1 2014-04-25 Radovan Obradovic <robradovic@mips.com>
2 Tom de Vries <tom@codesourcery.com>
4 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
6 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
7 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
8 new argument to find_all_hard_reg_sets call.
10 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
13 Use HOST_WIDE_INT_C for mask literal.
14 (aarch_rev16_shleft_mask_imm_p): Likewise.
16 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
19 * config/sparc/sparc.md (ashlsi3_extend): Delete.
21 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
24 * config/i386/i386-c.c (ix86_target_macros): Define
25 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
27 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
29 * configure.ac (tga_func): Remove.
30 (LIB_TLS_SPEC): Remove.
31 * configure: Regenerate.
32 * config.in: Regenerate.
33 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
35 2014-04-25 Richard Biener <rguenther@suse.de>
38 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
39 call stmt use/clobber sets during stmt walk instead of
40 walking the possibly incomplete set of caller edges.
42 2014-04-25 Richard Biener <rguenther@suse.de>
45 * passes.c (apply_ipa_transforms): Inline into only caller ...
46 (execute_one_pass): ... here. Properly bring in function
47 bodies for nodes we want to apply IPA transforms to.
49 2014-04-24 Cong Hou <congh@google.com>
51 PR tree-optimization/60896
52 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
53 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
54 (vect_mark_pattern_stmts): Set the def type of all statements in
55 PATTERN_DEF_SEQ as vect_internal_def.
57 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
59 * doc/extend.texi (PowerPC Built-in Functions): Document new
60 powerpc extended divide, bcd, pack/unpack 128-bit, builtin
62 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
64 * config/rs6000/predicates.md (const_0_to_3_operand): New
65 predicate to match 0..3 integer constants.
67 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
68 to support adding miscellaneous builtin functions.
69 (BU_DFP_MISC_2): Likewise.
70 (BU_P7_MISC_1): Likewise.
71 (BU_P7_MISC_2): Likewise.
72 (BU_P8V_MISC_3): Likewise.
73 (BU_MISC_1): Likewise.
74 (BU_MISC_2): Likewise.
75 (DIVWE): Add extended divide builtin functions.
83 (DXEX): Add decimal floating-point builtin functions.
95 (CDTBCD): Add new BCD builtin functions.
99 (BCDADD_LT): Likewise.
100 (BCDADD_EQ): Likewise.
101 (BCDADD_GT): Likewise.
102 (BCDADD_OV): Likewise.
104 (BCDSUB_LT): Likewise.
105 (BCDSUB_EQ): Likewise.
106 (BCDSUB_GT): Likewise.
107 (BCDSUB_OV): Likewise.
108 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
109 (UNPACK_TD): Likewise.
111 (UNPACK_TF): Likewise.
112 (UNPACK_TF_0): Likewise.
113 (UNPACK_TF_1): Likewise.
114 (PACK_V1TI): Likewise.
115 (UNPACK_V1TI): Likewise.
117 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
118 support for decimal floating point builtin functions.
119 (rs6000_expand_ternop_builtin): Add checks for the new builtin
120 functions that take constant arguments.
121 (rs6000_invalid_builtin): Add decimal floating point builtin
123 (rs6000_init_builtins): Setup long double, _Decimal64, and
124 _Decimal128 types for new builtin functions.
125 (builtin_function_type): Set the unsigned flags appropriately for
126 the new builtin functions.
127 (rs6000_opt_masks): Add support for decimal floating point builtin
130 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
131 floating point builtin functions.
132 (RS6000_BTM_COMMON): Likewise.
133 (RS6000_BTI_long_double): Likewise.
134 (RS6000_BTI_dfloat64): Likewise.
135 (RS6000_BTI_dfloat128): Likewise.
136 (long_double_type_internal_node): Likewise.
137 (dfloat64_type_internal_node): Likewise.
138 (dfloat128_type_internal_node): Likewise.
140 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
141 2.07 bcd arithmetic instructions.
142 (UNSPEC_BCDSUB): Likewise.
143 (UNSPEC_BCD_OVERFLOW): Likewise.
144 (UNSPEC_BCD_ADD_SUB): Likewise.
145 (bcd_add_sub): Likewise.
146 (BCD_TEST): Likewise.
147 (bcd<bcd_add_sub>): Likewise.
148 (bcd<bcd_add_sub>_test): Likewise.
149 (bcd<bcd_add_sub>_test2): Likewise.
150 (bcd<bcd_add_sub>_<code>): Likewise.
151 (peephole2 for combined bcd ops): Likewise.
153 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
154 decimal floating point builtin functions.
155 (UNSPEC_DENBCD): Likewise.
156 (UNSPEC_DXEX): Likewise.
157 (UNSPEC_DIEX): Likewise.
158 (UNSPEC_DSCLI): Likewise.
159 (UNSPEC_DSCRI): Likewise.
160 (D64_D128): Likewise.
161 (dfp_suffix): Likewise.
162 (dfp_ddedpd_<mode>): Likewise.
163 (dfp_denbcd_<mode>): Likewise.
164 (dfp_dxex_<mode>): Likewise.
165 (dfp_diex_<mode>): Likewise.
166 (dfp_dscli_<mode>): Likewise.
167 (dfp_dscri_<mode>): Likewise.
169 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
171 (UNSPEC_CDTBCD): Likewise.
172 (UNSPEC_CBCDTD): Likewise.
173 (UNSPEC_DIVE): Add support for new extended divide builtin
175 (UNSPEC_DIVEO): Likewise.
176 (UNSPEC_DIVEU): Likewise.
177 (UNSPEC_DIVEUO): Likewise.
178 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
179 pack/unpack 128-bit types.
180 (UNSPEC_PACK_128BIT): Likewise.
181 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
182 (udiv<mode>3): Use idiv_ldiv mode attribute.
183 (div<mode>3): Likewise.
184 (addg6s): Add new BCD builtin functions.
187 (UNSPEC_DIV_EXTEND): Add support for new extended divide
189 (div_extend): Likewise.
190 (div<div_extend>_<mode>"): Likewise.
191 (FP128_64): Add support for new builtin functions to pack/unpack
193 (unpack<mode>): Likewise.
194 (unpacktf_0): Likewise.
195 (unpacktf_1): Likewise.
196 (unpack<mode>_dm): Likewise.
197 (unpack<mode>_nodm): Likewise.
198 (pack<mode>): Likewise.
199 (unpackv1ti): Likewise.
200 (packv1ti): Likewise.
202 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
204 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
207 2014-04-24 Jakub Jelinek <jakub@redhat.com>
209 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
210 * gimplify.c (omp_is_private): Change last argument's type to int.
211 Only diagnose lastprivate if the simd argument is 1, only diagnose
212 linear if the simd argument is 2.
213 (gimplify_omp_for): Adjust omp_is_private callers. When adding
214 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
215 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
216 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
217 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
218 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
219 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
220 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
221 * tree-nested.c (convert_nonlocal_omp_clauses,
222 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
224 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
227 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
230 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
232 * flag-types.h (enum ivar_visibility): Add.
234 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
236 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
239 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
241 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
243 2014-04-24 Radovan Obradovic <robradovic@mips.com>
244 Tom de Vries <tom@codesourcery.com>
246 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
247 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
249 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
250 * emit-rtl.c (try_split): Same.
252 2014-04-24 Radovan Obradovic <robradovic@mips.com>
253 Tom de Vries <tom@codesourcery.com>
255 * common.opt (fuse-caller-save): New option.
257 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
259 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of elements
262 2014-04-24 Richard Biener <rguenther@suse.de>
264 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
265 during TER and instead use the sepops interface for expanding
266 non-GIMPLE_SINGLE_RHS.
268 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
270 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
271 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
273 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
275 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
276 assembler 64-bit option.
277 * configure: Regenerate.
279 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
281 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
282 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
283 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
284 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
285 (TARGET_CRYPTO): Take TARGET_SIMD into account.
287 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
289 * config/aarch64/aarch64-builtins.c
290 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
291 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
292 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
293 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
295 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
296 (Vrevsuff): New mode attribute.
298 2014-04-24 Terry Guo <terry.guo@arm.com>
300 * config/arm/arm.h (machine_function): Define variable
301 after_arm_reorg here.
302 * config/arm/arm.c (after_arm_reorg): Remove the definition.
303 (arm_split_constant): Update the way to access variable
306 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
308 2014-04-23 Tom de Vries <tom@codesourcery.com>
310 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
312 2014-04-23 David Malcolm <dmalcolm@redhat.com>
314 * is-a.h: Update comments to reflect the following changes to the
315 "pointerness" of the API, making the template parameter match the
316 return type, allowing use of is-a.h with typedefs of pointers.
317 (is_a_helper::cast): Return a T rather then a pointer to a T, so
318 that the return type matches the parameter to the is_a_helper.
320 (dyn_cast): Likewise.
322 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
323 pointer from the is-a.h API.
325 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
326 (is_a_helper <cgraph_node *>::test): ...this, matching change to
328 (is_a_helper <varpool_node>::test): Likewise, convert to...
329 (is_a_helper <varpool_node *>::test): ...this.
331 (varpool_first_variable): Update for removal of implicit pointer
333 (varpool_next_variable): Likewise.
334 (varpool_first_static_initializer): Likewise.
335 (varpool_next_static_initializer): Likewise.
336 (varpool_first_defined_variable): Likewise.
337 (varpool_next_defined_variable): Likewise.
338 (cgraph_first_defined_function): Likewise.
339 (cgraph_next_defined_function): Likewise.
340 (cgraph_first_function): Likewise.
341 (cgraph_next_function): Likewise.
342 (cgraph_first_function_with_gimple_body): Likewise.
343 (cgraph_next_function_with_gimple_body): Likewise.
344 (cgraph_alias_target): Likewise.
345 (varpool_alias_target): Likewise.
346 (cgraph_function_or_thunk_node): Likewise.
347 (varpool_variable_node): Likewise.
348 (symtab_real_symbol_p): Likewise.
349 * cgraphunit.c (referred_to_p): Likewise.
350 (analyze_functions): Likewise.
351 (handle_alias_pairs): Likewise.
352 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
353 * gimple-ssa.h (gimple_vuse_op): Likewise.
354 (gimple_vdef_op): Likewise.
355 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
356 * gimple.c (gimple_build_asm_1): Likewise.
357 (gimple_build_try): Likewise.
358 (gimple_build_resx): Likewise.
359 (gimple_build_eh_dispatch): Likewise.
360 (gimple_build_omp_for): Likewise.
361 (gimple_omp_for_set_clauses): Likewise.
363 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
364 (is_a_helper <gimple_statement_asm *>::test): ...this.
365 (is_a_helper <gimple_statement_bind>::test): Convert to...
366 (is_a_helper <gimple_statement_bind *>::test): ...this.
367 (is_a_helper <gimple_statement_call>::test): Convert to...
368 (is_a_helper <gimple_statement_call *>::test): ...this.
369 (is_a_helper <gimple_statement_catch>::test): Convert to...
370 (is_a_helper <gimple_statement_catch *>::test): ...this.
371 (is_a_helper <gimple_statement_resx>::test): Convert to...
372 (is_a_helper <gimple_statement_resx *>::test): ...this.
373 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
374 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
375 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
376 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
377 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
378 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
379 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
380 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
381 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
382 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
383 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
384 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
385 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
386 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
387 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
388 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
389 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
390 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
391 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
392 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
393 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
394 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
395 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
396 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
397 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
398 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
399 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
400 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
401 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
402 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
403 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
404 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
405 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
406 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
407 (is_a_helper <gimple_statement_phi>::test): Convert to...
408 (is_a_helper <gimple_statement_phi *>::test): ...this.
409 (is_a_helper <gimple_statement_transaction>::test): Convert to...
410 (is_a_helper <gimple_statement_transaction *>::test): ...this.
411 (is_a_helper <gimple_statement_try>::test): Convert to...
412 (is_a_helper <gimple_statement_try *>::test): ...this.
413 (is_a_helper <gimple_statement_wce>::test): Convert to...
414 (is_a_helper <gimple_statement_wce *>::test): ...this.
415 (is_a_helper <const gimple_statement_asm>::test): Convert to...
416 (is_a_helper <const gimple_statement_asm *>::test): ...this.
417 (is_a_helper <const gimple_statement_bind>::test): Convert to...
418 (is_a_helper <const gimple_statement_bind *>::test): ...this.
419 (is_a_helper <const gimple_statement_call>::test): Convert to...
420 (is_a_helper <const gimple_statement_call *>::test): ...this.
421 (is_a_helper <const gimple_statement_catch>::test): Convert to...
422 (is_a_helper <const gimple_statement_catch *>::test): ...this.
423 (is_a_helper <const gimple_statement_resx>::test): Convert to...
424 (is_a_helper <const gimple_statement_resx *>::test): ...this.
425 (is_a_helper <const gimple_statement_eh_dispatch>::test):
427 (is_a_helper <const gimple_statement_eh_dispatch *>::test):
429 (is_a_helper <const gimple_statement_eh_filter>::test): Convert
431 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
432 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
434 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
436 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
438 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
440 (is_a_helper <const gimple_statement_omp_return>::test): Convert
442 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
443 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
445 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
446 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
448 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
449 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
450 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
451 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert
453 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
454 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
456 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
457 (is_a_helper <const gimple_statement_omp_target>::test): Convert
459 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
460 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
462 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
463 (is_a_helper <const gimple_statement_omp_single>::test): Convert
465 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
466 (is_a_helper <const gimple_statement_omp_teams>::test): Convert
468 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
469 (is_a_helper <const gimple_statement_omp_task>::test): Convert
471 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
472 (is_a_helper <const gimple_statement_phi>::test): Convert to...
473 (is_a_helper <const gimple_statement_phi *>::test): ...this.
474 (is_a_helper <const gimple_statement_transaction>::test): Convert
476 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
477 (is_a_helper <const gimple_statement_with_ops>::test): Convert
479 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
480 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
481 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
482 (is_a_helper <const gimple_statement_with_memory_ops>::test):
484 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
486 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert
488 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
490 (gimple_use_ops): Update for removal of implicit pointer from the
492 (gimple_set_use_ops): Likewise.
493 (gimple_vuse): Likewise.
494 (gimple_vdef): Likewise.
495 (gimple_vuse_ptr): Likewise.
496 (gimple_vdef_ptr): Likewise.
497 (gimple_set_vuse): Likewise.
498 (gimple_set_vdef): Likewise.
499 (gimple_omp_return_set_lhs): Likewise.
500 (gimple_omp_return_lhs): Likewise.
501 (gimple_omp_return_lhs_ptr): Likewise.
502 (gimple_call_fntype): Likewise.
503 (gimple_call_set_fntype): Likewise.
504 (gimple_call_set_internal_fn): Likewise.
505 (gimple_call_use_set): Likewise.
506 (gimple_call_clobber_set): Likewise.
507 (gimple_bind_vars): Likewise.
508 (gimple_bind_set_vars): Likewise.
509 (gimple_bind_body_ptr): Likewise.
510 (gimple_bind_set_body): Likewise.
511 (gimple_bind_add_stmt): Likewise.
512 (gimple_bind_block): Likewise.
513 (gimple_bind_set_block): Likewise.
514 (gimple_asm_ninputs): Likewise.
515 (gimple_asm_noutputs): Likewise.
516 (gimple_asm_nclobbers): Likewise.
517 (gimple_asm_nlabels): Likewise.
518 (gimple_asm_input_op): Likewise.
519 (gimple_asm_input_op_ptr): Likewise.
520 (gimple_asm_output_op): Likewise.
521 (gimple_asm_output_op_ptr): Likewise.
522 (gimple_asm_set_output_op): Likewise.
523 (gimple_asm_clobber_op): Likewise.
524 (gimple_asm_set_clobber_op): Likewise.
525 (gimple_asm_label_op): Likewise.
526 (gimple_asm_set_label_op): Likewise.
527 (gimple_asm_string): Likewise.
528 (gimple_catch_types): Likewise.
529 (gimple_catch_types_ptr): Likewise.
530 (gimple_catch_handler_ptr): Likewise.
531 (gimple_catch_set_types): Likewise.
532 (gimple_catch_set_handler): Likewise.
533 (gimple_eh_filter_types): Likewise.
534 (gimple_eh_filter_types_ptr): Likewise.
535 (gimple_eh_filter_failure_ptr): Likewise.
536 (gimple_eh_filter_set_types): Likewise.
537 (gimple_eh_filter_set_failure): Likewise.
538 (gimple_eh_must_not_throw_fndecl): Likewise.
539 (gimple_eh_must_not_throw_set_fndecl): Likewise.
540 (gimple_eh_else_n_body_ptr): Likewise.
541 (gimple_eh_else_e_body_ptr): Likewise.
542 (gimple_eh_else_set_n_body): Likewise.
543 (gimple_eh_else_set_e_body): Likewise.
544 (gimple_try_eval_ptr): Likewise.
545 (gimple_try_cleanup_ptr): Likewise.
546 (gimple_try_set_eval): Likewise.
547 (gimple_try_set_cleanup): Likewise.
548 (gimple_wce_cleanup_ptr): Likewise.
549 (gimple_wce_set_cleanup): Likewise.
550 (gimple_phi_capacity): Likewise.
551 (gimple_phi_num_args): Likewise.
552 (gimple_phi_result): Likewise.
553 (gimple_phi_result_ptr): Likewise.
554 (gimple_phi_set_result): Likewise.
555 (gimple_phi_arg): Likewise.
556 (gimple_phi_set_arg): Likewise.
557 (gimple_resx_region): Likewise.
558 (gimple_resx_set_region): Likewise.
559 (gimple_eh_dispatch_region): Likewise.
560 (gimple_eh_dispatch_set_region): Likewise.
561 (gimple_omp_critical_name): Likewise.
562 (gimple_omp_critical_name_ptr): Likewise.
563 (gimple_omp_critical_set_name): Likewise.
564 (gimple_omp_for_clauses): Likewise.
565 (gimple_omp_for_clauses_ptr): Likewise.
566 (gimple_omp_for_set_clauses): Likewise.
567 (gimple_omp_for_collapse): Likewise.
568 (gimple_omp_for_index): Likewise.
569 (gimple_omp_for_index_ptr): Likewise.
570 (gimple_omp_for_set_index): Likewise.
571 (gimple_omp_for_initial): Likewise.
572 (gimple_omp_for_initial_ptr): Likewise.
573 (gimple_omp_for_set_initial): Likewise.
574 (gimple_omp_for_final): Likewise.
575 (gimple_omp_for_final_ptr): Likewise.
576 (gimple_omp_for_set_final): Likewise.
577 (gimple_omp_for_incr): Likewise.
578 (gimple_omp_for_incr_ptr): Likewise.
579 (gimple_omp_for_set_incr): Likewise.
580 (gimple_omp_for_pre_body_ptr): Likewise.
581 (gimple_omp_for_set_pre_body): Likewise.
582 (gimple_omp_parallel_clauses): Likewise.
583 (gimple_omp_parallel_clauses_ptr): Likewise.
584 (gimple_omp_parallel_set_clauses): Likewise.
585 (gimple_omp_parallel_child_fn): Likewise.
586 (gimple_omp_parallel_child_fn_ptr): Likewise.
587 (gimple_omp_parallel_set_child_fn): Likewise.
588 (gimple_omp_parallel_data_arg): Likewise.
589 (gimple_omp_parallel_data_arg_ptr): Likewise.
590 (gimple_omp_parallel_set_data_arg): Likewise.
591 (gimple_omp_task_clauses): Likewise.
592 (gimple_omp_task_clauses_ptr): Likewise.
593 (gimple_omp_task_set_clauses): Likewise.
594 (gimple_omp_task_child_fn): Likewise.
595 (gimple_omp_task_child_fn_ptr): Likewise.
596 (gimple_omp_task_set_child_fn): Likewise.
597 (gimple_omp_task_data_arg): Likewise.
598 (gimple_omp_task_data_arg_ptr): Likewise.
599 (gimple_omp_task_set_data_arg): Likewise.
600 (gimple_omp_taskreg_clauses): Likewise.
601 (gimple_omp_taskreg_clauses_ptr): Likewise.
602 (gimple_omp_taskreg_set_clauses): Likewise.
603 (gimple_omp_taskreg_child_fn): Likewise.
604 (gimple_omp_taskreg_child_fn_ptr): Likewise.
605 (gimple_omp_taskreg_set_child_fn): Likewise.
606 (gimple_omp_taskreg_data_arg): Likewise.
607 (gimple_omp_taskreg_data_arg_ptr): Likewise.
608 (gimple_omp_taskreg_set_data_arg): Likewise.
609 (gimple_omp_task_copy_fn): Likewise.
610 (gimple_omp_task_copy_fn_ptr): Likewise.
611 (gimple_omp_task_set_copy_fn): Likewise.
612 (gimple_omp_task_arg_size): Likewise.
613 (gimple_omp_task_arg_size_ptr): Likewise.
614 (gimple_omp_task_set_arg_size): Likewise.
615 (gimple_omp_task_arg_align): Likewise.
616 (gimple_omp_task_arg_align_ptr): Likewise.
617 (gimple_omp_task_set_arg_align): Likewise.
618 (gimple_omp_single_clauses): Likewise.
619 (gimple_omp_single_clauses_ptr): Likewise.
620 (gimple_omp_single_set_clauses): Likewise.
621 (gimple_omp_target_clauses): Likewise.
622 (gimple_omp_target_clauses_ptr): Likewise.
623 (gimple_omp_target_set_clauses): Likewise.
624 (gimple_omp_target_child_fn): Likewise.
625 (gimple_omp_target_child_fn_ptr): Likewise.
626 (gimple_omp_target_set_child_fn): Likewise.
627 (gimple_omp_target_data_arg): Likewise.
628 (gimple_omp_target_data_arg_ptr): Likewise.
629 (gimple_omp_target_set_data_arg): Likewise.
630 (gimple_omp_teams_clauses): Likewise.
631 (gimple_omp_teams_clauses_ptr): Likewise.
632 (gimple_omp_teams_set_clauses): Likewise.
633 (gimple_omp_sections_clauses): Likewise.
634 (gimple_omp_sections_clauses_ptr): Likewise.
635 (gimple_omp_sections_set_clauses): Likewise.
636 (gimple_omp_sections_control): Likewise.
637 (gimple_omp_sections_control_ptr): Likewise.
638 (gimple_omp_sections_set_control): Likewise.
639 (gimple_omp_for_set_cond): Likewise.
640 (gimple_omp_for_cond): Likewise.
641 (gimple_omp_atomic_store_set_val): Likewise.
642 (gimple_omp_atomic_store_val): Likewise.
643 (gimple_omp_atomic_store_val_ptr): Likewise.
644 (gimple_omp_atomic_load_set_lhs): Likewise.
645 (gimple_omp_atomic_load_lhs): Likewise.
646 (gimple_omp_atomic_load_lhs_ptr): Likewise.
647 (gimple_omp_atomic_load_set_rhs): Likewise.
648 (gimple_omp_atomic_load_rhs): Likewise.
649 (gimple_omp_atomic_load_rhs_ptr): Likewise.
650 (gimple_omp_continue_control_def): Likewise.
651 (gimple_omp_continue_control_def_ptr): Likewise.
652 (gimple_omp_continue_set_control_def): Likewise.
653 (gimple_omp_continue_control_use): Likewise.
654 (gimple_omp_continue_control_use_ptr): Likewise.
655 (gimple_omp_continue_set_control_use): Likewise.
656 (gimple_transaction_body_ptr): Likewise.
657 (gimple_transaction_label): Likewise.
658 (gimple_transaction_label_ptr): Likewise.
659 (gimple_transaction_set_body): Likewise.
660 (gimple_transaction_set_label): Likewise.
662 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
663 * ipa-inline-analysis.c (inline_write_summary): Likewise.
664 * ipa-ref.c (ipa_record_reference): Likewise.
665 * ipa-reference.c (analyze_function): Likewise.
666 (ipa_reference_write_optimization_summary): Likewise.
667 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
668 (address_taken_from_non_vtable_p): Likewise.
669 (comdat_can_be_unshared_p_1): Likewise.
670 * lto-cgraph.c (lto_output_ref): Likewise.
671 (add_references): Likewise.
672 (compute_ltrans_boundary): Likewise.
673 (output_symtab): Likewise.
674 (input_ref): Likewise.
675 (input_cgraph_1): Likewise.
676 (output_cgraph_opt_summary): Likewise.
677 * lto-streamer-out.c (lto_output): Likewise.
678 (output_symbol_p): Likewise.
679 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
680 (lsei_start_function_in_partition): Likewise.
681 (lsei_next_variable_in_partition): Likewise.
682 (lsei_start_variable_in_partition): Likewise.
683 * symtab.c (insert_to_assembler_name_hash): Likewise.
684 (unlink_from_assembler_name_hash): Likewise.
685 (symtab_unregister_node): Likewise.
686 (symtab_remove_node): Likewise.
687 (dump_symtab_node): Likewise.
688 (verify_symtab_base): Likewise.
689 (verify_symtab_node): Likewise.
690 (symtab_make_decl_local): Likewise.
691 (symtab_alias_ultimate_target): Likewise.
692 (symtab_resolve_alias): Likewise.
693 (symtab_get_symbol_partitioning_class): Likewise.
694 * tree-phinodes.c (allocate_phi_node): Likewise.
695 (reserve_phi_args_for_new_edge): Likewise.
696 (remove_phi_args): Likewise.
697 * varpool.c (varpool_node_for_asm): Likewise.
698 (varpool_remove_unreferenced_decls): Likewise.
700 2014-04-23 Jeff Law <law@redhat.com>
702 PR tree-optimization/60902
703 * tree-ssa-threadedge.c
704 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
705 invalidate outputs from statements that do not produce useful
706 outputs for threading.
708 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
710 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
711 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
712 machine descriptions for Stack Smashing Protector.
714 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
716 * aarch64.md (<optab>_rol<mode>3): New pattern.
717 (<optab>_rolsi3_uxtw): Likewise.
718 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
720 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
722 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
723 (arm_cortex_a12_tune): Likewise.
725 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
727 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
729 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
731 * config/arm/arm.md (arm_rev16si2): New pattern.
732 (arm_rev16si2_alt): Likewise.
733 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
735 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
737 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
738 (rev16<mode>2_alt): Likewise.
739 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
740 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
741 (aarch_rev16_shleft_mask_imm_p): Likewise.
742 (aarch_rev16_p_1): Likewise.
743 (aarch_rev16_p): Likewise.
744 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
745 (aarch_rev16_shright_mask_imm_p): Likewise.
746 (aarch_rev16_shleft_mask_imm_p): Likewise.
748 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
750 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
751 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
753 (cortex_a53_extra_costs): Likewise.
754 (cortex_a57_extra_costs): Likewise.
755 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
756 (cortexa7_extra_costs): Likewise.
757 (cortexa8_extra_costs): Likewise.
758 (cortexa12_extra_costs): Likewise.
759 (cortexa15_extra_costs): Likewise.
760 (v7m_extra_costs): Likewise.
761 (arm_new_rtx_costs): Handle BSWAP.
763 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
765 * config/arm/arm.c (cortexa8_extra_costs): New table.
766 (arm_cortex_a8_tune): New tuning struct.
767 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
769 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
771 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
773 2014-04-23 Richard Biener <rguenther@suse.de>
775 * Makefile.in (OBJS): Remove loop-unswitch.o.
776 * loop-unswitch.c: Delete.
777 * tree-pass.h (make_pass_rtl_unswitch): Remove.
778 * passes.def (pass_rtl_unswitch): Likewise.
779 * loop-init.c (gate_rtl_unswitch): Likewise.
780 (rtl_unswitch): Likewise.
781 (pass_data_rtl_unswitch): Likewise.
782 (pass_rtl_unswitch): Likewise.
783 (make_pass_rtl_unswitch): Likewise.
784 * rtl.h (reversed_condition): Likewise.
785 (compare_and_jump_seq): Likewise.
786 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
788 * loop-unroll.c (compare_and_jump_seq): Likewise.
790 2014-04-23 Richard Biener <rguenther@suse.de>
792 PR tree-optimization/60903
793 * tree-ssa-loop-im.c (analyze_memory_references): Remove
794 commented code block.
795 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
796 loop flags to newly created BBs and edges.
798 2014-04-23 Nick Clifton <nickc@redhat.com>
800 * config/msp430/msp430.c (msp430_handle_option): Move function
802 (msp430_option_override): Simplify mcu and mcpu option handling.
803 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
804 support for -mhwmult command line option.
805 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
806 -mhwmult command line option.
807 (msp430_hwmult_enabled): Delete.
808 (msp43o_output_labelref): Add support for -mhwmult command line
810 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
811 (umulsidi3): Likewise.
812 * config/msp430/msp430.opt (mmcu): Add Report attribute.
813 (mcpu, mlarge, msmall): Likewise.
814 (mhwmult): New option.
815 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
817 (msp430_is_f5_mcu): Remove prototype.
818 (msp430_use_f5_series_hwmult): Add prototype.
819 * config/msp430/msp430-opts.h: New file.
820 * common/config/msp430: New directory.
821 * common/config/msp430/msp430-common.c: New file.
822 * config.gcc (msp430): Remove target_has_targetm_common.
823 * doc/invoke.texi: Document -mhwmult command line option.
825 2014-04-23 Nick Clifton <nickc@redhat.com>
827 * config/i386/cygwin.h (ENDFILE_SPEC): Include
828 default-manifest.o if it can be found in the search path.
829 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
831 2014-04-23 Terry Guo <terry.guo@arm.com>
833 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
835 2014-04-23 Richard Biener <rguenther@suse.de>
838 * tree-inline.c (declare_return_variable): Use mark_addressable.
840 2014-04-23 Richard Biener <rguenther@suse.de>
843 * loop-init.c (loop_optimizer_init): Make sure to apply
844 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
846 2014-04-22 Jakub Jelinek <jakub@redhat.com>
849 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
851 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
852 if flag_sanitize_undefined_trap_on_error.
853 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
854 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
855 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
856 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
857 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
858 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
859 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
860 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
861 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
862 * ubsan.c (ubsan_instrument_unreachable): Return
863 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
864 (ubsan_expand_null_ifn): Emit __builtin_trap ()
865 if flag_sanitize_undefined_trap_on_error and
866 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
867 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
868 instrument_bool_enum_load): Emit __builtin_trap () if
869 flag_sanitize_undefined_trap_on_error and
870 __builtin_handle_*_abort () if !flag_sanitize_recover.
871 * doc/invoke.texi (-fsanitize-recover,
872 -fsanitize-undefined-trap-on-error): Document.
874 2014-04-22 Christian Bruel <christian.bruel@st.com>
876 * config/sh/sh.md (mov<mode>): Replace movQIHI.
877 Force immediates to SImode.
879 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
881 * config/nios2/nios2.md (UNSPEC_ROUND): New.
883 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
884 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
885 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
886 (nios2_fpu_insn): Add entry for round.
887 (N2FPU_NO_ERRNO_P): Define.
888 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
890 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
892 2014-04-22 Richard Henderson <rth@redhat.com>
894 * config/aarch64/aarch64 (addti3, subti3): New expanders.
895 (add<GPI>3_compare0): Remove leading * from name.
896 (add<GPI>3_carryin): Likewise.
897 (sub<GPI>3_compare0): Likewise.
898 (sub<GPI>3_carryin): Likewise.
899 (<su_optab>mulditi3): New expander.
900 (multi3): New expander.
901 (madd<GPI>): Remove leading * from name.
903 2014-04-22 Martin Jambor <mjambor@suse.cz>
905 * cgraphclones.c (cgraph_function_versioning): Copy
906 ipa_transforms_to_apply instead of asserting it is empty.
908 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
911 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
912 on count_exp to get mode.
914 2014-04-22 Andrew Pinski <apinski@cavium.com>
916 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
917 Handle TLS for ILP32.
918 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
919 (tlsie_small_<mode>): this and handle PTR.
920 (tlsie_small_sidi): New pattern.
921 (tlsle_small): Change to an expand to handle ILP32.
922 (tlsle_small_<mode>): New pattern.
923 (tlsdesc_small): Rename to ...
924 (tlsdesc_small_<mode>): this and handle PTR.
926 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
928 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
930 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
932 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
933 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
934 (aarch64_types_signed_poly_qualifiers): Likewise.
935 (aarch64_types_unsigned_signed_qualifiers): Likewise.
936 (aarch64_types_poly_signed_qualifiers): Likewise.
937 (TYPES_REINTERP_SS): Type macro added.
938 (TYPES_REINTERP_SU): Likewise.
939 (TYPES_REINTERP_SP): Likewise.
940 (TYPES_REINTERP_US): Likewise.
941 (TYPES_REINTERP_PS): Likewise.
942 (aarch64_fold_builtin): New expression folding added.
943 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
944 Declarations removed.
945 (REINTERP_SS): Declarations added.
946 (REINTERP_US): Likewise.
947 (REINTERP_PS): Likewise.
948 (REINTERP_SU): Likewise.
949 (REINTERP_SP): Likewise.
950 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
951 (vreinterpretq_p8_f64): Likewise.
952 (vreinterpret_p16_f64): Likewise.
953 (vreinterpretq_p16_f64): Likewise.
954 (vreinterpret_f32_f64): Likewise.
955 (vreinterpretq_f32_f64): Likewise.
956 (vreinterpret_f64_f32): Likewise.
957 (vreinterpret_f64_p8): Likewise.
958 (vreinterpret_f64_p16): Likewise.
959 (vreinterpret_f64_s8): Likewise.
960 (vreinterpret_f64_s16): Likewise.
961 (vreinterpret_f64_s32): Likewise.
962 (vreinterpret_f64_s64): Likewise.
963 (vreinterpret_f64_u8): Likewise.
964 (vreinterpret_f64_u16): Likewise.
965 (vreinterpret_f64_u32): Likewise.
966 (vreinterpret_f64_u64): Likewise.
967 (vreinterpretq_f64_f32): Likewise.
968 (vreinterpretq_f64_p8): Likewise.
969 (vreinterpretq_f64_p16): Likewise.
970 (vreinterpretq_f64_s8): Likewise.
971 (vreinterpretq_f64_s16): Likewise.
972 (vreinterpretq_f64_s32): Likewise.
973 (vreinterpretq_f64_s64): Likewise.
974 (vreinterpretq_f64_u8): Likewise.
975 (vreinterpretq_f64_u16): Likewise.
976 (vreinterpretq_f64_u32): Likewise.
977 (vreinterpretq_f64_u64): Likewise.
978 (vreinterpret_s64_f64): Likewise.
979 (vreinterpretq_s64_f64): Likewise.
980 (vreinterpret_u64_f64): Likewise.
981 (vreinterpretq_u64_f64): Likewise.
982 (vreinterpret_s8_f64): Likewise.
983 (vreinterpretq_s8_f64): Likewise.
984 (vreinterpret_s16_f64): Likewise.
985 (vreinterpretq_s16_f64): Likewise.
986 (vreinterpret_s32_f64): Likewise.
987 (vreinterpretq_s32_f64): Likewise.
988 (vreinterpret_u8_f64): Likewise.
989 (vreinterpretq_u8_f64): Likewise.
990 (vreinterpret_u16_f64): Likewise.
991 (vreinterpretq_u16_f64): Likewise.
992 (vreinterpret_u32_f64): Likewise.
993 (vreinterpretq_u32_f64): Likewise.
995 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
997 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
998 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
999 (vreinterpret_p8_s8): Likewise.
1000 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
1001 (vreinterpret_p8_s16): Likewise.
1002 (vreinterpret_p8_s32): Likewise.
1003 (vreinterpret_p8_s64): Likewise.
1004 (vreinterpret_p8_f32): Likewise.
1005 (vreinterpret_p8_u8): Likewise.
1006 (vreinterpret_p8_u16): Likewise.
1007 (vreinterpret_p8_u32): Likewise.
1008 (vreinterpret_p8_u64): Likewise.
1009 (vreinterpret_p8_p16): Likewise.
1010 (vreinterpretq_p8_s8): Likewise.
1011 (vreinterpretq_p8_s16): Likewise.
1012 (vreinterpretq_p8_s32): Likewise.
1013 (vreinterpretq_p8_s64): Likewise.
1014 (vreinterpretq_p8_f32): Likewise.
1015 (vreinterpretq_p8_u8): Likewise.
1016 (vreinterpretq_p8_u16): Likewise.
1017 (vreinterpretq_p8_u32): Likewise.
1018 (vreinterpretq_p8_u64): Likewise.
1019 (vreinterpretq_p8_p16): Likewise.
1020 (vreinterpret_p16_s8): Likewise.
1021 (vreinterpret_p16_s16): Likewise.
1022 (vreinterpret_p16_s32): Likewise.
1023 (vreinterpret_p16_s64): Likewise.
1024 (vreinterpret_p16_f32): Likewise.
1025 (vreinterpret_p16_u8): Likewise.
1026 (vreinterpret_p16_u16): Likewise.
1027 (vreinterpret_p16_u32): Likewise.
1028 (vreinterpret_p16_u64): Likewise.
1029 (vreinterpret_p16_p8): Likewise.
1030 (vreinterpretq_p16_s8): Likewise.
1031 (vreinterpretq_p16_s16): Likewise.
1032 (vreinterpretq_p16_s32): Likewise.
1033 (vreinterpretq_p16_s64): Likewise.
1034 (vreinterpretq_p16_f32): Likewise.
1035 (vreinterpretq_p16_u8): Likewise.
1036 (vreinterpretq_p16_u16): Likewise.
1037 (vreinterpretq_p16_u32): Likewise.
1038 (vreinterpretq_p16_u64): Likewise.
1039 (vreinterpretq_p16_p8): Likewise.
1040 (vreinterpret_f32_s8): Likewise.
1041 (vreinterpret_f32_s16): Likewise.
1042 (vreinterpret_f32_s32): Likewise.
1043 (vreinterpret_f32_s64): Likewise.
1044 (vreinterpret_f32_u8): Likewise.
1045 (vreinterpret_f32_u16): Likewise.
1046 (vreinterpret_f32_u32): Likewise.
1047 (vreinterpret_f32_u64): Likewise.
1048 (vreinterpret_f32_p8): Likewise.
1049 (vreinterpret_f32_p16): Likewise.
1050 (vreinterpretq_f32_s8): Likewise.
1051 (vreinterpretq_f32_s16): Likewise.
1052 (vreinterpretq_f32_s32): Likewise.
1053 (vreinterpretq_f32_s64): Likewise.
1054 (vreinterpretq_f32_u8): Likewise.
1055 (vreinterpretq_f32_u16): Likewise.
1056 (vreinterpretq_f32_u32): Likewise.
1057 (vreinterpretq_f32_u64): Likewise.
1058 (vreinterpretq_f32_p8): Likewise.
1059 (vreinterpretq_f32_p16): Likewise.
1060 (vreinterpret_s64_s8): Likewise.
1061 (vreinterpret_s64_s16): Likewise.
1062 (vreinterpret_s64_s32): Likewise.
1063 (vreinterpret_s64_f32): Likewise.
1064 (vreinterpret_s64_u8): Likewise.
1065 (vreinterpret_s64_u16): Likewise.
1066 (vreinterpret_s64_u32): Likewise.
1067 (vreinterpret_s64_u64): Likewise.
1068 (vreinterpret_s64_p8): Likewise.
1069 (vreinterpret_s64_p16): Likewise.
1070 (vreinterpretq_s64_s8): Likewise.
1071 (vreinterpretq_s64_s16): Likewise.
1072 (vreinterpretq_s64_s32): Likewise.
1073 (vreinterpretq_s64_f32): Likewise.
1074 (vreinterpretq_s64_u8): Likewise.
1075 (vreinterpretq_s64_u16): Likewise.
1076 (vreinterpretq_s64_u32): Likewise.
1077 (vreinterpretq_s64_u64): Likewise.
1078 (vreinterpretq_s64_p8): Likewise.
1079 (vreinterpretq_s64_p16): Likewise.
1080 (vreinterpret_u64_s8): Likewise.
1081 (vreinterpret_u64_s16): Likewise.
1082 (vreinterpret_u64_s32): Likewise.
1083 (vreinterpret_u64_s64): Likewise.
1084 (vreinterpret_u64_f32): Likewise.
1085 (vreinterpret_u64_u8): Likewise.
1086 (vreinterpret_u64_u16): Likewise.
1087 (vreinterpret_u64_u32): Likewise.
1088 (vreinterpret_u64_p8): Likewise.
1089 (vreinterpret_u64_p16): Likewise.
1090 (vreinterpretq_u64_s8): Likewise.
1091 (vreinterpretq_u64_s16): Likewise.
1092 (vreinterpretq_u64_s32): Likewise.
1093 (vreinterpretq_u64_s64): Likewise.
1094 (vreinterpretq_u64_f32): Likewise.
1095 (vreinterpretq_u64_u8): Likewise.
1096 (vreinterpretq_u64_u16): Likewise.
1097 (vreinterpretq_u64_u32): Likewise.
1098 (vreinterpretq_u64_p8): Likewise.
1099 (vreinterpretq_u64_p16): Likewise.
1100 (vreinterpret_s8_s16): Likewise.
1101 (vreinterpret_s8_s32): Likewise.
1102 (vreinterpret_s8_s64): Likewise.
1103 (vreinterpret_s8_f32): Likewise.
1104 (vreinterpret_s8_u8): Likewise.
1105 (vreinterpret_s8_u16): Likewise.
1106 (vreinterpret_s8_u32): Likewise.
1107 (vreinterpret_s8_u64): Likewise.
1108 (vreinterpret_s8_p8): Likewise.
1109 (vreinterpret_s8_p16): Likewise.
1110 (vreinterpretq_s8_s16): Likewise.
1111 (vreinterpretq_s8_s32): Likewise.
1112 (vreinterpretq_s8_s64): Likewise.
1113 (vreinterpretq_s8_f32): Likewise.
1114 (vreinterpretq_s8_u8): Likewise.
1115 (vreinterpretq_s8_u16): Likewise.
1116 (vreinterpretq_s8_u32): Likewise.
1117 (vreinterpretq_s8_u64): Likewise.
1118 (vreinterpretq_s8_p8): Likewise.
1119 (vreinterpretq_s8_p16): Likewise.
1120 (vreinterpret_s16_s8): Likewise.
1121 (vreinterpret_s16_s32): Likewise.
1122 (vreinterpret_s16_s64): Likewise.
1123 (vreinterpret_s16_f32): Likewise.
1124 (vreinterpret_s16_u8): Likewise.
1125 (vreinterpret_s16_u16): Likewise.
1126 (vreinterpret_s16_u32): Likewise.
1127 (vreinterpret_s16_u64): Likewise.
1128 (vreinterpret_s16_p8): Likewise.
1129 (vreinterpret_s16_p16): Likewise.
1130 (vreinterpretq_s16_s8): Likewise.
1131 (vreinterpretq_s16_s32): Likewise.
1132 (vreinterpretq_s16_s64): Likewise.
1133 (vreinterpretq_s16_f32): Likewise.
1134 (vreinterpretq_s16_u8): Likewise.
1135 (vreinterpretq_s16_u16): Likewise.
1136 (vreinterpretq_s16_u32): Likewise.
1137 (vreinterpretq_s16_u64): Likewise.
1138 (vreinterpretq_s16_p8): Likewise.
1139 (vreinterpretq_s16_p16): Likewise.
1140 (vreinterpret_s32_s8): Likewise.
1141 (vreinterpret_s32_s16): Likewise.
1142 (vreinterpret_s32_s64): Likewise.
1143 (vreinterpret_s32_f32): Likewise.
1144 (vreinterpret_s32_u8): Likewise.
1145 (vreinterpret_s32_u16): Likewise.
1146 (vreinterpret_s32_u32): Likewise.
1147 (vreinterpret_s32_u64): Likewise.
1148 (vreinterpret_s32_p8): Likewise.
1149 (vreinterpret_s32_p16): Likewise.
1150 (vreinterpretq_s32_s8): Likewise.
1151 (vreinterpretq_s32_s16): Likewise.
1152 (vreinterpretq_s32_s64): Likewise.
1153 (vreinterpretq_s32_f32): Likewise.
1154 (vreinterpretq_s32_u8): Likewise.
1155 (vreinterpretq_s32_u16): Likewise.
1156 (vreinterpretq_s32_u32): Likewise.
1157 (vreinterpretq_s32_u64): Likewise.
1158 (vreinterpretq_s32_p8): Likewise.
1159 (vreinterpretq_s32_p16): Likewise.
1160 (vreinterpret_u8_s8): Likewise.
1161 (vreinterpret_u8_s16): Likewise.
1162 (vreinterpret_u8_s32): Likewise.
1163 (vreinterpret_u8_s64): Likewise.
1164 (vreinterpret_u8_f32): Likewise.
1165 (vreinterpret_u8_u16): Likewise.
1166 (vreinterpret_u8_u32): Likewise.
1167 (vreinterpret_u8_u64): Likewise.
1168 (vreinterpret_u8_p8): Likewise.
1169 (vreinterpret_u8_p16): Likewise.
1170 (vreinterpretq_u8_s8): Likewise.
1171 (vreinterpretq_u8_s16): Likewise.
1172 (vreinterpretq_u8_s32): Likewise.
1173 (vreinterpretq_u8_s64): Likewise.
1174 (vreinterpretq_u8_f32): Likewise.
1175 (vreinterpretq_u8_u16): Likewise.
1176 (vreinterpretq_u8_u32): Likewise.
1177 (vreinterpretq_u8_u64): Likewise.
1178 (vreinterpretq_u8_p8): Likewise.
1179 (vreinterpretq_u8_p16): Likewise.
1180 (vreinterpret_u16_s8): Likewise.
1181 (vreinterpret_u16_s16): Likewise.
1182 (vreinterpret_u16_s32): Likewise.
1183 (vreinterpret_u16_s64): Likewise.
1184 (vreinterpret_u16_f32): Likewise.
1185 (vreinterpret_u16_u8): Likewise.
1186 (vreinterpret_u16_u32): Likewise.
1187 (vreinterpret_u16_u64): Likewise.
1188 (vreinterpret_u16_p8): Likewise.
1189 (vreinterpret_u16_p16): Likewise.
1190 (vreinterpretq_u16_s8): Likewise.
1191 (vreinterpretq_u16_s16): Likewise.
1192 (vreinterpretq_u16_s32): Likewise.
1193 (vreinterpretq_u16_s64): Likewise.
1194 (vreinterpretq_u16_f32): Likewise.
1195 (vreinterpretq_u16_u8): Likewise.
1196 (vreinterpretq_u16_u32): Likewise.
1197 (vreinterpretq_u16_u64): Likewise.
1198 (vreinterpretq_u16_p8): Likewise.
1199 (vreinterpretq_u16_p16): Likewise.
1200 (vreinterpret_u32_s8): Likewise.
1201 (vreinterpret_u32_s16): Likewise.
1202 (vreinterpret_u32_s32): Likewise.
1203 (vreinterpret_u32_s64): Likewise.
1204 (vreinterpret_u32_f32): Likewise.
1205 (vreinterpret_u32_u8): Likewise.
1206 (vreinterpret_u32_u16): Likewise.
1207 (vreinterpret_u32_u64): Likewise.
1208 (vreinterpret_u32_p8): Likewise.
1209 (vreinterpret_u32_p16): Likewise.
1210 (vreinterpretq_u32_s8): Likewise.
1211 (vreinterpretq_u32_s16): Likewise.
1212 (vreinterpretq_u32_s32): Likewise.
1213 (vreinterpretq_u32_s64): Likewise.
1214 (vreinterpretq_u32_f32): Likewise.
1215 (vreinterpretq_u32_u8): Likewise.
1216 (vreinterpretq_u32_u16): Likewise.
1217 (vreinterpretq_u32_u64): Likewise.
1218 (vreinterpretq_u32_p8): Likewise.
1219 (vreinterpretq_u32_p16): Likewise.
1221 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1223 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
1225 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator
1228 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
1229 (vqnegd_s64): Likewise.
1230 (vqabs_s64): Likewise.
1231 (vqabsd_s64): Likewise.
1233 2014-04-22 Richard Henderson <rth@redhat.com>
1235 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
1236 computation to the top of the loop.
1238 2014-04-22 Renlin <renlin.li@arm.com>
1239 Jiong Wang <jiong.wang@arm.com>
1241 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
1242 * config/aarch64/aarch64.c (aarch64_layout_frame)
1243 (aarch64_initial_elimination_offset): Likewise.
1245 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
1247 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
1250 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
1252 * machmode.h (bitwise_mode_for_mode): Declare.
1253 * stor-layout.h (bitwise_type_for_mode): Likewise.
1254 * stor-layout.c (bitwise_mode_for_mode): New function.
1255 (bitwise_type_for_mode): Likewise.
1256 * builtins.c (fold_builtin_memory_op): Use it instead of
1257 int_mode_for_mode and build_nonstandard_integer_type.
1259 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1261 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
1262 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
1263 (*-*-solaris2*): Simplify.
1264 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
1265 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
1266 *-*-solaris2.9* handling.
1268 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
1270 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
1271 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
1273 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
1274 * configure: Regenerate.
1276 * config/i386/sol2-9.h: Remove.
1278 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
1279 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
1280 Remove Solaris 9 references.
1282 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
1284 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
1285 (floatuns<GPI:mode><GPF:mode>2): Remove.
1286 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
1287 and floatuns conversions.
1288 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
1289 and floatuns conversions.
1290 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
1291 (w1,w2): New mode attributes for inequal width conversions.
1293 2014-04-22 Renlin Li <Renlin.Li@arm.com>
1295 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
1296 the output asm format.
1298 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
1300 * config/aarch64/aarch64-simd.md
1301 (aarch64_cm<optab>di): Always split.
1302 (*aarch64_cm<optab>di): New.
1303 (aarch64_cmtstdi): Always split.
1304 (*aarch64_cmtstdi): New.
1306 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1308 PR tree-optimization/60823
1309 * omp-low.c (ipa_simd_modify_function_body): Go through
1310 all SSA_NAMEs and for those refering to vector arguments
1311 which are going to be replaced adjust SSA_NAME_VAR and,
1312 if it is a default definition, change it into a non-default
1313 definition assigned at the beginning of function from new_decl.
1314 (ipa_simd_modify_stmt_ops): Rewritten.
1315 * tree-dfa.c (set_ssa_default_def): When removing default def,
1316 check for NULL loc instead of NULL *loc.
1318 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1320 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
1321 restrictions on core registers for DImode values in Thumb2.
1323 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1325 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
1326 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
1328 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1330 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
1331 (*iordi_notzesidi_di): Likewise.
1332 (*iordi_notsesidi_di): Likewise.
1334 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1336 * config/arm/arm-protos.h (tune_params): New struct members.
1337 * config/arm/arm.c: Initialise tune_params per processor.
1338 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
1339 for speed, based on new tune_params.
1341 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1343 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro
1345 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added
1347 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment
1349 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
1350 * config/aarch64/arm_neon.h (vrnd_f64): Added.
1351 (vrnda_f64): Likewise.
1352 (vrndi_f64): Likewise.
1353 (vrndm_f64): Likewise.
1354 (vrndn_f64): Likewise.
1355 (vrndp_f64): Likewise.
1356 (vrndx_f64): Likewise.
1358 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1360 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
1361 GET_MODE_SIZE argument is enum machine_mode.
1363 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1366 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
1367 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
1369 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
1372 * asan.c (asan_emit_stack_protection): Force the base to align to
1373 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
1374 appropriate bits if STRICT_ALIGNMENT.
1375 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
1377 (expand_used_vars): Leave a space in the stack frame for alignment
1378 if STRICT_ALIGNMENT.
1380 2014-04-21 David Malcolm <dmalcolm@redhat.com>
1382 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
1384 (gimple_store_p): Likewise.
1385 (gimple_assign_load_p): Likewise.
1386 (gimple_assign_cast_p): Likewise.
1387 (gimple_clobber_p): Likewise.
1389 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
1390 rather than a gimple.
1391 (gimple_assign_cast_p): Likewise.
1393 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
1396 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
1397 If mode is DDmode and TARGET_E500_DOUBLE allow move.
1399 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
1400 more debug information for E500 if -mdebug=reg.
1402 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
1405 * config/i386/i386.c (ix86_expand_builtin)
1406 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
1407 register for target RTX.
1408 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
1410 2014-04-18 Cong Hou <congh@google.com>
1412 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
1413 the widen-mult pattern by handling two operands with different sizes,
1414 and operands whose size is smaller than half of the result type.
1416 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
1418 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
1419 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
1420 (do_estimate_edge_time): Compute it.
1421 * ipa-inline.c (want_inline_small_function_p): Bypass
1422 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
1424 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
1426 * ipa-inline.c (spec_rem): New static variable.
1427 (dump_overall_stats): New function.
1428 (dump_inline_stats): New function.
1430 2014-04-18 Richard Henderson <rth@redhat.com>
1432 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
1433 to GET_MODE_SIZE, not a reg_class_t.
1435 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1437 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
1438 (vsx_xxmrglw_<mode>): Likewise.
1440 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
1443 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
1444 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
1445 (rs6000_init_hard_regno_mode_ok): Likewise.
1447 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
1449 * ipa-inline.c (inline_small_functions): Account only non-cold
1451 * doc/invoke.texi (inline-unit-growth): Update documentation.
1453 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
1455 * config/rs6000/rs6000.md (addti3, subti3): New.
1457 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
1460 * config/i386/i386.c (ix86_expand_clear): Remove outdated
1461 comment. Check optimize_insn_for_size_p instead of
1462 optimize_insn_for_speed_p.
1464 2014-04-17 Martin Jambor <mjambor@suse.cz>
1466 * gimple-iterator.c (gsi_start_edge): New function.
1467 * gimple-iterator.h (gsi_start_edge): Declare.
1468 * tree-sra.c (single_non_eh_succ): New function.
1469 (disqualify_ops_if_throwing_stmt): Renamed to
1470 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
1471 having one non-EH successor BB.
1472 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
1473 generate loads into replacements.
1474 (sra_modify_assign): Likewise and and also use the simple path for
1476 (sra_modify_function_body): Commit statements on edges.
1478 2014-04-17 Richard Biener <rguenther@suse.de>
1481 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
1482 comparison results and add clarifying comment.
1484 2014-04-17 Jakub Jelinek <jakub@redhat.com>
1486 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
1487 (blank_mode): Initialize it.
1488 (emit_mode_size_inline, emit_mode_nunits_inline,
1489 emit_mode_inner_inline): New functions.
1490 (emit_insn_modes_h): Call them and surround their output with
1491 #if GCC_VERSION >= 4001 ... #endif.
1492 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
1493 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
1494 mode_* arrays if the argument is __builtin_constant_p.
1495 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
1496 is enum machine_mode.
1498 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1500 * passes.c (opt_pass::execute): Adjust.
1501 (pass_manager::execute_pass_mode_switching): Likewise.
1502 (early_local_passes::execute): Likewise.
1503 (execute_one_pass): Pass cfun to the pass's execute method.
1504 * tree-pass.h (opt_pass::execute): Add function * argument.
1505 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
1506 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
1507 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
1508 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1509 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
1510 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
1511 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
1512 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
1513 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
1514 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
1515 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
1516 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
1517 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
1518 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
1519 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
1520 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
1521 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
1522 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
1523 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
1524 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1525 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1526 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1527 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1528 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1529 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1530 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1531 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
1532 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
1533 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
1536 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1538 * passes.c (opt_pass::gate): Take function * argument.
1539 (gate_all_early_local_passes): Merge into
1540 (early_local_passes::gate): this.
1541 (gate_all_early_optimizations): Merge into
1542 (all_early_optimizations::gate): this.
1543 (gate_all_optimizations): Mege into
1544 (all_optimizations::gate): this.
1545 (gate_all_optimizations_g): Merge into
1546 (all_optimizations_g::gate): this.
1547 (gate_rest_of_compilation): Mege into
1548 (rest_of_compilation::gate): this.
1549 (gate_postreload): Merge into
1550 (postreload::gate): this.
1551 (dump_one_pass): Pass cfun to the pass's gate method.
1552 (execute_ipa_summary_passes): Likewise.
1553 (execute_one_pass): Likewise.
1554 (ipa_write_summaries_2): Likewise.
1555 (ipa_write_optimization_summaries_1): Likewise.
1556 (ipa_read_summaries_1): Likewise.
1557 (ipa_read_optimization_summaries_1): Likewise.
1558 (execute_ipa_stmt_fixups): Likewise.
1559 * tree-pass.h (opt_pass::gate): Add function * argument.
1560 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
1561 combine-stack-adj.c, combine.c, compare-elim.c,
1562 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1563 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
1564 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
1565 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
1566 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
1567 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
1568 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
1569 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
1570 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
1571 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
1572 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
1573 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
1574 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
1575 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
1576 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
1577 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1578 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1579 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1580 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1581 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1582 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1583 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1584 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
1585 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
1586 var-tracking.c, vtable-verify.c, web.c: Adjust.
1588 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1590 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
1591 * configure: Regenerate.
1593 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1595 * passes.c (dump_one_pass): don't check pass->has_gate.
1596 (execute_ipa_summary_passes): Likewise.
1597 (execute_one_pass): Likewise.
1598 (ipa_write_summaries_2): Likewise.
1599 (ipa_write_optimization_summaries_1): Likewise.
1600 (ipa_read_optimization_summaries_1): Likewise.
1601 (execute_ipa_stmt_fixups): Likewise.
1602 * tree-pass.h (pass_data::has_gate): Remove.
1603 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
1604 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
1605 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
1606 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1607 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
1608 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
1609 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
1610 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
1611 gimple-low.c, gimple-ssa-isolate-paths.c,
1612 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
1613 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
1614 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
1615 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
1616 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
1617 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
1618 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
1619 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
1620 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
1621 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
1622 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
1623 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
1624 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1625 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1626 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1627 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1628 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1629 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1630 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1631 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
1632 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
1633 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
1636 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1638 * pass_manager.h (pass_manager::register_dump_files_1): Remove
1640 * passes.c (pass_manager::register_dump_files_1): Merge into
1641 (pass_manager::register_dump_files): this, and remove its handling of
1642 properties since the pass always has the properties anyway.
1643 (pass_manager::pass_manager): Adjust.
1645 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1647 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
1648 * passes.c (pass_manager::register_dump_files_1): Remove dead code
1649 dealing with properties.
1650 (pass_manager::register_dump_files): Adjust.
1652 2014-03-20 Mark Wielaard <mjw@redhat.com>
1654 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
1655 then represent the bound as normal constant value.
1657 2014-04-17 Jakub Jelinek <jakub@redhat.com>
1660 Forward port from 4.8 branch
1661 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
1663 * config/i386/bmiintrin.h (_blsi_u32): New.
1667 (_blsmsk_u32): Ditto.
1668 (_blsmsk_u64): Ditto.
1669 (_tzcnt_u32): Ditto.
1670 (_tzcnt_u64): Ditto.
1672 2014-04-17 Kito Cheng <kito@0xlab.org>
1674 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
1676 2014-04-17 Richard Biener <rguenther@suse.de>
1679 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
1680 boolean results for comparisons.
1682 2014-04-17 Richard Biener <rguenther@suse.de>
1684 PR tree-optimization/60836
1685 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
1686 initial PHI args to be gimple values.
1688 2014-04-17 Richard Biener <rguenther@suse.de>
1690 PR tree-optimization/60841
1691 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
1692 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
1693 of stmts to SLP build.
1694 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
1695 (vect_analyze_slp): Likewise.
1696 (vect_analyze_slp_instance): Likewise.
1697 (vect_build_slp_tree): Limit overall SLP tree growth.
1698 * tree-vectorizer.h (vect_analyze_data_refs,
1699 vect_analyze_slp): Adjust prototypes.
1701 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
1703 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
1706 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
1708 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
1709 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
1710 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
1711 for TARGET_SLOW_PSHUFB
1713 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
1715 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
1716 * config/i386/i386.c (intel_cost): Ditto.
1718 2014-04-17 Joey Ye <joey.ye@arm.com>
1720 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
1722 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
1724 * opts.c (common_handle_option): Disable -fipa-reference coorectly
1727 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
1729 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
1730 (type_all_derivations_known_p): New predicate.
1731 (type_all_ctors_visible_p): New predicate.
1732 (type_possibly_instantiated_p): New predicate.
1733 (get_odr_type): Compute all_derivations_known.
1734 (dump_odr_type): Dump the flag.
1735 (maybe_record_type): Cleanup.
1736 (record_target_from_binfo): Add bases_to_consider array;
1737 record bases for types w/o instances and skip CXX destructor.
1738 (possible_polymorphic_call_targets_1): Add bases_to_consider
1739 and consider_construction parameters; check if type may have instance.
1740 (get_polymorphic_call_info): Set maybe_in_construction to true
1741 when we know nothing.
1742 (record_targets_from_bases): Skip CXX destructors; they are
1743 never called for types in construction.
1744 (possible_polymorphic_call_targets): Do not record target when
1745 type may not have instance.
1747 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
1750 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
1751 external aliases alive, too.
1753 2014-04-16 Andrew Pinski <apinski@cavium.com>
1755 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
1758 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
1760 * final.c (compute_alignments): Do not apply loop alignment to a block
1761 falling through to the exit.
1763 2014-04-16 Catherine Moore <clm@codesourcery.com>
1765 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
1766 Adjust constraints for microMIPS store patterns.
1768 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
1770 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
1772 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
1774 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
1775 (append_use): Run at -O0.
1776 (append_vdef): Likewise.
1777 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
1778 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
1780 2014-04-16 Jakub Jelinek <jakub@redhat.com>
1782 PR tree-optimization/60844
1783 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
1784 (propagate_op_to_single_use, remove_visited_stmt_chain,
1785 linearize_expr, repropagate_negates, reassociate_bb): Use it
1786 instead of gsi_remove.
1788 2014-04-16 Martin Jambor <mjambor@suse.cz>
1790 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
1791 ipa_transforms_to_apply.
1792 (cgraph_function_versioning): Assert that old_node has empty
1793 ipa_transforms_to_apply.
1794 * trans-mem.c (ipa_tm_create_version): Likewise.
1795 * tree-inline.c (tree_function_versioning): Do not duplicate
1796 ipa_transforms_to_apply.
1798 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1801 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
1803 Pass necessary as flags on 64-bit Solaris/x86.
1804 Use lowercase relocs for x86_64-*-*.
1805 * configure: Regenerate.
1807 2014-04-15 Jan Hubicka <jh@suse.cz>
1809 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
1810 (maybe_record_node, likely_target_p): Use it.
1812 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1815 Revert following patch
1817 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
1820 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
1821 software floating point or no floating point registers, do not
1822 allow any type in the FPRs. Eliminate a test for SPE SIMD types
1823 in GPRs that occurs after we tested for GPRs that would never be
1826 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
1827 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
1828 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
1829 specifically allow DDmode, since that does not use the SPE SIMD
1832 2014-03-21 Mark Wielaard <mjw@redhat.com>
1834 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
1835 as unsigned or int depending on type and value used.
1837 2014-04-15 Richard Biener <rguenther@suse.de>
1839 PR rtl-optimization/56965
1840 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
1841 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
1843 * alias.c (true_dependence_1): Do not call
1844 nonoverlapping_component_refs_p.
1845 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
1846 nonoverlapping_component_refs_p.
1847 (indirect_refs_may_alias_p): Likewise.
1849 2014-04-15 Teresa Johnson <tejohnson@google.com>
1851 * cfg.c (dump_bb_info): Fix flags check.
1852 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
1854 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1856 PR rtl-optimization/60663
1857 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
1860 2014-04-15 Richard Biener <rguenther@suse.de>
1862 * lto-streamer.h (LTO_major_version): Bump to 4.
1864 2014-04-15 Richard Biener <rguenther@suse.de>
1866 * common.opt (lto_partition_model): New enum.
1867 (flto-partition=): Merge separate options with a single with argument,
1868 add -flto-partition=one support.
1869 * flag-types.h (enum lto_partition_model): Declare.
1870 * opts.c (finish_options): Remove duplicate -flto-partition=
1872 * lto-wrapper.c (run_gcc): Adjust.
1874 2014-04-15 Richard Biener <rguenther@suse.de>
1876 * alias.c (ncr_compar): New function.
1877 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
1879 2014-04-15 Richard Biener <rguenther@suse.de>
1881 * alias.c (record_component_aliases): Do not walk BINFOs.
1883 2014-04-15 Richard Biener <rguenther@suse.de>
1885 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
1886 Add struct function argument and adjust.
1887 (find_func_aliases_for_call): Likewise.
1888 (find_func_aliases): Likewise.
1889 (find_func_clobbers): Likewise.
1890 (intra_create_variable_infos): Likewise.
1891 (compute_points_to_sets): Likewise.
1892 (ipa_pta_execute): Adjust. Do not push/pop cfun.
1894 2014-04-15 Richard Biener <rguenther@suse.de>
1896 * tree.c (iterative_hash_expr): Use enum tree_code_class
1897 to store TREE_CODE_CLASS.
1898 (tree_block): Likewise.
1899 (tree_set_block): Likewise.
1900 * tree.h (fold_build_pointer_plus_loc): Use
1901 convert_to_ptrofftype_loc.
1903 2014-04-15 Jakub Jelinek <jakub@redhat.com>
1906 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
1909 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
1911 * cfgloop.h (struct loop): Move force_vectorize down.
1912 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
1913 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
1914 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
1915 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
1916 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
1917 * tree-core.h (enum annot_expr_kind): Add new kind values.
1918 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
1919 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
1921 * tree.def (ANNOTATE_EXPR): Tweak comment.
1923 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
1925 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
1928 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
1930 * tree.h (TYPE_IDENTIFIER): Declare.
1931 * tree.c (subrange_type_for_debug_p): Use it.
1932 * godump.c (go_format_type): Likewise.
1933 * dwarf2out.c (is_cxx_auto, modified_type_die,
1934 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
1935 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
1937 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
1940 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
1942 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
1944 * config/i386/i386.c (examine_argument): Return bool. Return true if
1945 parameter should be passed in memory.
1946 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
1947 (construct_container): Update calls to examine_argument.
1948 (function_arg_advance_64): Ditto.
1949 (return_in_memory_32): Merge with ix86_return_in_memory.
1950 (return_in_memory_64): Ditto.
1951 (return_in_memory_ms_64): Ditto.
1953 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
1955 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
1956 * coverage.c (coverage_compute_profile_id): Handle externally visible
1959 2014-04-14 Martin Jambor <mjambor@suse.cz>
1961 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
1962 DECL_DISREGARD_INLINE_LIMITS functions.
1964 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
1967 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
1969 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
1972 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
1973 optimize_insn_for_speed_p instead of
1974 optimize_function_for_speed_p.
1976 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
1978 * doc/invoke.texi (free): Document AArch64.
1980 2014-04-14 Richard Biener <rguenther@suse.de>
1982 PR tree-optimization/60042
1983 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
1984 (insert_into_preds_of_block): Do not prevent PHI insertion
1985 for REFERENCE exprs here ...
1986 (eliminate_dom_walker::before_dom_children): ... but prevent
1987 their use here under similar conditions when applied to the
1988 IL after PRE optimizations.
1990 2014-04-14 Richard Biener <rguenther@suse.de>
1992 * passes.def: Move early points-to after early SRA.
1994 2014-04-14 Richard Biener <rguenther@suse.de>
1996 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
1997 check for which sign-changes we allow when forwarding
1998 a converted value into a switch.
2000 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2002 * stor-layout.c (place_field): Finalize non-constant offset for the
2005 2014-04-14 Richard Biener <rguenther@suse.de>
2007 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
2009 (expand_switch_using_bit_tests_p): Likewise.
2010 (process_switch): Compute and pass on speed_p based on the
2012 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
2013 optimize_bb_for_speed_p.
2015 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2017 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
2018 * function.h (struct function): Rename has_force_vect_loops into
2019 has_force_vectorize_loops.
2020 * lto-streamer-in.c (input_cfg): Adjust for renaming.
2021 (input_struct_function_base): Likewise.
2022 * lto-streamer-out.c (output_cfg): Likewise.
2023 (output_struct_function_base): Likewise.
2024 * omp-low.c (expand_omp_simd): Likewise.
2025 * tree-cfg.c (move_sese_region_to_fn): Likewise.
2026 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
2027 (version_loop_for_if_conversion): Likewise.
2028 (tree_if_conversion): Likewise.
2029 (main_tree_if_conversion): Likewise.
2030 (gate_tree_if_conversion): Likewise.
2031 * tree-inline.c (copy_loops): Likewise.
2032 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
2033 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
2034 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
2035 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
2036 * tree-vectorizer.c (vectorize_loops): Likewise.
2037 * tree-vectorizer.h (unlimited_cost_model): Likewise.
2039 2014-04-14 Richard Biener <rguenther@suse.de>
2042 * lto-streamer-out.c (wrap_refs): New function.
2043 (lto_output): Wrap symbol references in global initializes in
2044 type-preserving MEM_REFs.
2046 2014-04-14 Christian Bruel <christian.bruel@st.com>
2048 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
2050 2014-04-14 Christian Bruel <christian.bruel@st.com>
2052 * config/sh/sh.md (setmemqi): New expand pattern.
2053 * config/sh/sh.h (CLEAR_RATIO): Define.
2054 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
2055 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
2057 2014-04-14 Richard Biener <rguenther@suse.de>
2060 * fold-const.c (negate_expr_p): Don't negate directional rounding
2062 (fold_negate_expr): Likewise.
2064 2014-04-14 Richard Biener <rguenther@suse.de>
2066 PR tree-optimization/59817
2067 PR tree-optimization/60453
2068 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
2069 recursion to catch all CHRECs in the scalar evolution and restrict
2070 the predicate for the remains appropriately.
2072 2014-04-12 Catherine Moore <clm@codesourcery.com>
2074 * config/mips/constraints.md: Add new register constraint "kb".
2075 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
2076 (*movhi_internal): Likewise.
2077 (*movqi_internal): Likewise.
2078 * config/mips/mips.h (M16_STORE_REGS): New register class.
2079 (REG_CLASS_NAMES): Add M16_STORE_REGS.
2080 (REG_CLASS_CONTENTS): Likewise.
2081 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
2083 2014-04-11 Tobias Burnus <burnus@net-b.de>
2086 * doc/invoke.texi (-Wformat-signedness): Document it.
2087 (Wformat=2): Mention that this enables -Wformat-signedness.
2089 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2091 * common/config/epiphany/epiphany-common.c
2092 (epiphany_option_optimization_table): Enable section anchors by
2093 default at -O1 or higher.
2094 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
2095 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
2096 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
2097 carries no extra cost.
2098 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
2099 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
2100 * config/epiphany/predicates.md (memclob_operand): New predicate.
2101 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
2102 Use memclob_operand predicate and X constraint for operand 3.
2104 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2106 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
2107 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
2110 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2112 PR rtl-optimization/60651
2113 * mode-switching.c (optimize_mode_switching): Make sure to emit
2114 sets of a lower numbered entity before sets of a higher numbered
2115 entity to a mode of the same or lower priority.
2116 When creating a seginfo for a basic block that starts with a code
2117 label, move the insertion point past the code label.
2118 (new_seginfo): Document and enforce requirement that
2119 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
2120 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
2121 * doc/tm.texi: Regenerate.
2123 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
2126 * config/arc/arc.c (arc_save_restore): Fix assert typo.
2128 2013-04-11 Jakub Jelinek <jakub@redhat.com>
2130 * BASE-VER: Set to 4.10.0.
2132 2014-04-11 Tobias Burnus <burnus@net-b.de>
2135 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
2136 * doc/gcc.texi (Service): Update description in the @menu
2137 * doc/invoke.texi (Option Summary): Remove misplaced and
2140 2014-04-11 Steve Ellcey <sellcey@mips.com>
2141 Jakub Jelinek <jakub@redhat.com>
2144 * expr.c (convert_move): Use emit_store_flag_force instead of
2145 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
2148 2014-04-11 Richard Biener <rguenther@suse.de>
2151 * varasm.c (assemble_alias): Avoid endless error reporting
2152 recursion by setting TREE_ASM_WRITTEN.
2154 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2156 * config/s390/s390.md: Add a splitter for NOT rtx.
2158 2014-04-11 Jakub Jelinek <jakub@redhat.com>
2160 PR rtl-optimization/60663
2161 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
2163 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
2164 Jakub Jelinek <jakub@redhat.com>
2167 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
2168 flag from decl_node to node.
2170 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2173 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
2174 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
2175 ameliorating the cases where it can be.
2177 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
2180 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2182 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2183 (loadsync_<mode>): Change mode.
2184 (load_quadpti, store_quadpti): New.
2185 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2186 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2187 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
2189 2014-04-09 Cong Hou <congh@google.com>
2192 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
2195 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2197 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
2198 instead of vnor to exploit possible fusion opportunity in the
2200 (altivec_expand_vec_perm_const_le): Likewise.
2202 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2204 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2205 (loadsync_<mode>): Change mode.
2206 (load_quadpti, store_quadpti): New.
2207 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2208 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2210 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
2213 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
2214 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
2215 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
2217 2014-04-08 Richard Biener <rguenther@suse.de>
2220 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
2221 a 64bit widest int print double-int similar to on HWI64 hosts.
2223 2014-04-08 Richard Biener <rguenther@suse.de>
2225 PR tree-optimization/60785
2226 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
2227 default defs properly.
2229 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
2231 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
2232 (Weffc++): Likewise.
2234 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
2236 * ipa-devirt.c (maybe_record_node): When node is not recorded,
2237 set completep to false rather than true.
2239 2014-04-07 Douglas B Rupp <rupp@adacore.com>
2242 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
2243 ARM_TARGET2_DWARF_FORMAT.
2245 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
2248 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
2249 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
2252 2014-04-07 Richard Biener <rguenther@suse.de>
2254 PR tree-optimization/60766
2255 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
2256 (may_eliminate_iv): Convert cand_value_at result to desired type.
2258 2014-04-07 Jason Merrill <jason@redhat.com>
2261 * common.opt (-fno-gnu-unique): Add.
2262 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
2264 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2266 * haifa-sched.c: Fix outdated function reference and minor
2267 grammar errors in introductory comment.
2269 2014-04-07 Richard Biener <rguenther@suse.de>
2272 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
2274 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
2276 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
2279 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
2280 size accounting for thunks.
2281 (pa_asm_output_mi_thunk): Use final_start_function() and
2282 final_end_function() to output function start and end directives.
2284 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
2286 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
2287 device specific ISA/ feature information. Remove short_sp and
2288 errata_skip ds. Add avr_device_specific_features enum to have device
2290 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
2291 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
2292 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
2293 updated device specific info.
2294 * config/avr/avr-mcus.def: Merge device specific details to
2295 dev_attribute field.
2296 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
2298 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
2299 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
2300 assembler if RMW isa supported by current device.
2301 * config/avr/genmultilib.awk: Update as device info structure changed.
2302 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
2304 2014-04-04 Cong Hou <congh@google.com>
2306 PR tree-optimization/60656
2307 * tree-vect-stmts.c (supportable_widening_operation):
2308 Fix a bug that elements in a vector with vect_used_by_reduction
2309 property are incorrectly reordered when the operation on it is not
2310 consistant with the one in reduction operation.
2312 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
2314 PR rtl-optimization/60155
2315 * gcse.c (record_set_data): New function.
2316 (single_set_gcse): New function.
2317 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
2318 (hoist_code): Likewise.
2319 (get_pressure_class_and_nregs): Likewise.
2321 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
2323 * explow.c (probe_stack_range): Emit a final optimization blockage.
2325 2014-04-04 Anthony Green <green@moxielogic.com>
2327 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
2330 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
2333 * lto-cgraph.c (input_overwrite_node): Check that partitioning
2334 flags are set only during streaming.
2335 * ipa.c (process_references, walk_polymorphic_call_targets,
2336 symtab_remove_unreachable_nodes): Drop bodies of always inline
2337 after early inlining.
2338 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
2340 2014-04-04 Jakub Jelinek <jakub@redhat.com>
2341 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2344 * dwarf2out.c (const_ok_for_output_1): Reject expressions
2347 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2350 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
2352 (cortex_a53_fdivd): Likewise.
2354 2014-04-04 Martin Jambor <mjambor@suse.cz>
2357 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
2359 * cgraph.c (clone_of_p): Also return true if thunks match.
2360 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
2361 cgraph_function_or_thunk_node and an obsolete comment.
2362 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
2364 (build_function_decl_skip_args): Likewise.
2365 (set_new_clone_decl_and_node_flags): New function.
2366 (duplicate_thunk_for_node): Likewise.
2367 (redirect_edge_duplicating_thunks): Likewise.
2368 (cgraph_clone_node): New parameter args_to_skip, pass it to
2369 redirect_edge_duplicating_thunks which is called instead of
2370 cgraph_redirect_edge_callee.
2371 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
2372 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
2374 2014-04-04 Jeff Law <law@redhat.com>
2377 * config/arm/predicates.md (const_int_I_operand): New predicate.
2378 (const_int_M_operand): Similarly.
2379 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
2381 (insv_t2, extv_reg, extzv_t2): Likewise.
2382 (load_multiple_with_writeback): Similarly for const_int_I_operand.
2383 (pop_multiple_with_writeback_and_return): Likewise.
2384 (vfp_pop_multiple_with_writeback): Likewise
2386 2014-04-04 Richard Biener <rguenther@suse.de>
2389 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
2390 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
2392 * gimplify.h (gimple_add_tmp_var_fn): Declare.
2393 * gimplify.c (gimple_add_tmp_var_fn): New function.
2394 * gimple-expr.h (create_tmp_reg_fn): Declare.
2395 * gimple-expr.c (create_tmp_reg_fn): New function.
2396 * gimple-low.c (record_vars_into): Don't change cfun.
2397 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
2398 code generation without cfun.
2400 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
2403 * Makefile.in (install-driver): Fix shell scripting.
2405 2014-04-03 Cong Hou <congh@google.com>
2407 PR tree-optimization/60505
2408 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
2409 threshold of number of iterations below which no vectorization
2411 * tree-vect-loop.c (new_loop_vec_info):
2412 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
2413 * tree-vect-loop.c (vect_analyze_loop_operations):
2414 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
2415 * tree-vect-loop.c (vect_transform_loop):
2416 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
2417 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
2418 of iterations of the loop and see if we should build the epilogue.
2420 2014-04-03 Richard Biener <rguenther@suse.de>
2422 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
2423 (streamer_tree_cache_create): Adjust.
2424 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
2425 to allow optional nodes array.
2426 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
2427 (streamer_tree_cache_append): Likewise.
2428 (streamer_tree_cache_create): Create nodes array optionally
2429 as specified by parameter.
2430 * lto-streamer-out.c (create_output_block): Avoid maintaining
2431 the node array in the writer cache.
2432 (DFS_write_tree): Remove assertion.
2433 (produce_asm_for_decls): Free the out decl state hash table early.
2434 * lto-streamer-in.c (lto_data_in_create): Adjust for
2435 streamer_tree_cache_create prototype change.
2437 2014-04-03 Richard Biener <rguenther@suse.de>
2439 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
2440 set TREE_CHAIN to NULL_TREE.
2442 2014-04-03 Richard Biener <rguenther@suse.de>
2444 PR tree-optimization/60740
2445 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
2446 over all GIMPLE_COND operands.
2448 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
2450 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
2451 (Weffc++): Remove Scott's numbering, merge lists and reference
2454 2014-04-03 Nick Clifton <nickc@redhat.com>
2456 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
2459 2014-04-03 Martin Jambor <mjambor@suse.cz>
2461 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
2462 mention gcc_unreachable before failing.
2463 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
2466 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
2469 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
2470 inconsistent code and instead mark the context inconsistent.
2471 (possible_polymorphic_call_targets): For inconsistent contexts
2472 return empty complete list.
2474 2014-04-02 Anthony Green <green@moxielogic.com>
2476 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
2477 (extendqisi2, extendhisi2): Define.
2478 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
2479 (WCHAR_TYPE): Change to unsigned int.
2481 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2483 PR tree-optimization/60733
2484 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
2485 insertion point for PHI candidates to be the end of the feeding
2486 block for the PHI argument.
2488 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
2490 PR rtl-optimization/60650
2491 * lra-constraints.c (process_alt_operands): Decrease reject for
2492 earlyclobber matching.
2494 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2496 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
2498 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2500 * config/spu/spu.c (pad_bb): Do not crash when the last
2501 insn is CODE_FOR_blockage.
2503 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2505 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
2506 lies outside the target mode.
2508 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2511 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
2512 software floating point or no floating point registers, do not
2513 allow any type in the FPRs. Eliminate a test for SPE SIMD types
2514 in GPRs that occurs after we tested for GPRs that would never be
2517 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
2518 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
2519 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
2520 specifically allow DDmode, since that does not use the SPE SIMD
2523 2014-04-02 Richard Biener <rguenther@suse.de>
2526 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
2527 MODE_INTs. Properly use negv_optab.
2528 (expand_abs): Likewise.
2530 2014-04-02 Richard Biener <rguenther@suse.de>
2533 * Makefile.in (install-driver): Guard extra installs with special
2536 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
2538 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
2541 2014-04-01 Richard Henderson <rth@redhat.com>
2544 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
2545 alternative enabled before register allocation.
2547 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
2549 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
2550 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
2552 (nios2_large_got_address): Remove unneeded 'sym' parameter.
2553 (nios2_got_address): Update nios2_large_got_address call site.
2554 (nios2_delegitimize_address): New function.
2555 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
2556 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
2557 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
2559 2014-04-01 Martin Husemann <martin@duskware.de>
2561 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
2564 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
2566 PR rtl-optimization/60604
2567 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
2568 check from register_operand.
2569 (register_operand): Redefine in terms of general_operand.
2570 (nonmemory_operand): Use register_operand for the non-constant cases.
2572 2014-04-01 Richard Biener <rguenther@suse.de>
2574 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
2576 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
2578 * doc/invoke.texi (mapp-regs): Clarify.
2580 2014-03-31 Ulrich Drepper <drepper@gmail.com>
2582 * config/i386/avx512fintrin.h (__v32hi): Define type.
2583 (__v64qi): Likewise.
2584 (_mm512_set1_epi8): Define.
2585 (_mm512_set1_epi16): Define.
2586 (_mm512_set4_epi32): Define.
2587 (_mm512_set4_epi64): Define.
2588 (_mm512_set4_pd): Define.
2589 (_mm512_set4_ps): Define.
2590 (_mm512_setr4_epi64): Define.
2591 (_mm512_setr4_epi32): Define.
2592 (_mm512_setr4_pd): Define.
2593 (_mm512_setr4_ps): Define.
2594 (_mm512_setzero_epi32): Define.
2596 2014-03-31 Martin Jambor <mjambor@suse.cz>
2599 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
2600 callsite_arguments_match_p. Updated all callers. Also check types of
2601 corresponding formal parameters and actual arguments.
2602 (not_all_callers_have_enough_arguments_p) Renamed to
2603 some_callers_have_mismatched_arguments_p.
2605 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
2607 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
2609 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
2612 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
2615 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
2617 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
2618 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
2620 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
2621 Use FMAMODE_NOVF512 mode iterator.
2622 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
2623 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
2624 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
2625 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
2627 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
2628 Use VF_128_256 mode iterator.
2629 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
2632 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
2634 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
2635 static chain if needed.
2637 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
2640 * lra-constraints.c (index_part_to_reg): New.
2641 (process_address): Use it.
2643 2014-03-27 Jeff Law <law@redhat.com>
2644 Jakub Jelinek <jakub@redhat.com>
2647 * expr.c (do_tablejump): Use simplify_gen_binary rather than
2648 gen_rtx_{PLUS,MULT} to build up the address expression.
2650 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
2651 creating non-canonical RTL.
2653 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
2656 * ipa-inline.c (want_inline_small_function_p): Short circuit large
2657 functions; reorganize to make cheap checks first.
2658 (inline_small_functions): Do not estimate growth when dumping;
2660 * ipa-inline.h (inline_summary): Add min_size.
2661 (growth_likely_positive): New function.
2662 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
2663 (set_cond_stmt_execution_predicate): Cleanup.
2664 (estimate_edge_size_and_time): Compute min_size.
2665 (estimate_calls_size_and_time): Likewise.
2666 (estimate_node_size_and_time): Likewise.
2667 (inline_update_overall_summary): Update min_size.
2668 (do_estimate_edge_time): Likewise.
2669 (do_estimate_edge_size): Update.
2670 (do_estimate_edge_hints): Update.
2671 (growth_likely_positive): New function.
2673 2014-03-28 Jakub Jelinek <jakub@redhat.com>
2676 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
2677 also if addr has VOIDmode.
2679 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2681 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
2682 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
2684 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
2685 instructions as well as AdvancedSIMD loads.
2687 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2689 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
2690 Use crypto_aese type.
2691 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
2692 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
2693 crypto_aese, crypto_aesmc. Move to types.md.
2694 * config/arm/types.md (crypto_aes): Split into crypto_aese,
2696 * config/arm/iterators.md (crypto_type): Likewise.
2698 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
2700 * cgraph.c: Include expr.h and tree-dfa.h.
2701 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
2704 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
2707 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
2708 regs from checking multi-reg pseudos.
2710 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2712 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
2714 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2716 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
2717 if it would clobber the stack pointer, even temporarily.
2719 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
2721 * mode-switching.c: Make small adjustments to the top comment.
2723 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
2725 * config/rs6000/constraints.md (wD constraint): New constraint to
2726 match the constant integer to get the top DImode/DFmode out of a
2727 vector in a VSX register.
2729 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
2730 match the constant integer to get the top DImode/DFmode out of a
2731 vector in a VSX register.
2733 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
2736 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
2739 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
2740 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
2742 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
2743 Optimize vec_extract of 64-bit values, where the value being
2744 extracted is in the top word, where we can use scalar
2745 instructions. Add direct move and store support. Combine the big
2746 endian/little endian vector select load support into a single insn.
2747 (vsx_extract_<mode>_internal1): Likewise.
2748 (vsx_extract_<mode>_internal2): Likewise.
2749 (vsx_extract_<mode>_load): Likewise.
2750 (vsx_extract_<mode>_store): Likewise.
2751 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
2752 combined into vsx_extract_<mode>_load.
2753 (vsx_extract_<mode>_one_le): Likewise.
2755 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
2756 define the top 64-bit vector element.
2758 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
2761 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
2762 Document vec_vbpermq builtin.
2765 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
2766 enable use of xxsldwi and xxpermdi builtin functions.
2767 (vec_xxpermdi): Likewise.
2769 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
2770 Document use of vec_xxsldwi and vec_xxpermdi builtins.
2772 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
2774 PR rtl-optimization/60650
2775 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
2777 (find_spills_for): New.
2778 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
2779 Spill all pseudos on the second iteration.
2781 2014-03-27 Marek Polacek <polacek@redhat.com>
2784 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
2787 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2789 * config/s390/s390.c (s390_can_use_return_insn): Check for
2790 call-saved FPRs on 31 bit.
2792 2014-03-27 Jakub Jelinek <jakub@redhat.com>
2795 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
2796 if they need regimplification, just drop them instead of
2797 calling gimple_regimplify_operands on them.
2799 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
2802 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
2803 (aarch64_frame_pointer_required): Adjust logic.
2804 (aarch64_can_eliminate): Adjust logic.
2805 (aarch64_override_options_after_change): Adjust logic.
2807 2014-03-27 Dehao Chen <dehao@google.com>
2809 * ipa-inline.c (early_inliner): Update node's inline info.
2811 2014-03-26 Dehao Chen <dehao@google.com>
2813 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
2814 compiler inserted conditional jumps for NAN float check.
2816 2014-03-26 Jakub Jelinek <jakub@redhat.com>
2818 * ubsan.h (ubsan_create_data): Change second argument's type
2819 to const location_t *.
2820 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
2822 (ubsan_create_data): Change second argument to const location_t *PLOC.
2823 Create Loc field whenever PLOC is non-NULL.
2824 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
2825 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
2829 * real.c (real_to_integer2): Change type of low to UHWI.
2831 2014-03-26 Tobias Burnus <burnus@net-b.de>
2833 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
2834 (CILK_SELF_SPECS): New define.
2835 (driver_self_specs): Use it.
2837 2014-03-26 Richard Biener <rguenther@suse.de>
2839 * tree-pretty-print.c (percent_K_format): Implement special
2840 case for LTO and its stripped down BLOCK tree.
2842 2014-03-26 Jakub Jelinek <jakub@redhat.com>
2845 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
2847 * tree-vrp.c (simplify_internal_call_using_ranges): If only
2848 one range is range_int_cst_p, but not both, at least optimize
2849 addition/subtraction of 0 and multiplication by 0 or 1.
2850 * gimple-fold.c (gimple_fold_call): Fold
2851 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
2852 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
2853 INTEGER_CSTs, try to fold at least x * 0 and y - y.
2855 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
2857 PR rtl-optimization/60452
2858 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
2859 <case REG>: Return 1 for invalid offsets from the frame pointer.
2861 2014-03-26 Marek Polacek <polacek@redhat.com>
2864 * doc/extend.texi (C Extensions): Mention variable-length arrays in
2867 2014-03-26 Marek Polacek <polacek@redhat.com>
2870 * doc/extend.texi (Designated Inits): Describe what happens to omitted
2873 2014-03-26 Marek Polacek <polacek@redhat.com>
2876 * ira-color.c (update_conflict_hard_regno_costs): Perform the
2877 multiplication in unsigned type.
2879 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
2881 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
2883 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
2885 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
2887 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
2890 * cif-code.def (UNREACHABLE) New code.
2891 * ipa-inline.c (inline_small_functions): Skip edges to
2892 __builtlin_unreachable.
2893 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
2894 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
2895 predicate to __bulitin_unreachable.
2896 (set_cond_stmt_execution_predicate): Fix issue when
2897 invert_tree_comparison returns ERROR_MARK.
2898 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
2899 propagate to inline clones.
2900 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
2902 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
2903 * cgraphclones.c (cgraph_clone_node): If call destination is already
2904 ureachable, do not redirect it back.
2905 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
2908 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
2910 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
2911 Do not modify inline clones.
2913 2014-03-25 Jakub Jelinek <jakub@redhat.com>
2915 * config/i386/i386.md (general_sext_operand): New mode attr.
2916 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
2917 don't generate (sign_extend (const_int)).
2918 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
2919 operands[2]. Use We constraint instead of <i> and
2920 <general_sext_operand> predicate instead of <general_operand>.
2921 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
2922 * config/i386/constraints.md (We): New constraint.
2923 * config/i386/predicates.md (x86_64_sext_operand,
2924 sext_operand): New predicates.
2926 2014-03-25 Martin Jambor <mjambor@suse.cz>
2929 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
2930 inconsistent devirtualizations to __builtin_unreachable.
2932 2014-03-25 Marek Polacek <polacek@redhat.com>
2935 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
2937 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
2939 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
2940 order of elements for big-endian.
2942 2014-03-25 Richard Biener <rguenther@suse.de>
2945 * gimplify-me.c (gimple_regimplify_operands): Update the
2948 2014-03-25 Martin Jambor <mjambor@suse.cz>
2951 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
2952 (lto_output_varpool_node): Likewise.
2953 (input_overwrite_node): Likewise.
2954 (input_varpool_node): Likewise.
2956 2014-03-25 Richard Biener <rguenther@suse.de>
2958 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
2959 (run_gcc): Likewise.
2961 2014-03-25 Jakub Jelinek <jakub@redhat.com>
2963 * combine.c (simplify_compare_const): Add MODE argument.
2964 Handle mode_width 0 as very large mode_width.
2965 (try_combine, simplify_comparison): Adjust callers.
2967 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
2968 type to avoid signed integer overflow.
2969 * explow.c (plus_constant): Likewise.
2971 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
2973 * doc/generic.texi: Correct typos.
2975 2014-03-24 Tobias Burnus <burnus@net-b.de>
2977 * doc/invoke.texi (-flto): Expand section about
2978 using static libraries with LTO.
2980 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2982 PR rtl-optimization/60501
2983 * optabs.def (addptr3_optab): New optab.
2984 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
2985 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
2986 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
2988 * lra.c (emit_add3_insn): Use the addptr pattern if available.
2990 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
2992 2014-03-24 Ulrich Drepper <drepper@gmail.com>
2994 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
2997 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
2998 (_mm256_undefined_ps): Define.
2999 (_mm256_undefined_pd): Define.
3000 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
3001 (_mm_undefined_pd): Define.
3002 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
3003 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
3004 (_mm512_undefined_ps): Define.
3005 (_mm512_undefined_pd): Define.
3006 Use _mm*_undefined_*.
3007 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
3009 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
3011 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
3012 (lshr_simd): DI mode added.
3013 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
3014 (aarch64_ushr_simddi): Likewise.
3015 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
3016 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
3017 (vshrd_n_u64): Likewise.
3019 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3021 * Makefile.in (s-macro_list): Depend on cc1.
3023 2014-03-23 Teresa Johnson <tejohnson@google.com>
3025 * ipa-utils.c (ipa_print_order): Use specified dump file.
3027 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
3029 PR rtl-optimization/60601
3030 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
3032 * gcc.c (eval_spec_function): Initialize save_growing_value.
3034 2014-03-22 Jakub Jelinek <jakub@redhat.com>
3037 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
3038 code == MINUS_EXPR, never swap op0 with op1.
3040 * toplev.c (init_local_tick): Avoid signed integer multiplication
3042 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
3043 shift by first operand's bitsize.
3045 2014-03-21 Jakub Jelinek <jakub@redhat.com>
3048 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
3050 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
3051 TARGET_ISA_64BIT_P(x).
3053 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3055 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
3056 pattern for vector nor instead of subtract from splat(-1).
3057 (altivec_expand_vec_perm_const_le): Likewise.
3059 2014-03-21 Richard Henderson <rth@twiddle.net>
3062 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
3063 related insns after epilogue_completed.
3065 2014-03-21 Martin Jambor <mjambor@suse.cz>
3068 * cgraph.h (symtab_node): New flag body_removed.
3069 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
3070 when removing bodies.
3071 * symtab.c (dump_symtab_base): Dump body_removed flag.
3072 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
3073 had their bodies removed.
3075 2014-03-21 Martin Jambor <mjambor@suse.cz>
3078 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
3081 2014-03-21 Richard Biener <rguenther@suse.de>
3083 PR tree-optimization/60577
3084 * tree-core.h (struct tree_base): Document nothrow_flag use
3085 in VAR_DECL_NONALIASED.
3086 * tree.h (VAR_DECL_NONALIASED): New.
3087 (may_be_aliased): Adjust.
3088 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
3090 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3092 * expr.c (expand_expr_real_1): Remove outdated comment.
3094 2014-03-20 Jakub Jelinek <jakub@redhat.com>
3097 * ira.c (adjust_cleared_regs): Call copy_rtx on
3098 *reg_equiv[REGNO (loc)].src_p before passing it to
3099 simplify_replace_fn_rtx.
3102 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
3103 into CONST, put pic register as first operand of PLUS. Use
3104 gen_const_mem for both 32-bit and 64-bit PIC got loads.
3106 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3108 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
3110 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3112 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
3113 around for store forwarding issue in the FPU on the UT699.
3114 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
3115 loads and operations if -mfix-ut699 is specified.
3116 (divtf3_hq): Tweak attribute.
3117 (sqrttf2_hq): Likewise.
3119 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3121 * calls.c (store_one_arg): Remove incorrect const qualification on the
3122 type of the temporary.
3123 * cfgexpand.c (expand_return): Likewise.
3124 * expr.c (expand_constructor): Likewise.
3125 (expand_expr_real_1): Likewise.
3127 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3129 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
3132 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
3135 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
3137 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
3139 * config/arm/aarch-common-protos.h
3140 (alu_cost_table): Fix spelling of "extend".
3141 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
3143 2014-03-19 Richard Biener <rguenther@suse.de>
3146 * tree-core.h (tree_type_common): Re-order pointer members
3147 to reduce recursion depth during GC walks.
3149 2014-03-19 Marek Polacek <polacek@redhat.com>
3152 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
3153 before accessing it.
3155 2014-03-19 Richard Biener <rguenther@suse.de>
3158 * lto-streamer-in.c (input_function): In WPA stage do not drop
3161 2014-03-19 Jakub Jelinek <jakub@redhat.com>
3163 PR tree-optimization/60559
3164 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
3165 with build_zero_cst assignment.
3167 2014-03-18 Kai Tietz <ktietz@redhat.com>
3169 PR rtl-optimization/56356
3170 * sdbout.c (sdbout_parms): Verify that parms'
3171 incoming argument is valid.
3172 (sdbout_reg_parms): Likewise.
3174 2014-03-18 Richard Henderson <rth@redhat.com>
3177 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
3178 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
3179 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
3181 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
3183 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
3184 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
3185 Italicize plugin event names in description. Explain that
3186 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
3187 Remind that no GCC functions should be called after PLUGIN_FINISH.
3188 Explain what pragmas with expansion are.
3190 2014-03-18 Martin Liska <mliska@suse.cz>
3192 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
3193 gimple call statement is update.
3194 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
3195 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
3197 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3200 * ubsan.c (ubsan_instrument_unreachable): Call
3201 initialize_sanitizer_builtins.
3202 (ubsan_pass): Likewise.
3205 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
3206 varpool_finalize_decl instead of rest_of_decl_compilation.
3208 2014-03-18 Richard Biener <rguenther@suse.de>
3210 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
3211 by using bitmap_and_compl instead of bitmap_and_compl_into.
3212 (df_rd_transfer_function): Likewise.
3214 2014-03-18 Richard Biener <rguenther@suse.de>
3216 * doc/lto.texi (fresolution): Fix typo.
3218 2014-03-18 Richard Biener <rguenther@suse.de>
3220 * doc/invoke.texi (flto): Update for changes in 4.9.
3222 2014-03-18 Richard Biener <rguenther@suse.de>
3224 * doc/loop.texi: Remove section on the removed lambda framework.
3225 Update loop docs with recent changes in preserving loop structure.
3227 2014-03-18 Richard Biener <rguenther@suse.de>
3229 * doc/lto.texi (-fresolution): Document.
3231 2014-03-18 Richard Biener <rguenther@suse.de>
3233 * doc/contrib.texi: Adjust my name.
3235 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3238 * internal-fn.c: Include diagnostic-core.h.
3239 (expand_BUILTIN_EXPECT): New function.
3240 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
3241 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
3242 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
3243 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
3245 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
3246 Revert 3 argument __builtin_expect code.
3247 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
3248 * gimple-fold.c (gimple_fold_call): Likewise.
3249 * tree.h (fold_builtin_expect): New prototype.
3250 * builtins.c (build_builtin_expect_predicate): Add predictor
3251 argument, if non-NULL, create 3 argument __builtin_expect.
3252 (fold_builtin_expect): No longer static. Add ARG2 argument,
3253 pass it through to build_builtin_expect_predicate.
3254 (fold_builtin_2): Adjust caller.
3255 (fold_builtin_3): Handle BUILT_IN_EXPECT.
3256 * internal-fn.def (BUILTIN_EXPECT): New.
3258 2014-03-18 Tobias Burnus <burnus@net-b.de>
3261 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
3262 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
3263 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
3265 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
3268 * predict.c (combine_predictions_for_bb): Fix up formatting.
3269 (expr_expected_value_1, expr_expected_value): Add predictor argument,
3270 fill what it points to if non-NULL.
3271 (tree_predict_by_opcode): Adjust caller, use the predictor.
3272 * predict.def (PRED_COMPARE_AND_SWAP): Add.
3274 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
3276 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
3277 proper constant for the store mode.
3279 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
3281 * symtab.c (change_decl_assembler_name): Fix transparent alias
3284 2014-03-16 Renlin Li <Renlin.Li@arm.com>
3286 * config/aarch64/aarch64.c: Correct the comments about the
3287 aarch64 stack layout.
3289 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
3291 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
3292 check for GF_OMP_FOR_KIND_FOR.
3294 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
3296 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
3297 ymm and zmm register names.
3299 2014-03-17 Jakub Jelinek <jakub@redhat.com>
3302 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
3303 note creation for the 2010-08-31 changes.
3305 2014-03-17 Marek Polacek <polacek@redhat.com>
3308 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
3309 as -fno-tree-loop-vectorize.
3310 (expand_omp_simd): Likewise.
3312 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
3314 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
3315 (eligible_for_call_delay): New prototype.
3316 * config/sparc/sparc.c (tls_call_delay): Rename into...
3317 (eligible_for_call_delay): ...this. Return false if the instruction
3318 cannot be put in the delay slot of a branch.
3319 (eligible_for_restore_insn): Simplify.
3320 (eligible_for_return_delay): Return false if the instruction cannot be
3321 put in the delay slot of a branch and simplify.
3322 (eligible_for_sibcall_delay): Return false if the instruction cannot be
3323 put in the delay slot of a branch.
3324 * config/sparc/sparc.md (fix_ut699): New attribute.
3325 (tls_call_delay): Delete.
3326 (in_call_delay): Reimplement.
3327 (eligible_for_sibcall_delay): Rename into...
3328 (in_sibcall_delay): ...this.
3329 (eligible_for_return_delay): Rename into...
3330 (in_return_delay): ...this.
3331 (in_branch_delay): Reimplement.
3332 (in_uncond_branch_delay): Delete.
3333 (in_annul_branch_delay): Delete.
3335 2014-03-14 Richard Henderson <rth@redhat.com>
3338 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
3339 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
3340 (*floathi<X87MODEF>2_i387_with_temp): Remove.
3341 (floathi splitters): Remove.
3342 (float<SWI48x>xf2): New pattern.
3343 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
3344 code that tried to handle DImode for 32-bit, but which was excluded
3345 by the pattern's condition. Drop allocation of stack temporary.
3346 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
3347 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
3348 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
3349 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
3350 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
3351 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
3352 (*float<SWI48><MODEF>2_sse_interunit): Remove.
3353 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
3354 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
3355 (*float<SWI48x><X87MODEF>2_i387): Remove.
3356 (all float _with_temp splitters): Remove.
3357 (*float<SWI48x><MODEF>2_i387): New pattern.
3358 (*float<SWI48><MODEF>2_sse): New pattern.
3359 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
3360 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
3362 2014-03-14 Jakub Jelinek <jakub@redhat.com>
3363 Marek Polacek <polacek@redhat.com>
3366 * common.opt (dump_base_name_prefixed): New Variable.
3367 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
3368 if x_dump_base_name_prefixed is already set, set it at the end.
3370 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
3372 PR rtl-optimization/60508
3373 * lra-constraints.c (get_reload_reg): Add new parameter
3375 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
3376 Pass the new parameter values.
3378 2014-03-14 Richard Biener <rguenther@suse.de>
3380 * common.opt: Revert unintented changes from r205065.
3383 2014-03-14 Richard Biener <rguenther@suse.de>
3386 * cfghooks.c (split_block): Properly adjust all loops the
3387 block was a latch of.
3389 2014-03-14 Martin Jambor <mjambor@suse.cz>
3392 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
3395 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
3398 * config/avr/avr.c (avr_set_current_function): Pass function name
3399 through default_strip_name_encoding before sanity checking instead
3400 of skipping the first char of the assembler name.
3402 2014-03-13 Richard Henderson <rth@redhat.com>
3405 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
3406 (ix86_force_to_memory, ix86_free_from_memory): Remove.
3407 * config/i386/i386-protos.h: Likewise.
3408 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
3409 in the expander instead of a splitter.
3410 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
3411 any possibility of requiring a memory.
3412 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
3413 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
3414 (fp branch splitters): Update for ix86_split_fp_branch.
3415 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
3416 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
3417 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
3418 (*fop_<MODEF>_2_i387): Remove f/r alternative.
3419 (*fop_<MODEF>_3_i387): Likewise.
3420 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
3421 (splitters for the fop_* register patterns): Remove.
3422 (fscalexf4_i387): Rename from *fscalexf4_i387.
3423 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
3425 2014-03-13 Jakub Jelinek <jakub@redhat.com>
3427 PR tree-optimization/59779
3428 * tree-dfa.c (get_ref_base_and_extent): Use double_int
3429 type for bitsize and maxsize instead of HOST_WIDE_INT.
3431 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
3433 PR rtl-optimization/57320
3434 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
3435 the CFG after thread_prologue_and_epilogue_insns.
3437 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
3439 PR rtl-optimization/57189
3440 * lra-constraints.c (process_alt_operands): Disfavor spilling
3443 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
3445 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
3447 2014-03-13 Jakub Jelinek <jakub@redhat.com>
3449 PR tree-optimization/59025
3451 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
3452 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
3454 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
3457 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
3458 calls of avr_out_plus_1.
3460 2014-03-13 Bin Cheng <bin.cheng@arm.com>
3462 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
3463 BB's single pred and update the father loop's latch info later.
3465 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
3467 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
3471 (VEC_base): Likewise.
3472 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
3473 registers, we need to swap double words in little endian mode.
3475 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
3476 to be a container mode for 128-bit integer operations added in ISA
3477 2.07. Unlike TImode and PTImode, the preferred register set is
3478 the Altivec/VMX registers for the 128-bit operations.
3480 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
3482 (rs6000_split_128bit_ok_p): Likewise.
3484 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
3485 macros for creating ISA 2.07 normal and overloaded builtin
3486 functions with 3 arguments.
3487 (BU_P8V_OVERLOAD_3): Likewise.
3488 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
3489 for use as overloaded functions.
3490 (VPERM_1TI_UNS): Likewise.
3491 (VSEL_1TI): Likewise.
3492 (VSEL_1TI_UNS): Likewise.
3493 (ST_INTERNAL_1ti): Likewise.
3494 (LD_INTERNAL_1ti): Likewise.
3495 (XXSEL_1TI): Likewise.
3496 (XXSEL_1TI_UNS): Likewise.
3497 (VPERM_1TI): Likewise.
3498 (VPERM_1TI_UNS): Likewise.
3499 (XXPERMDI_1TI): Likewise.
3500 (SET_1TI): Likewise.
3501 (LXVD2X_V1TI): Likewise.
3502 (STXVD2X_V1TI): Likewise.
3503 (VEC_INIT_V1TI): Likewise.
3504 (VEC_SET_V1TI): Likewise.
3505 (VEC_EXT_V1TI): Likewise.
3506 (EQV_V1TI): Likewise.
3507 (NAND_V1TI): Likewise.
3508 (ORC_V1TI): Likewise.
3509 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
3510 added in ISA 2.07. Add both normal 'altivec' builtins, and the
3512 (VADDUQM): Likewise.
3513 (VSUBCUQ): Likewise.
3514 (VADDEUQM): Likewise.
3515 (VADDECUQ): Likewise.
3516 (VSUBEUQM): Likewise.
3517 (VSUBECUQ): Likewise.
3519 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
3520 __int128_t and __uint128_t types.
3521 (__uint128_type): Likewise.
3522 (altivec_categorize_keyword): Add support for vector __int128_t,
3523 vector __uint128_t, vector __int128, and vector unsigned __int128
3524 as a container type for TImode operations that need to be done in
3525 VSX/Altivec registers.
3526 (rs6000_macro_to_expand): Likewise.
3527 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
3528 to support 128-bit integer instructions vaddcuq, vadduqm,
3529 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
3530 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
3532 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
3533 for V1TImode, and set up preferences to use VSX/Altivec registers.
3534 Setup VSX reload handlers.
3535 (rs6000_debug_reg_global): Likewise.
3536 (rs6000_init_hard_regno_mode_ok): Likewise.
3537 (rs6000_preferred_simd_mode): Likewise.
3538 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
3539 (easy_altivec_constant): Likewise.
3540 (output_vec_const_move): Likewise.
3541 (rs6000_expand_vector_set): Convert V1TImode set and extract to
3543 (rs6000_expand_vector_extract): Likewise.
3544 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
3546 (rs6000_const_vec): Add support for V1TImode.
3547 (rs6000_emit_le_vsx_load): Swap double words when loading or
3548 storing TImode/V1TImode.
3549 (rs6000_emit_le_vsx_store): Likewise.
3550 (rs6000_emit_le_vsx_move): Likewise.
3551 (rs6000_emit_move): Add support for V1TImode.
3552 (altivec_expand_ld_builtin): Likewise.
3553 (altivec_expand_st_builtin): Likewise.
3554 (altivec_expand_vec_init_builtin): Likewise.
3555 (altivec_expand_builtin): Likewise.
3556 (rs6000_init_builtins): Add support for V1TImode type. Add
3557 support for ISA 2.07 128-bit integer builtins. Define type names
3558 for the VSX/Altivec vector types.
3559 (altivec_init_builtins): Add support for overloaded vector
3560 functions with V1TImode type.
3561 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
3562 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
3564 (rs6000_split_128bit_ok_p): Likewise.
3565 (rs6000_handle_altivec_attribute): Create V1TImode from vector
3566 __int128_t and vector __uint128_t.
3568 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
3569 and mode attributes.
3576 (VS_scalar): Likewise.
3577 (VS_double): Likewise.
3578 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
3580 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
3581 we support the ISA 2.07 128-bit integer arithmetic instructions.
3582 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
3583 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
3584 and TImode types for use with the builtin functions.
3585 (V1TI_type_node): Likewise.
3586 (unsigned_V1TI_type_node): Likewise.
3587 (intTI_type_internal_node): Likewise.
3588 (uintTI_type_internal_node): Likewise.
3590 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
3591 128-bit builtin functions.
3592 (UNSPEC_VADDEUQM): Likewise.
3593 (UNSPEC_VADDECUQ): Likewise.
3594 (UNSPEC_VSUBCUQ): Likewise.
3595 (UNSPEC_VSUBEUQM): Likewise.
3596 (UNSPEC_VSUBECUQ): Likewise.
3597 (VM): Add V1TImode to vector mode iterators.
3599 (VI_unit): Likewise.
3600 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
3601 (altivec_vaddcuq): Likewise.
3602 (altivec_vsubuqm): Likewise.
3603 (altivec_vsubcuq): Likewise.
3604 (altivec_vaddeuqm): Likewise.
3605 (altivec_vaddecuq): Likewise.
3606 (altivec_vsubeuqm): Likewise.
3607 (altivec_vsubecuq): Likewise.
3609 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
3611 (BOOL_128): Likewise.
3612 (BOOL_REGS_OUTPUT): Likewise.
3613 (BOOL_REGS_OP1): Likewise.
3614 (BOOL_REGS_OP2): Likewise.
3615 (BOOL_REGS_UNARY): Likewise.
3616 (BOOL_REGS_AND_CR0): Likewise.
3618 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
3619 128-bit integer builtin support.
3620 (vec_vadduqm): Likewise.
3621 (vec_vaddecuq): Likewise.
3622 (vec_vaddeuqm): Likewise.
3623 (vec_vsubecuq): Likewise.
3624 (vec_vsubeuqm): Likewise.
3625 (vec_vsubcuq): Likewise.
3626 (vec_vsubuqm): Likewise.
3628 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3629 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
3630 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
3631 128-bit integer add/subtract to ISA 2.07.
3633 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
3635 * config/arc/arc.c (arc_predicate_delay_insns):
3636 Fix third argument passed to conditionalize_nonjump.
3638 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
3640 * config/aarch64/aarch64-builtins.c
3641 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
3642 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
3643 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
3644 instead of __builtin_lfloor.
3645 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
3647 2014-03-12 Jakub Jelinek <jakub@redhat.com>
3649 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
3650 (tree_ssa_ifcombine_bb_1): New function.
3651 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
3652 is an empty forwarder block to then_bb or vice versa and then_bb
3653 and else_bb are effectively swapped.
3655 2014-03-12 Christian Bruel <christian.bruel@st.com>
3658 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
3659 REG_CFA_DEF_CFA note.
3660 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
3661 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
3663 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
3665 PR tree-optimization/60454
3666 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
3668 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3670 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
3671 Do not define target_cpu_default2 to generic.
3672 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
3673 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
3674 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
3676 2014-03-12 Jakub Jelinek <jakub@redhat.com>
3677 Marc Glisse <marc.glisse@inria.fr>
3679 PR tree-optimization/60502
3680 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
3681 instead of build_low_bits_mask.
3683 2014-03-12 Jakub Jelinek <jakub@redhat.com>
3686 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
3687 if there are multiple uses, but op doesn't live on E edge.
3688 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
3689 clobber stmts before __builtin_unreachable.
3691 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
3693 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
3694 hard_frame_pointer_rtx.
3695 * cse.c (cse_insn): Remove volatile check.
3696 * cselib.c (cselib_process_insn): Likewise.
3697 * dse.c (scan_insn): Likewise.
3699 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
3701 * config/arc/arc.c (conditionalize_nonjump): New function,
3703 (arc_ifcvt): ... this.
3704 (arc_predicate_delay_insns): Use it.
3706 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
3708 * config/arc/predicates.md (extend_operand): During/after reload,
3709 allow const_int_operand.
3710 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
3711 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
3712 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
3714 (umulsi3_highpart_i): Likewise.
3716 2014-03-11 Richard Biener <rguenther@suse.de>
3718 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
3719 Add asserts to guard possible wrong-code bugs.
3721 2014-03-11 Richard Biener <rguenther@suse.de>
3723 PR tree-optimization/60429
3724 PR tree-optimization/60485
3725 * tree-ssa-structalias.c (set_union_with_increment): Properly
3726 take into account all fields that overlap the shifted vars.
3727 (do_sd_constraint): Likewise.
3728 (do_ds_constraint): Likewise.
3729 (get_constraint_for_ptr_offset): Likewise.
3731 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
3733 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
3734 (nios2_compute_frame_layout):
3735 Add calculation of cfun->machine->fp_save_offset.
3736 (nios2_expand_prologue): Correct setting of frame pointer register
3738 (nios2_expand_epilogue): Update recovery of stack pointer from
3739 frame pointer accordingly.
3740 (nios2_initial_elimination_offset): Update calculation of offset
3741 for eliminating to HARD_FRAME_POINTER_REGNUM.
3743 2014-03-10 Jakub Jelinek <jakub@redhat.com>
3746 * ipa.c (symtab_remove_unreachable_nodes): Don't call
3747 cgraph_get_create_node on VAR_DECLs.
3749 2014-03-10 Richard Biener <rguenther@suse.de>
3752 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
3754 2014-03-08 Douglas B Rupp <rupp@gnat.com>
3756 * config/vms/vms.opt (vms_float_format): New variable.
3758 2014-03-08 Tobias Burnus <burnus@net-b.de>
3760 * doc/invoke.texi (-fcilkplus): Update implementation status.
3762 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
3763 Richard Biener <rguenther@suse.de>
3765 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
3766 consistently accross all TUs.
3767 (run_gcc): Enable -fshort-double automatically at link at link-time
3768 and disallow override.
3770 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
3773 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
3774 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
3775 if they can't be used.
3777 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3779 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
3780 for Solaris 11/x86 ld.
3781 * configure: Regenerate.
3783 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3785 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
3786 (LIB_TLS_SPEC): Save as ld_tls_libs.
3787 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
3788 (HAVE_AS_IX86_TLSLDM): New test.
3789 * configure, config.in: Regenerate.
3790 * config/i386/i386.c (legitimize_tls_address): Fall back to
3791 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
3792 cannot support TLS_MODEL_LOCAL_DYNAMIC.
3793 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
3794 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
3796 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
3798 * common.opt (fira-loop-pressure): Mark as optimization.
3800 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
3802 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
3803 an OpenMP mappable type.
3805 2014-03-06 Matthias Klose <doko@ubuntu.com>
3807 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
3808 MULTILIB_OSDIRNAMES is not defined.
3810 2014-03-06 Jakub Jelinek <jakub@redhat.com>
3811 Meador Inge <meadori@codesourcery.com>
3814 * config/arm/arm.c (arm_tls_symbol_p): Remove.
3815 (arm_legitimize_address): Call legitimize_tls_address for any
3816 arm_tls_referenced_p expression, handle constant addend. Call it
3817 before testing for !TARGET_ARM.
3818 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
3820 2014-03-06 Richard Biener <rguenther@suse.de>
3826 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
3828 * tree-streamer.c (record_common_node): Assert we don't record
3829 nodes with type double.
3830 (preload_common_node): Skip type double, complex double and double
3831 pointer since it is now frontend dependent due to fshort-double option.
3833 2014-03-06 Richard Biener <rguenther@suse.de>
3835 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
3836 or -fno-lto is specified and the linker has full plugin support.
3837 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
3838 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
3839 * lto-wrapper.c (merge_and_complain): Merge compile-time
3840 optimization levels.
3841 (run_gcc): And pass it through to the link options.
3843 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
3847 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
3849 * cselib.c (remove_useless_values): Skip to avoid quadratic
3850 behavior if the condition moved from...
3851 (cselib_process_insn): ... here holds.
3853 2014-03-05 Jakub Jelinek <jakub@redhat.com>
3856 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
3857 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
3860 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
3861 (TM_H): Add x86-tune.def.
3863 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3865 * config/aarch64/aarch64.c (generic_tunings):
3866 Use cortexa57_extra_costs.
3868 2014-03-05 Jakub Jelinek <jakub@redhat.com>
3871 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
3872 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
3873 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
3876 2014-03-04 Heiher <r@hev.cc>
3878 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
3879 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
3881 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
3883 * config/i386/predicates.md (const2356_operand): Change to ...
3884 (const2367_operand): ... this.
3885 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
3887 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
3888 (*avx512pf_scatterpf<mode>sf): Ditto.
3889 (avx512pf_scatterpf<mode>df): Ditto.
3890 (*avx512pf_scatterpf<mode>df_mask): Ditto.
3891 (*avx512pf_scatterpf<mode>df): Ditto.
3892 * config/i386/i386.c (ix86_expand_builtin): Update
3893 incorrect hint operand error message.
3895 2014-03-04 Richard Biener <rguenther@suse.de>
3897 * lto-section-in.c (lto_get_section_data): Fix const cast.
3899 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
3901 * tree-streamer.c (record_common_node): Assert we don't record
3902 nodes with type double.
3903 (preload_common_node): Skip type double, complex double and double
3904 pointer since it is now frontend dependent due to fshort-double option.
3906 2014-03-04 Richard Biener <rguenther@suse.de>
3909 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
3910 (lto_input_toplevel_asms): Likewise.
3911 * lto-section-in.c (lto_get_section_data): Instead do it here
3914 2014-03-04 Richard Biener <rguenther@suse.de>
3916 PR tree-optimization/60382
3917 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
3918 dead PHIs a reduction.
3920 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
3922 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
3924 (_mm_prefetch): Move out of GCC target("sse") pragma.
3925 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
3926 GCC target("prfchw") pragma.
3927 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
3929 * config/i386/i386.c (ix86_option_override_internal): Enable
3930 -mprfchw with -mprefetchwt1.
3932 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
3934 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
3937 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
3939 * opts.h (CL_PCH_IGNORE): Define.
3940 * targhooks.c (option_affects_pch_p):
3941 Return false for options that have CL_PCH_IGNORE set.
3942 * opt-functions.awk: Process PchIgnore.
3943 * doc/options.texi: Document PchIgnore.
3945 * config/arc/arc.opt (misize): Add PchIgnore property.
3947 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3949 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
3950 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
3951 constraint on constants to permit them being loaded into
3952 GENERAL_REGS or BASE_REGS.
3954 2014-03-03 Nick Clifton <nickc@redhat.com>
3956 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
3957 anti-cacnonical alternatives.
3958 (negandhi3_real): New pattern.
3959 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
3961 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
3963 * config/avr/avr-mcus.def: Remove atxmega16x1.
3964 * config/avr/avr-tables.opt: Regenerate.
3965 * config/avr/t-multilib: Regenerate.
3966 * doc/avr-mmcu.texi: Regenerate.
3968 2014-03-03 Tobias Grosser <tobias@grosser.es>
3969 Mircea Namolaru <mircea.namolaru@inria.fr>
3971 PR tree-optimization/58028
3972 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
3975 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3977 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
3978 not handled by recognizers.
3980 2014-03-03 Jakub Jelinek <jakub@redhat.com>
3983 * function.c (expand_function_end): Don't emit
3984 clobber_return_register sequence if clobber_after is a BARRIER.
3985 * cfgexpand.c (construct_exit_block): Append instructions before
3986 return_label to prev_bb.
3988 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3990 * config/rs6000/constraints.md: Document reserved use of "wc".
3992 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
3995 * ipa.c (function_and_variable_visibility): When dissolving comdat
3996 group, also set all symbols to local.
3998 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
4003 2013-12-14 Jan Hubicka <jh@suse.cz>
4005 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
4007 2014-03-02 Jon Beniston <jon@beniston.com>
4013 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
4014 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
4015 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
4016 (simple_return, *simple_return): New patterns
4017 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
4018 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
4020 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
4022 * dwarf2out.c (gen_subprogram_die): Tidy.
4024 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
4027 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
4028 (*mov_t_msb_neg_negc): ... this new insn.
4030 2014-02-28 Jason Merrill <jason@redhat.com>
4033 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
4036 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
4039 * dwarf2out.c (decltype_auto_die): New static.
4040 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
4041 (gen_type_die_with_usage): Handle 'decltype(auto)'.
4042 (is_cxx_auto): Likewise.
4044 2014-02-28 Ian Bolton <ian.bolton@arm.com>
4046 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
4047 we are not using general regs only.
4049 2014-02-28 Richard Biener <rguenther@suse.de>
4052 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
4053 previous fix and only allow to remove trivial pre-headers
4054 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
4055 (remove_forwarder_block): Properly update the latch of a loop.
4057 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4060 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
4061 (cselib_preserved_hash_table): New.
4062 (preserve_constants_and_equivs): Move preserved vals to it.
4063 (cselib_find_slot): Look it up first.
4064 (cselib_init): Initialize it.
4065 (cselib_finish): Release it.
4066 (dump_cselib_table): Dump it.
4068 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4071 * cselib.c (remove_useless_values): Skip to avoid quadratic
4072 behavior if the condition moved from...
4073 (cselib_process_insn): ... here holds.
4075 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4078 * var-tracking.c (vt_initialize): Apply the same condition to
4079 preserve the CFA base value.
4081 2014-02-28 Joey Ye <joey.ye@arm.com>
4084 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
4085 if reload in progress or completed.
4087 2014-02-28 Tobias Burnus <burnus@net-b.de>
4090 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
4093 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
4095 * doc/tm.texi.in (Condition Code Status): Update documention for
4096 relative locations of cc0-setter and cc0-user.
4098 2014-02-27 Jeff Law <law@redhat.com>
4100 PR rtl-optimization/52714
4101 * combine.c (try_combine): When splitting an unrecognized PARALLEL
4102 into two independent simple sets, if I3 is a jump, ensure the
4103 pattern we place into I3 is a (set (pc) ...).
4105 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
4106 Jeff Law <law@redhat.com>
4108 PR rtl-optimization/49847
4109 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
4110 are in different blocks.
4111 * doc/tm.texi (Condition Code Status): Update documention for
4112 relative locations of cc0-setter and cc0-user.
4114 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
4117 * lra.c (lra_emit_add): Check SUBREG too.
4119 2014-02-27 Andreas Schwab <schwab@suse.de>
4121 * config/m68k/m68k.c (m68k_option_override): Disable
4122 -flive-range-shrinkage for classic m68k.
4123 (m68k_override_options_after_change): Likewise.
4125 2014-02-27 Marek Polacek <polacek@redhat.com>
4128 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
4129 -Wmaybe-uninitialized.
4131 2014-02-27 Alan Modra <amodra@gmail.com>
4134 * reload1.c (emit_input_reload_insns): When reload_override_in,
4135 set old to rl->in_reg when rl->in_reg is a subreg.
4137 2014-02-26 Richard Biener <rguenther@suse.de>
4140 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
4142 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4144 * common/config/i386/predicates.md (const1256_operand): Remove.
4145 (const2356_operand): New.
4146 (const_1_to_2_operand): Remove.
4147 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
4148 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
4149 (*avx512pf_gatherpf<mode>sf): Ditto.
4150 (avx512pf_gatherpf<mode>df): Ditto.
4151 (*avx512pf_gatherpf<mode>df_mask): Ditto.
4152 (*avx512pf_gatherpf<mode>df): Ditto.
4153 (avx512pf_scatterpf<mode>sf): Ditto.
4154 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
4155 (*avx512pf_scatterpf<mode>sf): Ditto.
4156 (avx512pf_scatterpf<mode>df): Ditto.
4157 (*avx512pf_scatterpf<mode>df_mask): Ditto.
4158 (*avx512pf_scatterpf<mode>df): Ditto.
4159 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
4161 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
4163 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
4164 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
4165 (_mm512_mask_testn_epi64_mask): Move to ...
4166 * config/i386/avx512cdintrin.h: Here.
4167 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
4168 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
4169 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
4170 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
4171 TARGET_AVX512F from TARGET_AVX512CD.
4173 2014-02-26 Richard Biener <rguenther@suse.de>
4176 * ipa.c (walk_polymorphic_call_targets): Properly guard
4177 call to inline_update_overall_summary.
4179 2014-02-26 Bin Cheng <bin.cheng@arm.com>
4182 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
4183 and latches only if requested. Fix latch if it is removed.
4184 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
4185 LOOPS_HAVE_PREHEADERS.
4187 2014-02-25 Andrew Pinski <apinski@cavium.com>
4189 * builtins.c (expand_builtin_thread_pointer): Create a new target
4190 when the target is NULL.
4192 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
4194 PR rtl-optimization/60317
4195 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4196 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4197 * lra-assigns.c: Include params.h.
4198 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
4199 other reload pseudos considerations.
4201 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4203 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
4204 to use canonical form for nor<mode>3.
4206 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4209 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
4212 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4214 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
4215 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
4216 (ix86_handle_option): Handle OPT_mprefetchwt1.
4217 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
4218 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
4220 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
4221 OPTION_MASK_ISA_PREFETCHWT1.
4222 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
4223 (PTA_PREFETCHWT1): New.
4224 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
4225 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
4226 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
4227 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
4228 (*prefetch_avx512pf_<mode>_: Change into ...
4229 (*prefetch_prefetchwt1_<mode>: This.
4230 * config/i386/i386.opt (mprefetchwt1): New.
4231 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
4232 (_mm_prefetch): Handle intent to write.
4233 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
4235 2014-02-25 Richard Biener <rguenther@suse.de>
4238 * emit-rtl.c (mem_attrs_htab): Remove.
4239 (mem_attrs_htab_hash): Likewise.
4240 (mem_attrs_htab_eq): Likewise.
4241 (set_mem_attrs): Always allocate new mem-attrs when something changed.
4242 (init_emit_once): Do not allocate mem_attrs_htab.
4244 2014-02-25 Richard Biener <rguenther@suse.de>
4247 * lto-opts.c (lto_write_options): Output non-explicit conservative
4248 -fwrapv, -fno-trapv and -fno-strict-overflow.
4249 * lto-wrapper.c (merge_and_complain): Handle merging those options.
4250 (run_gcc): And pass them through.
4252 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4254 * sel-sched.c (calculate_new_fences): New parameter ptime.
4255 Calculate it as a maximum over all fence cycles.
4256 (sel_sched_region_2): Adjust the call to calculate_new_fences.
4257 Print the final schedule timing when sched_verbose.
4259 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4261 PR rtl-optimization/60292
4262 * sel-sched.c (fill_vec_av_set): Do not reset target availability
4263 bit fot the fence instruction.
4265 2014-02-24 Alangi Derick <alangiderick@gmail.com>
4267 * calls.h: Fix typo in comment.
4269 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
4271 * config/pa/pa.c (pa_output_move_double): Don't valididate when
4272 adjusting offsetable addresses.
4274 2014-02-24 Guozhi Wei <carrot@google.com>
4276 * sparseset.h (sparseset_pop): Fix the wrong index.
4278 2014-02-24 Walter Lee <walt@tilera.com>
4280 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
4281 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
4283 * common/config/tilegx/tilegx-common.c
4284 (TARGET_DEFAULT_TARGET_FLAGS): Define.
4285 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
4287 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
4288 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
4289 (tilegx_gimplify_va_arg_expr): Handle big endian.
4290 (tilegx_expand_unaligned_load): Ditto.
4291 (tilegx_expand_unaligned_store): Ditto.
4292 (TARGET_RETURN_IN_MSB): New.
4293 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
4294 (TARGET_ENDIAN_DEFAULT): New.
4295 (TARGET_BIG_ENDIAN): Handle big endian.
4296 (BYTES_BIG_ENDIAN): Ditto.
4297 (WORDS_BIG_ENDIAN): Ditto.
4298 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
4301 * config/tilegx/tilegx.md (extv): Handle big endian.
4303 (insn_st<n>): Ditto.
4304 (insn_st<n>_add<bitsuffix>): Ditto.
4305 (insn_stnt<n>): Ditto.
4306 (insn_stnt<n>_add<bitsuffix>):Ditto.
4307 (vec_interleave_highv8qi): Handle big endian.
4308 (vec_interleave_highv8qi_be): New.
4309 (vec_interleave_highv8qi_le): New.
4310 (insn_v1int_h): Handle big endian.
4311 (vec_interleave_lowv8qi): Handle big endian.
4312 (vec_interleave_lowv8qi_be): New.
4313 (vec_interleave_lowv8qi_le): New.
4314 (insn_v1int_l): Handle big endian.
4315 (vec_interleave_highv4hi): Handle big endian.
4316 (vec_interleave_highv4hi_be): New.
4317 (vec_interleave_highv4hi_le): New.
4318 (insn_v2int_h): Handle big endian.
4319 (vec_interleave_lowv4hi): Handle big endian.
4320 (vec_interleave_lowv4hi_be): New.
4321 (vec_interleave_lowv4hi_le): New.
4322 (insn_v2int_l): Handle big endian.
4323 (vec_interleave_highv2si): Handle big endian.
4324 (vec_interleave_highv2si_be): New.
4325 (vec_interleave_highv2si_le): New.
4326 (insn_v4int_h): Handle big endian.
4327 (vec_interleave_lowv2si): Handle big endian.
4328 (vec_interleave_lowv2si_be): New.
4329 (vec_interleave_lowv2si_le): New.
4330 (insn_v4int_l): Handle big endian.
4331 * config/tilegx/tilegx.opt (mbig-endian): New option.
4332 (mlittle-endian): New option.
4333 * doc/install.texi: Document tilegxbe-linux.
4334 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
4336 2014-02-24 Martin Jambor <mjambor@suse.cz>
4339 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
4340 there are no parameter descriptors.
4342 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
4344 PR rtl-optimization/60268
4345 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
4346 initialization to ...
4347 (sched_rgn_init): ... here.
4348 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
4350 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4352 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
4355 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
4357 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
4360 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4362 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
4363 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
4365 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4367 * config/microblaze/predicates.md: Add cmp_op predicate.
4368 * config/microblaze/microblaze.md: Add branch_compare instruction
4369 which uses cmp_op predicate and emits cmp insn before branch.
4370 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
4371 to microblaze_expand_conditional_branch and consolidate logic.
4372 (microblaze_expand_conditional_branch): emit branch_compare
4373 insn instead of handling cmp op separate from branch insn.
4375 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4377 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
4380 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4382 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
4383 define_insn with define_expand and new define_insn
4384 *altivec_lve<VI_char>x_internal.
4385 (altivec_stve<VI_char>x): Replace define_insn with define_expand
4386 and new define_insn *altivec_stve<VI_char>x_internal.
4387 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
4389 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
4391 (altivec_expand_stvex_be): New function.
4393 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
4395 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
4396 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
4397 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
4398 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
4400 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
4403 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
4404 instead of emit_move_insn.
4406 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4408 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
4410 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
4411 gen_altivec_vsumsws.
4413 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4415 * config/rs6000/altivec.md (altivec_lvxl): Rename as
4416 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
4417 (altivec_lvxl_<mode>): New define_expand incorporating
4418 -maltivec=be semantics where needed.
4419 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
4420 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
4421 semantics where needed.
4422 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
4423 (altivec_stvx_<mode>): New define_expand incorporating
4424 -maltivec=be semantics where needed.
4425 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
4426 VM2 iterator instead of V4SI.
4427 (altivec_stvxl_<mode>): New define_expand incorporating
4428 -maltivec=be semantics where needed.
4429 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
4430 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
4431 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
4432 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
4433 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
4434 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
4435 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
4436 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
4437 ALTIVEC_BUILTIN_STVXL.
4438 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
4439 (altivec_expand_stvx_be): Likewise.
4440 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
4441 (altivec_expand_lvx_be): Likewise.
4442 (altivec_expand_stvx_be): Likewise.
4443 (altivec_expand_builtin): Add cases for
4444 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
4445 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
4446 (altivec_init_builtins): Add definitions for
4447 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
4448 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
4450 2014-02-21 Catherine Moore <clm@codesourcery.com>
4452 * doc/invoke.texi (mvirt, mno-virt): Document.
4453 * config/mips/mips.opt (mvirt): New option.
4454 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
4456 2014-02-21 Richard Biener <rguenther@suse.de>
4458 PR tree-optimization/60276
4459 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
4460 (STMT_VINFO_MIN_NEG_DIST): New macro.
4461 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
4462 STMT_VINFO_MIN_NEG_DIST.
4463 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
4464 made for negative dependence distances still hold.
4466 2014-02-21 Richard Biener <rguenther@suse.de>
4469 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
4470 DECL_INITIAL for globals not in the current function context.
4472 2014-02-21 Jakub Jelinek <jakub@redhat.com>
4474 PR tree-optimization/56490
4475 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
4476 * tree-ssa-uninit.c: Include params.h.
4477 (compute_control_dep_chain): Add num_calls argument, return false
4478 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
4479 num_calls to recursive call.
4480 (find_predicates): Change dep_chain into normal array,
4481 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
4482 variable and adjust compute_control_dep_chain caller.
4483 (find_def_preds): Likewise.
4485 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
4487 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
4488 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
4490 2014-02-21 Nick Clifton <nickc@redhat.com>
4492 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
4493 (pushhi1): Likewise.
4494 (popqi1): Add mode to pre_dec.
4497 2014-02-21 Jakub Jelinek <jakub@redhat.com>
4499 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
4500 mode for mask of V8SFmode permutation.
4502 2014-02-20 Richard Henderson <rth@redhat.com>
4505 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
4506 a new pseudo for OLDVAL.
4508 2014-02-20 Jakub Jelinek <jakub@redhat.com>
4511 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
4512 gen_reg_rtx if d->testing_p.
4513 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
4514 if d->testing_p and we will certainly return true.
4515 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
4518 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
4520 * emit-rtl.c (gen_reg_rtx): Assert that
4521 crtl->emit.regno_pointer_align_length is non-zero.
4523 2014-02-20 Richard Henderson <rth@redhat.com>
4526 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
4527 on failure the store back into EXPECT.
4529 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
4530 Sandra Loosemore <sandra@codesourcery.com>
4532 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
4533 * config/nios2/nios2.c (nios2_function_profiler): Add
4534 -fPIC (flag_pic == 2) support.
4535 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
4536 (nios2_large_offset_p): New function.
4537 (nios2_unspec_reloc_p): Move up position, update to use
4538 nios2_large_offset_p.
4539 (nios2_unspec_address): Remove function.
4540 (nios2_unspec_offset): New function.
4541 (nios2_large_got_address): New function.
4542 (nios2_got_address): Add large offset support.
4543 (nios2_legitimize_tls_address): Update usage of removed and new
4545 (nios2_symbol_binds_local_p): New function.
4546 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
4547 (nios2_legitimize_address): Update to use nios2_large_offset_p.
4548 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
4549 (nios2_print_operand): Merge H/L processing, add hiadj/lo
4550 processing for (const (unspec ...)).
4551 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
4553 2014-02-20 Richard Biener <rguenther@suse.de>
4555 * tree-cfg.c (replace_uses_by): Mark altered BBs before
4556 doing the substitution.
4557 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
4559 2014-02-20 Martin Jambor <mjambor@suse.cz>
4562 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
4563 info when checking whether lattices are bottom.
4565 2014-02-20 Richard Biener <rguenther@suse.de>
4568 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
4571 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
4574 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
4575 parameter specifying the scaling.
4576 (inline_call): Update.
4577 (want_inline_recursively): Guard division by zero.
4578 (recursive_inlining): Update.
4579 * ipa-inline.h (clone_inlined_nodes): Update.
4581 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
4584 * config/i386/i386.c (classify_argument): Pass structures of size
4585 64 bytes or less in register.
4587 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
4588 Kirill Yukhin <kirill.yukhin@intel.com>
4590 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
4591 (_mm_rcp28_round_ss): Ditto.
4592 (_mm_rsqrt28_round_sd): Ditto.
4593 (_mm_rsqrt28_round_ss): Ditto.
4594 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
4595 (_mm_rcp14_round_ss): Ditto.
4596 (_mm_rsqrt14_round_sd): Ditto.
4597 (_mm_rsqrt14_round_ss): Ditto.
4598 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
4599 the first input operand, get rid of match_dup.
4600 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
4602 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
4604 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
4605 operand as the first input operand, set type attribute.
4606 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
4608 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
4609 operand as the first input operand, set type attribute.
4611 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4613 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
4616 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
4619 * config/i386/i386.c (construct_container): Remove TFmode check
4620 for X86_64_INTEGER_CLASS.
4622 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
4625 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
4626 only when -Wpsabi is enabled.
4628 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
4631 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
4632 passing arrays in registers are the same as for structs, so remove the
4633 special case for them.
4635 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
4637 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
4638 destination type, extract only the valid bits if the source type is not
4639 integral and has a different mode.
4641 2014-02-19 Richard Biener <rguenther@suse.de>
4644 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
4647 2014-02-19 Richard Biener <rguenther@suse.de>
4650 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
4651 (ipa_modify_call_arguments): Emit an argument load explicitely and
4652 preserve virtual SSA form there and for the replacement call.
4653 Do not update SSA form nor free dominance info.
4655 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
4657 * ipa.c (function_and_variable_visibility): Also clear WEAK
4658 flag when disolving COMDAT_GROUP.
4660 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
4662 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
4663 * ipa-prop.c (ipa_set_jf_known_type): Return early when
4665 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
4666 do more sanity checks.
4667 (detect_type_change): Return true when giving up early.
4668 (compute_complex_assign_jump_func): Fix type parameter of
4669 ipa_set_ancestor_jf.
4670 (compute_complex_ancestor_jump_func): Likewise.
4671 (update_jump_functions_after_inlining): Fix updating of
4673 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
4675 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
4677 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
4678 inline clones when edge disappears.
4680 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
4683 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
4684 Split 64-bit moves into 2 patterns. Do not allow the use of
4685 direct move for TDmode in little endian, since the decimal value
4686 has little endian bytes within a word, but the 64-bit pieces are
4687 ordered in a big endian fashion, and normal subreg's of TDmode are
4689 (mov<mode>_64bit_dm): Likewise.
4690 (movtd_64bit_nodm): Likewise.
4692 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
4694 PR tree-optimization/60174
4695 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
4696 statement of an SSA_NAME that occurs in an abnormal PHI node.
4698 2014-02-18 Jakub Jelinek <jakub@redhat.com>
4701 * final.c (SEEN_BB): Remove.
4702 (SEEN_NOTE, SEEN_EMITTED): Renumber.
4703 (final_scan_insn): Don't force_source_line on second
4704 NOTE_INSN_BASIC_BLOCK.
4706 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
4709 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
4710 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
4711 (type_natural_mode): Warn ABI change when %zmm register is not
4712 available for AVX512F vector value passing.
4714 2014-02-18 Kai Tietz <ktietz@redhat.com>
4717 * config/i386/i386.c (ix86_expand_prologue): Use value in
4718 rax register as displacement when restoring %r10 or %rax.
4719 Fix wrong offset when restoring both registers.
4721 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
4723 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
4724 assertion with conditional return.
4726 2014-02-18 Jakub Jelinek <jakub@redhat.com>
4727 Uros Bizjak <ubizjak@gmail.com>
4730 * config/i386/driver-i386.c (host_detect_local_cpu): If
4731 YMM state is not saved by the OS, also clear has_f16c. Move
4732 CPUID 0x80000001 handling before YMM state saving checking.
4734 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
4736 PR rtl-optimization/58960
4737 * haifa-sched.c (alloc_global_sched_pressure_data): New,
4738 factored out from ...
4739 (sched_init): ... here.
4740 (free_global_sched_pressure_data): New, factored out from ...
4741 (sched_finish): ... here.
4742 * sched-int.h (free_global_sched_pressure_data): Declare.
4743 * sched-rgn.c (nr_regions_initial): New static global.
4744 (haifa_find_rgns): Initialize it.
4745 (schedule_region): Disable sched-pressure for the newly
4748 2014-02-17 Richard Biener <rguenther@suse.de>
4750 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
4751 release SSA defs of pattern stmts.
4753 2014-02-17 Richard Biener <rguenther@suse.de>
4755 * tree-inline.c (expand_call_inline): Release the virtual
4756 operand defined by the call we are about to inline.
4758 2014-02-17 Richard Biener <rguenther@suse.de>
4760 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
4762 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
4763 Ilya Tocar <ilya.tocar@intel.com>
4765 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
4766 arguments order in builtin.
4767 (_mm512_permutexvar_epi64): Ditto.
4768 (_mm512_mask_permutexvar_epi64): Ditto
4769 (_mm512_maskz_permutexvar_epi32): Ditto
4770 (_mm512_permutexvar_epi32): Ditto
4771 (_mm512_mask_permutexvar_epi32): Ditto
4773 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4775 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
4776 (p8_vmrgow): Likewise.
4778 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4780 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
4783 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
4786 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
4787 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
4788 into 64-bit and 32-bit moves. On 64-bit moves, add support for
4789 using direct move instructions on ISA 2.07. Also adjust
4790 instruction length for 64-bit.
4791 (mov<mode>_64bit, TFmode/TDmode): Likewise.
4792 (mov<mode>_32bit, TFmode/TDmode): Likewise.
4794 2014-02-15 Alan Modra <amodra@gmail.com>
4798 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
4799 find_replacement on parts of insn rtl that might be reloaded.
4801 2014-02-15 Richard Biener <rguenther@suse.de>
4803 PR tree-optimization/60183
4804 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
4805 (tree_ssa_phiprop): Calculate and free post-dominators.
4807 2014-02-14 Jeff Law <law@redhat.com>
4809 PR rtl-optimization/60131
4810 * ree.c (get_extended_src_reg): New function.
4811 (combine_reaching_defs): Use it rather than assuming location of REG.
4812 (find_and_remove_re): Verify first operand of extension is
4813 a REG before adding the insns to the copy list.
4815 2014-02-14 Roland McGrath <mcgrathr@google.com>
4817 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
4818 * configure: Regenerated.
4819 * config.in: Regenerated.
4820 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
4821 instead of ASM_SHORT.
4823 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
4824 Richard Earnshaw <rearnsha@arm.com>
4826 PR rtl-optimization/59535
4827 * lra-constraints.c (process_alt_operands): Encourage alternative
4828 when unassigned pseudo class is superset of the alternative class.
4829 (inherit_reload_reg): Don't inherit when optimizing for code size.
4830 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
4831 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
4832 modes not less than 4 for Thumb1.
4834 2014-02-14 Kyle McMartin <kyle@redhat.com>
4837 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
4839 2014-02-14 Richard Biener <rguenther@suse.de>
4841 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
4842 (get_frame_arg): Drop the assert with langhook types_compatible_p.
4843 Do not strip INDIRECT_REFs.
4845 2014-02-14 Richard Biener <rguenther@suse.de>
4848 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
4849 DECL_FUNCTION_SPECIFIC_TARGET.
4850 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
4851 * tree-streamer-out.c (pack_ts_target_option): Remove.
4852 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
4853 (write_ts_function_decl_tree_pointers): Do not stream
4854 DECL_FUNCTION_SPECIFIC_TARGET.
4855 * tree-streamer-in.c (unpack_ts_target_option): Remove.
4856 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
4857 (lto_input_ts_function_decl_tree_pointers): Do not stream
4858 DECL_FUNCTION_SPECIFIC_TARGET.
4860 2014-02-14 Jakub Jelinek <jakub@redhat.com>
4862 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
4863 (get_initial_def_for_induction, vectorizable_induction): Ignore
4864 debug stmts when looking for exit_phi.
4865 (vectorizable_live_operation): Fix up condition.
4867 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
4869 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
4870 nreverse() because it changes the content of original tree list.
4872 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
4874 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
4875 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
4877 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
4879 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
4880 GNU coding standards.
4882 2014-02-13 Jakub Jelinek <jakub@redhat.com>
4885 * dwarf2out.c (gen_subprogram_die): Don't call
4886 add_calling_convention_attribute if subr_die is old_die.
4888 2014-02-13 Sharad Singhai <singhai@google.com>
4890 * doc/optinfo.texi: Fix order of nodes.
4892 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
4894 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
4895 operands[2], not operands[3].
4897 2014-02-13 Richard Biener <rguenther@suse.de>
4900 * doc/install.texi (ISL): Update recommended version to 0.12.2,
4901 mention the possibility of an in-tree build.
4902 (CLooG): Update recommended version to 0.18.1, mention the
4903 possibility of an in-tree build and clarify that the ISL
4904 bundled with CLooG does not work.
4906 2014-02-13 Jakub Jelinek <jakub@redhat.com>
4909 * expr.c (compress_float_constant): If x is a hard register,
4910 extend into a pseudo and then move to x.
4912 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
4914 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
4915 caused by bad second argument to warning_at() with -mhotpatch and
4916 nested functions (e.g. with gfortran).
4918 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
4920 * opts.c (option_name): Remove "enabled by default" rider.
4922 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
4924 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
4926 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
4927 Uros Bizjak <ubizjak@gmail.com>
4930 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
4931 * configure: Regenerated.
4933 2014-02-12 Richard Biener <rguenther@suse.de>
4935 * vec.c (vec_prefix::calculate_allocation): Move as
4936 inline variant to vec.h.
4937 (vec_prefix::calculate_allocation_1): New out-of-line version.
4938 * vec.h (vec_prefix::calculate_allocation_1): Declare.
4939 (vec_prefix::m_has_auto_buf): Rename to ...
4940 (vec_prefix::m_using_auto_storage): ... this.
4941 (vec_prefix::calculate_allocation): Inline the easy cases
4942 and dispatch to calculate_allocation_1 which doesn't need the
4944 (va_heap::reserve): Use gcc_checking_assert.
4945 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
4946 m_using_auto_storage.
4947 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
4949 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
4950 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
4951 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
4953 2014-02-12 Richard Biener <rguenther@suse.de>
4955 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
4956 when we found a dependence.
4958 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
4960 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
4962 (maybe_fold_stmt): ... into this new function.
4963 * omp-low.c (lower_omp): Update comment.
4965 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
4968 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
4971 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
4973 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
4974 identifiers in comments.
4975 (cortexa53_extra_costs): Likewise.
4976 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
4977 (cortexa7_extra_costs): Likewise.
4978 (cortexa12_extra_costs): Likewise.
4979 (cortexa15_extra_costs): Likewise.
4980 (v7m_extra_costs): Likewise.
4982 2014-02-12 Richard Biener <rguenther@suse.de>
4985 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
4986 of posix_memalign being successful.
4987 (lower_stmt): Restrict lowering of posix_memalign to when
4988 -ftree-bit-ccp is enabled.
4990 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
4992 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
4994 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
4996 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
4998 PR rtl-optimization/60116
4999 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
5000 other_insn once the combination has been validated.
5002 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
5005 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
5007 * ipa-devirt.c: Include demangle.h
5008 (odr_violation_reported): New static variable.
5009 (add_type_duplicate): Update odr_violations.
5010 (maybe_record_node): Add completep parameter; update it.
5011 (record_target_from_binfo): Add COMPLETEP parameter;
5012 update it as needed.
5013 (possible_polymorphic_call_targets_1): Likewise.
5014 (struct polymorphic_call_target_d): Add nonconstruction_targets;
5015 rename FINAL to COMPLETE.
5016 (record_targets_from_bases): Sanity check we found the binfo;
5017 fix COMPLETEP updating.
5018 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
5019 parameter, fix computing of COMPLETEP.
5020 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
5021 at LTO time do demangling.
5022 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
5023 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
5025 (gimple_get_virt_method_for_binfo): Likewise.
5026 * gimple-fold.h (gimple_get_virt_method_for_binfo,
5027 gimple_get_virt_method_for_vtable): Update prototypes.
5029 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
5032 * genautomata.c (add_presence_absence): Fix typo with
5033 {final_}presence_list.
5035 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
5038 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
5039 for VSX/Altivec vectors that land in GPR registers.
5041 2014-02-11 Richard Henderson <rth@redhat.com>
5042 Jakub Jelinek <jakub@redhat.com>
5045 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
5046 around drhs if type conversion to lacc->type is not useless.
5048 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5050 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
5052 (cortex-a57.cortex-a53): Likewise.
5053 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
5055 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5057 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
5060 2014-02-11 Renlin Li <Renlin.Li@arm.com>
5062 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
5063 add_options_for_arm_vfp3.
5065 2014-02-11 Jeff Law <law@redhat.com>
5068 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
5069 object with an undesirable mode.
5071 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5074 * config/i386/sol2-9.h: New file.
5075 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
5076 *-*-solaris2.9*): Use it.
5078 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5080 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
5081 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
5083 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5085 * config/microblaze/microblaze.c: Extend mcpu version format
5087 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
5089 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
5091 2014-02-10 Richard Henderson <rth@redhat.com>
5094 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
5095 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
5096 ms-abi vs -mno-accumulate-outgoing-args.
5097 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
5098 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
5101 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
5104 * cfgexpand.c (expand_asm_operands): Attach source location to
5105 ASM_INPUT rtx objects.
5106 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
5108 2014-02-10 Nick Clifton <nickc@redhat.com>
5110 * config/mn10300/mn10300.c (popcount): New function.
5111 (mn10300_expand_prologue): Include saved registers in stack usage
5114 2014-02-10 Jeff Law <law@redhat.com>
5117 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
5118 when changing the SET_DEST of a prior insn to avoid an input reload.
5120 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5122 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
5123 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
5124 -mcall-openbsd, or -mcall-linux.
5125 (CC1_ENDIAN_BIG_SPEC): Remove.
5126 (CC1_ENDIAN_LITTLE_SPEC): Remove.
5127 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5128 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
5129 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
5130 and %cc1_endian_default.
5131 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5133 2014-02-10 Richard Biener <rguenther@suse.de>
5135 PR tree-optimization/60115
5136 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
5137 MEM_REF handling. Properly verify that the accesses are not
5138 out of the objects bound.
5140 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5142 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
5145 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
5147 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
5148 proper constants and fix formatting.
5149 (possible_polymorphic_call_targets): Fix formatting.
5151 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
5152 Ilya Tocar <ilya.tocar@intel.com>
5154 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
5155 (_mm512_loadu_epi32): Renamed into...
5156 (_mm512_loadu_si512): This.
5157 (_mm512_storeu_epi32): Renamed into...
5158 (_mm512_storeu_si512): This.
5159 (_mm512_maskz_ceil_ps): Removed.
5160 (_mm512_maskz_ceil_pd): Ditto.
5161 (_mm512_maskz_floor_ps): Ditto.
5162 (_mm512_maskz_floor_pd): Ditto.
5163 (_mm512_floor_round_ps): Ditto.
5164 (_mm512_floor_round_pd): Ditto.
5165 (_mm512_ceil_round_ps): Ditto.
5166 (_mm512_ceil_round_pd): Ditto.
5167 (_mm512_mask_floor_round_ps): Ditto.
5168 (_mm512_mask_floor_round_pd): Ditto.
5169 (_mm512_mask_ceil_round_ps): Ditto.
5170 (_mm512_mask_ceil_round_pd): Ditto.
5171 (_mm512_maskz_floor_round_ps): Ditto.
5172 (_mm512_maskz_floor_round_pd): Ditto.
5173 (_mm512_maskz_ceil_round_ps): Ditto.
5174 (_mm512_maskz_ceil_round_pd): Ditto.
5175 (_mm512_expand_pd): Ditto.
5176 (_mm512_expand_ps): Ditto.
5177 * config/i386/i386.c (ix86_builtins): Remove
5178 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
5179 (bdesc_args): Ditto.
5180 * config/i386/predicates.md (const1256_operand): New.
5181 (const_1_to_2_operand): Ditto.
5182 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
5183 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
5184 (*avx512pf_gatherpf<mode>sf): Ditto.
5185 (avx512pf_gatherpf<mode>df): Ditto.
5186 (*avx512pf_gatherpf<mode>df_mask): Ditto.
5187 (*avx512pf_gatherpf<mode>df): Ditto.
5188 (avx512pf_scatterpf<mode>sf): Ditto.
5189 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5190 (*avx512pf_scatterpf<mode>sf): Ditto.
5191 (avx512pf_scatterpf<mode>df): Ditto.
5192 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5193 (*avx512pf_scatterpf<mode>df): Ditto.
5194 (avx512f_expand<mode>): Removed.
5195 (<shift_insn><mode>3<mask_name>): Change predicate type.
5197 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5199 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
5200 not at the end of datarefs vector use ordered_remove to avoid
5201 reordering datarefs vector.
5204 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
5205 mark local addressable non-static vars as GOVD_PRIVATE
5206 instead of GOVD_LOCAL.
5207 * omp-low.c (lower_omp_for): Move gimple_bind_vars
5208 and BLOCK_VARS of gimple_bind_block to new_stmt rather
5212 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
5213 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
5214 assume_aligned or alloc_align attributes.
5215 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
5216 arguments. Handle also assume_aligned and alloc_align attributes.
5217 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
5218 calls to functions with assume_aligned or alloc_align attributes.
5219 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
5221 2014-02-08 Terry Guo <terry.guo@arm.com>
5223 * doc/invoke.texi: Document ARM -march=armv7e-m.
5225 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5227 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
5228 flag on __cilkrts_rethrow builtin.
5231 * ipa-cp.c (determine_versionability): Fail at -O0
5232 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
5233 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
5236 2014-02-04 Jakub Jelinek <jakub@redhat.com>
5239 * tree-inline.c (copy_forbidden): Fail for
5240 __attribute__((optimize (0))) functions.
5242 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5244 * varpool.c: Include pointer-set.h.
5245 (varpool_remove_unreferenced_decls): Variables in other partitions
5246 will not be output; be however careful to not lose information
5249 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5251 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
5252 lookup in the vtable constructor.
5254 2014-02-07 Jeff Law <law@redhat.com>
5257 * config/m68k/m68k.md (ashldi_extsi): Turn into a
5258 define_insn_and_split.
5260 * ipa-inline.c (inline_small_functions): Fix typos.
5262 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5264 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
5265 (s390_can_use_return_insn): Declare.
5266 * config/s390/s390.h (EPILOGUE_USES): Define.
5267 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
5269 (s390_chunkify_start): Handle return JUMP_LABELs.
5270 (s390_early_mach): Emit a main_pool instruction on the entry edge.
5271 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
5272 (s390_can_use_return_insn): New functions.
5273 (s390_fix_long_loop_prediction): Handle conditional returns.
5274 (TARGET_SET_UP_BY_PROLOGUE): Define.
5275 * config/s390/s390.md (ANY_RETURN): New code iterator.
5276 (*creturn, *csimple_return, return, simple_return): New patterns.
5278 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5280 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
5281 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
5282 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
5283 REG_CFA_RESTORE list when deciding not to restore a register.
5285 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5287 * config/s390/s390.c: Include tree-pass.h and context.h.
5288 (s390_early_mach): New function, split out from...
5289 (s390_emit_prologue): ...here.
5290 (pass_data_s390_early_mach): New pass structure.
5291 (pass_s390_early_mach): New class.
5292 (s390_option_override): Create and register early_mach pass.
5293 Move to end of file.
5295 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5297 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
5298 to match for the exit block.
5300 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5302 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
5303 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
5304 Reject misaligned operands.
5306 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5308 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
5310 2014-02-07 Richard Biener <rguenther@suse.de>
5313 * gimple-low.c (lower_builtin_posix_memalign): New function.
5314 (lower_stmt): Call it to lower posix_memalign in a way
5315 to make alignment info accessible.
5317 2014-02-07 Jakub Jelinek <jakub@redhat.com>
5320 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
5321 __builtin_setjmp_receiver.
5323 2014-02-07 Richard Biener <rguenther@suse.de>
5326 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
5327 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
5328 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
5329 Handle BUILT_IN_POSIX_MEMALIGN.
5330 (find_func_clobbers): Likewise.
5331 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
5332 (call_may_clobber_ref_p_1): Likewise.
5334 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5337 * ipa-devirt.c (record_target_from_binfo): Remove overactive
5340 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5343 * lto-cgraph.c (lto_output_node): Use
5344 symtab_get_symbol_partitioning_class.
5345 (lto_output_varpool_node): likewise.
5346 (symtab_get_symbol_partitioning_class): Move here from
5348 * cgraph.h (symbol_partitioning_class): Likewise.
5349 (symtab_get_symbol_partitioning_class): Declare.
5351 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5353 * ggc.h (ggc_internal_cleared_alloc): New macro.
5354 * vec.h (vec_safe_copy): Handle memory stats.
5355 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
5356 * target-globals.c (save_target_globals): Likewise.
5358 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5361 * expr.c (emit_move_resolve_push): Export; be bit more selective
5362 on when to clear alias set.
5363 * expr.h (emit_move_resolve_push): Declare.
5364 * function.h (struct function): Add tail_call_marked.
5365 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
5366 * config/i386/i386-protos.h (ix86_expand_push): Remove.
5367 * config/i386/i386.md (TImode move expander): De not call
5369 (FP push expanders): Preserve memory attributes.
5370 * config/i386/sse.md (push<mode>1): Remove.
5371 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
5372 (ix86_expand_push): Remove.
5373 * config/i386/mmx.md (push<mode>1): Remove.
5375 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5377 PR rtl-optimization/60030
5378 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
5379 lopart with paradoxical subreg before shifting it up by hprec.
5381 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5383 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
5384 Remove extra newline at end of file.
5385 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
5386 (arm_issue_rate): Handle cortexa57.
5387 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
5388 (cortex-a57.cortex-a53): Likewise.
5390 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5393 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
5394 don't record in REG_FRAME_RELATED_EXPR registers not set in that
5396 (arm_expand_prologue): Adjust all callers.
5397 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
5398 info, registers also at the lowest numbered registers side. Use
5399 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
5403 * var-tracking.c (adjust_mems): Before adding a SET to
5404 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
5406 2014-02-06 Alan Modra <amodra@gmail.com>
5409 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
5410 change SDmode to DDmode when lra_in_progress.
5412 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5415 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
5416 free_data_ref on the dr first, and before goto again also set dr
5417 to the next dr. For simd_lane_access, free old datarefs[i] before
5418 overwriting it. For get_vectype_for_scalar_type failure, don't
5419 free_data_ref if simd_lane_access.
5421 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
5424 * tree.h (opts_for_fn): New inline function.
5425 (opt_for_fn): Define.
5426 * config/i386/i386.c (ix86_function_regparm): Use
5427 opt_for_fn (decl, optimize) instead of optimize.
5429 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
5431 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
5432 for SYMBOL_REF in large memory model.
5434 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5436 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
5438 (cortex-a57): Likewise.
5439 (cortex-a57.cortex-a53): Likewise.
5441 2014-02-06 Yury Gribov <y.gribov@samsung.com>
5442 Kugan Vivekanandarajah <kuganv@linaro.org>
5444 * config/arm/arm.c (arm_vector_alignment_reachable): Check
5446 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
5448 2014-02-06 Richard Biener <rguenther@suse.de>
5450 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
5451 set_loop_copy and initialize_original_copy_tables.
5453 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
5455 * config/aarch64/aarch64-simd.md
5456 (aarch64_ashr_simddi): Change QI to SI.
5458 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
5459 Jakub Jelinek <jakub@redhat.com>
5462 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
5465 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5467 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
5468 CODE_FOR_altivec_vpku[hw]um to
5469 CODE_FOR_altivec_vpku[hw]um_direct.
5470 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
5471 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
5472 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
5473 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
5475 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5477 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
5478 generation for -maltivec=be.
5479 (altivec_vsumsws): Simplify redundant test.
5481 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5483 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
5484 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
5485 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
5486 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
5487 gen_altivec_vpkuwum.
5488 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
5490 (altivec_vpks<VI_char>ss): Likewise.
5491 (altivec_vpks<VI_char>us): Likewise.
5492 (altivec_vpku<VI_char>us): Likewise.
5493 (altivec_vpku<VI_char>um): Likewise.
5494 (altivec_vpku<VI_char>um_direct): New (copy of
5495 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
5497 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
5498 target is little endian and -maltivec=be is not specified.
5499 (*altivec_vupkhs<VU_char>_direct): New (copy of
5500 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
5501 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
5502 target is little endian and -maltivec=be is not specified.
5503 (*altivec_vupkls<VU_char>_direct): New (copy of
5504 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
5505 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
5506 little endian and -maltivec=be is not specified.
5507 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
5508 little endian and -maltivec=be is not specified.
5510 2014-02-05 Richard Henderson <rth@redhat.com>
5513 * combine-stack-adj.c: Revert r206943.
5514 * sched-int.h (struct deps_desc): Add last_args_size.
5515 * sched-deps.c (init_deps): Initialize it.
5516 (sched_analyze_insn): Add OUTPUT dependencies between insns that
5517 contain REG_ARGS_SIZE notes.
5519 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
5521 * lto-cgraph.c (asm_nodes_output): Make global.
5522 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
5523 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
5524 (driver_handle_option): Handle OPT_fwpa.
5526 2014-02-05 Jakub Jelinek <jakub@redhat.com>
5529 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
5530 a comment typo and formatting issue. If odr_hash hasn't been
5531 created, return vNULL and set *completep to false.
5534 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
5535 bb with no successors.
5537 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
5540 * doc/invoke.texi (-march): Clarify documentation for ARM.
5544 2014-02-05 Richard Biener <rguenther@suse.de>
5546 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
5547 when not vectorizing because of too many alias checks.
5548 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
5549 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
5551 2014-02-05 Nick Clifton <nickc@redhat.com>
5553 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
5554 accept extended registers in any mode when compiling for the MN10300.
5556 2014-02-05 Yury Gribov <y.gribov@samsung.com>
5558 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
5559 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
5560 sanitization attributes.
5561 (can_inline_edge_p): Likewise.
5562 (sanitize_attrs_match_for_inline_p): New function.
5564 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5566 * ipa-prop.c (detect_type_change): Shor circuit testing of
5567 type changes on THIS pointer.
5569 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
5572 * config/pa/pa.c (legitimize_tls_address): Return original address
5573 if not passed a SYMBOL_REF rtx.
5574 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
5576 (pa_emit_move_sequence): Simplify TLS source operands.
5577 (pa_legitimate_constant_p): Reject all TLS constants.
5578 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
5579 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
5581 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5583 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
5584 groups when we know they are controlled by LTO.
5585 * varasm.c (default_binds_local_p_1): If object is in other partition,
5586 it will be resolved locally.
5588 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
5590 * config/host-linux.c (linux_gt_pch_use_address): Don't
5591 use SSIZE_MAX because it is not always defined.
5593 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
5596 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
5597 threshold for pseudo splitting.
5598 (update_ebb_live_info): Process call argument hard registers and
5599 hard registers from insn definition too.
5600 (max_small_class_regs_num): New constant.
5601 (inherit_in_ebb): Update live hard regs through EBBs. Update
5602 reloads_num only for small register classes. Don't split for
5605 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
5608 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
5611 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5613 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
5616 2014-02-04 Marek Polacek <polacek@redhat.com>
5618 * gdbinit.in (pel): Define.
5620 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
5622 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
5625 2014-02-04 Richard Biener <rguenther@suse.de>
5628 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
5629 in function context local.
5630 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
5631 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
5632 similar to LTO_imported_decl_ref.
5634 2014-02-04 Jakub Jelinek <jakub@redhat.com>
5636 PR tree-optimization/60002
5637 * cgraphclones.c (build_function_decl_skip_args): Clear
5640 PR tree-optimization/60023
5641 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
5642 false to gsi_replace.
5643 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
5644 has been in some EH region and vec_stmt could throw, add
5645 vec_stmt into the same EH region.
5646 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
5647 has no lhs, ignore it.
5648 * internal-fn.c (expand_MASK_LOAD): Likewise.
5651 * tree-inline.c (copy_forbidden): Fail for
5652 __attribute__((optimize (0))) functions.
5655 * omp-low.c (simd_clone_struct_copy): If from->inbranch
5656 is set, copy one less argument.
5657 (expand_simd_clones): Don't subtract clone_info->inbranch
5658 from simd_clone_struct_alloc argument.
5660 PR rtl-optimization/57915
5661 * recog.c (simplify_while_replacing): If all unary/binary/relational
5662 operation arguments are constant, attempt to simplify those.
5665 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
5666 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
5668 2014-02-04 Richard Biener <rguenther@suse.de>
5670 PR tree-optimization/60012
5671 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
5672 TBAA disambiguation to all DDRs.
5674 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5677 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
5678 (LINK_SPEC): Use it for -shared, -shared-libgcc.
5680 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5683 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
5685 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5687 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
5688 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
5690 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5693 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
5694 to figure out targets of polymorphic calls with known decl.
5695 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
5696 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
5697 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
5698 (get_polymorphic_call_info): ... here.
5699 (get_polymorphic_call_info_from_invariant): New function.
5701 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5703 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
5704 lookup via vtable pointer; check for type consistency
5705 and turn inconsitent facts into UNREACHABLE.
5706 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
5707 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
5708 type inconsistent querries; return UNREACHABLE instead.
5710 2014-02-03 Richard Henderson <rth@twiddle.net>
5713 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
5714 already processed this node.
5715 (normalize_one_pred_1): Pass along mark_set.
5716 (normalize_one_pred): Create and destroy a pointer_set_t.
5717 (normalize_one_pred_chain): Likewise.
5719 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
5721 PR gcov-profile/58602
5722 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
5724 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5727 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
5728 -fno-devirtualize; try to devirtualize by the knowledge of
5729 virtual table pointer given by aggregate propagation.
5730 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
5731 (ipa_print_node_jump_functions): Dump also offset that
5732 is relevant for polymorphic calls.
5733 (determine_known_aggregate_parts): Add arg_type parameter; use it
5734 instead of determining the type from pointer type.
5735 (ipa_compute_jump_functions_for_edge): Update call of
5736 determine_known_aggregate_parts.
5737 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
5738 (gimple_get_virt_method_for_binfo): ... here; simplify using
5739 vtable_pointer_value_to_vtable.
5740 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
5741 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
5742 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
5743 (vtable_pointer_value_to_vtable): Break out from ...; handle also
5745 (vtable_pointer_value_to_binfo): ... here.
5746 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
5748 2014-02-03 Teresa Johnson <tejohnson@google.com>
5750 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
5751 redef of outer loop index variable.
5753 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
5757 * doc/extend.texi (Function Attributes): Typo.
5759 2014-02-03 Cong Hou <congh@google.com>
5761 PR tree-optimization/60000
5762 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
5763 if the vectorized statement is a store. A store statement can only
5764 appear at the end of pattern statements.
5766 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
5768 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
5769 (ix86_option_override_internal): Default long double to 64-bit for
5770 32-bit Bionic and to 128-bit for 64-bit Bionic.
5772 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
5773 TARGET_LONG_DOUBLE_128 is true.
5774 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
5776 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
5777 (mlong-double-64): Negate -mlong-double-128.
5778 (mlong-double-128): New option.
5780 * config/i386/i386-c.c (ix86_target_macros): Define
5781 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
5783 * doc/invoke.texi: Document -mlong-double-128.
5785 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
5787 PR rtl-optimization/60024
5788 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
5790 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
5792 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
5794 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
5796 PR rtl-optimization/57662
5797 * sel-sched.c (code_motion_path_driver): Do not mark already not
5798 existing blocks in the visiting bitmap.
5800 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
5802 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
5803 on the insn being emitted.
5805 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
5806 Will Deacon <will.deacon@arm.com>
5808 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
5810 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5812 * config/arm/arm-tables.opt: Regenerate.
5814 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5816 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
5817 for vector types other than V16QImode.
5818 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
5819 define_expand, and call altivec_expand_vec_perm_le when producing
5820 code with little endian element order.
5821 (*altivec_vperm_<mode>_internal): New insn having previous
5822 behavior of altivec_vperm_<mode>.
5823 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
5824 altivec_expand_vec_perm_le when producing code with little endian
5826 (*altivec_vperm_<mode>_uns_internal): New insn having previous
5827 behavior of altivec_vperm_<mode>_uns.
5829 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5831 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
5832 (altivec_vsumsws): Add handling for -maltivec=be with a little
5834 (altivec_vsumsws_direct): New.
5835 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
5836 gen_altivec_vsumsws.
5838 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
5840 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
5841 vtable_pointer_value_to_binfo): New functions.
5842 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
5843 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
5845 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
5847 * config/nios2/nios2.md (load_got_register): Initialize GOT
5848 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
5849 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
5851 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
5853 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
5854 preserverd by passthrough, do not propagate the type.
5856 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
5858 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
5859 (mips_atomic_assign_expand_fenv): New function.
5860 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
5862 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
5864 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
5865 (__builtin_mips_set_fcsr): Likewise.
5866 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
5867 MIPS_USI_FTYPE_VOID.
5868 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
5869 (mips16_expand_set_fcsr): Likewise.
5870 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
5871 (mips16_set_fcsr_stub): Likewise.
5872 (mips16_get_fcsr_one_only_stub): New class.
5873 (mips16_set_fcsr_one_only_stub): Likewise.
5874 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
5875 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
5876 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
5877 (hard_float): New availability predicate.
5878 (mips_builtins): Add get_fcsr and set_fcsr.
5879 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
5880 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
5881 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
5882 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
5883 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
5886 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
5888 * config/mips/mips.c (mips_one_only_stub): New class.
5889 (mips_need_mips16_rdhwr_p): Replace with...
5890 (mips16_rdhwr_stub): ...this new variable.
5891 (mips16_stub_call_address): New function.
5892 (mips16_rdhwr_one_only_stub): New class.
5893 (mips_expand_thread_pointer): Use mips16_stub_call_address.
5894 (mips_output_mips16_rdhwr): Delete.
5895 (mips_finish_stub): New function.
5896 (mips_code_end): Use it to handle rdhwr stubs.
5898 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
5901 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
5902 when calculating size of integer atomic types.
5904 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
5906 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
5908 2014-02-01 Jakub Jelinek <jakub@redhat.com>
5910 PR tree-optimization/60003
5911 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
5912 * profile.c (branch_prob): Use gimple_call_builtin_p
5913 to check for BUILT_IN_SETJMP_RECEIVER.
5914 * tree-inline.c (copy_bb): Call notice_special_calls.
5916 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
5919 * lra-constraints.c (process_alt_operands): Update reload_sum only
5922 2014-01-31 Richard Henderson <rth@redhat.com>
5925 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
5926 until after else_eh is processed.
5928 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
5930 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
5931 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
5932 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
5933 in smmintrin.h, remove them.
5934 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
5935 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
5936 * config/i386/i386.md (ROUND_SAE): Fix value.
5937 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
5938 (const48_operand): New.
5939 * config/i386/subst.md (round), (round_expand): Use
5940 const_4_or_8_to_11_operand.
5941 (round_saeonly), (round_saeonly_expand): Use const48_operand.
5943 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
5945 * config/i386/constraints.md (Yk): Swap meaning with k.
5946 * config/i386/i386.md (movhi_internal): Change Yk to k.
5947 (movqi_internal): Ditto.
5948 (*k<logic><mode>): Ditto.
5951 (kandn<mode>): Ditto.
5952 (*<code>hi_1): Ditto.
5953 (*<code>qi_1): Ditto.
5954 (kxnor<mode>): Ditto.
5955 (kortestzhi): Ditto.
5956 (kortestchi): Ditto.
5958 (*one_cmplhi2_1): Ditto.
5959 (*one_cmplqi2_1): Ditto.
5960 * config/i386/sse.md (): Change k to Yk.
5961 (avx512f_load<mode>_mask): Ditto.
5962 (avx512f_blendm<mode>): Ditto.
5963 (avx512f_store<mode>_mask): Ditto.
5964 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
5965 (avx512f_storedqu<mode>_mask): Ditto.
5966 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
5968 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
5969 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
5970 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
5971 (avx512f_maskcmp<mode>3): Ditto.
5972 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
5973 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
5974 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
5975 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
5976 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
5977 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
5978 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
5979 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
5980 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
5981 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
5982 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
5983 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
5984 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
5985 (vec_extract_lo_<mode>_maskm): Ditto.
5986 (vec_extract_hi_<mode>_maskm): Ditto.
5987 (avx512f_vternlog<mode>_mask): Ditto.
5988 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
5989 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
5990 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
5991 (avx512f_<code>v8div16qi2_mask): Ditto.
5992 (avx512f_<code>v8div16qi2_mask_store): Ditto.
5993 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
5994 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
5995 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
5996 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
5997 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
5998 (*avx512pf_gatherpf<mode>df_mask): Ditto.
5999 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6000 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6001 (avx512cd_maskb_vec_dupv8di): Ditto.
6002 (avx512cd_maskw_vec_dupv16si): Ditto.
6003 (avx512f_vpermi2var<mode>3_maskz): Ditto.
6004 (avx512f_vpermi2var<mode>3_mask): Ditto.
6005 (avx512f_vpermi2var<mode>3_mask): Ditto.
6006 (avx512f_vpermt2var<mode>3_maskz): Ditto.
6007 (*avx512f_gathersi<mode>): Ditto.
6008 (*avx512f_gathersi<mode>_2): Ditto.
6009 (*avx512f_gatherdi<mode>): Ditto.
6010 (*avx512f_gatherdi<mode>_2): Ditto.
6011 (*avx512f_scattersi<mode>): Ditto.
6012 (*avx512f_scatterdi<mode>): Ditto.
6013 (avx512f_compress<mode>_mask): Ditto.
6014 (avx512f_compressstore<mode>_mask): Ditto.
6015 (avx512f_expand<mode>_mask): Ditto.
6016 * config/i386/subst.md (mask): Change k to Yk.
6017 (mask_scalar_merge): Ditto.
6020 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
6022 * doc/extend.texi (Vector Extensions): Document ?: in C++.
6024 2014-01-31 Richard Biener <rguenther@suse.de>
6027 * builtins.c (fold_builtin_memory_op): Make sure to not
6028 use a floating-point mode or a boolean or enumeral type for
6031 2014-01-30 DJ Delorie <dj@redhat.com>
6033 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
6034 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
6035 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
6036 whenever main() has an epilogue.
6038 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6040 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
6041 unused variable "field".
6042 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
6043 (vsx_mergeh_<mode>): Likewise.
6044 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
6045 (altivec_vmrghh): Likewise.
6046 (altivec_vmrghw): Likewise.
6047 (altivec_vmrglb): Likewise.
6048 (altivec_vmrglh): Likewise.
6049 (altivec_vmrglw): Likewise.
6050 (altivec_vspltb): Add missing uses.
6051 (altivec_vsplth): Likewise.
6052 (altivec_vspltw): Likewise.
6053 (altivec_vspltsf): Likewise.
6055 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6058 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
6059 frame related instructions.
6061 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
6063 PR rtl-optimization/59959
6064 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
6065 any reload of register whose subreg is invalid.
6067 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6069 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
6070 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
6071 Add missing return type - void.
6073 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6075 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
6076 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
6077 remove element index adjustment for endian (now handled in vsx.md
6079 (altivec_expand_vec_perm_const): Use
6080 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
6081 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
6082 (vsx_xxspltw_<mode>): Adjust element index for little endian.
6083 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
6084 define_expand and a new define_insn *altivec_vspltb_internal;
6085 adjust for -maltivec=be on a little endian target.
6086 (altivec_vspltb_direct): New.
6087 (altivec_vsplth): Divide into a define_expand and a new
6088 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
6089 little endian target.
6090 (altivec_vsplth_direct): New.
6091 (altivec_vspltw): Divide into a define_expand and a new
6092 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
6093 little endian target.
6094 (altivec_vspltw_direct): New.
6095 (altivec_vspltsf): Divide into a define_expand and a new
6096 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
6097 a little endian target.
6099 2014-01-30 Richard Biener <rguenther@suse.de>
6101 PR tree-optimization/59993
6102 * tree-ssa-forwprop.c (associate_pointerplus): Check we
6103 can propagate form the earlier stmt and avoid the transform
6104 when the intermediate result is needed.
6106 2014-01-30 Alangi Derick <alangiderick@gmail.com>
6108 * README.Portability: Fix typo.
6110 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
6112 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
6113 comparison_operator with ordered_comparison_operator.
6115 2014-01-30 Nick Clifton <nickc@redhat.com>
6117 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
6118 Rename to mn10300_store_multiple_regs.
6119 * config/mn10300/mn10300.c: Likewise.
6120 * config/mn10300/mn10300.md (store_movm): Fix typo: call
6121 store_multiple_regs.
6122 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
6123 Call mn10300_store_multiple_regs.
6125 2014-01-30 Nick Clifton <nickc@redhat.com>
6126 DJ Delorie <dj@redhat.com>
6128 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
6129 %fp 2 to keep registers after it properly word-aligned.
6130 (rl78_alloc_physical_registers_umul): Handle the case where both
6131 input operands are the same.
6133 2014-01-30 Richard Biener <rguenther@suse.de>
6135 PR tree-optimization/59903
6136 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
6139 2014-01-30 Jason Merrill <jason@redhat.com>
6142 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
6145 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
6147 2014-01-30 Richard Biener <rguenther@suse.de>
6149 PR tree-optimization/59951
6150 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
6152 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
6155 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
6156 SFmode to DFmode case.
6158 2014-01-29 DJ Delorie <dj@redhat.com>
6160 * config/msp430/msp430.opt (-minrt): New.
6161 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
6163 (ENDFILE_SPEC): Likewise.
6165 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
6167 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
6168 (estimate_function_body_sizes): Use it.
6170 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
6173 * dwarf2out.c (is_cxx_auto): New.
6174 (is_base_type): Use it.
6175 (gen_type_die_with_usage): Likewise.
6177 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6179 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
6180 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
6181 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
6182 -maltivec=be with LE targets.
6183 (vsx_mergeh_<mode>): Likewise.
6184 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
6185 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
6186 (altivec_vmrghb): Replace with define_expand and new
6187 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
6188 (altivec_vmrghb_direct): New define_insn.
6189 (altivec_vmrghh): Replace with define_expand and new
6190 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
6191 (altivec_vmrghh_direct): New define_insn.
6192 (altivec_vmrghw): Replace with define_expand and new
6193 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
6194 (altivec_vmrghw_direct): New define_insn.
6195 (*altivec_vmrghsf): Adjust for endianness.
6196 (altivec_vmrglb): Replace with define_expand and new
6197 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
6198 (altivec_vmrglb_direct): New define_insn.
6199 (altivec_vmrglh): Replace with define_expand and new
6200 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
6201 (altivec_vmrglh_direct): New define_insn.
6202 (altivec_vmrglw): Replace with define_expand and new
6203 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
6204 (altivec_vmrglw_direct): New define_insn.
6205 (*altivec_vmrglsf): Adjust for endianness.
6206 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6207 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6208 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6209 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6210 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6211 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6212 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6213 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6215 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
6217 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
6218 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
6221 2014-01-29 Richard Biener <rguenther@suse.de>
6223 PR tree-optimization/58742
6224 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
6225 associate_pointerplus_align.
6226 (associate_pointerplus_diff): New function.
6227 (associate_pointerplus): Likewise. Call associate_pointerplus_align
6228 and associate_pointerplus_diff.
6230 2014-01-29 Richard Biener <rguenther@suse.de>
6232 * lto-streamer.h (LTO_major_version): Bump to 3.
6233 (LTO_minor_version): Reset to 0.
6235 2014-01-29 Renlin Li <Renlin.Li@arm.com>
6237 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
6238 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
6239 (arm_file_start): Generate correct asm header for armv7ve.
6240 * config/arm/bpabi.h: Add multilib support for armv7ve.
6241 * config/arm/driver-arm.c: Change the architectures of cortex-a7
6242 and cortex-a15 to armv7ve.
6243 * config/arm/t-aprofile: Add multilib support for armv7ve.
6244 * doc/invoke.texi: Document -march=armv7ve.
6246 2014-01-29 Richard Biener <rguenther@suse.de>
6248 PR tree-optimization/58742
6249 * tree-ssa-forwprop.c (associate_plusminus): Return true
6250 if we changed sth, defer EH cleanup to ...
6251 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
6252 (simplify_mult): New function.
6254 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6257 PR tree-optimization/59920
6258 * tree.c (build_common_builtin_nodes): Remove
6259 __builtin_setjmp_dispatcher initialization.
6260 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
6261 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
6262 instead of gsi_after_labels + manually skipping debug stmts.
6263 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
6264 ignore bbs with IFN_ABNORMAL_DISPATCHER.
6265 * tree-inline.c (copy_edges_for_bb): Remove
6266 can_make_abnormal_goto argument, instead add abnormal_goto_dest
6267 argument. Ignore computed_goto_p stmts. Don't call
6268 make_abnormal_goto_edges. If a call might need abnormal edges
6269 for non-local gotos, see if it already has an edge to
6270 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
6271 with true argument, don't do anything then, otherwise add
6272 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
6273 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
6275 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
6276 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
6277 (lower_stmt): Don't set data->calls_builtin_setjmp.
6278 (lower_builtin_setjmp): Adjust comment.
6279 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
6280 * tree-cfg.c (found_computed_goto): Remove.
6281 (factor_computed_gotos): Remove.
6282 (make_goto_expr_edges): Return bool, true for computed gotos.
6283 Don't call make_abnormal_goto_edges.
6284 (build_gimple_cfg): Don't set found_computed_goto, don't call
6285 factor_computed_gotos.
6286 (computed_goto_p): No longer static.
6287 (make_blocks): Don't set found_computed_goto.
6288 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
6289 (make_edges): If make_goto_expr_edges returns true, push bb
6290 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
6291 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
6292 vector. Record mapping between bbs and OpenMP regions if there
6293 are any, adjust make_gimple_omp_edges caller. Call
6294 handle_abnormal_edges.
6295 (make_abnormal_goto_edges): Remove.
6296 * tree-cfg.h (make_abnormal_goto_edges): Remove.
6297 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
6298 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
6299 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
6300 * internal-fn.def (ABNORMAL_DISPATCHER): New.
6301 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
6302 filling *region also set *region_idx to (*region)->entry->index.
6305 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
6306 For REGs set ORIGINAL_REGNO.
6308 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
6310 * doc/md.texi: Mention that a target shouldn't implement
6311 vec_widen_(s|u)mul_even/odd pair if it is less efficient
6314 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6316 PR tree-optimization/59594
6317 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
6318 a copy of the datarefs vector rather than the vector itself.
6320 2014-01-28 Jason Merrill <jason@redhat.com>
6323 * dwarf2out.c (auto_die): New static.
6324 (gen_type_die_with_usage): Handle C++1y 'auto'.
6325 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
6328 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
6331 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
6332 (SPEC_X32): Likewise.
6333 (SPEC_64): Likewise.
6334 * config/i386/i386.c (ix86_option_override_internal): Turn off
6335 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
6337 (x86_file_start): Output .code16gcc for TARGET_16BIT.
6338 * config/i386/i386.h (TARGET_16BIT): New macro.
6339 (TARGET_16BIT_P): Likewise.
6340 * config/i386/i386.opt: Add m16.
6341 * doc/invoke.texi: Document -m16.
6343 2014-01-28 Jakub Jelinek <jakub@redhat.com>
6345 PR preprocessor/59935
6346 * input.c (location_get_source_line): Bail out on when line number
6347 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
6349 2014-01-28 Richard Biener <rguenther@suse.de>
6351 PR tree-optimization/58742
6352 * tree-ssa-forwprop.c (associate_plusminus): Handle
6353 pointer subtraction of the form (T)(P + A) - (T)P.
6355 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6357 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
6360 2014-01-28 Richard Biener <rguenther@suse.de>
6363 2014-01-28 Richard Biener <rguenther@suse.de>
6365 PR rtl-optimization/45364
6366 PR rtl-optimization/59890
6367 * var-tracking.c (local_get_addr_clear_given_value): Handle
6368 already cleared slot.
6369 (val_reset): Handle not allocated local_get_addr_cache.
6370 (vt_find_locations): Use post-order on the inverted CFG.
6372 2014-01-28 Richard Biener <rguenther@suse.de>
6374 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
6376 2014-01-28 Richard Biener <rguenther@suse.de>
6378 PR rtl-optimization/45364
6379 PR rtl-optimization/59890
6380 * var-tracking.c (local_get_addr_clear_given_value): Handle
6381 already cleared slot.
6382 (val_reset): Handle not allocated local_get_addr_cache.
6383 (vt_find_locations): Use post-order on the inverted CFG.
6385 2014-01-28 Alan Modra <amodra@gmail.com>
6387 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
6388 * configure.ac <recursive call for build != host>: Define
6389 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
6390 and LD_FOR_BUILD too.
6391 * configure: Regenerate.
6393 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
6395 * config/i386/i386.c (get_builtin_code_for_version): Separate
6396 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
6397 Broadwell from Haswell.
6399 2014-01-27 Steve Ellcey <sellcey@mips.com>
6401 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
6402 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
6403 * config/mips/mips.c (mips_option_override): Change setting
6405 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
6406 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
6407 Change from Mask to Var.
6409 2014-01-27 Jeff Law <law@redhat.com>
6411 * ipa-inline.c (inline_small_functions): Fix typo.
6413 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
6415 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
6416 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
6417 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
6418 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
6419 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
6420 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
6421 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
6422 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
6423 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
6424 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
6425 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
6426 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
6427 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
6428 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
6429 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
6430 (_mm512_storeu_epi64): Ditto.
6431 (_mm512_cmpge_epi32_mask): Ditto.
6432 (_mm512_cmpge_epu32_mask): Ditto.
6433 (_mm512_cmpge_epi64_mask): Ditto.
6434 (_mm512_cmpge_epu64_mask): Ditto.
6435 (_mm512_cmple_epi32_mask): Ditto.
6436 (_mm512_cmple_epu32_mask): Ditto.
6437 (_mm512_cmple_epi64_mask): Ditto.
6438 (_mm512_cmple_epu64_mask): Ditto.
6439 (_mm512_cmplt_epi32_mask): Ditto.
6440 (_mm512_cmplt_epu32_mask): Ditto.
6441 (_mm512_cmplt_epi64_mask): Ditto.
6442 (_mm512_cmplt_epu64_mask): Ditto.
6443 (_mm512_cmpneq_epi32_mask): Ditto.
6444 (_mm512_cmpneq_epu32_mask): Ditto.
6445 (_mm512_cmpneq_epi64_mask): Ditto.
6446 (_mm512_cmpneq_epu64_mask): Ditto.
6447 (_mm512_expand_pd): Ditto.
6448 (_mm512_expand_ps): Ditto.
6449 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
6450 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
6451 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
6452 * config/i386/i386.c (ix86_builtins): Add
6453 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
6454 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
6455 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
6456 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
6457 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
6458 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
6459 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
6460 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
6461 IX86_BUILTIN_PMOVUSQW512_MEM.
6462 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
6463 __builtin_ia32_pmovsqd512mem_mask,
6464 __builtin_ia32_pmovqd512mem_mask,
6465 __builtin_ia32_pmovusqw512mem_mask,
6466 __builtin_ia32_pmovsqw512mem_mask,
6467 __builtin_ia32_pmovqw512mem_mask,
6468 __builtin_ia32_pmovusdw512mem_mask,
6469 __builtin_ia32_pmovsdw512mem_mask,
6470 __builtin_ia32_pmovdw512mem_mask,
6471 __builtin_ia32_pmovqb512mem_mask,
6472 __builtin_ia32_pmovusqb512mem_mask,
6473 __builtin_ia32_pmovsqb512mem_mask,
6474 __builtin_ia32_pmovusdb512mem_mask,
6475 __builtin_ia32_pmovsdb512mem_mask,
6476 __builtin_ia32_pmovdb512mem_mask.
6477 (bdesc_args): Add __builtin_ia32_expanddf512,
6478 __builtin_ia32_expandsf512.
6479 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
6480 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
6481 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
6482 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
6483 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
6484 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
6485 (avx512f_<code>v8div16qi2_mask_store): This.
6486 (avx512f_expand<mode>): New.
6488 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
6490 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
6492 (_mm512_mask_prefetch_i64gather_pd): Ditto.
6493 (_mm512_prefetch_i32scatter_pd): Ditto.
6494 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
6495 (_mm512_prefetch_i64scatter_pd): Ditto.
6496 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
6497 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
6498 (_mm512_mask_prefetch_i64gather_ps): Ditto.
6499 (_mm512_prefetch_i32scatter_ps): Ditto.
6500 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
6501 (_mm512_prefetch_i64scatter_ps): Ditto.
6502 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
6503 * config/i386/i386-builtin-types.def: Define
6504 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
6505 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
6506 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
6507 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
6508 IX86_BUILTIN_SCATTERPFQPD.
6509 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
6510 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
6511 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
6512 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
6513 __builtin_ia32_scatterpfqps.
6514 (ix86_expand_builtin): Expand new built-ins.
6515 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
6516 fix memory access data type.
6517 (*avx512pf_gatherpf<mode>_mask): Ditto.
6518 (*avx512pf_gatherpf<mode>): Ditto.
6519 (avx512pf_scatterpf<mode>): Ditto.
6520 (*avx512pf_scatterpf<mode>_mask): Ditto.
6521 (*avx512pf_scatterpf<mode>): Ditto.
6522 (GATHER_SCATTER_SF_MEM_MODE): New.
6523 (avx512pf_gatherpf<mode>df): Ditto.
6524 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6525 (*avx512pf_scatterpf<mode>df): Ditto.
6527 2014-01-27 Jakub Jelinek <jakub@redhat.com>
6530 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
6531 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
6534 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
6536 * common/config/arm/arm-common.c
6537 (arm_rewrite_mcpu): Handle multiple names.
6539 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
6541 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
6543 * gimple-builder.h (create_gimple_tmp): Delete.
6545 2014-01-27 Christian Bruel <christian.bruel@st.com>
6547 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
6550 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
6552 * config/pa/pa.md (call): Generate indirect long calls to non-local
6553 functions when outputing 32-bit code.
6554 (call_value): Likewise except for special call to buggy powf function.
6556 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
6557 portable runtime and PIC indirect calls.
6558 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
6559 and PIC call sequences. Use ldo instead of blr to set return register
6560 in PIC call sequence.
6562 2014-01-25 Walter Lee <walt@tilera.com>
6564 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
6565 avoid clobbering a live register.
6567 2014-01-25 Walter Lee <walt@tilera.com>
6569 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
6570 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
6571 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
6572 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
6574 2014-01-25 Walter Lee <walt@tilera.com>
6576 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
6577 arguments on even registers.
6578 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
6580 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
6581 (BIGGEST_ALIGNMENT): Ditto.
6582 (BIGGEST_FIELD_ALIGNMENT): Ditto.
6584 2014-01-25 Walter Lee <walt@tilera.com>
6586 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
6587 insns before bundling.
6588 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
6590 2014-01-25 Walter Lee <walt@tilera.com>
6592 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
6593 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
6594 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
6596 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
6598 * config/mips/constraints.md (kl): Delete.
6599 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
6600 define expands, using...
6601 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
6602 instructions for MIPS16.
6603 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
6604 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
6606 2014-01-25 Walter Lee <walt@tilera.com>
6608 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
6612 2014-01-25 Walter Lee <walt@tilera.com>
6614 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
6615 (TARGET_EXPAND_TO_RTL_HOOK): Define.
6617 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
6619 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
6622 2014-01-25 Jakub Jelinek <jakub@redhat.com>
6624 * print-rtl.c (in_call_function_usage): New var.
6625 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
6626 EXPR_LIST mode as mode and not as reg note name.
6629 * cfgloopmanip.c (copy_loop_info): If
6630 loop->warned_aggressive_loop_optimizations, make sure
6631 the flag is set in target loop too.
6633 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
6635 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
6637 * builtins.def: Likewise.
6638 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
6639 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
6640 * ira.c (ira_setup_eliminable_regset): Likewise.
6641 * omp-low.c (gate_expand_omp): Likewise.
6642 (execute_lower_omp): Likewise.
6643 (diagnose_sb_0): Likewise.
6644 (gate_diagnose_omp_blocks): Likewise.
6645 (simd_clone_clauses_extract): Likewise.
6648 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6650 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
6651 correction for little endian...
6652 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
6655 2014-01-24 Jeff Law <law@redhat.com>
6657 PR tree-optimization/59919
6658 * tree-vrp.c (find_assert_locations_1): Do not register asserts
6659 for non-returning calls.
6661 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
6663 * common/config/aarch64/aarch64-common.c
6664 (aarch64_rewrite_mcpu): Handle multiple names.
6665 * config/aarch64/aarch64.h
6666 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
6668 2014-01-24 Dodji Seketeli <dodji@redhat.com>
6670 * input.c (add_file_to_cache_tab): Handle the case where fopen
6673 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
6676 * config/i386/i386.md (pushsf splitter): Get stack adjustment
6677 from push operand if code of push isn't PRE_DEC.
6679 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
6682 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
6683 -mquad-memory-atomic. Update -mquad-memory documentation to say
6684 it is only used for non-atomic loads/stores.
6686 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
6687 -mquad-memory or -mquad-memory-atomic switches.
6689 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
6690 -mquad-memory-atomic to ISA 2.07 support.
6692 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
6693 to separate support of normal quad word memory operations (ldq, stq)
6694 from the atomic quad word memory operations.
6696 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
6697 support to separate non-atomic quad word operations from atomic
6698 quad word operations. Disable non-atomic quad word operations in
6699 little endian mode so that we don't have to swap words after the
6700 load and before the store.
6701 (quad_load_store_p): Add comment about atomic quad word support.
6702 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
6703 options printed with -mdebug=reg.
6705 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
6706 -mquad-memory-atomic as the test for whether we have quad word
6707 atomic instructions.
6708 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
6709 or -mp8-vector are used, allow byte/half-word atomic operations.
6711 * config/rs6000/sync.md (load_lockedti): Insure that the address
6712 is a proper indexed or indirect address for the lqarx instruction.
6713 On little endian systems, swap the hi/lo registers after the lqarx
6715 (load_lockedpti): Use indexed_or_indirect_operand predicate to
6716 insure the address is valid for the lqarx instruction.
6717 (store_conditionalti): Insure that the address is a proper indexed
6718 or indirect address for the stqcrx. instruction. On little endian
6719 systems, swap the hi/lo registers before doing the stqcrx.
6721 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
6722 insure the address is valid for the stqcrx. instruction.
6724 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
6725 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
6726 type of quad memory support is available.
6728 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
6731 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
6732 there is a danger of looping.
6734 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
6736 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
6737 force flag_ira_loop_pressure if set via command line.
6739 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
6741 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
6742 (ashr_simd): New builtin handling DI mode.
6743 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
6744 (aarch64_sshr_simddi): New match pattern.
6745 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
6746 (vshrd_n_s64): Likewise.
6747 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
6749 2014-01-23 Nick Clifton <nickc@redhat.com>
6751 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
6752 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
6753 favour of mcu specific scripts.
6754 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
6757 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
6758 Alex Velenko <Alex.Velenko@arm.com>
6760 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
6761 (vaddv_s16): Likewise.
6762 (vaddv_s32): Likewise.
6763 (vaddv_u8): Likewise.
6764 (vaddv_u16): Likewise.
6765 (vaddv_u32): Likewise.
6766 (vaddvq_s8): Likewise.
6767 (vaddvq_s16): Likewise.
6768 (vaddvq_s32): Likewise.
6769 (vaddvq_s64): Likewise.
6770 (vaddvq_u8): Likewise.
6771 (vaddvq_u16): Likewise.
6772 (vaddvq_u32): Likewise.
6773 (vaddvq_u64): Likewise.
6774 (vaddv_f32): Likewise.
6775 (vaddvq_f32): Likewise.
6776 (vaddvq_f64): Likewise.
6777 (vmaxv_f32): Likewise.
6778 (vmaxv_s8): Likewise.
6779 (vmaxv_s16): Likewise.
6780 (vmaxv_s32): Likewise.
6781 (vmaxv_u8): Likewise.
6782 (vmaxv_u16): Likewise.
6783 (vmaxv_u32): Likewise.
6784 (vmaxvq_f32): Likewise.
6785 (vmaxvq_f64): Likewise.
6786 (vmaxvq_s8): Likewise.
6787 (vmaxvq_s16): Likewise.
6788 (vmaxvq_s32): Likewise.
6789 (vmaxvq_u8): Likewise.
6790 (vmaxvq_u16): Likewise.
6791 (vmaxvq_u32): Likewise.
6792 (vmaxnmv_f32): Likewise.
6793 (vmaxnmvq_f32): Likewise.
6794 (vmaxnmvq_f64): Likewise.
6795 (vminv_f32): Likewise.
6796 (vminv_s8): Likewise.
6797 (vminv_s16): Likewise.
6798 (vminv_s32): Likewise.
6799 (vminv_u8): Likewise.
6800 (vminv_u16): Likewise.
6801 (vminv_u32): Likewise.
6802 (vminvq_f32): Likewise.
6803 (vminvq_f64): Likewise.
6804 (vminvq_s8): Likewise.
6805 (vminvq_s16): Likewise.
6806 (vminvq_s32): Likewise.
6807 (vminvq_u8): Likewise.
6808 (vminvq_u16): Likewise.
6809 (vminvq_u32): Likewise.
6810 (vminnmv_f32): Likewise.
6811 (vminnmvq_f32): Likewise.
6812 (vminnmvq_f64): Likewise.
6814 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
6816 * config/aarch64/aarch64-simd.md
6817 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
6818 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
6819 (*aarch64_mul3_elt<mode>): Likewise.
6820 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
6821 (*aarch64_mul3_elt_to_64v2df): Likewise.
6822 (*aarch64_mla_elt<mode>): Likewise.
6823 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
6824 (*aarch64_mls_elt<mode>): Likewise.
6825 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
6826 (*aarch64_fma4_elt<mode>): Likewise.
6827 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
6828 (*aarch64_fma4_elt_to_64v2df): Likewise.
6829 (*aarch64_fnma4_elt<mode>): Likewise.
6830 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
6831 (*aarch64_fnma4_elt_to_64v2df): Likewise.
6832 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
6833 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
6834 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
6835 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
6836 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
6837 (aarch64_sqdmull_lane<mode>_internal): Likewise.
6838 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
6840 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
6842 * config/aarch64/aarch64-simd.md
6843 (aarch64_be_checked_get_lane<mode>): New define_expand.
6844 * config/aarch64/aarch64-simd-builtins.def
6845 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
6846 New builtin definition.
6847 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
6848 Use new safe be builtin.
6850 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
6852 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
6854 (aarch64_be_st1<mode>): Likewise.
6855 (aarch_ld1<VALL:mode>): Define_expand modified.
6856 (aarch_st1<VALL:mode>): Likewise.
6857 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
6858 (UNSPEC_ST1): Likewise.
6860 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
6862 * config/microblaze/microblaze.md: Add trap insn and attribute
6864 2014-01-23 Dodji Seketeli <dodji@redhat.com>
6866 PR preprocessor/58580
6867 * input.h (location_get_source_line): Take an additional line_size
6869 (void diagnostics_file_cache_fini): Declare new function.
6870 * input.c (struct fcache): New type.
6871 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
6872 New static constants.
6873 (diagnostic_file_cache_init, total_lines_num)
6874 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
6875 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
6876 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
6877 (get_next_line, read_next_line, goto_next_line, read_line_num):
6878 New static function definitions.
6879 (diagnostic_file_cache_fini): New function.
6880 (location_get_source_line): Take an additional output line_len
6881 parameter. Re-write using lookup_or_add_file_to_cache_tab and
6883 * diagnostic.c (diagnostic_finish): Call
6884 diagnostic_file_cache_fini.
6885 (adjust_line): Take an additional input parameter for the length
6886 of the line, rather than calculating it with strlen.
6887 (diagnostic_show_locus): Adjust the use of
6888 location_get_source_line and adjust_line with respect to their new
6889 signature. While displaying a line now, do not stop at the first
6890 null byte. Rather, display the zero byte as a space and keep
6891 going until we reach the size of the line.
6892 * Makefile.in: Add vec.o to OBJS-libcommon
6894 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
6895 Ilya Tocar <ilya.tocar@intel.com>
6897 * config/i386/avx512fintrin.h (_mm512_kmov): New.
6898 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
6899 (__builtin_ia32_kmov16): Ditto.
6900 * config/i386/i386.md (UNSPEC_KMOV): New.
6903 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
6905 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
6906 (_mm512_storeu_si512): Ditto.
6908 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
6911 * rtl.h (get_referenced_operands): Declare.
6912 * recog.c (get_referenced_operands): New function.
6913 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
6914 operands have been referenced when recording LO_SUM references.
6916 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
6918 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
6920 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
6922 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
6923 Enable for generic and recent AMD targets.
6925 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
6927 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
6928 ARG_SIZE note when adjustment was eliminated.
6930 2014-01-22 Jeff Law <law@redhat.com>
6932 PR tree-optimization/59597
6933 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
6934 in file. Accept new argument REGISTERING and use it to modify
6935 dump output appropriately.
6936 (register_jump_thread): Corresponding changes.
6937 (mark_threaded_blocks): Reinstate code to cancel unprofitable
6938 thread paths involving joiner blocks. Add code to dump cancelled
6939 jump threading paths.
6941 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
6943 PR rtl-optimization/59477
6944 * lra-constraints.c (inherit_in_ebb): Process call for living hard
6945 regs. Update reloads_num and potential_reload_hard_regs for all insns.
6947 2014-01-22 Tom Tromey <tromey@redhat.com>
6949 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
6951 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
6953 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
6955 PR rtl-optimization/59896
6956 * lra-constraints.c (process_alt_operands): Check unused note for
6957 matched operands of insn with no output reloads.
6959 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
6961 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
6962 (mips_move_from_gpr_cost): Likewise.
6964 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
6966 PR rtl-optimization/59858
6967 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
6968 ira_class_hard_regs_num.
6969 (process_alt_operands): Increase reject for dying matched operand.
6971 2014-01-21 Jakub Jelinek <jakub@redhat.com>
6974 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
6975 smaller than size, perform several stores or loads and stores
6976 at dst + count - size to store or copy all of size bytes, rather
6977 than just last modesize bytes.
6979 2014-01-20 DJ Delorie <dj@redhat.com>
6981 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
6982 that CLOBBERs are REGs before propogating their values.
6984 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
6987 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
6988 (cgraph_inline_failed_type): New function.
6989 * cgraph.h (DEFCIFCODE): Add type.
6990 (cgraph_inline_failed_type_t): New enum.
6991 (cgraph_inline_failed_type): New prototype.
6992 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
6993 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
6994 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
6995 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
6996 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
6997 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
6998 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
6999 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
7000 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
7001 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
7002 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
7003 OPTIMIZATION_MISMATCH.
7004 * tree-inline.c (expand_call_inline): Emit errors during
7005 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
7007 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7010 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
7011 mode attribute in insn output.
7013 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
7015 * output.h (output_constant): Delete.
7016 * varasm.c (output_constant): Make private.
7018 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
7020 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
7022 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7025 * tree.h (fold_builtin_strcat): New prototype.
7026 * builtins.c (fold_builtin_strcat): No longer static. Add len
7027 argument, if non-NULL, don't call c_strlen. Optimize
7028 directly into __builtin_memcpy instead of __builtin_strcpy.
7029 (fold_builtin_2): Adjust fold_builtin_strcat caller.
7030 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
7032 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7034 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7035 for SImode_address_operand operands, having only a REG argument.
7037 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
7039 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
7040 loader name using mbig-endian.
7041 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
7043 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7045 * doc/invoke.texi (-march): Clarify documentation for AArch64.
7049 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
7051 * config/aarch64/aarch64-protos.h
7052 (aarch64_cannot_change_mode_class_ptr): Declare.
7053 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
7054 aarch64_cannot_change_mode_class_ptr): New.
7055 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
7056 backend hook aarch64_cannot_change_mode_class.
7058 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7060 * common/config/aarch64/aarch64-common.c
7061 (aarch64_handle_option): Don't handle any option order logic here.
7062 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
7063 selected_cpu, warn on architecture version mismatch.
7064 (aarch64_override_options): Fix parsing order for option strings.
7066 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7067 Iain Sandoe <iain@codesourcery.com>
7070 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
7071 warning. Amend comment to reflect current functionality.
7073 2014-01-20 Richard Biener <rguenther@suse.de>
7076 * builtins.c (fold_builtin_strcat): Remove case better handled
7077 by tree-ssa-strlen.c.
7079 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
7081 * config/aarch64/aarch64.opt
7082 (mcpu, march, mtune): Make case-insensitive.
7084 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7087 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7088 if operands[1] is a REG or ZERO_EXTEND of a REG.
7090 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
7092 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
7094 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
7096 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
7097 long non-pic millicode calls.
7099 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7101 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
7103 2014-01-19 Kito Cheng <kito@0xlab.org>
7105 * builtins.c (expand_movstr): Check movstr expand done or fail.
7107 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7108 H.J. Lu <hongjiu.lu@intel.com>
7111 * config/i386/i386.md (*lea<mode>): Zero-extend return register
7112 to DImode for zero-extended addresses.
7114 2014-01-19 Jakub Jelinek <jakub@redhat.com>
7116 PR rtl-optimization/57763
7117 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
7118 on the new indirect jump_insn and increment LABEL_NUSES (label).
7120 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
7124 * config.gcc (x86_archs): New variable.
7125 (x86_64_archs): Likewise.
7126 (x86_cpus): Likewise.
7127 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
7128 --with-arch/--with-cpu= options.
7129 Support --with-arch=/--with-cpu={nehalem,westmere,
7130 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
7132 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7134 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
7135 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
7137 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7139 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
7141 2014-01-18 Jakub Jelinek <jakub@redhat.com>
7144 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
7145 clear cpp_get_options (parse_in)->warn_unused_macros for
7146 ix86_target_macros_internal with cpp_define.
7148 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
7150 * jump.c (delete_related_insns): Keep (use (insn))s.
7151 * reorg.c (redundant_insn): Check for barriers too.
7153 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7155 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
7157 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
7159 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
7160 call to $$dyncall when TARGET_LONG_CALLS is true.
7162 2014-01-17 Jeff Law <law@redhat.com>
7164 * ree.c (combine_set_extension): Temporarily disable test for
7165 changing number of hard registers.
7167 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7170 * ipa-inline-analysis.c (inline_free_summary):
7171 Do not free summary of aliases.
7173 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7176 * gimplify.c (gimplify_expr): Use create_tmp_var
7177 instead of create_tmp_var_raw. If cond doesn't have
7178 integral type, don't add the IFN_ANNOTATE builtin at all.
7180 2014-01-17 Martin Jambor <mjambor@suse.cz>
7183 * ipa-cp.c (prev_edge_clone): New variable.
7184 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
7185 Also resize prev_edge_clone vector.
7186 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
7187 (ipcp_edge_removal_hook): New function.
7188 (ipcp_driver): Register ipcp_edge_removal_hook.
7190 2014-01-17 Andrew Pinski <apinski@cavium.com>
7191 Steve Ellcey <sellcey@mips.com>
7194 * config/mips/mips.c (mips_print_operand): Check operand mode instead
7197 2014-01-17 Jeff Law <law@redhat.com>
7200 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
7201 so that pass_ccp runs first.
7203 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7205 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
7206 (ix86_adjust_cost): Use !TARGET_XXX.
7207 (do_reorder_for_imul): Likewise.
7208 (swap_top_of_ready_list): Likewise.
7209 (ix86_sched_reorder): Likewise.
7211 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7213 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
7214 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
7215 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
7216 (intel_memset): New. Duplicate slm_memset.
7217 (intel_cost): New. Duplicate slm_cost.
7218 (m_INTEL): New macro.
7219 (processor_target_table): Add "intel".
7220 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
7221 with PROCESSOR_INTEL for "intel".
7222 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
7223 PROCESSOR_SILVERMONT.
7224 (ix86_issue_rate): Likewise.
7225 (ix86_adjust_cost): Likewise.
7226 (ia32_multipass_dfa_lookahead): Likewise.
7227 (swap_top_of_ready_list): Likewise.
7228 (ix86_sched_reorder): Likewise.
7229 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
7230 instead of TARGET_OPT_AGU.
7231 * config/i386/i386.h (TARGET_INTEL): New.
7232 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
7233 (processor_type): Add PROCESSOR_INTEL.
7234 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
7235 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
7237 2014-01-17 Marek Polacek <polacek@redhat.com>
7240 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
7243 2014-01-17 Richard Biener <rguenther@suse.de>
7245 PR tree-optimization/46590
7246 * opts.c (default_options_table): Add entries for
7247 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
7248 all enabled at -O1 but not for -Og.
7249 * common.opt (fbranch-count-reg): Remove Init(1).
7250 (fmove-loop-invariants): Likewise.
7251 (ftree-pta): Likewise.
7253 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7255 * config/i386/i386.c (ix86_data_alignment): For compatibility with
7256 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
7257 decls to at least the GCC 4.8 used alignments.
7260 * tree-nested.c (convert_nonlocal_reference_stmt,
7261 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
7262 of GIMPLE_BIND stmts, adjust associated decls.
7264 2014-01-17 Richard Biener <rguenther@suse.de>
7266 PR tree-optimization/46590
7267 * vec.h (vec<>::bseach): New member function implementing
7268 binary search according to C89 bsearch.
7269 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
7270 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
7271 bitmap pointer again. Make accesses_in_loop a flat array.
7272 (mem_ref_obstack): New global.
7273 (outermost_indep_loop): Adjust for mem_ref->stored changes.
7274 (mark_ref_stored): Likewise.
7275 (ref_indep_loop_p_2): Likewise.
7276 (set_ref_stored_in_loop): New helper function.
7277 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
7278 (memref_free): Adjust.
7279 (record_mem_ref_loc): Simplify.
7280 (gather_mem_refs_stmt): Adjust.
7281 (sort_locs_in_loop_postorder_cmp): New function.
7282 (analyze_memory_references): Sort accesses_in_loop after
7283 loop postorder number.
7284 (find_ref_loc_in_loop_cmp): New function.
7285 (for_all_locs_in_loop): Find relevant cluster of locs in
7286 accesses_in_loop and iterate without recursion.
7287 (execute_sm): Avoid uninit warning.
7288 (struct ref_always_accessed): Simplify.
7289 (ref_always_accessed::operator ()): Likewise.
7290 (ref_always_accessed_p): Likewise.
7291 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
7292 loop postorder numbers here.
7293 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
7296 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7299 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
7300 on decls for which assemble_alias has been called.
7302 2014-01-17 Nick Clifton <nickc@redhat.com>
7304 * config/msp430/msp430.opt: (mcpu): New option.
7305 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
7306 (msp430_option_override): Parse target_cpu. If the MCU name
7307 matches a generic string, clear target_mcu.
7308 (msp430_attr): Allow numeric interrupt values up to 63.
7309 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
7310 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
7312 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
7314 * config/msp430/msp430.md (popm): Use %J rather than %I.
7315 (addsi3): Use msp430_nonimmediate_operand for operand 2.
7316 (addhi_cy_i): Use immediate_operand for operand 2.
7317 * doc/invoke.texi: Document -mcpu option.
7319 2014-01-17 Richard Biener <rguenther@suse.de>
7321 PR rtl-optimization/38518
7322 * df.h (df_analyze_loop): Declare.
7323 * df-core.c: Include cfgloop.h.
7324 (df_analyze_1): Split out main part of df_analyze.
7325 (df_analyze): Adjust.
7326 (loop_inverted_post_order_compute): New function.
7327 (loop_post_order_compute): Likewise.
7328 (df_analyze_loop): New function avoiding whole-function
7330 * loop-invariant.c (find_defs): Use df_analyze_loop.
7331 (find_invariants): Adjust.
7332 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
7334 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
7336 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
7337 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
7339 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
7341 * ipa-ref.c (ipa_remove_stmt_references): Fix references
7342 traversal when removing references.
7344 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
7347 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
7349 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
7352 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
7353 pushing a reload for an autoinc when we had previously reloaded an
7354 inner part of the address.
7356 2014-01-16 Jakub Jelinek <jakub@redhat.com>
7358 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
7360 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
7361 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
7362 when not giving up or versioning for alias only because of
7364 (vect_analyze_data_ref_dependences): Set to true.
7365 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
7367 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
7368 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
7372 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
7375 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
7376 operand 0 predicate for gathers, use a new pseudo as subtarget.
7378 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7381 * lra-constraints.c (process_alt_operands): Add printing debug info.
7382 Check absence of input/output reloads for matched operands too.
7384 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7386 PR rtl-optimization/59835
7387 * ira.c (ira_init_register_move_cost): Increase cost for
7390 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
7392 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
7394 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
7397 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
7398 non-register objects. Use gen_(high/low)part more consistently.
7401 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
7404 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
7405 endian support, remove tests for WORDS_BIG_ENDIAN.
7406 (p8_mfvsrd_3_<mode>): Likewise.
7407 (reload_gpr_from_vsx<mode>): Likewise.
7408 (reload_gpr_from_vsxsf): Likewise.
7409 (p8_mfvsrd_4_disf): Likewise.
7411 2014-01-16 Richard Biener <rguenther@suse.de>
7413 PR rtl-optimization/46590
7414 * lcm.c (compute_antinout_edge): Use postorder iteration.
7415 (compute_laterin): Use inverted postorder iteration.
7417 2014-01-16 Nick Clifton <nickc@redhat.com>
7420 * varasm.c (output_constant): Return the number of bytes actually
7422 (output_constructor_array_range): Update the field size with the
7423 number of bytes emitted by output_constant.
7424 (output_constructor_regular_field): Likewise. Also do not
7425 complain if the total number of bytes emitted is now greater
7426 than the expected fieldpos.
7427 * output.h (output_constant): Update prototype and descriptive comment.
7429 2014-01-16 Marek Polacek <polacek@redhat.com>
7432 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
7433 it is error_mark_node.
7435 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
7437 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
7438 VALID_AVX256_REG_OR_OI_MODE.
7440 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
7442 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
7443 current procedure should be profiled.
7445 2014-01-15 Andrew Pinski <apinski@cavium.com>
7447 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
7448 of moving from/to the STACK_REG register class.
7450 2014-01-15 Richard Henderson <rth@redhat.com>
7453 * reginfo.c (global_regs_decl): Globalize.
7454 * rtl.h (global_regs_decl): Declare.
7455 * ira.c (do_reload): Diagnose frame_pointer_needed and it
7456 reserved via global_regs.
7458 2014-01-15 Teresa Johnson <tejohnson@google.com>
7460 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
7462 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
7464 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
7465 and vmulosh rather than call gen_vec_widen_smult_*.
7466 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
7467 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
7468 (vec_widen_smult_even_v16qi): Likewise.
7469 (vec_widen_umult_even_v8hi): Likewise.
7470 (vec_widen_smult_even_v8hi): Likewise.
7471 (vec_widen_umult_odd_v16qi): Likewise.
7472 (vec_widen_smult_odd_v16qi): Likewise.
7473 (vec_widen_umult_odd_v8hi): Likewise.
7474 (vec_widen_smult_odd_v8hi): Likewise.
7475 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
7476 vmuloub rather than call gen_vec_widen_umult_*.
7477 (vec_widen_umult_lo_v16qi): Likewise.
7478 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
7479 vmulosb rather than call gen_vec_widen_smult_*.
7480 (vec_widen_smult_lo_v16qi): Likewise.
7481 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
7482 rather than call gen_vec_widen_umult_*.
7483 (vec_widen_umult_lo_v8hi): Likewise.
7484 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
7485 rather than call gen_vec_widen_smult_*.
7486 (vec_widen_smult_lo_v8hi): Likewise.
7488 2014-01-15 Jeff Law <law@redhat.com>
7490 PR tree-optimization/59747
7491 * ree.c (find_and_remove_re): Properly handle case where a second
7492 eliminated extension requires widening a copy created for elimination
7493 of a prior extension.
7494 (combine_set_extension): Ensure that the number of hard regs needed
7495 for a destination register does not change when we widen it.
7497 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
7499 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
7500 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
7501 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
7502 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
7503 (avr-*-rtems*): Likewise.
7504 (bfin*-rtems*): Likewise.
7505 (moxie-*-rtems*): Likewise.
7506 (h8300-*-rtems*): Likewise.
7507 (i[34567]86-*-rtems*): Likewise.
7508 (lm32-*-rtems*): Likewise.
7509 (m32r-*-rtems*): Likewise.
7510 (m68k-*-rtems*): Likewise.
7511 (microblaze*-*-rtems*): Likewise.
7512 (mips*-*-rtems*): Likewise.
7513 (powerpc-*-rtems*): Likewise.
7514 (sh-*-rtems*): Likewise.
7515 (sparc-*-rtems*): Likewise.
7516 (sparc64-*-rtems*): Likewise.
7517 (v850-*-rtems*): Likewise.
7518 (m32c-*-rtems*): Likewise.
7520 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
7522 PR rtl-optimization/59511
7523 * ira.c (ira_init_register_move_cost): Use memory costs for some
7524 cases of register move cost calculations.
7525 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
7526 instead of BB frequency.
7527 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
7528 * lra-assigns.c (find_hard_regno_for): Ditto.
7530 2014-01-15 Richard Biener <rguenther@suse.de>
7532 PR tree-optimization/59822
7533 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
7534 (vectorizable_load): Use it to hoist defs of uses of invariant
7535 loads out of the loop.
7537 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
7538 Kugan Vivekanandarajah <kuganv@linaro.org>
7541 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
7544 2014-01-15 Richard Biener <rguenther@suse.de>
7546 PR rtl-optimization/59802
7547 * lcm.c (compute_available): Use inverted postorder to seed
7548 the initial worklist.
7550 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7553 * config/s390/s390.c (s390_preferred_reload_class): Don't return
7554 ADDR_REGS for invalid symrefs in non-PIC code.
7556 2014-01-15 Jakub Jelinek <jakub@redhat.com>
7559 * builtins.c (determine_block_size): Initialize *probable_max_size
7560 even if len_rtx is CONST_INT.
7562 2014-01-14 Andrew Pinski <apinski@cavium.com>
7564 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
7565 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
7566 (cortexa53_tunings): Likewise.
7567 (aarch64_sched_issue_rate): New function.
7568 (TARGET_SCHED_ISSUE_RATE): Define.
7570 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
7572 * ira-costs.c (find_costs_and_classes): Add missed
7573 ira_init_register_move_cost_if_necessary.
7575 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
7578 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
7580 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
7583 * config/i386/i386.c (type_natural_mode): Add a bool parameter
7584 to indicate if type is used for function return value. Warn ABI
7585 change if the vector mode isn't available for function return value.
7586 (ix86_function_arg_advance): Pass false to type_natural_mode.
7587 (ix86_function_arg): Likewise.
7588 (ix86_gimplify_va_arg): Likewise.
7589 (function_arg_32): Don't warn ABI change.
7590 (ix86_function_value): Pass true to type_natural_mode.
7591 (ix86_return_in_memory): Likewise.
7592 (ix86_struct_value_rtx): Removed.
7593 (TARGET_STRUCT_VALUE_RTX): Likewise.
7595 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7597 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
7598 converting a conditional jump into a conditional return.
7600 2014-01-14 Richard Biener <rguenther@suse.de>
7602 PR tree-optimization/58921
7603 PR tree-optimization/59006
7604 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
7605 hoisting invariant stmts.
7606 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
7607 invariant loads on the preheader edge if possible.
7609 2014-01-14 Joey Ye <joey.ye@arm.com>
7611 * doc/plugin.texi (Building GCC plugins): Update to C++.
7613 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
7615 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
7616 (_mm_rcp28_round_ss): Ditto.
7617 (_mm_rsqrt28_round_sd): Ditto.
7618 (_mm_rsqrt28_round_ss): Ditto.
7619 (_mm_rcp28_sd): Ditto.
7620 (_mm_rcp28_ss): Ditto.
7621 (_mm_rsqrt28_sd): Ditto.
7622 (_mm_rsqrt28_ss): Ditto.
7623 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
7624 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
7625 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
7626 (IX86_BUILTIN_RCP28SD): Ditto.
7627 (IX86_BUILTIN_RCP28SS): Ditto.
7628 (IX86_BUILTIN_RSQRT28SD): Ditto.
7629 (IX86_BUILTIN_RSQRT28SS): Ditto.
7630 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
7631 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
7632 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
7633 (ix86_expand_special_args_builtin): Expand new FTYPE.
7634 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
7635 (srcp14<mode>): Make insn unary.
7636 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
7637 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
7638 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
7639 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
7640 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
7641 Fix rounding: make it SAE only.
7642 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
7644 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
7646 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
7647 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
7648 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
7649 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
7650 (round_saeonly_mask_scalar_operand4): Ditto.
7651 (round_saeonly_mask_scalar_op3): Ditto.
7652 (round_saeonly_mask_scalar_op4): Ditto.
7654 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7656 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
7657 Implement -maltivec=be for vec_insert and vec_extract.
7659 2014-01-10 DJ Delorie <dj@redhat.com>
7661 * config/msp430/msp430.md (call_internal): Don't allow memory
7662 references with SP as the base register.
7663 (call_value_internal): Likewise.
7664 * config/msp430/constraints.md (Yc): New. For memory references
7665 that don't use SP as a base register.
7667 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
7668 "an integer without a # prefix"
7669 * config/msp430/msp430.md (epilogue_helper): Use it.
7671 2014-01-13 Jakub Jelinek <jakub@redhat.com>
7674 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
7675 AVX512F gather builtins.
7676 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
7677 on gather decls with INTEGER_TYPE masktype.
7678 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
7679 directly into the builtin rather than hoisting it before loop.
7681 PR tree-optimization/59387
7682 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
7683 (scev_const_prop): If folded_casts and type has undefined overflow,
7684 use force_gimple_operand instead of force_gimple_operand_gsi and
7685 for each added stmt if it is assign with
7686 arith_code_with_undefined_signed_overflow, call
7687 rewrite_to_defined_overflow.
7688 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
7689 gimple-fold.h instead.
7690 (arith_code_with_undefined_signed_overflow,
7691 rewrite_to_defined_overflow): Moved to ...
7692 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
7693 rewrite_to_defined_overflow): ... here. No longer static.
7694 Include gimplify-me.h.
7695 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
7696 rewrite_to_defined_overflow): New prototypes.
7698 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7700 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
7702 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
7704 * builtins.c (get_object_alignment_2): Minor tweak.
7705 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
7707 2014-01-13 Christian Bruel <christian.bruel@st.com>
7709 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
7710 optimized non constant lengths.
7712 2014-01-13 Jakub Jelinek <jakub@redhat.com>
7715 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
7716 load as __atomic_load_N if possible.
7718 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
7720 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
7722 (rs6000_expand_builtin): Adjust call.
7724 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
7727 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
7728 * config/rs6000/rs6000.c: Include target-globals.h.
7729 (rs6000_set_current_function): Instead of doing target_reinit
7730 unconditionally, use save_target_globals_default_opts and
7731 restore_target_globals.
7733 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
7735 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
7736 (rs6000_expand_builtin): Handle mffs and mtfsf.
7737 (rs6000_init_builtins): Define mffs and mtfsf.
7738 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
7739 (rs6000_mffs): New pattern.
7740 (rs6000_mtfsf): New pattern.
7742 2014-01-11 Bin Cheng <bin.cheng@arm.com>
7744 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
7745 Start narrowing with START. Apply candidate-use pair
7746 and check overall cost in narrowing.
7747 (iv_ca_prune): Pass new argument.
7749 2014-01-10 Jeff Law <law@redhat.com>
7752 * ree.c (combine_reaching_defs): Ensure the defining statement
7753 occurs before the extension when optimizing extensions with
7754 different source and destination hard registers.
7756 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
7759 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
7760 vtables into the type inheritance graph.
7762 2014-01-10 Jakub Jelinek <jakub@redhat.com>
7764 PR rtl-optimization/59754
7765 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
7766 modes in the REGNO != REGNO case.
7768 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7770 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
7772 2014-01-10 Jakub Jelinek <jakub@redhat.com>
7774 PR tree-optimization/59745
7775 * tree-predcom.c (tree_predictive_commoning_loop): Call
7776 free_affine_expand_cache if giving up because components is NULL.
7778 * target-globals.c (save_target_globals): Allocate < 4KB structs using
7779 GC in payload of target_globals struct instead of allocating them on
7780 the heap and the larger structs separately using GC.
7781 * target-globals.h (struct target_globals): Make regs, hard_regs,
7782 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
7783 of GTY((skip)) and change type to void *.
7784 (reset_target_globals): Cast loads from those fields to corresponding
7787 2014-01-10 Steve Ellcey <sellcey@mips.com>
7790 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
7791 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
7792 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
7794 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
7797 * aarch64-modes.def (CC_Zmode): New flags mode.
7798 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
7799 represents an equality.
7800 (aarch64_get_condition_code): Handle CC_Zmode.
7801 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
7803 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7805 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
7806 extraction in good case.
7808 2014-01-10 Richard Biener <rguenther@suse.de>
7810 PR tree-optimization/59374
7811 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
7812 checking after SLP discovery. Mark stmts not participating
7813 in any SLP instance properly.
7815 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7817 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
7818 when handling a SET rtx.
7820 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7822 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
7823 (cortex-a57): Likewise.
7824 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
7826 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7828 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
7829 non-iwmmxt builtins.
7831 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
7835 * ipa-devirt.c record_target_from_binfo): Take as argument
7836 stack of binfos and lookup matching one for virtual inheritance.
7837 (possible_polymorphic_call_targets_1): Update.
7839 2014-01-10 Huacai Chen <chenhc@lemote.com>
7841 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
7842 kernel strings for Loongson-2E/2F/3A.
7844 2014-01-10 Jakub Jelinek <jakub@redhat.com>
7847 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
7848 is_gimple_call before calling gimple_call_internal_p.
7850 2014-01-09 Steve Ellcey <sellcey@mips.com>
7852 * Makefile.in (TREE_FLOW_H): Remove.
7853 (TREE_SSA_H): Add file names from tree-flow.h.
7854 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
7855 * tree.h: Remove tree-flow.h reference.
7856 * hash-table.h: Remove tree-flow.h reference.
7857 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
7858 reference with tree-ssa-loop.h.
7860 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7862 * doc/invoke.texi: Add -maltivec={be,le} options, and document
7863 default element-order behavior for -maltivec.
7864 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
7865 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
7866 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
7867 when targeting big endian, at least for now.
7868 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
7870 2014-01-09 Jakub Jelinek <jakub@redhat.com>
7873 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
7874 var satisfies use_register_for_decl, just take into account type
7875 alignment, rather than decl alignment.
7877 PR tree-optimization/59622
7878 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
7879 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
7880 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
7881 Don't devirtualize for inplace at all. For targets.length () == 1,
7882 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
7884 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
7886 * config/i386/i386.md (cpu): Remove the unused btver1.
7888 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
7890 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
7892 2014-01-09 Jakub Jelinek <jakub@redhat.com>
7895 * tree-core.h (struct target_globals): New forward declaration.
7896 (struct tree_target_option): Add globals field.
7897 * tree.h (TREE_TARGET_GLOBALS): Define.
7898 (prepare_target_option_nodes_for_pch): New prototype.
7899 * target-globals.h (struct target_globals): Define even if
7901 * tree.c (prepare_target_option_node_for_pch,
7902 prepare_target_option_nodes_for_pch): New functions.
7903 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
7904 * config/i386/i386.c: Include target-globals.h.
7905 (ix86_set_current_function): Instead of doing target_reinit
7906 unconditionally, use save_target_globals_default_opts and
7907 restore_target_globals.
7909 2014-01-09 Richard Biener <rguenther@suse.de>
7911 PR tree-optimization/59715
7912 * tree-cfg.h (split_critical_edges): Declare.
7913 * tree-cfg.c (split_critical_edges): Export.
7914 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
7916 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
7918 * cfgexpand.c (expand_stack_vars): Optionally disable
7919 asan stack protection.
7920 (expand_used_vars): Likewise.
7921 (partition_stack_vars): Likewise.
7922 * asan.c (asan_emit_stack_protection): Optionally disable
7923 after return stack usage.
7924 (instrument_derefs): Optionally disable memory access instrumentation.
7925 (instrument_builtin_call): Likewise.
7926 (instrument_strlen_call): Likewise.
7927 (asan_protect_global): Optionally disable global variables protection.
7928 * doc/invoke.texi: Added doc for new options.
7929 * params.def: Added new options.
7930 * params.h: Likewise.
7932 2014-01-09 Jakub Jelinek <jakub@redhat.com>
7934 PR rtl-optimization/59724
7935 * ifcvt.c (cond_exec_process_if_block): Don't call
7936 flow_find_head_matching_sequence with 0 longest_match.
7937 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
7938 non-active insns if !stop_after.
7939 (try_head_merge_bb): Revert 2014-01-07 changes.
7941 2014-01-08 Jeff Law <law@redhat.com>
7943 * ree.c (get_sub_rtx): New function, extracted from...
7944 (merge_def_and_ext): Here.
7945 (combine_reaching_defs): Use get_sub_rtx.
7947 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
7949 * cgraph.h (varpool_variable_node): Do not choke on null node.
7951 2014-01-08 Catherine Moore <clm@codesourcery.com>
7953 * config/mips/mips.md (simple_return): Attempt to use JRC
7955 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
7957 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
7959 PR rtl-optimization/59137
7960 * reorg.c (steal_delay_list_from_target): Call update_block for
7962 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
7964 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7966 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
7967 two duplicate entries.
7969 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
7972 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
7974 * config/mips/mips.c (mips_truncated_op_cost): New function.
7975 (mips_rtx_costs): Adjust test for BADDU.
7976 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
7978 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
7980 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
7981 (*baddu_si): ...this new pattern.
7983 2014-01-08 Jakub Jelinek <jakub@redhat.com>
7986 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
7988 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
7991 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
7993 (expand_expr, expand_normal): Adjust.
7994 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
7995 inner_reference_p. Use inner_reference_p to expand inner references.
7996 (store_expr): Adjust.
7997 * cfgexpand.c (expand_call_stmt): Adjust.
7999 2014-01-08 Rong Xu <xur@google.com>
8001 * gcov-io.c (gcov_var): Move from gcov-io.h.
8002 (gcov_position): Ditto.
8003 (gcov_is_error): Ditto.
8004 (gcov_rewrite): Ditto.
8005 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
8006 only part to libgcc/libgcov.h.
8008 2014-01-08 Marek Polacek <polacek@redhat.com>
8011 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
8013 2014-01-08 Marek Polacek <polacek@redhat.com>
8016 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
8018 2014-01-08 Jakub Jelinek <jakub@redhat.com>
8020 PR rtl-optimization/59649
8021 * stor-layout.c (get_mode_bounds): For BImode return
8022 0 and STORE_FLAG_VALUE.
8024 2014-01-08 Richard Biener <rguenther@suse.de>
8027 * gimple.h (is_gimple_builtin_call): Remove.
8028 (gimple_builtin_call_types_compatible_p): New.
8029 (gimple_call_builtin_p): New overload.
8030 * gimple.c (is_gimple_builtin_call): Remove.
8031 (validate_call): Rename to ...
8032 (gimple_builtin_call_types_compatible_p): ... this and export. Also
8034 (validate_type): New static function.
8035 (gimple_call_builtin_p): New overload and adjust.
8036 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
8037 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
8038 (gimple_fold_stmt_to_constant_1): Likewise.
8039 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
8041 2014-01-08 Richard Biener <rguenther@suse.de>
8044 * gimplify.c (gimplify_expr): Gimplify register-register type
8045 VIEW_CONVERT_EXPRs to separate stmts.
8047 2014-01-07 Jeff Law <law@redhat.com>
8050 * ree.c (combine_set_extension): Handle case where source
8051 and destination registers in an extension insn are different.
8052 (combine_reaching_defs): Allow source and destination registers
8053 in extension to be different under limited circumstances.
8054 (add_removable_extension): Remove restriction that the
8055 source and destination registers in the extension are the same.
8056 (find_and_remove_re): Emit a copy from the extension's
8057 destination to its source after the defining insn if
8058 the source and destination registers are different.
8061 * ifcvt.c (merge_if_block): If we are merging a block with more than
8062 one successor with a block with no successors, remove any BARRIER
8063 after the second block.
8065 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
8067 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
8069 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
8072 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
8073 for 14-bit register offsets when INT14_OK_STRICT is false.
8075 2014-01-07 Roland Stigge <stigge@antcom.de>
8076 Michael Meissner <meissner@linux.vnet.ibm.com>
8079 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
8080 Only check TFmode for SPE constants. Don't check TImode or TDmode.
8082 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
8084 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
8087 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
8089 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
8090 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
8091 rtx is const0_rtx or not.
8093 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
8096 * target-globals.c (save_target_globals): Remove this_fn_optab
8098 * toplev.c: Include optabs.h.
8099 (target_reinit): Temporarily restore the global options if another
8100 set of options are in force.
8102 2014-01-07 Jakub Jelinek <jakub@redhat.com>
8104 PR rtl-optimization/58668
8105 * cfgcleanup.c (flow_find_cross_jump): Don't count
8106 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
8107 to determine what is counted.
8108 (flow_find_head_matching_sequence): Use active_insn_p to determine
8110 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
8112 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
8113 determine what is counted.
8115 PR tree-optimization/59643
8116 * tree-predcom.c (split_data_refs_to_components): If one dr is
8117 read and one write, determine_offset fails and the write isn't
8118 in the bad component, just put the read into the bad component.
8120 2014-01-07 Mike Stump <mikestump@comcast.net>
8121 Jakub Jelinek <jakub@redhat.com>
8124 * tree-core.h (struct tree_optimization_option): Change optabs
8125 type from unsigned char * to void *.
8126 * optabs.c (init_tree_optimization_optabs): Adjust
8127 TREE_OPTIMIZATION_OPTABS initialization.
8129 2014-01-06 Jakub Jelinek <jakub@redhat.com>
8132 * config/i386/i386.h (struct machine_function): Add
8133 no_drap_save_restore field.
8134 * config/i386/i386.c (ix86_save_reg): Use
8135 !cfun->machine->no_drap_save_restore instead of
8136 crtl->stack_realign_needed.
8137 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
8138 this function clears frame_pointer_needed. Set
8139 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
8140 and DRAP reg is needed.
8142 2014-01-06 Marek Polacek <polacek@redhat.com>
8145 * doc/implement-c.texi: Mention that other integer types are
8146 permitted as bit-field types in strictly conforming mode.
8148 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
8150 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
8153 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
8155 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
8157 2014-01-06 Martin Jambor <mjambor@suse.cz>
8160 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
8162 * ipa-prop.c (ipa_print_node_params): Fix indentation.
8164 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
8168 * var-tracking.c (add_stores): Preserve the value of the source even if
8169 we don't record the store.
8171 2014-01-06 Terry Guo <terry.guo@arm.com>
8173 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
8175 2014-01-05 Iain Sandoe <iain@codesourcery.com>
8178 * config/darwin.c (darwin_function_section): Adjust return values to
8179 correspond to optimisation changes made in r206070.
8181 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
8183 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
8184 from prefetch_block tune setting.
8185 (nocona_cost): Correct size of prefetch block to 64.
8187 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
8189 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
8190 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
8191 used to save the static chain register in the computation of the offset
8192 from which the FP registers need to be restored.
8194 2014-01-04 Jakub Jelinek <jakub@redhat.com>
8196 PR tree-optimization/59519
8197 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
8198 ICE if get_current_def (current_new_name) is already non-NULL, as long
8199 as it is a phi result of some other phi in *new_exit_bb that has
8202 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
8203 or vmovdqu* for misaligned_operand.
8204 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
8205 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
8206 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
8207 aligned_mem for AVX512F masked aligned load and store builtins and for
8210 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
8212 PR tree-optimization/59651
8213 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
8214 Address range for negative step should be added by TYPE_SIZE_UNIT.
8216 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
8218 * config/m68k/m68k.c (handle_move_double): Handle pushes with
8219 overlapping registers also for registers other than the stack pointer.
8221 2014-01-03 Marek Polacek <polacek@redhat.com>
8224 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
8227 2014-01-03 Jakub Jelinek <jakub@redhat.com>
8230 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
8233 * config/i386/i386.md (MODE_SIZE): New mode attribute.
8234 (push splitter): Use <P:MODE_SIZE> instead of
8235 GET_MODE_SIZE (<P:MODE>mode).
8236 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
8237 (mov -1, reg peephole2): Likewise.
8238 * config/i386/sse.md (*mov<mode>_internal,
8239 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
8240 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
8241 *<code><mode>3, *andnot<mode>3<mask_name>,
8242 <mask_codefor><code><mode>3<mask_name>): Likewise.
8243 * config/i386/subst.md (mask_mode512bit_condition,
8244 sd_mask_mode512bit_condition): Likewise.
8246 2014-01-02 Xinliang David Li <davidxl@google.com>
8248 PR tree-optimization/59303
8249 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
8250 (dump_predicates): Better output format.
8251 (pred_equal_p): New function.
8252 (is_neq_relop_p): Ditto.
8253 (is_neq_zero_form_p): Ditto.
8254 (pred_expr_equal_p): Ditto.
8255 (pred_neg_p): Ditto.
8256 (simplify_pred): Ditto.
8257 (simplify_preds_2): Ditto.
8258 (simplify_preds_3): Ditto.
8259 (simplify_preds_4): Ditto.
8260 (simplify_preds): Ditto.
8262 (push_to_worklist): Ditto.
8263 (get_pred_info_from_cmp): Ditto.
8264 (is_degenerated_phi): Ditto.
8265 (normalize_one_pred_1): Ditto.
8266 (normalize_one_pred): Ditto.
8267 (normalize_one_pred_chain): Ditto.
8268 (normalize_preds): Ditto.
8269 (normalize_cond_1): Remove function.
8270 (normalize_cond): Ditto.
8271 (is_gcond_subset_of): Ditto.
8272 (is_subset_of_any): Ditto.
8273 (is_or_set_subset_of): Ditto.
8274 (is_and_set_subset_of): Ditto.
8275 (is_norm_cond_subset_of): Ditto.
8276 (pred_chain_length_cmp): Ditto.
8277 (convert_control_dep_chain_into_preds): Type change.
8278 (find_predicates): Ditto.
8279 (find_def_preds): Ditto.
8280 (destroy_predicates_vecs): Ditto.
8281 (find_matching_predicates_in_rest_chains): Ditto.
8282 (use_pred_not_overlap_with_undef_path_pred): Ditto.
8283 (is_pred_expr_subset): Ditto.
8284 (is_pred_chain_subset_of): Ditto.
8285 (is_included_in): Ditto.
8286 (is_superset_of): Ditto.
8288 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8290 Update copyright years.
8292 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8294 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
8295 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
8296 config/arc/arc.md, config/arc/arc.opt,
8297 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
8298 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
8299 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
8300 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
8301 config/linux-protos.h, config/linux.c, config/winnt-c.c,
8302 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
8303 vtable-verify.c, vtable-verify.h: Use the standard form for the
8306 2014-01-02 Tobias Burnus <burnus@net-b.de>
8308 * gcc.c (process_command): Update copyright notice dates.
8309 * gcov-dump.c: Ditto.
8311 * doc/cpp.texi: Bump @copying's copyright year.
8312 * doc/cppinternals.texi: Ditto.
8313 * doc/gcc.texi: Ditto.
8314 * doc/gccint.texi: Ditto.
8315 * doc/gcov.texi: Ditto.
8316 * doc/install.texi: Ditto.
8317 * doc/invoke.texi: Ditto.
8319 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8321 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
8323 2014-01-01 Jakub Jelinek <jakub@redhat.com>
8325 * config/i386/sse.md (*mov<mode>_internal): Guard
8326 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
8328 PR rtl-optimization/59647
8329 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
8330 new_rtx into UNSIGNED_FLOAT rtxes.
8332 Copyright (C) 2014 Free Software Foundation, Inc.
8334 Copying and distribution of this file, with or without modification,
8335 are permitted in any medium without royalty provided the copyright
8336 notice and this notice are preserved.