fpr_gpr): Arrays to determine what insn to use if we can
19482 use direct move instructions.
19483 (reload_gpr_vsx): Likewise.
19484 (reload_vsx_gpr): Likewise.
19485 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
19486 information that is a simplification of register classes. Also
19487 precalculate direct move reload helpers.
19488 (direct_move_p): New function to return true if the operation can
19489 be done as a direct move instruciton.
19490 (quad_load_store_p): New function to return true if the operation
19491 is a quad memory operation.
19492 (rs6000_legitimize_address): If quad memory, only allow register
19493 indirect for TImode addresses.
19494 (rs6000_legitimate_address_p): Likewise.
19495 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
19496 (rs6000_reload_register_type): Likewise.
19497 (register_to_reg_type): Return register type.
19498 (rs6000_secondary_reload_simple_move): New helper function for
19499 secondary reload and secondary memory needed to identify anything
19500 that is a simple move, and does not need reloading.
19501 (rs6000_secondary_reload_direct_move): New helper function for
19502 secondary reload to identify cases that can be done with several
19503 instructions via the direct move instructions.
19504 (rs6000_secondary_reload_move): New helper function for secondary
19505 reload to identify moves between register types that can be done.
19506 (rs6000_secondary_reload): Add support for quad memory operations
19507 and for direct move.
19508 (rs6000_secondary_memory_needed): Likewise.
19509 (rs6000_debug_secondary_memory_needed): Change argument names.
19510 (rs6000_output_move_128bit): New function to return the move to
19511 use for 128-bit moves, including knowing about the various
19512 limitations of quad memory operations.
19514 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
19515 memory operations. call rs6000_output_move_128bit for the actual
19516 instruciton(s) to generate.
19517 (vsx_movti_64bit): Likewise.
19519 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
19520 (UNSPEC_P8V_MTVSRWZ): Likewise.
19521 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
19522 (UNSPEC_P8V_MTVSRD): Likewise.
19523 (UNSPEC_P8V_XXPERMDI): Likewise.
19524 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
19525 (UNSPEC_FUSION_GPR): Likewise.
19526 (FMOVE128_GPR): New iterator for direct move.
19527 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
19528 (f32_sv): Likewise.
19529 (f32_dm): Likewise.
19530 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
19531 loads and direct move instructions.
19532 (zero_extendsidi2_lfiwzx): Likewise.
19533 (extendsidi2_lfiwax): Likewise.
19534 (extendsidi2_nocell): Likewise.
19535 (floatsi<mode>2_lfiwax): Likewise.
19536 (lfiwax): Likewise.
19537 (floatunssi<mode>2_lfiwzx): Likewise.
19538 (lfiwzx): Likewise.
19539 (fix_trunc<mode>_stfiwx): Likewise.
19540 (fixuns_trunc<mode>_stfiwx): Likewise.
19541 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
19542 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
19543 (parity<mode>2_cmpb): Set length/type attr.
19544 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
19545 for 'mr.' to fast_compare.
19546 (bpermd_<mode>): Change type attr to popcnt.
19547 (p8_fmrgow_<mode>): New insns for power8 direct move support.
19548 (p8_mtvsrwz_1): Likewise.
19549 (p8_mtvsrwz_2): Likewise.
19550 (reload_fpr_from_gpr<mode>): Likewise.
19551 (p8_mtvsrd_1): Likewise.
19552 (p8_mtvsrd_2): Likewise.
19553 (p8_xxpermdi_<mode>): Likewise.
19554 (reload_vsx_from_gpr<mode>): Likewise.
19555 (reload_vsx_from_gprsf): Likewise.
19556 (p8_mfvsrd_3_<mode>): LIkewise.
19557 (reload_gpr_from_vsx<mode>): Likewise.
19558 (reload_gpr_from_vsxsf): Likewise.
19559 (p8_mfvsrd_4_disf): Likewise.
19560 (multi-word GPR splits): Do not split direct moves or quad memory
19563 2013-06-10 David Malcolm <dmalcolm@redhat.com>
19565 * tree-into-ssa.c (interesting_blocks): Make static.
19567 2013-06-10 Jakub Jelinek <jakub@redhat.com>
19570 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
19571 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
19572 Use DATA_ABI_ALIGNMENT for that case instead if defined.
19573 (get_variable_align): New function.
19574 (get_variable_section, emit_bss, emit_common,
19575 assemble_variable_contents, place_block_symbol): Use
19576 get_variable_align instead of DECL_ALIGN.
19577 (assemble_noswitch_variable): Add align argument, use it
19578 instead of DECL_ALIGN.
19579 (assemble_variable): Adjust caller. Use get_variable_align
19580 instead of DECL_ALIGN.
19581 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
19583 (DATA_ABI_ALIGNMENT): Define.
19584 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
19585 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
19586 opt is false, only return the psABI mandated alignment increase.
19587 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
19588 (DATA_ABI_ALIGNMENT): ... this.
19589 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
19590 (DATA_ABI_ALIGNMENT): ... this.
19591 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
19592 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
19593 (DATA_ABI_ALIGNMENT): ... this.
19594 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
19595 * doc/tm.texi: Regenerated.
19597 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
19599 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
19600 cmp_code to construct REG_EQUAL note.
19602 2013-06-09 Jakub Jelinek <jakub@redhat.com>
19605 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
19606 that operands[2] doesn't overlap with operands[0].
19608 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
19609 Jan Hubicka <jh@suse.cz>
19611 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
19612 hack to mark symbols as used.
19614 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
19616 PR rtl-optimization/57559
19617 * lra-constraints.c (process_alt_operands): Don't discourage
19618 memory with known offset for offsetable memory constraint.
19619 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
19621 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
19623 * varasm.c (struct oc_local_state): Reorder fields.
19624 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
19625 and adjust accordingly.
19626 (output_constructor): Reorder initialization code and adjust call to
19627 output_constructor_bitfield.
19629 2013-06-07 Jan Hubicka <jh@suse.cz>
19631 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
19633 2013-06-07 David Malcolm <dmalcolm@redhat.com>
19635 * tree-object-size.c (unknown): Make const.
19637 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19639 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
19640 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
19641 for last alternative in the cpu_facility attribute.
19643 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19646 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
19647 (xordi3): Change operand 2 constraint to arm_xordi_operand.
19648 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
19649 * config/arm/constraints.md (Dg): New constraint.
19650 * config/arm/neon.md (xordi3_neon): Remove.
19651 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
19652 * config/arm/predicates.md (arm_xordi_operand): New predicate.
19654 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19656 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
19657 Clean up alternatives.
19659 2013-06-07 Alan Modra <amodra@gmail.com>
19661 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
19664 2013-06-07 Alan Modra <amodra@gmail.com>
19666 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
19668 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19670 * config/arm/constraints.md (Df): New constraint.
19671 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
19672 Correct length attribute for last two alternatives.
19674 2013-06-07 Alan Modra <amodra@gmail.com>
19676 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
19677 override user -mfp-in-toc.
19678 (offsettable_ok_by_alignment): Consider just the current access
19679 rather than the whole object, unless BLKmode. Handle
19680 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
19681 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
19682 for -mcmodel=medium.
19683 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
19684 override user -mfp-in-toc or -msum-in-toc. Default to
19685 -mno-fp-in-toc for -mcmodel=medium.
19687 2013-06-06 DJ Delorie <dj@redhat.com>
19689 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
19690 TARGET_VALID_POINTER_MODE.
19692 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
19693 Pat Haugen <pthaugen@us.ibm.com>
19694 Peter Bergner <bergner@vnet.ibm.com>
19696 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
19697 Document new power8 builtins.
19699 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
19700 condition code register, to allow 128-bit logical operations to be
19701 done in the VSX or GPR registers.
19702 (nor<mode>3): Use the canonical form for nor.
19703 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
19704 vclz*, and vpopcnt* vector instructions.
19705 (nand<mode>3): Likewise.
19706 (orc<mode>3): Likewise.
19707 (clz<mode>2): LIkewise.
19708 (popcount<mode>2): Likewise.
19710 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
19711 that only the GPRs are recognized.
19713 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
19714 support for new power8 builtins.
19716 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
19718 (xscvdpspn): Likewise.
19724 (vpopcnt): Likewise.
19725 (vpopcntb): Likewise.
19726 (vpopcnth): Likewise.
19727 (vpopcntw): Likewise.
19728 (vpopcntd): Likewise.
19730 (vmrgew): Likewise.
19731 (vmrgow): Likewise.
19733 (eqv_v16qi3): Likewise.
19734 (eqv_v8hi3): Likewise.
19735 (eqv_v4si3): Likewise.
19736 (eqv_v2di3): Likewise.
19737 (eqv_v4sf3): Likewise.
19738 (eqv_v2df3): Likewise.
19740 (nand_v16qi3): Likewise.
19741 (nand_v8hi3): Likewise.
19742 (nand_v4si3): Likewise.
19743 (nand_v2di3): Likewise.
19744 (nand_v4sf3): Likewise.
19745 (nand_v2df3): Likewise.
19747 (orc_v16qi3): Likewise.
19748 (orc_v8hi3): Likewise.
19749 (orc_v4si3): Likewise.
19750 (orc_v2di3): Likewise.
19751 (orc_v4sf3): Likewise.
19752 (orc_v2df3): Likewise.
19754 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
19755 allow power8 quad mode in 64-bit.
19756 (rs6000_builtin_vectorized_function): Add support to vectorize
19757 ISA 2.07 count leading zeros, population count builtins.
19758 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
19759 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
19760 (builtin_function_type): Add vgbbd builtin function which takes an
19762 (altivec_expand_vec_perm_const): Add support for new power8 merge
19765 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
19766 that does not include TImdoe for use with 32-bit.
19767 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
19769 (UNSPEC_VSX_CVDPSPN): Likewise.
19770 (vsx_xscvdpspn): Likewise.
19771 (vsx_xscvspdpn): Likewise.
19772 (vsx_xscvdpspn_scalar): Likewise.
19773 (vsx_xscvspdpn_directmove): Likewise.
19774 (vsx_and<mode>3): Split logical operations into 32-bit and
19775 64-bit. Add support to do logical operations on TImode as well as
19776 VSX vector types. Allow logical operations to be done in either
19777 VSX registers or in general purpose registers in 64-bit mode. Add
19778 splitters if GPRs were used. For AND, add clobber of CCmode to
19779 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
19781 (vsx_and<mode>3_32bit): Likewise.
19782 (vsx_and<mode>3_64bit): Likewise.
19783 (vsx_ior<mode>3): Likewise.
19784 (vsx_ior<mode>3_32bit): Likewise.
19785 (vsx_ior<mode>3_64bit): Likewise.
19786 (vsx_xor<mode>3): Likewise.
19787 (vsx_xor<mode>3_32bit): Likewise.
19788 (vsx_xor<mode>3_64bit): Likewise.
19789 (vsx_one_cmpl<mode>2): Likewise.
19790 (vsx_one_cmpl<mode>2_32bit): Likewise.
19791 (vsx_one_cmpl<mode>2_64bit): Likewise.
19792 (vsx_nor<mode>3): Likewise.
19793 (vsx_nor<mode>3_32bit): Likewise.
19794 (vsx_nor<mode>3_64bit): Likewise.
19795 (vsx_andc<mode>3): Likewise.
19796 (vsx_andc<mode>3_32bit): Likewise.
19797 (vsx_andc<mode>3_64bit): Likewise.
19798 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
19799 and xxlorc instructions.
19800 (vsx_eqv<mode>3_64bit): Likewise.
19801 (vsx_nand<mode>3_32bit): Likewise.
19802 (vsx_nand<mode>3_64bit): Likewise.
19803 (vsx_orc<mode>3_32bit): Likewise.
19804 (vsx_orc<mode>3_64bit): Likewise.
19806 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
19808 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
19810 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
19811 (p8_vmrgow): Likewise.
19812 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
19813 GPRs to be split under VSX.
19814 (p8v_clz<mode>2): Add power8 count leading zero support.
19815 (p8v_popcount<mode>2): Add power8 population count support.
19816 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
19819 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
19822 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
19824 (vec_nand): Likewise.
19825 (vec_vclz): Likewise.
19826 (vec_vclzb): Likewise.
19827 (vec_vclzd): Likewise.
19828 (vec_vclzh): Likewise.
19829 (vec_vclzw): Likewise.
19830 (vec_vgbbd): Likewise.
19831 (vec_vmrgew): Likewise.
19832 (vec_vmrgow): Likewise.
19833 (vec_vpopcnt): Likewise.
19834 (vec_vpopcntb): Likewise.
19835 (vec_vpopcntd): Likewise.
19836 (vec_vpopcnth): Likewise.
19837 (vec_vpopcntw): Likewise.
19839 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
19841 PR rtl-optimization/57468
19842 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
19845 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
19847 PR rtl-optimization/57459
19848 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
19849 type when setting live regs.
19851 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
19853 * config/s390/s390.opt (mlra): New option.
19854 * config/s390/s390.c (s390_decompose_address): Check displacement
19855 for all registers for LRA.
19856 (s390_secondary_reload): Don't used secondary reloads for LRA.
19857 (s390_lra_p): New function.
19858 (TARGET_LRA_P): Define.
19859 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
19860 of attribute cpu_facility to zarch for the last alternative.
19861 (*cmpmem_short): Ditto.
19863 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
19865 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
19866 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
19867 (arm_expand_prologue): Likewise.
19869 2013-06-06 Teresa Johnson <tejohnson@google.com>
19872 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
19873 as this is now done by redirect_edge_and_branch_force.
19874 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
19875 barriers, and fix interaction with splitting.
19876 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
19877 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
19878 reflect changes made in the routine.
19879 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
19880 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
19881 since this is called in cfglayout mode, and replace partition fixup
19882 with assert as that is now done by force_nonfallthru_and_redirect.
19883 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
19884 already be marked with region crossing note.
19885 (insert_section_boundary_note): Make non-static, gate on flag
19886 has_bb_partition, rewrite to also check for multiple partitions.
19887 (rest_of_handle_reorder_blocks): Remove call to
19888 insert_section_boundary_note, now done later during free_cfg.
19889 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
19890 * bb-reorder.h (insert_section_boundary_note): Declare.
19891 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
19892 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
19893 invoke insert_section_boundary_note.
19894 (try_redirect_by_replacing_jump): Remove unnecessary
19895 check for region crossing note.
19896 (fixup_partition_crossing): New function.
19897 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
19898 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
19899 in non-cfglayout mode.
19900 (force_nonfallthru_and_redirect): Fixup partition boundaries,
19901 remove old code that tried to do this. Emit barrier correctly
19902 when we are in cfglayout mode.
19903 (last_bb_in_partition): New function.
19904 (rtl_split_edge): Correctly fixup partition boundaries.
19905 (commit_one_edge_insertion): Remove old code that tried to
19906 fixup region crossing edge since this is now handled in
19907 split_block, and set up insertion point correctly since
19908 block may now end in a jump.
19909 (verify_hot_cold_block_grouping): Guard against checking when not in
19910 linearized RTL mode.
19911 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
19913 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
19914 rtl_verify_flow_info, so not called in cfglayout mode.
19915 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
19916 (fixup_reorder_chain): Remove old code that attempted to fixup region
19917 crossing note as this is now handled in force_nonfallthru_and_redirect.
19918 (duplicate_insn_chain): Don't duplicate switch section notes.
19919 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
19921 * basic-block.h (emit_barrier_after_bb): Declare.
19923 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19925 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
19926 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
19927 arm_usatsihi): Adjust alternatives for arm_restrict_it.
19929 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19931 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
19933 * config/arm/ldmstm.md: Regenerate.
19935 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19937 * config/arm/sync.md (atomic_loaddi_1):
19938 Disable predication for arm_restrict_it.
19939 (arm_load_exclusive<mode>): Likewise.
19940 (arm_load_exclusivesi): Likewise.
19941 (arm_load_exclusivedi): Likewise.
19942 (arm_load_acquire_exclusive<mode>): Likewise.
19943 (arm_load_acquire_exclusivesi): Likewise.
19944 (arm_load_acquire_exclusivedi): Likewise.
19945 (arm_store_exclusive<mode>): Likewise.
19946 (arm_store_exclusive<mode>): Likewise.
19947 (arm_store_release_exclusivedi): Likewise.
19948 (arm_store_release_exclusive<mode>): Likewise.
19950 2013-06-06 Richard Biener <rguenther@suse.de>
19952 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
19954 (lto_tag_is_tree_code_p): Adjust.
19955 (lto_tag_is_gimple_code_p): Likewise.
19956 (lto_gimple_code_to_tag): Likewise.
19957 (lto_tag_to_gimple_code): Likewise.
19958 (lto_tree_code_to_tag): Likewise.
19959 (lto_tag_to_tree_code): Likewise.
19960 * data-streamer.h (streamer_write_hwi_in_range): Use
19961 uhwi streaming to stream the normalized range.
19962 (streamer_read_hwi_in_range): Likewise.
19964 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19966 * config/arm/arm.md (enabled_for_depr_it): New attribute.
19967 (predicable_short_it): Likewise.
19968 (predicated): Likewise.
19969 (enabled): Handle above.
19970 (define_cond_exec): Set predicated attribute to yes.
19972 2013-06-05 Mike Stump <mikestump@comcast.net>
19974 * gdbinit.in (__FUNCTION__): Add.
19976 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
19978 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
19979 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
19981 2013-06-05 Jan Hubicka <jh@suse.cz>
19983 * varasm.c (mark_decl_referenced): Revert the removal until targets
19986 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
19988 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
19989 instead of mark_decl_referenced.
19991 2013-06-05 Jan Hubicka <jh@suse.cz>
19993 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
19994 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
19995 and symtab_used_from_object_file_p.
19996 (cgraph_make_node_local_1): Clear forced_by_abi.
19997 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
19998 * cgraph.h (symtab_node_base): Add forced_by_abi.
19999 (decide_is_variable_needed): Remove.
20000 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
20001 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
20002 (decide_is_symbol_needed): ... this one; handle symbols in general;
20003 always analyze virtuals; honnor forced_by_abi.
20004 (cgraph_finalize_function): Update.
20005 (varpool_finalize_decl): Update.
20006 (symbol_defined_and_needed): Remove.
20007 (analyze_functions): Update.
20008 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
20009 output_refs, input_overwrite_node): Handle forced_by_abi.
20010 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
20011 (address_taken_from_non_vtable_p): ... this one.
20012 (comdat_can_be_unshared_p_1): New function.
20013 (cgraph_comdat_can_be_unshared_p): Rename to ...
20014 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
20015 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
20016 (function_and_variable_visibility): Clear forced_by_abi as needed.
20017 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
20018 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
20019 * symtab.c (dump_symtab_base): Dump forced_by_abi.
20020 * varpool.c (decide_is_variable_needed): Remove.
20022 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20024 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
20025 (arm_option_override): Override arm_restrict_it where appropriate.
20026 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
20027 * config/arm/arm.opt (mrestrict-it): New command-line option.
20028 * doc/invoke.texi: Document -mrestrict-it.
20030 2013-06-05 David Malcolm <dmalcolm@redhat.com>
20032 * tsan.c (tsan_atomic_table): Make const.
20034 2013-06-05 Richard Biener <rguenther@suse.de>
20036 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
20037 index associated with the tree we are supposed to replace.
20038 * tree-streamer-out.c (pack_ts_base_value_fields): Output
20039 TREE_ASM_WRITTEN as zero for everything but SSA names.
20041 2013-06-05 David Malcolm <dmalcolm@redhat.com>
20043 * tree-ssa-structalias.c (call_stmt_vars): Make static.
20045 2013-06-04 Jan Hubicka <jh@suse.cz>
20047 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
20048 (input_node, input_varpool_node): Handle correctly external same
20050 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
20051 nodes at ltrans stage.
20053 2013-06-04 Jan Hubicka <jh@suse.cz>
20055 * ipa-inline.c (update_caller_keys): Fix availability test.
20056 (update_callee_keys): Likewise.
20057 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
20058 to follow ELF standard.
20060 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
20062 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
20063 (mips64r5900el-*-elf*): New configurations.
20064 * config/mips/mips-cpus.def (r5900): New processor.
20065 * config/mips/mips-tables.opt: Regenerate.
20066 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
20067 (mips_issue_rate): Handle PROCESSOR_R5900.
20068 (mips_reorg_process_insns): Force reorder mode for the R5900.
20069 * config/mips/mips.h (TARGET_MIPS5900): Define.
20070 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
20072 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
20074 * config/mips/mips.md (processor): Add r5900.
20075 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
20077 2013-06-04 Ian Bolton <ian.bolton@arm.com>
20079 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
20080 into function to generate MOVI instruction.
20081 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
20082 (aarch64_preferred_simd_mode): Turn into wrapper.
20083 (aarch64_output_scalar_simd_mov_immediate): New function.
20084 * config/aarch64/aarch64-protos.h: Add prototype for above.
20086 2013-06-04 Ian Bolton <ian.bolton@arm.com>
20088 * config/aarch64/aarch64.c (simd_immediate_info): Remove
20089 element_char member.
20090 (sizetochar): Return signed char.
20091 (aarch64_simd_valid_immediate): Remove elchar and other
20092 unnecessary variables.
20093 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
20094 Calculate element_char as required.
20095 * config/aarch64/aarch64-protos.h: Update and move prototype
20096 for aarch64_output_simd_mov_immediate.
20097 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
20100 2013-06-04 Ian Bolton <ian.bolton@arm.com>
20102 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
20103 information completed by aarch64_simd_valid_immediate.
20104 (aarch64_legitimate_constant_p): Update arguments.
20105 (aarch64_simd_valid_immediate): Work with struct rather than many
20107 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
20108 (aarch64_simd_make_constant): Update arguments.
20109 (aarch64_output_simd_mov_immediate): Work with struct rather than
20110 many pointers. Output immediate directly rather than as operand.
20111 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
20113 * config/aarch64/constraints.md (Dn): Update arguments.
20115 2013-06-04 Ian Bolton <ian.bolton@arm.com>
20117 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
20119 (aarch64_simd_immediate_valid_for_move): Remove.
20120 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
20121 (aarch64_simd_make_constant): Update call.
20122 (aarch64_output_simd_mov_immediate): Update call.
20123 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
20125 * config/aarch64/constraints.md (Dn): Update call.
20127 2013-06-04 Ian Bolton <ian.bolton@arm.com>
20129 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
20130 return type to bool for prototype.
20131 (aarch64_legitimate_constant_p): Check for true instead of not -1.
20132 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
20133 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
20135 2013-06-04 Catherine Moore <clm@codesourcery.com>
20137 * config/mips/mips.opt (meva): New.
20138 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
20139 (ASM_SPEC): Handle -meva.
20140 * doc/invoke.texi (meva): Document.
20142 2013-06-04 Alan Modra <amodra@gmail.com>
20144 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
20147 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20149 * rtl.def: Add extra fourth optional field to define_cond_exec.
20150 * gensupport.c (process_one_cond_exec): Process attributes from
20152 * doc/md.texi: Document fourth field in define_cond_exec.
20154 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
20156 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
20157 out the processing order as in store_bit_field_1.
20159 2013-06-04 Jan Hubicka <jh@suse.cz>
20161 PR middle-end/57500
20162 * cgraphunit.c (cgraph_process_same_body_aliases): Create
20163 non-VAR_DECL node if it does not exist yet.
20165 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
20167 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
20168 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
20169 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
20170 target_cpu_default setting.
20172 2013-06-03 Teresa Johnson <tejohnson@google.com>
20174 * dumpfile.c (opt_info_switch_p): Change -fopt-info
20175 default to -fopt-info=optimized instead of all.
20176 * doc/invoke.texi: Ditto.
20177 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
20178 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
20179 (execute_vect_slp): Emit BB vectorization success under
20180 MSG_OPTIMIZED_LOCATIONS.
20181 * tree-vect-slp.c (vect_slp_transform_bb): Change
20182 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
20183 * tree-vect-loop.c (vect_transform_loop): Ditto.
20185 2013-06-03 Jason Merrill <jason@redhat.com>
20188 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
20189 Use TARGET_EXPR for C++.
20191 2013-06-03 Jakub Jelinek <jakub@redhat.com>
20193 PR rtl-optimization/57268
20194 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
20195 if DEBUG_INSN_P (insn).
20198 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
20200 PR rtl-optimization/57268
20201 * sched-deps.c (sched_analyze_2): Flush dependence lists if
20202 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
20204 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
20206 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
20207 (ix86_avoid_lea_for_addr): Likewise.
20208 (exact_dependency_1): Likewise.
20209 (ix86_adjust_cost): Likewise.
20210 (swap_top_of_ready_list): Fix formatting and !reload_completed check
20212 (do_reorder_for_imul): Fix typo, formatting and
20213 !reload_completed check removed.
20214 (ix86_sched_reorder): Fix typo and formatting.
20215 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
20218 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
20220 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
20222 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
20224 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
20225 <STRING_CST>: Likewise.
20226 <VECTOR_CST>: Likewise.
20228 2013-06-01 Janus Weil <janus@gcc.gnu.org>
20229 Mikael Morin <mikael@gcc.gnu.org>
20231 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
20232 * config.in: Regenerated.
20233 * configure: Regenerated.
20235 2013-06-01 Jan Hubicka <jh@suse.cz>
20237 PR middle-end/57366
20238 * cgraphunit.c (compile): When weakref is not supported,
20239 set up transparent aliases before final output pass.
20240 * varasm.c (assemble_alias): Do not try to do it here.
20242 2013-06-01 Jan Hubicka <jh@suse.cz>
20244 PR middle-end/57467
20245 * passes.c (for_per_function): Skip unanalyzed functions.
20247 2013-06-01 Jan Hubicka <jh@suse.cz>
20249 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
20250 (lto_symtab_merge_symbols_1): ... this one.
20251 (lto_symtab_merge_cgraph_nodes): Rename to ...
20252 (lto_symtab_merge_symbols): ... this one; simplify.
20253 * cgraph.c (same_body_aliases_done): Rename to ...
20254 (cpp_implicit_aliases_done): ... this one.
20255 (cgraph_create_function_alias): Update.
20256 (cgraph_same_body_alias): Update.
20257 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
20258 (verify_edge_corresponds_to_fndecl): Simplify.
20259 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
20260 (cgraph_node): Remove same_body_alias.
20261 (varpool_node): Remove alias_of and extra_name_alias.
20262 (same_body_aliases_done): Rename to ..
20263 (cpp_implicit_aliases_done): ... this one.
20264 (symtab_alias_ultimate_target): Add default parameter.
20265 (symtab_resolve_alias): New function.
20266 (fixup_same_cpp_alias_visibility): Declare.
20267 (cgraph_function_node): Add default parameter.
20268 (cgraph_node_asm_name): Likewise.
20269 (cgraph_function_or_thunk_node): Add default parameter; do
20270 not ICE when it is NULL.
20271 (varpool_variable_node): Likewise.
20272 * tree-emutls.c (create_emultls_var): Update.
20273 (ipa_lower_emutls): Update.
20274 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
20275 (cgraph_reset_node): Reset alias info.
20276 (cgraph_finalize_function): Update.
20277 (fixup_same_cpp_alias_visibility): Move to symtab.c.
20278 (analyze_function): Simplify.
20279 (cgraph_process_same_body_aliases): Simplify.
20280 (analyze_functions): Fixup same body aliases.
20281 (handle_alias_pairs): Simplify.
20282 (assemble_thunk): Update.
20283 (assemble_thunks_and_aliases): Update.
20284 (output_weakrefs): Rewrite.
20285 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
20286 (lto_output_varpool_node): Likewise.
20287 (compute_ltrans_boundary): Remve assert.
20288 (get_alias_symbol): New functoin.
20289 (input_node): Rewrite alias handling.
20290 (input_varpool_node): Likewise.
20291 * ipa-pure-const.c (propagate_pure_const): Fix formating.
20292 * ipa.c (process_references): Handle weakrefs correctly.
20293 (symtab_remove_unreachable_nodes): Likewise.
20294 * trans-mem.c (get_cg_data): Update.
20295 (ipa_tm_create_version_alias): Update.
20296 (ipa_tm_execute): Update.
20297 * symtab.c (dump_symtab_base): Dump aliases.
20298 (verify_symtab_base): Verify aliases.
20299 (symtab_node_availability): New function.
20300 (symtab_alias_ultimate_target): Simplify.
20301 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
20302 handle all the fixup cases.
20303 (symtab_resolve_alias): New function.
20304 * passes.c (ipa_write_summaries): Handle weakrefs.
20305 * varpool.c (varpool_analyze_node): Simplify.
20306 (assemble_aliases): Update.
20307 (varpool_create_variable_alias): Simplify.
20308 (varpool_extra_name_alias): Simplify.
20309 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
20310 (lto_symtab_merge_symbols): ... this one.
20312 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
20315 PR rtl-optimization/57268
20316 * sched-deps.c (sched_analyze_2): Flush dependence lists if
20317 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
20319 2013-06-01 Tobias Burnus <burnus@net-b.de>
20321 Partially reverted:
20322 2013-05-31 Tobias Burnus <burnus@net-b.de>
20324 PR middle-end/57073
20325 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
20328 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
20330 PR rtl-optimization/57268
20331 * sched-deps.c (sched_analyze_2): Flush dependence lists if
20332 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
20334 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
20336 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
20337 unordered comparison operators when -fno-trapping-math is in effect
20339 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
20340 and implement unordered comparison operators properly on the e500.
20342 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
20344 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
20345 for constant scalar integers.
20346 (simplify_relational_operation_1): Likewise.
20348 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
20350 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
20351 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
20354 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
20355 Igor Zamyatin <igor.zamyatin@intel.com>
20357 Silvermont (SLM) architecture performance tuning.
20358 * config/i386/i386.h (enum ix86_tune_indices): Add
20359 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
20360 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
20362 * config/i386/i386.c (initial_ix86_tune_features)
20363 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
20364 (ix86_lea_outperforms): Handle Silvermont tuning.
20365 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
20367 (ix86_use_lea_for_mov): Likewise.
20368 (ix86_avoid_lea_for_addr): Likewise.
20369 (ix86_lea_for_add_ok): Likewise.
20370 (exact_dependency_1): New function.
20371 (exact_store_load_dependency): Likewise.
20372 (ix86_adjust_cost): Handle Silvermont tuning.
20373 (do_reoder_for_imul): Likewise.
20374 (swap_top_of_ready_list): New function.
20375 (ix86_sched_reorder): Changed to handle Silvermont tuning.
20377 * config/i386/i386.md (peepholes that split memory operand in fp
20380 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
20382 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
20383 Remove un-necessary braces.
20385 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
20387 * config/aarch64/aarch64.c (aarch64_classify_symbol):
20388 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
20390 2013-05-31 Tobias Burnus <burnus@net-b.de>
20392 PR middle-end/57073
20393 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
20395 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20398 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
20399 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
20400 * config/arm/neon.md (iordi3_neon): Remove.
20401 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
20402 * config/arm/predicates.md (imm_for_neon_logic_operand):
20403 Move to earlier in the file.
20404 (neon_logic_op2): Likewise.
20405 (arm_iordi_operand_neon): New predicate.
20407 2013-05-31 Richard Biener <rguenther@suse.de>
20409 PR tree-optimization/57478
20410 PR tree-optimization/57453
20411 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
20414 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
20416 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
20417 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
20419 2013-05-30 Tobias Burnus <burnus@net-b.de>
20420 Thomas Koenig <tkoenig@gcc.gnu.org>
20422 PR middle-end/57073
20423 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
20424 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
20426 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
20428 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
20430 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
20432 * target.def (register_usage_leveling_p): New hook.
20433 * targhooks.c (default_register_usage_leveling_p): New.
20434 * targhooks.h (default_register_usage_leveling_p): New prototype.
20435 * lra-assigns.c (register_usage_leveling_p): Use the hook.
20436 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
20437 * doc/tm.texi: Update.
20438 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
20440 2013-05-30 Ian Bolton <ian.bolton@arm.com>
20442 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
20443 (*insv_reg<mode>): New define_insn.
20445 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
20447 PR rtl-optimization/57439
20448 * postreload.c (move2add_valid_value_p): Check that we have
20449 a zero subreg_regno_offset when accessing the register in
20450 the requested mode.
20452 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
20453 Igor Zamyatin <igor.zamyatin@intel.com>
20455 Silvermont (SLM) architecture pipeline model, tuning and
20457 * config.gcc: Add slm config options and target.
20459 * config/i386/slm.md: New.
20461 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
20463 * config/i386/i386-c.c (ix86_target_macros_internal): New case
20465 (ix86_target_macros_internal): Likewise.
20467 * config/i386/i386.c (slm_cost): New cost.
20468 (m_SLM): New macro flag.
20469 (initial_ix86_tune_features): Set m_SLM.
20470 (x86_accumulate_outgoing_args): Likewise.
20471 (x86_arch_always_fancy_math_387): Likewise.
20472 (processor_target_table): Add slm cost.
20473 (cpu_names): Add slm cpu name.
20474 (x86_option_override_internal): Set SLM ISA.
20475 (ix86_issue_rate): New case PROCESSOR_SLM.
20476 (ia32_multipass_dfa_lookahead): Likewise.
20477 (fold_builtin_cpu): Add slm.
20479 * config/i386/i386.h (TARGET_SLM): New target macro.
20480 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
20481 (processor_type): Add PROCESSOR_SLM.
20483 * config/i386/i386.md (cpu): Add new value "slm".
20484 (slm.md): Include slm.md.
20486 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
20487 Zhenqiang Chen <zhenqiang.chen@linaro.org>
20489 * config/arm/arm-protos.h: Add and update function protos.
20490 * config/arm/arm.c (use_simple_return_p): New added.
20491 (thumb2_expand_return): Check simple_return flag.
20492 * config/arm/arm.md: Add simple_return and conditional simple_return.
20493 * config/arm/iterators.md: Add iterator for return and simple_return.
20495 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
20497 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
20498 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
20499 (arm_emit_vfp_multi_reg_pop): Likewise.
20500 (thumb2_emit_ldrd_pop): Likewise.
20501 (arm_expand_epilogue): Add misc REG_CFA notes.
20502 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
20504 2013-05-29 Lawrence Crowl <crowl@google.com>
20506 * config/arm/t-arm: Update for below.
20508 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
20509 Change type to hash_table. Update dependent calls and types.
20511 * config/i386/t-cygming: Update for below.
20513 * config/i386/t-interix: Update for below.
20515 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
20516 Change type to hash_table. Update dependent calls and types.
20517 (i386_find_on_wrapper_list::wrappers): Likewise.
20519 * config/ia64/t-ia64: Update for below.
20521 * config/ia64/ia64.c (bundle_state_table):
20522 Change type to hash_table. Update dependent calls and types.
20524 * config/mips/mips.c (mips_reorg_process_insns::htab):
20525 Change type to hash_table. Update dependent calls and types.
20527 * config/sol2.c (solaris_comdat_htab):
20528 Change type to hash_table. Update dependent calls and types.
20530 * config/t-sol2: Update for above.
20532 2013-05-29 Teresa Johnson <tejohnson@google.com>
20534 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
20535 functions are not yet marked as defined.
20537 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
20538 Pat Haugen <pthaugen@us.ibm.com>
20539 Peter Bergner <bergner@vnet.ibm.com>
20541 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
20545 (vrotl<mode>3): Likewise.
20546 (vashl<mode>3): Likewise.
20547 (vlshr<mode>3): Likewise.
20548 (vashr<mode>3): Likewise.
20550 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
20551 support for power8 V2DI builtins.
20553 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
20554 power8 V2DI builtins.
20555 (vupkhsw): Likewise.
20556 (vupklsw): Likewise.
20557 (vaddudm): Likewise.
20558 (vminsd): Likewise.
20559 (vmaxsd): Likewise.
20560 (vminud): Likewise.
20561 (vmaxud): Likewise.
20562 (vpkudum): Likewise.
20563 (vpksdss): Likewise.
20564 (vpkudus): Likewise.
20565 (vpksdus): Likewise.
20570 (vsubudm): Likewise.
20571 (vcmpequd): Likewise.
20572 (vcmpgtsd): Likewise.
20573 (vcmpgtud): Likewise.
20574 (vcmpequd_p): Likewise.
20575 (vcmpgtsd_p): Likewise.
20576 (vcmpgtud_p): Likewise.
20577 (vupkhsw): Likewise.
20578 (vupklsw): Likewise.
20579 (vaddudm): Likewise.
20580 (vmaxsd): Likewise.
20581 (vmaxud): Likewise.
20582 (vminsd): Likewise.
20583 (vminud): Likewise.
20584 (vpksdss): Likewise.
20585 (vpksdus): Likewise.
20586 (vpkudum): Likewise.
20587 (vpkudus): Likewise.
20592 (vsubudm): Likewise.
20594 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
20595 support for power8 V2DI instructions.
20597 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
20598 power8 V2DI instructions. Combine pack and unpack insns to use an
20599 iterator for each mode. Check whether a particular mode supports
20600 Altivec instructions instead of just checking TARGET_ALTIVEC.
20601 (UNSPEC_VPKUWUM): Likewise.
20602 (UNSPEC_VPKSHSS): Likewise.
20603 (UNSPEC_VPKSWSS): Likewise.
20604 (UNSPEC_VPKUHUS): Likewise.
20605 (UNSPEC_VPKSHUS): Likewise.
20606 (UNSPEC_VPKUWUS): Likewise.
20607 (UNSPEC_VPKSWUS): Likewise.
20608 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
20609 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
20610 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
20611 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
20612 (UNSPEC_VUPKHSB): Likewise.
20613 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
20614 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
20615 (UNSPEC_VUPKHSH): Likewise.
20616 (UNSPEC_VUPKLSB): Likewise.
20617 (UNSPEC_VUPKLSH): Likewise.
20619 (VI_char): Likewise.
20620 (VI_scalar): Likewise.
20621 (VI_unit): Likewise.
20623 (VP_small): Likewise.
20624 (VP_small_lc): Likewise.
20625 (VU_char): Likewise.
20626 (add<mode>3): Likewise.
20627 (altivec_vaddcuw): Likewise.
20628 (altivec_vaddu<VI_char>s): Likewise.
20629 (altivec_vadds<VI_char>s): Likewise.
20630 (sub<mode>3): Likewise.
20631 (altivec_vsubcuw): Likewise.
20632 (altivec_vsubu<VI_char>s): Likewise.
20633 (altivec_vsubs<VI_char>s): Likewise.
20634 (altivec_vavgs<VI_char>): Likewise.
20635 (altivec_vcmpbfp): Likewise.
20636 (altivec_eq<mode>): Likewise.
20637 (altivec_gt<mode>): Likewise.
20638 (altivec_gtu<mode>): Likewise.
20639 (umax<mode>3): Likewise.
20640 (smax<mode>3): Likewise.
20641 (umin<mode>3): Likewise.
20642 (smin<mode>3): Likewise.
20643 (altivec_vpkuhum): Likewise.
20644 (altivec_vpkuwum): Likewise.
20645 (altivec_vpkshss): Likewise.
20646 (altivec_vpkswss): Likewise.
20647 (altivec_vpkuhus): Likewise.
20648 (altivec_vpkshus): Likewise.
20649 (altivec_vpkuwus): Likewise.
20650 (altivec_vpkswus): Likewise.
20651 (altivec_vpks<VI_char>ss): Likewise.
20652 (altivec_vpks<VI_char>us): Likewise.
20653 (altivec_vpku<VI_char>us): Likewise.
20654 (altivec_vpku<VI_char>um): Likewise.
20655 (altivec_vrl<VI_char>): Likewise.
20656 (altivec_vsl<VI_char>): Likewise.
20657 (altivec_vsr<VI_char>): Likewise.
20658 (altivec_vsra<VI_char>): Likewise.
20659 (altivec_vsldoi_<mode>): Likewise.
20660 (altivec_vupkhsb): Likewise.
20661 (altivec_vupkhs<VU_char>): Likewise.
20662 (altivec_vupkls<VU_char>): Likewise.
20663 (altivec_vupkhsh): Likewise.
20664 (altivec_vupklsb): Likewise.
20665 (altivec_vupklsh): Likewise.
20666 (altivec_vcmpequ<VI_char>_p): Likewise.
20667 (altivec_vcmpgts<VI_char>_p): Likewise.
20668 (altivec_vcmpgtu<VI_char>_p): Likewise.
20669 (abs<mode>2): Likewise.
20670 (vec_unpacks_hi_v16qi): Likewise.
20671 (vec_unpacks_hi_v8hi): Likewise.
20672 (vec_unpacks_lo_v16qi): Likewise.
20673 (vec_unpacks_hi_<VP_small_lc>): Likewise.
20674 (vec_unpacks_lo_v8hi): Likewise.
20675 (vec_unpacks_lo_<VP_small_lc>): Likewise.
20676 (vec_pack_trunc_v8h): Likewise.
20677 (vec_pack_trunc_v4si): Likewise.
20678 (vec_pack_trunc_<mode>): Likewise.
20680 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
20682 (vec_vmaxsd): Likewise.
20683 (vec_vmaxud): Likewise.
20684 (vec_vminsd): Likewise.
20685 (vec_vminud): Likewise.
20686 (vec_vpksdss): Likewise.
20687 (vec_vpksdus): Likewise.
20688 (vec_vpkudum): Likewise.
20689 (vec_vpkudus): Likewise.
20690 (vec_vrld): Likewise.
20691 (vec_vsld): Likewise.
20692 (vec_vsrad): Likewise.
20693 (vec_vsrd): Likewise.
20694 (vec_vsubudm): Likewise.
20695 (vec_vupkhsw): Likewise.
20696 (vec_vupklsw): Likewise.
20698 2013-05-29 Jan Hubicka <jh@suse.cz>
20700 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
20701 flags; reorder rest of fields in more consistent way.
20702 (varpool_node): Remove analyzed, finalized and alias.
20703 (cgraph_ndoe): Likewise.
20704 (symtab_alias_ultimate_target): New function.
20705 (cgraph_function_node): Move offline.
20706 (cgraph_reset_node): Declare.
20707 (cgraph_comdat_can_be_unshared_p): Remove.
20708 (varpool_remove_initializer): Declare.
20709 (varpool_first_defined_variable, varpool_next_defined_variable
20710 cgraph_first_defined_function, cgraph_next_defined_function): Update.
20711 (cgraph_function_with_gimple_body_p): Update.
20712 (varpool_all_refs_explicit_p): Update.
20713 (symtab_alias_target): New function.
20714 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
20715 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
20716 (cgraph_function_or_thunk_node): Simplify using
20717 symtab_alias_ultimate_target.
20718 (varpool_variable_node): Likewise.
20719 * cgraph.c (cgraph_create_function_alias): Update.
20720 (cgraph_add_thunk): Update.
20721 (cgraph_remove_node): Update.
20722 (dump_cgraph_node): Do not dump removed flags.
20723 (cgraph_function_body_availability): Update.
20724 (cgraph_propagate_frequency): Update.
20725 (verify_cgraph_node): Check sanity of local flag.
20726 (cgraph_function_node): Move here from cgraph.h; revamp for
20727 cgraph_function_or_thunk_node.
20728 * lto-symtab.c (lto_varpool_replace_node): Update.
20729 (lto_symtab_resolve_can_prevail_p): Update.
20730 (lto_symtab_merge_cgraph_nodes): Update.
20731 * ipa-cp.c (determine_versionability, initialize_node_lattices,
20732 propagate_constants_accross_call, devirtualization_time_bonus,
20733 ipcp_propagate_stage): Update.
20734 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
20735 * ipa-inline-transform.c (clone_inlined_nodes,
20736 preserve_function_body_p): Update.
20737 * ipa-reference.c (propagate): Update.
20738 (write_node_summary_p): Update.
20739 * toplev.c (wrapup_global_declaration_2): Update.
20740 * cgraphunit.c (cgraph_analyze_function): Rename to ...
20741 (analyze_function) ... this one.
20742 (cgraph_process_new_functions): Update.
20743 (cgraph_reset_node): Export.
20744 (cgraph_finalize_function): Update.
20745 (cgraph_add_new_function): Update.
20746 (process_function_and_variable_attributes): Update.
20747 (varpool_finalize_decl): Update.
20748 (symbol_finalized): Remove.
20749 (symbol_finalized_and_needed): Rename to ...
20750 (symbol_defined_and_needed): ... update.
20751 (cgraph_analyze_functions): Update.
20752 (handle_alias_pairs): Update.
20753 (mark_functions_to_output): Update.
20754 (assemble_thunk): Update.
20755 (output_in_order): Update.
20756 (output_weakrefs): Update.
20757 (finalize_compilation_unit): Update.
20758 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
20759 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
20760 input_node, input_varpool_node): Update.
20761 * dbxout.c (dbxout_expand_expr): Update.
20762 * cgraphclones.c (cgraph_clone_node): Update.
20763 (cgraph_copy_node_for_versioning): Update.
20764 (cgraph_materialize_clone): Update.
20765 (cgraph_materialize_all_clones): Update.
20766 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
20767 propagate_pure_const, propagate_nothrow): Update.
20768 * lto-streamer-out.c (lto_output, write_symbol): Update.
20769 * ipa-utils.c (ipa_reverse_postorder): Update.
20770 * ipa-inline.c (can_inline_edge_p): Update.
20771 (update_caller_keys, ipa_inline): Update.
20772 * dwarf2out.c (reference_to_unused,
20773 premark_types_used_by_global_vars_helper): Update.
20774 * tree-eh.c (tree_could_trap_p): Update.
20775 * ipa-split.c (consider_split, execute_split_functions): Update.
20776 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
20777 has_addr_references_p): Update; move ahead in file for better
20779 (process_references): Simplify.
20780 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
20781 bodies are removed.
20782 (cgraph_comdat_can_be_unshared_p): Make static.
20783 (cgraph_externally_visible_p): Update.
20784 (varpool_externally_visible_p): Update.
20785 (function_and_variable_visibility): Update.
20786 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
20787 ipa_tm_mark_force_output_node): Update.
20788 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
20789 estimate_edge_devirt_benefit, inline_generate_summary,
20790 inline_write_summary): Update.
20791 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
20792 * ipa-prop.c (ipa_compute_jump_functions): Update.
20793 (ipa_print_node_params, ipa_prop_read_section,
20794 ipa_update_after_lto_read, read_replacements_section): Update.
20795 * varasm.c (mark_decl_referenced): Update.
20796 (assemble_alias, dump_tm_clone_pairs): Update.
20797 * tree-inline.c (copy_bb): Update.
20798 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
20800 * symtab.c (dump_symtab_base): Print new flags.
20801 (verify_symtab_base): Verify new flags.
20802 (symtab_alias_ultimate_target): New function.
20803 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
20804 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
20806 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
20808 * i386.c (ix86_get_function_versions_dispatcher,
20809 ix86_generate_version_dispatcher_body): Update.
20810 (fold_builtin_cpu): Use varpool_add_new_variable.
20811 * varpool.c (varpool_remove_initializer): Break out from ...
20812 (varpool_remove_node): ... this one.
20813 (dump_varpool_node, varpool_node_for_asm,
20814 cgraph_variable_initializer_availability, varpool_analyze_node,
20815 varpool_assemble_decl, varpool_remove_unreferenced_decls,
20816 varpool_finalize_named_section_flags, varpool_create_variable_alias):
20819 2013-05-29 Jan Hubicka <jh@suse.cz>
20821 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
20823 2013-05-29 Easwaran Raman <eraman@google.com>
20825 PR tree-optimization/57442
20826 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
20827 when control exits the main loop.
20829 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
20831 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
20833 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
20834 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
20835 * rx/t-rx: Add rx100 under multi library matches option for nofpu
20838 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20840 PR tree-optimization/57441
20841 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
20842 Don't limit size of incr_vec to number of candidates.
20844 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
20846 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
20847 and mips16 directories.
20848 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
20849 (MULTILIB_DIRNAMES): Ditto.
20850 (MULTILIB_EXCEPTIONS): Add new exceptions.
20851 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
20852 (MULTILIB_DIRNAMES): Ditto.
20853 (MULTILIB_EXCEPTIONS): Add new exceptions.
20855 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
20856 Marcus Shawcroft <marcus.shawcroft@arm.com>
20858 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
20859 SYMBOL_TINY_ABSOLUTE.
20860 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
20861 SYMBOL_TINY_ABSOLUTE.
20862 (aarch64_expand_mov_immediate): Likewise.
20863 (aarch64_classify_symbol): Likewise.
20864 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
20865 Permit SYMBOL_TINY_ABSOLUTE.
20866 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
20868 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
20869 Marcus Shawcroft <marcus.shawcroft@arm.com>
20871 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
20872 Refactor if/switch. Replace gcc_assert with if.
20874 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
20876 * config/i386/i386.c (initial_ix86_tune_features): Enable
20877 FP Reassociation for AMD bdver1 and bdver2.
20879 2013-05-29 Martin Jambor <mjambor@suse.cz>
20881 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
20882 and IMAGPART_EXPR do not occur within other handled_components.
20884 2013-05-29 Richard Biener <rguenther@suse.de>
20886 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
20887 access on whether the use is in the BB we currently try to
20889 (vect_bb_vectorization_profitable_p): Pass the BB we currently
20890 vectorize to vect_bb_slp_scalar_cost.
20892 2013-05-29 Richard Biener <rguenther@suse.de>
20894 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
20895 computing scalar cost offsetted by stmts that are kept live
20897 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
20898 for computation of scalar cost.
20900 2013-05-28 Steve Ellcey <sellcey@mips.com>
20902 * config/mips/mips-cpus.def (mips32r2): Change processor type.
20904 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
20906 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
20907 array notation built-in reduction functions.
20908 * doc/passes.texi (Passes): Added documentation about changes done
20910 * doc/invoke.texi (C Dialect Options): Added documentation about
20911 the -fcilkplus flag.
20912 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
20913 (BUILTINS_DEF): Depend on cilkplus.def.
20914 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
20915 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
20916 * cilkplus.def: New file.
20918 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
20920 PR rtl-optimization/57439
20921 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
20923 2013-05-28 Easwaran Raman <eraman@google.com>
20925 PR tree-optimization/57337
20926 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
20927 (find_insert_point): Correctly identify the insertion point
20928 when two statements with the same UID is compared.
20930 2013-05-28 Richard Biener <rguenther@suse.de>
20932 PR tree-optimization/56787
20933 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
20934 from the list of data references.
20935 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
20937 (vect_analyze_loop_operations): Likewise.
20938 (vect_transform_loop): Remove clobbers.
20940 2013-05-28 Martin Jambor <mjambor@suse.cz>
20942 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
20943 and REALPART_EXPRs have scalar type.
20945 2013-05-28 Richard Biener <rguenther@suse.de>
20947 PR tree-optimization/57411
20948 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
20950 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
20951 virtual operand propagation.
20953 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
20955 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
20956 destination register for bmasksi_vis.
20957 (vector_init_bshuffle): Likewise.
20958 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
20960 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
20962 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
20963 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
20964 mode if the instruction isn't available in the original mode.
20965 * config/sparc/sparc.opt (mfix-ut699): New option.
20966 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
20967 (divdf3): Turn into expander.
20968 (divdf3_nofix): New insn.
20969 (divdf3_fix): Likewise.
20970 (divsf3): Disable if -mfix-ut699.
20971 (sqrtdf2): Turn into expander.
20972 (sqrtdf2_nofix): New insn.
20973 (sqrtdf2_fix): Likewise.
20974 (sqrtsf2): Disable if -mfix-ut699.
20976 2013-05-27 Richard Biener <rguenther@suse.de>
20978 PR middle-end/57412
20979 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
20980 block for the new loop.
20982 2013-05-27 Richard Biener <rguenther@suse.de>
20984 PR tree-optimization/57343
20985 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
20986 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
20987 (number_of_iterations_cond): Do not build the folded tree.
20989 2013-05-27 Richard Biener <rguenther@suse.de>
20992 PR middle-end/57381
20993 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
20994 OEP_CONSTANT_ADDRESS_OF retained.
20996 PR tree-optimization/57417
20997 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
20998 for unchanged base.
20999 (set_ssa_val_to): Compare addresses using
21000 get_addr_base_and_unit_offset.
21002 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
21004 PR rtl-optimization/56833
21005 * postreload.c (move2add_record_mode): New function.
21006 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
21007 (move2add_use_add2_insn): Use move2add_record_sym_value.
21008 (move2add_use_add3_insn): Likewise.
21009 (reload_cse_move2add): Use move2add_valid_value_p and
21010 move2add_record_mode. Invalidate call-clobbered and REG_INC
21011 affected regs by setting reg_mode to VOIDmode.
21012 (move2add_note_store): Don't pretend the inside of a SUBREG is
21013 the actual destination. Invalidate single/leading registers by
21014 setting reg_mode to VOIDmode.
21015 Use move2add_record_sym_value, move2add_valid_value_p and
21016 move2add_record_mode.
21018 2013-05-27 Richard Biener <rguenther@suse.de>
21020 PR tree-optimization/57396
21021 * tree-affine.c (double_int_constant_multiple_p): Properly
21022 return false for val == 0 and div != 0.
21024 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
21026 * config/mips/mips.h: Use #elif in preprocessor conditions.
21028 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
21031 * config/mips/constraints.md (kl): New constraint.
21032 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
21033 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
21034 constraint for operand 0. Split after CSE for MIPS16. Emit a move
21035 from LO for MIPS16.
21036 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
21038 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
21041 * config/mips/mips.c (mips_can_inline_p): New function.
21042 (TARGET_CAN_INLINE_P): Define.
21044 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
21046 * sched-int.h (ds_t, dw_t): Make unsigned int.
21047 Fix documentation that describes how all the ds_t bits are used.
21048 Reserve the last bit for delayed-branch scheduling.
21049 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
21050 (BITS_PER_DEP_WEAK): Fix definition and documentation.
21051 (gen_dep_weak_1): Remove prototype.
21052 * sched-deps.c (get_dep_weak_1): Make static.
21053 * target.def (speculate_insn, needs_block_p, gen_spec_check,
21054 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
21055 * doc/tm.texi: Regenerate.
21056 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
21058 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
21061 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
21063 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
21064 Sandra Loosemore <sandra@codesourcery.com>
21066 * config.gcc (powerpc-*): Allow native for with-cpu.
21068 2013-05-24 Jeff Law <law@redhat.com>
21070 PR tree-optimization/57124
21071 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
21072 conversion feeding a condition if the range has an overflow
21073 if -fstrict-overflow. Add warnings for when we do make the
21076 2013-05-24 Dehao Chen <dehao@google.com>
21078 * tree-cfg.c (locus_discrim_map): Fix the typo.
21079 (locus_discrim_hasher): Likewise.
21080 (locus_discrim_hasher::hash): Likewise.
21081 (locus_discrim_hasher::equal): Likewise.
21083 2013-05-24 Martin Jambor <mjambor@suse.cz>
21085 PR tree-optimization/57294
21086 * cgraph.h (ipa_record_stmt_references): Declare.
21087 * cgraphbuild.c (ipa_record_stmt_references): New function.
21088 (build_cgraph_edges): Use ipa_record_stmt_references.
21089 (rebuild_cgraph_edges): Likewise.
21090 (cgraph_rebuild_references): Likewise.
21091 * ipa-prop.c (ipa_modify_call_arguments): Discard references
21092 associated with the old statement and build references from the
21093 newly built statements.
21094 * ipa-ref.c (ipa_remove_stmt_references): New function.
21095 * ipa-ref.h (ipa_remove_stmt_references): Declare.
21097 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
21099 * lra-constraints.c (emit_spill_move): Use smaller mode for
21101 (check_and_process_move): Consider mem-reg moves for secondary
21103 (curr_insn_transform): Don't lose insns emitted before for
21104 secondary memory moves.
21105 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
21106 reg set up in the current insn.
21108 2013-05-24 Dehao Chen <dehao@google.com>
21110 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
21112 (locus_descrim_hasher::equal): Likewise.
21113 (build_gimple_cfg): New discriminator assignment algorithm.
21114 (make_edges): Likewise.
21115 (next_discriminator_for_locus): Likewise.
21116 (same_line_p): Likewise.
21117 (assign_discriminators): Likewise.
21118 (make_cond_expr_edges): Likewise.
21119 (make_gimple_switch_edges): Likewise.
21120 (make_goto_expr_edges): Likewise.
21121 (make_gimple_asm_edges): Likewise.
21123 2013-05-24 Ian Bolton <ian.bolton@arm.com>
21125 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
21126 X format specifier to only display bottom 16 bits.
21127 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
21128 immediate to match for operand 2, since it will be masked.
21130 2013-05-24 Richard Biener <rguenther@suse.de>
21132 PR tree-optimization/57287
21133 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
21134 all SSA names that occur in abnormal PHIs.
21136 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
21139 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
21140 that index is not negative.
21142 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
21144 PR rtl-optimization/55177
21145 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
21146 (simplify_byte_swapping_operation): New.
21147 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
21148 (simplify_relational_operation_1): Deal with BSWAP.
21150 2013-05-23 Richard Henderson <rth@redhat.com>
21153 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
21154 (ix86_reorg): Call it.
21156 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
21159 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
21160 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
21161 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
21163 2013-05-23 Christian Bruel <christian.bruel@st.com>
21166 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
21168 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
21169 Marcus Shawcroft <marcus.shawcroft@arm.com>
21171 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
21172 * config/aarch64/constraints.md (Usa): Remove.
21173 * doc/md.texi (AArch64 Usa): Remove.
21175 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
21176 Marcus Shawcroft <marcus.shawcroft@arm.com>
21178 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
21179 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
21180 * config/aarch64/predicates.md (aarch64_const_address): Remove.
21181 (aarch64_mov_operand): Use aarch64_mov_operand_p.
21183 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
21185 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
21186 instruction (AdvSIMD).
21187 * config/aarch64/aarch64-builtins.c
21188 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
21189 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
21191 2013-05-23 Martin Jambor <mjambor@suse.cz>
21193 PR middle-end/57347
21194 * tree.h (contains_bitfld_component_ref_p): Declare.
21195 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
21196 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
21198 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
21199 not access a bit-field. Assert all final offsets are byte-aligned.
21201 2013-05-23 Richard Biener <rguenther@suse.de>
21203 PR tree-optimization/57380
21204 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
21205 least one invariant or re-used load.
21206 * passes.c (init_optimization_passes): Move pass_phiprop before
21209 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
21211 * config/aarch64/aarch64-simd.md
21212 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
21214 2013-05-23 Richard Biener <rguenther@suse.de>
21216 PR middle-end/57381
21217 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
21218 OEP_CONSTANT_ADDRESS_OF retained.
21220 2013-05-23 Jakub Jelinek <jakub@redhat.com>
21222 PR middle-end/57344
21223 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
21224 don't lower unit. Handle unit not being always BITS_PER_WORD.
21226 2013-05-23 Richard Biener <rguenther@suse.de>
21228 PR rtl-optimization/57341
21229 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
21230 instead of true_dependence.
21232 2013-05-22 David Malcolm <dmalcolm@redhat.com>
21234 * bb-reorder.c (branch_threshold): Make const.
21235 (exec_threshold): Ditto.
21237 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
21238 Pat Haugen <pthaugen@us.ibm.com>
21239 Peter Bergner <bergner@vnet.ibm.com>
21241 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
21242 documentation for the power8 crypto builtins.
21244 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
21246 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
21247 macros for defining power8 builtin functions.
21248 (BU_P8V_AV_2): Likewise.
21249 (BU_P8V_AV_P): Likewise.
21250 (BU_P8V_VSX_1): Likewise.
21251 (BU_P8V_OVERLOAD_1): Likewise.
21252 (BU_P8V_OVERLOAD_2): Likewise.
21253 (BU_CRYPTO_1): Likewise.
21254 (BU_CRYPTO_2): Likewise.
21255 (BU_CRYPTO_3): Likewise.
21256 (BU_CRYPTO_OVERLOAD_1): Likewise.
21257 (BU_CRYPTO_OVERLOAD_2): Likewise.
21258 (XSCVSPDP): Fix typo, point to the correct instruction.
21259 (VCIPHER): Add power8 crypto builtins.
21260 (VCIPHERLAST): Likewise.
21261 (VNCIPHER): Likewise.
21262 (VNCIPHERLAST): Likewise.
21263 (VPMSUMB): Likewise.
21264 (VPMSUMH): Likewise.
21265 (VPMSUMW): Likewise.
21266 (VPERMXOR_V2DI): Likewise.
21267 (VPERMXOR_V4SI: Likewise.
21268 (VPERMXOR_V8HI: Likewise.
21269 (VPERMXOR_V16QI: Likewise.
21270 (VSHASIGMAW): Likewise.
21271 (VSHASIGMAD): Likewise.
21272 (VPMSUM): Likewise.
21273 (VPERMXOR): Likewise.
21274 (VSHASIGMA): Likewise.
21276 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
21277 __CRYPTO__ if the crypto instructions are available.
21278 (altivec_overloaded_builtins): Add support for overloaded power8
21281 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
21282 support for power8 crypto builtins.
21283 (builtin_function_type): Likewise.
21284 (altivec_init_builtins): Add support for builtins that take vector
21285 long long (V2DI) arguments.
21287 * config/rs6000/crypto.md: New file, define power8 crypto
21290 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
21291 Pat Haugen <pthaugen@us.ibm.com>
21292 Peter Bergner <bergner@vnet.ibm.com>
21294 * doc/invoke.texi (Option Summary): Add power8 options.
21295 (RS/6000 and PowerPC Options): Likewise.
21297 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
21298 constraints.md instead of rs6000.h. Reorder w* constraints. Add
21299 wm, wn, wr documentation.
21301 * config/rs6000/constraints.md (wm): New constraint for VSX
21302 registers if direct move instructions are enabled.
21303 (wn): New constraint for no registers.
21304 (wq): New constraint for quad word even GPR registers.
21305 (wr): New constraint if 64-bit instructions are enabled.
21306 (wv): New constraint if power8 vector instructions are enabled.
21307 (wQ): New constraint for quad word memory locations.
21309 * config/rs6000/predicates.md (const_0_to_15_operand): New
21310 constraint for 0..15 for crypto instructions.
21311 (gpc_reg_operand): If VSX allow registers in VSX registers as well
21312 as GPR and floating point registers.
21313 (int_reg_operand): New predicate to match only GPR registers.
21314 (base_reg_operand): New predicate to match base registers.
21315 (quad_int_reg_operand): New predicate to match even GPR registers
21316 for quad memory operations.
21317 (vsx_reg_or_cint_operand): New predicate to allow vector logical
21318 operations in both GPR and VSX registers.
21319 (quad_memory_operand): New predicate for quad memory operations.
21320 (reg_or_indexed_operand): New predicate for direct move support.
21322 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
21323 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
21324 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
21325 (POWERPC_MASKS): Add power8 options.
21326 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
21329 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
21330 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
21332 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
21333 (-mpower8-fusion): New power8 options.
21334 (-mpower8-fusion-sign): Likewise.
21335 (-mpower8-vector): Likewise.
21336 (-mcrypto): Likewise.
21337 (-mdirect-move): Likewise.
21338 (-mquad-memory): Likewise.
21340 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
21341 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
21343 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
21344 (rs6000_debug_vector_unit): Add p8_vector.
21345 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
21346 definitions. Also print fusion state.
21347 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
21348 (rs6000_builtin_mask_calculate): Add power8 builtin support.
21349 (rs6000_option_override_internal): Add support for power8.
21350 (rs6000_common_init_builtins): Add debugging for skipped builtins
21351 if -mdebug=builtin.
21352 (rs6000_adjust_cost): Add power8 support.
21353 (rs6000_issue_rate): Likewise.
21354 (insn_must_be_first_in_group): Likewise.
21355 (insn_must_be_last_in_group): Likewise.
21356 (force_new_group): Likewise.
21357 (rs6000_register_move_cost): Likewise.
21358 (rs6000_opt_masks): Likewise.
21360 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
21361 power8 capable assembler, default to power7 options.
21362 (TARGET_DIRECT_MOVE): Likewise.
21363 (TARGET_CRYPTO): Likewise.
21364 (TARGET_P8_VECTOR): Likewise.
21365 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
21366 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
21367 (VECTOR_MEM_P8_VECTOR_P): Likewise.
21368 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
21369 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
21370 (TARGET_XSCVDPSPN): Likewise.
21371 (TARGET_XSCVSPDPN): Likewsie.
21372 (TARGET_SYNC_HI_QI): Likewise.
21373 (TARGET_SYNC_TI): Likewise.
21374 (MASK_CRYPTO): Likewise.
21375 (MASK_DIRECT_MOVE): Likewise.
21376 (MASK_P8_FUSION): Likewise.
21377 (MASK_P8_VECTOR): Likewise.
21378 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
21379 temporary used by some of the direct move instructions to get two FP
21380 temporary registers does not force creation of a stack frame.
21381 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
21382 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
21383 that any VSX registers are tieable, even if they are also an
21384 Altivec vector mode.
21385 (r6000_reg_class_enum): Add wm, wr, wv constraints.
21386 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
21387 (RS6000_BTM_CRYPTO): Likewise.
21388 (RS6000_BTM_COMMON): Likewise.
21390 * config/rs6000/rs6000.md (cpu attribute): Add power8.
21391 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
21392 (enum rs6000_vector): Add power8 vector support.
21394 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
21398 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
21399 (any_sibcall_could_use_r3): this and handle indirect calls.
21400 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
21402 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21404 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
21406 2013-05-22 Richard Biener <rguenther@suse.de>
21408 PR middle-end/57349
21409 * profile.c (branch_prob): Do not split blocks that are
21410 abnormally receiving from ECF_RETURNS_TWICE functions.
21412 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
21414 * recog.c (offsettable_address_addr_space_p): Fix calculation of
21415 address mode. Move pointer mode initialization to the same place.
21417 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
21419 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
21420 while it has any effect.
21422 2013-05-21 Easwaran Raman <eraman@google.com>
21424 PR tree-optimization/57322
21425 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
21426 UID of the statement added to the BB to be 1.
21428 2013-05-21 Jakub Jelinek <jakub@redhat.com>
21430 PR tree-optimization/57331
21431 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
21432 of conversion from pointer type to integral type with integer.
21434 2013-05-21 Martin Jambor <mjambor@suse.cz>
21437 * ipa-prop.c (ipa_read_node_info): Process param_used and
21438 controlled_uses in the same order as when writing.
21440 2013-05-21 Magnus Granberg <baldrick@free.fr>
21443 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
21445 2013-05-21 Richard Biener <rguenther@suse.de>
21447 PR tree-optimization/57318
21448 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
21449 estimate stmts with side-effects as likely eliminated.
21451 2013-05-21 Richard Biener <rguenther@suse.de>
21453 PR tree-optimization/57330
21454 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
21455 preserve the call stmts fntype.
21457 2013-05-21 Richard Biener <rguenther@suse.de>
21459 PR tree-optimization/57303
21460 * tree-ssa-sink.c (statement_sink_location): Improve killing
21461 stmt detection and properly handle self-assignments.
21463 2013-05-21 Christian Bruel <christian.bruel@st.com>
21465 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
21466 spanning registers. LEAF_REG_REMAP is supported only for contiguous
21467 registers. Set register size out of the PARALLEL loop.
21469 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
21472 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
21473 (*fmasf4, *fmasf4_media): New insns.
21475 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
21477 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
21478 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
21479 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
21480 (mips_idiv_insns): Update the comments to say that the returned
21481 instruction counts are in units of BASE_INSN_LENGTH.
21482 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
21483 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
21484 using 2 rather than 4 as the length of indirect MIPS16 and
21485 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
21486 length of a NOP. Don't divide MIPS16 lengths by 2.
21487 (mips16_split_long_branches): Assume a branch is long if the
21488 length is greater than 4 rather than 8.
21489 * config/mips/mips.md (length): Give MIPS16 lengths directly,
21490 rather than multiplying them by 2. Multiply instruction counts
21491 by BASE_INSN_LENGTH rather than 4.
21492 (*jump_mips16, tls_get_tp_mips16_<mode>)
21493 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
21495 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
21497 * config/mips/mips.md (extended_mips16): Remove branch case.
21498 (length): Remove duplicated extended_mips16 test.
21500 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
21502 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
21504 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
21506 * recog.h (Recog_data): Rename to...
21507 (recog_data_d): ...this.
21508 (recog_data): Update accordingly.
21509 * recog.c (recog_data): Likewise.
21510 * reload.c (save_recog_data): Likewise.
21511 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
21512 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
21514 2013-05-17 Julian Brown <julian@codesourcery.com>
21516 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
21517 found in a REG_EQUAL note, invalidate it.
21519 2013-05-17 Easwaran Raman <eraman@google.com>
21521 * tree-ssa-reassoc.c (find_insert_point): New function.
21522 (insert_stmt_after): Likewise.
21523 (get_def_stmt): Likewise.
21524 (ensure_ops_are_available): Likewise.
21525 (not_dominated_by): Likewise.
21526 (rewrite_expr_tree): Do not move statements beyond what is
21527 necessary. Remove call to swap_ops_for_binary_stmt...
21528 (reassociate_bb): ... and move it here.
21529 (build_and_add_sum): Assign UIDs for new statements.
21530 (linearize_expr): Likewise.
21531 (do_reassoc): Renumber gimple statement UIDs.
21533 2013-05-17 Jan Hubicka <jh@suse.cz>
21535 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
21537 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
21538 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
21540 (output_weakrefs): Update.
21542 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
21543 Martin Jambor <mjambor@suse.cz>
21545 PR middle-end/57276
21546 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
21547 value that corresponds to the given aggval is found in values vector.
21549 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
21551 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
21552 sse, sse2, sse3, ssse3 and sse4a flags to options.
21554 2013-05-17 David Malcolm <dmalcolm@redhat.com>
21556 * gengtype-state.c: (s_expr_writer): New class, to handle
21557 prettifying of output layout of s-expressions.
21558 (state_writer): New class, to write out gtype.state.
21559 (state_written_type_count): Move this variable into member data of
21561 (s_expr_writer::s_expr_writer): New code: constructor for new class
21562 (state_writer::state_writer(): ditto
21563 (s_expr_writer::write_new_line): New function
21564 (s_expr_writer::write_any_indent): ditto
21565 (s_expr_writer::begin_s_expr): ditto
21566 (s_expr_writer::end_s_expr): ditto
21567 (write_state_fileloc): convert to method of state_writer...
21568 (state_writer:: write_state_fileloc): ...and use methods of
21569 s_expr_writer to write indentation into the gtype.state output file
21570 to visually represent the hierarchical structure of the list
21572 (write_state_fields): ditto, renaming to...
21573 (state_writer::write_state_fields)
21574 (write_state_a_string): ditto, renaming to...
21575 (state_writer::write_state_a_string)
21576 (write_state_string_option): ditto, renaming to...
21577 (state_writer::write_state_string_option)
21578 (write_state_type_option): ditto, renaming to...
21579 (state_writer::write_state_type_option)
21580 (write_state_nested_option): ditto, renaming to...
21581 (state_writer::write_state_nested_option)
21582 (write_state_option): ditto, renaming to...
21583 (state_writer::write_state_option)
21584 (write_state_options): ditto, renaming to...
21585 (state_writer::write_state_options)
21586 (write_state_lang_bitmap): ditto, renaming to...
21587 (state_writer::write_state_lang_bitmap)
21588 (write_state_version): ditto, renaming to...
21589 (state_writer::write_state_version)
21590 (write_state_scalar_type): ditto, renaming to...
21591 (state_writer::write_state_scalar_type)
21592 (write_state_string_type): ditto, renaming to...
21593 (state_writer::write_state_string_type)
21594 (write_state_undefined_type): ditto, renaming to...
21595 (state_writer::write_state_undefined_type)
21596 (write_state_struct_union_type): ditto, renaming to...
21597 (state_writer::write_state_struct_union_type)
21598 (write_state_struct_type): ditto, renaming to...
21599 (state_writer::write_state_struct_type)
21600 (write_state_user_struct_type): ditto, renaming to...
21601 (state_writer::write_state_user_struct_type)
21602 (write_state_lang_struct_type): ditto, renaming to...
21603 (state_writer::write_state_lang_struct_type)
21604 (write_state_param_struct_type): ditto, renaming to...
21605 (state_writer::write_state_param_struct_type)
21606 (write_state_pointer_type): ditto, renaming to...
21607 (state_writer::write_state_pointer_type)
21608 (write_state_array_type): ditto, renaming to...
21609 (state_writer::write_state_array_type)
21610 (write_state_gc_used): ditto, renaming to...
21611 (state_writer::write_state_gc_used)
21612 (write_state_common_type_content): ditto, renaming to...
21613 (state_writer::write_state_common_type_content)
21614 (write_state_type): ditto, renaming to...
21615 (state_writer::write_state_type)
21616 (write_state_pair_list): ditto, renaming to...
21617 (state_writer::write_state_pair_list)
21618 (write_state_pair): ditto, renaming to...
21619 (state_writer::write_state_pair)
21620 (write_state_typedefs): ditto, renaming to...
21621 (state_writer::write_state_typedefs)
21622 (write_state_structures): ditto, renaming to...
21623 (state_writer::write_state_structures)
21624 (write_state_param_structs): ditto, renaming to...
21625 (state_writer::write_state_param_structs)
21626 (write_state_variables): ditto, renaming to...
21627 (state_writer::write_state_variables)
21628 (write_state_srcdir): ditto, renaming to...
21629 (state_writer::write_state_srcdir)
21630 (write_state_files_list): ditto, renaming to...
21631 (state_writer::write_state_files_list)
21632 (write_state_languages): ditto, renaming to...
21633 (state_writer::write_state_languages)
21634 (write_state): create a state_writer instance and use it when
21635 writing out the state file
21637 2013-05-17 Mike Stump <mikestump@comcast.net>
21639 PR rtl-optimization/57304
21640 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
21641 accessing DF_REF_REAL_LOC.
21643 2013-05-17 Jakub Jelinek <jakub@redhat.com>
21645 PR rtl-optimization/57281
21646 PR rtl-optimization/57300
21647 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
21648 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
21649 what the other splitter did if the registers are dead.
21651 2013-05-17 Richard Biener <rguenther@suse.de>
21653 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
21656 2013-05-17 Jakub Jelinek <jakub@redhat.com>
21658 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
21661 2013-05-17 Marek Polacek <polacek@redhat.com>
21663 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
21664 length when doing non-zero store of storing '\0' to '\0'.
21666 2013-05-17 Jakub Jelinek <jakub@redhat.com>
21668 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
21669 vect_external_def oprnd1 with loop_vinfo, try to emit
21670 optional cast, negation and and stmts on the loop preheader
21671 edge instead of into the pattern def seq.
21673 PR tree-optimization/57051
21674 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
21675 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
21677 2013-05-16 Nick Clifton <nickc@redhat.com>
21679 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
21680 (rl78_is_naked_func): New function.
21681 (rl78_expand_prologue): Skip prologue generation for naked functions.
21682 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
21683 * doc/extend.texi (naked): Add RL78 to the list of processors
21684 that supports this attribute.
21686 2013-05-16 Jeff Law <law@redhat.com>
21688 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
21690 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
21692 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
21693 cache parameters using detect_caches_amd also for CYRIX,
21694 NSC and TM2 signatures.
21696 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
21697 Dzianis Kahanovich <mahatma@eu.by>
21701 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
21702 VIA/Centaur processors and determine their cache parameters
21703 using detect_caches_amd.
21705 2013-05-16 Teresa Johnson <tejohnson@google.com>
21707 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
21708 (rtl_verify_edges): New function.
21709 (rtl_verify_bb_insns): Ditto.
21710 (rtl_verify_bb_pointers): Ditto.
21711 (rtl_verify_bb_insn_chain): Ditto.
21712 (rtl_verify_fallthru): Ditto.
21713 (rtl_verify_bb_layout): Ditto.
21714 (rtl_verify_flow_info_1): Outline checks into new functions.
21715 (rtl_verify_flow_info): Ditto.
21717 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
21719 * cfghooks.c (copy_bbs): Add update_dominance argument.
21720 * cfghooks.h (copy_bbs): Update prototype.
21721 * tree-cfg.c (gimple_duplicate_sese_region):
21722 Add update_dominance argument.
21723 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
21724 * tree-ssa-loop-ch.c (copy_loop_headers): Update
21725 gimple_duplicate_sese_region call.
21726 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
21727 Update copy_bbs call.
21728 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
21729 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
21731 2013-05-16 Jakub Jelinek <jakub@redhat.com>
21733 * tree-vectorizer.h (NUM_PATTERNS): Increment.
21734 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
21735 vect_recog_rotate_pattern.
21736 (vect_recog_rotate_pattern): New function.
21738 2013-05-16 Jason Merrill <jason@redhat.com>
21740 * Makefile.in (LLINKER): New variable.
21741 (mostlyclean): Remove link mutex.
21742 * configure.ac: Handle --enable-link-mutex.
21743 * lock-and-run.sh: New script.
21745 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
21748 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
21751 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
21753 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
21755 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
21757 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
21758 * config/arm/arm.c (next_consecutive_mem): New function.
21759 (gen_movmem_ldrd_strd): Likewise.
21760 * config/arm/arm.md (movmemqi): Update condition and code.
21761 (unaligned_loaddi, unaligned_storedi): New patterns.
21763 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
21765 * config.gcc: Obsolete *-*-solaris2.9*.
21766 * doc/install.texi (Specific, *-*-solaris2*): Document it.
21768 2013-05-16 Richard Biener <rguenther@suse.de>
21770 * passes.c (init_optimization_passes): Move pass_parallelize_loops
21771 earlier, after GRAPHITE transforms and IV canonicalization.
21773 2013-05-16 Jakub Jelinek <jakub@redhat.com>
21775 * omp-low.c (extract_omp_for_data): For collapsed loops,
21776 if at least one of the loops is known at compile time to
21777 iterate zero times, set count to 0.
21778 (expand_omp_regimplify_p): New function.
21779 (expand_omp_for_generic): For collapsed loops, if at least
21780 one of the loops isn't known to iterate at least once,
21781 add runtime check with setting count to 0.
21782 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
21783 For unsigned types if it isn't known at compile time that
21784 the loop will iterate at least once, add runtime check to bypass
21785 the whole loop if initial condition isn't true.
21787 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
21789 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
21791 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
21793 PR middle-end/57286
21794 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
21795 transformations to avoid an infinite loop.
21797 2013-05-16 Marek Polacek <polacek@redhat.com>
21799 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
21801 2013-05-15 Leif Ekblad <leif@rdos.net>
21803 * config/i386/i386.c (ix86_decompose_address): Use
21804 DEFAULT_TLS_SEG_REG to access TLS segment register.
21805 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
21806 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
21807 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
21809 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
21812 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
21813 sibling calls to functions that would normally be lazily bound,
21814 unless $gp is call-clobbered.
21816 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
21818 * config/i386/i386.c (ix86_option_override_internal): Update
21819 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
21820 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
21821 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
21822 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
21824 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
21826 2013-05-15 Andreas Schwab <schwab@suse.de>
21828 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
21829 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
21832 2013-05-15 Teresa Johnson <tejohnson@google.com>
21834 * loop-unroll.c (report_unroll_peel): Check decision before
21835 emitting unroll/peel message.
21837 2013-05-15 Teresa Johnson <tejohnson@google.com>
21839 * function.h (has_bb_partition): New rtl_data flag.
21840 (bb_reorder_complete): Ditto.
21841 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
21842 instead of flag_reorder_blocks_and_partition.
21843 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
21844 with some enhancements.
21845 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
21846 * bb-reorder.c (connect_traces): Check for has_bb_partition
21847 instead of flag_reorder_blocks_and_partition.
21848 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
21849 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
21850 verify_hot_cold_block_grouping.
21851 (partition_hot_cold_basic_blocks): Set has_bb_partition.
21853 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
21856 * config/arm/predicates.md (call_insn_operand): New predicate.
21857 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
21858 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
21859 if insn is not a tail call.
21860 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
21862 * config/arm/arm.h (enum reg_class): New caller save register class.
21863 (REG_CLASS_NAMES): Likewise.
21864 (REG_CLASS_CONTENTS): Likewise.
21865 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
21868 2013-05-15 Richard Biener <rguenther@suse.de>
21870 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
21871 of MSG_OPTIMIZED_LOCATIONS.
21872 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
21873 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
21875 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
21876 of MSG_OPTIMIZED_LOCATIONS.
21877 (execute_vect_slp): Likewise.
21878 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
21879 (vect_create_cond_for_alias_checks): Likewise.
21880 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
21881 (vect_recog_widen_mult_pattern): Likewise.
21882 (vect_recog_widen_sum_pattern): Likewise.
21883 (vect_recog_over_widening_pattern): Likewise.
21884 (vect_recog_widen_shift_pattern): Likewise.
21885 (vect_recog_vector_vector_shift_pattern): Likewise.
21886 (vect_recog_divmod_pattern): Likewise.
21887 (vect_recog_mixed_size_cond_pattern): Likewise.
21888 (vect_recog_bool_pattern): Likewise.
21889 (vect_pattern_recog_1): Likewise.
21891 2013-05-15 Martin Jambor <mjambor@suse.cz>
21893 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
21894 non-functions to builtin_unreachable.
21895 * ipa-inline-transform.c (inline_call): Do not assert estimates were
21896 correct when new direct edges were discovered.
21898 2013-05-15 Martin Jambor <mjambor@suse.cz>
21900 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
21901 header, print symbol order instead of node uid, print more information
21902 about indirect edge targets.
21903 (ipa_make_edge_direct_to_target): Print symbol order instead of node
21905 (ipa_make_edge_direct_to_target): Likewise.
21906 (remove_described_reference): Likewise.
21907 (propagate_controlled_uses): Likewise.
21908 (ipa_print_node_params): Also print symbol order.
21909 (ipcp_transform_function): Print symbol order instead of node uids.
21910 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
21911 (cgraph_get_create_real_symbol_node): Likewise.
21912 * ipa-cp.c (print_lattice): Likewise.
21913 (print_all_lattices): Likewise.
21914 (determine_versionability): Likewise.
21915 (initialize_node_lattices): Likewise.
21916 (estimate_local_effects): Likewise.
21917 (update_profiling_info): Likewise.
21918 (create_specialized_node): Likewise.
21919 (perhaps_add_new_callers): Likewise.
21920 (decide_about_value): Likewise.
21921 (decide_whether_version_node): Likewise.
21922 (identify_dead_nodes): Likewise.
21923 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
21924 (dump_inline_summary): Likewise.
21925 (estimate_node_size_and_time): Likewise.
21926 (inline_analyze_function): Likewise.
21927 * ipa-inline.c (report_inline_failed_reason): Likewise.
21928 (want_early_inline_function_p): Likewise.
21929 (edge_badness): Likewise.
21930 (update_edge_key): Likewise.
21931 (inline_small_functions): Likewise. Add dumping of order to two other
21933 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
21934 instead of node uids.
21935 (propagate_pure_const): Likewise.
21936 (propagate_pure_const): Likewise.
21937 * ipa-utils.c (dump_cgraph_node_set): Likewise.
21938 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
21939 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
21941 * tree-pretty-print.c (dump_function_header): Likewise.
21942 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
21943 Print symbol order instead of node uids.
21945 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
21947 * config/s390/s390.c (s390_register_move_cost): Don't impose the
21948 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
21950 2013-05-15 Richard Biener <rguenther@suse.de>
21952 PR tree-optimization/57275
21953 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
21954 return value for fail to do runtime alias checks for gather loads.
21956 2013-05-15 Jan Hubicka <jh@suse.cz>
21960 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
21961 weakrefs are not external.
21962 (lto_symtab_merge_decls): Fix thinko when dealing with
21963 non-lto_symtab decls.
21964 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
21965 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
21966 * varpool.c (dump_varpool_node): Dump more flags.
21968 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
21970 * config/i386/i386.c (processor_alias_table): Add instruction
21971 FSGSBASE for AMD bdver3 architecture.
21973 2013-05-14 Jakub Jelinek <jakub@redhat.com>
21975 * tree.c (warn_deprecated_use): Print file:line using locus color.
21976 * diagnostic.c (diagnostic_report_current_module): Print file:line
21977 and file:line:column using locus color.
21979 2013-05-14 Mike Stump <mikestump@comcast.net>
21981 * gdbinit.in: Add __null.
21983 2013-05-14 Mike Stump <mikestump@comcast.net>
21985 * recog.h: Rename struct recog_data to Recog_data.
21986 * recog.c: Likewise.
21987 * reload.c (can_reload_into): Likewise.
21988 * config/picochip/picochip.c: Likewise.
21990 2013-05-14 Mike Stump <mikestump@comcast.net>
21992 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
21994 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
21996 * resource.h (struct resources): Remove unch_memory member.
21997 (CLEAR_RESOURCE): Don't clear unch_memory.
21998 * resource.c (mark_referenced_resources): Don't set it.
21999 (mark_set_resources): Likewise.
22000 (mark_target_live_regs): Don't clear it.
22001 (init_resource_info): Likewise.
22002 * reorg.c (resource_conflicts_p): Don't compare it.
22003 (redundant_insn): Don't set it.
22005 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
22007 * emit-rtl.c (next_label): Remove unused function.
22008 (skip_consecutive_labels, link_cc0_insns): Move to ...
22009 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
22010 only place where these functions are used, and make them static.
22012 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
22014 * fold-const.c (fold_negate_expr): Handle vectors.
22015 (fold_truth_not_expr): Make it static.
22016 (fold_invert_truthvalue): New static function.
22017 (invert_truthvalue_loc): Handle vectors. Do not call
22018 fold_truth_not_expr directly.
22019 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
22020 <TRUTH_NOT_EXPR>: Do not cast to boolean.
22021 (fold_comparison): Handle vector constants.
22022 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
22023 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
22024 * tree.h (fold_truth_not_expr): Remove declaration.
22026 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
22028 * config/aarch64/aarch64-simd.md
22029 (aarch64_vcond_internal<mode>): Rename to...
22030 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
22031 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
22032 float modes. Clarify all iterator modes.
22033 (vcond<mode><mode>): Use new name for vcond expanders.
22034 (vcond<v_cmp_result><mode>): Likewise.
22035 (vcondu<mode><mode>: Likewise.
22036 * config/aarch64/iterators.md (VDQF_COND): New.
22038 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
22041 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
22042 variable for the shift amount. Check that we shift by non-negative
22045 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
22048 * config/arm/arm.h (EPILOGUE_USES): Only return true
22049 for LR_REGNUM after epilogue_completed.
22051 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
22053 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
22054 is error_mark_node.
22056 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
22059 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
22060 and Solaris 11+/x86 with gld.
22061 * configure: Regenerate.
22063 2013-05-14 Jakub Jelinek <jakub@redhat.com>
22065 * expmed.c (expand_shift_1): Canonicalize rotates by
22066 constant bitsize / 2 to bitsize - 1.
22067 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
22068 case ROTATERT>: Likewise.
22071 2013-05-10 Jakub Jelinek <jakub@redhat.com>
22073 * config/i386/i386.md (rotateinv): New code attr.
22074 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
22075 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
22076 roll $31, %eax, etc.
22078 2013-05-14 Richard Biener <rguenther@suse.de>
22080 PR middle-end/57235
22081 * tree-eh.c (sink_clobbers): Give up for successors with
22082 multiple predecessors and no virtual uses.
22084 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
22086 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
22087 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
22089 2013-05-14 Jakub Jelinek <jakub@redhat.com>
22091 PR middle-end/57251
22092 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
22093 the case when both op0 and op1 have VOIDmode.
22095 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
22097 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
22098 in multiply-accumulate mode.
22100 2013-05-13 Guozhi Wei <carrot@google.com>
22102 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
22104 2013-05-13 Kai Tietz <ktietz@redhat.com>
22107 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
22108 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
22109 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
22110 * config/i386/i386.c (ix86_option_override_internal): Likewise.
22111 (ix86_expand_prologue): Likewise.
22112 (ix86_expand_split_stack_prologue): Likewise.
22113 (legitimate_pic_address_disp_p): Likewise.
22114 (legitimize_pic_address): Likewise.
22115 (legitimize_tls_address): Likewise.
22116 (legitimize_pe_coff_symbol): Likewise.
22117 (output_pic_addr_const): Likewise.
22118 (construct_plt_address): Likewise.
22119 (ix86_expand_call): Likewise.
22120 (x86_output_mi_thunk): Likewise.
22121 (x86_function_profiler): Likewise.
22123 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
22125 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
22126 similar switch cases.
22127 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
22128 (aarch64_simd_mov_to_<mode>low): Delete.
22129 (aarch64_simd_mov_to_<mode>high): Delete.
22130 (move_lo_quad_<mode>): Add w<-r alternative.
22131 (aarch64_simd_move_hi_quad_<mode>): Likewise.
22132 (aarch64_simd_mov_from_*): Update type attribute.
22133 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
22136 2013-05-13 Jan Hubicka <jh@suse.cz>
22138 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
22139 * config/i386/i386.c (ix86_compute_frame_layout,
22140 ix86_expand_epilogue, emit_i387_cw_initialization,
22141 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
22142 ix86_local_alignment): Fix use of size/speed predicates.
22144 2013-05-13 Jakub Jelinek <jakub@redhat.com>
22146 PR tree-optimization/45216
22147 PR tree-optimization/57157
22148 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
22149 the (-Y) & (B - 1) variant if OP is |.
22150 * expmed.c (expand_shift_1): For rotations by const0_rtx just
22151 return shifted. Use (-op1) & (prec - 1) as other_amount
22152 instead of prec - op1.
22154 2013-05-13 Martin Jambor <mjambor@suse.cz>
22156 PR middle-end/42371
22157 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
22158 (ipa_constant_data): New type.
22159 (ipa_jump_func): Use ipa_constant_data to hold information about
22160 constant jump functions.
22161 (ipa_get_jf_constant): Adjust to jump function type changes.
22162 (ipa_get_jf_constant_rdesc): New function.
22163 (ipa_param_descriptor): New field controlled_uses.
22164 (ipa_get_controlled_uses): New function.
22165 (ipa_set_controlled_uses): Likewise.
22166 * ipa-ref.h (ipa_find_reference): Declare.
22167 * ipa-prop.c (ipa_cst_ref_desc): New type.
22168 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
22170 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
22171 New parameter cs. Adjust all callers.
22172 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
22173 (remove_described_reference): New function.
22174 (jfunc_rdesc_usable): Likewise.
22175 (try_make_edge_direct_simple_call): Decrement controlled use count,
22176 attempt to remove reference if it hits zero.
22177 (combine_controlled_uses_counters): New function.
22178 (propagate_controlled_uses): Likewise.
22179 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
22180 (ipa_edge_duplication_hook): Duplicate reference descriptions.
22181 (ipa_print_node_params): Print described use counter.
22182 (ipa_write_jump_function): Adjust to jump function type changes.
22183 (ipa_read_jump_function): New parameter CS, pass it to
22184 ipa_set_jf_constant. Adjust caller.
22185 (ipa_write_node_info): Stream controlled use count
22186 (ipa_read_node_info): Likewise.
22187 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
22189 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
22190 count. Remove cloning-added reference if it reaches zero.
22191 * ipa-ref.c (ipa_find_reference): New function.
22193 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
22195 * config/i386/i386.c (processor_target_table): Modified default
22196 alignment values for AMD BD and BT architectures.
22198 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
22200 * tree-vect-generic.c (uniform_vector_p): Move ...
22201 * tree.c (uniform_vector_p): ... here.
22202 * tree.h (uniform_vector_p): Declare it.
22203 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
22206 2013-05-13 Jakub Jelinek <jakub@redhat.com>
22208 PR tree-optimization/57230
22209 * tree-ssa-strlen.c (handle_char_store): Record length for
22210 array store from STRING_CST.
22212 PR tree-optimization/57230
22213 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
22216 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
22218 * config/epiphany/epiphany.c (epiphany_init): Check size of
22219 NUM_MODES_FOR_MODE_SWITCHING.
22220 (epiphany_expand_prologue):
22221 Remove CONFIG_REGNUM initial value handling code.
22222 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
22223 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
22224 (emit_set_fp_mode, epiphany_mode_after): Likewise.
22225 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
22226 Don't return 1 for FP_MODE_NONE.
22227 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
22228 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
22229 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
22230 * config/epiphany/epiphany.md (save_config): New pattern.
22232 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
22234 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
22236 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
22238 * config/i386/i386.md (memory): Handle sseishft1.
22239 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
22240 (*vec_extractv2di_1): Ditto.
22242 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
22244 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
22247 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
22249 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
22250 Add mthumb/march=armv7-a multilib.
22251 Add mthumb/march=armv7-r multilib.
22252 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
22254 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
22256 * config/v850/t-rtems: Add more multilibs.
22258 2013-05-10 Richard Biener <rguenther@suse.de>
22260 PR tree-optimization/57214
22261 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
22262 not propagate from SSA names that occur in abnormal PHI nodes.
22264 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
22266 * stor-layout.c (element_precision): New function.
22267 * machmode.h (element_precision): Declare it.
22268 * tree.c (build_minus_one_cst): New function.
22269 (element_precision): Likewise.
22270 * tree.h (build_minus_one_cst): Declare new function.
22271 (element_precision): Likewise.
22272 * fold-const.c (operand_equal_p): Use element_precision.
22273 (fold_binary_loc): Handle vector types.
22274 * convert.c (convert_to_integer): Use element_precision.
22275 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
22278 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
22280 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
22281 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
22282 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
22283 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
22284 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
22285 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
22286 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
22287 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
22288 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
22289 (Uuw8): New constraints.
22290 (Usb4): Move into alphabetical order.
22291 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
22292 (sd8_operand, ub8_operand, uw8_operand): New predicates.
22293 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
22294 previously unnamed patterns.
22295 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
22296 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
22297 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
22298 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
22299 of set_attr_alternative/if_then_else. Use extended_mips16 instead
22300 of specific lengths.
22302 2013-05-10 Jakub Jelinek <jakub@redhat.com>
22304 * config/i386/i386.md (rotateinv): New code attr.
22305 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
22306 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
22307 roll $31, %eax, etc.
22309 PR tree-optimization/45216
22310 PR tree-optimization/57157
22311 * tree-ssa-forwprop.c (simplify_rotate): New function.
22312 (ssa_forward_propagate_and_combine): Call it.
22314 2013-05-10 Richard Biener <rguenther@suse.de>
22316 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
22317 disable peeling when we version for aliasing.
22318 (vector_alignment_reachable_p): Honor explicit user alignment.
22319 (vect_supportable_dr_alignment): Likewise.
22320 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
22321 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
22322 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
22323 then peeling to arrange for the cost-model check to come first.
22325 2013-05-10 Alan Modra <amodra@gmail.com>
22327 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
22328 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
22329 * configure: Regenerate.
22331 2013-05-10 Alan Modra <amodra@gmail.com>
22334 * varasm.c (default_elf_select_section): Move !DECL_P check..
22335 (get_named_section): ..to here before calling get_section_name.
22337 (default_section_type_flags): Add DECL_P check.
22338 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
22339 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
22341 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
22343 * config/epiphany/epiphany.c (epiphany_expand_prologue):
22344 When using gen_stack_adjust_str with a register offset, add a
22345 REG_FRAME_RELATED_EXPR note.
22347 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
22349 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
22350 (*vec_extractv4si_zext_mem): Ditto.
22351 (*vec_extractv2di): Add 0->x and x->x alternatives.
22352 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
22353 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
22355 2013-05-09 Jason Merrill <jason@redhat.com>
22357 N3639 C++1y VLA support
22358 * gimplify.c (gimplify_vla_decl): Don't touch an existing
22361 * tree.c (build_constructor_va): New.
22362 * tree.h: Declare it.
22364 2013-05-09 Martin Jambor <mjambor@suse.cz>
22367 * gimple-fold.c (canonicalize_constructor_val): Call
22368 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
22370 2013-05-09 Jan Hubicka <jh@suse.cz>
22371 Richard Biener <rguenther@suse.de>
22374 * symtab.c (symtab_make_decl_local): Do not add private names.
22376 2013-05-09 Jan Hubicka <jh@suse.cz>
22379 * symtab.c (insert_to_assembler_name_hash): Handle clones.
22380 (unlink_from_assembler_name_hash): Likewise.
22381 (symtab_prevail_in_asm_name_hash, symtab_register_node,
22382 symtab_unregister_node, symtab_initialize_asm_name_hash,
22383 change_decl_assembler_name): Update.
22385 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
22387 * config/aarch64/aarch64.md: New movtf split.
22388 (*movtf_aarch64): Update.
22389 (aarch64_movdi_tilow): Handle TF modes and rename to
22390 aarch64_movdi_<mode>low.
22391 (aarch64_movdi_tihigh): Handle TF modes and rename to
22392 aarch64_movdi_<mode>high
22393 (aarch64_movtihigh_di): Handle TF modes and rename to
22394 aarch64_mov<mode>high_di
22395 (aarch64_movtilow_di): Handle TF modes and rename to
22396 aarch64_mov<mode>low_di
22397 (aarch64_movtilow_tilow): Remove spurious whitespace.
22398 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
22400 (aarch64_print_operand): Update.
22402 2013-05-09 Alan Modra <amodra@gmail.com>
22404 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
22406 * configure: Regenerate.
22408 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
22410 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
22411 splitter preparation statements.
22412 * config/i386/sse.md (*vec_extract* splitters): Ditto.
22413 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
22416 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
22418 * gimple-ssa-strength-reduction.c (count_candidates): Change
22419 return value to int.
22420 (analyze_candidates_and_replace): Change type of length to int.
22422 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
22424 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
22425 (*vec_extract<mode>): Use VI12_128 mode iterator.
22426 (*vec_extract<mode>_mem): Ditto.
22427 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
22430 2013-05-08 Diego Novillo <dnovillo@google.com>
22435 2012-08-17 Diego Novillo <dnovillo@google.com>
22438 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
22439 * config.in: Regenerate.
22440 * configure: Regenerate.
22441 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
22443 2013-05-08 Jan Hubicka <jh@suse.cz>
22446 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
22447 * cgraph.h (symtab_node_base): Add unique_name.
22448 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
22449 input_overwrite_node, input_varpool_node): Stream unique_name.
22450 * cgraphclones.c (cgraph_create_virtual_clone,
22451 cgraph_function_versioning): Set unique_name.
22452 * ipa.c (function_and_variable_visibility): Set unique_name.
22454 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
22456 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
22457 (alloc_cand_and_find_basis): Restrict conditional candidate
22458 processing to CAND_MULTs.
22460 2013-05-08 Jan Hubicka <jh@suse.cz>
22463 lto-symtab.c (lto_symtab_symbol_p): New function.
22464 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
22465 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
22466 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
22467 Skip static symbols.
22469 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
22471 PR tree-optimization/57200
22472 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
22473 Only call inform if the preceding warning_at returns true.
22475 2013-05-07 Han Shen <shenhan@google.com>
22477 * cfgexpand.c (record_or_union_type_has_array_p): New function.
22478 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
22479 * common.opt (fstack-protector-strong): New option.
22480 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
22481 * doc/invoke.texi (Optimization Options): Document
22482 "-fstack-protector-strong".
22483 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
22485 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
22487 * config/mips/mips.c (mips_machine_reorg2): Return 0.
22489 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
22491 * ira.c (update_equiv_regs): Add insn having equiv memory even if
22492 it is not lhs of the insn.
22493 (setup_reg_equiv): Remove insn having equiv memory which it is not
22495 * lra-constraints.c (process_address): Try to improve generation
22496 code for address base + disp.
22497 (lra_constraints): Make correct the code for checking insn setting
22498 up backward equivalence. Remove insn only if it is in the init
22500 * lra-eliminations.c (update_reg_eliminate): Change return value.
22501 (lra_eliminate): Use the result.
22503 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
22505 * config/i386/sse.md (ssescalarnummask): New mode attribute.
22506 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
22507 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
22508 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
22509 register target operands.
22510 (*vec_extractv8hi_sse2): New pattern.
22511 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
22512 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
22513 (*vec_extract<mode>_mem): New insn and split pattern.
22515 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
22517 * config/arm/arm.c (arm_asan_shadow_offset): New function.
22518 (TARGET_ASAN_SHADOW_OFFSET): Define.
22519 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
22520 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
22522 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
22524 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
22525 (incr_vec_index): Return -1 if increment not found.
22526 (create_add_on_incoming_edge): Assert if increment not found.
22527 (record_increment): Limit number of increments recorded.
22528 (all_phi_incrs_profitable): Return false if an increment not found.
22529 (replace_profitable_candidates): Don't process increments that were
22531 (analyze_candidates_and_replace): Limit size of incr_vec.
22533 2013-05-07 Richard Biener <rguenther@suse.de>
22535 * calls.c (special_function_p): setjmp-like functions are leaf.
22536 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
22537 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
22539 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
22541 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
22542 (aarch64_simd_mov<mode>): New expander.
22543 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
22544 (aarch64_simd_mov_to_<mode>high): Likewise.
22545 (aarch64_simd_mov_from_<mode>low): Likewise.
22546 (aarch64_simd_mov_from_<mode>high): Likewise.
22547 (aarch64_dup_lane<mode>): Update.
22548 (aarch64_dup_lanedi): New instruction pattern.
22549 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
22550 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
22552 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
22554 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
22555 (replace_mult_candidate): Remove unnecessary argument; remove
22556 unnecessary parameter from call to introduce_cast_before_cand.
22557 (replace_unconditional_candidate): Remove unnecessary parameter
22558 from call to replace_mult_candidate.
22559 (replace_conditional_candidate): Likewise.
22560 (insert_initializers): Use make_temp_ssa_name.
22561 (introduce_cast_before_cand): Remove unnecessary argument; use
22562 make_temp_ssa_name.
22563 (replace_one_candidate): Remove unnecessary argument; remove
22564 unnecessary parameter from calls to introduce_cast_before_cand.
22565 (replace_profitable_candidates): Remove unnecessary parameters
22566 from calls to replace_one_candidate.
22568 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
22570 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
22571 phi def as possibly hiding a basis for a CAND_ADD whose operands
22572 have been commuted in the analysis.
22573 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
22575 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22577 * config/aarch64/aarch64.md
22578 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
22579 shift value between 0-4.
22581 2013-05-07 Richard Biener <rguenther@suse.de>
22583 * double-int.h (rshift): New overload.
22584 * double-int.c (rshift): New function.
22585 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
22586 (create_reference_ops_from_ref): Remove.
22587 (vn_reference_insert): Use shared ops for constructing the
22588 reference and copy it.
22590 2013-05-07 Richard Biener <rguenther@suse.de>
22592 PR middle-end/57190
22593 * tree-eh.c (sink_clobbers): Properly propagate
22594 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
22596 2013-05-07 Jakub Jelinek <jakub@redhat.com>
22598 PR tree-optimization/57149
22599 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
22600 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
22601 collect_phi_def_edges, execute_late_warn_uninitialized): Use
22602 uninit_undefined_value_p instead of ssa_undefined_value_p.
22605 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
22606 for modifier == EXPAND_INITIALIZER.
22608 2013-05-07 Anton Blanchard <anton@samba.org>
22610 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
22611 for powerpc64 little endian.
22612 * configure: Regenerate.
22614 2013-05-06 Graham Stott <grahams@btinternet.com>
22616 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
22617 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
22618 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
22619 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
22621 2013-05-06 Graham Stott <grahams@btinternet.com>
22623 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
22624 codes which allow non-lvalues.
22626 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
22628 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
22629 components are all 1s.
22630 (integer_minus_onep): New function.
22631 * tree.h (integer_minus_onep): Declare it.
22632 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
22633 integer_minus_onep instead of integer_all_onesp.
22635 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
22638 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
22639 variations of these patterns.
22641 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
22643 * config/i386/i386.md (isa): Add x64_sse4 member.
22644 (enabled): Handle x64_sse4.
22645 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
22646 instruction for 64bit SSE4_1 targets. Update insn attributes.
22647 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
22648 instruction for SSE4_1 targets. Update insn attributes.
22649 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
22650 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
22651 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
22653 (*vec_extractv4si): Rename from *sse4_1_pextrd.
22654 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
22655 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
22657 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
22660 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
22662 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
22664 * final.c (do_assembler_dialects): Don't handle curly braces and
22665 vertical bar escaped by % as dialect delimiters.
22666 (output_asm_insn): Print curly braces and vertical bar if escaped
22667 by % and ASSEMBLER_DIALECT defined.
22668 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
22669 * doc/tm.texi: Regenerated.
22671 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
22673 * config/mips/mips.c: Include tree-pass.h.
22674 (mips_reorg): Split in pre- and post-dbr_schedule parts.
22675 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
22676 (pass_mips_machine_reorg2): New machine specific pass.
22677 (insert_pass_mips_machine_reorg2): New pass plugin definition.
22678 (mips_option_override): Register the new pass.
22679 * rtl.h (cleanup_barriers): Remove prototype.
22680 (dbr_schedule): Likewise.
22681 * jump.c (cleanup_barriers): Make static.
22682 * reorg.c (dbr_schedule): Likewise.
22684 2013-05-06 Richard Biener <rguenther@suse.de>
22686 PR tree-optimization/57185
22687 * tree-parloops.c (add_field_for_reduction): Handle anonymous
22688 SSA names properly.
22690 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
22693 * config/i386/i386.c (add_parameter_dependencies): Add dependence
22694 between "first_arg" and "insn", not "last" and "insn".
22696 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
22698 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
22699 (find_candidates_in_block): Re-enable slsr_process_phi.
22700 (create_phi_basis): Fix double counting of candidate adjustment.
22702 2013-05-06 Richard Biener <rguenther@suse.de>
22704 PR middle-end/57147
22705 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
22706 the edge is also fallthru, preserve it and just clear the
22708 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
22709 also complex, preserve that and just clear the fallthru flag.
22710 * tree-inline.c (update_ssa_across_abnormal_edges): Also
22711 update virtual operands.
22713 2013-05-06 Alan Modra <amodra@gmail.com>
22715 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
22716 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
22717 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
22718 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
22719 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
22720 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
22722 2013-05-06 Alan Modra <amodra@gmail.com>
22724 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
22725 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
22726 (DEFAULT_ASM_ENDIAN): Define.
22727 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
22728 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
22729 Update -K PIC clause from sysv4.h.
22730 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
22731 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
22733 2013-05-06 Alan Modra <amodra@gmail.com>
22735 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
22736 twice for little-endian.
22737 (ashrdi3_no_power, ashrdi3): Support little-endian.
22739 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
22742 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
22743 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
22745 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
22746 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
22747 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
22750 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
22751 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
22753 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
22754 * config/pa/pa.opt: Make mbig-switch a no-op.
22755 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
22756 (CASE_VECTOR_MODE): Always return SImode.
22757 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
22758 for the !TARGET_BIG_SWITCH case.
22759 * config/pa/pa-linux.h: Likewise.
22760 * config/pa/pa-openbsd.h: Likewise.
22761 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
22762 * config/pa/pa.md (short_jump): Remove define_insn.
22763 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
22764 (casesi0): Remove define_insn.
22765 (type): Remove btable_branch.
22766 (pa_combine_type): Likewise.
22767 (in_nullified_branch_delay): Likewise.
22768 (in_call_delay): Likewise.
22769 (define_delay): Likewise.
22770 (define_insn_reservation "Z3"): Likewise.
22771 (define_insn_reservation "Z4"): Likewise.
22772 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
22773 (pa_adjust_insn_length): Remove adjustment for btable branches.
22774 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
22777 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
22779 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
22780 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
22781 Add m->r,x alternatives.
22782 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
22783 splitters using SWI48x mode iterator.
22784 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
22785 TARGET_64BIT. Add m->x alternative.
22786 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
22787 Add o->x alternative. Enable for TARGET_SSE.
22788 (sse_storeq): Remove expander.
22789 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
22790 with memory input operand.
22791 (*vec_extractv2di_1 splitter): New.
22792 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
22793 * config/i386/i386.md (ssevecmodelower): New mode attribute.
22795 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
22797 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
22798 (INT_LOWPART): Delete.
22799 (extract_MB): Adjust.
22800 (extract_ME): Adjust.
22801 (print_operand): Adjust.
22803 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
22805 * config/rs6000/predicates.md (reg_or_add_cint_operand,
22806 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
22807 (reg_or_logical_cint_operand, easy_fp_constant,
22808 logical_const_operand): Delete "CONST_DOUBLE" case.
22809 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
22810 "HOST_BITS_PER_WIDE_INT == 64" test.
22811 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
22812 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
22813 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
22814 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
22816 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
22817 CONST_DOUBLE DImode/VOIDmode case.
22818 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
22819 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
22820 CONST_DOUBLE VOIDmode case.
22821 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
22822 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
22823 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
22824 Delete CONST_DOUBLE case.
22825 (splitters for mov FMOVE64 const_double): Delete
22826 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
22827 "HOST_BITS_PER_WIDE_INT >= 64" test.
22828 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
22830 (mov DI const_double): Delete.
22832 2013-05-04 Jakub Jelinek <jakub@redhat.com>
22834 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
22835 on op shows all bits zero in mode of a lowpart subreg, return zero.
22837 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
22840 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
22841 to save TFmode registers and DImode to save TImode registers for
22842 caller save operations.
22843 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
22844 mark being partially clobbered since they only use the first
22847 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
22848 and TDmode only use the upper 64-bits of each VSX register.
22850 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
22852 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
22853 (find_candidates_in_block): Disable slsr_process_phi.
22855 2013-05-03 Guozhi Wei <carrot@google.com>
22857 * coverage.c (coverage_obj_init): Move the construction of gcov
22859 (build_init_ctor): ... here.
22861 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
22863 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
22864 (slsr_cand_d): Redefine def_phi.
22865 (stride_status, phi_adjust_status, count_phis_status): New enums.
22866 (find_phi_def): New.
22867 (find_basis_for_base_expr): New.
22868 (find_basis_for_candidate): Handle hidden bases.
22869 (alloc_cand_and_find_basis): Handle phi candidates.
22870 (slsr_process_phi): New.
22871 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
22872 (create_mul_imm_cand): Likewise.
22873 (create_add_ssa_cand): Exclude phi base candidates.
22874 (create_add_imm_cand): Likewise.
22875 (slsr_process_cast): Likewise.
22876 (slsr_process_copy): Likewise.
22877 (find_candidates_in_block): Handle phi candidates.
22878 (dump_candidate): Likewise.
22879 (unconditional_cands): Delete.
22880 (unconditional_cands_with_known_stride_p): Delete.
22881 (phi_dependent_cand_p): New.
22882 (cand_increment): Handle phi-dependent candidates.
22883 (replace_dependent): Delete.
22884 (replace_mult_candidate): New.
22885 (replace_unconditional_candidate): New.
22886 (incr_vec_index): Move to avoid forward reference.
22887 (create_add_on_incoming_edge): New.
22888 (create_phi_basis): New.
22889 (replace_dependents): Delete.
22890 (replace_conditional_candidate): New.
22891 (phi_add_costs): New.
22892 (replace_uncond_cands_and_profitable_phis): New.
22893 (record_increment): Handle phi adjustments.
22894 (record_phi_increments): New.
22895 (record_increments): Handle phi adjustments.
22896 (phi_incr_cost): New.
22897 (lowest_cost_path): Handle phis.
22898 (total_savings): Likewise.
22899 (analyze_increments): Likewise.
22900 (ncd_with_phi): New.
22901 (ncd_of_cand_and_phis): New.
22902 (nearest_common_dominator_for_cands): Handle phi increments.
22903 (all_phi_incrs_profitable): New.
22904 (replace_profitable_candidates): Handle phi-dependent candidates.
22905 (analyze_candidates_and_replace): Likewise.
22907 2013-05-03 Teresa Johnson <tejohnson@google.com>
22910 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
22911 do not exceed REG_BR_PROB_BASE.
22913 2013-05-03 Jeff Law <law@redhat.com>
22915 PR tree-optimization/57144
22916 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
22917 operand of the condition will bit into the new type when eliminating
22918 a cast feeding a condition.
22920 2013-05-03 Jakub Jelinek <jakub@redhat.com>
22922 PR rtl-optimization/57130
22923 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
22924 of COMPARE as in_code to the recursive call if needed.
22926 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
22928 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
22929 (enabled): Handle new members.
22930 * config/i386/sse.md (*vec_concatv2si): Merge from
22931 *vec_concatv2si_sse2 and vec_concatv2si_sse.
22932 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
22934 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
22936 PR tree-optimization/57027
22937 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
22938 for fnms opportunity, check we got the prerequisite kind
22939 of tree / gimple before using accessor functions.
22941 2013-05-03 Richard Biener <rguenther@suse.de>
22943 * double-int.h (lshift): New overload without precision
22944 and arith argument.
22945 (operator *=, operator +=, operator -=): Move ...
22946 * double-int.c (operator *=, operator +=, operator -=): ... here
22947 and implement more efficiently.
22948 (mul_double_with_sign): Remove.
22949 (lshift_double): Adjust to take unsinged shift argument, push
22950 dispatching code to callers.
22951 (mul_double_wide_with_sign): Add early out for callers that
22952 are not interested in high parts or overflow.
22953 (lshift): New function.
22954 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
22955 dispatch code here.
22956 (lrotate, rrotate): Use logical shifts.
22957 * expr.c (get_inner_reference): Use lshift.
22958 * fixed-value.c (do_fixed_divide): Likewise.
22959 * tree-dfa.c (get_ref_base_and_extent): Likewise.
22960 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
22961 (indirect_refs_may_alias_p): Likewise.
22962 (stmt_kills_ref_p_1): Likewise.
22964 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
22966 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
22968 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
22970 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
22971 scalar form of FABD instruction.
22973 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
22975 * lra-constraints.c (process_alt_operands): Add checking alt
22976 number to choose the best alternative.
22978 2013-05-02 Richard Biener <rguenther@suse.de>
22980 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
22981 bitmap and its handling.
22982 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
22984 2013-05-02 Richard Biener <rguenther@suse.de>
22986 PR middle-end/57140
22987 * tree-inline.c (copy_loops): Properly handle removed loops.
22988 (copy_cfg_body): Mark destination loops for fixup if source
22989 loops needed fixup.
22991 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
22994 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
22995 generating simple_return for naked functions.
22997 2013-05-02 Martin Jambor <mjambor@suse.cz>
22999 PR middle-end/56988
23000 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
23001 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
23003 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
23004 ipa_agg_replacement_value structures.
23005 (known_aggs_to_agg_replacement_list): Likewise.
23006 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
23007 (read_agg_replacement_chain): Likewise.
23008 (ipcp_transform_function): Also check that by_ref flags match.
23010 2013-05-02 Richard Biener <rguenther@suse.de>
23012 * graphds.h (struct graph): Add obstack member.
23013 * graphds.c (new_graph): Initialize obstack and allocate
23015 (add_edge): Allocate edge from the obstack.
23016 (free_graph): Free the obstack instead of all edges and vertices.
23018 2013-05-02 Teresa Johnson <tejohnson@google.com>
23020 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
23022 * cfg.c (update_bb_profile_for_threading): Ditto.
23023 * tree-inline.c (copy_bb): Ditto.
23024 (copy_edges_for_bb): Ditto.
23025 (initialize_cfun): Ditto.
23026 (copy_cfg_body): Ditto.
23027 (expand_call_inline): Ditto.
23028 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
23029 (estimate_node_size_and_time): Ditto.
23030 (inline_merge_summary): Ditto.
23031 * cgraphclones.c (cgraph_clone_edge): Ditto.
23032 (cgraph_clone_node): Ditto.
23033 * sched-rgn.c (compute_dom_prob_ps): Ditto.
23034 (compute_trg_info): Ditto.
23036 2013-05-02 Ian Bolton <ian.bolton@arm.com>
23038 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
23039 S reg when fp attribute set.
23040 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
23042 2013-05-02 Ian Bolton <ian.bolton@arm.com>
23044 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
23046 (*and_one_cmplsi3_compare0_uxtw): Likewise.
23047 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
23048 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
23050 2013-05-02 Richard Biener <rguenther@suse.de>
23052 * tree-scalar-evolution.c (scev_info_hasher): Remove.
23053 (struct instantiate_cache_entry): New type.
23054 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
23055 (struct instantiate_cache_type): New type.
23056 (set_instantiated_value, get_instantiated_value): Remove.
23057 (get_instantiated_value_entry): New function.
23058 (instantiate_scev_name): Use the new cache and adjust.
23059 (instantiate_scev_poly): Adjust.
23060 (instantiate_scev_binary): Likewise.
23061 (instantiate_array_ref): Likewise.
23062 (instantiate_scev_convert): Likewise.
23063 (instantiate_scev_not): Likewise.
23064 (instantiate_scev_3): Likewise.
23065 (instantiate_scev_2): Likewise.
23066 (instantiate_scev_r): Likewise.
23067 (instantiate_scev): Likewise.
23068 (resolve_mixers): Likewise.
23070 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
23073 * lra-constraints.c (best_small_class_operands_num): Remove.
23074 (process_alt_operands): Remove small_class_operands_num. Take
23075 small classes operands into losers and only if the operand is not
23076 matched. Modify debugging output.
23077 (curr_insn_transform): Remove best_small_class_operands_num.
23080 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
23082 * config/aarch64/aarch64-builtins.c
23083 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
23084 * config/aarch64/aarch64-simd-builtins.def
23085 (reduc_splus_): Add new modes.
23086 (reduc_uplus_): New.
23087 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
23088 (reduc_uplus_v4sf): Likewise.
23089 (reduc_splus_v4sf): Likewise.
23090 (aarch64_addv<mode>): Likewise.
23091 (reduc_uplus_<mode>): Likewise.
23092 (reduc_splus_<mode>): Likewise.
23093 (aarch64_addvv2di): Likewise.
23094 (reduc_uplus_v2di): Likewise.
23095 (reduc_splus_v2di): Likewise.
23096 (aarch64_addvv2si): Likewise.
23097 (reduc_uplus_v2si): Likewise.
23098 (reduc_splus_v2si): Likewise.
23099 (reduc_<sur>plus_<mode>): New.
23100 (reduc_<sur>plus_v2di): Likewise.
23101 (reduc_<sur>plus_v2si): Likewise.
23102 (reduc_<sur>plus_v4sf): Likewise.
23103 (aarch64_addpv4sf): Likewise.
23104 * config/aarch64/arm_neon.h
23105 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
23106 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
23107 add UNSPEC_SADDV, UNSPEC_UADDV.
23109 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
23111 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
23113 * config/aarch64/arm_neon.h
23114 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
23116 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
23118 * config/aarch64/aarch64-builtins
23119 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
23121 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
23123 * config/aarch64/aarch64-simd-builtins.def
23124 (reduc_smax_): New.
23125 (reduc_smin_): Likewise.
23126 (reduc_umax_): Likewise.
23127 (reduc_umin_): Likewise.
23128 (reduc_smax_nan_): Likewise.
23129 (reduc_smin_nan_): Likewise.
23132 (smax): Update for V2SF, V4SF and V2DF modes.
23135 (smin_nan): Likewise.
23136 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
23137 (<su><maxmin><mode>3): ...This, refactor.
23138 (s<maxmin><mode>3): New.
23139 (<maxmin_uns><mode>3): Likewise.
23140 (reduc_<maxmin_uns>_<mode>): Refactor.
23141 (reduc_<maxmin_uns>_v4sf): Likewise.
23142 (reduc_<maxmin_uns>_v2si): Likewise.
23143 (aarch64_<fmaxmin><mode>: Remove.
23144 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
23146 (vmin<q>_f<32,64>): Likewise.
23147 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
23149 (su): Add mappings for smax, smin, umax, umin.
23151 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
23152 (FMAXMIN): Rename as...
23153 (FMAXMIN_UNS): ...This.
23155 (fmaxminv): Likewise.
23156 (fmaxmin): Likewise.
23158 (maxmin_uns_op): Likewise.
23160 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
23162 * config/aarch64/arm_neon.h
23163 (vac<ge, gt><sd>_f<32, 64>): Rename to...
23164 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
23165 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
23167 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
23169 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
23170 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
23172 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
23174 * config/aarch64/aarch64-simd.md
23175 (vcond<mode>_internal): Handle special cases for constant masks.
23176 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
23177 (vcondu<mode><mode>): Likewise.
23178 (vcond<v_cmp_result><mode>): New.
23180 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
23182 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
23183 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
23184 * config/aarch64/aarch64-simd-builtins.def
23185 (cmeq): Update to BUILTIN_VALLDI.
23190 * config/aarch64/arm_neon.h
23191 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
23192 to builtins or C as appropriate.
23194 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
23196 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
23198 (cmhi): Rename to...
23200 * config/aarch64/aarch64-simd.md
23201 (simd_mode): Add SF.
23202 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
23203 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
23204 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
23205 (cstore<mode>_neg): ...This.
23206 * config/aarch64/iterators.md
23208 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
23209 (COMPARISONS): New.
23210 (UCOMPARISONS): Likewise.
23211 (optab): Add missing comparisons.
23217 (VCMP_S): Likewise.
23218 (VCMP_U): Likewise.
23219 (V_cmp_result): Add DF, SF modes.
23220 (v_cmp_result): Likewise.
23222 (vmtype): Likewise.
23223 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
23225 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
23227 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
23228 define_insn to define_insn_and_split.
23229 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
23230 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
23231 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
23232 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
23233 (thumb2_negscc): Likewise.
23235 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
23237 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
23239 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
23241 * config/arm/thumb2.md: Remove trailing whitespaces.
23243 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
23245 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
23246 Use gen_int_mode rather than GEN_INT.
23248 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
23250 * value-prof.c (stream_in_histogram_value): Remove the strayed
23253 2013-04-30 Richard Biener <rguenther@suse.de>
23255 PR middle-end/57122
23256 * cfghooks.c (split_edge): Properly check for the loop latch edge.
23258 2013-04-30 Richard Biener <rguenther@suse.de>
23260 PR middle-end/57107
23261 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
23263 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
23265 PR rtl-optimization/56957
23266 PR rtl-optimization/57105
23267 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
23268 variable. Use just INSN_UID for determining whether an insn
23269 should be only disconnected from the insn stream.
23270 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
23272 2013-04-30 Jakub Jelinek <jakub@redhat.com>
23274 PR tree-optimization/57104
23275 * tsan.c (instrument_expr): Don't instrument accesses to
23276 DECL_HARD_REGISTER VAR_DECLs.
23278 2013-04-30 Richard Biener <rguenther@suse.de>
23280 * function.h (loops_for_fn): New inline function.
23281 (set_loops_for_fn): Likewise.
23282 * cfgloop.h (place_new_loop): Add struct function parameter.
23283 (get_loop): Likewise.
23284 (get_loops): Likewise.
23285 (number_of_loops): Likewise.
23286 (fel_next): Adjust.
23287 (fel_init): Likewise.
23288 * cfg.c (get_loop_copy): Adjust.
23289 * cfgloop.c (flow_loops_dump): Likewise.
23290 (record_loop_exits): Likewise.
23291 (verify_loop_structure): Likewise.
23292 * cfgloopanal.c (mark_irreducible_loops): Likewise.
23293 (estimate_reg_pressure_cost): Likewise.
23294 (mark_loop_exit_edges): Likewise.
23295 * cfgloopmanip.c (place_new_loop): Likewise.
23296 (add_loop): Likewise.
23297 (duplicate_loop): Likewise.
23298 * graph.c (draw_cfg_nodes): Likewise.
23299 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
23300 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
23301 (extract_affine_chrec): Likewise.
23302 (build_scop_iteration_domain): Likewise.
23303 * graphite.c (graphite_initialize): Likewise.
23304 * ira-build.c (create_loop_tree_nodes): Likewise.
23305 (more_one_region_p): Likewise.
23306 (rebuild_regno_allocno_maps): Likewise.
23307 (mark_loops_for_removal): Likewise.
23308 (mark_all_loops_for_removal): Likewise.
23309 (remove_unnecessary_regions): Likewise.
23310 (ira_build): Likewise.
23311 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
23312 * loop-init.c (fix_loop_structure): Likewise.
23313 (gate_rtl_move_loop_invariants): Likewise.
23314 (gate_rtl_unswitch): Likewise.
23315 (gate_rtl_unroll_and_peel_loops): Likewise.
23316 (rtl_doloop): Likewise.
23317 * lto-streamer-in.c (input_cfg): Likewise.
23318 * lto-streamer-out.c (output_cfg): Likewise.
23319 * modulo-sched.c (sms_schedule): Likewise.
23320 * predict.c (tree_estimate_probability): Likewise.
23321 (tree_estimate_probability_driver): Likewise.
23322 (estimate_loops): Likewise.
23323 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
23324 (move_sese_region_to_fn): Likewise.
23325 (debug_loop_num): Likewise.
23326 * tree-chrec.c (chrec_evaluate): Likewise.
23327 (hide_evolution_in_other_loops_than_loop): Likewise.
23328 (chrec_component_in_loop_num): Likewise.
23329 (reset_evolution_in_loop): Likewise.
23330 (evolution_function_is_invariant_rec_p): Likewise.
23331 * tree-if-conv.c (main_tree_if_conversion): Likewise.
23332 * tree-inline.c (copy_loops): Likewise.
23333 (copy_cfg_body): Likewise.
23334 (tree_function_versioning): Likewise.
23335 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
23336 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
23338 (add_to_evolution_1): Likewise.
23339 (scev_const_prop): Likewise.
23340 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
23341 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
23342 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
23343 (tree_ssa_lim_initialize): Likewise.
23344 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
23345 (verify_loop_closed_ssa): Likewise.
23346 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
23347 (tree_ssa_loop_im): Likewise.
23348 (tree_ssa_loop_unswitch): Likewise.
23349 (tree_vectorize): Likewise.
23350 (check_data_deps): Likewise.
23351 (tree_ssa_loop_ivcanon): Likewise.
23352 (tree_ssa_loop_bounds): Likewise.
23353 (tree_complete_unroll): Likewise.
23354 (tree_complete_unroll_inner): Likewise.
23355 (tree_parallelize_loops): Likewise.
23356 (tree_ssa_loop_prefetch): Likewise.
23357 (tree_ssa_loop_ivopts): Likewise.
23358 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
23359 * tree-vectorizer.c (vectorize_loops): Likewise.
23361 2013-04-29 Mike Frysinger <vapier@gentoo.org>
23363 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
23364 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
23365 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
23366 with EABI_LINK_SPEC.
23368 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
23371 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
23374 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
23377 * lra-constraints.c (process_alt_operands): Discourage a bit more
23378 using memory for pseudos. Print cost dump for alternatives.
23379 Modify cost values for conflicts with early clobbers.
23380 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
23382 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
23385 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
23387 2013-04-29 Ian Bolton <ian.bolton@arm.com>
23389 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
23390 from/to S register.
23391 (movdi_aarch64): Support LDR/STR from/to D register.
23393 2013-04-29 Ian Bolton <ian.bolton@arm.com>
23395 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
23396 or higher by default.
23398 2013-04-29 Richard Biener <rguenther@suse.de>
23400 PR middle-end/57075
23401 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
23402 even if not adding abnormal edges for calls that can make
23405 2013-04-29 Richard Biener <rguenther@suse.de>
23407 PR middle-end/57103
23408 * tree-cfg.c (move_stmt_op): Fix condition under which to update
23410 (move_stmt_r): Remove redundant checking.
23412 2013-04-29 Teresa Johnson <tejohnson@google.com>
23415 * basic-block.h (apply_scale): New function.
23416 (apply_probability): Use apply_scale.
23417 * gimple-streamer-in.c (input_bb): Ditto.
23418 * lto-streamer-in.c (input_cfg): Ditto.
23419 * lto-cgraph.c (merge_profile_summaries): Ditto.
23420 * tree-optimize.c (execute_fixup_cfg): Ditto.
23421 * tree-inline.c (copy_bb): Update comment to use apply_scale.
23422 (copy_edges_for_bb): Ditto.
23423 (copy_cfg_body): Ditto.
23425 2013-04-29 Tom de Vries <tom@codesourcery.com>
23427 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
23428 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
23429 (tail_merge_optimize): Handle current_loops == NULL.
23431 2013-04-26 Jeff Law <law@redhat.com>
23433 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
23434 (simplify_cond_using_ranges): Generalize code to simplify
23435 COND_EXPRs where one argument is a constant and the other
23436 is an SSA_NAME created by an integral type conversion.
23438 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23440 * config/arm/arm.md (store_minmaxsi): Use only when
23441 optimize_insn_for_size_p.
23443 2013-04-29 Christian Bruel <christian.bruel@st.com>
23446 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
23448 2013-04-29 Richard Biener <rguenther@suse.de>
23450 PR middle-end/57089
23451 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
23452 loop tree make sure to schedule a fixup for the child as well.
23453 (expand_omp_for_generic): Properly add loops.
23454 (expand_omp_for_static_nochunk): Likewise.
23455 (expand_omp_for_static_chunk): Likewise.
23456 (expand_omp_for): For the degenerate case fixup loops.
23457 (expand_omp_sections): Fix default bb placement in loops.
23458 (expand_omp_atomic_pipeline): Properly add loops.
23460 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23462 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
23464 2013-04-29 Tom de Vries <tom@codesourcery.com>
23466 * tree-ssa-tail-merge.c: Update header comment.
23468 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
23470 * config/aarch64/arm_neon.h
23471 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
23472 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
23473 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
23474 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
23475 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
23476 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
23477 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
23478 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
23480 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
23482 * config/aarch64/aarch64-simd.md
23483 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
23484 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
23485 fix_trunc, fixuns_trunc.
23486 (ftrunc<VDQF:mode>2): New.
23487 * config/aarch64/iterators.md (optab): Add fix, fixuns.
23488 (fix_trunc_optab): New.
23490 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
23492 * config/aarch64/aarch64-builtins.c
23493 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
23494 iceilf, lround, iroundf.
23496 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
23499 * config/i386/i386.h (enum ix86_tune_indices)
23500 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
23501 New, split from X86_TUNE_INTER_UNIT_MOVES.
23502 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
23503 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
23504 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
23505 (TARGET_INTER_UNIT_MOVES): Remove.
23506 * config/i386/i386.c (initial_ix86_tune_features): Update.
23507 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
23508 (ix86_expand_convert_uns_didf_sse): Use
23509 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
23510 (ix86_expand_vector_init_one_nonzero): Ditto.
23511 (ix86_expand_vector_init_interleave): Ditto.
23512 (inline_secondary_memory_needed): Return true for moves from SSE class
23513 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
23514 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
23515 * config/i386/constraints.md (Yi, Ym): Depend on
23516 TARGET_INTER_UNIT_MOVES_TO_VEC.
23517 (Yj, Yn): New constraints.
23518 * config/i386/i386.md (*movdi_internal): Change constraints of
23519 operand 1 from Yi to Yj and from Ym to Yn.
23520 (*movsi_internal): Ditto.
23521 (*movdf_internal): Ditto.
23522 (*movsf_internal): Ditto.
23523 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
23524 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
23525 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
23526 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
23527 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
23528 * config/i386/sse.md (movdi_to_sse): Ditto.
23529 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
23530 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
23531 TARGET_INTER_UNIT_MOVES.
23532 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
23533 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
23534 instead of TARGET_INTER_UNIT_MOVES.
23535 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
23536 operand 1 from Yi to Yj and from Ym to Yn.
23538 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
23540 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
23541 (float_truncate_hi_): Likewise.
23542 (float_extend_lo_): Likewise.
23543 (float_truncate_lo_): Likewise.
23544 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
23545 (aarch64_float_extend_lo_v2df): Likewise.
23546 (vec_unpacks_hi_v4sf): Likewise.
23547 (aarch64_float_truncate_lo_v2sf): Likewise.
23548 (aarch64_float_truncate_hi_v4sf): Likewise.
23549 (vec_pack_trunc_v2df): Likewise.
23550 (vec_pack_trunc_df): Likewise.
23552 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
23554 * config/aarch64/aarch64-builtins.c
23555 (aarch64_fold_builtin): Fold float conversions.
23556 * config/aarch64/aarch64-simd-builtins.def
23557 (floatv2si, floatv4si, floatv2di): New.
23558 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
23559 * config/aarch64/aarch64-simd.md
23560 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
23561 * config/aarch64/iterators.md (FLOATUORS): New.
23562 (optab): Add float, floatuns.
23563 (su_optab): Likewise.
23565 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
23567 * config/aarch64/aarch64-builtins.c
23568 (aarch64_builtin_vectorized_function): Use new names for
23570 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
23571 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
23572 (fcvtzu): Split as...
23573 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
23574 (fcvtas): Split as...
23575 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
23576 (fcvtau): Split as...
23577 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
23578 (fcvtps): Split as...
23579 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
23580 (fcvtpu): Split as...
23581 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
23582 (fcvtms): Split as...
23583 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
23584 (fcvtmu): Split as...
23585 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
23586 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
23587 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
23588 (lfrintnusf, lfrintnudf): Likewise.
23589 * config/aarch64/aarch64-simd.md
23590 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
23592 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
23593 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
23594 (fcvt_pattern): Likewise.
23596 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
23598 * config/aarch64/aarch64-simd.md
23599 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
23600 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
23602 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
23604 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
23605 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
23606 (vrnd<a,m,n,p>_f32): Implement using builtins.
23607 (vrnd<i,x><q>_f<32, 64>): New.
23609 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
23611 * config/aarch64/aarch64-builtins.c
23612 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
23613 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
23614 (frintz): Rename to...
23616 (frintp): Rename to...
23618 (frintm): Rename to...
23620 (frinti): Rename to...
23621 (nearbyint): ...this.
23622 (frintx): Rename to...
23624 (frinta): Rename to...
23626 * config/aarch64/aarch64-simd.md
23627 (aarch64_frint<frint_suffix><mode>): Delete.
23628 (<frint_pattern><mode>2): Convert to insn.
23629 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
23630 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
23631 (frint_pattern): Likewise.
23632 (frint_suffix): Likewise.
23634 2013-04-29 Richard Biener <rguenther@suse.de>
23636 PR tree-optimization/57081
23637 * loop-init.c: Include tree-flow.h.
23638 (loop_optimizer_finalize): Free number of iteration estimates.
23639 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
23641 2013-04-29 Jakub Jelinek <jakub@redhat.com>
23643 PR tree-optimization/57083
23644 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
23645 non-singleton shift count range, zero extend low_bound for uns case.
23647 * config/i386/predicates.md (general_vector_operand): New predicate.
23648 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
23649 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
23650 if they aren't nonimmediate operands. If their original values
23651 satisfy const_vector_equal_evenodd_p, don't shift them.
23652 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
23653 predicates. For the SSE4.1 case force operands[{1,2}] into registers
23654 if not nonimmediate_operand.
23655 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
23656 instead of register_operand.
23657 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
23659 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
23661 * stor-layout.c (finalize_size_functions): Allocate a structure and
23662 reset cfun before dumping the functions.
23664 2013-04-27 Jakub Jelinek <jakub@redhat.com>
23666 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
23669 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
23670 use xop_pmacsdqh if uns_p.
23671 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
23672 the immediate rotate count.
23674 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
23676 * rtl.h (struct rtx_def): Add comment for field jump.
23677 (LRA_SUBREG_P): New macro.
23678 * recog.c (register_operand): Check LRA_SUBREG_P.
23679 * lra.c (lra): Add note at the end of RTL code. Align non-empty
23681 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
23682 (lra_final_code_change): Skip subreg change for operators.
23683 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
23684 if there are no operand changes.
23685 * lra-constraints.c (curr_insn_set): New.
23686 (match_reload): Set LRA_SUBREG_P.
23687 (emit_spill_move): Ditto.
23688 (check_and_process_move): Use curr_insn_set. Process only single
23689 set insns. Don't initialize sec_mem_p and change_p.
23690 (simplify_operand_subreg): Use LRA_SUBREG_P.
23691 (reg_in_class_p): New function.
23692 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
23693 of #ifdef. Add code to remove cycling.
23694 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
23695 non-null disp. Reload inner instead of disp when base and index
23696 are null. Try to put lo_sum into register.
23697 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
23698 (check_and_process_move): Move code for move cost check to
23699 simple_move_p. Remove equiv_substitution.
23700 (simple_move_p): New function.
23701 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
23702 curr_insn_set. Call check_and_process_move only for single set
23703 insns. Use the new function. Move call of check_and_process_move
23704 after operand equiv substitution and address process.
23706 2013-04-26 Jakub Jelinek <jakub@redhat.com>
23709 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
23710 with nonlocal goto receivers or returns twice calls, ignore
23711 unininitialized values from abnormal edges to nl goto receiver
23712 or returns twice call.
23714 2013-04-26 Jakub Jelinek <jakub@redhat.com>
23716 PR tree-optimization/57051
23717 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
23718 and VEC_RSHIFT_EXPR if shift count is a multiple of element
23721 2013-04-26 Richard Biener <rguenther@suse.de>
23723 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
23724 (expand_omp_taskreg): Likewise. Mark loops for fixup.
23725 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
23726 (fixup_loop_arrays_after_move): New function.
23727 (move_sese_region_to_fn): Properly outline the loop tree parts
23728 of the SESE region.
23730 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
23732 * config/i386/i386.md (type, unit): Fix long lines.
23734 2013-04-26 Richard Biener <rguenther@suse.de>
23736 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
23737 (lto-streamer-out.o): Likewise.
23738 * cfgloop.c (init_loops_structure): Export, add struct function
23739 argument and adjust.
23740 (flow_loops_find): Adjust.
23741 * cfgloop.h (enum loop_estimation): Add EST_LAST.
23742 (init_loops_structure): Declare.
23743 * lto-streamer-in.c: Include cfgloop.h.
23744 (input_cfg): Input the loop tree.
23745 * lto-streamer-out.c: Include cfgloop.h.
23746 (output_cfg): Output the loop tree.
23747 (output_struct_function_base): Do not drop PROP_loops.
23749 2013-03-26 Richard Biener <rguenther@suse.de>
23751 * tree-cfg.c (execute_build_cfg): Build the loop tree.
23752 (pass_build_cfg): Provide PROP_loops.
23753 (move_sese_region_to_fn): Remove loops that are outlined into fn
23755 * tree-inline.c: Include cfgloop.h.
23756 (initialize_cfun): Do not drop PROP_loops.
23757 (copy_loops): New function.
23758 (copy_cfg_body): Copy loop structure.
23759 (tree_function_versioning): Initialize destination loop tree.
23760 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
23761 (pass_parallelize_loops): Do IL verification.
23762 * loop-init.c (loop_optimizer_init): Fixup loops if required.
23763 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
23764 the CFG make sure we fixup loops as well.
23765 * tree-ssa-tail-merge.c: Include cfgloop.h.
23766 (replace_block_by): When merging loop latches mark loops for fixup.
23767 * lto-streamer-out.c (output_struct_function_base): Drop
23768 PROP_loops for now.
23769 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
23770 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
23771 * ipa-split.c: Include cfgloop.h.
23772 (split_function): Add the new return block to the loop tree root.
23773 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
23774 whether we have removed the forwarder block.
23775 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
23776 * cfgloop.h (place_new_loop): Declare.
23777 * cfgloopmanip.c (place_new_loop): Export.
23778 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
23779 (tree-switch-conversion.o): Likewise.
23780 (tree-complex.o): Likewise.
23781 (tree-inline.o): Likewise.
23782 (tree-ssa-tailmerge.o): Likewise.
23783 (ipa-split.o): Likewise.
23784 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
23785 (tree-ssa-copy.o): Likewise.
23786 * tree-switch-conversion.c: Include cfgloop.h
23787 (process_switch): If we emit a bit-test cascade, schedule loops
23789 * tree-complex.c: Include cfgloop.h.
23790 (expand_complex_div_wide): Properly add new basic-blocks to loops.
23791 * asan.c: Include cfgloop.h.
23792 (create_cond_insert_point): Properly add new basic-blocks to
23793 loops, schedule loop fixup.
23794 * cfgloop.c (verify_loop_structure): Check that looks are not
23796 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
23798 (expand_omp_for_generic): Likewise.
23799 (expand_omp_sections): Likewise.
23800 (expand_omp_atomic_pipeline): Schedule loops for fixup.
23801 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
23802 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
23803 is initialized, not when loops are present.
23804 * tree-parloops.c (parallelize_loops): Remove checking here.
23805 * passes.c (init_optimization_passes): Schedule a copy-propagation
23806 pass before complete unrolling of inner loops.
23808 2013-04-26 Jakub Jelinek <jakub@redhat.com>
23810 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
23811 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
23812 (colorize_init): Add argument to _WIN32 version.
23813 * toplev.c: Include diagnostic-color.h.
23814 (process_options): Default to -fdiagnostics-color=auto if
23815 GCC_COLORS env var is in the environment.
23816 * common.opt (fdiagnostics-color=): Add Var and Init.
23817 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
23818 env var is in the environment, the default is auto rather than never.
23820 * diagnostic.h (file_name_as_prefix): Add context argument.
23821 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
23822 the string as locus.
23823 * langhooks.c (lhd_print_error_function): Adjust caller.
23825 2013-04-25 Lawrence Crowl <crowl@google.com>
23827 * var-tracking.c (shared_hash_def::htab):
23828 Change type to hash_table. Update dependent calls and types.
23830 2013-04-25 Lawrence Crowl <crowl@google.com>
23832 * Makefile.in: Update as needed below.
23834 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
23835 Move declaration to after the type's method definitons.
23837 * attribs.c (htab_t scoped_attributes::attribute_hash):
23838 Change type to hash_table. Update dependent calls and types.
23840 * bitmap.c (htab_t bitmap_desc_hash):
23841 Change type to hash_table. Update dependent calls and types.
23843 * cselib.c (htab_t cselib_hash_table):
23844 Change type to hash_table. Update dependent calls and types.
23846 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
23847 (hash_string_slot_node): Move implementation into lto-streamer.h
23848 struct string_slot_hasher.
23849 (eq_string_slot_node): Likewise.
23851 * data-streamer-out.c: Update output_block::string_hash_table
23852 dependent calls and types.
23854 * dwarf2cfi.c (htab_t trace_index):
23855 Change type to hash_table. Update dependent calls and types.
23857 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
23858 Change type to hash_table. Update dependent calls and types.
23859 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
23860 (htab_t optimize_external_refs::map): Likewise.
23861 (htab_t output_comp_unit::extern_map): Likewise.
23862 (htab_t output_comdat_type_unit::extern_map): Likewise.
23863 (htab_t output_macinfo::macinfo_htab): Likewise.
23864 (htab_t optimize_location_lists::htab): Likewise.
23865 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
23867 * except.c (htab_t ehspec_hash_type):
23868 Change type to hash_table. Update dependent calls and types.
23869 (assign_filter_values::ttypes): Likewise.
23870 (assign_filter_values::ehspec): Likewise.
23871 (sjlj_assign_call_site_values::ar_hash): Likewise.
23872 (convert_to_eh_region_ranges::ar_hash): Likewise.
23874 * gcse.c (htab_t pre_ldst_table):
23875 Change type to hash_table. Update dependent calls and types.
23877 * ggc-common.c (htab_t saving_htab):
23878 Change type to hash_table. Update dependent calls and types.
23879 (htab_t loc_hash): Likewise.
23880 (htab_t ptr_hash): Likewise.
23881 (call_count): Rename ggc_call_count.
23882 (call_alloc): Rename ggc_call_alloc.
23883 (loc_descriptor): Rename make_loc_descriptor.
23884 (add_statistics): Rename ggc_add_statistics.
23886 * ggc-common.c (saving_htab):
23887 Change type to hash_table. Update dependent calls and types.
23889 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
23890 (push_gimplify_context): Likewise.
23891 (pop_gimplify_context): Likewise.
23892 (struct gimple_temp_hash_elt): Added.
23893 (struct gimplify_hasher): Likewise.
23894 (struct gimplify_ctx.temp_htab):
23895 Change type to hash_table. Update dependent calls and types.
23897 * gimple-fold.c: Include gimplify-ctx.h.
23899 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
23900 Change type to hash_table. Update dependent calls and types.
23901 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
23902 avoid potential global name collision.
23904 * gimplify.c: Include gimplify-ctx.h.
23905 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
23906 (htab_t gimplify_ctx::temp_htab):
23907 Update dependent calls and types for new type hash_table.
23908 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
23909 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
23911 * gimplify-ctx.h: New.
23912 (struct gimple_temp_hash_elt): Move from gimplify.c.
23913 (class gimplify_hasher): New.
23914 (struct gimplify_ctx): Move from gimple.h.
23915 (htab_t gimplify_ctx::temp_htab):
23916 Change type to hash_table. Update dependent calls and types.
23918 * graphite-clast-to-gimple.c: Include graphite-htab.h.
23919 (htab_t ivs_params::newivs_index):
23920 Change type to hash_table. Update dependent calls and types.
23921 (htab_t ivs_params::params_index): Likewise.
23922 (htab_t print_generated_program::params_index): Likewise.
23923 (htab_t gloog::newivs_index): Likewise.
23924 (htab_t gloog::params_index): Likewise.
23926 * graphite.c: Include graphite-htab.h.
23927 4htab_t graphite_transform_loops::bb_pbb_mapping):
23928 Change type to hash_table. Update dependent calls and types.
23930 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
23931 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
23932 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
23934 * graphite-dependences.c: Include graphite-htab.h.
23935 (loop_is_parallel_p): Change hash table type of parameter.
23937 * graphite-htab.h: New.
23938 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
23939 (extern find_pbb_via_hash): Move from graphite-poly.h.
23940 (extern loop_is_parallel_p): Move from graphite-poly.h.
23941 (extern get_loop_body_pbbs): Move from graphite-poly.h.
23943 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
23944 (extern loop_is_parallel_p): Move to graphite-htab.h.
23945 (extern get_loop_body_pbbs): Move to graphite-htab.h.
23947 * haifa-sched.c (htab_t delay_htab):
23948 Change type to hash_table. Update dependent calls and types.
23949 (htab_t delay_htab_i2): Likewise.
23951 * ira-color.c (htab_t allocno_hard_regs_htab):
23952 Change type to hash_table. Update dependent calls and types.
23954 * ira-costs.c (htab_t cost_classes_htab):
23955 Change type to hash_table. Update dependent calls and types.
23957 * loop-invariant.c (htab_t merge_identical_invariants::eq):
23958 Change type to hash_table. Update dependent calls and types.
23960 * loop-iv.c (htab_t bivs):
23961 Change type to hash_table. Update dependent calls and types.
23963 * loop-unroll.c (htab_t opt_info::insns_to_split):
23964 Change type to hash_table. Update dependent calls and types.
23965 (htab_t opt_info::insns_with_var_to_expand): Likewise.
23967 * lto-streamer.h (struct string_slot): Move from data-streamer.h
23968 (struct string_slot_hasher): New.
23969 (htab_t output_block::string_hash_table):
23970 Change type to hash_table. Update dependent calls and types.
23972 * lto-streamer-in.c (freeing_string_slot_hasher): New.
23973 (htab_t file_name_hash_table):
23974 Change type to hash_table. Update dependent calls and types.
23976 * lto-streamer-out.c: Update output_block::string_hash_table dependent
23979 * lto-streamer.c (htab_t tree_htab):
23980 Change type to hash_table. Update dependent calls and types.
23982 * omp-low.c: Include gimplify-ctx.h.
23984 * passes.c (htab_t name_to_pass_map):
23985 Change type to hash_table. Update dependent calls and types.
23986 (pass_traverse): Rename to passes_pass_traverse.
23988 * plugin.c (htab_t event_tab):
23989 Change type to hash_table. Update dependent calls and types.
23991 * postreload-gcse.c (htab_t expr_table):
23992 Change type to hash_table. Update dependent calls and types.
23993 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
23995 * sese.c (debug_rename_map_1): Make extern.
23996 (htab_t copy_bb_and_scalar_dependences::rename_map):
23997 Change type to hash_table. Update dependent calls and types.
23999 * sese.h (extern debug_rename_map): Move to .c file.
24001 * store-motion.c (htab_t store_motion_mems_table):
24002 Change type to hash_table. Update dependent calls and types.
24004 * trans-mem.c (htab_t tm_new_mem_hash):
24005 Change type to hash_table. Update dependent calls and types.
24007 * tree-browser.c (htab_t TB_up_ht):
24008 Change type to hash_table. Update dependent calls and types.
24010 * tree-cfg.c (htab_t discriminator_per_locus):
24011 Change type to hash_table. Update dependent calls and types.
24013 * tree-complex.c: Include tree-hasher.h
24014 (htab_t complex_variable_components):
24015 Change type to hash_table. Update dependent calls and types.
24017 * tree-eh.c (htab_t finally_tree):
24018 Change type to hash_table. Update dependent calls and types.
24020 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
24021 struct int_tree_hasher.
24022 (extern int_tree_map_eq): Likewise.
24023 (uid_decl_map_hash): Removed.
24024 (extern decl_tree_map_eq): Likewise.
24026 * tree-hasher.h: New.
24027 (struct int_tree_hasher): New.
24028 (typedef int_tree_htab_type): New.
24030 * tree-inline.c: Include gimplify-ctx.h.
24032 * tree-mudflap.c: Include gimplify-ctx.h.
24034 * tree-parloops.c: Include tree-hasher.h.
24035 (htab_t eliminate_local_variables_stmt::decl_address):
24036 Change type to hash_table. Update dependent calls and types.
24037 (htab_t separate_decls_in_region::decl_copies): Likewise.
24039 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
24040 Change type to hash_table. Update dependent calls and types.
24042 * tree-sra.c (candidates):
24043 Change type to hash_table. Update dependent calls and types.
24045 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
24047 (int_tree_map_hash): Likewise.
24049 * tree-ssa-dom.c (htab_t avail_exprs):
24050 Change type to hash_table. Update dependent calls and types.
24052 * tree-ssa-live.c (var_map_base_init::tree_to_index):
24053 Change type to hash_table. Update dependent calls and types.
24055 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
24056 Change type to hash_table. Update dependent calls and types.
24058 * tree-ssa-phiopt.c (seen_ssa_names):
24059 Change type to hash_table. Update dependent calls and types.
24061 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
24062 Change type to hash_table. Update dependent calls and types.
24064 * tree-ssa-uncprop.c (equiv):
24065 Change type to hash_table. Update dependent calls and types.
24067 2013-04-25 Jakub Jelinek <jakub@redhat.com>
24069 PR rtl-optimization/57003
24070 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
24071 call note_stores with kill_clobbered_value callback again after
24072 killing regs_invalidated_by_call.
24074 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
24076 * config/aarch64/aarch64-simd.md
24077 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
24078 (aarch64_simd_bsl<mode>): Likewise.
24079 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
24081 2013-04-25 Marek Polacek <polacek@redhat.com>
24083 PR tree-optimization/57066
24084 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
24086 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
24088 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
24090 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
24092 * config/aarch64/aarch64-builtins.c
24093 (aarch64_fold_builtin): New.
24094 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
24095 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
24096 * config/aarch64/aarch64-simd-builtins.def (abs): New.
24097 * config/aarch64/arm_neon.h
24098 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
24100 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
24101 Tejas Belagod <tejas.belagod@arm.com>
24103 * config/aarch64/aarch64-builtins.c
24104 (aarch64_gimple_fold_builtin): New.
24105 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
24106 * config/aarch64/aarch64-simd-builtins.def (addv): New.
24107 * config/aarch64/aarch64-simd.md (addpv4sf): New.
24108 (addvv4sf): Update.
24109 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
24111 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
24113 * config/aarch64/aarch64.md
24114 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
24116 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
24118 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
24119 (*ngcsi_uxtw): New pattern.
24121 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24122 Julian Brown <julian@codesourcery.com>
24124 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
24125 (TB_DREG): Add T_V4HF.
24126 (v4hf_UP): New macro.
24127 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
24128 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
24129 Handle initialisation of V4HF. Adjust initialisation of reinterpret
24131 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
24132 (arm_vector_mode_supported_p): Handle V4HF.
24133 (arm_mangle_map): Handle V4HFmode.
24134 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
24135 * config/arm/arm_neon_builtins.def: Add entries for
24136 vcvtv4hfv4sf, vcvtv4sfv4hf.
24137 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
24138 (neon_vcvtv4hfv4sf): Likewise.
24139 * config/arm/neon-gen.ml: Handle half-precision floating point
24141 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
24142 * config/arm/arm_neon.h: Regenerate.
24143 * config/arm/neon.ml (type elts): Add F16.
24144 (type vectype): Add T_float16x4, T_floatHF.
24145 (type vecmode): Add V4HF.
24146 (type features): Add Requires_FP_bit feature.
24147 (elt_width): Handle F16.
24148 (elt_class): Likewise.
24149 (elt_of_class_width): Likewise.
24150 (mode_of_elt): Refactor.
24151 (type_for_elt): Handle F16, fix error messages.
24152 (vectype_size): Handle T_float16x4.
24153 (vcvt_sh): New function.
24154 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
24155 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
24156 (string_of_mode): Handle V4HF.
24157 * doc/arm-neon-intrinsics.texi: Regenerate.
24159 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
24161 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
24162 format specifier in 'X' case.
24164 2013-04-25 Alan Modra <amodra@gmail.com>
24167 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
24168 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
24169 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
24170 Repeat for many other rotate/shift and mask patterns using subregs.
24171 Name lshiftrt insns.
24172 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
24173 on WORDS_BIG_ENDIAN.
24175 2013-04-25 Alan Modra <amodra@gmail.com>
24177 * config.gcc: Support little-endian powerpc-linux targets.
24178 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
24179 (LINK_OS_LINUX_SPEC): Define.
24180 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
24181 Preserve MASK_LITTLE_ENDIAN.
24182 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
24183 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
24184 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
24185 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
24186 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
24187 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
24188 Correct fp word order for little-endian. Don't shift toc entries
24189 smaller than a word for little-endian.
24190 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
24191 (bswapdi2 splits): Correct low-part subreg for little-endian.
24192 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
24193 low/high where such is correct only for be.
24194 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
24195 little-endian for -mcall-aixdesc.
24197 2013-04-25 Alan Modra <amodra@gmail.com>
24199 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
24200 replace_equiv_address_nv.
24202 2013-04-25 Alan Modra <amodra@gmail.com>
24204 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
24206 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
24209 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
24210 * rtl.h (struct rtx_def): ...
24212 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
24214 PR rtl-optimizations/57046
24215 * lra-constraints (split_reg): Set up lra_risky_transformations_p
24216 for multi-reg splits.
24218 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
24220 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
24222 2013-04-24 Sterling Augustine <saugustine@google.com>
24224 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
24225 (comp_dir_string, debug_str_dwo_section): New.
24226 (DEBUG_STR_DWO_SECTION): Rename to ...
24227 (DEBUG_DWO_STR_SECTION): ... this.
24228 (DEBUG_NORM_STR_SECTION): Delete.
24229 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
24230 (DEBUG_STR_DWO_SECTION_FLAGS): New.
24231 (find_AT_string): Move most logic to ...
24232 (find_AT_string_in_table): ... here. New.
24233 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
24234 add_skeleton_AT_string. Delete logic.
24235 (output_skeleton_debug_sections): Remove call to
24236 add_top_level_skeleton_die_attrs.
24237 (add_comp_dir_attribute): Move logic to comp_dir_string.
24238 (dwarf2out_init): Initialize debug_str_dwo_section.
24239 (output_indirect_string): Call find_string_form.
24240 (output_indirect_strings): Rewrite.
24241 (prune_unused_types): Empty skeleton_debug_str_hash.
24242 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
24243 (dwarf2out_finish): Call output_indirect_strings.
24245 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
24247 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
24249 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
24251 * rtl.h (struct rtx_def): Add comment for field jump.
24252 (LRA_SUBREG_P): New macro.
24253 * recog.c (register_operand): Check LRA_SUBREG_P.
24254 * lra.c (lra): Add note at the end of RTL code. Align non-empty
24256 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
24257 (lra_final_code_change): Skip subreg change for operators.
24258 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
24259 if there are no operand changes.
24260 * lra-constraints.c (curr_insn_set): New.
24261 (match_reload): Set LRA_SUBREG_P.
24262 (emit_spill_move): Ditto.
24263 (check_and_process_move): Use curr_insn_set. Process only single
24264 set insns. Don't initialize sec_mem_p and change_p.
24265 (simplify_operand_subreg): Use LRA_SUBREG_P.
24266 (reg_in_class_p): New function.
24267 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
24268 of #ifdef. Add code to remove cycling.
24269 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
24270 non-null disp. Reload inner instead of disp when base and index
24271 are null. Try to put lo_sum into register.
24272 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
24273 (check_and_process_move): Move code for move cost check to
24274 simple_move_p. Remove equiv_substitution.
24275 (simple_move_p): New function.
24276 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
24277 curr_insn_set. Call check_and_process_move only for single set
24278 insns. Use the new function. Move call of check_and_process_move
24279 after operand equiv substitution and address process.
24281 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
24283 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
24284 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
24285 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
24287 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
24289 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
24291 2013-04-24 Marek Polacek <polacek@redhat.com>
24293 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
24294 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
24295 (select_loops_exit_conditions): Likewise.
24296 (number_of_iterations_for_all_loops): Likewise.
24297 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
24298 (scev_analysis): Likewise.
24300 2013-04-02 Catherine Moore <clm@codesourcery.com>
24301 Chao-ying Fu <fu@mips.com>
24303 * config/mips/micromips.md (jraddiusp): New pattern.
24304 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
24305 instruction if possible.
24307 2013-04-24 Alan Modra <amodra@gmail.com>
24309 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
24311 2013-04-24 Julian Brown <julian@codesourcery.com>
24312 Chung-Lin Tang <cltang@codesourcery.com>
24314 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
24315 dependency behavior in enumeration type DIE generation. Add TODO note
24316 to comments about future DW_FORM_sdata/udata re-work of related code.
24318 2013-04-23 Lawrence Crowl <crowl@google.com>
24320 * Makefile.in: Update as needed below.
24322 * hash-table.h (class hash_table):
24323 Correct many methods with parameter types compare_type to the correct
24324 value_type. (Correct code was unlikely to notice the change.)
24325 (hash_table::elements_with_deleted) New.
24326 (class hashtable::iterator): New.
24327 (hashtable::begin()): New.
24328 (hashtable::end()): New.
24329 (FOR_EACH_HASH_TABLE_ELEMENT): New.
24331 * statistics.c (statistics_hashes):
24332 Change type to hash_table. Update dependent calls and types.
24334 * tree-into-ssa.c (var_infos):
24335 Change type to hash_table. Update dependent calls and types.
24337 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
24338 Change type to hash_table. Update dependent calls and types.
24340 * tree-ssa-loop-im.c (struct mem_ref.refs):
24341 Change type to hash_table. Update dependent calls and types.
24343 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
24344 Change type to hash_table. Update dependent calls and types.
24346 * tree-ssa-sccvn.c (vn_tables_s::nary):
24347 Change type to hash_table. Update dependent calls and types.
24348 (vn_tables_s::phis): Likewise.
24349 (vn_tables_s::references): Likewise.
24351 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
24352 (vn_reference_eq): Update parameter and return types.
24354 * tree-ssa-structalias.c (pointer_equiv_class_table):
24355 Change type to hash_table. Update dependent calls and types.
24356 (location_equiv_class_table): Likewise.
24358 * tree-vect-data-refs.c: Consequential changes for making
24359 peeling a hash_table.
24361 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
24362 (destroy_loop_vec_info): Dependent hash_table update.
24364 * tree-vectorizer.h (peeling_htab):
24365 Change type to hash_table. Update dependent calls and types.
24367 2013-04-23 Shiva Chen <shiva0217@gmail.com>
24369 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
24370 to check the register content is equal or not.
24371 * lra-constraints.c (match_reload): Use lra_assign_reg_val
24372 to assign register content record.
24373 * lra-eliminations.c (update_reg_eliminate): Use
24374 lra_update_reg_val_offset to update register content offset.
24375 * lra-int.h (struct lra_reg): Add offset member.
24376 (lra_reg_val_equal_p): New static inline function.
24377 (lra_update_reg_val_offset): New static inline function.
24378 (lra_assign_reg_val): New static inline function.
24379 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
24380 to assign register content record.
24381 (initialize_lra_reg_info_element): Initial offset to zero.
24383 2013-04-23 Catherine Moore <clm@codesourcery.com>
24385 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
24386 operands. Record compression.
24388 2013-04-23 Xinliang David Li <davidxl@google.com>
24390 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
24392 2013-04-23 Richard Biener <rguenther@suse.de>
24394 PR middle-end/57036
24395 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
24396 parameter, only add abnormal goto edges from the copied body
24397 if the call could perform abnormal gotos.
24398 (copy_cfg_body): Adjust.
24400 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
24402 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
24404 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
24406 * coretypes.h (gimple_stmt_iterator): Add struct to make
24409 2013-04-23 Richard Biener <rguenther@suse.de>
24411 PR tree-optimization/57026
24412 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
24413 from SSA names occuring in abnormal PHI nodes.
24415 2013-04-22 Andi Kleen <ak@linux.intel.com>
24417 * lto/lto.c (print_lto_report_1): Fix LTO report names.
24419 2013-04-22 Andi Kleen <ak@linux.intel.com>
24421 * lto/lto.c (print_lto_report_1): Declare early.
24422 (read_cgraph_and_symbols): Call print_lto_report_1 early.
24424 2013-04-22 Andi Kleen <ak@linux.intel.com>
24426 * common.opt (-flto-report-wpa): Add.
24427 * doc/invoke.texi (-flto-report-wpa): Add.
24428 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
24431 2013-04-22 Xinliang David Li <davidxl@google.com>
24433 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
24434 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
24435 * Makefile.in: New dependency
24437 David Daney <ddaney.cavm@gmail.com>
24439 * configure.ac (gcc_cv_as_micromips_support): Use the
24440 --fatal-warnings option.
24441 * configure: Regenerate.
24443 2013-04-22 Marek Polacek <polacek@redhat.com>
24446 * tsan.c (instrument_expr): Don't instrument expression
24447 in case its size is zero.
24449 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
24453 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
24455 * config/alpha/alpha.c (TARGET_LRA_P): New define.
24457 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
24459 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
24460 (gimple_stmt_iterator): New typedef.
24461 * gimple.h (gimple_stmt_iterator): Rename to...
24462 (gimple_stmt_iterator_d): ... This.
24463 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
24464 trees be valid for GIMPLE and GENERIC.
24465 (TARGET_GIMPLE_FOLD_BUILTIN): New.
24466 * gimple-fold.c (gimple_fold_call): Call target hook
24467 gimple_fold_builtin.
24468 * hooks.c (hook_bool_gsiptr_false): New.
24469 * hooks.h (hook_bool_gsiptr_false): New.
24470 * target.def (fold_stmt): New.
24471 * doc/tm.texi: Regenerate.
24473 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
24476 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
24477 a set sp if no stack realignment.
24479 2013-04-22 Nick Clifton <nickc@redhat.com>
24481 * config.gcc (tilegx-linux): Extend extra_objs rather than
24483 (tilepro-linux): Likewise.
24485 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
24487 * config/aarch64/aarch64-builtins.c
24489 (CF0, CF1, CF2, CF3, CF4, CF10): New.
24490 (VAR<1-12>): Add MAP parameter.
24491 (BUILTIN_*): Likewise.
24492 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
24493 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
24494 (aarch64_ushl_n<mode>): Likewise.
24495 (aarch64_sshr_n<mode>): Likewise.
24496 (aarch64_ushr_n<mode>): Likewise.
24497 (aarch64_<maxmin><mode>): Likewise.
24498 (aarch64_sqrt<mode>): Likewise.
24499 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
24500 (vshr<q>_n_*): Likewise.
24502 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
24504 * config/aarch64/aarch64-builtins.c
24505 (aarch64_simd_builtin_type_mode): Handle SF types.
24507 (BUILTIN_GPF): Define.
24508 (aarch64_init_simd_builtins): Handle SF types.
24509 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
24510 (frecps): Likewise.
24511 (frecpx): Likewise.
24512 * config/aarch64/aarch64-simd.md
24513 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
24514 (aarch64_frecpe<mode>): New.
24515 (aarch64_frecps<mode>): Likewise.
24516 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
24517 (v8type): Add frecp<esx>.
24518 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
24519 (aarch64_frecps<mode>): Likewise.
24520 * config/aarch64/iterators.md (FRECP): New.
24521 (frecp_suffix): Likewise.
24522 * config/aarch64/arm_neon.h
24523 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
24525 2013-04-22 Christian Bruel <christian.bruel@st.com>
24528 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
24529 (REG_CLASS_NAMES): Idem.
24530 (REG_CLASS_CONTENTS): Idem.
24531 (REGCLASS_HAS_FP_REG): Idem.
24532 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
24533 (sh_conditional_register_usage): Idem.
24535 2013-04-21 Jeff Law <law@redhat.com>
24537 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
24538 (ssa_forward_propagate_and_combine): Use it.
24540 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
24542 * lra.c: Update the flow chart diagram.
24544 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
24546 PR rtl-optimization/56847
24547 * lra-constraints.c (process_alt_operands): Discourage alternative
24548 with non-matche doffsettable memory constraint fro memory with
24551 2013-04-19 Richard Biener <rguenther@suse.de>
24553 PR tree-optimization/56982
24554 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
24556 * gimplify.c (gimplify_call_expr): Notice special calls.
24557 (gimplify_modify_expr): Likewise.
24558 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
24559 abnormal control flow receivers.
24560 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
24561 in the same way as cfun->has_nonlocal_labels.
24562 (gimple_purge_dead_abnormal_call_edges): Likewise.
24563 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
24564 receivers start a basic-block.
24566 2013-04-19 Richard Biener <rguenther@suse.de>
24568 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
24570 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
24571 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
24572 (SLP_TREE_LOAD_PERMUTATION): Add.
24573 (vect_transform_slp_perm_load): Adjust prototype.
24574 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
24575 (vect_free_slp_instance): Likewise.
24576 (vect_create_new_slp_node): Likewise.
24577 (vect_supported_slp_permutation_p): Remove.
24578 (vect_slp_rearrange_stmts): Adjust.
24579 (vect_supported_load_permutation_p): Likewise. Inline
24580 vect_supported_slp_permutation_p here.
24581 (vect_analyze_slp_instance): Compute load permutations per
24582 slp node instead of per instance.
24583 (vect_get_slp_defs): Adjust.
24584 (vect_transform_slp_perm_load): Likewise.
24585 (vect_schedule_slp_instance): Remove redundant code.
24586 (vect_schedule_slp): Remove hack for PR56270, add it ...
24587 * tree-vect-stmts.c (vectorizable_load): ... here, do not
24588 CSE loads for SLP. Adjust.
24590 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
24592 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
24593 spelling in two comments.
24595 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
24598 * config/arm/arm.c (load_multiple_sequence): Require SP
24599 as base register for loads if SP is in the register list.
24601 2013-04-19 Martin Jambor <mjambor@suse.cz>
24603 PR tree-optimization/56718
24604 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
24605 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
24606 and made public. Adjusted all callers.
24607 (ipa_intraprocedural_devirtualization): New function.
24608 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
24609 (ipa_intraprocedural_devirtualization): Likewise.
24610 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
24612 2013-04-19 Richard Biener <rguenther@suse.de>
24614 PR tree-optimization/57000
24615 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
24617 2013-04-19 Terry Guo <terry.guo@arm.com>
24619 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
24621 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
24622 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
24623 (cortex_m4_fmacs): Use new reservations.
24624 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
24626 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
24628 PR rtl-optimization/56999
24629 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
24631 (lra_coalesce): Remove split_origin_bitmap and related code.
24632 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
24633 ranges if necessary.
24635 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
24637 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
24639 (ix86_expand_call): Remove clobbered_registers array and use
24640 x86_64_ms_sysv_extra_clobbered_registers instead.
24641 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
24643 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
24645 * config/i386/i386.md (*call_rex64_ms_sysv): Use
24646 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
24647 (*call_value_rex64_ms_sysv): Ditto.
24649 2013-04-18 Cary Coutant <ccoutant@google.com>
24651 * dwarf2out.c (output_pubnames): Check die_perennial_p of
24652 parent instead of die_mark.
24654 2013-04-18 Diego Novillo <dnovillo@google.com>
24656 * gimple.c (create_gimple_tmp): New.
24657 (get_expr_type): New.
24658 (build_assign): New.
24659 (build_type_cast): New.
24660 * gimple.h (enum ssa_mode): Define.
24661 (gimple_seq_set_location): New.
24662 * asan.c (build_check_stmt): Change some gimple_build_* calls
24663 to use build_assign and build_type_cast.
24665 2013-04-18 Richard Biener <rguenther@suse.de>
24667 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
24668 handle negative step. Remove redundant checks.
24669 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
24670 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
24671 for negative step and grouped loads fail to vectorize.
24673 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
24675 * emit-rtl.c (reset_insn_used_flags): New function.
24676 (reset_all_used_flags): Use it.
24677 (verify_insn_sharing): New function.
24678 (verify_rtl_sharing): Fix verification for SEQUENCEs.
24680 2013-04-18 Jakub Jelinek <jakub@redhat.com>
24682 PR tree-optimization/56984
24683 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
24684 and (x >> M) >= N don't register any assertion if N << M is the
24687 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
24689 * lower-subreg.c (resolve_simple_move): If called self-recursive,
24690 do not delete_insn insns that have not yet been emitted, only
24691 unlink them with remove_insn.
24692 * df-scan.c (df_insn_delete): Revert r197492.
24694 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
24696 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
24697 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
24699 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
24701 * config/arm/arm.md (movsicc_insn): Convert define_insn into
24702 define_insn_and_split.
24703 (and_scc,ior_scc,negscc): Likewise.
24704 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
24706 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
24708 * config/arm/arm.c (use_return_insn): Return 0 for targets that
24709 can benefit from using a sequence of LDRD instructions in epilogue
24710 instead of a single LDM instruction.
24712 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
24715 * doc/extend.texi: Fix typo.
24717 2013-04-17 Richard Biener <rguenther@suse.de>
24719 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
24720 (vect_build_slp_tree): ... here.
24721 (vect_build_slp_tree_1): Compute which stmts of the SLP group
24722 match. Remove special-casing of mismatched complex loads.
24723 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
24724 re-try the match with swapped commutative operands.
24725 (vect_supported_load_permutation_p): Remove special-casing of
24726 mismatched complex loads.
24727 (vect_analyze_slp_instance): Adjust.
24729 2013-04-17 Richard Biener <rguenther@suse.de>
24731 PR rtl-optimization/56921
24732 * cfgloop.h (struct loop): Add simple_loop_desc member.
24733 (struct niter_desc): Mark with GTY(()).
24734 (simple_loop_desc): Do not use aux field but simple_loop_desc.
24735 * loop-iv.c (get_simple_loop_desc): Likewise.
24736 (free_simple_loop_desc): Likewise.
24739 2013-04-16 Richard Biener <rguenther@suse.de>
24741 PR rtl-optimization/56921
24742 * loop-init.c (pass_rtl_move_loop_invariants): Add
24743 TODO_do_not_ggc_collect to todo_flags_finish.
24744 (pass_rtl_unswitch): Same.
24745 (pass_rtl_unroll_and_peel_loops): Same.
24746 (pass_rtl_doloop): Same.
24748 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
24750 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
24751 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
24752 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
24754 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
24755 * tree-streamer.c (record_common_node): Adjust reference in comment.
24757 2013-04-17 Terry Guo <terry.guo@arm.com>
24759 * config/arm/cortex-m4.md: Add a new bypass.
24761 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
24763 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
24765 (*subs_<optab><mode>_multp2): New pattern.
24766 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
24767 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
24769 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
24771 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
24772 (*subs_mul_imm_<mode>): New pattern.
24774 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
24777 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
24778 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
24779 (vsx_movti_32bit): Change j->wa to O->wa.
24781 2013-04-16 Richard Biener <rguenther@suse.de>
24783 PR rtl-optimization/56921
24784 * loop-init.c (pass_rtl_move_loop_invariants): Add
24785 TODO_do_not_ggc_collect to todo_flags_finish.
24786 (pass_rtl_unswitch): Same.
24787 (pass_rtl_unroll_and_peel_loops): Same.
24788 (pass_rtl_doloop): Same.
24790 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
24792 * config/arm/arm.c (emit_multi_reg_push): New declaration
24793 for an existing function.
24794 (arm_emit_strd_push): New function.
24795 (arm_expand_prologue): Used here.
24796 (arm_emit_ldrd_pop): New function.
24797 (arm_expand_epilogue): Used here.
24798 (arm_get_frame_offsets): Update condition.
24799 (arm_emit_multi_reg_pop): Add a special case for load of a single
24800 register with writeback.
24802 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
24804 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
24807 2013-04-16 Richard Biener <rguenther@suse.de>
24809 PR tree-optimization/56756
24810 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
24811 (first_mem_ref_loc): New.
24812 (execute_sm): Place the load temporarily before a previous
24813 access instead of in the latch edge to ensure its SSA dependencies
24814 are defined at points dominating the load.
24816 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
24818 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
24819 correct fix by moving header and footer insn to the footer of
24820 the merged basic block. Clear BB_END of the merged-away block.
24822 PR middle-end/43631
24823 * emit-rtl.c (make_note_raw): New function.
24824 (link_insn_into_chain): New static inline function.
24825 (add_insn): Use it.
24826 (add_insn_before, add_insn_after): Factor insn chain linking code...
24827 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
24828 using link_insn_into_chain.
24829 (note_outside_basic_block_p): New helper function for emit_note_after
24830 and emit_note_before.
24831 (emit_note_after): Use nobb variant of add_insn_after if the note
24832 should not be contained in a basic block.
24833 (emit_note_before): Use nobb variant of add_insn_before if the note
24834 should not be contained in a basic block.
24835 (emit_note_copy): Use make_note_raw.
24836 (emit_note): Likewise.
24837 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
24838 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
24839 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
24840 the moved barrier the tail of the basic block it follows.
24841 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
24843 2013-04-15 Jakub Jelinek <jakub@redhat.com>
24845 PR tree-optimization/56962
24846 * gimple-ssa-strength-reduction.c (record_increment): Only set
24847 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
24848 either rhs1 or rhs2 is equal to c->base_expr.
24850 2013-04-15 Richard Biener <rguenther@suse.de>
24852 PR tree-optimization/56933
24853 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
24855 (GROUP_READ_WRITE_DEPENDENCE): Remove.
24856 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
24857 * tree-vect-data-refs.c (vect_analyze_group_access): Move
24858 dependence check ...
24859 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
24861 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
24862 GROUP_READ_WRITE_DEPENDENCE.
24864 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
24866 * emit-rtl.c (reset_all_used_flags): New function.
24867 (verify_rtl_sharing): Call reset_all_used_flags before and after
24868 performing the checks.
24870 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24872 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
24873 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
24874 * config/arm/constraints.md (De): New constraint.
24875 * config/arm/neon.md (anddi3_neon): Delete.
24876 (neon_vand<mode>): Expand to standard anddi3 pattern.
24877 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
24878 Move earlier in the file.
24879 (neon_inv_logic_op2): Likewise.
24880 (arm_anddi_operand_neon): New predicate.
24882 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
24884 * configure.ac (gcc_cv_ld_as_needed): Set
24885 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
24886 Use -z ignore, -z record on *-*-solaris2*.
24887 (HAVE_LD_AS_NEEDED): Update comment.
24888 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
24889 * configure: Regenerate.
24890 * config.in: Regenerate.
24891 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
24892 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
24893 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
24894 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
24895 equivalents. Fix markup.
24896 * doc/tm.texi: Regenerate.
24898 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
24900 * config/i386/i386.opt: New option mstack-protector-guard=.
24901 * config/i386/i386-opts.h: Add enum stack_protector_guard.
24902 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
24903 TARGET_SSP_TLS_GUARD.
24904 * config/i386/i386.c (ix86_option_override_internal): Set
24905 ix86_stack_protector_guard.
24906 * config/i386/i386.md (stack_protect_set): Enable for
24907 TARGET_SSP_TLS_GUARD only.
24908 (stack_protect_set_<mode>): Ditto.
24909 (stack_protect_test): Ditto.
24910 (stack_protect_test_<mode>): Ditto.
24911 * doc/invoke.texi (i386 Option): Document.
24913 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
24916 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
24917 (S_MODES): Set H_MODE bit.
24918 (SF_MODES): Set only S_MODE and SF_MODE bits.
24919 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
24920 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
24921 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
24922 <MODE_FLOAT>: Likewise.
24924 2013-04-15 Joey Ye <joey.ye@arm.com>
24926 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
24928 2013-04-15 Joey Ye <joey.ye@arm.com>
24930 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
24932 (thumb_far_jump_used_p): Count instruction size and set
24935 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
24937 * reorg.c (fill_simple_delay_slots): Reindent block of code.
24938 * resource.c (mark_target_live_regs): Reformat conditional block.
24940 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
24942 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
24943 notes, they are emitted only just before final.
24944 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
24946 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
24948 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
24949 * cfgrtl.c (delete_insn): Call it here instead.
24950 * lra-spills.c (lra_final_code_change): Use delete_insn.
24951 * haifa-sched.c (sched_remove_insn): Likewise.
24952 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
24953 returning to the nop pool.
24954 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
24955 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
24957 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
24959 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
24960 * doc/tm.texi: Regenerated.
24962 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
24964 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
24967 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
24969 * df-core.c (df_find_def): Compare register numbers.
24970 (df_find_use): Likewise.
24972 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
24975 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
24976 lra_in_progress for return.
24978 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
24980 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
24981 define_insn into define_insn_and_split and emit movsicc patterns.
24983 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
24985 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
24987 2013-04-12 Richard Biener <rguenther@suse.de>
24989 * tree-pass.h (TODO_do_not_ggc_collect): New.
24990 * passes.c (execute_one_ipa_transform_pass): Honor
24991 TODO_do_not_ggc_collect.
24992 (execute_one_pass): Likewise.
24995 2013-04-10 Richard Biener <rguenther@suse.de>
24997 * passes.c (init_optimization_passes): Remove reload pass.
24998 * ira.c (do_reload): Merge into ...
25000 (rest_of_handle_reload): Remove.
25001 (pass_reload): Likewise.
25002 * config/i386/i386.c (ix86_option_override): Refer to ira instead
25003 of reload for vzeroupper pass placement.
25005 2013-04-12 Jakub Jelinek <jakub@redhat.com>
25007 PR tree-optimization/56918
25008 PR tree-optimization/56920
25009 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
25010 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
25011 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
25012 use wide_mul_with_sign method.
25014 2013-04-12 Richard Biener <rguenther@suse.de>
25016 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
25017 not be considered a gimple constant.
25019 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
25021 * fold-const.c (const_binop): Handle vector shifts by a scalar.
25022 (fold_binary_loc): Call const_binop also for mixed vector-scalar
25025 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
25026 Jakub Jelinek <jakub@redhat.com>
25028 * opts.c: Include diagnostic-color.h.
25029 (common_handle_option): Handle OPT_fdiagnostics_color_.
25030 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
25031 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
25032 (diagnostic-color.o): New.
25033 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
25034 (diagnostic_color_rule): New enum.
25035 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
25036 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
25037 the location string.
25038 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
25039 either NULL, or color kind.
25040 * diagnostic-color.c: New file.
25041 * diagnostic-color.h: New file.
25042 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
25044 * doc/invoke.texi (-fdiagnostics-color): Document.
25045 * pretty-print.h (pp_show_color): Define.
25046 (struct pretty_print_info): Add show_color field.
25047 * diagnostic.c: Include diagnostic-color.h.
25048 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
25049 macros. Colorize error:, warning: etc. strings and also the location
25051 (diagnostic_show_locus): Colorize the caret line.
25052 * pretty-print.c: Include diagnostic-color.h.
25053 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
25054 inside of %< %> quotes or quoted through q format modifier.
25056 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
25058 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
25060 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25062 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
25063 code in CC_NZ mode.
25064 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
25067 2013-04-11 Marek Polacek <polacek@redhat.com>
25069 PR tree-optimization/48184
25070 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
25072 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
25074 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
25075 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
25076 (skip_simple_arithmetic): Tidy up.
25077 * tree.h (skip_simple_constant_arithmetic): Declare.
25079 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25081 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
25083 2013-04-11 Richard Biener <rguenther@suse.de>
25085 * tree-vect-loop.c (get_initial_def_for_induction): Properly
25086 generate vector constants.
25088 2013-04-11 Richard Biener <rguenther@suse.de>
25090 PR tree-optimization/56878
25091 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
25092 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
25094 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
25095 Prefer to align the DR with the most invariant base address.
25097 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
25099 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
25102 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
25104 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
25105 floating-point vector comparisons against 0.
25107 2013-04-11 Jakub Jelinek <jakub@redhat.com>
25109 PR tree-optimization/56899
25110 * fold-const.c (extract_muldiv_1): Apply distributive law
25111 only if TYPE_OVERFLOW_WRAPS (ctype).
25113 2013-04-11 Bin Cheng <bin.cheng@arm.com>
25116 * ira-costs.c (scan_one_insn): Check whether the source rtx of
25117 loading has side effect.
25119 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
25121 * config/sparc/sparc.c: Include tree-pass.h.
25122 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
25123 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
25124 head of file. Change return type. Split off gate function.
25125 (sparc_gate_work_around_errata): New function.
25126 (pass_work_around_errata): New pass definition.
25127 (insert_pass_work_around_errata) New pass insert definition to
25128 insert pass_work_around_errata just after delayed-branch scheduling.
25129 (sparc_option_override): Insert the pass.
25130 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
25132 2013-04-10 David S. Miller <davem@davemloft.net>
25134 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
25135 or -mcpu=hypersparc.
25137 * target.def (cstore_mode): New hook.
25138 * target.h: Include insn-codes.h
25139 * targhooks.c: Likewise.
25140 (default_cstore_mode): New function.
25141 * targhooks.h: Declare it.
25142 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
25143 * doc/tm.texi: Rebuild.
25144 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
25145 target hook, rather than inspecting the insn_data.
25146 * config/sparc/sparc.c (sparc_cstore_mode): New function.
25147 (TARGET_CSTORE_MODE): Redefine.
25148 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
25150 * config/sparc/predicates.md (cstore_result_operand): New special
25152 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
25153 Use it for operand 0.
25154 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
25155 (*snesi_special): Likewise.
25156 (*snesi_zero): Likewise.
25157 (*seqsi_zero): Likewise.
25158 (*sltu_insn): Likewise.
25159 (*sgeu_insn): Likewise.
25160 (*seqdi_special): Make operand 0 and comparison operation be of
25162 (*snedi_special): Likewise.
25163 (*snedi_special_vis3): Likewise.
25164 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
25165 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
25166 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
25167 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
25168 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
25169 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
25170 (*sltu_extend_sp64): Likewise.
25171 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
25172 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
25173 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
25174 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
25175 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
25177 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
25179 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
25180 (aarch64_start_file): Use the new function.
25182 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
25183 Jason Merrill <jason@redhat.com>
25185 * common.opt: Add -gdwarf.
25186 * opts.c (common_handle_option): Handle it.
25187 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
25189 2013-04-10 Richard Biener <rguenther@suse.de>
25191 * passes.c (execute_todo): Do not call ggc_collect conditional here.
25192 (execute_one_ipa_transform_pass): But unconditionally here.
25193 (execute_one_pass): And here.
25194 (init_optimization_passes): Remove reload pass.
25195 * tree-pass.h (TODO_ggc_collect): Remove.
25196 (pass_reload): Likewise.
25197 * ira.c (do_reload): Merge into ...
25199 (rest_of_handle_reload): Remove.
25200 (pass_reload): Likewise.
25201 * config/i386/i386.c (ix86_option_override): Refer to ira instead
25202 of reload for vzeroupper pass placement.
25203 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
25204 and todo_flags_finish of all passes.
25206 2013-04-10 Richard Biener <rguenther@suse.de>
25208 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
25209 first_const_oprnd field, rename first_def_type to first_op_type.
25210 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
25211 (vect_get_and_check_slp_defs): Always use the type of the
25212 operand. Allow mixed vect_external_def, vect_constant_def types.
25213 (vect_get_constant_vectors): Handle mixed vect_external_def,
25214 vect_constant_def types.
25216 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
25218 PR tree-optimization/55524
25219 * tree-ssa-math-opts.c
25220 (convert_mult_to_fma): Don't use an fms construct
25221 when we don't have an fms operation, but fnma, and it looks
25222 likely that we'll be able to use the latter.
25224 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
25226 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
25228 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
25229 inline fail caused by overwritable functions.
25231 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
25233 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
25234 unnecessary bits in the constant power of two case.
25236 2013-04-10 Richard Biener <rguenther@suse.de>
25238 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
25239 broken code swapping operands.
25240 (vect_build_slp_tree): Do not compute load permutations here.
25241 (vect_analyze_slp_instance): Compute load permutations here,
25242 after building the SLP tree.
25244 2013-04-09 Christian Bruel <christian.bruel@st.com>
25246 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
25247 of next/prev_real_insn.
25249 2013-04-09 Jan Hubicka <jh@suse.cz>
25251 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
25252 Drop aliased parameter.
25253 (function_and_variable_visibility): Do not handle alias pairs.
25254 * cgraph.c (varpool_externally_visible_p): Update prototype.
25255 * varpool.c (varpool_add_new_variable): Update.
25257 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
25259 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
25261 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
25263 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
25265 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
25267 2013-04-09 Marek Polacek <polacek@redhat.com>
25269 PR tree-optimization/48762
25270 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
25272 2013-04-09 Richard Biener <rguenther@suse.de>
25274 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
25276 (vect_build_slp_tree): Likewise.
25277 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
25278 calculating the cost of a SLP instance.
25279 (vect_analyze_slp_instance): Use it from here, after building
25282 2013-04-09 Jakub Jelinek <jakub@redhat.com>
25284 PR middle-end/56883
25285 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
25286 expand_omp_for_static_chunk): Use simple_p = true in
25287 force_gimple_operand_gsi calls when assigning to addressable decls.
25289 2013-04-09 Jeff Law <law@redhat.com>
25291 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
25292 when the boolean was created by converting a wider object which
25293 had a boolean range.
25295 2013-04-09 Richard Biener <rguenther@suse.de>
25297 * tree-vectorizer.h (slp_void_p): Remove.
25298 (slp_tree): Typedef before _slp_tree declaration.
25299 (struct _slp_tree): Use a vector of slp_tree as children.
25300 (vect_get_place_in_interleaving_chain): Remove.
25301 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
25303 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
25305 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
25306 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
25307 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
25308 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
25309 Use slp_node instead of slp_void_p and adjust.
25311 2013-04-09 Richard Biener <rguenther@suse.de>
25313 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
25314 work that is not necessary.
25316 2013-04-09 Jakub Jelinek <jakub@redhat.com>
25318 PR tree-optimization/56854
25319 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
25320 forward into clobber stmts if it would change MEM_REF lhs into
25323 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
25325 * tree.c (type_hash_lookup, type_hash_add): Make static.
25326 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
25328 2013-04-09 Richard Biener <rguenther@suse.de>
25330 * tree.h (unsave_expr_now): Remove.
25331 * tree-inline.c (mark_local_for_remap_r): Remove.
25332 (unsave_expr_1): Likewise.
25333 (unsave_r): Likewise.
25334 (unsave_expr_now): Likewise.
25335 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
25336 (propagate_tree_value): Likewise.
25338 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
25340 * doc/rtl.texi (sequence): Rewrite documentation to match the
25341 current use of SEQUENCE rtl objects.
25342 * rtl.def (SEQUENCE): Likewise.
25344 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
25345 Update documentation.
25346 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
25347 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
25349 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
25351 2013-04-08 Teresa Johnson <tejohnson@google.com>
25353 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
25354 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
25356 (estimate_edge_size_and_time): Add comment to suggest using rounding
25358 (estimate_node_size_and_time): Ditto.
25359 (remap_edge_change_prob): Use helper rounding divide methods.
25360 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
25361 (gimple_mod_pow2_value_transform): Ditto.
25362 (gimple_mod_subtract_transform): Ditto.
25363 (gimple_ic_transform): Ditto.
25364 (gimple_stringops_transform): Ditto.
25365 * stmt.c (conditional_probability): Ditto.
25366 (emit_case_dispatch_table): Ditto.
25367 * lto-cgraph.c (merge_profile_summaries): Ditto.
25368 * tree-optimize.c (execute_fixup_cfg): Ditto.
25369 * cfgcleanup.c (try_forward_edges): Ditto.
25370 * cfgloopmanip.c (scale_loop_profile): Ditto.
25372 (duplicate_loop_to_header_edge): Ditto.
25373 (lv_adjust_loop_entry_edge): Ditto.
25374 * tree-vect-loop.c (vect_transform_loop): Ditto.
25375 * profile.c (compute_branch_probabilities): Ditto.
25376 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
25377 * lto-streamer-in.c (input_cfg): Ditto.
25378 * gimple-streamer-in.c (input_bb): Ditto.
25379 * ipa-cp.c (update_profiling_info): Ditto.
25380 (update_specialized_profile): Ditto.
25381 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
25382 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
25384 * sched-rgn.c (compute_dom_prob_ps): Ditto.
25385 (compute_trg_info): Ditto.
25386 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
25387 (purge_dead_edges): Ditto.
25388 * loop-unswitch.c (unswitch_loop): Ditto.
25389 * cgraphclones.c (cgraph_clone_edge): Ditto.
25390 (cgraph_clone_node): Ditto.
25391 * tree-inline.c (copy_bb): Ditto.
25392 (copy_edges_for_bb): Ditto.
25393 (initialize_cfun): Ditto.
25394 (copy_cfg_body): Ditto.
25395 (expand_call_inline): Ditto.
25397 2013-04-08 Kai Tietz <ktietz@redhat.com>
25399 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
25400 TARGET_CYGWIN64 by TARGET_64BIT.
25402 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
25404 * config/epiphany/epiphany.md (GPR_1): New constant.
25405 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
25406 * config/epiphany/epiphany.c (gen_compare_reg):
25407 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
25408 is already in place.
25409 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
25410 Don't require being called during rtl expansion; If y operlaps r0,
25412 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
25413 (epiphany_expand_epilogue): Likewise.
25415 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
25416 Don't use CC_FPmode for ORDERED / UNORDERED.
25417 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
25419 * config/epiphany/constraints.md (CnL): New constraint.
25420 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
25421 * config/epiphany/predicates.md (add_operand): Allow 1024.
25423 * config/epiphany/epiphany.md (logical_op): New code iterator.
25424 (op_mnc): New code attribute.
25425 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
25426 (mov_f+1, mov_f+2): New peephole2 patterns.
25428 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
25429 (cstoresi4): Also allow re-use of zero result when doing a NE
25430 comparison to a non-zero operand.
25431 Use (clobber (scratch)) for first insn if the gpr output is not needed.
25433 * config/epiphany/epiphany.md (<insn_opname>v2si3):
25434 Use gen_addsi3_i / gen_subsi3_i.
25436 2013-04-08 Jakub Jelinek <jakub@redhat.com>
25440 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
25441 contain anything but clobbers, at most one __builtin_stack_restore,
25442 optionally debug stmts and final resx, and if it has at least one
25443 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
25444 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
25445 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
25446 which isn't defaut definition, remove them.
25447 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
25448 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
25449 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
25450 with MEM_REF LHS with SSA_NAME address.
25452 2013-04-08 Jeff Law <law@redhat.com>
25454 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
25456 2013-04-08 Richard Biener <rguenther@suse.de>
25458 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
25460 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
25461 determined vector type.
25462 (vect_analyze_data_refs): Likewise.
25463 (vect_get_new_vect_var): Adjust.
25464 (vect_create_destination_var): Preserve SSA name versions.
25465 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
25466 not dump anything here.
25468 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
25470 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
25471 Add member lr_slot_known.
25472 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
25474 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
25475 Remove code that sets lr_slot_offset according to what a previous
25476 version of epiphany_emit_save_restore used to do.
25477 (epiphany_emit_save_restore): When doing an lr save or restore,
25478 set/verify lr_slot_known and lr_slot_offset.
25480 2013-04-08 Xinyu Qi <xyqi@marvell.com>
25483 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
25486 2013-04-08 Richard Biener <rguenther@suse.de>
25488 * alias.c (find_base_term): Fix thinko in previous change.
25490 2013-04-08 Jakub Jelinek <jakub@redhat.com>
25492 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
25493 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
25494 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
25495 if possible to compute val.
25496 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
25497 For QImode integers don't require anything about precision. Use
25498 const_with_all_bytes_same to find out if the constant doesn't have
25499 repeated bytes in it.
25501 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
25503 * config/s390/s390.c (s390_expand_insv): Only accept insertions
25506 2013-04-08 Marek Polacek <polacek@redhat.com>
25508 PR rtl-optimization/48182
25509 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
25512 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25515 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
25516 nuses, make sure we have a label.
25518 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
25521 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
25522 (rs6000_emit_swdiv_low_precision): Remove.
25523 (rs6000_emit_swdiv): Rewrite to handle between one and four
25524 iterations of Newton-Raphson generally; modify required number of
25525 iterations for some cases.
25526 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
25528 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
25530 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
25531 set-but-unused variable.
25533 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
25534 basic blocks of released function bodies garbage-collectable.
25536 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
25537 (struct rtl_opt_pass): Add TODO_df_finish.
25539 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
25541 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
25543 * config/arm/constraints.md (q): New constraint.
25544 * config/arm/ldrdstrd.md: New file.
25545 * config/arm/arm.md (ldrdstrd.md) New include.
25546 (arm_movdi): Use "q" instead of "r" constraint
25547 for double-word memory access.
25548 (movdf_soft_insn): Likewise.
25549 * config/arm/vfp.md (movdi_vfp): Likewise.
25550 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
25551 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
25552 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
25553 (mem_ok_for_ldrd_strd): Likewise.
25554 (output_move_double): Update assertion.
25556 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
25558 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
25560 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
25562 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
25563 define_insn_and_split.
25564 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
25566 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
25568 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
25569 define_insn_and_split.
25570 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
25571 (shiftsi3_compare): New pattern.
25572 (rrx): New pattern.
25573 * config/arm/unspecs.md (UNSPEC_RRX): New.
25575 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
25577 * config/arm/arm.md (negdi_extendsidi): New pattern.
25578 (negdi_zero_extendsidi): Likewise.
25580 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
25582 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
25583 define_insn_and_split.
25584 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
25585 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
25587 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
25589 * config/arm/arm.md (arm_subdi3): Convert define_insn into
25590 define_insn_and_split.
25591 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
25592 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
25594 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
25596 * config/arm/arm.md (subsi3_carryin): New pattern.
25597 (subsi3_carryin_const): Likewise.
25598 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
25599 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
25601 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
25603 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
25605 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
25607 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
25608 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
25610 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
25612 * config/arm/arm.c (arm_expand_builtin): Change fcode
25613 type to unsigned int.
25615 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
25617 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
25619 2013-04-04 Ian Lance Taylor <iant@google.com>
25621 * doc/standards.texi (Standards): The Go frontend supports the Go 1
25624 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
25626 PR middle-end/56729
25627 * df-scan.c (df_insn_delete): Disable failing assert.
25629 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
25631 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
25632 New function prototype.
25633 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
25634 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
25635 (arm_builtin_vectorized_function): New function.
25637 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
25639 * config/arm/arm_neon_builtins.def: New file.
25640 * config/arm/arm.c (neon_builtin_data): Move contents to
25641 arm_neon_builtins.def.
25642 (enum arm_builtins): Include neon builtin definitions.
25643 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
25644 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
25646 2013-04-04 Marek Polacek <polacek@redhat.com>
25648 PR tree-optimization/48186
25649 * predict.c (maybe_hot_frequency_p): Return false if
25650 HOT_BB_FREQUENCY_FRACTION is 0.
25651 (cgraph_maybe_hot_edge_p): Likewise.
25653 2013-04-04 Richard Biener <rguenther@suse.de>
25655 PR tree-optimization/56826
25656 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
25659 2013-04-04 Richard Biener <rguenther@suse.de>
25661 PR tree-optimization/56213
25662 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
25663 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
25665 2013-04-04 Richard Biener <rguenther@suse.de>
25667 PR tree-optimization/56837
25668 * tree-loop-distribution.c (classify_partition): For non-zero
25669 values require that the value has the same precision as its
25670 mode to be useful as memset value.
25672 2013-04-03 Nick Clifton <nickc@redhat.com>
25674 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
25675 (fmssf4): Use fmsf.s on E3V5 architectures.
25676 (fnmasf4): Use fnmaf.s on E3V5 architectures.
25677 (fnmssf4): Use fnmsf.s on E3V5 architectures.
25679 2013-04-03 Jeff Law <law@redhat.com>
25681 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
25682 (lra-eliminations.o): Likewise.
25684 2013-04-03 Teresa Johnson <tejohnson@google.com>
25686 * gcov-io.c (compute_working_sets): Moved most of body of old
25687 compute_working_sets here from profile.c.
25688 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
25689 (gcov_working_set_t): Moved typedef here from basic-block.h
25690 (compute_working_set): Declare.
25691 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
25692 (get_working_sets): Renamed from compute_working_set,
25693 replace most of body with call to new compute_working_sets.
25694 (get_exec_counts): Replace call to compute_working_sets
25695 to get_working_sets.
25696 * profile.h (get_working_sets): Renamed from compute_working_set.
25697 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
25698 to get_working_sets.
25699 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
25700 * gcov-dump.c (dump_working_sets): New function.
25702 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
25704 * hwint.c (sext_hwi, zext_hwi): New functions.
25705 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
25706 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
25707 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
25708 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
25709 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
25710 (sext_hwi, zext_hwi): New functions.
25712 2013-04-03 Jeff Law <law@redhat.com>
25714 PR tree-optimization/56799
25715 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
25716 back test for widening conversion erroneously dropped in prior change.
25718 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
25721 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
25722 instead of next_real_insn.
25724 2013-04-03 Marek Polacek <polacek@redhat.com>
25727 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
25729 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
25732 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
25734 (thumb1_output_casesi): Likewise.
25735 (thumb2_output_casesi): Likewise.
25737 2013-04-03 Richard Biener <rguenther@suse.de>
25739 PR tree-optimization/56817
25740 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
25742 (tree_unroll_loops_completely_1): ... new function to manually
25743 walk the loop tree, properly defering outer loops of unrolled
25744 loops to later iterations.
25746 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
25748 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
25749 (vectorizable_load): Likewise.
25750 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
25751 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
25753 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
25755 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
25758 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
25760 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
25762 2013-04-03 Bin Cheng <bin.cheng@arm.com>
25764 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
25766 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
25768 PR tree-optimization/56790
25769 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
25772 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
25774 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
25776 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
25777 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
25780 2013-04-03 Jakub Jelinek <jakub@redhat.com>
25783 * tree.h (force_folding_builtin_constant_p): New decl.
25784 * builtins.c (force_folding_builtin_constant_p): New variable.
25785 (fold_builtin_constant_p): Fold immediately also if
25786 force_folding_builtin_constant_p.
25788 2013-04-03 Richard Biener <rguenther@suse.de>
25790 PR tree-optimization/56812
25791 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
25792 DRs of the same interleaving chain are independent.
25794 2013-04-02 Jason Merrill <jason@redhat.com>
25796 * gdbinit.in (pbb): Use debug fn.
25798 2013-04-02 Lawrence Crowl <crowl@google.com>
25800 * sese.h (struct ivtype_map_elt_s): Remove unused.
25801 (extern debug_ivtype_map): Remove unused.
25802 (extern eq_ivtype_map_elts): Remove unused.
25803 * sese.c (debug_ivtype_map): Removed unused.
25804 (debug_ivtype_map_1): Removed unused.
25805 (debug_ivtype_elt): Remove unused.
25806 (eq_ivtype_map_elts): Remove unused.
25808 2013-04-02 Kai Tietz <ktietz@redhat.com>
25811 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
25812 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
25813 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
25815 (legitimize_pe_coff_symbol): Likewise.
25816 (is_imported_p): New helper-function.
25817 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
25818 for Windows x64 targets.
25819 (ix86_expand_prologue): Optimize for pe-coff targets.
25820 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
25821 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
25822 medium/large code-model.
25823 (legitimize_pic_address): Likewise.
25824 (legitimize_tls_address): Likewise.
25825 (ix86_expand_call): Likewise.
25826 (x86_output_mi_thunk): Likewise.
25827 (get_dllimport_decl): Add new beimport argument.
25828 (construct_plt_address): Don't assert for x64 pe-coff targets.
25829 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
25831 (SYMBOL_FLAG_STUBVAR): New macro.
25832 (SYMBOL_REF_STUBVAR_P): Likewise.
25833 * config/i386/winnt.c (stub_list): New structure.
25834 (stub_head): New local variable.
25835 (i386_pe_record_stub): New function.
25836 (i386_pe_file_end): Emit refptr-stubs.
25838 2013-04-02 Jakub Jelinek <jakub@redhat.com>
25840 PR rtl-optimization/56745
25841 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
25842 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
25845 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
25846 and both of them are MEM_REFs, just compare first argument for
25847 equality and attempt to deal even with differing offsets.
25850 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
25851 of gimple_clobber_p to be MEM_REF.
25852 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
25853 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
25854 after gimplification.
25855 * asan.c (get_mem_ref_of_assignment): Don't instrument
25856 gimple_clobber_p stmts.
25857 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
25858 gimple_clobber_p stmt if they have MEM_REF lhs and
25859 are dead because of another gimple_clobber_p stmt.
25860 * tree-ssa-live.c (clear_unused_block_pointer): Treat
25861 gimple_clobber_p stmts like debug stmts.
25862 (remove_unused_locals): Remove clobbers with MEM_REF lhs
25863 that refer to unused VAR_DECLs or uninitialized values.
25864 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
25865 gimple_clobber_p stmts if they refer to removed parameters.
25866 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
25869 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
25871 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
25872 using SWI48 mode attribute.
25874 2013-04-02 Wei Mi <wmi@google.com>
25876 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
25877 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
25878 *<rotate_insn><mode>3_mask in i386.md.
25880 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
25882 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
25884 2013-04-02 Richard Biener <rguenther@suse.de>
25886 PR tree-optimization/56778
25887 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
25888 Runtime alias tests are not supported for gather loads.
25889 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
25890 stmts referenced from SSA operands before updating SSA form.
25892 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
25893 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
25895 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
25896 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
25897 * config/arm/cortex-a53.md: New file.
25898 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
25899 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
25900 * config/arm/arm.c (arm_issue_rate): Likewise.
25901 * config/arm/arm-tune.md: Regenerate
25902 * config/arm/arm-tables.opt: Regenerate.
25903 * config/arm/arm-cores.def: Add cortex-a53.
25905 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
25907 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
25910 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
25912 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
25913 scalar load/store operations using B/H registers.
25914 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
25916 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
25918 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
25920 * config/aarch64/aarch64.c
25921 (aarch64_simd_scalar_immediate_valid_for_move): New.
25922 * config/aarch64/aarch64-protos.h
25923 (aarch64_simd_scalar_immediate_valid_for_move): New.
25924 * config/aarch64/constraints.md (Dh, Dq): New.
25925 * config/aarch64/iterators.md (hq): New.
25927 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
25929 * reorg.c (get_branch_condition): Deal with conditional returns.
25930 (fill_simple_delay_slots): Remove dead code dealing with jumps.
25932 2013-04-01 Wei Mi <wmi@google.com>
25934 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
25935 Truncate operand 2 using %b asm operand modifier.
25936 (*<shift_insn><mode>3_mask): Ditto.
25937 (*<rotate_insn><mode>3_mask): Ditto.
25939 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
25941 PR middle-end/56798
25942 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
25944 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
25946 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
25948 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
25950 2013-03-30 Lawrence Crowl <crowl@google.com>
25952 * dse.c (clear_alias_sets): Remove never set.
25953 (disqualified_clear_alias_sets): Remove never set.
25954 (clear_alias_mode_pool): Remove never set.
25955 (dse_step0): Remove condition that is never true.
25956 (canon_address): Remove condition that is never true.
25957 (dse_step7): Remove condition that is never true.
25958 (rest_of_handle_dse): Remove condition that is never true.
25959 (rest_of_handle_dse::did_global): Remove never read from above.
25960 (dse_step2_spill): Remove never called from above.
25961 (dse_step5_spill): Remove never called from above.
25963 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
25965 * doc/md.texi (Standard Names) <casesi>: Update documentation for
25966 JUMP_TABLE_DATA changes.
25967 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
25968 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
25969 (Insns) <jump_table_data>: New entry.
25970 * doc/tm.texi: Regenerate.
25972 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
25974 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
25975 for table jump at the end of a basic block using tablejump_p.
25976 * targhooks.c (default_invalid_within_doloop): Likewise.
25977 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
25978 target hook implementation that is identical to the default hook.
25979 (rs6000_invalid_within_doloop): Remove.
25981 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
25982 unused variable from tablejump_p call.
25984 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
25985 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
25986 (INSN_DELETED_P): Likewise.
25987 (emit_jump_table_data): New prototype.
25988 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
25989 after 4th as unused.
25990 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
25991 * sched-vis.c (print_insn): Likewise.
25992 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
25993 insn for compatibility with back ends that use next_active_insn to
25994 identify jump table data.
25995 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
25996 (remove_insn): Likewise.
25997 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
25999 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
26000 (emit_jump_table_data): New function.
26002 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
26003 basic block, a JUMP_TABLE_DATA never is.
26004 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
26005 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
26006 off from code handling real insns.
26007 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
26008 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
26009 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
26010 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
26011 is not a NONDEBUG_INSN_P.
26012 * ira-costs.c (scan_one_insn): Likewise.
26013 * jump.c (mark_all_labels): Likewise.
26014 (mark_jump_label_1): Likewise.
26015 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
26016 * lra.c (get_insn_freq): Expect all insns reaching here to be in
26018 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
26019 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
26020 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
26021 JUMP_TABLE_DATA_P insns.
26022 (calculate_elim_costs_all_insns): Likewise.
26023 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
26024 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
26025 (delete_output_reload): Code style fixups.
26026 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
26027 insn flags on this non-insn.
26028 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
26029 as scheduling barriers, for pre-change compatibility.
26030 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
26031 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
26033 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
26034 redundant JUMP_TABLE_DATA_P test.
26035 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
26036 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
26037 (frv_for_each_packet): Likewise.
26038 * config/i386/i386.c (min_insn_size): Likewise.
26039 (ix86_avoid_jump_mispredicts): Likewise.
26040 * config/m32r/m32r.c (m32r_is_insn): Likewise.
26041 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
26042 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
26043 (mips16_insn_length): Robustify.
26044 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
26045 (mips16_split_long_branches): Likewise.
26046 * config/pa/pa.c (pa_combine_instructions): Likewise.
26047 * config/rs6000/rs6000.c (get_next_active_insn): Treat
26048 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
26049 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
26050 as contributing to pool range lengths.
26051 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
26052 Remove redundant JUMP_TABLE_DATA_P test.
26053 (sh_loop_align): Likewise.
26054 (split_branches): Likewise.
26055 (sh_insn_length_adjustment): Likewise.
26056 * config/spu/spu.c (get_branch_target): Likewise.
26058 2013-03-29 Jan Hubicka <jh@suse.cz>
26060 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
26061 gcov streaming; stream hot bb threshold to ltrans.
26062 * predict.c (get_hot_bb_threshold): Break out from ....
26063 (maybe_hot_count_p): ... here.
26064 (set_hot_bb_threshold): New function.
26065 * lto-section-in.c (lto_section_name): Add profile.
26066 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
26067 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
26068 and data-streamer.h
26069 (histogram_entry): New structure.
26070 (histogram, histogram_pool): New global vars.
26071 (histogram_hash): New structure.
26072 (histogram_hash::hash): New method.
26073 (histogram_hash::equal): Likewise.
26074 (account_time_size): New function.
26075 (cmp_counts): New function.
26076 (dump_histogram): New function.
26077 (ipa_profile_generate_summary): New function.
26078 (ipa_profile_write_summary): New function.
26079 (ipa_profile_read_summary): New function.
26080 (ipa_profile): Decide on threshold.
26081 (pass_ipa_profile): Add ipa_profile_write_summary and
26082 ipa_profile_read_summary.
26083 * Makefile.in (ipa.o): Update dependencies.
26084 * lto-streamer.h (LTO_section_ipa_profile): New section.
26086 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
26088 * tree.h (VAR_P): New.
26090 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
26093 * doc/invoke.texi ([-fwhole-program]): Fix typo.
26095 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
26097 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
26098 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
26099 (control_flow_insn_p): Likewise.
26100 * cfgrtl.c (duplicate_insn_chain): Likewise.
26101 * final.c (get_attr_length_1): Likewise.
26102 (shorten_branches): Likewise.
26103 (final_scan_insn): Likewise.
26104 * function.c (instantiate_virtual_regs): Likewise.
26105 * gcse.c (insert_insn_end_basic_block): Likewise.
26106 * ira-costs.c (scan_one_insn): Likewise.
26107 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
26108 * lra.c (check_rtl): Likewise.
26109 * reload1.c (elimination_costs_in_insn): Likewise.
26110 * reorg.c (follow_jumps): Likewise.
26112 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
26113 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
26114 (thumb_far_jump_used_p): Likewise.
26115 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
26116 (workaround_speculation): Likewise.
26117 (add_sched_insns_for_speculation): Likewise.
26118 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
26119 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
26120 (frv_for_each_packet): Likewise.
26121 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
26122 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
26123 (final_emit_insn_group_barriers): Likewise.
26124 * config/m32r/m32r.c (m32r_is_insn): Likewise.
26125 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
26126 (mips16_insn_length): Likewise.
26127 * config/pa/pa.c (pa_reorg): Likewise.
26128 (pa_combine_instructions): Likewise.
26129 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
26130 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
26131 (sh_reorg): Likewise.
26132 (split_branches): Likewise.
26133 * config/spu/spu.c (get_branch_target): Likewise.
26135 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
26138 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
26140 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
26141 Fix declaration name.
26143 2013-03-28 Lawrence Crowl <crowl@google.com>
26145 * graphds.h (struct graph.indicies): Remove unused.
26146 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
26147 (SCOP_ORIGINAL_PDDRS): Remove unused.
26148 * sese.h (extern insert_loop_close_phis): Removed unused.
26149 (extern insert_guard_phis): Removed unused.
26150 (extern ivtype_map_elt_info): Removed unused.
26151 (new_ivtype_map_elt): Removed unused.
26152 * sese.c (ivtype_map_elt_info): Removed unused.
26154 2013-03-28 Lawrence Crowl <crowl@google.com>
26156 * Makefile.in: Add several missing include dependences.
26158 (test-dump.o): New. This object is not added to any executable,
26159 but is present for ad-hoc testing.
26161 (debug (const bitmap_head_def &)): New.
26162 (debug (const bitmap_head_def *)): New.
26164 (extern debug (const bitmap_head_def &)): New.
26165 (extern debug (const bitmap_head_def *)): New.
26167 (debug (edge_def &)): New.
26168 (debug (edge_def *)): New.
26170 (debug (basic_block_def &)): New.
26171 (debug (basic_block_def *)): New.
26173 (dump_node (const_tree, int, FILE *)): Correct source file.
26175 (debug (die_struct &)): New.
26176 (debug (die_struct *)): New.
26178 (extern debug (die_struct &)): New.
26179 (extern debug (die_struct *)): New.
26180 * gimple-pretty-print.c
26181 (debug (gimple_statement_d &)): New.
26182 (debug (gimple_statement_d *)): New.
26183 * gimple-pretty-print.h
26184 (extern debug (gimple_statement_d &)): New.
26185 (extern debug (gimple_statement_d *)): New.
26187 (debug (ira_allocno_copy &)): New.
26188 (debug (ira_allocno_copy *)): New.
26189 (debug (ira_allocno &)): New.
26190 (debug (ira_allocno *)): New.
26192 (extern debug (ira_allocno_copy &)): New.
26193 (extern debug (ira_allocno_copy *)): New.
26194 (extern debug (ira_allocno &)): New.
26195 (extern debug (ira_allocno *)): New.
26197 (debug (live_range &)): New.
26198 (debug (live_range *)): New.
26200 (debug (lra_live_range &)): New.
26201 (debug (lra_live_range *)): New.
26203 (debug (lra_live_range &)): New.
26204 (debug (lra_live_range *)): New.
26206 (debug (omega_pb_d &)): New.
26207 (debug (omega_pb_d *)): New.
26209 (extern debug (omega_pb_d &)): New.
26210 (extern debug (omega_pb_d *)): New.
26212 (debug (const rtx_def &)): New.
26213 (debug (const rtx_def *)): New.
26215 (debug_tree (tree): Move within file.
26216 (debug_raw (const tree_node &)): New.
26217 (debug_raw (const tree_node *)): New.
26218 (dump_tree_via_hooks (const tree_node *, int)): New.
26219 (debug (const tree_node &)): New.
26220 (debug (const tree_node *)): New.
26221 (debug_verbose (const tree_node &)): New.
26222 (debug_verbose (const tree_node *)): New.
26223 (debug_head (const tree_node &)): New.
26224 (debug_head (const tree_node *)): New.
26225 (debug_body (const tree_node &)): New.
26226 (debug_body (const tree_node *)): New.
26227 (debug_vec_tree (tree): Move and reimplement in terms of dump.
26228 (debug (vec<tree, va_gc> &)): New.
26229 (debug (vec<tree, va_gc> *)): New.
26231 (extern debug (const rtx_def &)): New.
26232 (extern debug (const rtx_def *)): New.
26234 (debug_raw (simple_bitmap_def &)): New.
26235 (debug_raw (simple_bitmap_def *)): New.
26236 (debug (simple_bitmap_def &)): New.
26237 (debug (simple_bitmap_def *)): New.
26239 (extern debug (simple_bitmap_def &)): New.
26240 (extern debug (simple_bitmap_def *)): New.
26241 (extern debug_raw (simple_bitmap_def &)): New.
26242 (extern debug_raw (simple_bitmap_def *)): New.
26244 (debug (vinsn_def &)): New.
26245 (debug (vinsn_def *)): New.
26246 (debug_verbose (vinsn_def &)): New.
26247 (debug_verbose (vinsn_def *)): New.
26248 (debug (expr_def &)): New.
26249 (debug (expr_def *)): New.
26250 (debug_verbose (expr_def &)): New.
26251 (debug_verbose (expr_def *)): New.
26252 (debug (vec<rtx> &)): New.
26253 (debug (vec<rtx> *)): New.
26255 (extern debug (vinsn_def &)): New.
26256 (extern debug (vinsn_def *)): New.
26257 (extern debug_verbose (vinsn_def &)): New.
26258 (extern debug_verbose (vinsn_def *)): New.
26259 (extern debug (expr_def &)): New.
26260 (extern debug (expr_def *)): New.
26261 (extern debug_verbose (expr_def &)): New.
26262 (extern debug_verbose (expr_def *)): New.
26263 (extern debug (vec<rtx> &)): New.
26264 (extern debug (vec<rtx> *)): New.
26266 (_list_iter_cond_expr): Make inline instead of static.
26268 (debug (sreal &)): New.
26269 (debug (sreal *)): New.
26271 (extern debug (sreal &)): New.
26272 (extern debug (sreal *)): New.
26274 (extern debug_raw (const tree_node &)): New.
26275 (extern debug_raw (const tree_node *)): New.
26276 (extern debug (const tree_node &)): New.
26277 (extern debug (const tree_node *)): New.
26278 (extern debug_verbose (const tree_node &)): New.
26279 (extern debug_verbose (const tree_node *)): New.
26280 (extern debug_head (const tree_node &)): New.
26281 (extern debug_head (const tree_node *)): New.
26282 (extern debug_body (const tree_node &)): New.
26283 (extern debug_body (const tree_node *)): New.
26284 (extern debug (vec<tree, va_gc> &)): New.
26285 (extern debug (vec<tree, va_gc> *)): New.
26287 (debug (struct loop &)): New.
26288 (debug (struct loop *)): New.
26289 (debug_verbose (struct loop &)): New.
26290 (debug_verbose (struct loop *)): New.
26291 * tree-dump.c: Add header dependence.
26293 (extern debug (struct loop &)): New.
26294 (extern debug (struct loop *)): New.
26295 (extern debug_verbose (struct loop &)): New.
26296 (extern debug_verbose (struct loop *)): New.
26298 (debug (data_reference &)): New.
26299 (debug (data_reference *)): New.
26300 (debug (vec<data_reference_p> &)): New.
26301 (debug (vec<data_reference_p> *)): New.
26302 (debug (vec<ddr_p> &)): New.
26303 (debug (vec<ddr_p> *)): New.
26305 (extern debug (data_reference &)): New.
26306 (extern debug (data_reference *)): New.
26307 (extern debug (vec<data_reference_p> &)): New.
26308 (extern debug (vec<data_reference_p> *)): New.
26309 (extern debug (vec<ddr_p> &)): New.
26310 (extern debug (vec<ddr_p> *)): New.
26312 (debug (pt_solution &)): New.
26313 (debug (pt_solution *)): New.
26315 (extern debug (pt_solution &)): New.
26316 (extern debug (pt_solution *)): New.
26318 (debug (_var_map &)): New.
26319 (debug (_var_map *)): New.
26320 (debug (tree_live_info_d &)): New.
26321 (debug (tree_live_info_d *)): New.
26323 (extern debug (_var_map &)): New.
26324 (extern debug (_var_map *)): New.
26325 (extern debug (tree_live_info_d &)): New.
26326 (extern debug (tree_live_info_d *)): New.
26328 2013-03-28 Jan Hubicka <jh@suse.cz>
26330 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
26332 2013-03-28 Ian Bolton <ian.bolton@arm.com>
26334 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
26335 record only when desired or required.
26337 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
26339 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
26340 *vec_extractv2di_1_rex64. Use x64 isa attribute.
26342 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
26344 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
26345 (*andsi3_compare0_uxtw): New pattern.
26346 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
26347 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
26349 2013-03-28 Jan Hubicka <jh@suse.cz>
26351 * data-streamer-in.c (streamer_read_gcov_count): New function.
26352 * gimple-streamer-out.c: Include value-prof.h.
26353 (output_gimple_stmt): Output histogram.
26354 (output_bb): Use streamer_write_gcov_count.
26355 * value-prof.c: Include data-streamer.h
26356 (dump_histogram_value): Add HIST_TYPE_MAX.
26357 (stream_out_histogram_value): New function.
26358 (stream_in_histogram_value): New function.
26359 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
26360 (stream_out_histogram_value, stream_in_histogram_value): Declare.
26361 * data-streamer-out.c (streamer_write_gcov_count): New function.
26362 (streamer_write_gcov_count_stream): New function.
26363 * lto-cgraph.c (lto_output_edge): Update counter streaming.
26364 (lto_output_node): Likewise.
26365 (input_node, input_edge): Likewise.
26366 * lto-streamer-out.c (output_cfg): Update streaming.
26367 * lto-streamer-in.c (input_cfg): Likewise.
26368 * data-streamer.h (streamer_write_gcov_count,
26369 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
26370 * gimple-streamer-in.c: Include value-prof.h
26371 (input_gimple_stmt): Input histograms.
26372 (input_bb): Update profile streaming.
26374 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
26376 * genmodes.c (emit_max_int): New function.
26377 (emit_insn_modes_h): Added call to emit_max_function.
26378 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
26380 * machmode.def: Fixed comment.
26382 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
26384 * combine.c (try_combine): Removed useless assert.
26385 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
26387 2013-03-28 Marek Polacek <polacek@redhat.com>
26388 Richard Biener <rguenther@suse.de>
26390 PR tree-optimization/56695
26391 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
26392 build signed result of a vector comparison.
26393 * tree-cfg.c (verify_gimple_comparison): Check that a result
26394 of a vector comparison has signed type.
26396 2013-03-28 Richard Biener <rguenther@suse.de>
26398 PR tree-optimization/37021
26399 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
26400 do not restrict gaps between groups.
26401 * tree-vect-stmts.c (vectorizable_load): Properly account for
26402 a gap between groups.
26404 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
26406 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
26407 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
26410 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
26412 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
26413 * doc/extend.texi (Named Address Spaces): Ditto.
26414 (Variable Attributes): Ditto.
26416 2013-03-27 Kai Tietz <ktietz@redhat.com>
26418 * config.build: Add support for cygwin x64 target.
26419 * config.gcc: Likewise.
26420 * config.host: Likewise.
26421 * configure.ac: Likewise
26422 * configure: Regenerated.
26424 2013-03-27 Kai Tietz <ktietz@redhat.com>
26426 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
26427 * config/i386/t-cygwin-w64: New file.
26428 * config/i386/cygwin-w64.h: New file.
26429 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
26430 and add support for x64-cygwin target.
26431 (CPP_SPEC): Likewise.
26432 (CXX_WRAP_SPEC_LIST): Undefine before define.
26433 (LIBGCJ_SONAME): Use 15 as version.
26435 2013-03-27 Richard Biener <rguenther@suse.de>
26437 PR tree-optimization/56716
26438 * tree-ssa-structalias.c (perform_var_substitution): Adjust
26439 dumping for ref nodes.
26441 2013-03-27 Martin Jambor <mjambor@suse.cz>
26443 PR tree-optimization/55334
26444 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
26445 restricted pointers to arrays.
26447 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
26449 * Makefile.in (.SUFFIXES): Add .cc.
26450 (.c.o): Apply same recipe for implicit rule .cc.o.
26452 2013-03-27 Richard Biener <rguenther@suse.de>
26454 PR tree-optimization/37021
26455 * tree-vect-data-refs.c (vect_check_strided_load): Allow
26456 REALPART/IMAGPART_EXPRs around the supported refs.
26457 * tree-ssa-structalias.c (find_func_aliases): Assume that
26458 floating-point values are not used to transfer pointers.
26460 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
26462 * target.def (TARGET_HAS_IFUNC_P): New target hook.
26463 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
26464 * doc/tm.texi: Regenerate.
26465 * targhooks.h (default_has_ifunc_p): New.
26466 * targhooks.c (default_has_ifunc_p): Ditto.
26467 * config/linux-protos.h: New file.
26468 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
26469 hook for linux which disables support of indirect functions in android.
26470 * config/linux-android.c: New file.
26471 * config/t-linux-android.c: Ditto.
26472 * config.gcc: Added new object file linux-android.o.
26473 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
26474 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
26475 * varasm.c (do_assemble_alias): Likewise.
26476 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
26477 doesn't support indirect functions.
26478 * configure: Regenerate.
26480 2013-03-27 Bin Cheng <bin.cheng@arm.com>
26483 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
26484 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
26487 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26489 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
26491 2013-03-27 Terry Guo <terry.guo@arm.com>
26493 * config/arm/arm-cores.def: Added core cortex-r7.
26494 * config/arm/arm-tune.md: Regenerated.
26495 * config/arm/arm-tables.opt: Regenerated.
26496 * doc/invoke.texi: Added entry for core cortex-r7.
26498 2013-03-27 Walter Lee <walt@tilera.com>
26500 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
26501 double-decrement of next_scratch_regno.
26503 2013-03-27 Walter Lee <walt@tilera.com>
26505 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
26507 (insn_v1mulus): Ditto.
26508 (insn_v2muls): Ditto.
26510 2013-03-27 Walter Lee <walt@tilera.com>
26512 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
26513 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
26515 2013-03-27 Walter Lee <walt@tilera.com>
26517 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
26518 (*sibcall_value): Ditto.
26520 2013-03-27 Walter Lee <walt@tilera.com>
26522 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
26523 (insn_mnz_v8qi): ... this ...
26524 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
26526 (insn_v<n>mnz): Replaced by ...
26527 (insn_v1mnz): ... this ...
26528 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
26530 (insn_mz_<mode>): Replaced by ...
26531 (insn_mz_v8qi): ... this ...
26532 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
26534 (insn_v<n>mz): Replaced by ...
26535 (insn_v1mz): ... this ...
26536 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
26539 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
26541 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
26543 2013-03-26 Roland McGrath <mcgrathr@google.com>
26545 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
26546 than fprintf with a non-constant, non-format string.
26548 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
26550 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
26551 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
26552 operand 0 predicate.
26553 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
26554 attribute. Use general_x64nomem_operand as operand 1 predicate.
26555 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
26556 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
26557 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
26558 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
26559 (mov<mode>_insv_1): Remove expander. Merge insn with
26560 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
26561 Use general_x64nomem_operand as operand 1 predicate.
26562 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
26563 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
26564 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
26565 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
26566 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
26567 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
26568 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
26569 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
26570 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
26571 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
26572 (general_x64nomem_operand): Ditto.
26574 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
26576 * config/rtems.opt: Add -pthread option.
26578 2013-03-26 Richard Biener <rguenther@suse.de>
26580 * alias.c (find_base_term): Avoid redundant and not used recursion.
26581 (base_alias_check): Get the initial base term from the caller.
26582 (true_dependence_1): Compute and pass base terms to base_alias_check.
26583 (write_dependence_p): Likewise.
26584 (may_alias_p): Likewise.
26586 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
26588 * config/aarch64/aarch64.c (aarch64_classify_address): Support
26589 PC-relative load in SI modes and above only.
26591 2013-03-26 Xinyu Qi <xyqi@marvell.com>
26593 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
26594 * config/arm/iwmmxt.md (WCGR0): Update.
26595 (WCGR1, WCGR2, WCGR3): Likewise.
26597 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
26599 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
26600 Use x64 and nox64 isa attributes.
26602 2013-03-26 Richard Biener <rguenther@suse.de>
26604 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
26605 alignment computations and rely on get_object_alignment_1
26606 for the !TYPE_P case.
26607 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
26609 2013-03-26 Walter Lee <walt@tilera.com>
26611 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
26612 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
26614 2013-03-25 Jeff Law <law@redhat.com>
26616 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
26617 check for INTEGRAL_TYPE_P that was missing due to checking in
26618 wrong version of prior patch.
26620 2013-03-25 Walter Lee <walt@tilera.com>
26622 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
26623 TILEGX_INSN_SHUFFLEBYTES1.
26624 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
26626 (tilegx_builtins): Ditto.
26627 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
26629 2013-03-25 Walter Lee <walt@tilera.com>
26631 * config/tilegx/tilegx.md (floatsisf2): New pattern.
26632 (floatunssisf2): New pattern.
26633 (floatsidf2): New pattern.
26634 (floatunssidf2): New pattern.
26636 2013-03-25 Walter Lee <walt@tilera.com>
26638 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
26639 tests for constraint J, K, N, P.
26641 2013-03-25 Walter Lee <walt@tilera.com>
26643 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
26644 Use indirect/pcrel encoding.
26645 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
26648 2013-03-25 Steve Ellcey <sellcey@mips.com>
26650 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
26651 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
26652 * config/mips/mips.c (mips_option_override): Set IMADD default.
26653 * config/mips/mips.h (PTF_AVOID_IMADD): New.
26654 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
26655 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
26656 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
26657 * doc/invoke.texi (-mimadd/-mno-imadd): New.
26659 2013-03-25 Jeff Law <law@redhat.com>
26661 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
26662 slightly to avoid creating and folding useless trees. Simplify
26663 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
26665 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
26667 * config/i386/i386.md (*zero_extendsidi2): Merge with
26668 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
26669 * config/i386/predicates.md (x86_64_zext_operand): Rename from
26670 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
26671 targets. Clarify comment.
26673 2013-03-25 Martin Jambor <mjambor@suse.cz>
26675 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
26676 pass-through jump functions differently.
26677 (ipa_read_jump_function): Likewise. Also use setter functions to set
26680 2013-03-25 Martin Jambor <mjambor@suse.cz>
26682 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
26683 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
26685 (ipa_get_indirect_edge_target): New function.
26686 (devirtualization_time_bonus): New parameter known_aggs, pass it to
26687 ipa_get_indirect_edge_target. Update all callers.
26688 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
26689 ipa_get_indirect_edge_target_1 instead of calling
26690 ipa_get_indirect_edge_target.
26691 (create_specialized_node): Pass aggvlas to
26692 ipcp_discover_new_direct_edges.
26694 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
26696 * config/arm/arm.md (f_sels, f_seld): New types.
26697 (*cmov<mode>): New pattern.
26698 * config/arm/predicates.md (arm_vsel_comparison_operator): New
26701 2013-03-25 Kai Tietz <ktietz@redhat.com>
26703 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
26704 POSIX-printf for mingw-hosted builds.
26706 2013-03-25 Richard Biener <rguenther@suse.de>
26708 PR middle-end/56694
26709 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
26710 must-not-throw stmt location.
26712 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
26714 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
26715 Emit load-acquire versions when acq is true.
26716 (arm_emit_store_exclusive): Add rel parameter.
26717 Emit store-release versions when rel is true.
26718 (arm_split_compare_and_swap): Use acquire-release instructions
26720 of barriers when appropriate.
26721 (arm_split_atomic_op): Likewise.
26722 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
26723 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
26724 (VUNSPEC_SLX): Likewise.
26725 (VUNSPEC_LDA): Likewise.
26726 (VUNSPEC_STL): Likewise.
26727 * config/arm/sync.md (atomic_load<mode>): New pattern.
26728 (atomic_store<mode>): Likewise.
26729 (arm_load_acquire_exclusive<mode>): Likewise.
26730 (arm_load_acquire_exclusivesi): Likewise.
26731 (arm_load_acquire_exclusivedi): Likewise.
26732 (arm_store_release_exclusive<mode>): Likewise.
26734 2013-03-25 Catherine Moore <clm@codesourcery.com>
26736 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
26737 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
26738 * config/mip/predicates.md (lwsp_swsp_operand,
26739 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
26740 sb16_operand, db4_operand, db7_operand, ib3_operand,
26741 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
26742 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
26743 andi16_operand): New predicates.
26744 * config/mips/mips.md (compression): New attribute.
26745 (enabled): New attribute.
26746 (length): Consider compression in computing length.
26747 (shift_compression): New code attribute.
26748 (*add<mode>3): New operands. Record compression.
26749 (sub<mode>3): Likewise.
26750 (one_cmpl<mode>2): Likewise.
26751 (*and<mode>3): Likewise.
26752 (*ior<mode>3): Likewise.
26753 (unnamed pattern for xor): Likewise.
26754 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
26755 (*<optab><mode>3): Likewise.
26756 (*mov<mode>_internal: Likewise.
26757 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
26758 (mips_unsigned_immediate_p): New.
26759 (umips_lwsp_swsp_address_p): New.
26760 (m16_based_address_p): New.
26761 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
26762 (mips_unsigned_immediate_p): New prototype.
26763 (lwsp_swsp_address_p): New prototype.
26764 (m16_based_address_p): New prototype.
26765 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
26766 (mips_signed_immediate_p): New function.
26767 (m16_based_address_p): New function.
26768 (lwsp_swsp_address_p): New function.
26769 (mips_print_operand_punctuation): Recognize short delay slot insns
26770 for microMIPS.add<mode>3"
26772 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
26775 * config/arm/iterators.md (v_cmp_result): New mode attribute.
26776 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
26778 2013-03-25 Richard Biener <rguenther@suse.de>
26780 PR tree-optimization/56689
26781 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
26784 2013-03-25 Richard Biener <rguenther@suse.de>
26786 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
26788 (memory_references): Likewise.
26789 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
26790 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
26791 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
26792 (gather_mem_refs_in_loops): Fold into ...
26793 (analyze_memory_references): ... this. Move initialization
26794 to tree_ssa_lim_initialize.
26795 (fill_always_executed_in): Rename to ...
26796 (fill_always_executed_in_1): ... this.
26797 (fill_always_executed_in): Move contains_call computation to
26798 this new function from ...
26799 (tree_ssa_lim_initialize): ... here.
26800 (tree_ssa_lim): Call fill_always_executed_in.
26802 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
26804 * postreload.c (reload_combine): Fix code detecting returns.
26806 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
26808 * function.c (emit_use_return_register_into_block): On cc0 targets,
26809 do not emit the sequence between cc0 setter and user.
26811 2013-03-25 Kai Tietz <ktietz@redhat.com>
26813 * config/i386/predicates.md (local_symbolic_operand): Interpret
26814 dll-imported symbols as none-local.
26816 2013-03-25 Richard Biener <rguenther@suse.de>
26818 * tree-ssa-loop-im.c (struct depend): Remove.
26819 (struct lim_aux_data): Make depends a vec of gimples.
26820 (free_lim_aux_data): Adjust.
26821 (add_dependency): Likewise.
26822 (set_level): Likewise.
26824 2013-03-25 Richard Biener <rguenther@suse.de>
26826 PR middle-end/56434
26827 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
26828 the pointer returned by calls with ECF_MALLOC set.
26830 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
26832 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
26834 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
26836 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
26837 using MMXMODE mode iterator.
26838 (*move<mode>_internal): Merge with *movv2sf_internal and
26839 *movv2sf_internal_rex64 using MMXMODE mode iterator.
26841 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
26843 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
26844 (record_last_mem_set_info): Likewise.
26846 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
26847 of XNEWVEC followed by memset.
26848 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
26850 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
26852 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
26853 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
26854 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
26855 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
26856 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
26857 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
26858 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
26859 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
26860 BARRIER_P instead of GET_CODE.
26862 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
26864 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
26865 inaccuracy in the probing code.
26867 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
26868 (ctrapdi4): Likewise.
26870 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
26872 * calls.c (expand_call): Add missing guard to code handling return
26873 of non-BLKmode structures in MSB.
26874 * function.c (expand_function_end): Likewise.
26876 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
26878 * combine.c (try_combine): Adjust comment. Do not add the set of
26879 insn #0 if the destination indirectly is set or dies in insn #2.
26880 Tidy up code to distribute a new note.
26882 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
26884 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
26885 also for alternatives 16 and 17.
26887 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
26889 * config/i386/sse.md (*mov<mode>_internal): Merge with
26890 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
26891 Emit insn template depending on type attribute. Use
26892 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
26893 movd instead of movq mnemonic for interunit moves. Rewrite mode
26894 attribute calculation. Remove unit attribute calculation.
26895 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
26896 Set prefix_data16 attribute for DImode ssemov types.
26897 Use Ym instead of y for SSE-MMX conversion alternatives.
26898 Reorder operand constraints.
26900 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
26902 * df.h (df_insn_delete): Adjust prototype.
26903 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
26904 and let it decide whether mark the basic block dirty.
26905 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
26906 * df-scan.c (df_insn_info_delete): New helper function, split
26907 off from df_insn_delete.
26908 (df_scan_free_bb_info): Use it.
26909 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
26911 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
26912 that the insn is actually an insn and it has a non-NULL basic block.
26913 Do not mark basic block dirty if only deleting a DEBUG_INSN.
26915 2013-03-22 Richard Biener <rguenther@suse.de>
26917 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
26919 (mem_ref_alloc): Do not allocate them.
26920 (refs_independent_p): Do not query or maintain a cache.
26922 2013-03-22 Richard Biener <rguenther@suse.de>
26924 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
26925 (gather_mem_refs_in_loops): Do not compute it.
26926 (analyze_memory_references): Do not allocate it.
26927 (tree_ssa_lim_finalize): Do not free it.
26928 (for_all_locs_in_loop): Do not query all_refs_in_loop.
26930 2013-03-22 Richard Biener <rguenther@suse.de>
26932 * is-a.h (as_a): Use gcc_checking_assert.
26934 2013-03-22 Ian Bolton <ian.bolton@arm.com>
26936 * config/aarch64/aarch64.c (aarch64_print_operand): New
26937 format specifier for printing a constant in hex.
26938 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
26939 format specifier for printing second operand.
26941 2013-03-22 Richard Biener <rguenther@suse.de>
26943 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
26945 (gather_mem_refs_in_loops): Perform store accumulation here.
26946 (create_vop_ref_mapping_loop): Remove.
26947 (create_vop_ref_mapping): Likewise.
26948 (analyze_memory_references): Initialize refs_stored_in_loop.
26949 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
26950 (record_indep_loop): Remove.
26951 (record_dep_loop): New function.
26952 (ref_indep_loop_p_1): Adjust to only walk over references
26953 in the loop, not its subloops.
26954 (ref_indep_loop_p): Rename to ...
26955 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
26956 maintaining a more fine-grained cache.
26957 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
26958 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
26960 2013-03-22 Richard Biener <rguenther@suse.de>
26962 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
26963 (struct mem_ref): Make accesses_in_loop a vec of a vec of
26964 aggregate mem_ref_loc.
26965 (free_mem_ref_locs): Inline into ...
26966 (memref_free): ... this and adjust.
26967 (mem_ref_alloc): Adjust.
26968 (mem_ref_locs_alloc): Remove.
26969 (record_mem_ref_loc): Adjust.
26970 (get_all_locs_in_loop): Rewrite into ...
26971 (for_all_locs_in_loop): ... this iterator.
26972 (rewrite_mem_ref_loc): New functor.
26973 (rewrite_mem_refs): Use for_all_locs_in_loop.
26974 (sm_set_flag_if_changed): New functor.
26975 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
26976 (ref_always_accessed): New functor.
26977 (ref_always_accessed_p): Use for_all_locs_in_loop.
26979 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
26981 * tree-pass.h (PROP_gimple_lvec): New.
26982 * passes.c (dump_properties): Handle PROP_gimple_lvec.
26983 (init_optimization_passes): Move pass_lower_vector.
26984 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
26986 (pass_lower_vector): Provide PROP_gimple_lvec.
26987 (pass_lower_vector_ssa): Likewise.
26988 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
26990 2013-03-21 Mark Wielaard <mjw@redhat.com>
26992 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
26994 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
26996 * config/i386/i386.md (*movdi_internal): Disparage slightly
26997 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
26998 conversion alternatives.
27000 2013-03-21 Jakub Jelinek <jakub@redhat.com>
27002 PR middle-end/48087
27003 * diagnostic.def (DK_WERROR): New kind.
27004 * diagnostic.h (werrorcount): Define.
27005 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
27006 promoted to DK_ERROR, increment DK_WERROR counter instead of
27008 * toplev.c (toplev_main): Call print_ignored_options even if
27009 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
27010 even if just werrorcount is non-zero.
27013 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
27015 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
27016 (string_cst_pool_decl): New function.
27017 (optimize_one_addr_into_implicit_ptr): New function.
27018 (resolve_addr_in_expr): Optimize DWARF location expression
27019 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
27020 which doesn't live in memory, but has DW_AT_location or
27021 DW_AT_const_value, or refers to a string literal, into
27022 DW_OP_GNU_implicit_pointer.
27023 (optimize_location_into_implicit_ptr): New function.
27024 (resolve_addr): If removing DW_AT_location of a variable because
27025 it was DW_OP_addr of address of the variable, but the variable doesn't
27026 live in memory, try to emit const value attribute for the initializer.
27028 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
27030 * tree.h (VECTOR_TYPE_P): New macro.
27031 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
27032 TYPE_MODE): Use it.
27033 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
27034 VEC_COND_EXPR cannot be lvalues.
27035 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
27037 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
27039 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
27040 Restrict the transformation to equal modes.
27042 2013-03-21 Richard Biener <rguenther@suse.de>
27044 PR tree-optimization/39326
27045 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
27046 (MEM_ANALYZABLE): Adjust.
27047 (record_mem_ref_loc): Move bitmap ops ...
27048 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
27049 unanalyzable refs, do not record locations for it.
27050 (analyze_memory_references): Allocate ref zero as shared
27052 (refs_independent_p): Do not test for unanalyzed mems here.
27053 (ref_indep_loop_p_1): Special-case disambiguation against
27054 the unanalyzed ref.
27055 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
27057 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
27059 * config/arm/arm-protos.h (tune_params): Add
27060 prefer_neon_for_64bits field.
27061 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
27062 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
27063 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
27064 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
27065 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
27066 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
27067 (arm_option_override): Handle -mneon-for-64bits new option.
27068 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
27069 (prefer_neon_for_64bits): Declare new variable.
27070 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
27071 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
27072 (arch_enabled): Handle new arch types. Remove support for onlya8
27074 (one_cmpldi2): Use new arch names.
27075 (zero_extend<mode>di2, extend<mode>di2): Ditto.
27076 * config/arm/arm.opt (mneon-for-64bits): Add option.
27077 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
27078 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
27079 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
27081 * doc/invoke.texi (-mneon-for-64bits): Document.
27083 2013-03-21 Richard Biener <rguenther@suse.de>
27085 PR tree-optimization/39326
27086 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
27087 (sort_bbs_in_loop_postorder_cmp): New function.
27088 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
27090 2013-03-21 Richard Biener <rguenther@suse.de>
27092 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
27093 (vect_insert_into_interleaving_chain): Likewise.
27094 (vect_drs_dependent_in_basic_block): Inline ...
27095 (vect_slp_analyze_data_ref_dependence): ... here. New function,
27097 (vect_analyze_data_ref_dependence): ... here. Simplify.
27098 (vect_check_interleaving): Simplify.
27099 (vect_analyze_data_ref_dependences): Likewise. Split out ...
27100 (vect_slp_analyze_data_ref_dependences): ... this new function.
27101 (dr_group_sort_cmp): New function.
27102 (vect_analyze_data_ref_accesses): Compute data-reference groups
27103 here instead of in vect_analyze_data_ref_dependence. Use
27104 a more efficient algorithm.
27105 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
27106 vect_slp_analyze_data_ref_dependences. Call
27107 vect_analyze_data_ref_accesses earlier.
27108 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
27109 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
27110 (vect_slp_analyze_data_ref_dependences): New prototype.
27112 2013-03-21 Richard Biener <rguenther@suse.de>
27114 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
27115 ref is stored in the loop.
27116 (find_refs_for_sm): Walk only over all stores.
27117 (store_motion_loop): Allocate from lim_bitmap_obstack.
27118 (store_motion): Likewise.
27120 2013-03-21 Richard Biener <rguenther@suse.de>
27122 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
27123 Update virtual SSA form.
27125 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
27127 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
27128 * configure: Regenerate.
27129 * config.in: Regenerate.
27130 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
27131 if !HAVE_LD_EH_FRAME_CIEV3.
27133 2013-03-21 Richard Biener <rguenther@suse.de>
27135 * tree-cfg.c (verify_expr_no_block): New function.
27136 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
27137 nor DECL_VALUE_EXPR have locations with associated blocks.
27138 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
27139 (clear_unused_block_pointer): Remove code dealing with
27140 blocks in DECL_DEBUG_EXPR locations.
27142 2013-03-21 Richard Biener <rguenther@suse.de>
27144 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
27145 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
27146 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
27147 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
27148 instead of DECL_DEBUG_EXPR_IS_FROM.
27149 * gimplify.c (gimplify_modify_expr): Likewise.
27150 * tree-cfg.c (verify_expr_location_1): Likewise.
27151 * tree-complex.c (create_one_component_var): Likewise.
27152 * tree-sra.c (create_access_replacement): Likewise.
27153 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
27154 (clear_unused_block_pointer): Likewise.
27155 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
27156 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
27157 * var-tracking.c (var_debug_decl): Likewise.
27158 (track_expr_p): Likewise.
27159 * tree-inline.c (add_local_variables): Likewise. Set
27160 DECL_HAS_DEBUG_EXPR_P after copying it.
27161 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
27162 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
27164 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
27167 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
27168 * configure: Regenerate.
27169 * config.in: Regenerate.
27170 * config/i386/i386.md (*movdf_internal): Use
27171 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
27172 movd instead of movq mnemonic for interunit moves.
27173 (*movdi_internal): Ditto.
27175 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
27177 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
27178 (abd<mode>_3): New pattern.
27179 (aba<mode>_3): New pattern.
27180 (fabd<mode>_3): New pattern.
27182 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
27184 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
27185 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
27186 occurrence of REGISTER_PREFIX as its empty string.
27188 2013-03-20 Jeff Law <law@redhat.com>
27190 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
27191 addititional equivalences for equality comparisons between an SSA_NAME
27192 and a constant where the SSA_NAME was set from a widening conversion.
27194 2013-03-20 Walter Lee <walt@tilera.com>
27196 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
27198 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
27200 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
27201 depending on type attribute.
27202 (*movti_internal): Ditto.
27203 (*movtf_internal): Ditto.
27204 (*movxf_internal): Ditto.
27205 (*movdf_internal): Ditto.
27206 (*movsf_internal): Ditto.
27208 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
27210 * config/i386/i386.md (*movti_internal): Set prefix attribute to
27211 maybe_vex for sselog1 and ssemov types.
27212 (*movdi_internal): Reorder operand constraints.
27213 (*movsi_internal): Ditto. Set prefix attribute to
27214 maybe_vex for sselog1 and ssemov types.
27215 (*movtf_internal): Set prefix attribute to maybe_vex
27216 for sselog1 and ssemov types.
27217 (*movdf_internal): Ditto. Set prefix_data16 attribute for
27218 DImode ssemov types. Reorder operand constraints.
27219 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
27220 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
27221 attribute for SImode ssemov types. Reorder operand constraints.
27223 2013-03-20 Martin Jambor <mjambor@suse.cz>
27225 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
27226 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
27228 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
27230 * config/rs6000/predicates.md (indexed_address, update_address_mem
27231 update_indexed_address_mem): New predicates.
27232 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
27233 attribute for load/store instructions.
27234 * config/rs6000/dfp.md (movsd_store): Likewise.
27235 (movsd_load): Likewise.
27236 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
27237 (unnamed HI->DI extend define_insn): Likewise.
27238 (unnamed SI->DI extend define_insn): Likewise.
27239 (unnamed QI->SI extend define_insn): Likewise.
27240 (unnamed QI->HI extend define_insn): Likewise.
27241 (unnamed HI->SI extend define_insn): Likewise.
27242 (unnamed HI->SI extend define_insn): Likewise.
27243 (extendsfdf2_fpr): Likewise.
27244 (movsi_internal1): Likewise.
27245 (movsi_internal1_single): Likewise.
27246 (movhi_internal): Likewise.
27247 (movqi_internal): Likewise.
27248 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
27249 attribute for load/store instructions.
27250 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
27252 (mov<mode>_softfloat): Likewise.
27253 (mov<mode>_hardfloat32): Likewise.
27254 (mov<mode>_hardfloat64): Likewise.
27255 (mov<mode>_softfloat64): Likewise.
27256 (movdi_internal32): Likewise.
27257 (movdi_internal64): Likewise.
27258 (probe_stack_<mode>): Likewise.
27260 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
27262 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
27263 floating point, and decimal floating point to reload iterator.
27265 * config/rs6000/constraints.md (wl constraint): New constraints to
27266 return FLOAT_REGS if certain options are used to reduce the number
27267 of separate patterns that exist in the file.
27268 (wx constraint): Likewise.
27269 (wz constraint): Likewise.
27271 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
27272 -mdebug=reg, print wg, wl, wx, and wz constraints.
27273 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
27274 Initialize the reload functions for 64-bit binary/decimal floating
27276 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
27277 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
27278 create the buffer on the stack to overcome not having a 32-bit
27280 (rs6000_emit_move): Likewise.
27281 (rs6000_secondary_memory_needed_rtx): Likewise.
27282 (rs6000_alloc_sdmode_stack_slot): Likewise.
27283 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
27284 via xxlxor, just like DFmode 0.0.
27286 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
27287 define as 1 if we are running on a power7 or newer.
27288 (enum r6000_reg_class_enum): Add new constraints.
27290 * config/rs6000/dfp.md (movsd): Delete, combine with binary
27291 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
27292 with other moves by using conditional constraits (wg). Use LFIWZX
27293 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
27294 (movsd splitter): Likewise.
27295 (movsd_hardfloat): Likewise.
27296 (movsd_softfloat): Likewise.
27298 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
27299 binary and decimal floating point moves.
27300 (fmove_ok): New attributes to combine binary and decimal floating
27301 point moves, and to combine power6x (mfpgpr) moves along normal
27303 (real_value_to_target): Likewise.
27304 (f32_lr): Likewise.
27305 (f32_lm): Likewise.
27306 (f32_li): Likewise.
27307 (f32_sr): Likewise.
27308 (f32_sm): Likewise.
27309 (f32_si): Likewise.
27310 (movsf): Combine binary and decimal floating point moves. Combine
27311 power6x (mfpgpr) moves with other moves by using conditional
27312 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
27313 (mov<mode> for SFmode/SDmode); Likewise.
27314 (SFmode/SDmode splitters): Likewise.
27315 (movsf_hardfloat): Likewise.
27316 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
27317 (movsf_softfloat): Likewise.
27318 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
27320 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
27321 wx and wz constraints.
27323 * config/rs6000/constraints.md (wg constraint): New constraint to
27324 return FLOAT_REGS if -mmfpgpr (power6x) was used.
27326 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
27329 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
27330 -mdebug=reg, print wg, wl, wx, and wz constraints.
27331 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
27332 Initialize the reload functions for 64-bit binary/decimal floating
27334 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
27335 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
27336 create the buffer on the stack to overcome not having a 32-bit
27338 (rs6000_emit_move): Likewise.
27339 (rs6000_secondary_memory_needed_rtx): Likewise.
27340 (rs6000_alloc_sdmode_stack_slot): Likewise.
27341 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
27342 via xxlxor, just like DFmode 0.0.
27344 * config/rs6000/dfp.md (movdd): Delete, combine with binary
27345 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
27346 with other moves by using conditional constraits (wg). Use LFIWZX
27347 and STFIWX for loading SDmode on power7.
27348 (movdd splitters): Likewise.
27349 (movdd_hardfloat32): Likewise.
27350 (movdd_softfloat32): Likewise.
27351 (movdd_hardfloat64_mfpgpr): Likewise.
27352 (movdd_hardfloat64): Likewise.
27353 (movdd_softfloat64): Likewise.
27355 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
27356 64-bit binary and decimal floating point moves.
27357 (FMOVE64X): Likewise.
27358 (movdf): Combine 64-bit binary and decimal floating point moves.
27359 Combine power6x (mfpgpr) moves with other moves by using
27360 conditional constraits (wg).
27361 (mov<mode> for DFmode/DDmode): Likewise.
27362 (DFmode/DDmode splitters): Likewise.
27363 (movdf_hardfloat32): Likewise.
27364 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
27365 (movdf_softfloat32): Likewise.
27366 (movdf_hardfloat64_mfpgpr): Likewise.
27367 (movdf_hardfloat64): Likewise.
27368 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
27369 (movdf_softfloat64): Likewise.
27370 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
27371 (reload_<mode>_load): Move to later in the file so they aren't in
27372 the middle of the floating point move insns.
27373 (reload_<mode>_store): Likewise.
27375 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
27378 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
27379 constraint if -mdebug=reg.
27380 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
27381 Enable using dd reload support if needed.
27383 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
27384 binary and decimal floating point moves in rs6000.md.
27385 (movtd_internal): Likewise.
27387 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
27388 decimal floating point moves.
27390 (movtf_internal): Likewise.
27391 (mov<mode>_internal, TDmode/TFmode): Likewise.
27392 (movtf_softfloat): Likewise.
27393 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
27395 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
27396 movdi_internal64, using wg constraint for move direct operations.
27397 (movdi_internal64): Likewise.
27399 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
27400 MODES_TIEABLE_P for selected modes. Print the numerical value of
27401 the various virtual registers. Use GPR/FPR first/last values,
27402 instead of hard coding the register numbers. Print which modes
27403 have reload functions registered.
27404 (rs6000_option_override_internal): If -mdebug=reg, trace the options
27405 settings before/after setting cpu, target and subtarget settings.
27406 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
27407 and for secondary reload failures in rs6000_secondary_reload_inner.
27408 (rs6000_secondary_reload_fail): Likewise.
27409 (rs6000_secondary_reload_inner): Likewise.
27411 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
27412 macros for first/last GPR and FPR registers.
27413 (LAST_GPR_REGNO): Likewise.
27414 (FIRST_FPR_REGNO): Likewise.
27415 (LAST_FPR_REGNO): Likewise.
27417 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
27418 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
27419 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
27420 (vcond<mode><mode>): Likewise.
27421 (vcondu<mode><mode>): Likewise.
27422 (vector_gtu<mode>): Likewise.
27423 (vector_gte<mode>): Likewise.
27424 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
27425 to prevent the compiler from converting DImode operations to TImode.
27426 (ior<mode>3): Likewise.
27427 (and<mode>3): Likewise.
27428 (one_cmpl<mode>2): Likewise.
27429 (nor<mode>3): Likewise.
27430 (andc<mode>3): Likewise.
27432 * config/rs6000/constraints.md (wt constraint): New constraint
27433 that returns VSX_REGS if TImode is allowed in VSX registers.
27435 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
27436 constant under VSX.
27438 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
27439 similar to TImode, but it is restricted to being in the GPRs.
27441 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
27442 TImode to occupy a single VSX register.
27444 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
27445 -mvsx-timode for power7/power8.
27446 (power7 cpu): Likewise.
27447 (power8 cpu): Likewise.
27449 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
27450 sure that TFmode/TDmode take up two registers if they are ever
27451 allowed in the upper VSX registers.
27452 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
27454 (rs6000_init_hard_regno_mode_ok): Likewise.
27455 (rs6000_debug_reg_global): Add debugging for PTImode and wt
27456 constraint. Print if LRA is turned on.
27457 (rs6000_option_override_internal): Give an error if -mvsx-timode
27458 and VSX is not enabled.
27459 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
27460 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
27461 to reg+offset addressing. Use PTImode when checking offset
27462 addresses for validity.
27463 (reg_offset_addressing_ok_p): Likewise.
27464 (rs6000_legitimate_offset_address_p): Likewise.
27465 (rs6000_legitimize_address): Likewise.
27466 (rs6000_legitimize_reload_address): Likewise.
27467 (rs6000_legitimate_address_p): Likewise.
27468 (rs6000_eliminate_indexed_memrefs): Likewise.
27469 (rs6000_emit_move): Likewise.
27470 (rs6000_secondary_reload): Likewise.
27471 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
27472 reloads to fpr registers to continue to use reg+offset addressing,
27473 but 64-bit reloads to altivec registers need reg+reg addressing.
27474 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
27475 it. Treat LO_SUM like a PLUS operation.
27476 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
27477 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
27478 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
27479 registers to share a register with a smaller sized type, since VSX
27480 puts scalars in the upper 64-bits.
27481 (print_operand): Add support for PTImode.
27482 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
27483 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
27484 registers, but don't have arithmetic support.
27485 (rs6000_memory_move_cost): Add test for VSX.
27486 (rs6000_opt_masks): Add -mvsx-timode.
27488 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
27491 (VSr): Use wt constraint for TImode.
27492 (VSv): Drop TImode support.
27493 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
27494 (vsx_movti_64bit): Likewise.
27495 (vsx_movti_32bit): Likewise.
27496 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
27497 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
27498 one '?' on the appropriate output constraint. Do not allow TImode
27499 logical operations on 32-bit systems.
27500 (vsx_ior<mode>3): Likewise.
27501 (vsx_xor<mode>3): Likewise.
27502 (vsx_one_cmpl<mode>2): Likewise.
27503 (vsx_nor<mode>3): Likewise.
27504 (vsx_andc<mode>3): Likewise.
27505 (vsx_concat_<mode>): Likewise.
27506 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
27508 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
27509 OPTION_MASK_VSX_TIMODE.
27510 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
27511 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
27513 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
27514 (TI2 iterator): New iterator for TImode, PTImode.
27515 (wd mode attribute): Add values for vector types.
27516 (movti_string): Replace TI move operations with operations for TImode
27517 and PTImode. Add support for TImode being allowed in VSX registers.
27518 (mov<mode>_string, TImode/PTImode): Likewise.
27519 (movti_ppc64): Likewise.
27520 (mov<mode>_ppc64, TImode/PTImode): Likewise.
27521 (TI mode splitters): Likewise.
27523 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
27526 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
27528 PR tree-optimization/56355
27529 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
27530 Also handle integers with undefined overflow.
27532 2013-03-20 Catherine Moore <clm@codesourcery.com>
27533 Maciej W. Rozycki <macro@codesourcery.com>
27534 Tom de Vries <tom@codesourcery.com>
27535 Nathan Sidwell <nathan@codesourcery.com>
27536 Iain Sandoe <iain@codesourcery.com>
27537 Nathan Froyd <froydnj@codesourcery.com>
27538 Chao-ying Fu <fu@mips.com>
27540 * doc/extend.texi: (micromips, nomicromips, nocompression):
27541 Document new function attributes.
27542 * doc/invoke.texi (minterlink-compressed, mmicromips,
27543 m14k, m14ke, m14kec): Document new options.
27544 (minterlink-mips16): Update documentation.
27545 * doc/md.texi (ZC, ZD): Document new constraints.
27546 * configure.ac (gcc_cv_as_micromips): Check if linker
27547 supports the .set micromips directive.
27548 * configure: Regenerate.
27549 * config.in: Regenerate.
27550 * config/mips/mips-tables.opt: Regenerate.
27551 * config/mips/micromips.md: New file.
27552 * constraints.md (ZC, ZD): New constraints.
27553 * config/mips/predicates.md (movep_src_register): New predicate.
27554 (movep_src_operand): New predicate.
27555 (non_volatile_mem_operand): New predicate.
27556 * config/mips/mips.md (multimem): New type.
27557 (length): Differentiate between 17-bit and 18-bit branch offsets.
27558 (MOVEP1, MOVEP2): New mode iterator.
27559 (mov_<load>l): Use ZC constraint.
27560 (mov_<load>r): Likewise.
27561 (mov_<store>l): Likewise.
27562 (mov_<store>r): Likewise.
27563 (*branch_equality<mode>_inverted): Add microMIPS support.
27564 (*branch_equality<mode>): Likewise.
27565 (*jump_absolute): Likewise.
27566 (indirect_jump_<mode>): Likewise.
27567 (tablejump_<mode>): Likewise.
27568 (<optab>_internal): Likewise.
27569 (sibcall_internal): Likewise.
27570 (sibcall_value_internal): Likewise.
27571 (prefetch): Use constraint ZD.
27572 * config/mips/mips.opt (minterlink-compressed): New option.
27573 (minterlink-mips16): Now an alias for minterlink-compressed.
27574 (mmicromips): New option.
27575 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
27576 (compare_and_swap_12): Likewise.
27577 (sync_add<mode>): Likewise.
27578 (sync_<optab>_12): Likewise.
27579 (sync_old_<optab>_12): Likewise.
27580 (sync_new_<optab>_12): Likewise.
27581 (sync_nand_12): Likewise.
27582 (sync_old_nand_12): Likewise.
27583 (sync_new_nand_12): Likewise.
27584 (sync_sub<mode>): Likewise.
27585 (sync_old_add<mode>): Likewise.
27586 (sync_old_sub<mode>): Likewise.
27587 (sync_new_add<mode>): Likewise.
27588 (sync_new_sub<mode>): Likewise.
27589 (sync_<optab><mode>): Likewise.
27590 (sync_old_<optab><mode>): Likewise.
27591 (sync_new_<optab><mode>): Likewise.
27592 (sync_nand<mode>): Likewise.
27593 (sync_old_nand<mode>): Likewise.
27594 (sync_new_nand<mode>): Likewise.
27595 (sync_lock_test_and_set<mode>): Likewise.
27596 (test_and_set_12): Likewise.
27597 (atomic_compare_and_swap<mode>): Likewise.
27598 (atomic_exchange<mode>_llsc): Likewise.
27599 (atomic_fetch_add<mode>_llsc): Likewise.
27600 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
27601 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
27602 (umips_save_restore_pattern_p): Likewise.
27603 (umips_load_store_pair_p): Likewise.
27604 (umips_output_load_store_pair): Likewise.
27605 (umips_movep_target_p): Likewise.
27606 (umips_12bit_offset_address_p): Likewise.
27607 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
27608 (mips_base_mips16): Rename this...
27609 (mips_base_compression_flags): ...to this. Update all uses.
27610 (mips_attribute_table): Add micromips, nomicromips and nocompression.
27611 (mips_mips16_decl_p): Delete.
27612 (mips_nomips16_decl_p): Delete.
27613 (mips_get_compress_on_flags): New function.
27614 (mips_get_compress_off_flags): New function.
27615 (mips_get_compress_mode): New function.
27616 (mips_get_compress_on_name): New function.
27617 (mips_get_compress_off_name): New function.
27618 (mips_insert_attributes): Support multiple compression types.
27619 (mips_merge_decl_attributes): Likewise.
27620 (umips_12bit_offset_address_p): New function.
27621 (mips_start_function_definition): Emit .set micromips directive.
27622 (mips_call_may_need_jalx_p): New function.
27623 (mips_function_ok_for_sibcall): Add microMIPS support.
27624 (mips_print_operand_punctuation): Support short delay slots and
27626 (umips_swm_mask, umips_swm_encoding): New.
27627 (umips_build_save_restore): New function.
27628 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
27629 (was_mips16_p): Remove.
27630 (old_compression_mode): New.
27631 (mips_set_compression_mode): New function.
27632 (mips_set_current_function): Add microMIPS support.
27633 (mips_option_override): Likewise.
27634 (umips_save_restore_pattern_p): New function.
27635 (umips_output_save_restore): New function.
27636 (umips_load_store_pair_p_1): New function.
27637 (umips_load_store_pair_p): New function.
27638 (umips_output_load_store_pair_1): New function.
27639 (umips_output_load_store_pair): New function.
27640 (umips_movep_target_p) New function.
27641 (mips_prepare_pch_save): Add microMIPS support.
27642 * config/mips/mips.h (TARGET_COMPRESSION): New.
27643 (TARGET_CPU_CPP_BUILTINS): Update macro
27644 to use new compression flags and to support microMIPS.
27645 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
27646 (MIPS_ARCH_FLOAT_SPEC): Likewise.
27647 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
27648 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
27649 (ASM_SPEC): Support mmicromips and mno-micromips.
27650 (M16STORE_REG_P): New macro.
27651 (MIPS_CALL): Support TARGET_MICROMIPS.
27652 (MICROMIPS_J): New macro.
27653 (mips_base_mips16): Rename this...
27654 (mips_base_compression_flags): ...to this.
27655 (UMIPS_12BIT_OFFSET_P): New macro.
27656 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
27657 (MULTILIB_DIRNAMES): Likewise.
27658 2013-03-20 Richard Biener <rguenther@suse.de>
27660 PR tree-optimization/56661
27661 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
27662 the result does not have to be distinct.
27664 2013-03-20 Richard Biener <rguenther@suse.de>
27666 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
27669 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
27670 Steven Bosscher <steven@gcc.gnu.org>
27672 PR rtl-optimization/56605
27673 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
27675 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
27678 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
27679 that require movd instead of movq.
27681 2013-03-20 Richard Biener <rguenther@suse.de>
27683 * tree-ssa-structalias.c (struct variable_info): Add pointer
27684 to the first field of an aggregate with sub-vars. Make
27685 this and the pointer to the next subfield its ID.
27686 (vi_next): New function.
27687 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
27688 storedanything_id, integer_id): Increment by one.
27689 (new_var_info, get_call_vi, lookup_call_clobber_vi,
27690 get_call_clobber_vi): Adjust.
27691 (solution_set_expand): Simplify and speedup.
27692 (solution_set_add): Inline into ...
27693 (set_union_with_increment): ... this. Adjust accordingly.
27694 (do_sd_constraint): Likewise.
27695 (do_ds_constraint): Likewise.
27696 (do_complex_constraint): Simplify.
27697 (build_pred_graph): Adjust.
27698 (solve_graph): Likewise. Simplify and speedup.
27699 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
27700 get_constraint_for_component_ref, get_constraint_for_1,
27701 first_vi_for_offset, first_or_preceding_vi_for_offset,
27702 create_function_info_for, create_variable_info_for_1,
27703 create_variable_info_for, intra_create_variable_infos): Adjust.
27704 (init_base_vars): Push NULL for ID zero.
27705 (compute_points_to_sets): Adjust.
27707 2013-03-20 Richard Biener <rguenther@suse.de>
27709 * cfgloop.c (verify_loop_structure): Streamline and avoid
27710 ICEing on corrupt loop tree.
27711 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
27714 2013-03-20 Richard Biener <rguenther@suse.de>
27716 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
27717 check whether an SSA update is needed.
27719 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
27721 * config/mips/constraints.md (T): Rename to...
27725 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
27726 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
27728 2013-03-19 Ian Bolton <ian.bolton@arm.com>
27730 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
27731 (*subsi3_carryin_uxtw): Likewise.
27733 2013-03-19 Ian Bolton <ian.bolton@arm.com>
27735 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
27736 (*rorsi3_insn_uxtw): Likewise.
27738 2013-03-19 Ian Bolton <ian.bolton@arm.com>
27740 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
27741 (*extrsi5_insn_uxtw): Likewise.
27743 2013-03-19 Richard Biener <rguenther@suse.de>
27745 PR tree-optimization/56273
27746 * passes.c (init_optimization_passes): Move second VRP after DOM.
27748 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
27750 * config/i386/i386.md (*movti_internal): Merge from
27751 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
27752 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
27753 nox64 isa attributes.
27755 2013-03-18 Richard Biener <rguenther@suse.de>
27757 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
27759 (merge_node_constraints): Likewise.
27760 (build_succ_graph): Likewise.
27761 (valid_graph_edge): Inline into single caller.
27762 (unify_nodes): Likewise. Use bitmap_set_bit return value
27764 (scc_visit): Fix formatting and variable use.
27765 (do_sd_constraint): Use gcc_checking_assert.
27766 (do_ds_constraint): Likewise.
27767 (do_complex_constraint): Likewise.
27768 (condense_visit): Likewise. Cleanup.
27769 (dump_pred_graph): New function.
27770 (perform_var_substitution): Dump the pred-graph before
27771 variable substitution.
27772 (find_equivalent_node): Use gcc_checking_assert.
27773 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
27775 2013-03-18 Richard Biener <rguenther@suse.de>
27777 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
27778 Remove cond_expr_stmt_list argument and do not gimplify the
27780 (vect_loop_versioning): Adjust.
27781 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
27782 Cleanup to use less temporaries.
27783 (vect_create_data_ref_ptr): Cleanup.
27785 2013-03-18 Jakub Jelinek <jakub@redhat.com>
27787 PR tree-optimization/56635
27788 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
27789 require types_compatible_p types.
27791 2013-03-18 Nick Clifton <nickc@redhat.com>
27793 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
27794 spurious backslash.
27796 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
27797 Add missing line to comment describing function.
27799 2013-03-18 Richard Biener <rguenther@suse.de>
27801 PR tree-optimization/56210
27802 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
27803 Handle string / character search functions.
27804 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
27806 2013-03-18 Richard Biener <rguenther@suse.de>
27808 PR middle-end/56483
27809 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
27810 and implement properly.
27811 * gimple.h (gimple_cond_single_var_p): Remove.
27813 2013-03-18 Richard Biener <rguenther@suse.de>
27815 * tree-data-ref.h (find_data_references_in_loop): Declare.
27816 * tree-data-ref.c (get_references_in_stmt): Use a stack
27817 vector pre-allocated in the callers.
27818 (find_data_references_in_stmt): Adjust.
27819 (graphite_find_data_references_in_stmt): Likewise.
27820 (create_rdg_vertices): Likewise.
27821 (find_data_references_in_loop): Export.
27822 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
27823 Compute dependences here...
27824 (vect_analyze_data_refs): ...not here. When we encounter
27825 a non-vectorizable data reference in basic-block vectorization
27826 truncate the data reference vector. Do not bother to
27827 fixup data-dependence information for gather loads.
27828 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
27829 of data references, as reported.
27831 2013-03-18 Richard Biener <rguenther@suse.de>
27833 PR tree-optimization/3713
27834 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
27835 has_constants and expr.
27836 (stmt_has_constants): Properly valueize SSA names when deciding
27837 whether the stmt has constants.
27839 2013-03-18 Richard Biener <rguenther@suse.de>
27841 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
27842 whole function when there is nothing to do.
27843 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
27844 * tree-vectorizer.c (vectorize_loops): Update virtual and
27845 loop-closed SSA once.
27846 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
27848 2013-03-18 Richard Biener <rguenther@suse.de>
27850 PR middle-end/56113
27851 * domwalk.c (bb_postorder): New global static.
27852 (cmp_bb_postorder): New function.
27853 (walk_dominator_tree): Replace scheme imposing an order for
27854 visiting dominator sons by one sorting them at the time they
27855 are pushed on the stack.
27857 2013-03-18 Richard Biener <rguenther@suse.de>
27859 PR tree-optimization/39326
27860 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
27861 (struct mem_ref): Replace mem member with ao_ref typed member.
27862 (MEM_ANALYZABLE): Adjust.
27863 (memref_eq): Likewise.
27864 (mem_ref_alloc): Likewise.
27865 (gather_mem_refs_stmt): Likewise.
27866 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
27867 (execute_sm_if_changed_flag_set): Adjust.
27868 (execute_sm): Likewise.
27869 (ref_always_accessed_p): Likewise.
27870 (refs_independent_p): Likewise.
27871 (can_sm_ref_p): Likewise.
27873 2013-03-18 Jakub Jelinek <jakub@redhat.com>
27876 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
27877 return 1 even for !unsignedp.
27879 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
27881 * config/i386/i386.md (isa): Add x64 and nox64.
27882 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
27883 (*pushtf): Enable *roF alternative for x64 isa only.
27884 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
27885 mode attribute of integer alternatives to DImode for TARGET_64BIT.
27886 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
27887 (*movtf_internal): Merge from *movtf_internal_rex64 and
27888 *movtf_internal_sse. Use x64 and nox64 isa attributes.
27889 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
27890 nox64 isa attributes.
27891 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
27892 nox64 isa attributes.
27893 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
27895 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
27897 * config/alpha/alpha.c (TARGET_LRA_P): New define.
27899 2013-03-17 Jakub Jelinek <jakub@redhat.com>
27902 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
27903 class names. Remove trailing comma after "ALL_REGS".
27905 2013-03-16 Jan Hubicka <jh@suse.cz>
27907 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
27908 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
27909 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
27910 of cgraph_get_create_node.
27911 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
27913 2013-03-16 Jason Merrill <jason@redhat.com>
27916 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
27917 with DW_AT_default_value.
27919 2013-03-16 Jakub Jelinek <jakub@redhat.com>
27921 * BASE-VER: Set to 4.9.0.
27923 2013-03-14 Andi Kleen <ak@linux.intel.com>
27926 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
27927 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
27928 Document _x* TSX intrinsics.
27930 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
27931 David Holsgrove <david.holsgrove@xilinx.com>
27933 * configure.ac: Add MicroBlaze TLS support detection.
27934 * configure: Regenerate.
27935 * config/microblaze/microblaze-protos.h
27936 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
27937 symbol_mentioned_p, label_mentioned_p): Add prototypes.
27938 * config/microblaze/microblaze.c (microblaze_address_type): Add
27939 ADDRESS_TLS and tls_reloc address types.
27940 (microblaze_address_info): Add tls_reloc.
27941 (TARGET_HAVE_TLS): Define.
27942 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
27943 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
27944 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
27945 load_tls_operand, microblaze_call_tls_get_addr,
27946 microblaze_legitimize_tls_address): New functions.
27947 (microblaze_classify_unspec): Handle UNSPEC_TLS.
27948 (get_base_reg): Use microblaze_tls_symbol_p.
27949 (microblaze_classify_address): Handle TLS.
27950 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
27951 label_mentioned_p and microblaze_tls_referenced_p.
27952 (microblaze_legitimize_address): Handle TLS.
27953 (microblaze_address_insns): Handle ADDRESS_TLS.
27954 (pic_address_needs_scratch): Handle TLS.
27955 (print_operand_address): Handle TLS.
27956 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
27957 (microblaze_expand_move): Handle TLS.
27958 (microblaze_legitimate_constant_p): Check
27959 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
27960 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
27961 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
27962 (PIC_OFFSET_TABLE_REGNUM): Set.
27963 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
27964 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
27965 (addsi3, movsi_internal2, movdf_internal): Update constraints
27966 * config/microblaze/predicates.md (arith_plus_operand): Define
27967 (move_operand): Redefine as move_src_operand,
27968 check microblaze_tls_referenced_p.
27970 2013-03-14 Ian Bolton <ian.bolton@arm.com>
27972 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
27973 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
27975 2013-03-14 Ian Bolton <ian.bolton@arm.com>
27977 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
27980 2013-03-14 Jakub Jelinek <jakub@redhat.com>
27982 PR tree-optimization/53265
27983 * common.opt (Waggressive-loop-optimizations): New option.
27984 * tree-ssa-loop-niter.c: Include tree-pass.h.
27985 (do_warn_aggressive_loop_optimizations): New function.
27986 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
27987 if number_of_latch_executions returned constant.
27988 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
27989 early. If number_of_latch_executions returned constant, set
27990 nb_iterations_upper_bound back to it.
27991 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
27993 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
27994 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
27996 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
27997 (MULTILIB_OSDIRNAMES): Set.
27998 * genmultilib: If defaultosdirname doesn't start with :: , set
27999 defaultosdirname2 instead, clear it and emit two . multilib_raw
28000 entries instead of just one.
28002 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
28004 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
28005 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
28006 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
28007 (SUBTARGET_OVERRIDE_OPTIONS): New.
28009 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
28012 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
28013 (musermode): Convert to Var(TARGET_USERMODE).
28014 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
28015 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
28016 * config/sh/sh.c (sh_option_override): Use
28017 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
28018 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
28020 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
28022 (udivsi3_i4_single, divsi3_i4_single): Use
28023 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
28025 2013-03-13 Dave Korn <dave.korn.cygwin@gmail.com>
28027 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
28030 2013-03-13 Richard Biener <rguenther@suse.de>
28032 PR tree-optimization/56608
28033 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
28034 calls when vectorizing basic-blocks.
28036 2013-03-13 Jakub Jelinek <jakub@redhat.com>
28039 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
28042 2013-03-12 Jakub Jelinek <jakub@redhat.com>
28044 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
28046 2013-03-11 Jan Hubicka <jh@suse.cz>
28049 * lto-streamer-out.c (output_symbol_p): Skip references from
28050 constructors of external variables.
28052 2013-03-11 Jan Hubicka <jh@suse.cz>
28054 PR middle-end/56571
28055 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
28057 * emit-rtl.c (verify_rtx_sharing): Likewise.
28058 (copy_insn_1): Likewise.
28059 * rtl.c (copy_rtx): Likewise.
28061 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
28064 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
28065 output_operand_lossage message.
28067 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
28070 * arm.c (shift_op): Validate RTL pattern on the fly.
28071 (arm_print_operand, case 'S'): Don't use shift_operator to validate
28074 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
28077 * config/pa/pa.md (call_value): Check for calls to powf and direct to
28078 new call patterns that clobber %fr12.
28079 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
28080 split and postreload patterns.
28081 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
28082 registers %fr12 and %fr12R as call used.
28084 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
28086 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
28087 (canon_address, record_store, replace_read, check_mem_read_rtx,
28088 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
28089 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
28090 rest_of_handle_dse): Likewise.
28092 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
28094 PR middle-end/56524
28095 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
28097 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
28098 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
28099 (save_optabs_if_changed): Replace with...
28100 (init_tree_optimization_optabs): ...this.
28101 * optabs.c (save_optabs_if_changed): Rename to...
28102 (init_tree_optimization_optabs): ...this. Take the optimization node
28103 as argument. Do nothing if the base optabs are already correct.
28104 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
28105 to recompute optabs.
28106 * function.h (function): Remove optabs field.
28107 * function.c (invoke_set_current_function_hook): Call
28108 init_tree_optimization_optabs. Use the result to initialize
28111 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
28113 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
28114 if GTMA_HAS_NO_INSTRUMENTATION.
28115 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
28116 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
28117 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
28118 * gimple-pretty-print.c (dump_gimple_transaction): Handle
28119 GTMA_HAS_NO_INSTRUMENTATION.
28121 2013-03-08 Jakub Jelinek <jakub@redhat.com>
28123 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
28126 2013-03-08 Marek Polacek <polacek@redhat.com>
28127 Jakub Jelinek <jakub@redhat.com>
28129 PR tree-optimization/56478
28130 * predict.c (is_comparison_with_loop_invariant_p): Change the
28131 type of loop_step to tree.
28132 (predict_loops): Adjust.
28133 (predict_iv_comparison): Perform the computations on double_ints.
28135 2013-03-08 Richard Biener <rguenther@suse.de>
28137 PR tree-optimization/56570
28138 * tree-cfg.c (verify_expr_location_1): Verify locations for
28140 * tree-sra.c (create_access_replacement): Strip locations
28141 from DECL_DEBUG_EXPRs.
28143 2013-03-08 Richard Biener <rguenther@suse.de>
28145 * tree-inline.c (expand_call_inline): Do not associate
28146 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
28147 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
28149 2013-03-08 Richard Biener <rguenther@suse.de>
28151 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
28152 or block changes with -Og. Fix for location / block encoding
28153 changes and PHI arguments with locations.
28155 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
28157 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
28159 (struct output_info): Likewise.
28160 (register_overhead): Remove bad gcc_assert.
28161 (bitmap_find_bit): If there is only a single bitmap element, do not
28162 count a miss as a search.
28163 (print_statistics): Update for counter type changes.
28164 (dump_bitmap_statistics): Likewise. Print headers such that they
28165 are properly lined up with the printed counters.
28167 2013-03-07 Jakub Jelinek <jakub@redhat.com>
28169 PR tree-optimization/56559
28170 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
28171 check that it has only a single use.
28173 2013-03-07 Richard Biener <rguenther@suse.de>
28175 * doc/invoke.texi (fwhole-program): Discourage use in combination
28178 2013-03-06 Jakub Jelinek <jakub@redhat.com>
28180 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
28182 PR tree-optimization/56539
28183 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
28184 instead of GSI_CONTINUE_LINKING as last argument to
28185 force_gimple_operand_gsi. Adjust function comment.
28187 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
28190 PR middle-end/56548
28191 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
28192 promoted mode, convert the result back to the original mode.
28194 2013-03-06 Richard Biener <rguenther@suse.de>
28196 PR middle-end/56294
28197 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
28198 (insert_updated_phi_nodes_compare_uids): New function.
28199 (update_ssa): Sort symbols_to_rename after UID before
28200 traversing it to insert PHI nodes.
28202 2013-03-06 Richard Biener <rguenther@suse.de>
28204 PR middle-end/50494
28205 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
28206 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
28209 2013-02-13 Richard Biener <rguenther@suse.de>
28212 * varasm.c (output_constant_def_1): Get the decl representing
28213 the constant as argument.
28214 (output_constant_def): Wrap output_constant_def_1.
28215 (make_decl_rtl): Use output_constant_def_1 with the decl
28216 representing the constant.
28217 (build_constant_desc): Optionally re-use a decl already
28218 representing the constant.
28219 (tree_output_constant_def): Adjust.
28221 2013-03-06 Joey Ye <joey.ye@arm.com>
28224 * gcc.c (convert_white_space): New function.
28225 (main): Handles white space in function name.
28227 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
28230 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
28231 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
28232 to SH_DIV_CALL_TABLE for TARGET_SH2.
28233 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
28235 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
28236 call-table options.
28238 2013-03-05 Sterling Augustine <saugustine@google.com>
28239 Cary Coutant <ccoutant@google.com>
28242 * dwarf2out.c (resolve_addr): Don't call
28243 remove_loc_list_addr_table_entries a second time for the same
28246 2013-03-05 Jakub Jelinek <jakub@redhat.com>
28249 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
28250 (avoid_complex_debug_insns): New function.
28251 (expand_debug_locations): Call it.
28253 PR rtl-optimization/56484
28254 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
28255 lifetimes of hard registers on small register class machines.
28257 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
28259 * config/microblaze/microblaze-protos.h: Rename
28260 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
28261 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
28263 (microblaze_fast_interrupt_function_p): New function.
28264 (microblaze_is_interrupt_handler): Rename to
28265 microblaze_is_interrupt_variant and add fast_interrupt check.
28266 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
28267 (save_restore_insns): Likewise.
28268 (compute_frame_size): Likewise.
28269 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
28270 (microblaze_globalize_label): Likewise.
28271 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
28272 * config/microblaze/microblaze.md: Use wrapper
28273 microblaze_is_interrupt_variant.
28275 2013-03-05 Kai Tietz <ktietz@redhat.com>
28277 * sdbout.c (sdbout_one_type): Switch to current function's section
28278 supporting cold/hot.
28280 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
28282 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
28285 2013-03-05 Jakub Jelinek <jakub@redhat.com>
28287 PR middle-end/56461
28288 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
28289 if VALGRIND_GET_VBITS is defined, temporarily make object
28290 memory all defined, and restore previous valgrind addressability
28291 and definability afterwards. Free this_object at the end.
28293 PR middle-end/56461
28294 * lra.c (lra): Call lra_clear_live_ranges if live_p,
28295 right before calling lra_create_live_ranges, also call it
28296 when clearing live_p. Only call lra_clear_live_ranges
28297 at the end if live_p.
28299 PR middle-end/56461
28300 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
28302 2013-03-05 Richard Biener <rguenther@suse.de>
28304 PR tree-optimization/56521
28305 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
28308 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
28311 * except.h (remove_unreachable_eh_regions): New prototype.
28312 * except.c (remove_eh_handler_splicer): New function, split out
28313 of remove_eh_handler.
28314 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
28315 warning about running it on many EH regions one at a time.
28316 (remove_unreachable_eh_regions_worker): New function, walk the
28317 EH tree in depth-first order and remove non-marked regions.
28318 (remove_unreachable_eh_regions): New function.
28319 * tree-eh.c (mark_reachable_handlers): New function, split out
28320 from remove_unreachable_handlers.
28321 (remove_unreachable_handlers): Use mark_reachable_handlers and
28322 remove_unreachable_eh_regions.
28323 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
28324 and remove_unreachable_eh_regions.
28326 2013-03-05 Richard Biener <rguenther@suse.de>
28328 PR middle-end/56525
28329 * loop-init.c (fix_loop_structure): Remove loops in two stages,
28330 not freeing them until the end.
28332 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
28334 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
28336 2013-03-05 Richard Biener <rguenther@suse.de>
28338 PR tree-optimization/56270
28339 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
28340 of loads after scheduling an SLP instance.
28342 2013-03-05 Jakub Jelinek <jakub@redhat.com>
28344 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
28346 (check_gcc_parallelize): Run guality.exp as a separate job from
28347 vect.exp with unsorted.exp and $(dg_target_exps) separately from
28348 struct-layout-1.exp with stackalign.exp.
28350 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
28352 PR middle-end/56461
28353 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
28354 load_index sbitmap even if some bit in it isn't set.
28356 PR middle-end/56461
28357 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
28358 (discover_iteration_bound_by_body_walk): Change queues to
28359 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
28360 spelling in comment. Call safe_push on queues[bound_index] directly.
28361 Release queues[queue_index] in every iteration unconditionally.
28362 Release bounds vector.
28364 PR middle-end/56461
28365 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
28366 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
28367 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
28370 2013-03-05 Richard Biener <rguenther@suse.de>
28373 * tree-inline.c (remap_blocks_to_null): New function.
28374 (expand_call_inline): When expanding a call stmt without
28375 an associated block inline remap all callee blocks to NULL.
28377 2013-03-05 Jakub Jelinek <jakub@redhat.com>
28379 PR rtl-optimization/56494
28380 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
28381 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
28382 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
28384 PR middle-end/56461
28385 * sel-sched-ir.c (free_sched_pools): Release
28386 succs_info_pool.stack[succs_info_pool.max_top] vectors too
28387 if succs_info_pool.max_top isn't -1.
28390 * opts.c (opts_obstack, opts_concat): Moved to...
28391 * opts-common.c (opts_obstack, opts_concat): ... here.
28393 2013-03-04 Jakub Jelinek <jakub@redhat.com>
28395 PR middle-end/56461
28396 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
28398 2013-03-04 Martin Jambor <mjambor@suse.cz>
28400 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
28401 all appropriate places.
28403 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
28405 PR tree-optimization/56424
28406 * ipa-split.c (split_function): Do not set the RSO flag if result is
28407 not by reference and its type is a register type.
28409 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
28411 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
28412 (microblaze_legitimate_pic_operand): Likewise
28413 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
28414 new function microblaze_legitimate_pic_operand
28415 * config/microblaze/microblaze-protos.h
28416 (microblaze_legitimate_pic_operand): Declare.
28418 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28420 * config/microblaze/predicates.md (call_insn_simple_operand):
28421 New predicate for supported rtx code types.
28422 * config/microblaze/microblaze.md (call_internal1): Use
28423 call_insn_simple_operand predicate.
28425 2013-03-04 Jakub Jelinek <jakub@redhat.com>
28427 PR middle-end/56461
28428 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
28429 partitions.ordered_remove.
28431 PR middle-end/56461
28432 * tree-vect-stmts.c (vectorizable_conversion): Don't call
28433 vec_oprnds0.create (1) for modifier == NONE.
28435 PR middle-end/56461
28436 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
28437 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
28438 vec_oprnds1 right before pushing anything to it for
28441 PR middle-end/56461
28442 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
28443 set nbbs to 0 instead of having separate code path.
28444 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
28445 instead of false as last argument if returning NULL.
28447 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
28449 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
28450 the attribute is now called "target" instead of "option".
28451 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
28452 * doc/tm.texi.in (Target Attributes): Likewise document the correct
28453 attribute/pragma name for TARGET_OPTION_VALID_P and
28454 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
28455 * doc/tm.texi: Regenerated.
28457 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
28459 * config/microblaze/microblaze.c:
28460 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
28461 * config/microblaze/microblaze.h: Add -mxl-reorder to
28463 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
28464 instructions emitted if TARGET_REORDER.
28465 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
28466 or 0 for -m/-mno case, but initialises as 2 to detect default use case
28469 2013-03-01 Xinliang David Li <davidxl@google.com>
28471 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
28474 2013-03-01 Jakub Jelinek <jakub@redhat.com>
28476 PR middle-end/56461
28477 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
28478 vector even when returning true. Fix up function comment formatting.
28480 PR middle-end/56461
28481 * ira-build.c (ira_loop_nodes_count): New variable.
28482 (create_loop_tree_nodes): Initialize it.
28483 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
28485 PR middle-end/56461
28486 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
28487 method on dr_chain and result_chain.
28488 * tree-vect-stmts.c (vectorizable_store): Only call
28489 result_chain.create if j == 0.
28491 PR middle-end/56461
28492 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
28493 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
28494 before overwriting it.
28496 2013-03-01 Tobias Burnus <burnus@net-b.de>
28498 * doc/extended.texi (C Extensions): Change order in @menu
28500 (Other MIPS Built-in Functions): Move last MIPS entry before
28501 "picoChip Built-in Functions".
28502 (SH Built-in Functions): Move after RX Built-in Functions.
28503 * doc/gcc.texi (Introduction): Change order in @menu
28505 * doc/md.texi (Constraints): Ditto.
28506 * gty.texi (Type Information): Ditto.
28507 (User-provided marking routines for template types): Make
28509 * doc/invoke.texi (AArch64 Options): Move before
28510 "Adapteva Epiphany Options".
28512 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
28513 Jakub Jelinek <jakub@redhat.com>
28516 * asan.c (gate_asan): Lookup no_sanitize_address instead of
28517 no_address_safety_analysis attribute.
28518 * doc/extend.texi (no_address_safety_attribute): Rename to
28519 no_sanitize_address attribute, mention no_address_safety_analysis
28520 attribute as deprecated alias.
28522 2013-02-28 Jakub Jelinek <jakub@redhat.com>
28524 PR middle-end/56461
28525 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
28526 type to vec<vec<tree> > *.
28527 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
28528 to be vec<tree> instead of vec<tree> *, set vec_defs
28529 to vNULL and call vec_defs.create (number_of_vects), adjust other
28531 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
28532 vectorizable_condition): Adjust vect_get_slp_defs callers.
28534 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
28536 * config/aarch64/aarch64.c
28537 (aarch64_float_const_representable): Remove unused variable.
28539 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
28541 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
28543 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
28545 * config/aarch64/aarch64-builtins.c
28546 (aarch64_init_simd_builtins): Make static.
28548 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
28550 * config/aarch64/aarch64.c
28551 (aarch64_simd_make_constant): Make static.
28553 2013-02-28 Martin Jambor <mjambor@suse.cz>
28555 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
28556 with no initialization to the RHS of debug statements.
28558 2013-02-28 Martin Jambor <mjambor@suse.cz>
28560 PR tree-optimization/56294
28561 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
28563 (get_access_replacement): Do not call create_access_replacement.
28564 Assert a replacement exists.
28565 (get_repl_default_def_ssa_name): Create the replacement declaration
28568 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
28570 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
28571 final_end_function.
28573 2013-02-28 Marek Polacek <polacek@redhat.com>
28575 PR rtl-optimization/56466
28576 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
28577 if we're changing a loop.
28578 (peel_loops_completely): Likewise.
28580 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
28583 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
28585 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
28588 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
28589 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
28590 INTX_FTYPE_FX, FX_FTYPE_INTX.
28591 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
28593 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
28595 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
28596 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
28597 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
28598 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
28599 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
28600 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
28601 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
28602 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
28603 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
28604 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
28605 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
28606 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
28607 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
28608 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
28609 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
28610 (avrxmega6): Increase max flash segments from 5 to 6.
28611 * config/avr/t-multilib: Regenerate.
28612 * config/avr/avr-tables.opt: Regenerate.
28613 * doc/avr-mmcu.texi: Regenerate.
28615 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
28617 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
28618 (avr_device_to_arch): Rename to avr_device_to_ld.
28619 (avr_device_to_as): New prototype.
28620 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
28621 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
28622 * config/avr/driver-avr.c (avr_device_to_as): New.
28623 (avr_device_to_arch): Rename to avr_device_to_ld.
28625 2013-02-27 Jakub Jelinek <jakub@redhat.com>
28627 PR middle-end/56461
28628 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
28629 method on dr_chain and result_chain.
28631 PR middle-end/56461
28632 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
28633 pointer_set_destroy on not_executed_last_iteration.
28635 PR middle-end/56461
28636 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
28638 PR middle-end/56461
28639 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
28640 FOR_EACH_DEFINED_FUNCTION when freeing state.
28642 PR middle-end/56461
28643 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
28645 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
28648 PR middle-end/56461
28649 * ipa-cp.c (decide_whether_version_node): Call vec_free on
28650 known_aggs[i].items and release known_aggs vector.
28652 PR middle-end/56461
28653 * ipa-reference.c (propagate): Free node_info even for alias nodes.
28655 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28657 * config/microblaze/microblaze.c (microblaze_emit_compare):
28658 Use xor for EQ/NE comparisions.
28659 * config/microblaze/microblaze.md (cstoresf4): Add constraints
28660 (cbranchsf4): Adjust operator to comparison_operator.
28662 2013-02-27 Jakub Jelinek <jakub@redhat.com>
28664 PR middle-end/56461
28665 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
28667 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
28668 vec_safe_push, always update *slot.
28669 (redirect_edge_var_map_clear): Use vec_free.
28670 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
28671 (free_var_map_entry): Use vec_free.
28672 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
28673 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
28675 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
28677 PR middle-end/45472
28678 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
28679 when the may_trap_p bit of the exprs being merged differs.
28680 Reorder tests for speculativeness in the logical and operator.
28682 2013-02-27 Jakub Jelinek <jakub@redhat.com>
28684 * incpath.c (add_standard_paths): Use reconcat instead of concat
28685 where appropriate and avoid leaking memory.
28687 * opts.h: Include obstack.h.
28688 (opts_concat): New prototype.
28689 (opts_obstack): New declaration.
28690 * opts.c (opts_concat): New function.
28691 (opts_obstack): New variable.
28692 (init_options_struct): Call gcc_init_obstack on opts_obstack.
28693 (finish_options): Use opts_concat instead of concat
28694 and XOBNEWVEC instead of XNEWVEC.
28695 * opts-common.c (generate_canonical_option, decode_cmdline_option,
28696 generate_option): Likewise.
28697 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
28698 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
28701 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
28702 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
28704 2013-02-26 Jakub Jelinek <jakub@redhat.com>
28706 PR middle-end/56461
28707 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
28709 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
28711 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
28712 (arm_block_move_unaligned_straight): Likewise.
28713 (arm_adjust_block_mem): Likewise.
28715 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
28718 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
28719 temp, cond and label.
28720 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
28723 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
28724 * config/c6x/c6x.h (dbx_register_map): Update declaration.
28727 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
28728 of prologue/epilogue functions.
28731 * config/tilegx/tilegx.c (tilegx_expand_prologue):
28732 Remove unused variable cfa_offset.
28733 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
28736 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
28737 type promotion to unsigned.
28740 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
28741 for HOST_WIDE_INT of 32 bit / same size as int.
28742 (arm_block_move_unaligned_straight): Likewise.
28743 (arm_adjust_block_mem): Likewise.
28746 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
28749 2013-02-26 Marek Polacek <polacek@redhat.com>
28751 PR tree-optimization/56426
28752 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
28754 2013-02-26 Richard Biener <rguenther@suse.de>
28757 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
28758 unused variable loops.
28760 2013-02-26 Jakub Jelinek <jakub@redhat.com>
28762 PR tree-optimization/56448
28763 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
28764 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
28765 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
28766 later operands of the references, or even first operand for
28767 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
28769 PR tree-optimization/56443
28770 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
28771 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
28772 to type_for_mode langhook.
28774 2013-02-25 Matt Turner <mattst88@gmail.com>
28776 * doc/invoke.texi: Document r4700.
28778 2013-02-25 Richard Biener <rguenther@suse.de>
28780 PR tree-optimization/56175
28781 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
28783 (simplify_bitwise_binary): ... here. Also guard the conversion
28784 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
28786 2013-02-25 Catherine Moore <clm@codesourcery.com>
28789 2013-02-24 Catherine Moore <clm@codesourcery.com>
28790 Maciej W. Rozycki <macro@codesourcery.com>
28791 Tom de Vries <tom@codesourcery.com>
28792 Nathan Sidwell <nathan@codesourcery.com>
28793 Iain Sandoe <iain@codesourcery.com>
28794 Nathan Froyd <froydnj@codesourcery.com>
28795 Chao-ying Fu <fu@mips.com>
28797 * doc/extend.texi: (micromips, nomicromips, nocompression):
28798 Document new function attributes.
28799 * doc/invoke.texi (minterlink-compressed, mmicromips,
28800 m14k, m14ke, m14kec): Document new options.
28801 (minterlink-mips16): Update documentation.
28802 * doc/md.texi (ZC, ZD): Document new constraints.
28803 * configure.ac (gcc_cv_as_micromips): Check if linker
28804 supports the .set micromips directive.
28805 * configure: Regenerate.
28806 * config.in: Regenerate.
28807 * config/mips/mips-tables.opt: Regenerate.
28808 * config/mips/micromips.md: New file.
28809 * constraints.md (ZC, AD): New constraints.
28810 * config/mips/predicates.md (movep_src_register): New predicate.
28811 (movep_src_operand): New predicate.
28812 (non_volatile_mem_operand): New predicate.
28813 * config/mips/mips.md (multimem): New type.
28814 (length): Differentiate between 17-bit and 18-bit branch offsets.
28815 (MOVEP1, MOVEP2): New mode iterator.
28816 (mov_<load>l): Use ZC constraint.
28817 (mov_<load>r): Likewise.
28818 (mov_<store>l): Likewise.
28819 (mov_<store>r): Likewise.
28820 (*branch_equality<mode>_inverted): Add microMIPS support.
28821 (*branch_equality<mode>): Likewise.
28822 (*jump_absolute): Likewise.
28823 (indirect_jump_<mode>): Likewise.
28824 (tablejump_<mode>): Likewise.
28825 (<optab>_internal): Likewise.
28826 (sibcall_internal): Likewise.
28827 (sibcall_value_internal): Likewise.
28828 (prefetch): Use constraint ZD.
28829 * config/mips/mips.opt (minterlink-compressed): New option.
28830 (minterlink-mips16): Now an alias for minterlink-compressed.
28831 (mmicromips): New option.
28832 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
28833 (compare_and_swap_12): Likewise.
28834 (sync_add<mode>): Likewise.
28835 (sync_<optab>_12): Likewise.
28836 (sync_old_<optab>_12): Likewise.
28837 (sync_new_<optab>_12): Likewise.
28838 (sync_nand_12): Likewise.
28839 (sync_old_nand_12): Likewise.
28840 (sync_new_nand_12): Likewise.
28841 (sync_sub<mode>): Likewise.
28842 (sync_old_add<mode>): Likewise.
28843 (sync_old_sub<mode>): Likewise.
28844 (sync_new_add<mode>): Likewise.
28845 (sync_new_sub<mode>): Likewise.
28846 (sync_<optab><mode>): Likewise.
28847 (sync_old_<optab><mode>): Likewise.
28848 (sync_new_<optab><mode>): Likewise.
28849 (sync_nand<mode>): Likewise.
28850 (sync_old_nand<mode>): Likewise.
28851 (sync_new_nand<mode>): Likewise.
28852 (sync_lock_test_and_set<mode>): Likewise.
28853 (test_and_set_12): Likewise.
28854 (atomic_compare_and_swap<mode>): Likewise.
28855 (atomic_exchange<mode>_llsc): Likewise.
28856 (atomic_fetch_add<mode>_llsc): Likewise.
28857 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
28858 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
28859 (umips_save_restore_pattern_p): Likewise.
28860 (umips_load_store_pair_p): Likewise.
28861 (umips_output_load_store_pair): Likewise.
28862 (umips_movep_target_p): Likewise.
28863 (umips_12bit_offset_address_p): Likewise.
28864 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
28865 (mips_base_mips16): Rename this...
28866 (mips_base_compression_flags): ...to this. Update all uses.
28867 (mips_attribute_table): Add micromips, nomicromips and nocompression.
28868 (mips_mips16_decl_p): Delete.
28869 (mips_nomips16_decl_p): Delete.
28870 (mips_get_compress_on_flags): New function.
28871 (mips_get_compress_off_flags): New function.
28872 (mips_get_compress_mode): New function.
28873 (mips_get_compress_on_name): New function.
28874 (mips_get_compress_off_name): New function.
28875 (mips_insert_attributes): Support multiple compression types.
28876 (mips_merge_decl_attributes): Likewise.
28877 (umips_12bit_offset_address_p): New function.
28878 (mips_start_function_definition): Emit .set micromips directive.
28879 (mips_call_may_need_jalx_p): New function.
28880 (mips_function_ok_for_sibcall): Add microMIPS support.
28881 (mips_print_operand_punctuation): Support short delay slots and
28883 (umips_swm_mask, umips_swm_encoding): New.
28884 (umips_build_save_restore): New function.
28885 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
28886 (was_mips16_p): Remove.
28887 (old_compression_mode): New.
28888 (mips_set_compression_mode): New function.
28889 (mips_set_current_function): Add microMIPS support.
28890 (mips_option_override): Likewise.
28891 (umips_save_restore_pattern_p): New function.
28892 (umips_output_save_restore): New function.
28893 (umips_load_store_pair_p_1): New function.
28894 (umips_load_store_pair_p): New function.
28895 (umips_output_load_store_pair_1): New function.
28896 (umips_output_load_store_pair): New function.
28897 (umips_movep_target_p) New function.
28898 (mips_prepare_pch_save): Add microMIPS support.
28899 * config/mips/mips.h (TARGET_COMPRESSION): New.
28900 (TARGET_CPU_CPP_BUILTINS): Update macro
28901 to use new compression flags and to support microMIPS.
28902 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
28903 (MIPS_ARCH_FLOAT_SPEC): Likewise.
28904 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
28905 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
28906 (ASM_SPEC): Support mmicromips and mno-micromips.
28907 (M16STORE_REG_P): New macro.
28908 (MIPS_CALL): Support TARGET_MICROMIPS.
28909 (MICROMIPS_J): New macro.
28910 (mips_base_mips16): Rename this...
28911 (mips_base_compression_flags): ...to this.
28912 (UMIPS_12BIT_OFFSET_P): New macro.
28913 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
28914 (MULTILIB_DIRNAMES): Likewise.
28916 2013-02-25 Tom de Vries <tom@codesourcery.com>
28918 PR rtl-optimization/56131
28919 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
28920 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
28921 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
28923 2013-02-25 Tobias Burnus <burnus@net-b.de>
28925 * doc/invoke.texi (-fsanitize=): Move from optimization
28926 to debugging options.
28928 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
28930 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
28932 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
28933 Alexander Monakov <amonakov@ispras.ru>
28935 PR middle-end/56077
28936 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
28937 flush pending lists also on non-jumps. Adjust comment.
28939 2013-02-24 Catherine Moore <clm@codesourcery.com>
28940 Maciej W. Rozycki <macro@codesourcery.com>
28941 Tom de Vries <tom@codesourcery.com>
28942 Nathan Sidwell <nathan@codesourcery.com>
28943 Iain Sandoe <iain@codesourcery.com>
28944 Nathan Froyd <froydnj@codesourcery.com>
28945 Chao-ying Fu <fu@mips.com>
28947 * doc/extend.texi: (micromips, nomicromips, nocompression):
28948 Document new function attributes.
28949 * doc/invoke.texi (minterlink-compressed, mmicromips,
28950 m14k, m14ke, m14kec): Document new options.
28951 (minterlink-mips16): Update documentation.
28952 * doc/md.texi (ZC, ZD): Document new constraints.
28953 * configure.ac (gcc_cv_as_micromips): Check if linker
28954 supports the .set micromips directive.
28955 * configure: Regenerate.
28956 * config.in: Regenerate.
28957 * config/mips/mips-tables.opt: Regenerate.
28958 * config/mips/micromips.md: New file.
28959 * constraints.md (ZC, AD): New constraints.
28960 * config/mips/predicates.md (movep_src_register): New predicate.
28961 (movep_src_operand): New predicate.
28962 (non_volatile_mem_operand): New predicate.
28963 * config/mips/mips.md (multimem): New type.
28964 (length): Differentiate between 17-bit and 18-bit branch offsets.
28965 (MOVEP1, MOVEP2): New mode iterator.
28966 (mov_<load>l): Use ZC constraint.
28967 (mov_<load>r): Likewise.
28968 (mov_<store>l): Likewise.
28969 (mov_<store>r): Likewise.
28970 (*branch_equality<mode>_inverted): Add microMIPS support.
28971 (*branch_equality<mode>): Likewise.
28972 (*jump_absolute): Likewise.
28973 (indirect_jump_<mode>): Likewise.
28974 (tablejump_<mode>): Likewise.
28975 (<optab>_internal): Likewise.
28976 (sibcall_internal): Likewise.
28977 (sibcall_value_internal): Likewise.
28978 (prefetch): Use constraint ZD.
28979 * config/mips/mips.opt (minterlink-compressed): New option.
28980 (minterlink-mips16): Now an alias for minterlink-compressed.
28981 (mmicromips): New option.
28982 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
28983 (compare_and_swap_12): Likewise.
28984 (sync_add<mode>): Likewise.
28985 (sync_<optab>_12): Likewise.
28986 (sync_old_<optab>_12): Likewise.
28987 (sync_new_<optab>_12): Likewise.
28988 (sync_nand_12): Likewise.
28989 (sync_old_nand_12): Likewise.
28990 (sync_new_nand_12): Likewise.
28991 (sync_sub<mode>): Likewise.
28992 (sync_old_add<mode>): Likewise.
28993 (sync_old_sub<mode>): Likewise.
28994 (sync_new_add<mode>): Likewise.
28995 (sync_new_sub<mode>): Likewise.
28996 (sync_<optab><mode>): Likewise.
28997 (sync_old_<optab><mode>): Likewise.
28998 (sync_new_<optab><mode>): Likewise.
28999 (sync_nand<mode>): Likewise.
29000 (sync_old_nand<mode>): Likewise.
29001 (sync_new_nand<mode>): Likewise.
29002 (sync_lock_test_and_set<mode>): Likewise.
29003 (test_and_set_12): Likewise.
29004 (atomic_compare_and_swap<mode>): Likewise.
29005 (atomic_exchange<mode>_llsc): Likewise.
29006 (atomic_fetch_add<mode>_llsc): Likewise.
29007 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
29008 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
29009 (umips_save_restore_pattern_p): Likewise.
29010 (umips_load_store_pair_p): Likewise.
29011 (umips_output_load_store_pair): Likewise.
29012 (umips_movep_target_p): Likewise.
29013 (umips_12bit_offset_address_p): Likewise.
29014 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
29015 (mips_base_mips16): Rename this...
29016 (mips_base_compression_flags): ...to this. Update all uses.
29017 (mips_attribute_table): Add micromips, nomicromips and nocompression.
29018 (mips_mips16_decl_p): Delete.
29019 (mips_nomips16_decl_p): Delete.
29020 (mips_get_compress_on_flags): New function.
29021 (mips_get_compress_off_flags): New function.
29022 (mips_get_compress_mode): New function.
29023 (mips_get_compress_on_name): New function.
29024 (mips_get_compress_off_name): New function.
29025 (mips_insert_attributes): Support multiple compression types.
29026 (mips_merge_decl_attributes): Likewise.
29027 (umips_12bit_offset_address_p): New function.
29028 (mips_start_function_definition): Emit .set micromips directive.
29029 (mips_call_may_need_jalx_p): New function.
29030 (mips_function_ok_for_sibcall): Add microMIPS support.
29031 (mips_print_operand_punctuation): Support short delay slots and
29033 (umips_swm_mask, umips_swm_encoding): New.
29034 (umips_build_save_restore): New function.
29035 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
29036 (was_mips16_p): Remove.
29037 (old_compression_mode): New.
29038 (mips_set_compression_mode): New function.
29039 (mips_set_current_function): Add microMIPS support.
29040 (mips_option_override): Likewise.
29041 (umips_save_restore_pattern_p): New function.
29042 (umips_output_save_restore): New function.
29043 (umips_load_store_pair_p_1): New function.
29044 (umips_load_store_pair_p): New function.
29045 (umips_output_load_store_pair_1): New function.
29046 (umips_output_load_store_pair): New function.
29047 (umips_movep_target_p) New function.
29048 (mips_prepare_pch_save): Add microMIPS support.
29049 * config/mips/mips.h (TARGET_COMPRESSION): New.
29050 (TARGET_CPU_CPP_BUILTINS): Update macro
29051 to use new compression flags and to support microMIPS.
29052 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
29053 (MIPS_ARCH_FLOAT_SPEC): Likewise.
29054 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
29055 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
29056 (ASM_SPEC): Support mmicromips and mno-micromips.
29057 (M16STORE_REG_P): New macro.
29058 (MIPS_CALL): Support TARGET_MICROMIPS.
29059 (MICROMIPS_J): New macro.
29060 (mips_base_mips16): Rename this...
29061 (mips_base_compression_flags): ...to this.
29062 (UMIPS_12BIT_OFFSET_P): New macro.
29063 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
29064 (MULTILIB_DIRNAMES): Likewise.
29066 2013-02-24 Jakub Jelinek <jakub@redhat.com>
29069 * target-globals.c (save_target_globals): For init_reg_sets and
29070 target_reinit remporarily set this_fn_optabs to this_target_optabs.
29072 2013-02-22 James Greenhalgh <james.greenhalgh@arm.com>
29074 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
29075 * config/aarch64/t-aarch64
29076 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
29078 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
29080 PR inline-asm/56148
29081 * lra-constraints.c (process_alt_operands): Reload operand
29082 conflicting with earlier clobber only if no more other conflicting
29085 2013-02-22 Jakub Jelinek <jakub@redhat.com>
29088 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
29089 if not linking a shared library.
29091 2013-02-22 Seth LaForge <sethml@google.com>
29093 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
29095 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
29097 * config/arm/arm.md (split for extendsidi): Update condition.
29098 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
29099 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
29100 (qhs_zextenddi_cstr): Likewise.
29102 2013-02-21 Jakub Jelinek <jakub@redhat.com>
29104 PR middle-end/56420
29105 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
29106 avoid signed wrapping.
29107 (expand_mult): Handle properly multiplication by
29108 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
29109 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
29110 in the compiler if coeff is HOST_WIDE_INT_MIN.
29111 (expand_divmod): Don't make ext_op1 static, change it's type to
29112 uhwi. Avoid undefined behavior in -INTVAL (op1).
29114 PR rtl-optimization/50339
29115 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
29117 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
29118 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
29119 into splitting_ashiftrt field.
29120 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
29122 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
29125 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
29127 PR middle-end/56108
29128 * trans-mem.c (execute_tm_mark): Do not expand transactions that
29129 are sure to go irrevocable.
29131 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
29133 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
29134 scalars are valid operands.
29136 2013-02-21 Martin Jambor <mjambor@suse.cz>
29138 PR tree-optimization/56310
29139 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
29140 only matching indices and non-negative final offsets.
29141 (intersect_aggregates_with_edge): Pass src_idx to
29142 agg_replacements_to_vector. Pass src_idx insstead of index to
29143 intersect_with_agg_replacements.
29145 2013-02-21 Martin Jambor <mjambor@suse.cz>
29147 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
29148 instead of hard-wired defaults.
29150 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
29152 * doc/invoke.texi (MIPS Options): Update documentation of the
29153 floating-point multiply-accumulate instruction restrictions.
29155 2013-02-21 Kostya Serebryany <kcc@google.com>
29157 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
29158 asan_shadow_offset on x86_64 linux.
29160 2013-02-21 Richard Biener <rguenther@suse.de>
29162 PR tree-optimization/56415
29164 2013-02-11 Richard Biener <rguenther@suse.de>
29166 PR tree-optimization/56273
29167 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
29170 2013-02-21 Jakub Jelinek <jakub@redhat.com>
29173 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
29176 PR inline-asm/56405
29177 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
29178 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
29180 2013-02-20 Jan Hubicka <jh@suse.cz>
29182 PR tree-optimization/56265
29183 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
29184 when target is referenced for first time.
29186 2013-02-20 Richard Biener <rguenther@suse.de>
29188 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
29189 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
29190 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
29191 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
29192 not return anything.
29193 (rename_ssa_copies): Do not remove unused locals.
29194 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
29195 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
29196 * passes.c (execute_function_todo): Do not schedule unused locals
29197 removal if cleanup_tree_cfg did something.
29198 * tree-ssa-live.c (remove_unused_locals): Dump statistics
29199 about the number of removed locals.
29201 2013-02-20 Richard Biener <rguenther@suse.de>
29203 PR tree-optimization/56398
29204 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
29206 2013-02-20 Martin Jambor <mjambor@suse.cz>
29208 PR tree-optimization/55334
29209 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
29210 restricted pointers to arrays.
29212 2013-02-20 Richard Biener <rguenther@suse.de>
29213 Jakub Jelinek <jakub@redhat.com>
29215 PR tree-optimization/56396
29216 * tree-ssa-ccp.c (n_const_val): New static variable.
29217 (get_value): Return NULL for SSA names we don't have a lattice
29219 (ccp_initialize): Initialize n_const_val.
29220 * tree-ssa-copy.c (n_copy_of): New static variable.
29221 (init_copy_prop): Initialize n_copy_of.
29222 (get_value): Return NULL_TREE for SSA names we don't have a
29225 2013-02-20 Martin Jambor <mjambor@suse.cz>
29227 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
29229 2013-02-20 Richard Biener <rguenther@suse.de>
29231 * genpreds.c (write_lookup_constraint): Do not compare first
29232 letter of the constraint again.
29234 2013-02-20 Richard Biener <rguenther@suse.de>
29236 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
29238 (get_use_iv_cost): Terminate hashtable walk when coming across
29241 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
29243 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
29244 reassociation for avx2 targets.
29246 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
29248 * config/microblaze/microblaze.c: microblaze_has_clz = 0
29249 Add version check for v8.10.a to enable microblaze_has_clz
29250 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
29251 version and TARGET_PATTERN_COMPARE check
29252 * config/microblaze/microblaze.md: New clzsi2 instruction
29254 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
29256 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
29257 function before branching.
29259 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
29261 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
29263 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
29265 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
29267 PR middle-end/55889
29268 * sel-sched.c: Include ira.h.
29269 (implicit_clobber_conflict_p): New function.
29270 (moveup_expr): Use it.
29271 * Makefile.in (sel-sched.o): Depend on ira.h.
29273 2013-02-19 Richard Biener <rguenther@suse.de>
29275 PR tree-optimization/56384
29276 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
29277 (vn_hash_type): Split out from ...
29278 (vn_hash_constant_with_type): ... here.
29279 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
29280 (vn_phi_eq): Compare types from vn_phi_s structure.
29281 (vn_phi_lookup): Populate vn_phi_s type.
29282 (vn_phi_insert): Likewise.
29284 2013-02-19 Jakub Jelinek <jakub@redhat.com>
29286 PR tree-optimization/56350
29287 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
29288 if haven't found reduction or nested cycle operand, rather than
29289 asserting we must find it.
29291 PR tree-optimization/56381
29292 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
29295 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
29296 Jakub Jelinek <jakub@redhat.com>
29299 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
29300 (swap_optab_enable): Same.
29301 (init_all_optabs): Use argument instead of global.
29302 * tree.h (struct tree_optimization_option): New field target_optabs.
29303 * expr.h (init_all_optabs): Add argument to prototype.
29304 (TREE_OPTIMIZATION_OPTABS): New.
29305 (save_optabs_if_changed): Protoize.
29306 * optabs.h: Declare this_fn_optabs.
29307 * optabs.c (save_optabs_if_changed): New.
29308 Declare this_fn_optabs.
29309 (init_optabs): Add argument to init_all_optabs() call.
29310 * function.c (invoke_set_current_function_hook): Handle per
29312 * function.h (struct function): New field optabs.
29313 * config/mips/mips.c (mips_set_mips16_mode): Handle when
29314 optimization_current_node has changed.
29315 * target-globals.h (save_target_globals_default_opts): Protoize.
29316 * target-globals.c (save_target_globals_default_opts): New.
29318 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29321 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
29322 registers %fr12 and %fr12R as call used.
29325 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
29326 and HImode, require all displacements to be an integer multiple of
29328 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
29329 only allow QImode and HImode when reload is in progress and strict is
29330 true. Likewise for symbolic addresses. Use base14_operand to check
29331 displacements in REG+BASE addresses.
29333 2013-02-18 Richard Biener <rguenther@suse.de>
29335 PR tree-optimization/56366
29336 * tree-vect-loop.c (get_initial_def_for_induction): Properly
29337 handle sign-conversion of outer-loop initial induction value.
29339 2013-02-18 Richard Biener <rguenther@suse.de>
29341 PR middle-end/56349
29342 * cfghooks.c (merge_blocks): If we merge a latch into another
29343 block adjust references to it.
29344 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
29345 (verify_loop_structure): Verify that a recorded latch is in fact
29348 2013-02-18 Richard Biener <rguenther@suse.de>
29350 PR tree-optimization/56321
29351 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
29352 order SSA name release and virtual operand unlinking.
29354 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
29356 * config/microblaze/microblaze.md (save_stack_block): Define.
29357 (restore_stack_block): Likewise.
29359 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
29361 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
29362 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
29363 * config/microblaze/microblaze.c (microblaze_option_override):
29364 Bail out early for PIC modes when target does not support PIC.
29366 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
29368 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
29369 Replace with a microblaze version.
29370 (microblaze_trampoline_init): Adapt for microblaze.
29371 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
29374 2013-02-16 Jakub Jelinek <jakub@redhat.com>
29375 Dodji Seketeli <dodji@redhat.com>
29378 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
29379 (instrument_mem_region_access): Do not forget to always put
29380 instrumentation of the of 'base' and 'base + len' in a "if (len !=
29381 0) statement, even for cases where either 'base' or 'base + len'
29382 are not instrumented -- because they have been previously
29383 instrumented. Simplify the logic by putting all the statements
29384 instrument 'base + len' inside a sequence, and then insert that
29385 sequence right before the current insertion point. Then, to
29386 instrument 'base + len', just get an iterator on that statement.
29387 And do not forget to update the pointer to iterator the function
29388 received as argument.
29390 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
29392 PR rtl-optimization/56348
29393 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
29395 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
29397 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
29398 (clean_graph_dump_file): Pass base to start_graph_dump.
29400 2013-02-14 Richard Henderson <rth@redhat.com>
29403 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
29405 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
29407 * collect2-aix.h: Define F_LOADONLY.
29409 2013-02-14 Richard Biener <rguenther@suse.de>
29412 * varasm.c (output_constant_def_1): Get the decl representing
29413 the constant as argument.
29414 (output_constant_def): Wrap output_constant_def_1.
29415 (make_decl_rtl): Use output_constant_def_1 with the decl
29416 representing the constant.
29417 (build_constant_desc): Optionally re-use a decl already
29418 representing the constant.
29419 (tree_output_constant_def): Adjust.
29421 2013-02-14 Dodji Seketeli <dodji@redhat.com>
29424 * asan.c (instrument_builtin_call): Really put the length of the
29425 second source argument into src1_len.
29427 2013-02-13 Jakub Jelinek <jakub@redhat.com>
29429 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
29430 argument. If it is false, don't create edge from then_bb to
29432 (insert_if_then_before_iter): Pass true to it.
29433 (build_check_stmt): Pass false to it.
29434 (transform_statements): Flush hash table only on extended basic
29435 block boundaries, rather than at the beginning of every bb.
29436 Don't flush hash table on nonfreeing_call_p calls.
29437 * tree-flow.h (nonfreeing_call_p): New prototype.
29438 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
29440 2013-02-13 David S. Miller <davem@davemloft.net>
29442 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
29444 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
29447 * ira.c (max_regno_before_ira): Move from ...
29449 (fix_reg_equiv_init): Use max_regno_before_ira instead of
29452 2013-02-13 Jakub Jelinek <jakub@redhat.com>
29454 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
29456 2013-02-13 Richard Biener <rguenther@suse.de>
29459 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
29460 globals in MEM_REFs.
29462 2013-02-13 Richard Biener <rguenther@suse.de>
29464 * loop-init.c (loop_optimizer_init): Clear loop state when
29465 re-initializing preserved loops.
29466 * loop-unswitch.c (unswitch_single_loop): Return whether
29467 we unswitched the loop. Do not verify loop state here.
29468 (unswitch_loops): When we unswitched a loop discover new loops.
29470 2013-02-13 Kostya Serebryany <kcc@google.com>
29472 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
29474 * sanitizer.def: Rename __asan_init to __asan_init_v1.
29476 2013-02-12 Dodji Seketeli <dodji@redhat.com>
29478 Avoid instrumenting duplicated memory access in the same basic block
29479 * Makefile.in (asan.o): Add new dependency on hash-table.h
29480 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
29481 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
29482 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
29483 (free_mem_ref_resources, has_mem_ref_been_instrumented)
29484 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
29485 (get_mem_ref_of_assignment): New functions.
29486 (get_mem_refs_of_builtin_call): Extract from
29487 instrument_builtin_call and tweak a little bit to make it fit with
29489 (instrument_builtin_call): Use the new
29490 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
29491 of is_gimple_builtin_call.
29492 (instrument_derefs, instrument_mem_region_access): Insert the
29493 instrumented memory reference into the hash table.
29494 (maybe_instrument_assignment): Renamed instrument_assignment into
29495 this, and change it to advance the iterator when instrumentation
29496 actually happened and return true in that case. This makes it
29497 homogeneous with maybe_instrument_assignment, and thus give a
29498 chance to callers to be more 'regular'.
29499 (transform_statements): Clear the memory reference hash table
29500 whenever we enter a new BB, when we cross a function call, or when
29501 we are done transforming statements. Use
29502 maybe_instrument_assignment instead of instrumentation. No more
29503 need to special case maybe_instrument_assignment and advance the
29504 iterator after calling it; it's now handled just like
29505 maybe_instrument_call. Update comment.
29507 2013-02-13 Richard Biener <rguenther@suse.de>
29509 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
29510 Fix loop discovery code.
29512 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
29514 PR inline-asm/56148
29515 * lra-constraints.c (process_alt_operands): Match early clobber
29516 operand with itself. Check conflicts with earlyclobber only if
29517 the operand is not reloaded. Prefer to reload conflicting operand
29518 if earlyclobber and matching operands are the same.
29520 2013-02-12 Richard Biener <rguenther@suse.de>
29523 * lto-streamer-out.c (write_symbol): Do not output symbols
29524 for hard register variables.
29526 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
29529 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
29530 (umulsidi3_insn, mulsidi3_insn): New insns.
29532 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
29534 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
29535 (struct tune_params): Add vec_costs field.
29536 * config/arm/arm.c (arm_builtin_vectorization_cost)
29537 (arm_add_stmt_cost): New functions.
29538 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
29539 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
29540 (arm_default_vec_cost): New struct of type cpu_vec_costs.
29541 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
29542 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
29543 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
29544 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
29546 2013-02-12 Richard Biener <rguenther@suse.de>
29549 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
29550 decls again if possible.
29552 2013-02-12 Richard Biener <rguenther@suse.de>
29554 PR middle-end/56288
29555 * tree-ssa.c (verify_ssa_name): Fix check, move
29556 SSA_NAME_IN_FREE_LIST check up.
29558 2013-02-12 Jakub Jelinek <jakub@redhat.com>
29559 Steven Bosscher <steven@gcc.gnu.org>
29561 PR rtl-optimization/56151
29562 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
29563 equal to op0 or op1, and last_insn pattern is CODE operation
29564 with MEM dest and one of the operands matches that MEM.
29566 2013-02-11 Sriraman Tallam <tmsriram@google.com>
29568 * doc/extend.texi: Document Function Multiversioning and "default"
29569 parameter string to target attribute.
29570 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
29571 target attribute parameter is "default".
29572 (ix86_compare_version_priority): Remove checks for target attribute.
29573 (ix86_mangle_function_version_assembler_name): Change error to sorry.
29574 Remove check for target attribute equal to NULL. Add assert.
29575 (ix86_generate_version_dispatcher_body): Change error to sorry.
29577 2013-02-11 Iain Sandoe <iain@codesourcery.com>
29578 Jack Howarth <howarth@bromo.med.uc.edu>
29579 Patrick Marlier <patrick.marlier@gmail.com>
29582 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
29583 define ENDFILE_SPEC as TM_DESTRUCTOR.
29584 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
29586 2013-02-11 Alexander Potapenko <glider@google.com>
29587 Jack Howarth <howarth@bromo.med.uc.edu>
29588 Jakub Jelinek <jakub@redhat.com>
29591 * config/darwin.c (cdtor_record): Rename ctor_record.
29592 (sort_cdtor_records): Rename sort_ctor_records.
29593 (finalize_dtors): New routine to sort destructors by
29594 priority before use in assemble_integer.
29595 (machopic_asm_out_destructor): Use finalize_dtors if needed.
29597 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
29599 PR rtl-optimization/56275
29600 * simplify-rtx.c (avoid_constant_pool_reference): Check that
29601 offset is non-negative and less than cmode size before
29602 calling simplify_subreg.
29604 2013-02-11 Richard Biener <rguenther@suse.de>
29606 PR tree-optimization/56264
29607 * cfgloop.h (fix_loop_structure): Adjust prototype.
29608 * loop-init.c (fix_loop_structure): Return the number of
29609 newly discovered loops.
29610 * tree-cfgcleanup.c (repair_loop_structures): When new loops
29611 are discovered, do a full loop-closed SSA rewrite.
29613 2013-02-11 Richard Biener <rguenther@suse.de>
29615 PR tree-optimization/56273
29616 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
29618 (check_array_ref): Fix missing newline in dumps.
29619 (search_for_addr_array): Likewise.
29621 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
29623 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
29625 2013-02-09 Jakub Jelinek <jakub@redhat.com>
29628 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
29630 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
29632 PR rtl-optimization/56246
29633 * lra-constraints.c (simplify_operand_subreg): Try to reuse
29635 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
29636 constraints are satisfied.
29638 2013-02-08 Jeff Law <law@redhat.com>
29641 * emit-rtl.c (reg_is_parm_p): New function.
29642 * regs.h (reg_is_parm_p): New prototype.
29643 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
29644 callee-clobbered registers.
29646 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
29649 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
29650 If there is no implicit builtin declaration, just return NULL.
29652 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
29654 * config/i386/sse.md (FMAMODEM): New mode iterator.
29655 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
29656 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
29658 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
29660 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
29661 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
29662 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
29664 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
29666 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
29667 (microblaze*-*-elf): Likewise.
29668 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
29670 * config/microblaze/microblaze-c.c: Add builtin defines for
29671 _LITTLE_ENDIAN and _BIG_ENDIAN.
29672 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
29673 add to TARGET_DEFAULT flags.
29674 Expand ASM_SPEC and LINK_SPEC.
29675 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
29676 * config/microblaze/microblaze.md: Update extendsidi2 and
29677 movdi_internal instructions to use low-order / high-order reg
29679 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
29680 options and inversemask / mask of LITTLE_ENDIAN.
29681 * config/microblaze/t-microblaze: Expand multilib options to
29682 include mlittle-endian (le) and update exceptions patterns.
29684 2013-02-08 Jakub Jelinek <jakub@redhat.com>
29686 PR rtl-optimization/56195
29687 * lra-constraints.c (get_reload_reg): Don't reuse regs
29688 if they have smaller mode than requested, if they have
29689 wider mode than requested, try to return a SUBREG.
29691 PR tree-optimization/56250
29692 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
29693 if type is unsigned and code isn't MULT_EXPR.
29695 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
29697 PR tree-optimization/56064
29698 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
29699 bits according to mode.
29700 * fixed-value.h (fixed_from_double_int)
29701 (const_fixed_from_double_int): Adjust comments.
29703 2013-02-08 Richard Biener <rguenther@suse.de>
29706 * lto-streamer.h (struct data_in): Remove current_file, current_line
29707 and current_col members.
29708 * lto-streamer-out.c (lto_output_location): Stream changed bits
29709 en-block for efficiency.
29710 * lto-streamer-in.c (clear_line_info): Remove.
29711 (lto_input_location): Cache current file, line and column
29712 globally via local statics. Read changed bits en-block.
29713 (input_function): Do not call clear_line_info.
29714 (lto_read_body): Likewise.
29715 (lto_input_toplevel_asms): Likewise.
29717 2013-02-08 Michael Matz <matz@suse.de>
29719 PR tree-optimization/52448
29720 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
29721 (nt_call_phase): New static.
29722 (add_or_mark_expr): Only mark accesses with newer phase than any
29724 (nonfreeing_call_p): New.
29725 (nt_init_block): Update nt_call_phase, mark blocks as visited.
29726 (nt_fini_block): Keep blocks marked as visited.
29727 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
29729 2013-02-08 Richard Biener <rguenther@suse.de>
29731 * ira.c (ira): Free broken dominator information.
29733 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
29735 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
29737 2013-02-08 Marek Polacek <polacek@redhat.com>
29739 * cfgloop.c (verify_loop_structure): Add more checking of headers.
29741 2013-02-08 Richard Biener <rguenther@suse.de>
29743 PR middle-end/56181
29744 * cfgloop.h (flow_loops_find): Adjust.
29745 (bb_loop_header_p): Declare.
29746 * cfgloop.c (bb_loop_header_p): New function split out from ...
29747 (flow_loops_find): ... here. Adjust function signature,
29748 support incremental loop structure update.
29749 (verify_loop_structure): Cleanup. Verify a loop is a loop.
29750 * cfgloopmanip.c (fix_loop_structure): Move ...
29751 * loop-init.c (fix_loop_structure): ... here.
29752 (apply_loop_flags): Split out from ...
29753 (loop_optimizer_init): ... here.
29754 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
29755 in incremental mode, only remove dead loops here.
29757 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
29760 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
29761 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
29762 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
29763 (*round<mode>3.libgcc): New insns for fixed-modes.
29764 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
29765 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
29766 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
29767 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
29768 implementations. Define to __builtin_avr_absFX,
29769 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
29770 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
29771 __builtin_avr_countlsFX, respectively.
29772 * config/avr/avr-c.c (target.h): Include it.
29773 (enum avr_builtin_id): New enum.
29774 (avr_resolve_overloaded_builtin): New static function.
29775 (avr_register_target_pragmas): Use it to set
29776 targetm.resolve_overloaded_builtin.
29777 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
29778 tree nodes used by DEF_BUILTIN.
29779 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
29780 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
29781 <AVR_BUILTIN_xxBITS>: Same.
29783 2013-02-08 Richard Biener <rguenther@suse.de>
29785 * cfgloop.c (verify_loop_structure): Properly handle
29786 a loop exiting to another loop header.
29787 * ira-int.h (ira_loops): Remove.
29788 * ira.c (ira_loops): Remove.
29789 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
29790 (do_reload): Use loop_optimizer_finalize.
29791 * ira-build.c (create_loop_tree_nodes): Use get_loops and
29792 number_of_loops to access the loop tree.
29793 (more_one_region_p): Likewise.
29794 (finish_loop_tree_nodes): Likewise.
29795 (rebuild_regno_allocno_maps): Likewise.
29796 (mark_loops_for_removal): Likewise.
29797 (mark_all_loops_for_removal): Likewise.
29798 (remove_unnecessary_regions): Likewise.
29799 (ira_build): Likewise.
29800 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
29802 2013-02-08 Richard Biener <rguenther@suse.de>
29804 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
29805 * ipa-pure-const.c (analyze_function): Avoid calling
29806 mark_irreducible_loops twice.
29807 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
29809 2013-02-07 David S. Miller <davem@davemloft.net>
29811 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
29813 * var-tracking.c (vt_add_function_parameter): Test the presence of
29814 HAVE_window_save properly and do not remap argument registers when
29815 we have a leaf function.
29817 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
29820 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
29822 * config/i386/i386.c (ix86_print_operand): Ditto.
29824 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
29826 * lra-constraints.c (process_alt_operands): Fix recently added comment.
29828 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
29830 PR rtl-optimization/56225
29831 * lra-constraints.c (process_alt_operands): Check that reload hard
29832 reg can hold value for strict_low_part.
29834 2013-02-07 Jakub Jelinek <jakub@redhat.com>
29837 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
29838 dwarf2out_end_function.
29839 (in_first_function_p, maybe_at_text_label_p,
29840 first_loclabel_num_not_at_text_label): New variables.
29841 (dwarf2out_var_location): In the first function find out
29842 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
29843 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
29846 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
29848 PR rtl-optimization/56178
29849 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
29850 SUBREG of a register. Tidy up related block of code.
29851 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
29852 note if the source is a register or a SUBREG of a register.
29854 2013-02-07 Jakub Jelinek <jakub@redhat.com>
29857 * config/rs6000/rs6000.md (ptrm): New mode attr.
29858 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
29859 call_value_indirect_aix<pttrsize>,
29860 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
29863 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
29865 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
29866 if -bnortl. Convert to strcmp and strncmp.
29868 2013-02-07 Alan Modra <amodra@gmail.com>
29871 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
29872 addresses won't wrap when offsetting.
29873 (rs6000_secondary_reload): Provide secondary reloads needed for
29874 wrapping LO_SUM addresses.
29876 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
29878 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
29879 MACH, just __MACH__.
29881 2013-02-06 Richard Biener <rguenther@suse.de>
29883 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
29884 instead of calling fix_loop_structure.
29886 2013-02-06 Jakub Jelinek <jakub@redhat.com>
29888 PR middle-end/56217
29889 * omp-low.c (use_pointer_for_field): Return false if
29890 lower_send_shared_vars doesn't generate any copy-out code.
29892 2013-02-06 Tom de Vries <tom@codesourcery.com>
29894 PR rtl-optimization/56131
29895 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
29896 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
29897 of the label is NULL. Add comment.
29899 2013-02-05 Jakub Jelinek <jakub@redhat.com>
29901 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
29904 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
29905 (STATIC_LIBTSAN_LIBS): Likewise.
29906 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
29907 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
29908 is defined, don't add anything else beyond that.
29909 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
29910 (LINK_COMMAND_SPEC): Use them.
29912 PR tree-optimization/56205
29913 * tree-stdarg.c (check_all_va_list_escapes): Return true if
29914 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
29915 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
29917 2013-02-05 Richard Biener <rguenther@suse.de>
29919 PR tree-optimization/53342
29920 PR tree-optimization/53185
29921 * tree-vectorizer.h (vect_check_strided_load): Remove.
29922 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
29923 not disallow peeling for vectorized strided loads.
29924 (vect_check_strided_load): Make static and simplify.
29925 (vect_analyze_data_refs): Adjust.
29926 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
29927 correctly when vectorizing strided loads.
29929 2013-02-05 Richard Biener <rguenther@suse.de>
29931 * doc/install.texi: Refer to ISL, not PPL.
29933 2013-02-05 Jan Hubicka <jh@suse.cz>
29935 PR tree-optimization/55789
29936 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
29938 2013-02-05 Jan Hubicka <jh@suse.cz>
29940 PR tree-optimization/55789
29941 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
29942 the dead call anyway.
29944 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
29947 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
29949 2013-02-04 Alexander Potapenko <glider@google.com>
29950 Jack Howarth <howarth@bromo.med.uc.edu>
29951 Jakub Jelinek <jakub@redhat.com>
29954 * config/darwin.c (sort_ctor_records): Stabilized qsort
29955 on constructor priority by using original position.
29956 (finalize_ctors): New routine to sort constructors by
29957 priority before use in assemble_integer.
29958 (machopic_asm_out_constructor): Use finalize_ctors if needed.
29960 2013-02-04 Jakub Jelinek <jakub@redhat.com>
29963 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
29964 about visibility on artificial decls.
29965 * config/sol2.c (solaris_assemble_visibility): Likewise.
29967 2013-02-04 Kai Tietz <ktietz@redhat.com>
29970 * config/i386/i386.c (function_value_ms_64): Add additional valtype
29971 argument and improve checking of return-argument types for 16-byte
29973 (ix86_function_value_1): Add additional valtype argument on call
29974 of function_value_64.
29975 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
29976 handling infunction_value_64 function.
29978 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
29980 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
29982 2013-02-04 Richard Biener <rguenther@suse.de>
29984 PR tree-optimization/56188
29985 * tree-ssa-structalias.c (label_visit): Consider case with
29986 initially non-empty points-to set.
29987 (perform_var_substitution): Dump node mapping and clean up.
29989 2013-02-04 Richard Guenther <rguenther@suse.de>
29992 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
29993 node prevail as last resort.
29994 (lto_symtab_merge_decls): Remove guard on LTRANS here.
29995 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
29997 2013-02-04 Richard Biener <rguenther@suse.de>
29999 PR tree-optimization/56113
30000 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
30002 (equiv_class_lookup_or_add): ... this.
30003 (label_visit): Adjust and fix error in previous patch.
30004 (perform_var_substitution): Adjust.
30006 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
30008 * config/sh/divtab.c: Fix formatting and comments throughout the file.
30009 * config/sh/sh4-300.md: Likewise.
30010 * config/sh/sh4a.md: Likewise.
30011 * config/sh/constraints.md: Likewise.
30012 * config/sh/sh.md: Likewise.
30013 * config/sh/netbsd-elf.h: Likewise.
30014 * config/sh/predicates.md: Likewise.
30015 * config/sh/sh-protos.h: Likewise.
30016 * config/sh/ushmedia.h: Likewise.
30017 * config/sh/linux.h: Likewise.
30018 * config/sh/sh.c: Likewise.
30019 * config/sh/superh.h: Likewise.
30020 * config/sh/elf.h: Likewise.
30021 * config/sh/sh4.md: Likewise.
30022 * config/sh/sh.h: Likewise.
30024 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
30026 * config/pa/constraints.md: Adjust unused letters. Change "T"
30027 constraint to match_test floating_point_store_memory_operand().
30028 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
30029 (base14_operand): New.
30030 (floating_point_store_memory_operand): New.
30031 (integer_store_memory_operand): Revise to use base14_operand and
30032 reg_plus_base_memory_operand.
30033 (move_dest_operand): Allow symbolic_memory_operands.
30034 (symbolic_memory_operand): Check for LO_SOM.
30035 (symbolic_operand): Change default case to break.
30036 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
30037 CONST_DOUBLE values to be reloaded by putting them into memory when
30038 the destination is a floating point register.
30039 (movdf): Remove code to handle CONST_DOUBLE.
30041 (reload_indf_r1): New.
30042 (reload_insf_r1): New.
30043 Consistently use "Q" and "T" constraints with integer and floating
30044 point move instructions, respectively.
30045 (movdi): Remove FAIL.
30046 Change predicate for source operand unamed DImode move from
30047 general_operand to move_src_operand.
30048 (umulsidi3): Change predicate for destination operand to
30050 Likewise for similar unamed patterns.
30051 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
30052 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
30053 (hppa_legitimize_address): Simplify mask calculation.
30054 (pa_emit_move_sequence): Revised handling of secondary reloads from
30055 REG+D addresses for floating point loads and stores. Directly handle
30056 loading CONST0_RTX (mode) to a floating point register.
30057 (pa_secondary_reload): Handle reloading DF and SFmode constant values
30058 to floating point registers. Don't restrict secondary reloads to
30059 floating point registers to integer modes. Revise some comments and
30061 (TARGET_LEGITIMATE_ADDRESS_P): Define.
30062 (pa_legitimate_address_p): New.
30063 (pa_legitimize_reload_address): New.
30064 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
30065 (STRICT_REG_OK_FOR_BASE_P): New.
30066 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
30067 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
30069 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
30070 Andrew Dixie <andrewd@gentrack.com>
30072 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
30075 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
30077 * expmed.c (extract_bit_field_1): Pass the full width of the
30078 structure to get_best_reg_extraction_insn.
30080 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
30083 * configure.ac (use_cxa_atexit): Add AIX.
30084 * configure: Regenerate.
30086 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
30088 2013-02-01 Jakub Jelinek <jakub@redhat.com>
30091 * final.c (need_profile_function): New variable.
30092 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
30093 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
30094 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
30095 notes, targetm.asm_out.function_prologue doesn't emit anything,
30096 HAVE_prologue and profiler should be emitted before prologue,
30097 set need_profile_function instead of emitting it.
30098 (final_scan_insn): If need_profile_function, emit
30099 profile_function on the first NOTE_INSN_BASIC_BLOCK or
30100 NOTE_INSN_FUNCTION_BEG note.
30102 2013-02-01 Richard Henderson <rth@redhat.com>
30104 * config/rs6000/rs6000.md (smulditi3): New.
30107 * config/alpha/alpha.md (umulditi3): New.
30109 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
30111 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
30112 (ASM_OUTPUT_ALIGNED_LOCAL): New.
30114 2013-02-01 Richard Biener <rguenther@suse.de>
30116 PR tree-optimization/56113
30117 * tree-ssa-structalias.c (label_visit): Reduce work for
30118 single-predecessor nodes.
30120 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
30122 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
30123 range isn't testing for zero.
30125 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
30127 PR middle-end/56113
30128 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
30130 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
30131 Nick Clifton <nickc@redhat.com>
30133 * config/v850/constraints.md (Q): Define as a memory constraint.
30134 * config/v850/predicates.md (label_ref_operand): New predicate.
30135 (e3v5_shift_operand): New predicate.
30136 (ior_operator): New predicate.
30137 * config/v850/t-v850: Add e3v5 multilib.
30138 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
30139 (v850_gen_movdi): Prototype.
30140 * config/v850/v850.c: Add support for e3v5 architecture.
30141 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
30143 (construct_save_jarl): Add e3v5 long JARL support.
30144 (v850_adjust_insn_length): New function. Adjust length of call
30145 insns when using e3v5 instructions.
30146 (v850_gen_movdi): New function: Generate instructions to move a
30148 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
30149 (CPP_SPEC): Define __v850e3v5__ as appropriate.
30150 (TARGET_USE_FPU): Enable for e3v5.
30151 (CONST_OK_FOR_W): New macro.
30152 (ADJUST_INSN_LENGTH): Define.
30153 * config/v850/v850.md (UNSPEC_LOOP): Define.
30154 (attr cpu): Add v850e3v5.
30155 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
30156 (movdi): New pattern.
30157 (movdi_internal): New pattern.
30158 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
30159 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
30160 (cstoresf4): Likewise.
30161 (cstoredf4): Likewise.
30162 (insv): New pattern.
30163 (rotlso3_a): New pattern.
30164 (rotlsi3_b): New pattern
30165 (rotlsi3_v850e3v5): New pattern.
30166 (doloop_begin): New pattern.
30167 (fix_loop_counter): New pattern.
30168 (doloop_end): New pattern.
30169 (branch_normal): Add e3v5 long branch support.
30170 (branch_invert): Likewise.
30171 (branch_z_normal): Likewise.
30172 (branch_z_invert): Likewise.
30173 (branch_nz_normal): Likewise.
30174 (branch_nz_invert): Likewise.
30175 (call_internal_short): Add e3v5 register-indirect JARL support.
30176 (call_internal_long): Likewise.
30177 (call_value_internal_short): Likewise.
30178 (call_value_internal_long): Likewise.
30179 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
30180 (mloop): New option.
30181 * config.gcc: Add support for configuring v840e3v5 target.
30182 * doc/invoke.texi: Document new v850 specific command line options.
30184 2013-01-31 Paul Koning <ni1d@arrl.net>
30188 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
30189 children if parent is a class.
30190 (prune_unused_types_prune): Don't add DW_AT_declaration.
30192 2013-01-31 Richard Biener <rguenther@suse.de>
30194 PR tree-optimization/56157
30195 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
30196 match up operand with SLP child.
30198 2013-01-31 Jason Merrill <jason@redhat.com>
30201 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
30202 parameters the first time.
30203 (gen_scheduled_generic_parms_dies): Check completeness here.
30205 2013-01-31 Richard Biener <rguenther@suse.de>
30207 PR middle-end/53073
30208 * common.opt (faggressive-loop-optimizations): New flag,
30209 enabled by default.
30210 * doc/invoke.texi (faggressive-loop-optimizations): Document.
30211 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
30212 infer_loop_bounds_from_undefined by it.
30214 2013-01-31 Richard Biener <rguenther@suse.de>
30216 PR tree-optimization/56150
30217 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
30218 visit virtual operands.
30219 (find_uses_to_rename_bb): Likewise.
30221 2013-01-31 Richard Biener <rguenther@suse.de>
30223 PR tree-optimization/56150
30224 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
30225 mixed store non-store stmts.
30227 2013-01-30 Jakub Jelinek <jakub@redhat.com>
30230 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
30231 LIBASAN_EARLY_SPEC is defined.
30232 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
30233 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
30235 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
30238 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
30239 invalid args instead of ICEing on it.
30240 (ix86_valid_target_attribute_tree): Return error_mark_node if
30241 ix86_valid_target_attribute_inner_p failed.
30242 (ix86_valid_target_attribute_p): Return false only if
30243 ix86_valid_target_attribute_tree returned error_mark_node. Allow
30244 target("default") attribute.
30245 (sorted_attr_string): Change argument from const char * to tree,
30246 merge in all target attribute arguments rather than just one.
30247 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
30248 instead of free. Avoid using strcat.
30249 (ix86_mangle_function_version_assembler_name): Mangle
30250 target("default") as if no target attribute is present. Adjust
30251 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
30252 instead of xmalloc and XDELETEVEC instead of free.
30253 (ix86_function_versions): Don't return true if one of the decls
30254 doesn't have target attribute. If they don't and one of the decls
30255 is DECL_FUNCTION_VERSIONED, report an error. Adjust
30256 sorted_attr_string caller. Use XDELETEVEC instead of free.
30257 (ix86_supports_function_versions): Remove.
30258 (make_name): Fix up formatting.
30259 (make_dispatcher_decl): Remove resolver_name and its initialization.
30260 Avoid leaking memory.
30261 (is_function_default_version): Return true if there is
30262 target("default") attribute rather than no target attribute at all.
30263 (make_resolver_func): Avoid leaking memory.
30264 (ix86_generate_version_dispatcher_body): Likewise.
30265 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
30266 * target.def (supports_function_versions): Remove.
30267 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
30268 * doc/tm.texi: Regenerated.
30270 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
30272 PR rtl-optimization/56144
30273 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
30274 for values with side effects.
30276 2013-01-30 Richard Biener <rguenther@suse.de>
30278 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
30279 (sparseset_pop): Likewise.
30280 * cfganal.c (compute_idf): Likewise. Increase work-stack size
30281 to be able to use quick_push in the worker loop.
30283 2013-01-30 Marek Polacek <polacek@redhat.com>
30285 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
30287 2013-01-30 Richard Biener <rguenther@suse.de>
30290 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
30292 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
30294 PR tree-optimization/56064
30295 * fixed-value.c (fixed_from_double_int): New function.
30296 * fixed-value.h (fixed_from_double_int): New prototype.
30297 (const_fixed_from_double_int): New static inline function.
30298 * fold-const.c (native_interpret_fixed): New static function.
30299 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
30300 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
30301 (native_encode_fixed): New static function.
30302 (native_encode_expr) <FIXED_CST>: Use it.
30303 (native_interpret_int): Move double_int worker code to...
30304 * double-int.c (double_int::from_buffer): ...this new static method.
30305 * double-int.h (double_int::from_buffer): Prototype it.
30307 2013-01-30 Richard Biener <rguenther@suse.de>
30309 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
30310 New pointer-map and obstack.
30311 (init_alias_vars): Allocate pointer-map and obstack.
30312 (delete_points_to_sets): Free them.
30313 (find_what_var_points_to): Cache result.
30314 (find_what_p_points_to): Adjust for changed interface of
30315 find_what_var_points_to.
30316 (compute_points_to_sets): Likewise.
30317 (ipa_pta_execute): Likewise.
30319 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
30321 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
30322 * configure: Regenerate.
30323 * config.in: Regenerate.
30324 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
30325 #nobits/#progbits if supported.
30327 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
30330 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
30331 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
30332 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
30334 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
30336 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
30337 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
30339 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
30341 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
30342 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
30344 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
30346 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
30348 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
30349 * config/arm/cortex-a7.md: New bypasses using
30350 arm_mac_accumulator_is_result.
30352 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
30354 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
30355 (cortex_a7_neon_mla): Likewise.
30356 (cortex_a7_fpfmad): New reservation.
30357 (cortex_a7_fpmacs): Use ffmas and update required units.
30358 (cortex_a7_fpmuld): Update required units and latency.
30359 (cortex_a7_fpmacd): Likewise.
30360 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
30361 (cortex_a7_neon). Likewise.
30362 (bypass) Update participating units.
30364 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
30366 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
30367 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
30369 * config/arm/vfp11.md (vfp_farith): Use ffmas.
30370 (vfp_fmul): Use ffmad.
30371 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
30372 (cortex_r4_fmacd): Use ffmad.
30373 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
30374 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
30375 (cortex_a9_fmacd): Use ffmad.
30376 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
30377 (cortex_a8_vfp_macd): Use ffmad.
30378 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
30379 (cortex_a5_fpmacd): Use ffmad.
30380 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
30381 (cortex_a15_vfp_macd): Use ffmad.
30382 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
30384 2013-01-29 Jason Merrill <jason@redhat.com>
30387 * varasm.c (default_assemble_visibility): Don't warn about
30388 visibility on artificial decls.
30390 2013-01-29 Richard Biener <rguenther@suse.de>
30392 PR tree-optimization/56113
30393 * tree-ssa-structalias.c (equiv_class_lookup): Also return
30395 (label_visit): Free duplicate bitmaps and record the leader instead.
30396 (perform_var_substitution): Adjust.
30398 2013-01-29 Richard Biener <rguenther@suse.de>
30400 PR tree-optimization/55270
30401 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
30402 the CFG, schedule loops for fixup.
30404 2013-01-29 Nick Clifton <nickc@redhat.com>
30406 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
30409 2013-01-28 Leif Ekblad <leif@rdos.net>
30411 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
30412 * config/i386/i386.h (TARGET_RDOS): New macro.
30413 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
30414 * config/i386/i386.c (ix86_option_override_internal): For 64bit
30415 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
30416 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
30417 DEFAULT_LARGE_SECTION_THRESHOLD.
30418 * config/i386/i386.md (R14_REG, R15_REG): New constants.
30419 * config/i386/rdos.h: New file.
30420 * config/i386/rdos64.h: New file.
30422 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
30425 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
30428 2013-01-28 Jakub Jelinek <jakub@redhat.com>
30430 PR rtl-optimization/56117
30431 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
30432 call cselib_lookup_from_insn on the MEM before calling
30433 add_insn_mem_dependence.
30435 2013-01-28 Richard Biener <rguenther@suse.de>
30437 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
30438 to a stmt that didn't have one.
30439 (copy_phis_for_bb): Likewise for PHI arguments.
30440 (copy_debug_stmt): Likewise for debug stmts.
30442 2013-01-28 Richard Biener <rguenther@suse.de>
30444 PR tree-optimization/56034
30445 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
30446 (partition_builtin_p): Adjust.
30447 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
30448 it is the last partition.
30449 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
30450 up the vertex for the definition.
30451 (classify_partition): Classify whether a partition is a
30452 PKIND_REDUCTION, thus has uses outside of the loop.
30453 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
30454 Merge all PKIND_REDUCTION partitions into the last partition.
30455 (tree_loop_distribution): Seed partitions from reductions as well.
30457 2013-01-28 Jakub Jelinek <jakub@redhat.com>
30459 PR tree-optimization/56125
30460 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
30461 pow(x,c) into sqrt(x) * powi(x, n/2) or
30462 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
30463 optimizing for size.
30464 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
30465 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
30468 PR tree-optimization/56094
30469 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
30470 to UNKNOWN_LOCATION while gimplifying expr.
30472 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
30475 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
30476 operand 0 in movabs insn template for -masm=intel asm alternative.
30477 (*movabs<mode>_2): Ditto for operand 1.
30479 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
30482 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
30485 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
30487 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
30490 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
30492 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
30493 Undef to avoid warning.
30495 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
30497 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
30498 * configure: Regenerate.
30500 2013-01-25 Jakub Jelinek <jakub@redhat.com>
30502 PR tree-optimization/56098
30503 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
30504 for stmts with volatile ops.
30505 (cond_store_replacement): Don't optimize if assign has volatile ops.
30506 (cond_if_else_store_replacement_1): Don't optimize if either
30507 then_assign or else_assign have volatile ops.
30508 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
30511 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
30513 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
30515 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
30517 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
30518 missing ':' in asm example.
30520 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
30522 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
30523 entries into lane and laneq entries.
30524 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
30525 Remove AdvSIMD scalar modes.
30526 (aarch64_sq<r>dmulh_laneq<mode>): New.
30527 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
30529 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
30530 builtin implementations to relfect changes in RTL in aarch64-simd.md.
30531 * config/aarch64/iterators.md (VCOND): New.
30534 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
30537 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
30538 Add NULL LIBNAME argument to existing definitions.
30539 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
30540 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
30541 * config/avr/avr.c (DEF_BUILTIN): Same.
30542 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
30543 (avr_expand_builtin): Expand to a vanilla call if a libgcc
30544 implementation is available (DECL_ASSEMBLER_NAME is set).
30545 (avr_fold_absfx): New static function.
30546 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
30547 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
30548 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
30549 AVR_BUILTIN_ABSLLK.
30550 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
30551 (abshk, absk, abslk, absllk): Provide as static inline functions.
30553 2013-01-25 Marek Polacek <polacek@redhat.com>
30555 PR tree-optimization/56035
30556 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
30558 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
30560 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
30561 (*movtf_internal_rex64): Add (!o,C) alternative
30562 (*movxf_internal_rex64): Ditto.
30563 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
30565 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
30567 * doc/invoke.texi: fix typo.
30568 * doc/objc.texi: fix typo.
30570 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
30572 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
30573 for the first two alternatives.
30575 2013-01-24 Diego Novillo <dnovillo@google.com>
30577 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
30578 (ggc-zone.o): Remove.
30579 * configure.ac: Remove option --with-gc.
30580 * configure: Re-generate.
30581 * doc/install.texi: Remove documentation for --with-gc.
30582 * gengtype.c (write_enum_defn): Remove. Update all users.
30583 (write_Types_process_field): Remove generation of gt_e_* argument.
30584 (output_type_enum): Remove. Update all users.
30585 (write_enum_defn): Remove. Update all users.
30586 (enum alloc_zone): Remove. Update all users.
30587 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
30588 * ggc-common.c (ggc_splay_alloc): Remove first argument.
30589 Update all callers.
30590 (struct ptr_data): Remove field TYPE. Update all users.
30591 (gt_pch_note_object): Remove argument TYPE. Update all users.
30592 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
30594 * ggc-none.c (ggc_alloc_typed_stat): Remove.
30595 (struct alloc_zone): Remove.
30596 (ggc_internal_alloc_zone_stat): Remove.
30597 (ggc_internal_cleared_alloc_zone_stat): Remove.
30598 * ggc-page.c (ggc_alloc_typed_stat): Remove.
30599 (ggc_pch_count_object): Remove last argument. Update all users.
30600 (ggc_pch_alloc_object): Remove last argument. Update all users.
30601 (struct alloc_zone): Remove.
30602 * ggc-zone.c: Remove.
30603 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
30604 (struct alloc_zone): Remove.
30605 (ggc_alloc_typed_stat): Remove.
30606 (ggc_alloc_typed): Remove.
30607 (ggc_splay_alloc): Remove first argument.
30608 (rtl_zone): Remove. Update all users.
30609 (tree_zone): Remove. Update all users.
30610 (tree_id_zone): Remove. Update all users.
30611 (ggc_internal_zone_alloc_stat): Remove. Update all users.
30612 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
30613 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
30614 * tree-ssanames.c: Remove references to zone allocator in comments.
30616 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
30618 * config/avr/avr.c (avr_out_fract): Make register numbers that
30619 might be outside of source operand signed.
30621 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
30623 * config/i386/constraints.md (Yf): New constraint.
30624 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
30625 of f constraint to conditionaly disable x87 register preferences.
30626 (*movdf_internal): Ditto.
30627 (*movsf_internal): Ditto.
30629 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
30631 PR inline-asm/55934
30632 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
30633 that have operands with impossible constraints.
30634 Add a FIXME for a speed-up opportunity.
30635 * lra-constraints.c (process_alt_operands): Verify that a class
30636 selected from constraints on asms is valid for the operand mode.
30637 (curr_insn_transform): Remove incorrect comment.
30639 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
30641 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
30642 TOC operand is a valid symbol ref in the constant pool.
30644 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
30646 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
30648 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
30651 * config/avr/stdfix.h: New file.
30652 * t-avr (stdfix-gcc.h): New rule to build it.
30653 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
30655 2013-01-23 Kostya Serebryany <kcc@google.com>
30657 * config/darwin.h: remove dependency on
30658 CoreFoundation (asan on Mac OS).
30660 2013-01-23 Jakub Jelinek <jakub@redhat.com>
30663 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
30664 instead of cmpdi_operand for first comparison operand.
30665 Don't assert that comparison operands aren't both constants.
30667 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
30669 * doc/install.texi (Downloading the Source): Update references to
30670 downloading separate components.
30672 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
30674 * doc/extend.texi (__int128): Improve grammar.
30676 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
30679 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
30680 alternative to (o,r).
30681 (*movdi_internal_rex64): Remove (!o,n) alternative.
30682 (DImode immediate->memory splitter): Remove.
30683 (DImode immediate->memory peephole2): Remove.
30684 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
30685 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
30686 alternative to (!o,*r).
30687 (*movtf_internal_sse): New pattern.
30688 (*movxf_internal_rex64): New pattern.
30689 (*movxf_internal): Disable for TARGET_64BIT.
30690 (*movdf_internal_rex64): Remove (!o,F) alternative.
30692 2013-01-22 Jakub Jelinek <jakub@redhat.com>
30694 PR middle-end/56074
30695 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
30696 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
30697 * tree-vect-loop-manip.c (find_loop_location): Also ignore
30698 stmt locations where LOCATION_LOCUS of the stmt location is
30699 UNKNOWN_LOCATION or BUILTINS_LOCATION.
30702 * config/i386/i386.md (UNSPEC_STOS): New.
30703 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
30704 *strsetqi_1): Add UNSPEC_STOS.
30706 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
30709 * doc/invoke.texi: Remove left over -Wsynth example.
30711 2013-01-21 Jakub Jelinek <jakub@redhat.com>
30713 PR tree-optimization/56051
30714 * fold-const.c (fold_binary_loc): Don't fold
30715 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
30716 a narrowing conversion, or widening conversion from signed
30719 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
30721 PR rtl-optimization/56023
30722 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
30723 dependent on debug instruction.
30725 2013-01-21 Martin Jambor <mjambor@suse.cz>
30727 PR middle-end/56022
30728 * function.c (allocate_struct_function): Call
30729 invoke_set_current_function_hook earlier.
30731 2013-01-21 Jakub Jelinek <jakub@redhat.com>
30733 * reload1.c (init_reload): Only initialize reload_obstack
30734 during the first call.
30736 2013-01-21 Marek Polacek <polacek@redhat.com>
30738 * cfgloop.c (verify_loop_structure): Fix up grammar.
30740 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
30742 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
30743 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
30745 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
30748 * config/arm/marvell-pj4.md: Update copyright year.
30749 Fix up use of alu to alu_reg and simple_alu_imm.
30751 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
30753 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
30755 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
30758 * lra-constraints.c (curr_insn_transform): Don't reuse original
30759 insn for secondary memory move when memory mode should be different.
30761 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
30763 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
30764 atomic_storedi_1): New patterns.
30766 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
30768 btver2 pipeline descriptions.
30769 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
30771 * config/i386/i386.md (btver2_decode): New type attributes.
30772 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
30774 * config/i386/btver2.md: New file describing btver2 pipelines.
30776 2013-01-19 Andrew Pinski <apinski@cavium.com>
30778 PR tree-optimization/52631
30779 * tree-ssa-sccvn (visit_use): Before looking up the original
30780 statement, try looking up the simplified expression.
30782 2013-01-19 Anthony Green <green@moxielogic.com>
30784 * config/moxie/moxie.c (moxie_expand_prologue): Set
30785 current_function_static_stack_size.
30787 2013-01-18 Jakub Jelinek <jakub@redhat.com>
30789 PR tree-optimization/56029
30790 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
30791 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
30793 2013-01-18 Sharad Singhai <singhai@google.com>
30795 PR tree-optimization/55995
30796 * dumpfile.c (dump_loc): Print location only if available.
30797 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
30799 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
30802 * lra-constraints.c (curr_insn_transform): Reuse original insn for
30803 secondary memory move.
30804 (inherit_reload_reg): Use rclass instead of cl for
30805 check_secondary_memory_needed_p.
30807 2013-01-18 Jakub Jelinek <jakub@redhat.com>
30809 PR middle-end/56015
30810 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
30811 the case where writing real complex part of target modifies op1.
30813 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
30815 * config/aarch64/aarch64-simd.md
30816 (aarch64_vcond_internal<mode>): Handle unordered cases.
30817 * config/aarch64/iterators.md (v_cmp_result): New.
30819 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
30820 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
30822 * config/arm/marvell-pj4.md: New file.
30823 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
30824 * config/arm/arm.md (generic_sched): Add marvell_pj4.
30825 (generic_vfp): Likewise.
30826 * config/arm/arm-cores.def: Add marvell-pj4.
30827 * config/arm/arm-tune.md: Regenerate.
30828 * config/arm/arm-tables.opt: Regenerate.
30829 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
30830 * doc/invoke.texi: Document marvell-pj4.
30832 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
30834 * config/aarch64/arm_neon.h: Map scalar types to standard types.
30836 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
30841 * var-tracking.c (negative_power_of_two_p): New.
30842 (global_get_addr_cache, local_get_addr_cache): New.
30843 (get_addr_from_global_cache, get_addr_from_local_cache): New.
30844 (vt_canonicalize_addr): Rewrite using the above. Adjust the
30846 (vt_stack_offset_p): Remove.
30847 (vt_canon_true_dep): Always canonicalize loc's address.
30848 (clobber_overlapping_mems): Make sure we have a MEM.
30849 (local_get_addr_clear_given_value): New.
30850 (val_reset): Clear local cached entries.
30851 (compute_bb_dataflow): Create and release the local cache.
30852 Disable duplicate MEMs clobbering.
30853 (emit_notes_in_bb): Clobber MEMs likewise.
30854 (vt_emit_notes): Create and release the local cache.
30855 (vt_initialize, vt_finalize): Create and release the global
30856 cache, respectively.
30857 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
30859 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
30861 PR libmudflap/53359
30862 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
30863 not found in the symtab.
30865 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
30868 PR rtl-optimization/55547
30869 PR rtl-optimization/53827
30872 * alias.c (offset_overlap_p): New, factored out of...
30873 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
30874 the conservative special case for symbolic constants. Don't
30875 adjust zero sizes on alignment.
30877 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
30879 PR rtl-optimization/52573
30880 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
30881 REG_UNUSED for the same register.
30883 2013-01-17 Richard Biener <rguenther@suse.de>
30884 Marek Polacek <polacek@redhat.com>
30886 PR rtl-optimization/55833
30887 * loop-unswitch.c (unswitch_loops): Move loop verification...
30888 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
30889 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
30890 Set it to true when we're removing a loop from hierarchy tree in
30891 an irreducible region.
30892 (fix_bb_placements): Adjust caller.
30893 (fix_loop_placements): Likewise.
30895 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
30897 * config/avr/builtins.def (DEF_BUILTIN): Factor out
30898 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
30899 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
30900 Remove ID. Adjust comments.
30901 * config/avr/avr-c.c (avr_builtin_name): Remove.
30902 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
30903 * config/avr/avr.c (avr_tolower): New static function.
30904 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
30905 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
30906 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
30909 2013-01-17 Jan Hubicka <jh@suse.cz>
30911 PR tree-optimization/55273
30912 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
30914 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
30917 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
30918 store through atomic_store<mode>_1.
30919 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
30921 2013-01-17 Martin Jambor <mjambor@suse.cz>
30923 PR tree-optimizations/55264
30924 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
30925 for virtual methods.
30926 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
30927 virtual methods before inlining is over.
30928 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
30930 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
30933 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
30935 PR rtl-optimization/56005
30936 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
30937 pending reads for prefetch.
30939 2013-01-16 Ian Bolton <ian.bolton@arm.com>
30941 * config/aarch64/aarch64.md
30942 (*cstoresi_neg_uxtw): New pattern.
30943 (*cmovsi_insn_uxtw): New pattern.
30944 (*<optab>si3_uxtw): New pattern.
30945 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
30946 (*<optab>si3_insn_uxtw): New pattern.
30947 (*bswapsi2_uxtw): New pattern.
30949 2013-01-16 Richard Biener <rguenther@suse.de>
30951 * tree-inline.c (tree_function_versioning): Remove set but
30952 never used variable.
30954 2013-01-16 Richard Biener <rguenther@suse.de>
30956 PR tree-optimization/55964
30957 * tree-flow.h (rename_variables_in_loop): Remove.
30958 (rename_variables_in_bb): Likewise.
30959 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
30960 (copy_loop_before): Adjust and delete update-ssa status.
30961 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
30962 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
30963 (rename_variables_in_loop): Remove.
30964 (slpeel_update_phis_for_duplicate_loop): Likewise.
30965 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
30966 use available cfg machinery instead of duplicating it.
30967 Update PHI nodes and perform poor-mans SSA update here.
30968 (slpeel_tree_peel_loop_to_edge): Adjust.
30970 2013-01-16 Richard Biener <rguenther@suse.de>
30972 PR tree-optimization/54767
30973 PR tree-optimization/53465
30974 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
30975 (vrp_visit_phi_node): For PHI arguments coming via backedges
30976 drop all symbolical range information.
30977 (execute_vrp): Compute backedges.
30979 2013-01-16 Richard Biener <rguenther@suse.de>
30981 * doc/install.texi: Update CLooG and ISL requirements to
30984 2013-01-16 Christian Bruel <christian.bruel@st.com>
30987 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
30988 (broken_move): Handle UNSPECV_SP_SWITCH_B.
30989 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
30991 2013-01-16 DJ Delorie <dj@redhat.com>
30993 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
30994 (UNSPECV_SP_SWITCH_E): New.
30995 (sp_switch_1): Change to an unspec.
30996 (sp_switch_2): Change to an unspec. Don't use post-inc when we
30999 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
31001 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
31002 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
31003 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
31004 (expand_mem_thread_fence): Ditto.
31005 (expand_mem_signal_fence): Ditto.
31006 (expand_atomic_load): Ditto.
31007 (expand_atomic_store): Ditto.
31009 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
31011 PR rtl-optimization/55547
31012 PR rtl-optimization/53827
31015 * alias.c (memrefs_conflict_p): Set sizes to negative after
31018 2013-01-15 Jakub Jelinek <jakub@redhat.com>
31021 * function.c (thread_prologue_and_epilogue_insns): Always
31022 add crtl->drap_reg to set_up_by_prologue.set, even if
31023 stack_realign_drap is false.
31025 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
31027 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
31028 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
31029 *call): Fix indention.
31031 2013-01-15 Tom de Vries <tom@codesourcery.com>
31034 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
31037 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
31039 PR rtl-optimization/55153
31040 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
31042 2013-01-15 Martin Jambor <mjambor@suse.cz>
31044 PR tree-optimization/55920
31045 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
31046 accesses as grp_to_be_debug_replaced.
31048 2013-01-15 Jakub Jelinek <jakub@redhat.com>
31050 PR tree-optimization/55920
31051 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
31052 there is non-useless type conversion needed from debug rhs to lhs,
31053 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
31055 2013-01-15 Joseph Myers <joseph@codesourcery.com>
31056 Mikael Pettersson <mikpe@it.uu.se>
31059 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
31061 (ASM_OUTPUT_CASE_LABEL): Remove.
31062 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
31063 * final.c (shorten_branches): Update alignment of labels before
31064 jump tables if CASE_VECTOR_SHORTEN_MODE.
31066 2013-01-15 Richard Biener <rguenther@suse.de>
31069 * system.h: Do not include gmp.h for building host tools.
31071 2013-01-15 Richard Biener <rguenther@suse.de>
31073 PR middle-end/55882
31074 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
31075 account for bitpos when computing alignment.
31077 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
31079 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
31080 (ix86_target_macros_internal): Likewise.
31082 * config/i386/i386.c (m_CORE2I7): Removed.
31083 (m_CORE_HASWELL): New macro.
31084 (m_CORE_ALL): Likewise.
31085 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
31086 (initial_ix86_arch_features): Likewise.
31087 (processor_target_table): Initializations for Core avx2.
31088 (cpu_names): New names "core-avx2".
31089 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
31090 PROCESSOR_CORE_HASWELL.
31091 (ix86_issue_rate): New case.
31092 (ia32_multipass_dfa_lookahead): Likewise.
31093 (ix86_sched_init_global): Likewise.
31095 * config/i386/i386.h (TARGET_HASWELL): New macro.
31096 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
31097 (processor_type): New PROCESSOR_HASWELL.
31099 2013-01-15 Jakub Jelinek <jakub@redhat.com>
31101 PR tree-optimization/55955
31102 * tree-vect-loop.c (vectorizable_reduction): Give up early on
31103 *SHIFT_EXPR and *ROTATE_EXPR codes.
31105 PR tree-optimization/48766
31106 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
31107 -ftrapv disable -fwrapv.
31109 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
31112 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
31113 etc. to 1 and not to __flash.
31114 Use LL suffix for __INT24_MAX__ with -mint8.
31115 Use ULL suffix for __UINT24_MAX__ with -mint8.
31117 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
31119 * config/avr/avr-arch.h
31120 (struct base_arch_s): Use typedef avr_arch_t instead.
31121 (struct arch_info_s): Use typedef avr_arch_info_t instead.
31122 (struct mcu_type_s): Use typedef avr_mcu_t instead.
31123 * config/avr/avr.c: Same.
31124 * config/avr/avr-devices.c: Same.
31125 * config/avr/driver-avr.c: Same.
31126 * config/avr/gen-avr-mmcu-texi.c: Same.
31127 * config/avr/avr-mcus.def: Adjust comment.
31129 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
31131 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
31132 * config/aarch64/iterators.md (VALLDI): New.
31134 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
31135 Andi Kleen <ak@linux.intel.com>
31138 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
31139 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
31142 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
31144 * config/avr/avr-stdint.h: Remove trailing blanks.
31145 * config/avr/avr-log.h: Same.
31146 * config/avr/avr-arch.h: Same.
31147 * config/avr/avr-devices.c: Same.
31148 * config/avr/avr-dimode.md: Same.
31149 * config/avr/predicates.md: Same.
31150 * config/avr/avr-c.c: Same. And fix typo.
31152 * config/avr/avr-protos.h: Same. And:
31153 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
31154 (init_cumulative_args): Rename to avr_init_cumulative_args.
31155 (expand_prologue): Rename to avr_expand_prologue.
31156 (expand_epilogue): Rename to avr_expand_epilogue.
31157 (adjust_insn_length): Rename to avr_adjust_insn_length.
31158 (notice_update_cc): Rename to avr_notice_update_cc.
31159 (final_prescan_insn): Rename to avr_final_prescan_insn.
31160 * config/avr/avr.c: Same.
31161 * config/avr/avr.h: Same.
31162 * config/avr/avr.md: Remove trailing blanks.
31163 (prologue): Use avr_expand_prologue.
31164 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
31166 2013-01-14 Richard Biener <rguenther@suse.de>
31168 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
31169 verify_location, collect_subblocks): New functions.
31170 (verify_gimple_in_cfg): Verify that locations only reference
31171 BLOCKs in the functions BLOCK tree.
31173 2013-01-14 Richard Biener <rguenther@suse.de>
31175 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
31177 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
31179 (insert_out_of_ssa_copy_on_edge): Likewise.
31180 (rewrite_close_phi_out_of_ssa): Likewise.
31181 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
31183 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
31184 propagated constants.
31185 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
31188 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
31190 * config/avr/avr-modes.def: Add GPL copyright notice.
31192 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
31194 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
31195 MEMMODEL_MASK to determine memory model.
31196 (atomic_store<mode>): Ditto from operands[2].
31197 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
31199 2013-01-13 Jakub Jelinek <jakub@redhat.com>
31202 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
31203 (fold_gimple_assign): Don't call unshare_expr here.
31204 (fold_ctor_reference): Call unshare_expr.
31206 2013-01-13 Terry Guo <terry.guo@arm.com>
31208 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
31209 * doc/fragments.texi: Document MULTILIB_REUSE.
31210 * gcc.c (multilib_reuse): New internal spec.
31211 (set_multilib_dir): Also search multilib from multilib_reuse.
31212 * genmultilib (tmpmultilib3): Refactor code.
31213 (tmpmultilib4): Ditto.
31214 (multilib_reuse): New multilib argument.
31216 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
31218 * Makefile.in: Update copyright.
31220 2013-01-12 Tom de Vries <tom@codesourcery.com>
31222 PR middle-end/55890
31223 * calls.c (expand_call): Check if arg_nr is valid.
31225 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
31227 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
31228 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
31229 documentation. Add missing '__' in front of
31230 __builtin_ia32_packssdw256.
31232 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
31235 * config/s390/s390.c (s390_preferred_reload_class): Do not return
31236 NO_REGS for larl operands.
31237 (s390_reload_larl_operand): Use s390_load_address instead of
31240 2013-01-11 Richard Biener <rguenther@suse.de>
31242 * tree-cfg.c (verify_node_sharing_1): Split out from ...
31243 (verify_node_sharing): ... here.
31244 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
31246 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
31248 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
31249 Substitute TREECHECKING.
31250 * configure: Regenerate.
31251 * Makefile.in (TREECHECKING): New.
31253 2013-01-11 Richard Guenther <rguenther@suse.de>
31255 PR tree-optimization/44061
31256 * tree-vrp.c (extract_range_basic): Compute zero as
31257 value-range for __builtin_constant_p of function parameters.
31259 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
31261 Update copyright years.
31263 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
31265 PR rtl-optimization/55672
31266 * lra-eliminations.c (mark_not_eliminable): Permit addition with
31267 const to be eliminable.
31269 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
31271 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
31272 * configure: Regenerate.
31274 2013-01-10 Richard Biener <rguenther@suse.de>
31276 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
31278 2013-01-10 Richard Biener <rguenther@suse.de>
31281 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
31282 locations for virtual PHI arguments.
31283 (rewrite_update_phi_arguments): Likewise.
31285 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
31287 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
31290 2013-01-10 Jakub Jelinek <jakub@redhat.com>
31292 PR tree-optimization/55921
31293 * tree-complex.c (expand_complex_asm): New function.
31294 (expand_complex_operations_1): Call it for GIMPLE_ASM.
31296 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
31299 * config/s390/s390.c (s390_symref_operand_p)
31300 (s390_loadrelative_operand_p): Merge the two functions.
31301 (s390_check_qrst_address, print_operand_address): Add parameters
31302 to s390_loadrelative_operand_p invokation.
31303 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
31304 (s390_reload_larl_operand, s390_secondary_reload): Use
31305 s390_loadrelative_operand_p instead of s390_symref_operand_p.
31306 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
31308 2013-01-09 Mike Stump <mikestump@comcast.net>
31310 * dse.c (record_store): Remove unnecessary assert.
31312 2013-01-09 Jan Hubicka <jh@suse.cz>
31314 PR tree-optimization/55569
31315 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
31316 * cfgloop.h (scale_loop_profile): Likewise.
31318 2013-01-09 Jan Hubicka <jh@suse.cz>
31321 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
31323 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
31325 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
31327 PR middle-end/55114
31328 * expr.h (maybe_emit_group_store): Declare.
31329 * expr.c (maybe_emit_group_store): New function.
31330 * builtins.c (expand_builtin_int_roundingfn): Call it.
31331 (expand_builtin_int_roundingfn_2): Likewise.
31333 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
31335 PR rtl-optimization/55829
31336 * lra-constraints.c (match_reload): Add code for absent output.
31337 (curr_insn_transform): Add code for reloads of matched inputs
31340 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
31342 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
31343 attribute of movddup insn to DF.
31344 (*vec_interleave_lowv2df): Ditto.
31345 (vec_dupv2df): Ditto.
31347 2013-01-09 Jan Hubicka <jh@suse.cz>
31349 PR tree-optimiation/55875
31350 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
31351 EVERY_ITERATION parameter.
31352 (number_of_iterations_exit): Check if exit is executed every iteration.
31353 (idx_infer_loop_bounds): Similarly here.
31354 (n_of_executions_at_most): Simplify
31355 to only test for cases where statement is dominated by the
31356 particular bound; handle correctly the "postdominance" test.
31357 (scev_probably_wraps_p): Use max loop iterations info
31358 as a global bound first.
31360 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
31361 Nick Clifton <nickc@redhat.com>
31363 * config/v850/v850.md (cbranchsf4): New pattern.
31364 (cstoresf4): New pattern.
31365 (cbranchdf4): New pattern.
31366 (cstoredf4): New pattern.
31367 (movsicc): Disallow floating point comparisons.
31368 (cmpsf_le_insn): Fix order of operators.
31369 (cmpsf_lt_insn): Likewise.
31370 (cmpsf_eq_insn): Likewise.
31371 (cmpdf_le_insn): Likewise.
31372 (cmpdf_lt_insn): Likewise.
31373 (cmpdf_eq_insn): Likewise.
31374 (cmpsf_ge_insn): Use LE comparison.
31375 (cmpdf_ge_insn): Likewise.
31376 (cmpsf_gt_insn): Use LT comparison.
31377 (cmpdf_gt_insn): Likewise.
31378 (cmpsf_ne_insn): Delete pattern.
31379 (cmpdf_ne_insn): Delete pattern.
31380 * config/v850/v850.c (v850_gen_float_compare): Use
31381 gen_cmpdf_eq_insn for NE comparison.
31382 (v850_float_z_comparison_operator)
31383 (v850_float_nz_comparison_operator): Move from here ...
31384 * config/v850/predicates.md: ... to here. Move GT and GE
31385 comparisons into v850_float_z_comparison_operator.
31386 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
31388 (v850_float_nz_comparison_operator): Likewise.
31390 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
31392 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
31393 with calls to gen_insvsi/gen_insvdi.
31395 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
31397 * config/i386/i386.c (initial_ix86_tune_features): Set up
31398 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
31400 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
31401 Jakub Jelinek <jakub@redhat.com>
31403 PR tree-optimization/48189
31404 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
31405 If nitercst is 0, don't predict the exit edge.
31407 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
31409 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
31410 in asm_fprintf with reg_names.
31411 (aarch64_print_operand_address): Likewise.
31412 (aarch64_return_addr): Likewise.
31413 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
31415 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
31417 * config/pa/pa.h (VAL_U6_BITS_P): Define.
31418 (INT_U6_BITS): Likewise.
31419 * config/pa/predicates.md (uint6_operand): New predicate.
31420 (shift5_operand, shift6_operand): Likewise.
31421 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
31423 (lshrdi3): Use shift6_operand.
31424 (shrpsi4, shrpdi4): New insn patterns.
31425 (extzv): Delete expander.
31426 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
31427 predicates in unamed zero extract patterns. Tighten common constraint.
31428 (extv): Delete expander.
31429 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
31430 predicates in unamed sign extract patterns. Tighten common constraint.
31431 (insv): Delete expander.
31432 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
31433 predicates in unamed insert patterns. Tighten common constraint.
31434 Change uint32_operand predicate to uint6_operand predicate in unamed
31435 DImode pattern to insert constant values of type 1...1xxxx.
31437 2013-01-04 Jan Hubicka <jh@suse.cz>
31439 PR tree-optimization/55823
31440 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
31443 2013-01-08 Jakub Jelinek <jakub@redhat.com>
31444 Uros Bizjak <ubizjak@gmail.com>
31446 PR rtl-optimization/55845
31447 * df-problems.c (can_move_insns_across): Stop scanning at
31448 volatile_insn_p source instruction or give up if
31449 across_from .. across_to range contains any volatile_insn_p
31452 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
31454 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
31455 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
31457 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
31458 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
31460 2013-01-08 Jakub Jelinek <jakub@redhat.com>
31463 * asan.c (asan_clear_shadow): New function.
31464 (asan_emit_stack_protection): Use it.
31466 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
31468 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
31469 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
31470 with tab instead of space.
31472 2013-01-08 Nick Clifton <nickc@redhat.com>
31474 * config/rl78/rl78.c (rl78_expand_prologue): Always select
31475 register bank 0 at the start of an interrupt handler.
31476 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
31479 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
31481 * config/aarch64/aarch64-simd.md
31482 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
31483 (aarch64_simd_bsl): Likewise.
31484 (aarch64_vcond_internal<mode>): Likewise.
31485 (vcond<mode><mode>): Likewise.
31486 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
31487 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
31489 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
31491 * config/aarch64/aarch64-builtins.c
31492 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
31494 2013-01-08 Martin Jambor <mjambor@suse.cz>
31497 * tree-sra.c (analyze_access_subtree): Return true also after
31498 potentially creating a debug-only replacement.
31500 2013-01-08 Jakub Jelinek <jakub@redhat.com>
31502 PR middle-end/55890
31503 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
31505 PR tree-optimization/54120
31506 * tree-vrp.c (range_fits_type_p): Don't allow
31507 src_precision < precision from signed vr to unsigned_p
31508 if vr->min or vr->max is negative.
31509 (simplify_float_conversion_using_ranges): Test can_float_p
31510 against CODE_FOR_nothing.
31512 2013-01-08 Jakub Jelinek <jakub@redhat.com>
31513 Richard Biener <rguenther@suse.de>
31515 PR middle-end/55851
31516 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
31517 types instead of just INTEGER_TYPE types.
31519 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
31521 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
31524 2013-01-07 Steve Ellcey <sellcey@mips.com>
31527 * config/mips/mips.opt: Change mad to mmad to match documentation.
31529 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
31532 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
31533 .progmemx.data now.
31535 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
31538 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
31539 (avr_addrspace_t): Add .section_name field.
31540 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
31542 (avr_addrspace): Same. Initialize .section_name. Remove last
31543 NULL entry. Put __memx into .progmemx.data.
31544 (progmem_section_prefix): Remove.
31545 (avr_asm_init_sections): No need to initialize progmem_section.
31546 (avr_asm_named_section): Use avr_addrspace[].section_name to get
31547 section name prefix.
31548 (avr_asm_select_section): Ditto. And use get_unnamed_section to
31549 retrieve the progmem section.
31550 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
31551 boundary to run over avr_addrspace[].
31552 (avr_register_target_pragmas): Ditto.
31554 2013-01-06 Jakub Jelinek <jakub@redhat.com>
31556 * varasm.c (output_constant_def_contents): For asan_protect_global
31557 protected strings, adjust DECL_ALIGN if needed, before testing for
31559 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
31560 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
31562 (output_object_block): For asan protected decls, emit asan padding
31563 after their contents.
31564 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
31565 (asan_finish_file): Test it here instead.
31567 2013-01-07 Nick Clifton <nickc@redhat.com>
31568 Matthias Klose <doko@debian.org>
31569 Doug Kwan <dougkwan@google.com>
31570 H.J. Lu <hongjiu.lu@intel.com>
31573 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
31575 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
31577 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
31579 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
31581 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
31583 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
31586 * doc/install.texi (Cross-Compiler-Specific Options): Document
31589 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
31591 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
31592 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
31593 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
31594 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
31595 vqmovun_high_s64): Fix source operand number and update copyright.
31597 2013-01-07 Richard Biener <rguenther@suse.de>
31599 PR middle-end/55890
31600 * gimple.h (gimple_call_builtin_p): New overload.
31601 * gimple.c (validate_call): New function.
31602 (gimple_call_builtin_p): Likewise.
31603 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
31604 Use gimple_call_builtin_p.
31605 (find_func_clobbers): Likewise.
31606 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
31607 (strlen_optimize_stmt): Likewise.
31609 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
31611 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
31612 (vld1q_dup_*): Likewise.
31613 (vld1_*): Likewise.
31614 (vld1q_*): Likewise.
31615 (vld1_lane_*): Likewise.
31616 (vld1q_lane_*): Likewise.
31618 2013-01-07 Richard Biener <rguenther@suse.de>
31620 * lto-streamer.h (LTO_minor_version): Bump to 2.
31622 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
31624 * config/aarch64/aarch64-protos.h
31625 (aarch64_const_double_zero_rtx_p): Rename to...
31626 (aarch64_float_const_zero_rtx_p): ...this.
31627 (aarch64_float_const_representable_p): New.
31628 (aarch64_output_simd_mov_immediate): Likewise.
31629 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
31630 move immediate case.
31631 * config/aarch64/aarch64.c
31632 (aarch64_const_double_zero_rtx_p): Rename to...
31633 (aarch64_float_const_zero_rtx_p): ...this.
31634 (aarch64_print_operand): Allow printing of new constants.
31635 (aarch64_valid_floating_const): New.
31636 (aarch64_legitimate_constant_p): Check for valid floating-point
31638 (aarch64_simd_valid_immediate): Likewise.
31639 (aarch64_vect_float_const_representable_p): New.
31640 (aarch64_float_const_representable_p): Likewise.
31641 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
31642 (aarch64_output_simd_mov_immediate): New.
31643 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
31644 (*movdf_aarch64): Likewise.
31645 * config/aarch64/constraints.md (Ufc): New.
31646 (Y): call aarch64_float_const_zero_rtx.
31647 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
31649 2013-01-07 Richard Biener <rguenther@suse.de>
31651 PR tree-optimization/55888
31652 PR tree-optimization/55862
31653 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
31654 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
31655 not if it is contained therein.
31657 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
31659 * config/avr/t-avr: Typo.
31661 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
31664 * config/avr/t-avr: Don't automatically rebuild
31665 $(srcdir)/config/avr/t-multilib
31666 $(srcdir)/config/avr/avr-tables.opt
31667 $(srcdir)/doc/avr-mmcu.texi
31668 (avr-mcus): New phony target to build them on request.
31669 (s-avr-mlib, s-avr-mmcu-texi): Remove.
31670 * avr/avr-mcus.def: Adjust comments.
31672 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
31674 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
31676 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
31678 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
31680 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
31682 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
31684 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
31686 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
31687 to generate profiling.
31688 * config/rs6000/aix64.h (LIB_SPEC): Same.
31690 2013-01-04 Andrew Pinski <apinski@cavium.com>
31692 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
31694 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
31696 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
31698 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
31700 (ix86_expand_move): Ditto.
31701 (ix86_zero_extend_to_Pmode): Ditto.
31702 (ix86_expand_call): Ditto.
31703 (ix86_expand_special_args_builtin): Ditto.
31704 (ix86_expand_builtin): Ditto.
31706 2013-01-04 Richard Biener <rguenther@suse.de>
31708 PR tree-optimization/55862
31709 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
31710 translating them through PHI nodes.
31712 2013-01-04 Martin Jambor <mjambor@suse.cz>
31714 PR tree-optimization/55755
31715 * tree-sra.c (sra_modify_assign): Do not check that an access has no
31716 children when trying to avoid producing a VIEW_CONVERT_EXPR.
31718 2013-01-04 Marek Polacek <polacek@redhat.com>
31720 PR middle-end/55859
31721 * opts.c (default_options_optimization): Clarify error message.
31723 2013-01-04 Richard Biener <rguenther@suse.de>
31725 PR middle-end/55863
31726 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
31729 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
31732 * config/pa/pa.md (movsi): Revert previous change.
31733 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
31736 2013-01-03 Richard Henderson <rth@redhat.com>
31738 * config/i386/i386.c (ix86_expand_move): Always assign to op1
31739 after eliminating TLS symbols.
31741 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
31744 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
31745 * graphite-poly.c (debug_gmp_value): Likewise.
31747 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
31750 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
31751 selected code model, define __code_mode_small__, __code_model_medium__,
31752 __code_model_large__, __code_model_32__ or __code_model_kernel__.
31753 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
31754 xchg temporary register with %k. Declare temporary register as
31756 [__x86_64__]: For medium and large code models, preserve %rbx register.
31758 2013-01-03 Richard Biener <rguenther@suse.de>
31760 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
31761 (dump_subscript): Adjust.
31762 (finalize_ddr_dependent): Do not dump redundant info.
31763 (analyze_siv_subscript): Adjust.
31764 (subscript_dependence_tester): Likewise.
31765 (compute_affine_dependence): Likewise.
31767 2013-01-03 Richard Biener <rguenther@suse.de>
31770 2013-01-03 Richard Biener <rguenther@suse.de>
31772 PR tree-optimization/55857
31773 * tree-vect-stmts.c (vectorizable_load): Do not setup
31774 re-alignment for invariant loads.
31776 2013-01-02 Richard Biener <rguenther@suse.de>
31778 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
31779 invariant load do not generate a vector load from the scalar location.
31781 2013-01-03 Richard Biener <rguenther@suse.de>
31783 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
31784 for not vectorizing.
31785 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
31786 not build INDIRECT_REFs, call get_name once only.
31787 (vect_create_data_ref_ptr): Likewise. Dump base object kind
31788 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
31790 2013-01-03 Richard Biener <rguenther@suse.de>
31792 PR tree-optimization/55857
31793 * tree-vect-stmts.c (vectorizable_load): Do not setup
31794 re-alignment for invariant loads.
31796 2013-01-03 Richard Biener <rguenther@suse.de>
31799 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
31800 prefer a built-in decl.
31802 2013-01-03 Jakub Jelinek <jakub@redhat.com>
31804 * gcc.c (process_command): Update copyright notice dates.
31805 * gcov.c (print_version): Likewise.
31806 * gcov-dump.c (print_version): Likewise.
31808 PR rtl-optimization/55838
31809 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
31810 iv0.step, iv1.step and step.
31812 2013-01-03 Jakub Jelinek <jakub@redhat.com>
31813 Marc Glisse <marc.glisse@inria.fr>
31815 PR tree-optimization/55832
31816 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
31817 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
31818 integer_{one,zero}_node.
31820 2013-01-03 Jakub Jelinek <jakub@redhat.com>
31823 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
31824 * var-tracking.c (reverse_op): Don't add reverse ops to
31825 VALUEs that have already
31826 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
31828 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
31830 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
31832 2013-01-02 Teresa Johnson <tejohnson@google.com>
31834 * dumpfile.c (dump_loc): Print filename with location.
31835 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
31836 new location_t parameter to emit complete unroll message with
31837 new dump framework.
31838 (canonicalize_loop_induction_variables): Compute loops location
31839 and pass to try_unroll_loop_completely.
31840 * loop-unroll.c (report_unroll_peel): New function.
31841 (peel_loops_completely): Use new dump format with location
31842 for main dumpfile message, and invoke report_unroll_peel on success.
31843 (decide_unrolling_and_peeling): Ditto.
31844 (decide_peel_once_rolling): Remove old dumpfile message subsumed
31845 by report_unroll_peel.
31846 (decide_peel_completely): Ditto.
31847 (decide_unroll_constant_iterations): Ditto.
31848 (decide_unroll_runtime_iterations): Ditto.
31849 (decide_peel_simple): Ditto.
31850 (decide_unroll_stupid): Ditto.
31851 * cfgloop.c (get_loop_location): New function.
31852 * cfgloop.h (get_loop_location): Declare.
31854 2013-01-02 Sriraman Tallam <tmsriram@google.com>
31856 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
31859 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
31861 PR middle-end/55198
31862 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
31863 BLKmode objects when EXPAND_MEMORY is specified.
31865 2013-01-02 Sriraman Tallam <tmsriram@google.com>
31867 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
31869 (fold_builtin_cpu): Do not share cpu model decls across statements.
31871 2013-01-02 Jason Merrill <jason@redhat.com>
31874 * tree.c (build_array_type_1): Revert earlier change.
31876 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
31878 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
31880 * config/aarch64/aarch64-tune.md: Re-generate.
31882 2013-01-02 Richard Biener <rguenther@suse.de>
31884 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
31885 invariant load do not generate a vector load from the scalar location.
31887 2013-01-02 Richard Biener <rguenther@suse.de>
31890 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
31891 * configure: Regenerate.
31893 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
31895 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
31896 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
31897 (expand_builtin_int_roundingfn_2): Keep the original target around
31898 for the fallback case.
31900 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
31902 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
31903 to be clear for sign changes.
31905 2013-01-01 Jan Hubicka <jh@suse.cz>
31907 * ipa-inline-analysis.c: Fix formatting.
31909 2013-01-01 Jakub Jelinek <jakub@redhat.com>
31911 PR tree-optimization/55831
31912 * tree-vect-loop.c (get_initial_def_for_induction): Use
31913 gsi_after_labels instead of gsi_start_bb.
31915 Copyright (C) 2013 Free Software Foundation, Inc.
31917 Copying and distribution of this file, with or without modification,
31918 are permitted in any medium without royalty provided the copyright
31919 notice and this notice are preserved.