1 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
3 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
4 * configure: Regenerate.
5 * config/sparc/sparc.opt (muser-mode): New option.
6 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
8 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
9 * doc/invoke.texi (SPARC options): Document -muser-mode.
11 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
13 * cselib.c (find_slot_memmode): Delete.
14 (cselib_hasher): Change compare_type to a struct.
15 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
17 (preserve_constants_and_equivs): Adjust for new compare_type.
18 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
19 (wrap_constant): Delete.
20 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
22 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
24 * doc/install.texi (Building with profile feedback): Remove
27 2014-04-26 Tom de Vries <tom@codesourcery.com>
29 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
32 2014-04-25 Cary Coutant <ccoutant@google.com>
35 * dwarf2out.c (should_move_die_to_comdat): A type definition
36 can contain a subprogram definition, but don't move it to a
38 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
39 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
41 (clone_tree_hash): Rename to...
42 (clone_tree_partial): ...this; change callers. Copy
43 DW_TAG_subprogram DIEs as declarations.
44 (copy_decls_walk): Don't copy children of a declaration into a
47 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
50 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
53 2014-04-25 Jiong Wang <jiong.wang@arm.com>
55 * config/arm/predicates.md (call_insn_operand): Add long_call check.
56 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
58 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
61 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
63 * config/arm/arm.c (arm_cortex_a8_tune): Initialise
66 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
68 PR tree-optimization/60930
69 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
70 creating a multiply candidate by folding two constant
71 multiplicands when the result overflows.
73 2014-04-25 Jakub Jelinek <jakub@redhat.com>
75 PR tree-optimization/60960
76 * tree-vect-generic.c (expand_vector_operation): Only call
77 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
79 2014-04-25 Tom de Vries <tom@codesourcery.com>
81 * expr.c (clobber_reg_mode): New function.
82 * expr.h (clobber_reg): New function.
84 2014-04-25 Tom de Vries <tom@codesourcery.com>
86 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
89 2014-04-25 Radovan Obradovic <robradovic@mips.com>
90 Tom de Vries <tom@codesourcery.com>
92 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
94 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
95 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
96 new argument to find_all_hard_reg_sets call.
98 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
100 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
101 Use HOST_WIDE_INT_C for mask literal.
102 (aarch_rev16_shleft_mask_imm_p): Likewise.
104 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
107 * config/sparc/sparc.md (ashlsi3_extend): Delete.
109 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
111 PR preprocessor/56540
112 * config/i386/i386-c.c (ix86_target_macros): Define
113 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
115 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
117 * configure.ac (tga_func): Remove.
118 (LIB_TLS_SPEC): Remove.
119 * configure: Regenerate.
120 * config.in: Regenerate.
121 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
123 2014-04-25 Richard Biener <rguenther@suse.de>
126 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
127 call stmt use/clobber sets during stmt walk instead of
128 walking the possibly incomplete set of caller edges.
130 2014-04-25 Richard Biener <rguenther@suse.de>
133 * passes.c (apply_ipa_transforms): Inline into only caller ...
134 (execute_one_pass): ... here. Properly bring in function
135 bodies for nodes we want to apply IPA transforms to.
137 2014-04-24 Cong Hou <congh@google.com>
139 PR tree-optimization/60896
140 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
141 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
142 (vect_mark_pattern_stmts): Set the def type of all statements in
143 PATTERN_DEF_SEQ as vect_internal_def.
145 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
147 * doc/extend.texi (PowerPC Built-in Functions): Document new
148 powerpc extended divide, bcd, pack/unpack 128-bit, builtin
150 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
152 * config/rs6000/predicates.md (const_0_to_3_operand): New
153 predicate to match 0..3 integer constants.
155 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
156 to support adding miscellaneous builtin functions.
157 (BU_DFP_MISC_2): Likewise.
158 (BU_P7_MISC_1): Likewise.
159 (BU_P7_MISC_2): Likewise.
160 (BU_P8V_MISC_3): Likewise.
161 (BU_MISC_1): Likewise.
162 (BU_MISC_2): Likewise.
163 (DIVWE): Add extended divide builtin functions.
171 (DXEX): Add decimal floating-point builtin functions.
183 (CDTBCD): Add new BCD builtin functions.
187 (BCDADD_LT): Likewise.
188 (BCDADD_EQ): Likewise.
189 (BCDADD_GT): Likewise.
190 (BCDADD_OV): Likewise.
192 (BCDSUB_LT): Likewise.
193 (BCDSUB_EQ): Likewise.
194 (BCDSUB_GT): Likewise.
195 (BCDSUB_OV): Likewise.
196 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
197 (UNPACK_TD): Likewise.
199 (UNPACK_TF): Likewise.
200 (UNPACK_TF_0): Likewise.
201 (UNPACK_TF_1): Likewise.
202 (PACK_V1TI): Likewise.
203 (UNPACK_V1TI): Likewise.
205 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
206 support for decimal floating point builtin functions.
207 (rs6000_expand_ternop_builtin): Add checks for the new builtin
208 functions that take constant arguments.
209 (rs6000_invalid_builtin): Add decimal floating point builtin
211 (rs6000_init_builtins): Setup long double, _Decimal64, and
212 _Decimal128 types for new builtin functions.
213 (builtin_function_type): Set the unsigned flags appropriately for
214 the new builtin functions.
215 (rs6000_opt_masks): Add support for decimal floating point builtin
218 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
219 floating point builtin functions.
220 (RS6000_BTM_COMMON): Likewise.
221 (RS6000_BTI_long_double): Likewise.
222 (RS6000_BTI_dfloat64): Likewise.
223 (RS6000_BTI_dfloat128): Likewise.
224 (long_double_type_internal_node): Likewise.
225 (dfloat64_type_internal_node): Likewise.
226 (dfloat128_type_internal_node): Likewise.
228 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
229 2.07 bcd arithmetic instructions.
230 (UNSPEC_BCDSUB): Likewise.
231 (UNSPEC_BCD_OVERFLOW): Likewise.
232 (UNSPEC_BCD_ADD_SUB): Likewise.
233 (bcd_add_sub): Likewise.
234 (BCD_TEST): Likewise.
235 (bcd<bcd_add_sub>): Likewise.
236 (bcd<bcd_add_sub>_test): Likewise.
237 (bcd<bcd_add_sub>_test2): Likewise.
238 (bcd<bcd_add_sub>_<code>): Likewise.
239 (peephole2 for combined bcd ops): Likewise.
241 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
242 decimal floating point builtin functions.
243 (UNSPEC_DENBCD): Likewise.
244 (UNSPEC_DXEX): Likewise.
245 (UNSPEC_DIEX): Likewise.
246 (UNSPEC_DSCLI): Likewise.
247 (UNSPEC_DSCRI): Likewise.
248 (D64_D128): Likewise.
249 (dfp_suffix): Likewise.
250 (dfp_ddedpd_<mode>): Likewise.
251 (dfp_denbcd_<mode>): Likewise.
252 (dfp_dxex_<mode>): Likewise.
253 (dfp_diex_<mode>): Likewise.
254 (dfp_dscli_<mode>): Likewise.
255 (dfp_dscri_<mode>): Likewise.
257 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
259 (UNSPEC_CDTBCD): Likewise.
260 (UNSPEC_CBCDTD): Likewise.
261 (UNSPEC_DIVE): Add support for new extended divide builtin
263 (UNSPEC_DIVEO): Likewise.
264 (UNSPEC_DIVEU): Likewise.
265 (UNSPEC_DIVEUO): Likewise.
266 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
267 pack/unpack 128-bit types.
268 (UNSPEC_PACK_128BIT): Likewise.
269 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
270 (udiv<mode>3): Use idiv_ldiv mode attribute.
271 (div<mode>3): Likewise.
272 (addg6s): Add new BCD builtin functions.
275 (UNSPEC_DIV_EXTEND): Add support for new extended divide
277 (div_extend): Likewise.
278 (div<div_extend>_<mode>"): Likewise.
279 (FP128_64): Add support for new builtin functions to pack/unpack
281 (unpack<mode>): Likewise.
282 (unpacktf_0): Likewise.
283 (unpacktf_1): Likewise.
284 (unpack<mode>_dm): Likewise.
285 (unpack<mode>_nodm): Likewise.
286 (pack<mode>): Likewise.
287 (unpackv1ti): Likewise.
288 (packv1ti): Likewise.
290 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
292 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
295 2014-04-24 Jakub Jelinek <jakub@redhat.com>
297 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
298 * gimplify.c (omp_is_private): Change last argument's type to int.
299 Only diagnose lastprivate if the simd argument is 1, only diagnose
300 linear if the simd argument is 2.
301 (gimplify_omp_for): Adjust omp_is_private callers. When adding
302 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
303 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
304 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
305 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
306 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
307 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
308 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
309 * tree-nested.c (convert_nonlocal_omp_clauses,
310 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
312 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
315 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
318 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
320 * flag-types.h (enum ivar_visibility): Add.
322 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
324 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
327 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
329 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
331 2014-04-24 Radovan Obradovic <robradovic@mips.com>
332 Tom de Vries <tom@codesourcery.com>
334 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
335 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
337 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
338 * emit-rtl.c (try_split): Same.
340 2014-04-24 Radovan Obradovic <robradovic@mips.com>
341 Tom de Vries <tom@codesourcery.com>
343 * common.opt (fuse-caller-save): New option.
345 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
347 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of elements
350 2014-04-24 Richard Biener <rguenther@suse.de>
352 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
353 during TER and instead use the sepops interface for expanding
354 non-GIMPLE_SINGLE_RHS.
356 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
358 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
359 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
361 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
363 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
364 assembler 64-bit option.
365 * configure: Regenerate.
367 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
369 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
370 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
371 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
372 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
373 (TARGET_CRYPTO): Take TARGET_SIMD into account.
375 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
377 * config/aarch64/aarch64-builtins.c
378 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
379 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
380 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
381 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
383 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
384 (Vrevsuff): New mode attribute.
386 2014-04-24 Terry Guo <terry.guo@arm.com>
388 * config/arm/arm.h (machine_function): Define variable
389 after_arm_reorg here.
390 * config/arm/arm.c (after_arm_reorg): Remove the definition.
391 (arm_split_constant): Update the way to access variable
394 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
396 2014-04-23 Tom de Vries <tom@codesourcery.com>
398 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
400 2014-04-23 David Malcolm <dmalcolm@redhat.com>
402 * is-a.h: Update comments to reflect the following changes to the
403 "pointerness" of the API, making the template parameter match the
404 return type, allowing use of is-a.h with typedefs of pointers.
405 (is_a_helper::cast): Return a T rather then a pointer to a T, so
406 that the return type matches the parameter to the is_a_helper.
408 (dyn_cast): Likewise.
410 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
411 pointer from the is-a.h API.
413 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
414 (is_a_helper <cgraph_node *>::test): ...this, matching change to
416 (is_a_helper <varpool_node>::test): Likewise, convert to...
417 (is_a_helper <varpool_node *>::test): ...this.
419 (varpool_first_variable): Update for removal of implicit pointer
421 (varpool_next_variable): Likewise.
422 (varpool_first_static_initializer): Likewise.
423 (varpool_next_static_initializer): Likewise.
424 (varpool_first_defined_variable): Likewise.
425 (varpool_next_defined_variable): Likewise.
426 (cgraph_first_defined_function): Likewise.
427 (cgraph_next_defined_function): Likewise.
428 (cgraph_first_function): Likewise.
429 (cgraph_next_function): Likewise.
430 (cgraph_first_function_with_gimple_body): Likewise.
431 (cgraph_next_function_with_gimple_body): Likewise.
432 (cgraph_alias_target): Likewise.
433 (varpool_alias_target): Likewise.
434 (cgraph_function_or_thunk_node): Likewise.
435 (varpool_variable_node): Likewise.
436 (symtab_real_symbol_p): Likewise.
437 * cgraphunit.c (referred_to_p): Likewise.
438 (analyze_functions): Likewise.
439 (handle_alias_pairs): Likewise.
440 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
441 * gimple-ssa.h (gimple_vuse_op): Likewise.
442 (gimple_vdef_op): Likewise.
443 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
444 * gimple.c (gimple_build_asm_1): Likewise.
445 (gimple_build_try): Likewise.
446 (gimple_build_resx): Likewise.
447 (gimple_build_eh_dispatch): Likewise.
448 (gimple_build_omp_for): Likewise.
449 (gimple_omp_for_set_clauses): Likewise.
451 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
452 (is_a_helper <gimple_statement_asm *>::test): ...this.
453 (is_a_helper <gimple_statement_bind>::test): Convert to...
454 (is_a_helper <gimple_statement_bind *>::test): ...this.
455 (is_a_helper <gimple_statement_call>::test): Convert to...
456 (is_a_helper <gimple_statement_call *>::test): ...this.
457 (is_a_helper <gimple_statement_catch>::test): Convert to...
458 (is_a_helper <gimple_statement_catch *>::test): ...this.
459 (is_a_helper <gimple_statement_resx>::test): Convert to...
460 (is_a_helper <gimple_statement_resx *>::test): ...this.
461 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
462 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
463 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
464 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
465 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
466 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
467 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
468 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
469 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
470 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
471 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
472 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
473 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
474 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
475 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
476 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
477 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
478 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
479 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
480 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
481 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
482 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
483 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
484 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
485 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
486 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
487 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
488 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
489 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
490 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
491 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
492 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
493 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
494 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
495 (is_a_helper <gimple_statement_phi>::test): Convert to...
496 (is_a_helper <gimple_statement_phi *>::test): ...this.
497 (is_a_helper <gimple_statement_transaction>::test): Convert to...
498 (is_a_helper <gimple_statement_transaction *>::test): ...this.
499 (is_a_helper <gimple_statement_try>::test): Convert to...
500 (is_a_helper <gimple_statement_try *>::test): ...this.
501 (is_a_helper <gimple_statement_wce>::test): Convert to...
502 (is_a_helper <gimple_statement_wce *>::test): ...this.
503 (is_a_helper <const gimple_statement_asm>::test): Convert to...
504 (is_a_helper <const gimple_statement_asm *>::test): ...this.
505 (is_a_helper <const gimple_statement_bind>::test): Convert to...
506 (is_a_helper <const gimple_statement_bind *>::test): ...this.
507 (is_a_helper <const gimple_statement_call>::test): Convert to...
508 (is_a_helper <const gimple_statement_call *>::test): ...this.
509 (is_a_helper <const gimple_statement_catch>::test): Convert to...
510 (is_a_helper <const gimple_statement_catch *>::test): ...this.
511 (is_a_helper <const gimple_statement_resx>::test): Convert to...
512 (is_a_helper <const gimple_statement_resx *>::test): ...this.
513 (is_a_helper <const gimple_statement_eh_dispatch>::test):
515 (is_a_helper <const gimple_statement_eh_dispatch *>::test):
517 (is_a_helper <const gimple_statement_eh_filter>::test): Convert
519 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
520 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
522 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
524 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
526 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
528 (is_a_helper <const gimple_statement_omp_return>::test): Convert
530 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
531 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
533 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
534 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
536 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
537 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
538 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
539 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert
541 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
542 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
544 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
545 (is_a_helper <const gimple_statement_omp_target>::test): Convert
547 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
548 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
550 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
551 (is_a_helper <const gimple_statement_omp_single>::test): Convert
553 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
554 (is_a_helper <const gimple_statement_omp_teams>::test): Convert
556 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
557 (is_a_helper <const gimple_statement_omp_task>::test): Convert
559 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
560 (is_a_helper <const gimple_statement_phi>::test): Convert to...
561 (is_a_helper <const gimple_statement_phi *>::test): ...this.
562 (is_a_helper <const gimple_statement_transaction>::test): Convert
564 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
565 (is_a_helper <const gimple_statement_with_ops>::test): Convert
567 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
568 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
569 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
570 (is_a_helper <const gimple_statement_with_memory_ops>::test):
572 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
574 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert
576 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
578 (gimple_use_ops): Update for removal of implicit pointer from the
580 (gimple_set_use_ops): Likewise.
581 (gimple_vuse): Likewise.
582 (gimple_vdef): Likewise.
583 (gimple_vuse_ptr): Likewise.
584 (gimple_vdef_ptr): Likewise.
585 (gimple_set_vuse): Likewise.
586 (gimple_set_vdef): Likewise.
587 (gimple_omp_return_set_lhs): Likewise.
588 (gimple_omp_return_lhs): Likewise.
589 (gimple_omp_return_lhs_ptr): Likewise.
590 (gimple_call_fntype): Likewise.
591 (gimple_call_set_fntype): Likewise.
592 (gimple_call_set_internal_fn): Likewise.
593 (gimple_call_use_set): Likewise.
594 (gimple_call_clobber_set): Likewise.
595 (gimple_bind_vars): Likewise.
596 (gimple_bind_set_vars): Likewise.
597 (gimple_bind_body_ptr): Likewise.
598 (gimple_bind_set_body): Likewise.
599 (gimple_bind_add_stmt): Likewise.
600 (gimple_bind_block): Likewise.
601 (gimple_bind_set_block): Likewise.
602 (gimple_asm_ninputs): Likewise.
603 (gimple_asm_noutputs): Likewise.
604 (gimple_asm_nclobbers): Likewise.
605 (gimple_asm_nlabels): Likewise.
606 (gimple_asm_input_op): Likewise.
607 (gimple_asm_input_op_ptr): Likewise.
608 (gimple_asm_output_op): Likewise.
609 (gimple_asm_output_op_ptr): Likewise.
610 (gimple_asm_set_output_op): Likewise.
611 (gimple_asm_clobber_op): Likewise.
612 (gimple_asm_set_clobber_op): Likewise.
613 (gimple_asm_label_op): Likewise.
614 (gimple_asm_set_label_op): Likewise.
615 (gimple_asm_string): Likewise.
616 (gimple_catch_types): Likewise.
617 (gimple_catch_types_ptr): Likewise.
618 (gimple_catch_handler_ptr): Likewise.
619 (gimple_catch_set_types): Likewise.
620 (gimple_catch_set_handler): Likewise.
621 (gimple_eh_filter_types): Likewise.
622 (gimple_eh_filter_types_ptr): Likewise.
623 (gimple_eh_filter_failure_ptr): Likewise.
624 (gimple_eh_filter_set_types): Likewise.
625 (gimple_eh_filter_set_failure): Likewise.
626 (gimple_eh_must_not_throw_fndecl): Likewise.
627 (gimple_eh_must_not_throw_set_fndecl): Likewise.
628 (gimple_eh_else_n_body_ptr): Likewise.
629 (gimple_eh_else_e_body_ptr): Likewise.
630 (gimple_eh_else_set_n_body): Likewise.
631 (gimple_eh_else_set_e_body): Likewise.
632 (gimple_try_eval_ptr): Likewise.
633 (gimple_try_cleanup_ptr): Likewise.
634 (gimple_try_set_eval): Likewise.
635 (gimple_try_set_cleanup): Likewise.
636 (gimple_wce_cleanup_ptr): Likewise.
637 (gimple_wce_set_cleanup): Likewise.
638 (gimple_phi_capacity): Likewise.
639 (gimple_phi_num_args): Likewise.
640 (gimple_phi_result): Likewise.
641 (gimple_phi_result_ptr): Likewise.
642 (gimple_phi_set_result): Likewise.
643 (gimple_phi_arg): Likewise.
644 (gimple_phi_set_arg): Likewise.
645 (gimple_resx_region): Likewise.
646 (gimple_resx_set_region): Likewise.
647 (gimple_eh_dispatch_region): Likewise.
648 (gimple_eh_dispatch_set_region): Likewise.
649 (gimple_omp_critical_name): Likewise.
650 (gimple_omp_critical_name_ptr): Likewise.
651 (gimple_omp_critical_set_name): Likewise.
652 (gimple_omp_for_clauses): Likewise.
653 (gimple_omp_for_clauses_ptr): Likewise.
654 (gimple_omp_for_set_clauses): Likewise.
655 (gimple_omp_for_collapse): Likewise.
656 (gimple_omp_for_index): Likewise.
657 (gimple_omp_for_index_ptr): Likewise.
658 (gimple_omp_for_set_index): Likewise.
659 (gimple_omp_for_initial): Likewise.
660 (gimple_omp_for_initial_ptr): Likewise.
661 (gimple_omp_for_set_initial): Likewise.
662 (gimple_omp_for_final): Likewise.
663 (gimple_omp_for_final_ptr): Likewise.
664 (gimple_omp_for_set_final): Likewise.
665 (gimple_omp_for_incr): Likewise.
666 (gimple_omp_for_incr_ptr): Likewise.
667 (gimple_omp_for_set_incr): Likewise.
668 (gimple_omp_for_pre_body_ptr): Likewise.
669 (gimple_omp_for_set_pre_body): Likewise.
670 (gimple_omp_parallel_clauses): Likewise.
671 (gimple_omp_parallel_clauses_ptr): Likewise.
672 (gimple_omp_parallel_set_clauses): Likewise.
673 (gimple_omp_parallel_child_fn): Likewise.
674 (gimple_omp_parallel_child_fn_ptr): Likewise.
675 (gimple_omp_parallel_set_child_fn): Likewise.
676 (gimple_omp_parallel_data_arg): Likewise.
677 (gimple_omp_parallel_data_arg_ptr): Likewise.
678 (gimple_omp_parallel_set_data_arg): Likewise.
679 (gimple_omp_task_clauses): Likewise.
680 (gimple_omp_task_clauses_ptr): Likewise.
681 (gimple_omp_task_set_clauses): Likewise.
682 (gimple_omp_task_child_fn): Likewise.
683 (gimple_omp_task_child_fn_ptr): Likewise.
684 (gimple_omp_task_set_child_fn): Likewise.
685 (gimple_omp_task_data_arg): Likewise.
686 (gimple_omp_task_data_arg_ptr): Likewise.
687 (gimple_omp_task_set_data_arg): Likewise.
688 (gimple_omp_taskreg_clauses): Likewise.
689 (gimple_omp_taskreg_clauses_ptr): Likewise.
690 (gimple_omp_taskreg_set_clauses): Likewise.
691 (gimple_omp_taskreg_child_fn): Likewise.
692 (gimple_omp_taskreg_child_fn_ptr): Likewise.
693 (gimple_omp_taskreg_set_child_fn): Likewise.
694 (gimple_omp_taskreg_data_arg): Likewise.
695 (gimple_omp_taskreg_data_arg_ptr): Likewise.
696 (gimple_omp_taskreg_set_data_arg): Likewise.
697 (gimple_omp_task_copy_fn): Likewise.
698 (gimple_omp_task_copy_fn_ptr): Likewise.
699 (gimple_omp_task_set_copy_fn): Likewise.
700 (gimple_omp_task_arg_size): Likewise.
701 (gimple_omp_task_arg_size_ptr): Likewise.
702 (gimple_omp_task_set_arg_size): Likewise.
703 (gimple_omp_task_arg_align): Likewise.
704 (gimple_omp_task_arg_align_ptr): Likewise.
705 (gimple_omp_task_set_arg_align): Likewise.
706 (gimple_omp_single_clauses): Likewise.
707 (gimple_omp_single_clauses_ptr): Likewise.
708 (gimple_omp_single_set_clauses): Likewise.
709 (gimple_omp_target_clauses): Likewise.
710 (gimple_omp_target_clauses_ptr): Likewise.
711 (gimple_omp_target_set_clauses): Likewise.
712 (gimple_omp_target_child_fn): Likewise.
713 (gimple_omp_target_child_fn_ptr): Likewise.
714 (gimple_omp_target_set_child_fn): Likewise.
715 (gimple_omp_target_data_arg): Likewise.
716 (gimple_omp_target_data_arg_ptr): Likewise.
717 (gimple_omp_target_set_data_arg): Likewise.
718 (gimple_omp_teams_clauses): Likewise.
719 (gimple_omp_teams_clauses_ptr): Likewise.
720 (gimple_omp_teams_set_clauses): Likewise.
721 (gimple_omp_sections_clauses): Likewise.
722 (gimple_omp_sections_clauses_ptr): Likewise.
723 (gimple_omp_sections_set_clauses): Likewise.
724 (gimple_omp_sections_control): Likewise.
725 (gimple_omp_sections_control_ptr): Likewise.
726 (gimple_omp_sections_set_control): Likewise.
727 (gimple_omp_for_set_cond): Likewise.
728 (gimple_omp_for_cond): Likewise.
729 (gimple_omp_atomic_store_set_val): Likewise.
730 (gimple_omp_atomic_store_val): Likewise.
731 (gimple_omp_atomic_store_val_ptr): Likewise.
732 (gimple_omp_atomic_load_set_lhs): Likewise.
733 (gimple_omp_atomic_load_lhs): Likewise.
734 (gimple_omp_atomic_load_lhs_ptr): Likewise.
735 (gimple_omp_atomic_load_set_rhs): Likewise.
736 (gimple_omp_atomic_load_rhs): Likewise.
737 (gimple_omp_atomic_load_rhs_ptr): Likewise.
738 (gimple_omp_continue_control_def): Likewise.
739 (gimple_omp_continue_control_def_ptr): Likewise.
740 (gimple_omp_continue_set_control_def): Likewise.
741 (gimple_omp_continue_control_use): Likewise.
742 (gimple_omp_continue_control_use_ptr): Likewise.
743 (gimple_omp_continue_set_control_use): Likewise.
744 (gimple_transaction_body_ptr): Likewise.
745 (gimple_transaction_label): Likewise.
746 (gimple_transaction_label_ptr): Likewise.
747 (gimple_transaction_set_body): Likewise.
748 (gimple_transaction_set_label): Likewise.
750 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
751 * ipa-inline-analysis.c (inline_write_summary): Likewise.
752 * ipa-ref.c (ipa_record_reference): Likewise.
753 * ipa-reference.c (analyze_function): Likewise.
754 (ipa_reference_write_optimization_summary): Likewise.
755 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
756 (address_taken_from_non_vtable_p): Likewise.
757 (comdat_can_be_unshared_p_1): Likewise.
758 * lto-cgraph.c (lto_output_ref): Likewise.
759 (add_references): Likewise.
760 (compute_ltrans_boundary): Likewise.
761 (output_symtab): Likewise.
762 (input_ref): Likewise.
763 (input_cgraph_1): Likewise.
764 (output_cgraph_opt_summary): Likewise.
765 * lto-streamer-out.c (lto_output): Likewise.
766 (output_symbol_p): Likewise.
767 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
768 (lsei_start_function_in_partition): Likewise.
769 (lsei_next_variable_in_partition): Likewise.
770 (lsei_start_variable_in_partition): Likewise.
771 * symtab.c (insert_to_assembler_name_hash): Likewise.
772 (unlink_from_assembler_name_hash): Likewise.
773 (symtab_unregister_node): Likewise.
774 (symtab_remove_node): Likewise.
775 (dump_symtab_node): Likewise.
776 (verify_symtab_base): Likewise.
777 (verify_symtab_node): Likewise.
778 (symtab_make_decl_local): Likewise.
779 (symtab_alias_ultimate_target): Likewise.
780 (symtab_resolve_alias): Likewise.
781 (symtab_get_symbol_partitioning_class): Likewise.
782 * tree-phinodes.c (allocate_phi_node): Likewise.
783 (reserve_phi_args_for_new_edge): Likewise.
784 (remove_phi_args): Likewise.
785 * varpool.c (varpool_node_for_asm): Likewise.
786 (varpool_remove_unreferenced_decls): Likewise.
788 2014-04-23 Jeff Law <law@redhat.com>
790 PR tree-optimization/60902
791 * tree-ssa-threadedge.c
792 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
793 invalidate outputs from statements that do not produce useful
794 outputs for threading.
796 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
798 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
799 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
800 machine descriptions for Stack Smashing Protector.
802 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
804 * aarch64.md (<optab>_rol<mode>3): New pattern.
805 (<optab>_rolsi3_uxtw): Likewise.
806 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
808 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
810 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
811 (arm_cortex_a12_tune): Likewise.
813 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
815 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
817 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
819 * config/arm/arm.md (arm_rev16si2): New pattern.
820 (arm_rev16si2_alt): Likewise.
821 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
823 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
825 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
826 (rev16<mode>2_alt): Likewise.
827 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
828 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
829 (aarch_rev16_shleft_mask_imm_p): Likewise.
830 (aarch_rev16_p_1): Likewise.
831 (aarch_rev16_p): Likewise.
832 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
833 (aarch_rev16_shright_mask_imm_p): Likewise.
834 (aarch_rev16_shleft_mask_imm_p): Likewise.
836 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
838 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
839 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
841 (cortex_a53_extra_costs): Likewise.
842 (cortex_a57_extra_costs): Likewise.
843 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
844 (cortexa7_extra_costs): Likewise.
845 (cortexa8_extra_costs): Likewise.
846 (cortexa12_extra_costs): Likewise.
847 (cortexa15_extra_costs): Likewise.
848 (v7m_extra_costs): Likewise.
849 (arm_new_rtx_costs): Handle BSWAP.
851 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
853 * config/arm/arm.c (cortexa8_extra_costs): New table.
854 (arm_cortex_a8_tune): New tuning struct.
855 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
857 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
859 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
861 2014-04-23 Richard Biener <rguenther@suse.de>
863 * Makefile.in (OBJS): Remove loop-unswitch.o.
864 * loop-unswitch.c: Delete.
865 * tree-pass.h (make_pass_rtl_unswitch): Remove.
866 * passes.def (pass_rtl_unswitch): Likewise.
867 * loop-init.c (gate_rtl_unswitch): Likewise.
868 (rtl_unswitch): Likewise.
869 (pass_data_rtl_unswitch): Likewise.
870 (pass_rtl_unswitch): Likewise.
871 (make_pass_rtl_unswitch): Likewise.
872 * rtl.h (reversed_condition): Likewise.
873 (compare_and_jump_seq): Likewise.
874 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
876 * loop-unroll.c (compare_and_jump_seq): Likewise.
878 2014-04-23 Richard Biener <rguenther@suse.de>
880 PR tree-optimization/60903
881 * tree-ssa-loop-im.c (analyze_memory_references): Remove
882 commented code block.
883 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
884 loop flags to newly created BBs and edges.
886 2014-04-23 Nick Clifton <nickc@redhat.com>
888 * config/msp430/msp430.c (msp430_handle_option): Move function
890 (msp430_option_override): Simplify mcu and mcpu option handling.
891 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
892 support for -mhwmult command line option.
893 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
894 -mhwmult command line option.
895 (msp430_hwmult_enabled): Delete.
896 (msp43o_output_labelref): Add support for -mhwmult command line
898 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
899 (umulsidi3): Likewise.
900 * config/msp430/msp430.opt (mmcu): Add Report attribute.
901 (mcpu, mlarge, msmall): Likewise.
902 (mhwmult): New option.
903 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
905 (msp430_is_f5_mcu): Remove prototype.
906 (msp430_use_f5_series_hwmult): Add prototype.
907 * config/msp430/msp430-opts.h: New file.
908 * common/config/msp430: New directory.
909 * common/config/msp430/msp430-common.c: New file.
910 * config.gcc (msp430): Remove target_has_targetm_common.
911 * doc/invoke.texi: Document -mhwmult command line option.
913 2014-04-23 Nick Clifton <nickc@redhat.com>
915 * config/i386/cygwin.h (ENDFILE_SPEC): Include
916 default-manifest.o if it can be found in the search path.
917 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
919 2014-04-23 Terry Guo <terry.guo@arm.com>
921 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
923 2014-04-23 Richard Biener <rguenther@suse.de>
926 * tree-inline.c (declare_return_variable): Use mark_addressable.
928 2014-04-23 Richard Biener <rguenther@suse.de>
931 * loop-init.c (loop_optimizer_init): Make sure to apply
932 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
934 2014-04-22 Jakub Jelinek <jakub@redhat.com>
937 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
939 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
940 if flag_sanitize_undefined_trap_on_error.
941 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
942 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
943 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
944 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
945 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
946 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
947 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
948 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
949 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
950 * ubsan.c (ubsan_instrument_unreachable): Return
951 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
952 (ubsan_expand_null_ifn): Emit __builtin_trap ()
953 if flag_sanitize_undefined_trap_on_error and
954 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
955 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
956 instrument_bool_enum_load): Emit __builtin_trap () if
957 flag_sanitize_undefined_trap_on_error and
958 __builtin_handle_*_abort () if !flag_sanitize_recover.
959 * doc/invoke.texi (-fsanitize-recover,
960 -fsanitize-undefined-trap-on-error): Document.
962 2014-04-22 Christian Bruel <christian.bruel@st.com>
964 * config/sh/sh.md (mov<mode>): Replace movQIHI.
965 Force immediates to SImode.
967 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
969 * config/nios2/nios2.md (UNSPEC_ROUND): New.
971 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
972 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
973 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
974 (nios2_fpu_insn): Add entry for round.
975 (N2FPU_NO_ERRNO_P): Define.
976 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
978 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
980 2014-04-22 Richard Henderson <rth@redhat.com>
982 * config/aarch64/aarch64 (addti3, subti3): New expanders.
983 (add<GPI>3_compare0): Remove leading * from name.
984 (add<GPI>3_carryin): Likewise.
985 (sub<GPI>3_compare0): Likewise.
986 (sub<GPI>3_carryin): Likewise.
987 (<su_optab>mulditi3): New expander.
988 (multi3): New expander.
989 (madd<GPI>): Remove leading * from name.
991 2014-04-22 Martin Jambor <mjambor@suse.cz>
993 * cgraphclones.c (cgraph_function_versioning): Copy
994 ipa_transforms_to_apply instead of asserting it is empty.
996 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
999 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
1000 on count_exp to get mode.
1002 2014-04-22 Andrew Pinski <apinski@cavium.com>
1004 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
1005 Handle TLS for ILP32.
1006 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
1007 (tlsie_small_<mode>): this and handle PTR.
1008 (tlsie_small_sidi): New pattern.
1009 (tlsle_small): Change to an expand to handle ILP32.
1010 (tlsle_small_<mode>): New pattern.
1011 (tlsdesc_small): Rename to ...
1012 (tlsdesc_small_<mode>): this and handle PTR.
1014 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1016 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
1018 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1020 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
1021 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
1022 (aarch64_types_signed_poly_qualifiers): Likewise.
1023 (aarch64_types_unsigned_signed_qualifiers): Likewise.
1024 (aarch64_types_poly_signed_qualifiers): Likewise.
1025 (TYPES_REINTERP_SS): Type macro added.
1026 (TYPES_REINTERP_SU): Likewise.
1027 (TYPES_REINTERP_SP): Likewise.
1028 (TYPES_REINTERP_US): Likewise.
1029 (TYPES_REINTERP_PS): Likewise.
1030 (aarch64_fold_builtin): New expression folding added.
1031 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
1032 Declarations removed.
1033 (REINTERP_SS): Declarations added.
1034 (REINTERP_US): Likewise.
1035 (REINTERP_PS): Likewise.
1036 (REINTERP_SU): Likewise.
1037 (REINTERP_SP): Likewise.
1038 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
1039 (vreinterpretq_p8_f64): Likewise.
1040 (vreinterpret_p16_f64): Likewise.
1041 (vreinterpretq_p16_f64): Likewise.
1042 (vreinterpret_f32_f64): Likewise.
1043 (vreinterpretq_f32_f64): Likewise.
1044 (vreinterpret_f64_f32): Likewise.
1045 (vreinterpret_f64_p8): Likewise.
1046 (vreinterpret_f64_p16): Likewise.
1047 (vreinterpret_f64_s8): Likewise.
1048 (vreinterpret_f64_s16): Likewise.
1049 (vreinterpret_f64_s32): Likewise.
1050 (vreinterpret_f64_s64): Likewise.
1051 (vreinterpret_f64_u8): Likewise.
1052 (vreinterpret_f64_u16): Likewise.
1053 (vreinterpret_f64_u32): Likewise.
1054 (vreinterpret_f64_u64): Likewise.
1055 (vreinterpretq_f64_f32): Likewise.
1056 (vreinterpretq_f64_p8): Likewise.
1057 (vreinterpretq_f64_p16): Likewise.
1058 (vreinterpretq_f64_s8): Likewise.
1059 (vreinterpretq_f64_s16): Likewise.
1060 (vreinterpretq_f64_s32): Likewise.
1061 (vreinterpretq_f64_s64): Likewise.
1062 (vreinterpretq_f64_u8): Likewise.
1063 (vreinterpretq_f64_u16): Likewise.
1064 (vreinterpretq_f64_u32): Likewise.
1065 (vreinterpretq_f64_u64): Likewise.
1066 (vreinterpret_s64_f64): Likewise.
1067 (vreinterpretq_s64_f64): Likewise.
1068 (vreinterpret_u64_f64): Likewise.
1069 (vreinterpretq_u64_f64): Likewise.
1070 (vreinterpret_s8_f64): Likewise.
1071 (vreinterpretq_s8_f64): Likewise.
1072 (vreinterpret_s16_f64): Likewise.
1073 (vreinterpretq_s16_f64): Likewise.
1074 (vreinterpret_s32_f64): Likewise.
1075 (vreinterpretq_s32_f64): Likewise.
1076 (vreinterpret_u8_f64): Likewise.
1077 (vreinterpretq_u8_f64): Likewise.
1078 (vreinterpret_u16_f64): Likewise.
1079 (vreinterpretq_u16_f64): Likewise.
1080 (vreinterpret_u32_f64): Likewise.
1081 (vreinterpretq_u32_f64): Likewise.
1083 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1085 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
1086 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
1087 (vreinterpret_p8_s8): Likewise.
1088 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
1089 (vreinterpret_p8_s16): Likewise.
1090 (vreinterpret_p8_s32): Likewise.
1091 (vreinterpret_p8_s64): Likewise.
1092 (vreinterpret_p8_f32): Likewise.
1093 (vreinterpret_p8_u8): Likewise.
1094 (vreinterpret_p8_u16): Likewise.
1095 (vreinterpret_p8_u32): Likewise.
1096 (vreinterpret_p8_u64): Likewise.
1097 (vreinterpret_p8_p16): Likewise.
1098 (vreinterpretq_p8_s8): Likewise.
1099 (vreinterpretq_p8_s16): Likewise.
1100 (vreinterpretq_p8_s32): Likewise.
1101 (vreinterpretq_p8_s64): Likewise.
1102 (vreinterpretq_p8_f32): Likewise.
1103 (vreinterpretq_p8_u8): Likewise.
1104 (vreinterpretq_p8_u16): Likewise.
1105 (vreinterpretq_p8_u32): Likewise.
1106 (vreinterpretq_p8_u64): Likewise.
1107 (vreinterpretq_p8_p16): Likewise.
1108 (vreinterpret_p16_s8): Likewise.
1109 (vreinterpret_p16_s16): Likewise.
1110 (vreinterpret_p16_s32): Likewise.
1111 (vreinterpret_p16_s64): Likewise.
1112 (vreinterpret_p16_f32): Likewise.
1113 (vreinterpret_p16_u8): Likewise.
1114 (vreinterpret_p16_u16): Likewise.
1115 (vreinterpret_p16_u32): Likewise.
1116 (vreinterpret_p16_u64): Likewise.
1117 (vreinterpret_p16_p8): Likewise.
1118 (vreinterpretq_p16_s8): Likewise.
1119 (vreinterpretq_p16_s16): Likewise.
1120 (vreinterpretq_p16_s32): Likewise.
1121 (vreinterpretq_p16_s64): Likewise.
1122 (vreinterpretq_p16_f32): Likewise.
1123 (vreinterpretq_p16_u8): Likewise.
1124 (vreinterpretq_p16_u16): Likewise.
1125 (vreinterpretq_p16_u32): Likewise.
1126 (vreinterpretq_p16_u64): Likewise.
1127 (vreinterpretq_p16_p8): Likewise.
1128 (vreinterpret_f32_s8): Likewise.
1129 (vreinterpret_f32_s16): Likewise.
1130 (vreinterpret_f32_s32): Likewise.
1131 (vreinterpret_f32_s64): Likewise.
1132 (vreinterpret_f32_u8): Likewise.
1133 (vreinterpret_f32_u16): Likewise.
1134 (vreinterpret_f32_u32): Likewise.
1135 (vreinterpret_f32_u64): Likewise.
1136 (vreinterpret_f32_p8): Likewise.
1137 (vreinterpret_f32_p16): Likewise.
1138 (vreinterpretq_f32_s8): Likewise.
1139 (vreinterpretq_f32_s16): Likewise.
1140 (vreinterpretq_f32_s32): Likewise.
1141 (vreinterpretq_f32_s64): Likewise.
1142 (vreinterpretq_f32_u8): Likewise.
1143 (vreinterpretq_f32_u16): Likewise.
1144 (vreinterpretq_f32_u32): Likewise.
1145 (vreinterpretq_f32_u64): Likewise.
1146 (vreinterpretq_f32_p8): Likewise.
1147 (vreinterpretq_f32_p16): Likewise.
1148 (vreinterpret_s64_s8): Likewise.
1149 (vreinterpret_s64_s16): Likewise.
1150 (vreinterpret_s64_s32): Likewise.
1151 (vreinterpret_s64_f32): Likewise.
1152 (vreinterpret_s64_u8): Likewise.
1153 (vreinterpret_s64_u16): Likewise.
1154 (vreinterpret_s64_u32): Likewise.
1155 (vreinterpret_s64_u64): Likewise.
1156 (vreinterpret_s64_p8): Likewise.
1157 (vreinterpret_s64_p16): Likewise.
1158 (vreinterpretq_s64_s8): Likewise.
1159 (vreinterpretq_s64_s16): Likewise.
1160 (vreinterpretq_s64_s32): Likewise.
1161 (vreinterpretq_s64_f32): Likewise.
1162 (vreinterpretq_s64_u8): Likewise.
1163 (vreinterpretq_s64_u16): Likewise.
1164 (vreinterpretq_s64_u32): Likewise.
1165 (vreinterpretq_s64_u64): Likewise.
1166 (vreinterpretq_s64_p8): Likewise.
1167 (vreinterpretq_s64_p16): Likewise.
1168 (vreinterpret_u64_s8): Likewise.
1169 (vreinterpret_u64_s16): Likewise.
1170 (vreinterpret_u64_s32): Likewise.
1171 (vreinterpret_u64_s64): Likewise.
1172 (vreinterpret_u64_f32): Likewise.
1173 (vreinterpret_u64_u8): Likewise.
1174 (vreinterpret_u64_u16): Likewise.
1175 (vreinterpret_u64_u32): Likewise.
1176 (vreinterpret_u64_p8): Likewise.
1177 (vreinterpret_u64_p16): Likewise.
1178 (vreinterpretq_u64_s8): Likewise.
1179 (vreinterpretq_u64_s16): Likewise.
1180 (vreinterpretq_u64_s32): Likewise.
1181 (vreinterpretq_u64_s64): Likewise.
1182 (vreinterpretq_u64_f32): Likewise.
1183 (vreinterpretq_u64_u8): Likewise.
1184 (vreinterpretq_u64_u16): Likewise.
1185 (vreinterpretq_u64_u32): Likewise.
1186 (vreinterpretq_u64_p8): Likewise.
1187 (vreinterpretq_u64_p16): Likewise.
1188 (vreinterpret_s8_s16): Likewise.
1189 (vreinterpret_s8_s32): Likewise.
1190 (vreinterpret_s8_s64): Likewise.
1191 (vreinterpret_s8_f32): Likewise.
1192 (vreinterpret_s8_u8): Likewise.
1193 (vreinterpret_s8_u16): Likewise.
1194 (vreinterpret_s8_u32): Likewise.
1195 (vreinterpret_s8_u64): Likewise.
1196 (vreinterpret_s8_p8): Likewise.
1197 (vreinterpret_s8_p16): Likewise.
1198 (vreinterpretq_s8_s16): Likewise.
1199 (vreinterpretq_s8_s32): Likewise.
1200 (vreinterpretq_s8_s64): Likewise.
1201 (vreinterpretq_s8_f32): Likewise.
1202 (vreinterpretq_s8_u8): Likewise.
1203 (vreinterpretq_s8_u16): Likewise.
1204 (vreinterpretq_s8_u32): Likewise.
1205 (vreinterpretq_s8_u64): Likewise.
1206 (vreinterpretq_s8_p8): Likewise.
1207 (vreinterpretq_s8_p16): Likewise.
1208 (vreinterpret_s16_s8): Likewise.
1209 (vreinterpret_s16_s32): Likewise.
1210 (vreinterpret_s16_s64): Likewise.
1211 (vreinterpret_s16_f32): Likewise.
1212 (vreinterpret_s16_u8): Likewise.
1213 (vreinterpret_s16_u16): Likewise.
1214 (vreinterpret_s16_u32): Likewise.
1215 (vreinterpret_s16_u64): Likewise.
1216 (vreinterpret_s16_p8): Likewise.
1217 (vreinterpret_s16_p16): Likewise.
1218 (vreinterpretq_s16_s8): Likewise.
1219 (vreinterpretq_s16_s32): Likewise.
1220 (vreinterpretq_s16_s64): Likewise.
1221 (vreinterpretq_s16_f32): Likewise.
1222 (vreinterpretq_s16_u8): Likewise.
1223 (vreinterpretq_s16_u16): Likewise.
1224 (vreinterpretq_s16_u32): Likewise.
1225 (vreinterpretq_s16_u64): Likewise.
1226 (vreinterpretq_s16_p8): Likewise.
1227 (vreinterpretq_s16_p16): Likewise.
1228 (vreinterpret_s32_s8): Likewise.
1229 (vreinterpret_s32_s16): Likewise.
1230 (vreinterpret_s32_s64): Likewise.
1231 (vreinterpret_s32_f32): Likewise.
1232 (vreinterpret_s32_u8): Likewise.
1233 (vreinterpret_s32_u16): Likewise.
1234 (vreinterpret_s32_u32): Likewise.
1235 (vreinterpret_s32_u64): Likewise.
1236 (vreinterpret_s32_p8): Likewise.
1237 (vreinterpret_s32_p16): Likewise.
1238 (vreinterpretq_s32_s8): Likewise.
1239 (vreinterpretq_s32_s16): Likewise.
1240 (vreinterpretq_s32_s64): Likewise.
1241 (vreinterpretq_s32_f32): Likewise.
1242 (vreinterpretq_s32_u8): Likewise.
1243 (vreinterpretq_s32_u16): Likewise.
1244 (vreinterpretq_s32_u32): Likewise.
1245 (vreinterpretq_s32_u64): Likewise.
1246 (vreinterpretq_s32_p8): Likewise.
1247 (vreinterpretq_s32_p16): Likewise.
1248 (vreinterpret_u8_s8): Likewise.
1249 (vreinterpret_u8_s16): Likewise.
1250 (vreinterpret_u8_s32): Likewise.
1251 (vreinterpret_u8_s64): Likewise.
1252 (vreinterpret_u8_f32): Likewise.
1253 (vreinterpret_u8_u16): Likewise.
1254 (vreinterpret_u8_u32): Likewise.
1255 (vreinterpret_u8_u64): Likewise.
1256 (vreinterpret_u8_p8): Likewise.
1257 (vreinterpret_u8_p16): Likewise.
1258 (vreinterpretq_u8_s8): Likewise.
1259 (vreinterpretq_u8_s16): Likewise.
1260 (vreinterpretq_u8_s32): Likewise.
1261 (vreinterpretq_u8_s64): Likewise.
1262 (vreinterpretq_u8_f32): Likewise.
1263 (vreinterpretq_u8_u16): Likewise.
1264 (vreinterpretq_u8_u32): Likewise.
1265 (vreinterpretq_u8_u64): Likewise.
1266 (vreinterpretq_u8_p8): Likewise.
1267 (vreinterpretq_u8_p16): Likewise.
1268 (vreinterpret_u16_s8): Likewise.
1269 (vreinterpret_u16_s16): Likewise.
1270 (vreinterpret_u16_s32): Likewise.
1271 (vreinterpret_u16_s64): Likewise.
1272 (vreinterpret_u16_f32): Likewise.
1273 (vreinterpret_u16_u8): Likewise.
1274 (vreinterpret_u16_u32): Likewise.
1275 (vreinterpret_u16_u64): Likewise.
1276 (vreinterpret_u16_p8): Likewise.
1277 (vreinterpret_u16_p16): Likewise.
1278 (vreinterpretq_u16_s8): Likewise.
1279 (vreinterpretq_u16_s16): Likewise.
1280 (vreinterpretq_u16_s32): Likewise.
1281 (vreinterpretq_u16_s64): Likewise.
1282 (vreinterpretq_u16_f32): Likewise.
1283 (vreinterpretq_u16_u8): Likewise.
1284 (vreinterpretq_u16_u32): Likewise.
1285 (vreinterpretq_u16_u64): Likewise.
1286 (vreinterpretq_u16_p8): Likewise.
1287 (vreinterpretq_u16_p16): Likewise.
1288 (vreinterpret_u32_s8): Likewise.
1289 (vreinterpret_u32_s16): Likewise.
1290 (vreinterpret_u32_s32): Likewise.
1291 (vreinterpret_u32_s64): Likewise.
1292 (vreinterpret_u32_f32): Likewise.
1293 (vreinterpret_u32_u8): Likewise.
1294 (vreinterpret_u32_u16): Likewise.
1295 (vreinterpret_u32_u64): Likewise.
1296 (vreinterpret_u32_p8): Likewise.
1297 (vreinterpret_u32_p16): Likewise.
1298 (vreinterpretq_u32_s8): Likewise.
1299 (vreinterpretq_u32_s16): Likewise.
1300 (vreinterpretq_u32_s32): Likewise.
1301 (vreinterpretq_u32_s64): Likewise.
1302 (vreinterpretq_u32_f32): Likewise.
1303 (vreinterpretq_u32_u8): Likewise.
1304 (vreinterpretq_u32_u16): Likewise.
1305 (vreinterpretq_u32_u64): Likewise.
1306 (vreinterpretq_u32_p8): Likewise.
1307 (vreinterpretq_u32_p16): Likewise.
1309 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1311 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
1313 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator
1316 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
1317 (vqnegd_s64): Likewise.
1318 (vqabs_s64): Likewise.
1319 (vqabsd_s64): Likewise.
1321 2014-04-22 Richard Henderson <rth@redhat.com>
1323 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
1324 computation to the top of the loop.
1326 2014-04-22 Renlin <renlin.li@arm.com>
1327 Jiong Wang <jiong.wang@arm.com>
1329 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
1330 * config/aarch64/aarch64.c (aarch64_layout_frame)
1331 (aarch64_initial_elimination_offset): Likewise.
1333 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
1335 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
1338 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
1340 * machmode.h (bitwise_mode_for_mode): Declare.
1341 * stor-layout.h (bitwise_type_for_mode): Likewise.
1342 * stor-layout.c (bitwise_mode_for_mode): New function.
1343 (bitwise_type_for_mode): Likewise.
1344 * builtins.c (fold_builtin_memory_op): Use it instead of
1345 int_mode_for_mode and build_nonstandard_integer_type.
1347 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1349 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
1350 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
1351 (*-*-solaris2*): Simplify.
1352 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
1353 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
1354 *-*-solaris2.9* handling.
1356 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
1358 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
1359 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
1361 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
1362 * configure: Regenerate.
1364 * config/i386/sol2-9.h: Remove.
1366 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
1367 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
1368 Remove Solaris 9 references.
1370 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
1372 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
1373 (floatuns<GPI:mode><GPF:mode>2): Remove.
1374 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
1375 and floatuns conversions.
1376 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
1377 and floatuns conversions.
1378 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
1379 (w1,w2): New mode attributes for inequal width conversions.
1381 2014-04-22 Renlin Li <Renlin.Li@arm.com>
1383 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
1384 the output asm format.
1386 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
1388 * config/aarch64/aarch64-simd.md
1389 (aarch64_cm<optab>di): Always split.
1390 (*aarch64_cm<optab>di): New.
1391 (aarch64_cmtstdi): Always split.
1392 (*aarch64_cmtstdi): New.
1394 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1396 PR tree-optimization/60823
1397 * omp-low.c (ipa_simd_modify_function_body): Go through
1398 all SSA_NAMEs and for those refering to vector arguments
1399 which are going to be replaced adjust SSA_NAME_VAR and,
1400 if it is a default definition, change it into a non-default
1401 definition assigned at the beginning of function from new_decl.
1402 (ipa_simd_modify_stmt_ops): Rewritten.
1403 * tree-dfa.c (set_ssa_default_def): When removing default def,
1404 check for NULL loc instead of NULL *loc.
1406 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1408 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
1409 restrictions on core registers for DImode values in Thumb2.
1411 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1413 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
1414 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
1416 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1418 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
1419 (*iordi_notzesidi_di): Likewise.
1420 (*iordi_notsesidi_di): Likewise.
1422 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1424 * config/arm/arm-protos.h (tune_params): New struct members.
1425 * config/arm/arm.c: Initialise tune_params per processor.
1426 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
1427 for speed, based on new tune_params.
1429 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1431 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro
1433 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added
1435 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment
1437 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
1438 * config/aarch64/arm_neon.h (vrnd_f64): Added.
1439 (vrnda_f64): Likewise.
1440 (vrndi_f64): Likewise.
1441 (vrndm_f64): Likewise.
1442 (vrndn_f64): Likewise.
1443 (vrndp_f64): Likewise.
1444 (vrndx_f64): Likewise.
1446 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1448 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
1449 GET_MODE_SIZE argument is enum machine_mode.
1451 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1454 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
1455 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
1457 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
1460 * asan.c (asan_emit_stack_protection): Force the base to align to
1461 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
1462 appropriate bits if STRICT_ALIGNMENT.
1463 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
1465 (expand_used_vars): Leave a space in the stack frame for alignment
1466 if STRICT_ALIGNMENT.
1468 2014-04-21 David Malcolm <dmalcolm@redhat.com>
1470 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
1472 (gimple_store_p): Likewise.
1473 (gimple_assign_load_p): Likewise.
1474 (gimple_assign_cast_p): Likewise.
1475 (gimple_clobber_p): Likewise.
1477 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
1478 rather than a gimple.
1479 (gimple_assign_cast_p): Likewise.
1481 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
1484 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
1485 If mode is DDmode and TARGET_E500_DOUBLE allow move.
1487 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
1488 more debug information for E500 if -mdebug=reg.
1490 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
1493 * config/i386/i386.c (ix86_expand_builtin)
1494 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
1495 register for target RTX.
1496 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
1498 2014-04-18 Cong Hou <congh@google.com>
1500 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
1501 the widen-mult pattern by handling two operands with different sizes,
1502 and operands whose size is smaller than half of the result type.
1504 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
1506 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
1507 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
1508 (do_estimate_edge_time): Compute it.
1509 * ipa-inline.c (want_inline_small_function_p): Bypass
1510 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
1512 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
1514 * ipa-inline.c (spec_rem): New static variable.
1515 (dump_overall_stats): New function.
1516 (dump_inline_stats): New function.
1518 2014-04-18 Richard Henderson <rth@redhat.com>
1520 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
1521 to GET_MODE_SIZE, not a reg_class_t.
1523 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1525 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
1526 (vsx_xxmrglw_<mode>): Likewise.
1528 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
1531 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
1532 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
1533 (rs6000_init_hard_regno_mode_ok): Likewise.
1535 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
1537 * ipa-inline.c (inline_small_functions): Account only non-cold
1539 * doc/invoke.texi (inline-unit-growth): Update documentation.
1541 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
1543 * config/rs6000/rs6000.md (addti3, subti3): New.
1545 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
1548 * config/i386/i386.c (ix86_expand_clear): Remove outdated
1549 comment. Check optimize_insn_for_size_p instead of
1550 optimize_insn_for_speed_p.
1552 2014-04-17 Martin Jambor <mjambor@suse.cz>
1554 * gimple-iterator.c (gsi_start_edge): New function.
1555 * gimple-iterator.h (gsi_start_edge): Declare.
1556 * tree-sra.c (single_non_eh_succ): New function.
1557 (disqualify_ops_if_throwing_stmt): Renamed to
1558 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
1559 having one non-EH successor BB.
1560 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
1561 generate loads into replacements.
1562 (sra_modify_assign): Likewise and and also use the simple path for
1564 (sra_modify_function_body): Commit statements on edges.
1566 2014-04-17 Richard Biener <rguenther@suse.de>
1569 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
1570 comparison results and add clarifying comment.
1572 2014-04-17 Jakub Jelinek <jakub@redhat.com>
1574 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
1575 (blank_mode): Initialize it.
1576 (emit_mode_size_inline, emit_mode_nunits_inline,
1577 emit_mode_inner_inline): New functions.
1578 (emit_insn_modes_h): Call them and surround their output with
1579 #if GCC_VERSION >= 4001 ... #endif.
1580 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
1581 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
1582 mode_* arrays if the argument is __builtin_constant_p.
1583 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
1584 is enum machine_mode.
1586 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1588 * passes.c (opt_pass::execute): Adjust.
1589 (pass_manager::execute_pass_mode_switching): Likewise.
1590 (early_local_passes::execute): Likewise.
1591 (execute_one_pass): Pass cfun to the pass's execute method.
1592 * tree-pass.h (opt_pass::execute): Add function * argument.
1593 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
1594 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
1595 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
1596 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1597 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
1598 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
1599 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
1600 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
1601 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
1602 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
1603 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
1604 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
1605 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
1606 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
1607 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
1608 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
1609 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
1610 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
1611 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
1612 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1613 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1614 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1615 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1616 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1617 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1618 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1619 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
1620 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
1621 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
1624 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1626 * passes.c (opt_pass::gate): Take function * argument.
1627 (gate_all_early_local_passes): Merge into
1628 (early_local_passes::gate): this.
1629 (gate_all_early_optimizations): Merge into
1630 (all_early_optimizations::gate): this.
1631 (gate_all_optimizations): Mege into
1632 (all_optimizations::gate): this.
1633 (gate_all_optimizations_g): Merge into
1634 (all_optimizations_g::gate): this.
1635 (gate_rest_of_compilation): Mege into
1636 (rest_of_compilation::gate): this.
1637 (gate_postreload): Merge into
1638 (postreload::gate): this.
1639 (dump_one_pass): Pass cfun to the pass's gate method.
1640 (execute_ipa_summary_passes): Likewise.
1641 (execute_one_pass): Likewise.
1642 (ipa_write_summaries_2): Likewise.
1643 (ipa_write_optimization_summaries_1): Likewise.
1644 (ipa_read_summaries_1): Likewise.
1645 (ipa_read_optimization_summaries_1): Likewise.
1646 (execute_ipa_stmt_fixups): Likewise.
1647 * tree-pass.h (opt_pass::gate): Add function * argument.
1648 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
1649 combine-stack-adj.c, combine.c, compare-elim.c,
1650 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1651 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
1652 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
1653 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
1654 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
1655 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
1656 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
1657 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
1658 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
1659 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
1660 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
1661 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
1662 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
1663 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
1664 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
1665 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1666 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1667 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1668 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1669 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1670 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1671 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1672 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
1673 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
1674 var-tracking.c, vtable-verify.c, web.c: Adjust.
1676 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1678 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
1679 * configure: Regenerate.
1681 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1683 * passes.c (dump_one_pass): don't check pass->has_gate.
1684 (execute_ipa_summary_passes): Likewise.
1685 (execute_one_pass): Likewise.
1686 (ipa_write_summaries_2): Likewise.
1687 (ipa_write_optimization_summaries_1): Likewise.
1688 (ipa_read_optimization_summaries_1): Likewise.
1689 (execute_ipa_stmt_fixups): Likewise.
1690 * tree-pass.h (pass_data::has_gate): Remove.
1691 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
1692 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
1693 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
1694 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1695 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
1696 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
1697 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
1698 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
1699 gimple-low.c, gimple-ssa-isolate-paths.c,
1700 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
1701 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
1702 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
1703 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
1704 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
1705 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
1706 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
1707 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
1708 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
1709 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
1710 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
1711 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
1712 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1713 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1714 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1715 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1716 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1717 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1718 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1719 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
1720 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
1721 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
1724 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1726 * pass_manager.h (pass_manager::register_dump_files_1): Remove
1728 * passes.c (pass_manager::register_dump_files_1): Merge into
1729 (pass_manager::register_dump_files): this, and remove its handling of
1730 properties since the pass always has the properties anyway.
1731 (pass_manager::pass_manager): Adjust.
1733 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1735 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
1736 * passes.c (pass_manager::register_dump_files_1): Remove dead code
1737 dealing with properties.
1738 (pass_manager::register_dump_files): Adjust.
1740 2014-03-20 Mark Wielaard <mjw@redhat.com>
1742 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
1743 then represent the bound as normal constant value.
1745 2014-04-17 Jakub Jelinek <jakub@redhat.com>
1748 Forward port from 4.8 branch
1749 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
1751 * config/i386/bmiintrin.h (_blsi_u32): New.
1755 (_blsmsk_u32): Ditto.
1756 (_blsmsk_u64): Ditto.
1757 (_tzcnt_u32): Ditto.
1758 (_tzcnt_u64): Ditto.
1760 2014-04-17 Kito Cheng <kito@0xlab.org>
1762 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
1764 2014-04-17 Richard Biener <rguenther@suse.de>
1767 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
1768 boolean results for comparisons.
1770 2014-04-17 Richard Biener <rguenther@suse.de>
1772 PR tree-optimization/60836
1773 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
1774 initial PHI args to be gimple values.
1776 2014-04-17 Richard Biener <rguenther@suse.de>
1778 PR tree-optimization/60841
1779 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
1780 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
1781 of stmts to SLP build.
1782 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
1783 (vect_analyze_slp): Likewise.
1784 (vect_analyze_slp_instance): Likewise.
1785 (vect_build_slp_tree): Limit overall SLP tree growth.
1786 * tree-vectorizer.h (vect_analyze_data_refs,
1787 vect_analyze_slp): Adjust prototypes.
1789 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
1791 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
1794 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
1796 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
1797 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
1798 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
1799 for TARGET_SLOW_PSHUFB
1801 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
1803 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
1804 * config/i386/i386.c (intel_cost): Ditto.
1806 2014-04-17 Joey Ye <joey.ye@arm.com>
1808 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
1810 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
1812 * opts.c (common_handle_option): Disable -fipa-reference coorectly
1815 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
1817 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
1818 (type_all_derivations_known_p): New predicate.
1819 (type_all_ctors_visible_p): New predicate.
1820 (type_possibly_instantiated_p): New predicate.
1821 (get_odr_type): Compute all_derivations_known.
1822 (dump_odr_type): Dump the flag.
1823 (maybe_record_type): Cleanup.
1824 (record_target_from_binfo): Add bases_to_consider array;
1825 record bases for types w/o instances and skip CXX destructor.
1826 (possible_polymorphic_call_targets_1): Add bases_to_consider
1827 and consider_construction parameters; check if type may have instance.
1828 (get_polymorphic_call_info): Set maybe_in_construction to true
1829 when we know nothing.
1830 (record_targets_from_bases): Skip CXX destructors; they are
1831 never called for types in construction.
1832 (possible_polymorphic_call_targets): Do not record target when
1833 type may not have instance.
1835 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
1838 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
1839 external aliases alive, too.
1841 2014-04-16 Andrew Pinski <apinski@cavium.com>
1843 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
1846 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
1848 * final.c (compute_alignments): Do not apply loop alignment to a block
1849 falling through to the exit.
1851 2014-04-16 Catherine Moore <clm@codesourcery.com>
1853 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
1854 Adjust constraints for microMIPS store patterns.
1856 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
1858 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
1860 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
1862 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
1863 (append_use): Run at -O0.
1864 (append_vdef): Likewise.
1865 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
1866 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
1868 2014-04-16 Jakub Jelinek <jakub@redhat.com>
1870 PR tree-optimization/60844
1871 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
1872 (propagate_op_to_single_use, remove_visited_stmt_chain,
1873 linearize_expr, repropagate_negates, reassociate_bb): Use it
1874 instead of gsi_remove.
1876 2014-04-16 Martin Jambor <mjambor@suse.cz>
1878 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
1879 ipa_transforms_to_apply.
1880 (cgraph_function_versioning): Assert that old_node has empty
1881 ipa_transforms_to_apply.
1882 * trans-mem.c (ipa_tm_create_version): Likewise.
1883 * tree-inline.c (tree_function_versioning): Do not duplicate
1884 ipa_transforms_to_apply.
1886 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1889 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
1891 Pass necessary as flags on 64-bit Solaris/x86.
1892 Use lowercase relocs for x86_64-*-*.
1893 * configure: Regenerate.
1895 2014-04-15 Jan Hubicka <jh@suse.cz>
1897 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
1898 (maybe_record_node, likely_target_p): Use it.
1900 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1903 Revert following patch
1905 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
1908 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
1909 software floating point or no floating point registers, do not
1910 allow any type in the FPRs. Eliminate a test for SPE SIMD types
1911 in GPRs that occurs after we tested for GPRs that would never be
1914 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
1915 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
1916 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
1917 specifically allow DDmode, since that does not use the SPE SIMD
1920 2014-03-21 Mark Wielaard <mjw@redhat.com>
1922 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
1923 as unsigned or int depending on type and value used.
1925 2014-04-15 Richard Biener <rguenther@suse.de>
1927 PR rtl-optimization/56965
1928 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
1929 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
1931 * alias.c (true_dependence_1): Do not call
1932 nonoverlapping_component_refs_p.
1933 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
1934 nonoverlapping_component_refs_p.
1935 (indirect_refs_may_alias_p): Likewise.
1937 2014-04-15 Teresa Johnson <tejohnson@google.com>
1939 * cfg.c (dump_bb_info): Fix flags check.
1940 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
1942 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1944 PR rtl-optimization/60663
1945 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
1948 2014-04-15 Richard Biener <rguenther@suse.de>
1950 * lto-streamer.h (LTO_major_version): Bump to 4.
1952 2014-04-15 Richard Biener <rguenther@suse.de>
1954 * common.opt (lto_partition_model): New enum.
1955 (flto-partition=): Merge separate options with a single with argument,
1956 add -flto-partition=one support.
1957 * flag-types.h (enum lto_partition_model): Declare.
1958 * opts.c (finish_options): Remove duplicate -flto-partition=
1960 * lto-wrapper.c (run_gcc): Adjust.
1962 2014-04-15 Richard Biener <rguenther@suse.de>
1964 * alias.c (ncr_compar): New function.
1965 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
1967 2014-04-15 Richard Biener <rguenther@suse.de>
1969 * alias.c (record_component_aliases): Do not walk BINFOs.
1971 2014-04-15 Richard Biener <rguenther@suse.de>
1973 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
1974 Add struct function argument and adjust.
1975 (find_func_aliases_for_call): Likewise.
1976 (find_func_aliases): Likewise.
1977 (find_func_clobbers): Likewise.
1978 (intra_create_variable_infos): Likewise.
1979 (compute_points_to_sets): Likewise.
1980 (ipa_pta_execute): Adjust. Do not push/pop cfun.
1982 2014-04-15 Richard Biener <rguenther@suse.de>
1984 * tree.c (iterative_hash_expr): Use enum tree_code_class
1985 to store TREE_CODE_CLASS.
1986 (tree_block): Likewise.
1987 (tree_set_block): Likewise.
1988 * tree.h (fold_build_pointer_plus_loc): Use
1989 convert_to_ptrofftype_loc.
1991 2014-04-15 Jakub Jelinek <jakub@redhat.com>
1994 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
1997 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
1999 * cfgloop.h (struct loop): Move force_vectorize down.
2000 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
2001 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
2002 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
2003 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
2004 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
2005 * tree-core.h (enum annot_expr_kind): Add new kind values.
2006 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
2007 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
2009 * tree.def (ANNOTATE_EXPR): Tweak comment.
2011 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2013 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
2016 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
2018 * tree.h (TYPE_IDENTIFIER): Declare.
2019 * tree.c (subrange_type_for_debug_p): Use it.
2020 * godump.c (go_format_type): Likewise.
2021 * dwarf2out.c (is_cxx_auto, modified_type_die,
2022 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
2023 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
2025 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2028 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
2030 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
2032 * config/i386/i386.c (examine_argument): Return bool. Return true if
2033 parameter should be passed in memory.
2034 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
2035 (construct_container): Update calls to examine_argument.
2036 (function_arg_advance_64): Ditto.
2037 (return_in_memory_32): Merge with ix86_return_in_memory.
2038 (return_in_memory_64): Ditto.
2039 (return_in_memory_ms_64): Ditto.
2041 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2043 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
2044 * coverage.c (coverage_compute_profile_id): Handle externally visible
2047 2014-04-14 Martin Jambor <mjambor@suse.cz>
2049 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
2050 DECL_DISREGARD_INLINE_LIMITS functions.
2052 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
2055 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
2057 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
2060 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
2061 optimize_insn_for_speed_p instead of
2062 optimize_function_for_speed_p.
2064 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
2066 * doc/invoke.texi (free): Document AArch64.
2068 2014-04-14 Richard Biener <rguenther@suse.de>
2070 PR tree-optimization/60042
2071 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
2072 (insert_into_preds_of_block): Do not prevent PHI insertion
2073 for REFERENCE exprs here ...
2074 (eliminate_dom_walker::before_dom_children): ... but prevent
2075 their use here under similar conditions when applied to the
2076 IL after PRE optimizations.
2078 2014-04-14 Richard Biener <rguenther@suse.de>
2080 * passes.def: Move early points-to after early SRA.
2082 2014-04-14 Richard Biener <rguenther@suse.de>
2084 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
2085 check for which sign-changes we allow when forwarding
2086 a converted value into a switch.
2088 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2090 * stor-layout.c (place_field): Finalize non-constant offset for the
2093 2014-04-14 Richard Biener <rguenther@suse.de>
2095 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
2097 (expand_switch_using_bit_tests_p): Likewise.
2098 (process_switch): Compute and pass on speed_p based on the
2100 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
2101 optimize_bb_for_speed_p.
2103 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2105 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
2106 * function.h (struct function): Rename has_force_vect_loops into
2107 has_force_vectorize_loops.
2108 * lto-streamer-in.c (input_cfg): Adjust for renaming.
2109 (input_struct_function_base): Likewise.
2110 * lto-streamer-out.c (output_cfg): Likewise.
2111 (output_struct_function_base): Likewise.
2112 * omp-low.c (expand_omp_simd): Likewise.
2113 * tree-cfg.c (move_sese_region_to_fn): Likewise.
2114 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
2115 (version_loop_for_if_conversion): Likewise.
2116 (tree_if_conversion): Likewise.
2117 (main_tree_if_conversion): Likewise.
2118 (gate_tree_if_conversion): Likewise.
2119 * tree-inline.c (copy_loops): Likewise.
2120 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
2121 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
2122 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
2123 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
2124 * tree-vectorizer.c (vectorize_loops): Likewise.
2125 * tree-vectorizer.h (unlimited_cost_model): Likewise.
2127 2014-04-14 Richard Biener <rguenther@suse.de>
2130 * lto-streamer-out.c (wrap_refs): New function.
2131 (lto_output): Wrap symbol references in global initializes in
2132 type-preserving MEM_REFs.
2134 2014-04-14 Christian Bruel <christian.bruel@st.com>
2136 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
2138 2014-04-14 Christian Bruel <christian.bruel@st.com>
2140 * config/sh/sh.md (setmemqi): New expand pattern.
2141 * config/sh/sh.h (CLEAR_RATIO): Define.
2142 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
2143 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
2145 2014-04-14 Richard Biener <rguenther@suse.de>
2148 * fold-const.c (negate_expr_p): Don't negate directional rounding
2150 (fold_negate_expr): Likewise.
2152 2014-04-14 Richard Biener <rguenther@suse.de>
2154 PR tree-optimization/59817
2155 PR tree-optimization/60453
2156 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
2157 recursion to catch all CHRECs in the scalar evolution and restrict
2158 the predicate for the remains appropriately.
2160 2014-04-12 Catherine Moore <clm@codesourcery.com>
2162 * config/mips/constraints.md: Add new register constraint "kb".
2163 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
2164 (*movhi_internal): Likewise.
2165 (*movqi_internal): Likewise.
2166 * config/mips/mips.h (M16_STORE_REGS): New register class.
2167 (REG_CLASS_NAMES): Add M16_STORE_REGS.
2168 (REG_CLASS_CONTENTS): Likewise.
2169 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
2171 2014-04-11 Tobias Burnus <burnus@net-b.de>
2174 * doc/invoke.texi (-Wformat-signedness): Document it.
2175 (Wformat=2): Mention that this enables -Wformat-signedness.
2177 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2179 * common/config/epiphany/epiphany-common.c
2180 (epiphany_option_optimization_table): Enable section anchors by
2181 default at -O1 or higher.
2182 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
2183 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
2184 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
2185 carries no extra cost.
2186 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
2187 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
2188 * config/epiphany/predicates.md (memclob_operand): New predicate.
2189 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
2190 Use memclob_operand predicate and X constraint for operand 3.
2192 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2194 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
2195 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
2198 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2200 PR rtl-optimization/60651
2201 * mode-switching.c (optimize_mode_switching): Make sure to emit
2202 sets of a lower numbered entity before sets of a higher numbered
2203 entity to a mode of the same or lower priority.
2204 When creating a seginfo for a basic block that starts with a code
2205 label, move the insertion point past the code label.
2206 (new_seginfo): Document and enforce requirement that
2207 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
2208 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
2209 * doc/tm.texi: Regenerate.
2211 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
2214 * config/arc/arc.c (arc_save_restore): Fix assert typo.
2216 2013-04-11 Jakub Jelinek <jakub@redhat.com>
2218 * BASE-VER: Set to 4.10.0.
2220 2014-04-11 Tobias Burnus <burnus@net-b.de>
2223 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
2224 * doc/gcc.texi (Service): Update description in the @menu
2225 * doc/invoke.texi (Option Summary): Remove misplaced and
2228 2014-04-11 Steve Ellcey <sellcey@mips.com>
2229 Jakub Jelinek <jakub@redhat.com>
2232 * expr.c (convert_move): Use emit_store_flag_force instead of
2233 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
2236 2014-04-11 Richard Biener <rguenther@suse.de>
2239 * varasm.c (assemble_alias): Avoid endless error reporting
2240 recursion by setting TREE_ASM_WRITTEN.
2242 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2244 * config/s390/s390.md: Add a splitter for NOT rtx.
2246 2014-04-11 Jakub Jelinek <jakub@redhat.com>
2248 PR rtl-optimization/60663
2249 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
2251 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
2252 Jakub Jelinek <jakub@redhat.com>
2255 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
2256 flag from decl_node to node.
2258 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2261 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
2262 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
2263 ameliorating the cases where it can be.
2265 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
2268 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2270 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2271 (loadsync_<mode>): Change mode.
2272 (load_quadpti, store_quadpti): New.
2273 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2274 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2275 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
2277 2014-04-09 Cong Hou <congh@google.com>
2280 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
2283 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2285 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
2286 instead of vnor to exploit possible fusion opportunity in the
2288 (altivec_expand_vec_perm_const_le): Likewise.
2290 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2292 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2293 (loadsync_<mode>): Change mode.
2294 (load_quadpti, store_quadpti): New.
2295 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2296 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2298 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
2301 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
2302 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
2303 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
2305 2014-04-08 Richard Biener <rguenther@suse.de>
2308 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
2309 a 64bit widest int print double-int similar to on HWI64 hosts.
2311 2014-04-08 Richard Biener <rguenther@suse.de>
2313 PR tree-optimization/60785
2314 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
2315 default defs properly.
2317 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
2319 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
2320 (Weffc++): Likewise.
2322 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
2324 * ipa-devirt.c (maybe_record_node): When node is not recorded,
2325 set completep to false rather than true.
2327 2014-04-07 Douglas B Rupp <rupp@adacore.com>
2330 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
2331 ARM_TARGET2_DWARF_FORMAT.
2333 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
2336 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
2337 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
2340 2014-04-07 Richard Biener <rguenther@suse.de>
2342 PR tree-optimization/60766
2343 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
2344 (may_eliminate_iv): Convert cand_value_at result to desired type.
2346 2014-04-07 Jason Merrill <jason@redhat.com>
2349 * common.opt (-fno-gnu-unique): Add.
2350 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
2352 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2354 * haifa-sched.c: Fix outdated function reference and minor
2355 grammar errors in introductory comment.
2357 2014-04-07 Richard Biener <rguenther@suse.de>
2360 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
2362 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
2364 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
2367 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
2368 size accounting for thunks.
2369 (pa_asm_output_mi_thunk): Use final_start_function() and
2370 final_end_function() to output function start and end directives.
2372 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
2374 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
2375 device specific ISA/ feature information. Remove short_sp and
2376 errata_skip ds. Add avr_device_specific_features enum to have device
2378 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
2379 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
2380 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
2381 updated device specific info.
2382 * config/avr/avr-mcus.def: Merge device specific details to
2383 dev_attribute field.
2384 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
2386 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
2387 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
2388 assembler if RMW isa supported by current device.
2389 * config/avr/genmultilib.awk: Update as device info structure changed.
2390 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
2392 2014-04-04 Cong Hou <congh@google.com>
2394 PR tree-optimization/60656
2395 * tree-vect-stmts.c (supportable_widening_operation):
2396 Fix a bug that elements in a vector with vect_used_by_reduction
2397 property are incorrectly reordered when the operation on it is not
2398 consistant with the one in reduction operation.
2400 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
2402 PR rtl-optimization/60155
2403 * gcse.c (record_set_data): New function.
2404 (single_set_gcse): New function.
2405 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
2406 (hoist_code): Likewise.
2407 (get_pressure_class_and_nregs): Likewise.
2409 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
2411 * explow.c (probe_stack_range): Emit a final optimization blockage.
2413 2014-04-04 Anthony Green <green@moxielogic.com>
2415 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
2418 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
2421 * lto-cgraph.c (input_overwrite_node): Check that partitioning
2422 flags are set only during streaming.
2423 * ipa.c (process_references, walk_polymorphic_call_targets,
2424 symtab_remove_unreachable_nodes): Drop bodies of always inline
2425 after early inlining.
2426 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
2428 2014-04-04 Jakub Jelinek <jakub@redhat.com>
2429 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2432 * dwarf2out.c (const_ok_for_output_1): Reject expressions
2435 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2438 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
2440 (cortex_a53_fdivd): Likewise.
2442 2014-04-04 Martin Jambor <mjambor@suse.cz>
2445 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
2447 * cgraph.c (clone_of_p): Also return true if thunks match.
2448 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
2449 cgraph_function_or_thunk_node and an obsolete comment.
2450 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
2452 (build_function_decl_skip_args): Likewise.
2453 (set_new_clone_decl_and_node_flags): New function.
2454 (duplicate_thunk_for_node): Likewise.
2455 (redirect_edge_duplicating_thunks): Likewise.
2456 (cgraph_clone_node): New parameter args_to_skip, pass it to
2457 redirect_edge_duplicating_thunks which is called instead of
2458 cgraph_redirect_edge_callee.
2459 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
2460 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
2462 2014-04-04 Jeff Law <law@redhat.com>
2465 * config/arm/predicates.md (const_int_I_operand): New predicate.
2466 (const_int_M_operand): Similarly.
2467 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
2469 (insv_t2, extv_reg, extzv_t2): Likewise.
2470 (load_multiple_with_writeback): Similarly for const_int_I_operand.
2471 (pop_multiple_with_writeback_and_return): Likewise.
2472 (vfp_pop_multiple_with_writeback): Likewise
2474 2014-04-04 Richard Biener <rguenther@suse.de>
2477 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
2478 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
2480 * gimplify.h (gimple_add_tmp_var_fn): Declare.
2481 * gimplify.c (gimple_add_tmp_var_fn): New function.
2482 * gimple-expr.h (create_tmp_reg_fn): Declare.
2483 * gimple-expr.c (create_tmp_reg_fn): New function.
2484 * gimple-low.c (record_vars_into): Don't change cfun.
2485 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
2486 code generation without cfun.
2488 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
2491 * Makefile.in (install-driver): Fix shell scripting.
2493 2014-04-03 Cong Hou <congh@google.com>
2495 PR tree-optimization/60505
2496 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
2497 threshold of number of iterations below which no vectorization
2499 * tree-vect-loop.c (new_loop_vec_info):
2500 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
2501 * tree-vect-loop.c (vect_analyze_loop_operations):
2502 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
2503 * tree-vect-loop.c (vect_transform_loop):
2504 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
2505 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
2506 of iterations of the loop and see if we should build the epilogue.
2508 2014-04-03 Richard Biener <rguenther@suse.de>
2510 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
2511 (streamer_tree_cache_create): Adjust.
2512 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
2513 to allow optional nodes array.
2514 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
2515 (streamer_tree_cache_append): Likewise.
2516 (streamer_tree_cache_create): Create nodes array optionally
2517 as specified by parameter.
2518 * lto-streamer-out.c (create_output_block): Avoid maintaining
2519 the node array in the writer cache.
2520 (DFS_write_tree): Remove assertion.
2521 (produce_asm_for_decls): Free the out decl state hash table early.
2522 * lto-streamer-in.c (lto_data_in_create): Adjust for
2523 streamer_tree_cache_create prototype change.
2525 2014-04-03 Richard Biener <rguenther@suse.de>
2527 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
2528 set TREE_CHAIN to NULL_TREE.
2530 2014-04-03 Richard Biener <rguenther@suse.de>
2532 PR tree-optimization/60740
2533 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
2534 over all GIMPLE_COND operands.
2536 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
2538 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
2539 (Weffc++): Remove Scott's numbering, merge lists and reference
2542 2014-04-03 Nick Clifton <nickc@redhat.com>
2544 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
2547 2014-04-03 Martin Jambor <mjambor@suse.cz>
2549 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
2550 mention gcc_unreachable before failing.
2551 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
2554 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
2557 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
2558 inconsistent code and instead mark the context inconsistent.
2559 (possible_polymorphic_call_targets): For inconsistent contexts
2560 return empty complete list.
2562 2014-04-02 Anthony Green <green@moxielogic.com>
2564 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
2565 (extendqisi2, extendhisi2): Define.
2566 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
2567 (WCHAR_TYPE): Change to unsigned int.
2569 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2571 PR tree-optimization/60733
2572 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
2573 insertion point for PHI candidates to be the end of the feeding
2574 block for the PHI argument.
2576 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
2578 PR rtl-optimization/60650
2579 * lra-constraints.c (process_alt_operands): Decrease reject for
2580 earlyclobber matching.
2582 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2584 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
2586 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2588 * config/spu/spu.c (pad_bb): Do not crash when the last
2589 insn is CODE_FOR_blockage.
2591 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2593 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
2594 lies outside the target mode.
2596 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2599 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
2600 software floating point or no floating point registers, do not
2601 allow any type in the FPRs. Eliminate a test for SPE SIMD types
2602 in GPRs that occurs after we tested for GPRs that would never be
2605 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
2606 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
2607 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
2608 specifically allow DDmode, since that does not use the SPE SIMD
2611 2014-04-02 Richard Biener <rguenther@suse.de>
2614 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
2615 MODE_INTs. Properly use negv_optab.
2616 (expand_abs): Likewise.
2618 2014-04-02 Richard Biener <rguenther@suse.de>
2621 * Makefile.in (install-driver): Guard extra installs with special
2624 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
2626 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
2629 2014-04-01 Richard Henderson <rth@redhat.com>
2632 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
2633 alternative enabled before register allocation.
2635 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
2637 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
2638 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
2640 (nios2_large_got_address): Remove unneeded 'sym' parameter.
2641 (nios2_got_address): Update nios2_large_got_address call site.
2642 (nios2_delegitimize_address): New function.
2643 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
2644 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
2645 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
2647 2014-04-01 Martin Husemann <martin@duskware.de>
2649 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
2652 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
2654 PR rtl-optimization/60604
2655 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
2656 check from register_operand.
2657 (register_operand): Redefine in terms of general_operand.
2658 (nonmemory_operand): Use register_operand for the non-constant cases.
2660 2014-04-01 Richard Biener <rguenther@suse.de>
2662 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
2664 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
2666 * doc/invoke.texi (mapp-regs): Clarify.
2668 2014-03-31 Ulrich Drepper <drepper@gmail.com>
2670 * config/i386/avx512fintrin.h (__v32hi): Define type.
2671 (__v64qi): Likewise.
2672 (_mm512_set1_epi8): Define.
2673 (_mm512_set1_epi16): Define.
2674 (_mm512_set4_epi32): Define.
2675 (_mm512_set4_epi64): Define.
2676 (_mm512_set4_pd): Define.
2677 (_mm512_set4_ps): Define.
2678 (_mm512_setr4_epi64): Define.
2679 (_mm512_setr4_epi32): Define.
2680 (_mm512_setr4_pd): Define.
2681 (_mm512_setr4_ps): Define.
2682 (_mm512_setzero_epi32): Define.
2684 2014-03-31 Martin Jambor <mjambor@suse.cz>
2687 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
2688 callsite_arguments_match_p. Updated all callers. Also check types of
2689 corresponding formal parameters and actual arguments.
2690 (not_all_callers_have_enough_arguments_p) Renamed to
2691 some_callers_have_mismatched_arguments_p.
2693 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
2695 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
2697 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
2700 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
2703 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
2705 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
2706 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
2708 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
2709 Use FMAMODE_NOVF512 mode iterator.
2710 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
2711 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
2712 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
2713 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
2715 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
2716 Use VF_128_256 mode iterator.
2717 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
2720 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
2722 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
2723 static chain if needed.
2725 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
2728 * lra-constraints.c (index_part_to_reg): New.
2729 (process_address): Use it.
2731 2014-03-27 Jeff Law <law@redhat.com>
2732 Jakub Jelinek <jakub@redhat.com>
2735 * expr.c (do_tablejump): Use simplify_gen_binary rather than
2736 gen_rtx_{PLUS,MULT} to build up the address expression.
2738 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
2739 creating non-canonical RTL.
2741 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
2744 * ipa-inline.c (want_inline_small_function_p): Short circuit large
2745 functions; reorganize to make cheap checks first.
2746 (inline_small_functions): Do not estimate growth when dumping;
2748 * ipa-inline.h (inline_summary): Add min_size.
2749 (growth_likely_positive): New function.
2750 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
2751 (set_cond_stmt_execution_predicate): Cleanup.
2752 (estimate_edge_size_and_time): Compute min_size.
2753 (estimate_calls_size_and_time): Likewise.
2754 (estimate_node_size_and_time): Likewise.
2755 (inline_update_overall_summary): Update min_size.
2756 (do_estimate_edge_time): Likewise.
2757 (do_estimate_edge_size): Update.
2758 (do_estimate_edge_hints): Update.
2759 (growth_likely_positive): New function.
2761 2014-03-28 Jakub Jelinek <jakub@redhat.com>
2764 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
2765 also if addr has VOIDmode.
2767 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2769 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
2770 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
2772 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
2773 instructions as well as AdvancedSIMD loads.
2775 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2777 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
2778 Use crypto_aese type.
2779 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
2780 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
2781 crypto_aese, crypto_aesmc. Move to types.md.
2782 * config/arm/types.md (crypto_aes): Split into crypto_aese,
2784 * config/arm/iterators.md (crypto_type): Likewise.
2786 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
2788 * cgraph.c: Include expr.h and tree-dfa.h.
2789 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
2792 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
2795 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
2796 regs from checking multi-reg pseudos.
2798 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2800 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
2802 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2804 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
2805 if it would clobber the stack pointer, even temporarily.
2807 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
2809 * mode-switching.c: Make small adjustments to the top comment.
2811 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
2813 * config/rs6000/constraints.md (wD constraint): New constraint to
2814 match the constant integer to get the top DImode/DFmode out of a
2815 vector in a VSX register.
2817 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
2818 match the constant integer to get the top DImode/DFmode out of a
2819 vector in a VSX register.
2821 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
2824 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
2827 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
2828 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
2830 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
2831 Optimize vec_extract of 64-bit values, where the value being
2832 extracted is in the top word, where we can use scalar
2833 instructions. Add direct move and store support. Combine the big
2834 endian/little endian vector select load support into a single insn.
2835 (vsx_extract_<mode>_internal1): Likewise.
2836 (vsx_extract_<mode>_internal2): Likewise.
2837 (vsx_extract_<mode>_load): Likewise.
2838 (vsx_extract_<mode>_store): Likewise.
2839 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
2840 combined into vsx_extract_<mode>_load.
2841 (vsx_extract_<mode>_one_le): Likewise.
2843 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
2844 define the top 64-bit vector element.
2846 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
2849 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
2850 Document vec_vbpermq builtin.
2853 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
2854 enable use of xxsldwi and xxpermdi builtin functions.
2855 (vec_xxpermdi): Likewise.
2857 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
2858 Document use of vec_xxsldwi and vec_xxpermdi builtins.
2860 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
2862 PR rtl-optimization/60650
2863 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
2865 (find_spills_for): New.
2866 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
2867 Spill all pseudos on the second iteration.
2869 2014-03-27 Marek Polacek <polacek@redhat.com>
2872 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
2875 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2877 * config/s390/s390.c (s390_can_use_return_insn): Check for
2878 call-saved FPRs on 31 bit.
2880 2014-03-27 Jakub Jelinek <jakub@redhat.com>
2883 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
2884 if they need regimplification, just drop them instead of
2885 calling gimple_regimplify_operands on them.
2887 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
2890 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
2891 (aarch64_frame_pointer_required): Adjust logic.
2892 (aarch64_can_eliminate): Adjust logic.
2893 (aarch64_override_options_after_change): Adjust logic.
2895 2014-03-27 Dehao Chen <dehao@google.com>
2897 * ipa-inline.c (early_inliner): Update node's inline info.
2899 2014-03-26 Dehao Chen <dehao@google.com>
2901 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
2902 compiler inserted conditional jumps for NAN float check.
2904 2014-03-26 Jakub Jelinek <jakub@redhat.com>
2906 * ubsan.h (ubsan_create_data): Change second argument's type
2907 to const location_t *.
2908 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
2910 (ubsan_create_data): Change second argument to const location_t *PLOC.
2911 Create Loc field whenever PLOC is non-NULL.
2912 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
2913 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
2917 * real.c (real_to_integer2): Change type of low to UHWI.
2919 2014-03-26 Tobias Burnus <burnus@net-b.de>
2921 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
2922 (CILK_SELF_SPECS): New define.
2923 (driver_self_specs): Use it.
2925 2014-03-26 Richard Biener <rguenther@suse.de>
2927 * tree-pretty-print.c (percent_K_format): Implement special
2928 case for LTO and its stripped down BLOCK tree.
2930 2014-03-26 Jakub Jelinek <jakub@redhat.com>
2933 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
2935 * tree-vrp.c (simplify_internal_call_using_ranges): If only
2936 one range is range_int_cst_p, but not both, at least optimize
2937 addition/subtraction of 0 and multiplication by 0 or 1.
2938 * gimple-fold.c (gimple_fold_call): Fold
2939 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
2940 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
2941 INTEGER_CSTs, try to fold at least x * 0 and y - y.
2943 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
2945 PR rtl-optimization/60452
2946 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
2947 <case REG>: Return 1 for invalid offsets from the frame pointer.
2949 2014-03-26 Marek Polacek <polacek@redhat.com>
2952 * doc/extend.texi (C Extensions): Mention variable-length arrays in
2955 2014-03-26 Marek Polacek <polacek@redhat.com>
2958 * doc/extend.texi (Designated Inits): Describe what happens to omitted
2961 2014-03-26 Marek Polacek <polacek@redhat.com>
2964 * ira-color.c (update_conflict_hard_regno_costs): Perform the
2965 multiplication in unsigned type.
2967 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
2969 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
2971 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
2973 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
2975 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
2978 * cif-code.def (UNREACHABLE) New code.
2979 * ipa-inline.c (inline_small_functions): Skip edges to
2980 __builtlin_unreachable.
2981 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
2982 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
2983 predicate to __bulitin_unreachable.
2984 (set_cond_stmt_execution_predicate): Fix issue when
2985 invert_tree_comparison returns ERROR_MARK.
2986 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
2987 propagate to inline clones.
2988 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
2990 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
2991 * cgraphclones.c (cgraph_clone_node): If call destination is already
2992 ureachable, do not redirect it back.
2993 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
2996 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
2998 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
2999 Do not modify inline clones.
3001 2014-03-25 Jakub Jelinek <jakub@redhat.com>
3003 * config/i386/i386.md (general_sext_operand): New mode attr.
3004 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
3005 don't generate (sign_extend (const_int)).
3006 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
3007 operands[2]. Use We constraint instead of <i> and
3008 <general_sext_operand> predicate instead of <general_operand>.
3009 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
3010 * config/i386/constraints.md (We): New constraint.
3011 * config/i386/predicates.md (x86_64_sext_operand,
3012 sext_operand): New predicates.
3014 2014-03-25 Martin Jambor <mjambor@suse.cz>
3017 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
3018 inconsistent devirtualizations to __builtin_unreachable.
3020 2014-03-25 Marek Polacek <polacek@redhat.com>
3023 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
3025 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
3027 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
3028 order of elements for big-endian.
3030 2014-03-25 Richard Biener <rguenther@suse.de>
3033 * gimplify-me.c (gimple_regimplify_operands): Update the
3036 2014-03-25 Martin Jambor <mjambor@suse.cz>
3039 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
3040 (lto_output_varpool_node): Likewise.
3041 (input_overwrite_node): Likewise.
3042 (input_varpool_node): Likewise.
3044 2014-03-25 Richard Biener <rguenther@suse.de>
3046 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
3047 (run_gcc): Likewise.
3049 2014-03-25 Jakub Jelinek <jakub@redhat.com>
3051 * combine.c (simplify_compare_const): Add MODE argument.
3052 Handle mode_width 0 as very large mode_width.
3053 (try_combine, simplify_comparison): Adjust callers.
3055 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
3056 type to avoid signed integer overflow.
3057 * explow.c (plus_constant): Likewise.
3059 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
3061 * doc/generic.texi: Correct typos.
3063 2014-03-24 Tobias Burnus <burnus@net-b.de>
3065 * doc/invoke.texi (-flto): Expand section about
3066 using static libraries with LTO.
3068 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3070 PR rtl-optimization/60501
3071 * optabs.def (addptr3_optab): New optab.
3072 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
3073 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
3074 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
3076 * lra.c (emit_add3_insn): Use the addptr pattern if available.
3078 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
3080 2014-03-24 Ulrich Drepper <drepper@gmail.com>
3082 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
3085 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
3086 (_mm256_undefined_ps): Define.
3087 (_mm256_undefined_pd): Define.
3088 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
3089 (_mm_undefined_pd): Define.
3090 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
3091 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
3092 (_mm512_undefined_ps): Define.
3093 (_mm512_undefined_pd): Define.
3094 Use _mm*_undefined_*.
3095 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
3097 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
3099 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
3100 (lshr_simd): DI mode added.
3101 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
3102 (aarch64_ushr_simddi): Likewise.
3103 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
3104 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
3105 (vshrd_n_u64): Likewise.
3107 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3109 * Makefile.in (s-macro_list): Depend on cc1.
3111 2014-03-23 Teresa Johnson <tejohnson@google.com>
3113 * ipa-utils.c (ipa_print_order): Use specified dump file.
3115 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
3117 PR rtl-optimization/60601
3118 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
3120 * gcc.c (eval_spec_function): Initialize save_growing_value.
3122 2014-03-22 Jakub Jelinek <jakub@redhat.com>
3125 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
3126 code == MINUS_EXPR, never swap op0 with op1.
3128 * toplev.c (init_local_tick): Avoid signed integer multiplication
3130 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
3131 shift by first operand's bitsize.
3133 2014-03-21 Jakub Jelinek <jakub@redhat.com>
3136 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
3138 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
3139 TARGET_ISA_64BIT_P(x).
3141 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3143 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
3144 pattern for vector nor instead of subtract from splat(-1).
3145 (altivec_expand_vec_perm_const_le): Likewise.
3147 2014-03-21 Richard Henderson <rth@twiddle.net>
3150 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
3151 related insns after epilogue_completed.
3153 2014-03-21 Martin Jambor <mjambor@suse.cz>
3156 * cgraph.h (symtab_node): New flag body_removed.
3157 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
3158 when removing bodies.
3159 * symtab.c (dump_symtab_base): Dump body_removed flag.
3160 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
3161 had their bodies removed.
3163 2014-03-21 Martin Jambor <mjambor@suse.cz>
3166 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
3169 2014-03-21 Richard Biener <rguenther@suse.de>
3171 PR tree-optimization/60577
3172 * tree-core.h (struct tree_base): Document nothrow_flag use
3173 in VAR_DECL_NONALIASED.
3174 * tree.h (VAR_DECL_NONALIASED): New.
3175 (may_be_aliased): Adjust.
3176 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
3178 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3180 * expr.c (expand_expr_real_1): Remove outdated comment.
3182 2014-03-20 Jakub Jelinek <jakub@redhat.com>
3185 * ira.c (adjust_cleared_regs): Call copy_rtx on
3186 *reg_equiv[REGNO (loc)].src_p before passing it to
3187 simplify_replace_fn_rtx.
3190 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
3191 into CONST, put pic register as first operand of PLUS. Use
3192 gen_const_mem for both 32-bit and 64-bit PIC got loads.
3194 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3196 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
3198 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3200 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
3201 around for store forwarding issue in the FPU on the UT699.
3202 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
3203 loads and operations if -mfix-ut699 is specified.
3204 (divtf3_hq): Tweak attribute.
3205 (sqrttf2_hq): Likewise.
3207 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3209 * calls.c (store_one_arg): Remove incorrect const qualification on the
3210 type of the temporary.
3211 * cfgexpand.c (expand_return): Likewise.
3212 * expr.c (expand_constructor): Likewise.
3213 (expand_expr_real_1): Likewise.
3215 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3217 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
3220 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
3223 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
3225 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
3227 * config/arm/aarch-common-protos.h
3228 (alu_cost_table): Fix spelling of "extend".
3229 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
3231 2014-03-19 Richard Biener <rguenther@suse.de>
3234 * tree-core.h (tree_type_common): Re-order pointer members
3235 to reduce recursion depth during GC walks.
3237 2014-03-19 Marek Polacek <polacek@redhat.com>
3240 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
3241 before accessing it.
3243 2014-03-19 Richard Biener <rguenther@suse.de>
3246 * lto-streamer-in.c (input_function): In WPA stage do not drop
3249 2014-03-19 Jakub Jelinek <jakub@redhat.com>
3251 PR tree-optimization/60559
3252 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
3253 with build_zero_cst assignment.
3255 2014-03-18 Kai Tietz <ktietz@redhat.com>
3257 PR rtl-optimization/56356
3258 * sdbout.c (sdbout_parms): Verify that parms'
3259 incoming argument is valid.
3260 (sdbout_reg_parms): Likewise.
3262 2014-03-18 Richard Henderson <rth@redhat.com>
3265 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
3266 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
3267 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
3269 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
3271 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
3272 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
3273 Italicize plugin event names in description. Explain that
3274 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
3275 Remind that no GCC functions should be called after PLUGIN_FINISH.
3276 Explain what pragmas with expansion are.
3278 2014-03-18 Martin Liska <mliska@suse.cz>
3280 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
3281 gimple call statement is update.
3282 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
3283 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
3285 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3288 * ubsan.c (ubsan_instrument_unreachable): Call
3289 initialize_sanitizer_builtins.
3290 (ubsan_pass): Likewise.
3293 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
3294 varpool_finalize_decl instead of rest_of_decl_compilation.
3296 2014-03-18 Richard Biener <rguenther@suse.de>
3298 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
3299 by using bitmap_and_compl instead of bitmap_and_compl_into.
3300 (df_rd_transfer_function): Likewise.
3302 2014-03-18 Richard Biener <rguenther@suse.de>
3304 * doc/lto.texi (fresolution): Fix typo.
3306 2014-03-18 Richard Biener <rguenther@suse.de>
3308 * doc/invoke.texi (flto): Update for changes in 4.9.
3310 2014-03-18 Richard Biener <rguenther@suse.de>
3312 * doc/loop.texi: Remove section on the removed lambda framework.
3313 Update loop docs with recent changes in preserving loop structure.
3315 2014-03-18 Richard Biener <rguenther@suse.de>
3317 * doc/lto.texi (-fresolution): Document.
3319 2014-03-18 Richard Biener <rguenther@suse.de>
3321 * doc/contrib.texi: Adjust my name.
3323 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3326 * internal-fn.c: Include diagnostic-core.h.
3327 (expand_BUILTIN_EXPECT): New function.
3328 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
3329 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
3330 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
3331 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
3333 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
3334 Revert 3 argument __builtin_expect code.
3335 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
3336 * gimple-fold.c (gimple_fold_call): Likewise.
3337 * tree.h (fold_builtin_expect): New prototype.
3338 * builtins.c (build_builtin_expect_predicate): Add predictor
3339 argument, if non-NULL, create 3 argument __builtin_expect.
3340 (fold_builtin_expect): No longer static. Add ARG2 argument,
3341 pass it through to build_builtin_expect_predicate.
3342 (fold_builtin_2): Adjust caller.
3343 (fold_builtin_3): Handle BUILT_IN_EXPECT.
3344 * internal-fn.def (BUILTIN_EXPECT): New.
3346 2014-03-18 Tobias Burnus <burnus@net-b.de>
3349 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
3350 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
3351 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
3353 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
3356 * predict.c (combine_predictions_for_bb): Fix up formatting.
3357 (expr_expected_value_1, expr_expected_value): Add predictor argument,
3358 fill what it points to if non-NULL.
3359 (tree_predict_by_opcode): Adjust caller, use the predictor.
3360 * predict.def (PRED_COMPARE_AND_SWAP): Add.
3362 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
3364 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
3365 proper constant for the store mode.
3367 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
3369 * symtab.c (change_decl_assembler_name): Fix transparent alias
3372 2014-03-16 Renlin Li <Renlin.Li@arm.com>
3374 * config/aarch64/aarch64.c: Correct the comments about the
3375 aarch64 stack layout.
3377 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
3379 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
3380 check for GF_OMP_FOR_KIND_FOR.
3382 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
3384 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
3385 ymm and zmm register names.
3387 2014-03-17 Jakub Jelinek <jakub@redhat.com>
3390 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
3391 note creation for the 2010-08-31 changes.
3393 2014-03-17 Marek Polacek <polacek@redhat.com>
3396 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
3397 as -fno-tree-loop-vectorize.
3398 (expand_omp_simd): Likewise.
3400 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
3402 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
3403 (eligible_for_call_delay): New prototype.
3404 * config/sparc/sparc.c (tls_call_delay): Rename into...
3405 (eligible_for_call_delay): ...this. Return false if the instruction
3406 cannot be put in the delay slot of a branch.
3407 (eligible_for_restore_insn): Simplify.
3408 (eligible_for_return_delay): Return false if the instruction cannot be
3409 put in the delay slot of a branch and simplify.
3410 (eligible_for_sibcall_delay): Return false if the instruction cannot be
3411 put in the delay slot of a branch.
3412 * config/sparc/sparc.md (fix_ut699): New attribute.
3413 (tls_call_delay): Delete.
3414 (in_call_delay): Reimplement.
3415 (eligible_for_sibcall_delay): Rename into...
3416 (in_sibcall_delay): ...this.
3417 (eligible_for_return_delay): Rename into...
3418 (in_return_delay): ...this.
3419 (in_branch_delay): Reimplement.
3420 (in_uncond_branch_delay): Delete.
3421 (in_annul_branch_delay): Delete.
3423 2014-03-14 Richard Henderson <rth@redhat.com>
3426 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
3427 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
3428 (*floathi<X87MODEF>2_i387_with_temp): Remove.
3429 (floathi splitters): Remove.
3430 (float<SWI48x>xf2): New pattern.
3431 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
3432 code that tried to handle DImode for 32-bit, but which was excluded
3433 by the pattern's condition. Drop allocation of stack temporary.
3434 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
3435 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
3436 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
3437 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
3438 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
3439 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
3440 (*float<SWI48><MODEF>2_sse_interunit): Remove.
3441 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
3442 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
3443 (*float<SWI48x><X87MODEF>2_i387): Remove.
3444 (all float _with_temp splitters): Remove.
3445 (*float<SWI48x><MODEF>2_i387): New pattern.
3446 (*float<SWI48><MODEF>2_sse): New pattern.
3447 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
3448 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
3450 2014-03-14 Jakub Jelinek <jakub@redhat.com>
3451 Marek Polacek <polacek@redhat.com>
3454 * common.opt (dump_base_name_prefixed): New Variable.
3455 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
3456 if x_dump_base_name_prefixed is already set, set it at the end.
3458 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
3460 PR rtl-optimization/60508
3461 * lra-constraints.c (get_reload_reg): Add new parameter
3463 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
3464 Pass the new parameter values.
3466 2014-03-14 Richard Biener <rguenther@suse.de>
3468 * common.opt: Revert unintented changes from r205065.
3471 2014-03-14 Richard Biener <rguenther@suse.de>
3474 * cfghooks.c (split_block): Properly adjust all loops the
3475 block was a latch of.
3477 2014-03-14 Martin Jambor <mjambor@suse.cz>
3480 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
3483 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
3486 * config/avr/avr.c (avr_set_current_function): Pass function name
3487 through default_strip_name_encoding before sanity checking instead
3488 of skipping the first char of the assembler name.
3490 2014-03-13 Richard Henderson <rth@redhat.com>
3493 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
3494 (ix86_force_to_memory, ix86_free_from_memory): Remove.
3495 * config/i386/i386-protos.h: Likewise.
3496 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
3497 in the expander instead of a splitter.
3498 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
3499 any possibility of requiring a memory.
3500 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
3501 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
3502 (fp branch splitters): Update for ix86_split_fp_branch.
3503 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
3504 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
3505 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
3506 (*fop_<MODEF>_2_i387): Remove f/r alternative.
3507 (*fop_<MODEF>_3_i387): Likewise.
3508 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
3509 (splitters for the fop_* register patterns): Remove.
3510 (fscalexf4_i387): Rename from *fscalexf4_i387.
3511 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
3513 2014-03-13 Jakub Jelinek <jakub@redhat.com>
3515 PR tree-optimization/59779
3516 * tree-dfa.c (get_ref_base_and_extent): Use double_int
3517 type for bitsize and maxsize instead of HOST_WIDE_INT.
3519 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
3521 PR rtl-optimization/57320
3522 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
3523 the CFG after thread_prologue_and_epilogue_insns.
3525 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
3527 PR rtl-optimization/57189
3528 * lra-constraints.c (process_alt_operands): Disfavor spilling
3531 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
3533 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
3535 2014-03-13 Jakub Jelinek <jakub@redhat.com>
3537 PR tree-optimization/59025
3539 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
3540 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
3542 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
3545 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
3546 calls of avr_out_plus_1.
3548 2014-03-13 Bin Cheng <bin.cheng@arm.com>
3550 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
3551 BB's single pred and update the father loop's latch info later.
3553 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
3555 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
3559 (VEC_base): Likewise.
3560 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
3561 registers, we need to swap double words in little endian mode.
3563 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
3564 to be a container mode for 128-bit integer operations added in ISA
3565 2.07. Unlike TImode and PTImode, the preferred register set is
3566 the Altivec/VMX registers for the 128-bit operations.
3568 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
3570 (rs6000_split_128bit_ok_p): Likewise.
3572 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
3573 macros for creating ISA 2.07 normal and overloaded builtin
3574 functions with 3 arguments.
3575 (BU_P8V_OVERLOAD_3): Likewise.
3576 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
3577 for use as overloaded functions.
3578 (VPERM_1TI_UNS): Likewise.
3579 (VSEL_1TI): Likewise.
3580 (VSEL_1TI_UNS): Likewise.
3581 (ST_INTERNAL_1ti): Likewise.
3582 (LD_INTERNAL_1ti): Likewise.
3583 (XXSEL_1TI): Likewise.
3584 (XXSEL_1TI_UNS): Likewise.
3585 (VPERM_1TI): Likewise.
3586 (VPERM_1TI_UNS): Likewise.
3587 (XXPERMDI_1TI): Likewise.
3588 (SET_1TI): Likewise.
3589 (LXVD2X_V1TI): Likewise.
3590 (STXVD2X_V1TI): Likewise.
3591 (VEC_INIT_V1TI): Likewise.
3592 (VEC_SET_V1TI): Likewise.
3593 (VEC_EXT_V1TI): Likewise.
3594 (EQV_V1TI): Likewise.
3595 (NAND_V1TI): Likewise.
3596 (ORC_V1TI): Likewise.
3597 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
3598 added in ISA 2.07. Add both normal 'altivec' builtins, and the
3600 (VADDUQM): Likewise.
3601 (VSUBCUQ): Likewise.
3602 (VADDEUQM): Likewise.
3603 (VADDECUQ): Likewise.
3604 (VSUBEUQM): Likewise.
3605 (VSUBECUQ): Likewise.
3607 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
3608 __int128_t and __uint128_t types.
3609 (__uint128_type): Likewise.
3610 (altivec_categorize_keyword): Add support for vector __int128_t,
3611 vector __uint128_t, vector __int128, and vector unsigned __int128
3612 as a container type for TImode operations that need to be done in
3613 VSX/Altivec registers.
3614 (rs6000_macro_to_expand): Likewise.
3615 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
3616 to support 128-bit integer instructions vaddcuq, vadduqm,
3617 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
3618 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
3620 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
3621 for V1TImode, and set up preferences to use VSX/Altivec registers.
3622 Setup VSX reload handlers.
3623 (rs6000_debug_reg_global): Likewise.
3624 (rs6000_init_hard_regno_mode_ok): Likewise.
3625 (rs6000_preferred_simd_mode): Likewise.
3626 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
3627 (easy_altivec_constant): Likewise.
3628 (output_vec_const_move): Likewise.
3629 (rs6000_expand_vector_set): Convert V1TImode set and extract to
3631 (rs6000_expand_vector_extract): Likewise.
3632 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
3634 (rs6000_const_vec): Add support for V1TImode.
3635 (rs6000_emit_le_vsx_load): Swap double words when loading or
3636 storing TImode/V1TImode.
3637 (rs6000_emit_le_vsx_store): Likewise.
3638 (rs6000_emit_le_vsx_move): Likewise.
3639 (rs6000_emit_move): Add support for V1TImode.
3640 (altivec_expand_ld_builtin): Likewise.
3641 (altivec_expand_st_builtin): Likewise.
3642 (altivec_expand_vec_init_builtin): Likewise.
3643 (altivec_expand_builtin): Likewise.
3644 (rs6000_init_builtins): Add support for V1TImode type. Add
3645 support for ISA 2.07 128-bit integer builtins. Define type names
3646 for the VSX/Altivec vector types.
3647 (altivec_init_builtins): Add support for overloaded vector
3648 functions with V1TImode type.
3649 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
3650 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
3652 (rs6000_split_128bit_ok_p): Likewise.
3653 (rs6000_handle_altivec_attribute): Create V1TImode from vector
3654 __int128_t and vector __uint128_t.
3656 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
3657 and mode attributes.
3664 (VS_scalar): Likewise.
3665 (VS_double): Likewise.
3666 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
3668 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
3669 we support the ISA 2.07 128-bit integer arithmetic instructions.
3670 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
3671 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
3672 and TImode types for use with the builtin functions.
3673 (V1TI_type_node): Likewise.
3674 (unsigned_V1TI_type_node): Likewise.
3675 (intTI_type_internal_node): Likewise.
3676 (uintTI_type_internal_node): Likewise.
3678 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
3679 128-bit builtin functions.
3680 (UNSPEC_VADDEUQM): Likewise.
3681 (UNSPEC_VADDECUQ): Likewise.
3682 (UNSPEC_VSUBCUQ): Likewise.
3683 (UNSPEC_VSUBEUQM): Likewise.
3684 (UNSPEC_VSUBECUQ): Likewise.
3685 (VM): Add V1TImode to vector mode iterators.
3687 (VI_unit): Likewise.
3688 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
3689 (altivec_vaddcuq): Likewise.
3690 (altivec_vsubuqm): Likewise.
3691 (altivec_vsubcuq): Likewise.
3692 (altivec_vaddeuqm): Likewise.
3693 (altivec_vaddecuq): Likewise.
3694 (altivec_vsubeuqm): Likewise.
3695 (altivec_vsubecuq): Likewise.
3697 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
3699 (BOOL_128): Likewise.
3700 (BOOL_REGS_OUTPUT): Likewise.
3701 (BOOL_REGS_OP1): Likewise.
3702 (BOOL_REGS_OP2): Likewise.
3703 (BOOL_REGS_UNARY): Likewise.
3704 (BOOL_REGS_AND_CR0): Likewise.
3706 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
3707 128-bit integer builtin support.
3708 (vec_vadduqm): Likewise.
3709 (vec_vaddecuq): Likewise.
3710 (vec_vaddeuqm): Likewise.
3711 (vec_vsubecuq): Likewise.
3712 (vec_vsubeuqm): Likewise.
3713 (vec_vsubcuq): Likewise.
3714 (vec_vsubuqm): Likewise.
3716 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3717 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
3718 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
3719 128-bit integer add/subtract to ISA 2.07.
3721 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
3723 * config/arc/arc.c (arc_predicate_delay_insns):
3724 Fix third argument passed to conditionalize_nonjump.
3726 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
3728 * config/aarch64/aarch64-builtins.c
3729 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
3730 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
3731 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
3732 instead of __builtin_lfloor.
3733 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
3735 2014-03-12 Jakub Jelinek <jakub@redhat.com>
3737 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
3738 (tree_ssa_ifcombine_bb_1): New function.
3739 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
3740 is an empty forwarder block to then_bb or vice versa and then_bb
3741 and else_bb are effectively swapped.
3743 2014-03-12 Christian Bruel <christian.bruel@st.com>
3746 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
3747 REG_CFA_DEF_CFA note.
3748 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
3749 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
3751 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
3753 PR tree-optimization/60454
3754 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
3756 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3758 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
3759 Do not define target_cpu_default2 to generic.
3760 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
3761 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
3762 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
3764 2014-03-12 Jakub Jelinek <jakub@redhat.com>
3765 Marc Glisse <marc.glisse@inria.fr>
3767 PR tree-optimization/60502
3768 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
3769 instead of build_low_bits_mask.
3771 2014-03-12 Jakub Jelinek <jakub@redhat.com>
3774 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
3775 if there are multiple uses, but op doesn't live on E edge.
3776 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
3777 clobber stmts before __builtin_unreachable.
3779 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
3781 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
3782 hard_frame_pointer_rtx.
3783 * cse.c (cse_insn): Remove volatile check.
3784 * cselib.c (cselib_process_insn): Likewise.
3785 * dse.c (scan_insn): Likewise.
3787 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
3789 * config/arc/arc.c (conditionalize_nonjump): New function,
3791 (arc_ifcvt): ... this.
3792 (arc_predicate_delay_insns): Use it.
3794 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
3796 * config/arc/predicates.md (extend_operand): During/after reload,
3797 allow const_int_operand.
3798 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
3799 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
3800 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
3802 (umulsi3_highpart_i): Likewise.
3804 2014-03-11 Richard Biener <rguenther@suse.de>
3806 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
3807 Add asserts to guard possible wrong-code bugs.
3809 2014-03-11 Richard Biener <rguenther@suse.de>
3811 PR tree-optimization/60429
3812 PR tree-optimization/60485
3813 * tree-ssa-structalias.c (set_union_with_increment): Properly
3814 take into account all fields that overlap the shifted vars.
3815 (do_sd_constraint): Likewise.
3816 (do_ds_constraint): Likewise.
3817 (get_constraint_for_ptr_offset): Likewise.
3819 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
3821 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
3822 (nios2_compute_frame_layout):
3823 Add calculation of cfun->machine->fp_save_offset.
3824 (nios2_expand_prologue): Correct setting of frame pointer register
3826 (nios2_expand_epilogue): Update recovery of stack pointer from
3827 frame pointer accordingly.
3828 (nios2_initial_elimination_offset): Update calculation of offset
3829 for eliminating to HARD_FRAME_POINTER_REGNUM.
3831 2014-03-10 Jakub Jelinek <jakub@redhat.com>
3834 * ipa.c (symtab_remove_unreachable_nodes): Don't call
3835 cgraph_get_create_node on VAR_DECLs.
3837 2014-03-10 Richard Biener <rguenther@suse.de>
3840 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
3842 2014-03-08 Douglas B Rupp <rupp@gnat.com>
3844 * config/vms/vms.opt (vms_float_format): New variable.
3846 2014-03-08 Tobias Burnus <burnus@net-b.de>
3848 * doc/invoke.texi (-fcilkplus): Update implementation status.
3850 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
3851 Richard Biener <rguenther@suse.de>
3853 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
3854 consistently accross all TUs.
3855 (run_gcc): Enable -fshort-double automatically at link at link-time
3856 and disallow override.
3858 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
3861 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
3862 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
3863 if they can't be used.
3865 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3867 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
3868 for Solaris 11/x86 ld.
3869 * configure: Regenerate.
3871 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3873 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
3874 (LIB_TLS_SPEC): Save as ld_tls_libs.
3875 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
3876 (HAVE_AS_IX86_TLSLDM): New test.
3877 * configure, config.in: Regenerate.
3878 * config/i386/i386.c (legitimize_tls_address): Fall back to
3879 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
3880 cannot support TLS_MODEL_LOCAL_DYNAMIC.
3881 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
3882 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
3884 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
3886 * common.opt (fira-loop-pressure): Mark as optimization.
3888 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
3890 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
3891 an OpenMP mappable type.
3893 2014-03-06 Matthias Klose <doko@ubuntu.com>
3895 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
3896 MULTILIB_OSDIRNAMES is not defined.
3898 2014-03-06 Jakub Jelinek <jakub@redhat.com>
3899 Meador Inge <meadori@codesourcery.com>
3902 * config/arm/arm.c (arm_tls_symbol_p): Remove.
3903 (arm_legitimize_address): Call legitimize_tls_address for any
3904 arm_tls_referenced_p expression, handle constant addend. Call it
3905 before testing for !TARGET_ARM.
3906 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
3908 2014-03-06 Richard Biener <rguenther@suse.de>
3914 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
3916 * tree-streamer.c (record_common_node): Assert we don't record
3917 nodes with type double.
3918 (preload_common_node): Skip type double, complex double and double
3919 pointer since it is now frontend dependent due to fshort-double option.
3921 2014-03-06 Richard Biener <rguenther@suse.de>
3923 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
3924 or -fno-lto is specified and the linker has full plugin support.
3925 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
3926 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
3927 * lto-wrapper.c (merge_and_complain): Merge compile-time
3928 optimization levels.
3929 (run_gcc): And pass it through to the link options.
3931 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
3935 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
3937 * cselib.c (remove_useless_values): Skip to avoid quadratic
3938 behavior if the condition moved from...
3939 (cselib_process_insn): ... here holds.
3941 2014-03-05 Jakub Jelinek <jakub@redhat.com>
3944 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
3945 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
3948 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
3949 (TM_H): Add x86-tune.def.
3951 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3953 * config/aarch64/aarch64.c (generic_tunings):
3954 Use cortexa57_extra_costs.
3956 2014-03-05 Jakub Jelinek <jakub@redhat.com>
3959 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
3960 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
3961 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
3964 2014-03-04 Heiher <r@hev.cc>
3966 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
3967 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
3969 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
3971 * config/i386/predicates.md (const2356_operand): Change to ...
3972 (const2367_operand): ... this.
3973 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
3975 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
3976 (*avx512pf_scatterpf<mode>sf): Ditto.
3977 (avx512pf_scatterpf<mode>df): Ditto.
3978 (*avx512pf_scatterpf<mode>df_mask): Ditto.
3979 (*avx512pf_scatterpf<mode>df): Ditto.
3980 * config/i386/i386.c (ix86_expand_builtin): Update
3981 incorrect hint operand error message.
3983 2014-03-04 Richard Biener <rguenther@suse.de>
3985 * lto-section-in.c (lto_get_section_data): Fix const cast.
3987 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
3989 * tree-streamer.c (record_common_node): Assert we don't record
3990 nodes with type double.
3991 (preload_common_node): Skip type double, complex double and double
3992 pointer since it is now frontend dependent due to fshort-double option.
3994 2014-03-04 Richard Biener <rguenther@suse.de>
3997 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
3998 (lto_input_toplevel_asms): Likewise.
3999 * lto-section-in.c (lto_get_section_data): Instead do it here
4002 2014-03-04 Richard Biener <rguenther@suse.de>
4004 PR tree-optimization/60382
4005 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
4006 dead PHIs a reduction.
4008 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
4010 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
4012 (_mm_prefetch): Move out of GCC target("sse") pragma.
4013 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
4014 GCC target("prfchw") pragma.
4015 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
4017 * config/i386/i386.c (ix86_option_override_internal): Enable
4018 -mprfchw with -mprefetchwt1.
4020 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
4022 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
4025 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
4027 * opts.h (CL_PCH_IGNORE): Define.
4028 * targhooks.c (option_affects_pch_p):
4029 Return false for options that have CL_PCH_IGNORE set.
4030 * opt-functions.awk: Process PchIgnore.
4031 * doc/options.texi: Document PchIgnore.
4033 * config/arc/arc.opt (misize): Add PchIgnore property.
4035 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4037 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
4038 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
4039 constraint on constants to permit them being loaded into
4040 GENERAL_REGS or BASE_REGS.
4042 2014-03-03 Nick Clifton <nickc@redhat.com>
4044 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
4045 anti-cacnonical alternatives.
4046 (negandhi3_real): New pattern.
4047 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
4049 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
4051 * config/avr/avr-mcus.def: Remove atxmega16x1.
4052 * config/avr/avr-tables.opt: Regenerate.
4053 * config/avr/t-multilib: Regenerate.
4054 * doc/avr-mmcu.texi: Regenerate.
4056 2014-03-03 Tobias Grosser <tobias@grosser.es>
4057 Mircea Namolaru <mircea.namolaru@inria.fr>
4059 PR tree-optimization/58028
4060 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
4063 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4065 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
4066 not handled by recognizers.
4068 2014-03-03 Jakub Jelinek <jakub@redhat.com>
4071 * function.c (expand_function_end): Don't emit
4072 clobber_return_register sequence if clobber_after is a BARRIER.
4073 * cfgexpand.c (construct_exit_block): Append instructions before
4074 return_label to prev_bb.
4076 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4078 * config/rs6000/constraints.md: Document reserved use of "wc".
4080 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
4083 * ipa.c (function_and_variable_visibility): When dissolving comdat
4084 group, also set all symbols to local.
4086 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
4091 2013-12-14 Jan Hubicka <jh@suse.cz>
4093 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
4095 2014-03-02 Jon Beniston <jon@beniston.com>
4101 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
4102 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
4103 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
4104 (simple_return, *simple_return): New patterns
4105 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
4106 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
4108 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
4110 * dwarf2out.c (gen_subprogram_die): Tidy.
4112 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
4115 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
4116 (*mov_t_msb_neg_negc): ... this new insn.
4118 2014-02-28 Jason Merrill <jason@redhat.com>
4121 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
4124 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
4127 * dwarf2out.c (decltype_auto_die): New static.
4128 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
4129 (gen_type_die_with_usage): Handle 'decltype(auto)'.
4130 (is_cxx_auto): Likewise.
4132 2014-02-28 Ian Bolton <ian.bolton@arm.com>
4134 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
4135 we are not using general regs only.
4137 2014-02-28 Richard Biener <rguenther@suse.de>
4140 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
4141 previous fix and only allow to remove trivial pre-headers
4142 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
4143 (remove_forwarder_block): Properly update the latch of a loop.
4145 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4148 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
4149 (cselib_preserved_hash_table): New.
4150 (preserve_constants_and_equivs): Move preserved vals to it.
4151 (cselib_find_slot): Look it up first.
4152 (cselib_init): Initialize it.
4153 (cselib_finish): Release it.
4154 (dump_cselib_table): Dump it.
4156 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4159 * cselib.c (remove_useless_values): Skip to avoid quadratic
4160 behavior if the condition moved from...
4161 (cselib_process_insn): ... here holds.
4163 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4166 * var-tracking.c (vt_initialize): Apply the same condition to
4167 preserve the CFA base value.
4169 2014-02-28 Joey Ye <joey.ye@arm.com>
4172 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
4173 if reload in progress or completed.
4175 2014-02-28 Tobias Burnus <burnus@net-b.de>
4178 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
4181 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
4183 * doc/tm.texi.in (Condition Code Status): Update documention for
4184 relative locations of cc0-setter and cc0-user.
4186 2014-02-27 Jeff Law <law@redhat.com>
4188 PR rtl-optimization/52714
4189 * combine.c (try_combine): When splitting an unrecognized PARALLEL
4190 into two independent simple sets, if I3 is a jump, ensure the
4191 pattern we place into I3 is a (set (pc) ...).
4193 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
4194 Jeff Law <law@redhat.com>
4196 PR rtl-optimization/49847
4197 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
4198 are in different blocks.
4199 * doc/tm.texi (Condition Code Status): Update documention for
4200 relative locations of cc0-setter and cc0-user.
4202 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
4205 * lra.c (lra_emit_add): Check SUBREG too.
4207 2014-02-27 Andreas Schwab <schwab@suse.de>
4209 * config/m68k/m68k.c (m68k_option_override): Disable
4210 -flive-range-shrinkage for classic m68k.
4211 (m68k_override_options_after_change): Likewise.
4213 2014-02-27 Marek Polacek <polacek@redhat.com>
4216 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
4217 -Wmaybe-uninitialized.
4219 2014-02-27 Alan Modra <amodra@gmail.com>
4222 * reload1.c (emit_input_reload_insns): When reload_override_in,
4223 set old to rl->in_reg when rl->in_reg is a subreg.
4225 2014-02-26 Richard Biener <rguenther@suse.de>
4228 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
4230 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4232 * common/config/i386/predicates.md (const1256_operand): Remove.
4233 (const2356_operand): New.
4234 (const_1_to_2_operand): Remove.
4235 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
4236 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
4237 (*avx512pf_gatherpf<mode>sf): Ditto.
4238 (avx512pf_gatherpf<mode>df): Ditto.
4239 (*avx512pf_gatherpf<mode>df_mask): Ditto.
4240 (*avx512pf_gatherpf<mode>df): Ditto.
4241 (avx512pf_scatterpf<mode>sf): Ditto.
4242 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
4243 (*avx512pf_scatterpf<mode>sf): Ditto.
4244 (avx512pf_scatterpf<mode>df): Ditto.
4245 (*avx512pf_scatterpf<mode>df_mask): Ditto.
4246 (*avx512pf_scatterpf<mode>df): Ditto.
4247 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
4249 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
4251 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
4252 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
4253 (_mm512_mask_testn_epi64_mask): Move to ...
4254 * config/i386/avx512cdintrin.h: Here.
4255 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
4256 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
4257 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
4258 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
4259 TARGET_AVX512F from TARGET_AVX512CD.
4261 2014-02-26 Richard Biener <rguenther@suse.de>
4264 * ipa.c (walk_polymorphic_call_targets): Properly guard
4265 call to inline_update_overall_summary.
4267 2014-02-26 Bin Cheng <bin.cheng@arm.com>
4270 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
4271 and latches only if requested. Fix latch if it is removed.
4272 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
4273 LOOPS_HAVE_PREHEADERS.
4275 2014-02-25 Andrew Pinski <apinski@cavium.com>
4277 * builtins.c (expand_builtin_thread_pointer): Create a new target
4278 when the target is NULL.
4280 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
4282 PR rtl-optimization/60317
4283 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4284 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4285 * lra-assigns.c: Include params.h.
4286 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
4287 other reload pseudos considerations.
4289 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4291 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
4292 to use canonical form for nor<mode>3.
4294 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4297 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
4300 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4302 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
4303 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
4304 (ix86_handle_option): Handle OPT_mprefetchwt1.
4305 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
4306 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
4308 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
4309 OPTION_MASK_ISA_PREFETCHWT1.
4310 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
4311 (PTA_PREFETCHWT1): New.
4312 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
4313 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
4314 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
4315 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
4316 (*prefetch_avx512pf_<mode>_: Change into ...
4317 (*prefetch_prefetchwt1_<mode>: This.
4318 * config/i386/i386.opt (mprefetchwt1): New.
4319 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
4320 (_mm_prefetch): Handle intent to write.
4321 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
4323 2014-02-25 Richard Biener <rguenther@suse.de>
4326 * emit-rtl.c (mem_attrs_htab): Remove.
4327 (mem_attrs_htab_hash): Likewise.
4328 (mem_attrs_htab_eq): Likewise.
4329 (set_mem_attrs): Always allocate new mem-attrs when something changed.
4330 (init_emit_once): Do not allocate mem_attrs_htab.
4332 2014-02-25 Richard Biener <rguenther@suse.de>
4335 * lto-opts.c (lto_write_options): Output non-explicit conservative
4336 -fwrapv, -fno-trapv and -fno-strict-overflow.
4337 * lto-wrapper.c (merge_and_complain): Handle merging those options.
4338 (run_gcc): And pass them through.
4340 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4342 * sel-sched.c (calculate_new_fences): New parameter ptime.
4343 Calculate it as a maximum over all fence cycles.
4344 (sel_sched_region_2): Adjust the call to calculate_new_fences.
4345 Print the final schedule timing when sched_verbose.
4347 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4349 PR rtl-optimization/60292
4350 * sel-sched.c (fill_vec_av_set): Do not reset target availability
4351 bit fot the fence instruction.
4353 2014-02-24 Alangi Derick <alangiderick@gmail.com>
4355 * calls.h: Fix typo in comment.
4357 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
4359 * config/pa/pa.c (pa_output_move_double): Don't valididate when
4360 adjusting offsetable addresses.
4362 2014-02-24 Guozhi Wei <carrot@google.com>
4364 * sparseset.h (sparseset_pop): Fix the wrong index.
4366 2014-02-24 Walter Lee <walt@tilera.com>
4368 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
4369 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
4371 * common/config/tilegx/tilegx-common.c
4372 (TARGET_DEFAULT_TARGET_FLAGS): Define.
4373 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
4375 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
4376 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
4377 (tilegx_gimplify_va_arg_expr): Handle big endian.
4378 (tilegx_expand_unaligned_load): Ditto.
4379 (tilegx_expand_unaligned_store): Ditto.
4380 (TARGET_RETURN_IN_MSB): New.
4381 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
4382 (TARGET_ENDIAN_DEFAULT): New.
4383 (TARGET_BIG_ENDIAN): Handle big endian.
4384 (BYTES_BIG_ENDIAN): Ditto.
4385 (WORDS_BIG_ENDIAN): Ditto.
4386 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
4389 * config/tilegx/tilegx.md (extv): Handle big endian.
4391 (insn_st<n>): Ditto.
4392 (insn_st<n>_add<bitsuffix>): Ditto.
4393 (insn_stnt<n>): Ditto.
4394 (insn_stnt<n>_add<bitsuffix>):Ditto.
4395 (vec_interleave_highv8qi): Handle big endian.
4396 (vec_interleave_highv8qi_be): New.
4397 (vec_interleave_highv8qi_le): New.
4398 (insn_v1int_h): Handle big endian.
4399 (vec_interleave_lowv8qi): Handle big endian.
4400 (vec_interleave_lowv8qi_be): New.
4401 (vec_interleave_lowv8qi_le): New.
4402 (insn_v1int_l): Handle big endian.
4403 (vec_interleave_highv4hi): Handle big endian.
4404 (vec_interleave_highv4hi_be): New.
4405 (vec_interleave_highv4hi_le): New.
4406 (insn_v2int_h): Handle big endian.
4407 (vec_interleave_lowv4hi): Handle big endian.
4408 (vec_interleave_lowv4hi_be): New.
4409 (vec_interleave_lowv4hi_le): New.
4410 (insn_v2int_l): Handle big endian.
4411 (vec_interleave_highv2si): Handle big endian.
4412 (vec_interleave_highv2si_be): New.
4413 (vec_interleave_highv2si_le): New.
4414 (insn_v4int_h): Handle big endian.
4415 (vec_interleave_lowv2si): Handle big endian.
4416 (vec_interleave_lowv2si_be): New.
4417 (vec_interleave_lowv2si_le): New.
4418 (insn_v4int_l): Handle big endian.
4419 * config/tilegx/tilegx.opt (mbig-endian): New option.
4420 (mlittle-endian): New option.
4421 * doc/install.texi: Document tilegxbe-linux.
4422 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
4424 2014-02-24 Martin Jambor <mjambor@suse.cz>
4427 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
4428 there are no parameter descriptors.
4430 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
4432 PR rtl-optimization/60268
4433 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
4434 initialization to ...
4435 (sched_rgn_init): ... here.
4436 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
4438 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4440 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
4443 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
4445 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
4448 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4450 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
4451 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
4453 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4455 * config/microblaze/predicates.md: Add cmp_op predicate.
4456 * config/microblaze/microblaze.md: Add branch_compare instruction
4457 which uses cmp_op predicate and emits cmp insn before branch.
4458 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
4459 to microblaze_expand_conditional_branch and consolidate logic.
4460 (microblaze_expand_conditional_branch): emit branch_compare
4461 insn instead of handling cmp op separate from branch insn.
4463 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4465 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
4468 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4470 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
4471 define_insn with define_expand and new define_insn
4472 *altivec_lve<VI_char>x_internal.
4473 (altivec_stve<VI_char>x): Replace define_insn with define_expand
4474 and new define_insn *altivec_stve<VI_char>x_internal.
4475 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
4477 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
4479 (altivec_expand_stvex_be): New function.
4481 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
4483 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
4484 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
4485 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
4486 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
4488 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
4491 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
4492 instead of emit_move_insn.
4494 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4496 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
4498 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
4499 gen_altivec_vsumsws.
4501 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4503 * config/rs6000/altivec.md (altivec_lvxl): Rename as
4504 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
4505 (altivec_lvxl_<mode>): New define_expand incorporating
4506 -maltivec=be semantics where needed.
4507 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
4508 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
4509 semantics where needed.
4510 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
4511 (altivec_stvx_<mode>): New define_expand incorporating
4512 -maltivec=be semantics where needed.
4513 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
4514 VM2 iterator instead of V4SI.
4515 (altivec_stvxl_<mode>): New define_expand incorporating
4516 -maltivec=be semantics where needed.
4517 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
4518 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
4519 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
4520 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
4521 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
4522 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
4523 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
4524 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
4525 ALTIVEC_BUILTIN_STVXL.
4526 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
4527 (altivec_expand_stvx_be): Likewise.
4528 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
4529 (altivec_expand_lvx_be): Likewise.
4530 (altivec_expand_stvx_be): Likewise.
4531 (altivec_expand_builtin): Add cases for
4532 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
4533 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
4534 (altivec_init_builtins): Add definitions for
4535 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
4536 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
4538 2014-02-21 Catherine Moore <clm@codesourcery.com>
4540 * doc/invoke.texi (mvirt, mno-virt): Document.
4541 * config/mips/mips.opt (mvirt): New option.
4542 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
4544 2014-02-21 Richard Biener <rguenther@suse.de>
4546 PR tree-optimization/60276
4547 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
4548 (STMT_VINFO_MIN_NEG_DIST): New macro.
4549 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
4550 STMT_VINFO_MIN_NEG_DIST.
4551 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
4552 made for negative dependence distances still hold.
4554 2014-02-21 Richard Biener <rguenther@suse.de>
4557 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
4558 DECL_INITIAL for globals not in the current function context.
4560 2014-02-21 Jakub Jelinek <jakub@redhat.com>
4562 PR tree-optimization/56490
4563 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
4564 * tree-ssa-uninit.c: Include params.h.
4565 (compute_control_dep_chain): Add num_calls argument, return false
4566 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
4567 num_calls to recursive call.
4568 (find_predicates): Change dep_chain into normal array,
4569 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
4570 variable and adjust compute_control_dep_chain caller.
4571 (find_def_preds): Likewise.
4573 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
4575 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
4576 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
4578 2014-02-21 Nick Clifton <nickc@redhat.com>
4580 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
4581 (pushhi1): Likewise.
4582 (popqi1): Add mode to pre_dec.
4585 2014-02-21 Jakub Jelinek <jakub@redhat.com>
4587 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
4588 mode for mask of V8SFmode permutation.
4590 2014-02-20 Richard Henderson <rth@redhat.com>
4593 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
4594 a new pseudo for OLDVAL.
4596 2014-02-20 Jakub Jelinek <jakub@redhat.com>
4599 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
4600 gen_reg_rtx if d->testing_p.
4601 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
4602 if d->testing_p and we will certainly return true.
4603 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
4606 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
4608 * emit-rtl.c (gen_reg_rtx): Assert that
4609 crtl->emit.regno_pointer_align_length is non-zero.
4611 2014-02-20 Richard Henderson <rth@redhat.com>
4614 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
4615 on failure the store back into EXPECT.
4617 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
4618 Sandra Loosemore <sandra@codesourcery.com>
4620 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
4621 * config/nios2/nios2.c (nios2_function_profiler): Add
4622 -fPIC (flag_pic == 2) support.
4623 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
4624 (nios2_large_offset_p): New function.
4625 (nios2_unspec_reloc_p): Move up position, update to use
4626 nios2_large_offset_p.
4627 (nios2_unspec_address): Remove function.
4628 (nios2_unspec_offset): New function.
4629 (nios2_large_got_address): New function.
4630 (nios2_got_address): Add large offset support.
4631 (nios2_legitimize_tls_address): Update usage of removed and new
4633 (nios2_symbol_binds_local_p): New function.
4634 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
4635 (nios2_legitimize_address): Update to use nios2_large_offset_p.
4636 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
4637 (nios2_print_operand): Merge H/L processing, add hiadj/lo
4638 processing for (const (unspec ...)).
4639 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
4641 2014-02-20 Richard Biener <rguenther@suse.de>
4643 * tree-cfg.c (replace_uses_by): Mark altered BBs before
4644 doing the substitution.
4645 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
4647 2014-02-20 Martin Jambor <mjambor@suse.cz>
4650 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
4651 info when checking whether lattices are bottom.
4653 2014-02-20 Richard Biener <rguenther@suse.de>
4656 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
4659 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
4662 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
4663 parameter specifying the scaling.
4664 (inline_call): Update.
4665 (want_inline_recursively): Guard division by zero.
4666 (recursive_inlining): Update.
4667 * ipa-inline.h (clone_inlined_nodes): Update.
4669 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
4672 * config/i386/i386.c (classify_argument): Pass structures of size
4673 64 bytes or less in register.
4675 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
4676 Kirill Yukhin <kirill.yukhin@intel.com>
4678 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
4679 (_mm_rcp28_round_ss): Ditto.
4680 (_mm_rsqrt28_round_sd): Ditto.
4681 (_mm_rsqrt28_round_ss): Ditto.
4682 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
4683 (_mm_rcp14_round_ss): Ditto.
4684 (_mm_rsqrt14_round_sd): Ditto.
4685 (_mm_rsqrt14_round_ss): Ditto.
4686 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
4687 the first input operand, get rid of match_dup.
4688 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
4690 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
4692 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
4693 operand as the first input operand, set type attribute.
4694 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
4696 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
4697 operand as the first input operand, set type attribute.
4699 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4701 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
4704 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
4707 * config/i386/i386.c (construct_container): Remove TFmode check
4708 for X86_64_INTEGER_CLASS.
4710 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
4713 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
4714 only when -Wpsabi is enabled.
4716 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
4719 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
4720 passing arrays in registers are the same as for structs, so remove the
4721 special case for them.
4723 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
4725 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
4726 destination type, extract only the valid bits if the source type is not
4727 integral and has a different mode.
4729 2014-02-19 Richard Biener <rguenther@suse.de>
4732 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
4735 2014-02-19 Richard Biener <rguenther@suse.de>
4738 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
4739 (ipa_modify_call_arguments): Emit an argument load explicitely and
4740 preserve virtual SSA form there and for the replacement call.
4741 Do not update SSA form nor free dominance info.
4743 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
4745 * ipa.c (function_and_variable_visibility): Also clear WEAK
4746 flag when disolving COMDAT_GROUP.
4748 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
4750 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
4751 * ipa-prop.c (ipa_set_jf_known_type): Return early when
4753 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
4754 do more sanity checks.
4755 (detect_type_change): Return true when giving up early.
4756 (compute_complex_assign_jump_func): Fix type parameter of
4757 ipa_set_ancestor_jf.
4758 (compute_complex_ancestor_jump_func): Likewise.
4759 (update_jump_functions_after_inlining): Fix updating of
4761 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
4763 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
4765 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
4766 inline clones when edge disappears.
4768 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
4771 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
4772 Split 64-bit moves into 2 patterns. Do not allow the use of
4773 direct move for TDmode in little endian, since the decimal value
4774 has little endian bytes within a word, but the 64-bit pieces are
4775 ordered in a big endian fashion, and normal subreg's of TDmode are
4777 (mov<mode>_64bit_dm): Likewise.
4778 (movtd_64bit_nodm): Likewise.
4780 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
4782 PR tree-optimization/60174
4783 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
4784 statement of an SSA_NAME that occurs in an abnormal PHI node.
4786 2014-02-18 Jakub Jelinek <jakub@redhat.com>
4789 * final.c (SEEN_BB): Remove.
4790 (SEEN_NOTE, SEEN_EMITTED): Renumber.
4791 (final_scan_insn): Don't force_source_line on second
4792 NOTE_INSN_BASIC_BLOCK.
4794 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
4797 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
4798 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
4799 (type_natural_mode): Warn ABI change when %zmm register is not
4800 available for AVX512F vector value passing.
4802 2014-02-18 Kai Tietz <ktietz@redhat.com>
4805 * config/i386/i386.c (ix86_expand_prologue): Use value in
4806 rax register as displacement when restoring %r10 or %rax.
4807 Fix wrong offset when restoring both registers.
4809 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
4811 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
4812 assertion with conditional return.
4814 2014-02-18 Jakub Jelinek <jakub@redhat.com>
4815 Uros Bizjak <ubizjak@gmail.com>
4818 * config/i386/driver-i386.c (host_detect_local_cpu): If
4819 YMM state is not saved by the OS, also clear has_f16c. Move
4820 CPUID 0x80000001 handling before YMM state saving checking.
4822 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
4824 PR rtl-optimization/58960
4825 * haifa-sched.c (alloc_global_sched_pressure_data): New,
4826 factored out from ...
4827 (sched_init): ... here.
4828 (free_global_sched_pressure_data): New, factored out from ...
4829 (sched_finish): ... here.
4830 * sched-int.h (free_global_sched_pressure_data): Declare.
4831 * sched-rgn.c (nr_regions_initial): New static global.
4832 (haifa_find_rgns): Initialize it.
4833 (schedule_region): Disable sched-pressure for the newly
4836 2014-02-17 Richard Biener <rguenther@suse.de>
4838 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
4839 release SSA defs of pattern stmts.
4841 2014-02-17 Richard Biener <rguenther@suse.de>
4843 * tree-inline.c (expand_call_inline): Release the virtual
4844 operand defined by the call we are about to inline.
4846 2014-02-17 Richard Biener <rguenther@suse.de>
4848 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
4850 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
4851 Ilya Tocar <ilya.tocar@intel.com>
4853 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
4854 arguments order in builtin.
4855 (_mm512_permutexvar_epi64): Ditto.
4856 (_mm512_mask_permutexvar_epi64): Ditto
4857 (_mm512_maskz_permutexvar_epi32): Ditto
4858 (_mm512_permutexvar_epi32): Ditto
4859 (_mm512_mask_permutexvar_epi32): Ditto
4861 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4863 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
4864 (p8_vmrgow): Likewise.
4866 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4868 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
4871 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
4874 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
4875 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
4876 into 64-bit and 32-bit moves. On 64-bit moves, add support for
4877 using direct move instructions on ISA 2.07. Also adjust
4878 instruction length for 64-bit.
4879 (mov<mode>_64bit, TFmode/TDmode): Likewise.
4880 (mov<mode>_32bit, TFmode/TDmode): Likewise.
4882 2014-02-15 Alan Modra <amodra@gmail.com>
4886 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
4887 find_replacement on parts of insn rtl that might be reloaded.
4889 2014-02-15 Richard Biener <rguenther@suse.de>
4891 PR tree-optimization/60183
4892 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
4893 (tree_ssa_phiprop): Calculate and free post-dominators.
4895 2014-02-14 Jeff Law <law@redhat.com>
4897 PR rtl-optimization/60131
4898 * ree.c (get_extended_src_reg): New function.
4899 (combine_reaching_defs): Use it rather than assuming location of REG.
4900 (find_and_remove_re): Verify first operand of extension is
4901 a REG before adding the insns to the copy list.
4903 2014-02-14 Roland McGrath <mcgrathr@google.com>
4905 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
4906 * configure: Regenerated.
4907 * config.in: Regenerated.
4908 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
4909 instead of ASM_SHORT.
4911 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
4912 Richard Earnshaw <rearnsha@arm.com>
4914 PR rtl-optimization/59535
4915 * lra-constraints.c (process_alt_operands): Encourage alternative
4916 when unassigned pseudo class is superset of the alternative class.
4917 (inherit_reload_reg): Don't inherit when optimizing for code size.
4918 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
4919 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
4920 modes not less than 4 for Thumb1.
4922 2014-02-14 Kyle McMartin <kyle@redhat.com>
4925 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
4927 2014-02-14 Richard Biener <rguenther@suse.de>
4929 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
4930 (get_frame_arg): Drop the assert with langhook types_compatible_p.
4931 Do not strip INDIRECT_REFs.
4933 2014-02-14 Richard Biener <rguenther@suse.de>
4936 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
4937 DECL_FUNCTION_SPECIFIC_TARGET.
4938 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
4939 * tree-streamer-out.c (pack_ts_target_option): Remove.
4940 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
4941 (write_ts_function_decl_tree_pointers): Do not stream
4942 DECL_FUNCTION_SPECIFIC_TARGET.
4943 * tree-streamer-in.c (unpack_ts_target_option): Remove.
4944 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
4945 (lto_input_ts_function_decl_tree_pointers): Do not stream
4946 DECL_FUNCTION_SPECIFIC_TARGET.
4948 2014-02-14 Jakub Jelinek <jakub@redhat.com>
4950 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
4951 (get_initial_def_for_induction, vectorizable_induction): Ignore
4952 debug stmts when looking for exit_phi.
4953 (vectorizable_live_operation): Fix up condition.
4955 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
4957 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
4958 nreverse() because it changes the content of original tree list.
4960 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
4962 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
4963 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
4965 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
4967 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
4968 GNU coding standards.
4970 2014-02-13 Jakub Jelinek <jakub@redhat.com>
4973 * dwarf2out.c (gen_subprogram_die): Don't call
4974 add_calling_convention_attribute if subr_die is old_die.
4976 2014-02-13 Sharad Singhai <singhai@google.com>
4978 * doc/optinfo.texi: Fix order of nodes.
4980 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
4982 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
4983 operands[2], not operands[3].
4985 2014-02-13 Richard Biener <rguenther@suse.de>
4988 * doc/install.texi (ISL): Update recommended version to 0.12.2,
4989 mention the possibility of an in-tree build.
4990 (CLooG): Update recommended version to 0.18.1, mention the
4991 possibility of an in-tree build and clarify that the ISL
4992 bundled with CLooG does not work.
4994 2014-02-13 Jakub Jelinek <jakub@redhat.com>
4997 * expr.c (compress_float_constant): If x is a hard register,
4998 extend into a pseudo and then move to x.
5000 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
5002 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
5003 caused by bad second argument to warning_at() with -mhotpatch and
5004 nested functions (e.g. with gfortran).
5006 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
5008 * opts.c (option_name): Remove "enabled by default" rider.
5010 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
5012 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
5014 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
5015 Uros Bizjak <ubizjak@gmail.com>
5018 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
5019 * configure: Regenerated.
5021 2014-02-12 Richard Biener <rguenther@suse.de>
5023 * vec.c (vec_prefix::calculate_allocation): Move as
5024 inline variant to vec.h.
5025 (vec_prefix::calculate_allocation_1): New out-of-line version.
5026 * vec.h (vec_prefix::calculate_allocation_1): Declare.
5027 (vec_prefix::m_has_auto_buf): Rename to ...
5028 (vec_prefix::m_using_auto_storage): ... this.
5029 (vec_prefix::calculate_allocation): Inline the easy cases
5030 and dispatch to calculate_allocation_1 which doesn't need the
5032 (va_heap::reserve): Use gcc_checking_assert.
5033 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
5034 m_using_auto_storage.
5035 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
5037 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
5038 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
5039 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
5041 2014-02-12 Richard Biener <rguenther@suse.de>
5043 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
5044 when we found a dependence.
5046 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
5048 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
5050 (maybe_fold_stmt): ... into this new function.
5051 * omp-low.c (lower_omp): Update comment.
5053 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
5056 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
5059 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
5061 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
5062 identifiers in comments.
5063 (cortexa53_extra_costs): Likewise.
5064 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
5065 (cortexa7_extra_costs): Likewise.
5066 (cortexa12_extra_costs): Likewise.
5067 (cortexa15_extra_costs): Likewise.
5068 (v7m_extra_costs): Likewise.
5070 2014-02-12 Richard Biener <rguenther@suse.de>
5073 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
5074 of posix_memalign being successful.
5075 (lower_stmt): Restrict lowering of posix_memalign to when
5076 -ftree-bit-ccp is enabled.
5078 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
5080 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
5082 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
5084 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
5086 PR rtl-optimization/60116
5087 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
5088 other_insn once the combination has been validated.
5090 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
5093 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
5095 * ipa-devirt.c: Include demangle.h
5096 (odr_violation_reported): New static variable.
5097 (add_type_duplicate): Update odr_violations.
5098 (maybe_record_node): Add completep parameter; update it.
5099 (record_target_from_binfo): Add COMPLETEP parameter;
5100 update it as needed.
5101 (possible_polymorphic_call_targets_1): Likewise.
5102 (struct polymorphic_call_target_d): Add nonconstruction_targets;
5103 rename FINAL to COMPLETE.
5104 (record_targets_from_bases): Sanity check we found the binfo;
5105 fix COMPLETEP updating.
5106 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
5107 parameter, fix computing of COMPLETEP.
5108 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
5109 at LTO time do demangling.
5110 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
5111 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
5113 (gimple_get_virt_method_for_binfo): Likewise.
5114 * gimple-fold.h (gimple_get_virt_method_for_binfo,
5115 gimple_get_virt_method_for_vtable): Update prototypes.
5117 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
5120 * genautomata.c (add_presence_absence): Fix typo with
5121 {final_}presence_list.
5123 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
5126 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
5127 for VSX/Altivec vectors that land in GPR registers.
5129 2014-02-11 Richard Henderson <rth@redhat.com>
5130 Jakub Jelinek <jakub@redhat.com>
5133 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
5134 around drhs if type conversion to lacc->type is not useless.
5136 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5138 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
5140 (cortex-a57.cortex-a53): Likewise.
5141 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
5143 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5145 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
5148 2014-02-11 Renlin Li <Renlin.Li@arm.com>
5150 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
5151 add_options_for_arm_vfp3.
5153 2014-02-11 Jeff Law <law@redhat.com>
5156 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
5157 object with an undesirable mode.
5159 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5162 * config/i386/sol2-9.h: New file.
5163 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
5164 *-*-solaris2.9*): Use it.
5166 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5168 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
5169 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
5171 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5173 * config/microblaze/microblaze.c: Extend mcpu version format
5175 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
5177 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
5179 2014-02-10 Richard Henderson <rth@redhat.com>
5182 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
5183 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
5184 ms-abi vs -mno-accumulate-outgoing-args.
5185 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
5186 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
5189 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
5192 * cfgexpand.c (expand_asm_operands): Attach source location to
5193 ASM_INPUT rtx objects.
5194 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
5196 2014-02-10 Nick Clifton <nickc@redhat.com>
5198 * config/mn10300/mn10300.c (popcount): New function.
5199 (mn10300_expand_prologue): Include saved registers in stack usage
5202 2014-02-10 Jeff Law <law@redhat.com>
5205 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
5206 when changing the SET_DEST of a prior insn to avoid an input reload.
5208 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5210 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
5211 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
5212 -mcall-openbsd, or -mcall-linux.
5213 (CC1_ENDIAN_BIG_SPEC): Remove.
5214 (CC1_ENDIAN_LITTLE_SPEC): Remove.
5215 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5216 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
5217 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
5218 and %cc1_endian_default.
5219 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5221 2014-02-10 Richard Biener <rguenther@suse.de>
5223 PR tree-optimization/60115
5224 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
5225 MEM_REF handling. Properly verify that the accesses are not
5226 out of the objects bound.
5228 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5230 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
5233 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
5235 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
5236 proper constants and fix formatting.
5237 (possible_polymorphic_call_targets): Fix formatting.
5239 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
5240 Ilya Tocar <ilya.tocar@intel.com>
5242 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
5243 (_mm512_loadu_epi32): Renamed into...
5244 (_mm512_loadu_si512): This.
5245 (_mm512_storeu_epi32): Renamed into...
5246 (_mm512_storeu_si512): This.
5247 (_mm512_maskz_ceil_ps): Removed.
5248 (_mm512_maskz_ceil_pd): Ditto.
5249 (_mm512_maskz_floor_ps): Ditto.
5250 (_mm512_maskz_floor_pd): Ditto.
5251 (_mm512_floor_round_ps): Ditto.
5252 (_mm512_floor_round_pd): Ditto.
5253 (_mm512_ceil_round_ps): Ditto.
5254 (_mm512_ceil_round_pd): Ditto.
5255 (_mm512_mask_floor_round_ps): Ditto.
5256 (_mm512_mask_floor_round_pd): Ditto.
5257 (_mm512_mask_ceil_round_ps): Ditto.
5258 (_mm512_mask_ceil_round_pd): Ditto.
5259 (_mm512_maskz_floor_round_ps): Ditto.
5260 (_mm512_maskz_floor_round_pd): Ditto.
5261 (_mm512_maskz_ceil_round_ps): Ditto.
5262 (_mm512_maskz_ceil_round_pd): Ditto.
5263 (_mm512_expand_pd): Ditto.
5264 (_mm512_expand_ps): Ditto.
5265 * config/i386/i386.c (ix86_builtins): Remove
5266 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
5267 (bdesc_args): Ditto.
5268 * config/i386/predicates.md (const1256_operand): New.
5269 (const_1_to_2_operand): Ditto.
5270 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
5271 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
5272 (*avx512pf_gatherpf<mode>sf): Ditto.
5273 (avx512pf_gatherpf<mode>df): Ditto.
5274 (*avx512pf_gatherpf<mode>df_mask): Ditto.
5275 (*avx512pf_gatherpf<mode>df): Ditto.
5276 (avx512pf_scatterpf<mode>sf): Ditto.
5277 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5278 (*avx512pf_scatterpf<mode>sf): Ditto.
5279 (avx512pf_scatterpf<mode>df): Ditto.
5280 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5281 (*avx512pf_scatterpf<mode>df): Ditto.
5282 (avx512f_expand<mode>): Removed.
5283 (<shift_insn><mode>3<mask_name>): Change predicate type.
5285 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5287 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
5288 not at the end of datarefs vector use ordered_remove to avoid
5289 reordering datarefs vector.
5292 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
5293 mark local addressable non-static vars as GOVD_PRIVATE
5294 instead of GOVD_LOCAL.
5295 * omp-low.c (lower_omp_for): Move gimple_bind_vars
5296 and BLOCK_VARS of gimple_bind_block to new_stmt rather
5300 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
5301 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
5302 assume_aligned or alloc_align attributes.
5303 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
5304 arguments. Handle also assume_aligned and alloc_align attributes.
5305 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
5306 calls to functions with assume_aligned or alloc_align attributes.
5307 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
5309 2014-02-08 Terry Guo <terry.guo@arm.com>
5311 * doc/invoke.texi: Document ARM -march=armv7e-m.
5313 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5315 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
5316 flag on __cilkrts_rethrow builtin.
5319 * ipa-cp.c (determine_versionability): Fail at -O0
5320 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
5321 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
5324 2014-02-04 Jakub Jelinek <jakub@redhat.com>
5327 * tree-inline.c (copy_forbidden): Fail for
5328 __attribute__((optimize (0))) functions.
5330 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5332 * varpool.c: Include pointer-set.h.
5333 (varpool_remove_unreferenced_decls): Variables in other partitions
5334 will not be output; be however careful to not lose information
5337 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5339 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
5340 lookup in the vtable constructor.
5342 2014-02-07 Jeff Law <law@redhat.com>
5345 * config/m68k/m68k.md (ashldi_extsi): Turn into a
5346 define_insn_and_split.
5348 * ipa-inline.c (inline_small_functions): Fix typos.
5350 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5352 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
5353 (s390_can_use_return_insn): Declare.
5354 * config/s390/s390.h (EPILOGUE_USES): Define.
5355 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
5357 (s390_chunkify_start): Handle return JUMP_LABELs.
5358 (s390_early_mach): Emit a main_pool instruction on the entry edge.
5359 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
5360 (s390_can_use_return_insn): New functions.
5361 (s390_fix_long_loop_prediction): Handle conditional returns.
5362 (TARGET_SET_UP_BY_PROLOGUE): Define.
5363 * config/s390/s390.md (ANY_RETURN): New code iterator.
5364 (*creturn, *csimple_return, return, simple_return): New patterns.
5366 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5368 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
5369 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
5370 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
5371 REG_CFA_RESTORE list when deciding not to restore a register.
5373 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5375 * config/s390/s390.c: Include tree-pass.h and context.h.
5376 (s390_early_mach): New function, split out from...
5377 (s390_emit_prologue): ...here.
5378 (pass_data_s390_early_mach): New pass structure.
5379 (pass_s390_early_mach): New class.
5380 (s390_option_override): Create and register early_mach pass.
5381 Move to end of file.
5383 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5385 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
5386 to match for the exit block.
5388 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5390 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
5391 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
5392 Reject misaligned operands.
5394 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5396 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
5398 2014-02-07 Richard Biener <rguenther@suse.de>
5401 * gimple-low.c (lower_builtin_posix_memalign): New function.
5402 (lower_stmt): Call it to lower posix_memalign in a way
5403 to make alignment info accessible.
5405 2014-02-07 Jakub Jelinek <jakub@redhat.com>
5408 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
5409 __builtin_setjmp_receiver.
5411 2014-02-07 Richard Biener <rguenther@suse.de>
5414 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
5415 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
5416 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
5417 Handle BUILT_IN_POSIX_MEMALIGN.
5418 (find_func_clobbers): Likewise.
5419 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
5420 (call_may_clobber_ref_p_1): Likewise.
5422 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5425 * ipa-devirt.c (record_target_from_binfo): Remove overactive
5428 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5431 * lto-cgraph.c (lto_output_node): Use
5432 symtab_get_symbol_partitioning_class.
5433 (lto_output_varpool_node): likewise.
5434 (symtab_get_symbol_partitioning_class): Move here from
5436 * cgraph.h (symbol_partitioning_class): Likewise.
5437 (symtab_get_symbol_partitioning_class): Declare.
5439 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5441 * ggc.h (ggc_internal_cleared_alloc): New macro.
5442 * vec.h (vec_safe_copy): Handle memory stats.
5443 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
5444 * target-globals.c (save_target_globals): Likewise.
5446 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5449 * expr.c (emit_move_resolve_push): Export; be bit more selective
5450 on when to clear alias set.
5451 * expr.h (emit_move_resolve_push): Declare.
5452 * function.h (struct function): Add tail_call_marked.
5453 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
5454 * config/i386/i386-protos.h (ix86_expand_push): Remove.
5455 * config/i386/i386.md (TImode move expander): De not call
5457 (FP push expanders): Preserve memory attributes.
5458 * config/i386/sse.md (push<mode>1): Remove.
5459 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
5460 (ix86_expand_push): Remove.
5461 * config/i386/mmx.md (push<mode>1): Remove.
5463 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5465 PR rtl-optimization/60030
5466 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
5467 lopart with paradoxical subreg before shifting it up by hprec.
5469 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5471 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
5472 Remove extra newline at end of file.
5473 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
5474 (arm_issue_rate): Handle cortexa57.
5475 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
5476 (cortex-a57.cortex-a53): Likewise.
5478 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5481 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
5482 don't record in REG_FRAME_RELATED_EXPR registers not set in that
5484 (arm_expand_prologue): Adjust all callers.
5485 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
5486 info, registers also at the lowest numbered registers side. Use
5487 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
5491 * var-tracking.c (adjust_mems): Before adding a SET to
5492 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
5494 2014-02-06 Alan Modra <amodra@gmail.com>
5497 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
5498 change SDmode to DDmode when lra_in_progress.
5500 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5503 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
5504 free_data_ref on the dr first, and before goto again also set dr
5505 to the next dr. For simd_lane_access, free old datarefs[i] before
5506 overwriting it. For get_vectype_for_scalar_type failure, don't
5507 free_data_ref if simd_lane_access.
5509 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
5512 * tree.h (opts_for_fn): New inline function.
5513 (opt_for_fn): Define.
5514 * config/i386/i386.c (ix86_function_regparm): Use
5515 opt_for_fn (decl, optimize) instead of optimize.
5517 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
5519 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
5520 for SYMBOL_REF in large memory model.
5522 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5524 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
5526 (cortex-a57): Likewise.
5527 (cortex-a57.cortex-a53): Likewise.
5529 2014-02-06 Yury Gribov <y.gribov@samsung.com>
5530 Kugan Vivekanandarajah <kuganv@linaro.org>
5532 * config/arm/arm.c (arm_vector_alignment_reachable): Check
5534 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
5536 2014-02-06 Richard Biener <rguenther@suse.de>
5538 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
5539 set_loop_copy and initialize_original_copy_tables.
5541 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
5543 * config/aarch64/aarch64-simd.md
5544 (aarch64_ashr_simddi): Change QI to SI.
5546 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
5547 Jakub Jelinek <jakub@redhat.com>
5550 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
5553 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5555 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
5556 CODE_FOR_altivec_vpku[hw]um to
5557 CODE_FOR_altivec_vpku[hw]um_direct.
5558 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
5559 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
5560 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
5561 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
5563 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5565 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
5566 generation for -maltivec=be.
5567 (altivec_vsumsws): Simplify redundant test.
5569 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5571 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
5572 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
5573 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
5574 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
5575 gen_altivec_vpkuwum.
5576 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
5578 (altivec_vpks<VI_char>ss): Likewise.
5579 (altivec_vpks<VI_char>us): Likewise.
5580 (altivec_vpku<VI_char>us): Likewise.
5581 (altivec_vpku<VI_char>um): Likewise.
5582 (altivec_vpku<VI_char>um_direct): New (copy of
5583 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
5585 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
5586 target is little endian and -maltivec=be is not specified.
5587 (*altivec_vupkhs<VU_char>_direct): New (copy of
5588 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
5589 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
5590 target is little endian and -maltivec=be is not specified.
5591 (*altivec_vupkls<VU_char>_direct): New (copy of
5592 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
5593 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
5594 little endian and -maltivec=be is not specified.
5595 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
5596 little endian and -maltivec=be is not specified.
5598 2014-02-05 Richard Henderson <rth@redhat.com>
5601 * combine-stack-adj.c: Revert r206943.
5602 * sched-int.h (struct deps_desc): Add last_args_size.
5603 * sched-deps.c (init_deps): Initialize it.
5604 (sched_analyze_insn): Add OUTPUT dependencies between insns that
5605 contain REG_ARGS_SIZE notes.
5607 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
5609 * lto-cgraph.c (asm_nodes_output): Make global.
5610 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
5611 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
5612 (driver_handle_option): Handle OPT_fwpa.
5614 2014-02-05 Jakub Jelinek <jakub@redhat.com>
5617 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
5618 a comment typo and formatting issue. If odr_hash hasn't been
5619 created, return vNULL and set *completep to false.
5622 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
5623 bb with no successors.
5625 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
5628 * doc/invoke.texi (-march): Clarify documentation for ARM.
5632 2014-02-05 Richard Biener <rguenther@suse.de>
5634 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
5635 when not vectorizing because of too many alias checks.
5636 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
5637 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
5639 2014-02-05 Nick Clifton <nickc@redhat.com>
5641 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
5642 accept extended registers in any mode when compiling for the MN10300.
5644 2014-02-05 Yury Gribov <y.gribov@samsung.com>
5646 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
5647 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
5648 sanitization attributes.
5649 (can_inline_edge_p): Likewise.
5650 (sanitize_attrs_match_for_inline_p): New function.
5652 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5654 * ipa-prop.c (detect_type_change): Shor circuit testing of
5655 type changes on THIS pointer.
5657 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
5660 * config/pa/pa.c (legitimize_tls_address): Return original address
5661 if not passed a SYMBOL_REF rtx.
5662 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
5664 (pa_emit_move_sequence): Simplify TLS source operands.
5665 (pa_legitimate_constant_p): Reject all TLS constants.
5666 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
5667 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
5669 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5671 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
5672 groups when we know they are controlled by LTO.
5673 * varasm.c (default_binds_local_p_1): If object is in other partition,
5674 it will be resolved locally.
5676 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
5678 * config/host-linux.c (linux_gt_pch_use_address): Don't
5679 use SSIZE_MAX because it is not always defined.
5681 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
5684 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
5685 threshold for pseudo splitting.
5686 (update_ebb_live_info): Process call argument hard registers and
5687 hard registers from insn definition too.
5688 (max_small_class_regs_num): New constant.
5689 (inherit_in_ebb): Update live hard regs through EBBs. Update
5690 reloads_num only for small register classes. Don't split for
5693 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
5696 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
5699 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5701 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
5704 2014-02-04 Marek Polacek <polacek@redhat.com>
5706 * gdbinit.in (pel): Define.
5708 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
5710 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
5713 2014-02-04 Richard Biener <rguenther@suse.de>
5716 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
5717 in function context local.
5718 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
5719 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
5720 similar to LTO_imported_decl_ref.
5722 2014-02-04 Jakub Jelinek <jakub@redhat.com>
5724 PR tree-optimization/60002
5725 * cgraphclones.c (build_function_decl_skip_args): Clear
5728 PR tree-optimization/60023
5729 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
5730 false to gsi_replace.
5731 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
5732 has been in some EH region and vec_stmt could throw, add
5733 vec_stmt into the same EH region.
5734 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
5735 has no lhs, ignore it.
5736 * internal-fn.c (expand_MASK_LOAD): Likewise.
5739 * tree-inline.c (copy_forbidden): Fail for
5740 __attribute__((optimize (0))) functions.
5743 * omp-low.c (simd_clone_struct_copy): If from->inbranch
5744 is set, copy one less argument.
5745 (expand_simd_clones): Don't subtract clone_info->inbranch
5746 from simd_clone_struct_alloc argument.
5748 PR rtl-optimization/57915
5749 * recog.c (simplify_while_replacing): If all unary/binary/relational
5750 operation arguments are constant, attempt to simplify those.
5753 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
5754 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
5756 2014-02-04 Richard Biener <rguenther@suse.de>
5758 PR tree-optimization/60012
5759 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
5760 TBAA disambiguation to all DDRs.
5762 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5765 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
5766 (LINK_SPEC): Use it for -shared, -shared-libgcc.
5768 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5771 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
5773 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5775 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
5776 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
5778 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5781 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
5782 to figure out targets of polymorphic calls with known decl.
5783 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
5784 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
5785 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
5786 (get_polymorphic_call_info): ... here.
5787 (get_polymorphic_call_info_from_invariant): New function.
5789 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5791 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
5792 lookup via vtable pointer; check for type consistency
5793 and turn inconsitent facts into UNREACHABLE.
5794 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
5795 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
5796 type inconsistent querries; return UNREACHABLE instead.
5798 2014-02-03 Richard Henderson <rth@twiddle.net>
5801 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
5802 already processed this node.
5803 (normalize_one_pred_1): Pass along mark_set.
5804 (normalize_one_pred): Create and destroy a pointer_set_t.
5805 (normalize_one_pred_chain): Likewise.
5807 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
5809 PR gcov-profile/58602
5810 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
5812 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5815 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
5816 -fno-devirtualize; try to devirtualize by the knowledge of
5817 virtual table pointer given by aggregate propagation.
5818 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
5819 (ipa_print_node_jump_functions): Dump also offset that
5820 is relevant for polymorphic calls.
5821 (determine_known_aggregate_parts): Add arg_type parameter; use it
5822 instead of determining the type from pointer type.
5823 (ipa_compute_jump_functions_for_edge): Update call of
5824 determine_known_aggregate_parts.
5825 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
5826 (gimple_get_virt_method_for_binfo): ... here; simplify using
5827 vtable_pointer_value_to_vtable.
5828 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
5829 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
5830 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
5831 (vtable_pointer_value_to_vtable): Break out from ...; handle also
5833 (vtable_pointer_value_to_binfo): ... here.
5834 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
5836 2014-02-03 Teresa Johnson <tejohnson@google.com>
5838 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
5839 redef of outer loop index variable.
5841 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
5845 * doc/extend.texi (Function Attributes): Typo.
5847 2014-02-03 Cong Hou <congh@google.com>
5849 PR tree-optimization/60000
5850 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
5851 if the vectorized statement is a store. A store statement can only
5852 appear at the end of pattern statements.
5854 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
5856 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
5857 (ix86_option_override_internal): Default long double to 64-bit for
5858 32-bit Bionic and to 128-bit for 64-bit Bionic.
5860 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
5861 TARGET_LONG_DOUBLE_128 is true.
5862 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
5864 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
5865 (mlong-double-64): Negate -mlong-double-128.
5866 (mlong-double-128): New option.
5868 * config/i386/i386-c.c (ix86_target_macros): Define
5869 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
5871 * doc/invoke.texi: Document -mlong-double-128.
5873 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
5875 PR rtl-optimization/60024
5876 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
5878 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
5880 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
5882 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
5884 PR rtl-optimization/57662
5885 * sel-sched.c (code_motion_path_driver): Do not mark already not
5886 existing blocks in the visiting bitmap.
5888 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
5890 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
5891 on the insn being emitted.
5893 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
5894 Will Deacon <will.deacon@arm.com>
5896 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
5898 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5900 * config/arm/arm-tables.opt: Regenerate.
5902 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5904 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
5905 for vector types other than V16QImode.
5906 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
5907 define_expand, and call altivec_expand_vec_perm_le when producing
5908 code with little endian element order.
5909 (*altivec_vperm_<mode>_internal): New insn having previous
5910 behavior of altivec_vperm_<mode>.
5911 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
5912 altivec_expand_vec_perm_le when producing code with little endian
5914 (*altivec_vperm_<mode>_uns_internal): New insn having previous
5915 behavior of altivec_vperm_<mode>_uns.
5917 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5919 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
5920 (altivec_vsumsws): Add handling for -maltivec=be with a little
5922 (altivec_vsumsws_direct): New.
5923 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
5924 gen_altivec_vsumsws.
5926 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
5928 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
5929 vtable_pointer_value_to_binfo): New functions.
5930 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
5931 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
5933 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
5935 * config/nios2/nios2.md (load_got_register): Initialize GOT
5936 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
5937 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
5939 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
5941 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
5942 preserverd by passthrough, do not propagate the type.
5944 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
5946 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
5947 (mips_atomic_assign_expand_fenv): New function.
5948 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
5950 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
5952 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
5953 (__builtin_mips_set_fcsr): Likewise.
5954 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
5955 MIPS_USI_FTYPE_VOID.
5956 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
5957 (mips16_expand_set_fcsr): Likewise.
5958 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
5959 (mips16_set_fcsr_stub): Likewise.
5960 (mips16_get_fcsr_one_only_stub): New class.
5961 (mips16_set_fcsr_one_only_stub): Likewise.
5962 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
5963 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
5964 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
5965 (hard_float): New availability predicate.
5966 (mips_builtins): Add get_fcsr and set_fcsr.
5967 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
5968 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
5969 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
5970 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
5971 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
5974 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
5976 * config/mips/mips.c (mips_one_only_stub): New class.
5977 (mips_need_mips16_rdhwr_p): Replace with...
5978 (mips16_rdhwr_stub): ...this new variable.
5979 (mips16_stub_call_address): New function.
5980 (mips16_rdhwr_one_only_stub): New class.
5981 (mips_expand_thread_pointer): Use mips16_stub_call_address.
5982 (mips_output_mips16_rdhwr): Delete.
5983 (mips_finish_stub): New function.
5984 (mips_code_end): Use it to handle rdhwr stubs.
5986 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
5989 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
5990 when calculating size of integer atomic types.
5992 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
5994 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
5996 2014-02-01 Jakub Jelinek <jakub@redhat.com>
5998 PR tree-optimization/60003
5999 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
6000 * profile.c (branch_prob): Use gimple_call_builtin_p
6001 to check for BUILT_IN_SETJMP_RECEIVER.
6002 * tree-inline.c (copy_bb): Call notice_special_calls.
6004 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
6007 * lra-constraints.c (process_alt_operands): Update reload_sum only
6010 2014-01-31 Richard Henderson <rth@redhat.com>
6013 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
6014 until after else_eh is processed.
6016 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
6018 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
6019 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
6020 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
6021 in smmintrin.h, remove them.
6022 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
6023 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
6024 * config/i386/i386.md (ROUND_SAE): Fix value.
6025 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
6026 (const48_operand): New.
6027 * config/i386/subst.md (round), (round_expand): Use
6028 const_4_or_8_to_11_operand.
6029 (round_saeonly), (round_saeonly_expand): Use const48_operand.
6031 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
6033 * config/i386/constraints.md (Yk): Swap meaning with k.
6034 * config/i386/i386.md (movhi_internal): Change Yk to k.
6035 (movqi_internal): Ditto.
6036 (*k<logic><mode>): Ditto.
6039 (kandn<mode>): Ditto.
6040 (*<code>hi_1): Ditto.
6041 (*<code>qi_1): Ditto.
6042 (kxnor<mode>): Ditto.
6043 (kortestzhi): Ditto.
6044 (kortestchi): Ditto.
6046 (*one_cmplhi2_1): Ditto.
6047 (*one_cmplqi2_1): Ditto.
6048 * config/i386/sse.md (): Change k to Yk.
6049 (avx512f_load<mode>_mask): Ditto.
6050 (avx512f_blendm<mode>): Ditto.
6051 (avx512f_store<mode>_mask): Ditto.
6052 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
6053 (avx512f_storedqu<mode>_mask): Ditto.
6054 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
6056 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
6057 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
6058 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
6059 (avx512f_maskcmp<mode>3): Ditto.
6060 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
6061 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
6062 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
6063 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
6064 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
6065 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
6066 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
6067 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
6068 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
6069 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
6070 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
6071 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
6072 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
6073 (vec_extract_lo_<mode>_maskm): Ditto.
6074 (vec_extract_hi_<mode>_maskm): Ditto.
6075 (avx512f_vternlog<mode>_mask): Ditto.
6076 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
6077 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
6078 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
6079 (avx512f_<code>v8div16qi2_mask): Ditto.
6080 (avx512f_<code>v8div16qi2_mask_store): Ditto.
6081 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
6082 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
6083 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
6084 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
6085 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6086 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6087 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6088 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6089 (avx512cd_maskb_vec_dupv8di): Ditto.
6090 (avx512cd_maskw_vec_dupv16si): Ditto.
6091 (avx512f_vpermi2var<mode>3_maskz): Ditto.
6092 (avx512f_vpermi2var<mode>3_mask): Ditto.
6093 (avx512f_vpermi2var<mode>3_mask): Ditto.
6094 (avx512f_vpermt2var<mode>3_maskz): Ditto.
6095 (*avx512f_gathersi<mode>): Ditto.
6096 (*avx512f_gathersi<mode>_2): Ditto.
6097 (*avx512f_gatherdi<mode>): Ditto.
6098 (*avx512f_gatherdi<mode>_2): Ditto.
6099 (*avx512f_scattersi<mode>): Ditto.
6100 (*avx512f_scatterdi<mode>): Ditto.
6101 (avx512f_compress<mode>_mask): Ditto.
6102 (avx512f_compressstore<mode>_mask): Ditto.
6103 (avx512f_expand<mode>_mask): Ditto.
6104 * config/i386/subst.md (mask): Change k to Yk.
6105 (mask_scalar_merge): Ditto.
6108 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
6110 * doc/extend.texi (Vector Extensions): Document ?: in C++.
6112 2014-01-31 Richard Biener <rguenther@suse.de>
6115 * builtins.c (fold_builtin_memory_op): Make sure to not
6116 use a floating-point mode or a boolean or enumeral type for
6119 2014-01-30 DJ Delorie <dj@redhat.com>
6121 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
6122 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
6123 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
6124 whenever main() has an epilogue.
6126 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6128 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
6129 unused variable "field".
6130 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
6131 (vsx_mergeh_<mode>): Likewise.
6132 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
6133 (altivec_vmrghh): Likewise.
6134 (altivec_vmrghw): Likewise.
6135 (altivec_vmrglb): Likewise.
6136 (altivec_vmrglh): Likewise.
6137 (altivec_vmrglw): Likewise.
6138 (altivec_vspltb): Add missing uses.
6139 (altivec_vsplth): Likewise.
6140 (altivec_vspltw): Likewise.
6141 (altivec_vspltsf): Likewise.
6143 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6146 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
6147 frame related instructions.
6149 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
6151 PR rtl-optimization/59959
6152 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
6153 any reload of register whose subreg is invalid.
6155 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6157 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
6158 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
6159 Add missing return type - void.
6161 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6163 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
6164 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
6165 remove element index adjustment for endian (now handled in vsx.md
6167 (altivec_expand_vec_perm_const): Use
6168 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
6169 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
6170 (vsx_xxspltw_<mode>): Adjust element index for little endian.
6171 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
6172 define_expand and a new define_insn *altivec_vspltb_internal;
6173 adjust for -maltivec=be on a little endian target.
6174 (altivec_vspltb_direct): New.
6175 (altivec_vsplth): Divide into a define_expand and a new
6176 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
6177 little endian target.
6178 (altivec_vsplth_direct): New.
6179 (altivec_vspltw): Divide into a define_expand and a new
6180 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
6181 little endian target.
6182 (altivec_vspltw_direct): New.
6183 (altivec_vspltsf): Divide into a define_expand and a new
6184 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
6185 a little endian target.
6187 2014-01-30 Richard Biener <rguenther@suse.de>
6189 PR tree-optimization/59993
6190 * tree-ssa-forwprop.c (associate_pointerplus): Check we
6191 can propagate form the earlier stmt and avoid the transform
6192 when the intermediate result is needed.
6194 2014-01-30 Alangi Derick <alangiderick@gmail.com>
6196 * README.Portability: Fix typo.
6198 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
6200 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
6201 comparison_operator with ordered_comparison_operator.
6203 2014-01-30 Nick Clifton <nickc@redhat.com>
6205 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
6206 Rename to mn10300_store_multiple_regs.
6207 * config/mn10300/mn10300.c: Likewise.
6208 * config/mn10300/mn10300.md (store_movm): Fix typo: call
6209 store_multiple_regs.
6210 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
6211 Call mn10300_store_multiple_regs.
6213 2014-01-30 Nick Clifton <nickc@redhat.com>
6214 DJ Delorie <dj@redhat.com>
6216 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
6217 %fp 2 to keep registers after it properly word-aligned.
6218 (rl78_alloc_physical_registers_umul): Handle the case where both
6219 input operands are the same.
6221 2014-01-30 Richard Biener <rguenther@suse.de>
6223 PR tree-optimization/59903
6224 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
6227 2014-01-30 Jason Merrill <jason@redhat.com>
6230 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
6233 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
6235 2014-01-30 Richard Biener <rguenther@suse.de>
6237 PR tree-optimization/59951
6238 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
6240 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
6243 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
6244 SFmode to DFmode case.
6246 2014-01-29 DJ Delorie <dj@redhat.com>
6248 * config/msp430/msp430.opt (-minrt): New.
6249 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
6251 (ENDFILE_SPEC): Likewise.
6253 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
6255 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
6256 (estimate_function_body_sizes): Use it.
6258 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
6261 * dwarf2out.c (is_cxx_auto): New.
6262 (is_base_type): Use it.
6263 (gen_type_die_with_usage): Likewise.
6265 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6267 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
6268 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
6269 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
6270 -maltivec=be with LE targets.
6271 (vsx_mergeh_<mode>): Likewise.
6272 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
6273 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
6274 (altivec_vmrghb): Replace with define_expand and new
6275 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
6276 (altivec_vmrghb_direct): New define_insn.
6277 (altivec_vmrghh): Replace with define_expand and new
6278 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
6279 (altivec_vmrghh_direct): New define_insn.
6280 (altivec_vmrghw): Replace with define_expand and new
6281 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
6282 (altivec_vmrghw_direct): New define_insn.
6283 (*altivec_vmrghsf): Adjust for endianness.
6284 (altivec_vmrglb): Replace with define_expand and new
6285 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
6286 (altivec_vmrglb_direct): New define_insn.
6287 (altivec_vmrglh): Replace with define_expand and new
6288 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
6289 (altivec_vmrglh_direct): New define_insn.
6290 (altivec_vmrglw): Replace with define_expand and new
6291 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
6292 (altivec_vmrglw_direct): New define_insn.
6293 (*altivec_vmrglsf): Adjust for endianness.
6294 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6295 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6296 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6297 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6298 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6299 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6300 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6301 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6303 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
6305 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
6306 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
6309 2014-01-29 Richard Biener <rguenther@suse.de>
6311 PR tree-optimization/58742
6312 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
6313 associate_pointerplus_align.
6314 (associate_pointerplus_diff): New function.
6315 (associate_pointerplus): Likewise. Call associate_pointerplus_align
6316 and associate_pointerplus_diff.
6318 2014-01-29 Richard Biener <rguenther@suse.de>
6320 * lto-streamer.h (LTO_major_version): Bump to 3.
6321 (LTO_minor_version): Reset to 0.
6323 2014-01-29 Renlin Li <Renlin.Li@arm.com>
6325 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
6326 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
6327 (arm_file_start): Generate correct asm header for armv7ve.
6328 * config/arm/bpabi.h: Add multilib support for armv7ve.
6329 * config/arm/driver-arm.c: Change the architectures of cortex-a7
6330 and cortex-a15 to armv7ve.
6331 * config/arm/t-aprofile: Add multilib support for armv7ve.
6332 * doc/invoke.texi: Document -march=armv7ve.
6334 2014-01-29 Richard Biener <rguenther@suse.de>
6336 PR tree-optimization/58742
6337 * tree-ssa-forwprop.c (associate_plusminus): Return true
6338 if we changed sth, defer EH cleanup to ...
6339 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
6340 (simplify_mult): New function.
6342 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6345 PR tree-optimization/59920
6346 * tree.c (build_common_builtin_nodes): Remove
6347 __builtin_setjmp_dispatcher initialization.
6348 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
6349 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
6350 instead of gsi_after_labels + manually skipping debug stmts.
6351 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
6352 ignore bbs with IFN_ABNORMAL_DISPATCHER.
6353 * tree-inline.c (copy_edges_for_bb): Remove
6354 can_make_abnormal_goto argument, instead add abnormal_goto_dest
6355 argument. Ignore computed_goto_p stmts. Don't call
6356 make_abnormal_goto_edges. If a call might need abnormal edges
6357 for non-local gotos, see if it already has an edge to
6358 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
6359 with true argument, don't do anything then, otherwise add
6360 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
6361 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
6363 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
6364 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
6365 (lower_stmt): Don't set data->calls_builtin_setjmp.
6366 (lower_builtin_setjmp): Adjust comment.
6367 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
6368 * tree-cfg.c (found_computed_goto): Remove.
6369 (factor_computed_gotos): Remove.
6370 (make_goto_expr_edges): Return bool, true for computed gotos.
6371 Don't call make_abnormal_goto_edges.
6372 (build_gimple_cfg): Don't set found_computed_goto, don't call
6373 factor_computed_gotos.
6374 (computed_goto_p): No longer static.
6375 (make_blocks): Don't set found_computed_goto.
6376 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
6377 (make_edges): If make_goto_expr_edges returns true, push bb
6378 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
6379 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
6380 vector. Record mapping between bbs and OpenMP regions if there
6381 are any, adjust make_gimple_omp_edges caller. Call
6382 handle_abnormal_edges.
6383 (make_abnormal_goto_edges): Remove.
6384 * tree-cfg.h (make_abnormal_goto_edges): Remove.
6385 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
6386 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
6387 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
6388 * internal-fn.def (ABNORMAL_DISPATCHER): New.
6389 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
6390 filling *region also set *region_idx to (*region)->entry->index.
6393 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
6394 For REGs set ORIGINAL_REGNO.
6396 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
6398 * doc/md.texi: Mention that a target shouldn't implement
6399 vec_widen_(s|u)mul_even/odd pair if it is less efficient
6402 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6404 PR tree-optimization/59594
6405 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
6406 a copy of the datarefs vector rather than the vector itself.
6408 2014-01-28 Jason Merrill <jason@redhat.com>
6411 * dwarf2out.c (auto_die): New static.
6412 (gen_type_die_with_usage): Handle C++1y 'auto'.
6413 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
6416 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
6419 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
6420 (SPEC_X32): Likewise.
6421 (SPEC_64): Likewise.
6422 * config/i386/i386.c (ix86_option_override_internal): Turn off
6423 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
6425 (x86_file_start): Output .code16gcc for TARGET_16BIT.
6426 * config/i386/i386.h (TARGET_16BIT): New macro.
6427 (TARGET_16BIT_P): Likewise.
6428 * config/i386/i386.opt: Add m16.
6429 * doc/invoke.texi: Document -m16.
6431 2014-01-28 Jakub Jelinek <jakub@redhat.com>
6433 PR preprocessor/59935
6434 * input.c (location_get_source_line): Bail out on when line number
6435 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
6437 2014-01-28 Richard Biener <rguenther@suse.de>
6439 PR tree-optimization/58742
6440 * tree-ssa-forwprop.c (associate_plusminus): Handle
6441 pointer subtraction of the form (T)(P + A) - (T)P.
6443 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6445 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
6448 2014-01-28 Richard Biener <rguenther@suse.de>
6451 2014-01-28 Richard Biener <rguenther@suse.de>
6453 PR rtl-optimization/45364
6454 PR rtl-optimization/59890
6455 * var-tracking.c (local_get_addr_clear_given_value): Handle
6456 already cleared slot.
6457 (val_reset): Handle not allocated local_get_addr_cache.
6458 (vt_find_locations): Use post-order on the inverted CFG.
6460 2014-01-28 Richard Biener <rguenther@suse.de>
6462 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
6464 2014-01-28 Richard Biener <rguenther@suse.de>
6466 PR rtl-optimization/45364
6467 PR rtl-optimization/59890
6468 * var-tracking.c (local_get_addr_clear_given_value): Handle
6469 already cleared slot.
6470 (val_reset): Handle not allocated local_get_addr_cache.
6471 (vt_find_locations): Use post-order on the inverted CFG.
6473 2014-01-28 Alan Modra <amodra@gmail.com>
6475 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
6476 * configure.ac <recursive call for build != host>: Define
6477 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
6478 and LD_FOR_BUILD too.
6479 * configure: Regenerate.
6481 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
6483 * config/i386/i386.c (get_builtin_code_for_version): Separate
6484 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
6485 Broadwell from Haswell.
6487 2014-01-27 Steve Ellcey <sellcey@mips.com>
6489 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
6490 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
6491 * config/mips/mips.c (mips_option_override): Change setting
6493 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
6494 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
6495 Change from Mask to Var.
6497 2014-01-27 Jeff Law <law@redhat.com>
6499 * ipa-inline.c (inline_small_functions): Fix typo.
6501 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
6503 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
6504 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
6505 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
6506 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
6507 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
6508 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
6509 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
6510 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
6511 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
6512 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
6513 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
6514 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
6515 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
6516 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
6517 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
6518 (_mm512_storeu_epi64): Ditto.
6519 (_mm512_cmpge_epi32_mask): Ditto.
6520 (_mm512_cmpge_epu32_mask): Ditto.
6521 (_mm512_cmpge_epi64_mask): Ditto.
6522 (_mm512_cmpge_epu64_mask): Ditto.
6523 (_mm512_cmple_epi32_mask): Ditto.
6524 (_mm512_cmple_epu32_mask): Ditto.
6525 (_mm512_cmple_epi64_mask): Ditto.
6526 (_mm512_cmple_epu64_mask): Ditto.
6527 (_mm512_cmplt_epi32_mask): Ditto.
6528 (_mm512_cmplt_epu32_mask): Ditto.
6529 (_mm512_cmplt_epi64_mask): Ditto.
6530 (_mm512_cmplt_epu64_mask): Ditto.
6531 (_mm512_cmpneq_epi32_mask): Ditto.
6532 (_mm512_cmpneq_epu32_mask): Ditto.
6533 (_mm512_cmpneq_epi64_mask): Ditto.
6534 (_mm512_cmpneq_epu64_mask): Ditto.
6535 (_mm512_expand_pd): Ditto.
6536 (_mm512_expand_ps): Ditto.
6537 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
6538 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
6539 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
6540 * config/i386/i386.c (ix86_builtins): Add
6541 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
6542 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
6543 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
6544 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
6545 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
6546 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
6547 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
6548 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
6549 IX86_BUILTIN_PMOVUSQW512_MEM.
6550 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
6551 __builtin_ia32_pmovsqd512mem_mask,
6552 __builtin_ia32_pmovqd512mem_mask,
6553 __builtin_ia32_pmovusqw512mem_mask,
6554 __builtin_ia32_pmovsqw512mem_mask,
6555 __builtin_ia32_pmovqw512mem_mask,
6556 __builtin_ia32_pmovusdw512mem_mask,
6557 __builtin_ia32_pmovsdw512mem_mask,
6558 __builtin_ia32_pmovdw512mem_mask,
6559 __builtin_ia32_pmovqb512mem_mask,
6560 __builtin_ia32_pmovusqb512mem_mask,
6561 __builtin_ia32_pmovsqb512mem_mask,
6562 __builtin_ia32_pmovusdb512mem_mask,
6563 __builtin_ia32_pmovsdb512mem_mask,
6564 __builtin_ia32_pmovdb512mem_mask.
6565 (bdesc_args): Add __builtin_ia32_expanddf512,
6566 __builtin_ia32_expandsf512.
6567 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
6568 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
6569 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
6570 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
6571 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
6572 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
6573 (avx512f_<code>v8div16qi2_mask_store): This.
6574 (avx512f_expand<mode>): New.
6576 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
6578 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
6580 (_mm512_mask_prefetch_i64gather_pd): Ditto.
6581 (_mm512_prefetch_i32scatter_pd): Ditto.
6582 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
6583 (_mm512_prefetch_i64scatter_pd): Ditto.
6584 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
6585 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
6586 (_mm512_mask_prefetch_i64gather_ps): Ditto.
6587 (_mm512_prefetch_i32scatter_ps): Ditto.
6588 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
6589 (_mm512_prefetch_i64scatter_ps): Ditto.
6590 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
6591 * config/i386/i386-builtin-types.def: Define
6592 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
6593 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
6594 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
6595 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
6596 IX86_BUILTIN_SCATTERPFQPD.
6597 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
6598 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
6599 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
6600 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
6601 __builtin_ia32_scatterpfqps.
6602 (ix86_expand_builtin): Expand new built-ins.
6603 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
6604 fix memory access data type.
6605 (*avx512pf_gatherpf<mode>_mask): Ditto.
6606 (*avx512pf_gatherpf<mode>): Ditto.
6607 (avx512pf_scatterpf<mode>): Ditto.
6608 (*avx512pf_scatterpf<mode>_mask): Ditto.
6609 (*avx512pf_scatterpf<mode>): Ditto.
6610 (GATHER_SCATTER_SF_MEM_MODE): New.
6611 (avx512pf_gatherpf<mode>df): Ditto.
6612 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6613 (*avx512pf_scatterpf<mode>df): Ditto.
6615 2014-01-27 Jakub Jelinek <jakub@redhat.com>
6618 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
6619 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
6622 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
6624 * common/config/arm/arm-common.c
6625 (arm_rewrite_mcpu): Handle multiple names.
6627 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
6629 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
6631 * gimple-builder.h (create_gimple_tmp): Delete.
6633 2014-01-27 Christian Bruel <christian.bruel@st.com>
6635 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
6638 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
6640 * config/pa/pa.md (call): Generate indirect long calls to non-local
6641 functions when outputing 32-bit code.
6642 (call_value): Likewise except for special call to buggy powf function.
6644 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
6645 portable runtime and PIC indirect calls.
6646 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
6647 and PIC call sequences. Use ldo instead of blr to set return register
6648 in PIC call sequence.
6650 2014-01-25 Walter Lee <walt@tilera.com>
6652 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
6653 avoid clobbering a live register.
6655 2014-01-25 Walter Lee <walt@tilera.com>
6657 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
6658 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
6659 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
6660 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
6662 2014-01-25 Walter Lee <walt@tilera.com>
6664 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
6665 arguments on even registers.
6666 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
6668 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
6669 (BIGGEST_ALIGNMENT): Ditto.
6670 (BIGGEST_FIELD_ALIGNMENT): Ditto.
6672 2014-01-25 Walter Lee <walt@tilera.com>
6674 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
6675 insns before bundling.
6676 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
6678 2014-01-25 Walter Lee <walt@tilera.com>
6680 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
6681 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
6682 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
6684 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
6686 * config/mips/constraints.md (kl): Delete.
6687 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
6688 define expands, using...
6689 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
6690 instructions for MIPS16.
6691 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
6692 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
6694 2014-01-25 Walter Lee <walt@tilera.com>
6696 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
6700 2014-01-25 Walter Lee <walt@tilera.com>
6702 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
6703 (TARGET_EXPAND_TO_RTL_HOOK): Define.
6705 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
6707 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
6710 2014-01-25 Jakub Jelinek <jakub@redhat.com>
6712 * print-rtl.c (in_call_function_usage): New var.
6713 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
6714 EXPR_LIST mode as mode and not as reg note name.
6717 * cfgloopmanip.c (copy_loop_info): If
6718 loop->warned_aggressive_loop_optimizations, make sure
6719 the flag is set in target loop too.
6721 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
6723 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
6725 * builtins.def: Likewise.
6726 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
6727 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
6728 * ira.c (ira_setup_eliminable_regset): Likewise.
6729 * omp-low.c (gate_expand_omp): Likewise.
6730 (execute_lower_omp): Likewise.
6731 (diagnose_sb_0): Likewise.
6732 (gate_diagnose_omp_blocks): Likewise.
6733 (simd_clone_clauses_extract): Likewise.
6736 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6738 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
6739 correction for little endian...
6740 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
6743 2014-01-24 Jeff Law <law@redhat.com>
6745 PR tree-optimization/59919
6746 * tree-vrp.c (find_assert_locations_1): Do not register asserts
6747 for non-returning calls.
6749 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
6751 * common/config/aarch64/aarch64-common.c
6752 (aarch64_rewrite_mcpu): Handle multiple names.
6753 * config/aarch64/aarch64.h
6754 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
6756 2014-01-24 Dodji Seketeli <dodji@redhat.com>
6758 * input.c (add_file_to_cache_tab): Handle the case where fopen
6761 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
6764 * config/i386/i386.md (pushsf splitter): Get stack adjustment
6765 from push operand if code of push isn't PRE_DEC.
6767 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
6770 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
6771 -mquad-memory-atomic. Update -mquad-memory documentation to say
6772 it is only used for non-atomic loads/stores.
6774 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
6775 -mquad-memory or -mquad-memory-atomic switches.
6777 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
6778 -mquad-memory-atomic to ISA 2.07 support.
6780 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
6781 to separate support of normal quad word memory operations (ldq, stq)
6782 from the atomic quad word memory operations.
6784 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
6785 support to separate non-atomic quad word operations from atomic
6786 quad word operations. Disable non-atomic quad word operations in
6787 little endian mode so that we don't have to swap words after the
6788 load and before the store.
6789 (quad_load_store_p): Add comment about atomic quad word support.
6790 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
6791 options printed with -mdebug=reg.
6793 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
6794 -mquad-memory-atomic as the test for whether we have quad word
6795 atomic instructions.
6796 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
6797 or -mp8-vector are used, allow byte/half-word atomic operations.
6799 * config/rs6000/sync.md (load_lockedti): Insure that the address
6800 is a proper indexed or indirect address for the lqarx instruction.
6801 On little endian systems, swap the hi/lo registers after the lqarx
6803 (load_lockedpti): Use indexed_or_indirect_operand predicate to
6804 insure the address is valid for the lqarx instruction.
6805 (store_conditionalti): Insure that the address is a proper indexed
6806 or indirect address for the stqcrx. instruction. On little endian
6807 systems, swap the hi/lo registers before doing the stqcrx.
6809 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
6810 insure the address is valid for the stqcrx. instruction.
6812 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
6813 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
6814 type of quad memory support is available.
6816 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
6819 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
6820 there is a danger of looping.
6822 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
6824 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
6825 force flag_ira_loop_pressure if set via command line.
6827 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
6829 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
6830 (ashr_simd): New builtin handling DI mode.
6831 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
6832 (aarch64_sshr_simddi): New match pattern.
6833 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
6834 (vshrd_n_s64): Likewise.
6835 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
6837 2014-01-23 Nick Clifton <nickc@redhat.com>
6839 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
6840 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
6841 favour of mcu specific scripts.
6842 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
6845 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
6846 Alex Velenko <Alex.Velenko@arm.com>
6848 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
6849 (vaddv_s16): Likewise.
6850 (vaddv_s32): Likewise.
6851 (vaddv_u8): Likewise.
6852 (vaddv_u16): Likewise.
6853 (vaddv_u32): Likewise.
6854 (vaddvq_s8): Likewise.
6855 (vaddvq_s16): Likewise.
6856 (vaddvq_s32): Likewise.
6857 (vaddvq_s64): Likewise.
6858 (vaddvq_u8): Likewise.
6859 (vaddvq_u16): Likewise.
6860 (vaddvq_u32): Likewise.
6861 (vaddvq_u64): Likewise.
6862 (vaddv_f32): Likewise.
6863 (vaddvq_f32): Likewise.
6864 (vaddvq_f64): Likewise.
6865 (vmaxv_f32): Likewise.
6866 (vmaxv_s8): Likewise.
6867 (vmaxv_s16): Likewise.
6868 (vmaxv_s32): Likewise.
6869 (vmaxv_u8): Likewise.
6870 (vmaxv_u16): Likewise.
6871 (vmaxv_u32): Likewise.
6872 (vmaxvq_f32): Likewise.
6873 (vmaxvq_f64): Likewise.
6874 (vmaxvq_s8): Likewise.
6875 (vmaxvq_s16): Likewise.
6876 (vmaxvq_s32): Likewise.
6877 (vmaxvq_u8): Likewise.
6878 (vmaxvq_u16): Likewise.
6879 (vmaxvq_u32): Likewise.
6880 (vmaxnmv_f32): Likewise.
6881 (vmaxnmvq_f32): Likewise.
6882 (vmaxnmvq_f64): Likewise.
6883 (vminv_f32): Likewise.
6884 (vminv_s8): Likewise.
6885 (vminv_s16): Likewise.
6886 (vminv_s32): Likewise.
6887 (vminv_u8): Likewise.
6888 (vminv_u16): Likewise.
6889 (vminv_u32): Likewise.
6890 (vminvq_f32): Likewise.
6891 (vminvq_f64): Likewise.
6892 (vminvq_s8): Likewise.
6893 (vminvq_s16): Likewise.
6894 (vminvq_s32): Likewise.
6895 (vminvq_u8): Likewise.
6896 (vminvq_u16): Likewise.
6897 (vminvq_u32): Likewise.
6898 (vminnmv_f32): Likewise.
6899 (vminnmvq_f32): Likewise.
6900 (vminnmvq_f64): Likewise.
6902 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
6904 * config/aarch64/aarch64-simd.md
6905 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
6906 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
6907 (*aarch64_mul3_elt<mode>): Likewise.
6908 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
6909 (*aarch64_mul3_elt_to_64v2df): Likewise.
6910 (*aarch64_mla_elt<mode>): Likewise.
6911 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
6912 (*aarch64_mls_elt<mode>): Likewise.
6913 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
6914 (*aarch64_fma4_elt<mode>): Likewise.
6915 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
6916 (*aarch64_fma4_elt_to_64v2df): Likewise.
6917 (*aarch64_fnma4_elt<mode>): Likewise.
6918 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
6919 (*aarch64_fnma4_elt_to_64v2df): Likewise.
6920 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
6921 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
6922 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
6923 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
6924 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
6925 (aarch64_sqdmull_lane<mode>_internal): Likewise.
6926 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
6928 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
6930 * config/aarch64/aarch64-simd.md
6931 (aarch64_be_checked_get_lane<mode>): New define_expand.
6932 * config/aarch64/aarch64-simd-builtins.def
6933 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
6934 New builtin definition.
6935 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
6936 Use new safe be builtin.
6938 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
6940 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
6942 (aarch64_be_st1<mode>): Likewise.
6943 (aarch_ld1<VALL:mode>): Define_expand modified.
6944 (aarch_st1<VALL:mode>): Likewise.
6945 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
6946 (UNSPEC_ST1): Likewise.
6948 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
6950 * config/microblaze/microblaze.md: Add trap insn and attribute
6952 2014-01-23 Dodji Seketeli <dodji@redhat.com>
6954 PR preprocessor/58580
6955 * input.h (location_get_source_line): Take an additional line_size
6957 (void diagnostics_file_cache_fini): Declare new function.
6958 * input.c (struct fcache): New type.
6959 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
6960 New static constants.
6961 (diagnostic_file_cache_init, total_lines_num)
6962 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
6963 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
6964 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
6965 (get_next_line, read_next_line, goto_next_line, read_line_num):
6966 New static function definitions.
6967 (diagnostic_file_cache_fini): New function.
6968 (location_get_source_line): Take an additional output line_len
6969 parameter. Re-write using lookup_or_add_file_to_cache_tab and
6971 * diagnostic.c (diagnostic_finish): Call
6972 diagnostic_file_cache_fini.
6973 (adjust_line): Take an additional input parameter for the length
6974 of the line, rather than calculating it with strlen.
6975 (diagnostic_show_locus): Adjust the use of
6976 location_get_source_line and adjust_line with respect to their new
6977 signature. While displaying a line now, do not stop at the first
6978 null byte. Rather, display the zero byte as a space and keep
6979 going until we reach the size of the line.
6980 * Makefile.in: Add vec.o to OBJS-libcommon
6982 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
6983 Ilya Tocar <ilya.tocar@intel.com>
6985 * config/i386/avx512fintrin.h (_mm512_kmov): New.
6986 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
6987 (__builtin_ia32_kmov16): Ditto.
6988 * config/i386/i386.md (UNSPEC_KMOV): New.
6991 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
6993 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
6994 (_mm512_storeu_si512): Ditto.
6996 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
6999 * rtl.h (get_referenced_operands): Declare.
7000 * recog.c (get_referenced_operands): New function.
7001 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
7002 operands have been referenced when recording LO_SUM references.
7004 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
7006 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
7008 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
7010 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
7011 Enable for generic and recent AMD targets.
7013 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
7015 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
7016 ARG_SIZE note when adjustment was eliminated.
7018 2014-01-22 Jeff Law <law@redhat.com>
7020 PR tree-optimization/59597
7021 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
7022 in file. Accept new argument REGISTERING and use it to modify
7023 dump output appropriately.
7024 (register_jump_thread): Corresponding changes.
7025 (mark_threaded_blocks): Reinstate code to cancel unprofitable
7026 thread paths involving joiner blocks. Add code to dump cancelled
7027 jump threading paths.
7029 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
7031 PR rtl-optimization/59477
7032 * lra-constraints.c (inherit_in_ebb): Process call for living hard
7033 regs. Update reloads_num and potential_reload_hard_regs for all insns.
7035 2014-01-22 Tom Tromey <tromey@redhat.com>
7037 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
7039 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
7041 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
7043 PR rtl-optimization/59896
7044 * lra-constraints.c (process_alt_operands): Check unused note for
7045 matched operands of insn with no output reloads.
7047 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
7049 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
7050 (mips_move_from_gpr_cost): Likewise.
7052 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
7054 PR rtl-optimization/59858
7055 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
7056 ira_class_hard_regs_num.
7057 (process_alt_operands): Increase reject for dying matched operand.
7059 2014-01-21 Jakub Jelinek <jakub@redhat.com>
7062 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
7063 smaller than size, perform several stores or loads and stores
7064 at dst + count - size to store or copy all of size bytes, rather
7065 than just last modesize bytes.
7067 2014-01-20 DJ Delorie <dj@redhat.com>
7069 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
7070 that CLOBBERs are REGs before propogating their values.
7072 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
7075 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
7076 (cgraph_inline_failed_type): New function.
7077 * cgraph.h (DEFCIFCODE): Add type.
7078 (cgraph_inline_failed_type_t): New enum.
7079 (cgraph_inline_failed_type): New prototype.
7080 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
7081 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
7082 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
7083 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
7084 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
7085 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
7086 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
7087 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
7088 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
7089 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
7090 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
7091 OPTIMIZATION_MISMATCH.
7092 * tree-inline.c (expand_call_inline): Emit errors during
7093 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
7095 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7098 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
7099 mode attribute in insn output.
7101 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
7103 * output.h (output_constant): Delete.
7104 * varasm.c (output_constant): Make private.
7106 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
7108 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
7110 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7113 * tree.h (fold_builtin_strcat): New prototype.
7114 * builtins.c (fold_builtin_strcat): No longer static. Add len
7115 argument, if non-NULL, don't call c_strlen. Optimize
7116 directly into __builtin_memcpy instead of __builtin_strcpy.
7117 (fold_builtin_2): Adjust fold_builtin_strcat caller.
7118 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
7120 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7122 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7123 for SImode_address_operand operands, having only a REG argument.
7125 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
7127 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
7128 loader name using mbig-endian.
7129 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
7131 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7133 * doc/invoke.texi (-march): Clarify documentation for AArch64.
7137 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
7139 * config/aarch64/aarch64-protos.h
7140 (aarch64_cannot_change_mode_class_ptr): Declare.
7141 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
7142 aarch64_cannot_change_mode_class_ptr): New.
7143 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
7144 backend hook aarch64_cannot_change_mode_class.
7146 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7148 * common/config/aarch64/aarch64-common.c
7149 (aarch64_handle_option): Don't handle any option order logic here.
7150 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
7151 selected_cpu, warn on architecture version mismatch.
7152 (aarch64_override_options): Fix parsing order for option strings.
7154 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7155 Iain Sandoe <iain@codesourcery.com>
7158 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
7159 warning. Amend comment to reflect current functionality.
7161 2014-01-20 Richard Biener <rguenther@suse.de>
7164 * builtins.c (fold_builtin_strcat): Remove case better handled
7165 by tree-ssa-strlen.c.
7167 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
7169 * config/aarch64/aarch64.opt
7170 (mcpu, march, mtune): Make case-insensitive.
7172 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7175 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7176 if operands[1] is a REG or ZERO_EXTEND of a REG.
7178 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
7180 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
7182 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
7184 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
7185 long non-pic millicode calls.
7187 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7189 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
7191 2014-01-19 Kito Cheng <kito@0xlab.org>
7193 * builtins.c (expand_movstr): Check movstr expand done or fail.
7195 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7196 H.J. Lu <hongjiu.lu@intel.com>
7199 * config/i386/i386.md (*lea<mode>): Zero-extend return register
7200 to DImode for zero-extended addresses.
7202 2014-01-19 Jakub Jelinek <jakub@redhat.com>
7204 PR rtl-optimization/57763
7205 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
7206 on the new indirect jump_insn and increment LABEL_NUSES (label).
7208 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
7212 * config.gcc (x86_archs): New variable.
7213 (x86_64_archs): Likewise.
7214 (x86_cpus): Likewise.
7215 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
7216 --with-arch/--with-cpu= options.
7217 Support --with-arch=/--with-cpu={nehalem,westmere,
7218 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
7220 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7222 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
7223 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
7225 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7227 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
7229 2014-01-18 Jakub Jelinek <jakub@redhat.com>
7232 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
7233 clear cpp_get_options (parse_in)->warn_unused_macros for
7234 ix86_target_macros_internal with cpp_define.
7236 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
7238 * jump.c (delete_related_insns): Keep (use (insn))s.
7239 * reorg.c (redundant_insn): Check for barriers too.
7241 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7243 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
7245 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
7247 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
7248 call to $$dyncall when TARGET_LONG_CALLS is true.
7250 2014-01-17 Jeff Law <law@redhat.com>
7252 * ree.c (combine_set_extension): Temporarily disable test for
7253 changing number of hard registers.
7255 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7258 * ipa-inline-analysis.c (inline_free_summary):
7259 Do not free summary of aliases.
7261 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7264 * gimplify.c (gimplify_expr): Use create_tmp_var
7265 instead of create_tmp_var_raw. If cond doesn't have
7266 integral type, don't add the IFN_ANNOTATE builtin at all.
7268 2014-01-17 Martin Jambor <mjambor@suse.cz>
7271 * ipa-cp.c (prev_edge_clone): New variable.
7272 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
7273 Also resize prev_edge_clone vector.
7274 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
7275 (ipcp_edge_removal_hook): New function.
7276 (ipcp_driver): Register ipcp_edge_removal_hook.
7278 2014-01-17 Andrew Pinski <apinski@cavium.com>
7279 Steve Ellcey <sellcey@mips.com>
7282 * config/mips/mips.c (mips_print_operand): Check operand mode instead
7285 2014-01-17 Jeff Law <law@redhat.com>
7288 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
7289 so that pass_ccp runs first.
7291 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7293 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
7294 (ix86_adjust_cost): Use !TARGET_XXX.
7295 (do_reorder_for_imul): Likewise.
7296 (swap_top_of_ready_list): Likewise.
7297 (ix86_sched_reorder): Likewise.
7299 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7301 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
7302 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
7303 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
7304 (intel_memset): New. Duplicate slm_memset.
7305 (intel_cost): New. Duplicate slm_cost.
7306 (m_INTEL): New macro.
7307 (processor_target_table): Add "intel".
7308 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
7309 with PROCESSOR_INTEL for "intel".
7310 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
7311 PROCESSOR_SILVERMONT.
7312 (ix86_issue_rate): Likewise.
7313 (ix86_adjust_cost): Likewise.
7314 (ia32_multipass_dfa_lookahead): Likewise.
7315 (swap_top_of_ready_list): Likewise.
7316 (ix86_sched_reorder): Likewise.
7317 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
7318 instead of TARGET_OPT_AGU.
7319 * config/i386/i386.h (TARGET_INTEL): New.
7320 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
7321 (processor_type): Add PROCESSOR_INTEL.
7322 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
7323 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
7325 2014-01-17 Marek Polacek <polacek@redhat.com>
7328 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
7331 2014-01-17 Richard Biener <rguenther@suse.de>
7333 PR tree-optimization/46590
7334 * opts.c (default_options_table): Add entries for
7335 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
7336 all enabled at -O1 but not for -Og.
7337 * common.opt (fbranch-count-reg): Remove Init(1).
7338 (fmove-loop-invariants): Likewise.
7339 (ftree-pta): Likewise.
7341 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7343 * config/i386/i386.c (ix86_data_alignment): For compatibility with
7344 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
7345 decls to at least the GCC 4.8 used alignments.
7348 * tree-nested.c (convert_nonlocal_reference_stmt,
7349 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
7350 of GIMPLE_BIND stmts, adjust associated decls.
7352 2014-01-17 Richard Biener <rguenther@suse.de>
7354 PR tree-optimization/46590
7355 * vec.h (vec<>::bseach): New member function implementing
7356 binary search according to C89 bsearch.
7357 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
7358 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
7359 bitmap pointer again. Make accesses_in_loop a flat array.
7360 (mem_ref_obstack): New global.
7361 (outermost_indep_loop): Adjust for mem_ref->stored changes.
7362 (mark_ref_stored): Likewise.
7363 (ref_indep_loop_p_2): Likewise.
7364 (set_ref_stored_in_loop): New helper function.
7365 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
7366 (memref_free): Adjust.
7367 (record_mem_ref_loc): Simplify.
7368 (gather_mem_refs_stmt): Adjust.
7369 (sort_locs_in_loop_postorder_cmp): New function.
7370 (analyze_memory_references): Sort accesses_in_loop after
7371 loop postorder number.
7372 (find_ref_loc_in_loop_cmp): New function.
7373 (for_all_locs_in_loop): Find relevant cluster of locs in
7374 accesses_in_loop and iterate without recursion.
7375 (execute_sm): Avoid uninit warning.
7376 (struct ref_always_accessed): Simplify.
7377 (ref_always_accessed::operator ()): Likewise.
7378 (ref_always_accessed_p): Likewise.
7379 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
7380 loop postorder numbers here.
7381 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
7384 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7387 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
7388 on decls for which assemble_alias has been called.
7390 2014-01-17 Nick Clifton <nickc@redhat.com>
7392 * config/msp430/msp430.opt: (mcpu): New option.
7393 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
7394 (msp430_option_override): Parse target_cpu. If the MCU name
7395 matches a generic string, clear target_mcu.
7396 (msp430_attr): Allow numeric interrupt values up to 63.
7397 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
7398 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
7400 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
7402 * config/msp430/msp430.md (popm): Use %J rather than %I.
7403 (addsi3): Use msp430_nonimmediate_operand for operand 2.
7404 (addhi_cy_i): Use immediate_operand for operand 2.
7405 * doc/invoke.texi: Document -mcpu option.
7407 2014-01-17 Richard Biener <rguenther@suse.de>
7409 PR rtl-optimization/38518
7410 * df.h (df_analyze_loop): Declare.
7411 * df-core.c: Include cfgloop.h.
7412 (df_analyze_1): Split out main part of df_analyze.
7413 (df_analyze): Adjust.
7414 (loop_inverted_post_order_compute): New function.
7415 (loop_post_order_compute): Likewise.
7416 (df_analyze_loop): New function avoiding whole-function
7418 * loop-invariant.c (find_defs): Use df_analyze_loop.
7419 (find_invariants): Adjust.
7420 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
7422 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
7424 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
7425 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
7427 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
7429 * ipa-ref.c (ipa_remove_stmt_references): Fix references
7430 traversal when removing references.
7432 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
7435 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
7437 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
7440 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
7441 pushing a reload for an autoinc when we had previously reloaded an
7442 inner part of the address.
7444 2014-01-16 Jakub Jelinek <jakub@redhat.com>
7446 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
7448 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
7449 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
7450 when not giving up or versioning for alias only because of
7452 (vect_analyze_data_ref_dependences): Set to true.
7453 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
7455 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
7456 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
7460 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
7463 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
7464 operand 0 predicate for gathers, use a new pseudo as subtarget.
7466 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7469 * lra-constraints.c (process_alt_operands): Add printing debug info.
7470 Check absence of input/output reloads for matched operands too.
7472 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7474 PR rtl-optimization/59835
7475 * ira.c (ira_init_register_move_cost): Increase cost for
7478 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
7480 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
7482 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
7485 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
7486 non-register objects. Use gen_(high/low)part more consistently.
7489 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
7492 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
7493 endian support, remove tests for WORDS_BIG_ENDIAN.
7494 (p8_mfvsrd_3_<mode>): Likewise.
7495 (reload_gpr_from_vsx<mode>): Likewise.
7496 (reload_gpr_from_vsxsf): Likewise.
7497 (p8_mfvsrd_4_disf): Likewise.
7499 2014-01-16 Richard Biener <rguenther@suse.de>
7501 PR rtl-optimization/46590
7502 * lcm.c (compute_antinout_edge): Use postorder iteration.
7503 (compute_laterin): Use inverted postorder iteration.
7505 2014-01-16 Nick Clifton <nickc@redhat.com>
7508 * varasm.c (output_constant): Return the number of bytes actually
7510 (output_constructor_array_range): Update the field size with the
7511 number of bytes emitted by output_constant.
7512 (output_constructor_regular_field): Likewise. Also do not
7513 complain if the total number of bytes emitted is now greater
7514 than the expected fieldpos.
7515 * output.h (output_constant): Update prototype and descriptive comment.
7517 2014-01-16 Marek Polacek <polacek@redhat.com>
7520 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
7521 it is error_mark_node.
7523 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
7525 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
7526 VALID_AVX256_REG_OR_OI_MODE.
7528 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
7530 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
7531 current procedure should be profiled.
7533 2014-01-15 Andrew Pinski <apinski@cavium.com>
7535 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
7536 of moving from/to the STACK_REG register class.
7538 2014-01-15 Richard Henderson <rth@redhat.com>
7541 * reginfo.c (global_regs_decl): Globalize.
7542 * rtl.h (global_regs_decl): Declare.
7543 * ira.c (do_reload): Diagnose frame_pointer_needed and it
7544 reserved via global_regs.
7546 2014-01-15 Teresa Johnson <tejohnson@google.com>
7548 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
7550 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
7552 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
7553 and vmulosh rather than call gen_vec_widen_smult_*.
7554 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
7555 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
7556 (vec_widen_smult_even_v16qi): Likewise.
7557 (vec_widen_umult_even_v8hi): Likewise.
7558 (vec_widen_smult_even_v8hi): Likewise.
7559 (vec_widen_umult_odd_v16qi): Likewise.
7560 (vec_widen_smult_odd_v16qi): Likewise.
7561 (vec_widen_umult_odd_v8hi): Likewise.
7562 (vec_widen_smult_odd_v8hi): Likewise.
7563 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
7564 vmuloub rather than call gen_vec_widen_umult_*.
7565 (vec_widen_umult_lo_v16qi): Likewise.
7566 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
7567 vmulosb rather than call gen_vec_widen_smult_*.
7568 (vec_widen_smult_lo_v16qi): Likewise.
7569 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
7570 rather than call gen_vec_widen_umult_*.
7571 (vec_widen_umult_lo_v8hi): Likewise.
7572 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
7573 rather than call gen_vec_widen_smult_*.
7574 (vec_widen_smult_lo_v8hi): Likewise.
7576 2014-01-15 Jeff Law <law@redhat.com>
7578 PR tree-optimization/59747
7579 * ree.c (find_and_remove_re): Properly handle case where a second
7580 eliminated extension requires widening a copy created for elimination
7581 of a prior extension.
7582 (combine_set_extension): Ensure that the number of hard regs needed
7583 for a destination register does not change when we widen it.
7585 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
7587 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
7588 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
7589 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
7590 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
7591 (avr-*-rtems*): Likewise.
7592 (bfin*-rtems*): Likewise.
7593 (moxie-*-rtems*): Likewise.
7594 (h8300-*-rtems*): Likewise.
7595 (i[34567]86-*-rtems*): Likewise.
7596 (lm32-*-rtems*): Likewise.
7597 (m32r-*-rtems*): Likewise.
7598 (m68k-*-rtems*): Likewise.
7599 (microblaze*-*-rtems*): Likewise.
7600 (mips*-*-rtems*): Likewise.
7601 (powerpc-*-rtems*): Likewise.
7602 (sh-*-rtems*): Likewise.
7603 (sparc-*-rtems*): Likewise.
7604 (sparc64-*-rtems*): Likewise.
7605 (v850-*-rtems*): Likewise.
7606 (m32c-*-rtems*): Likewise.
7608 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
7610 PR rtl-optimization/59511
7611 * ira.c (ira_init_register_move_cost): Use memory costs for some
7612 cases of register move cost calculations.
7613 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
7614 instead of BB frequency.
7615 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
7616 * lra-assigns.c (find_hard_regno_for): Ditto.
7618 2014-01-15 Richard Biener <rguenther@suse.de>
7620 PR tree-optimization/59822
7621 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
7622 (vectorizable_load): Use it to hoist defs of uses of invariant
7623 loads out of the loop.
7625 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
7626 Kugan Vivekanandarajah <kuganv@linaro.org>
7629 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
7632 2014-01-15 Richard Biener <rguenther@suse.de>
7634 PR rtl-optimization/59802
7635 * lcm.c (compute_available): Use inverted postorder to seed
7636 the initial worklist.
7638 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7641 * config/s390/s390.c (s390_preferred_reload_class): Don't return
7642 ADDR_REGS for invalid symrefs in non-PIC code.
7644 2014-01-15 Jakub Jelinek <jakub@redhat.com>
7647 * builtins.c (determine_block_size): Initialize *probable_max_size
7648 even if len_rtx is CONST_INT.
7650 2014-01-14 Andrew Pinski <apinski@cavium.com>
7652 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
7653 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
7654 (cortexa53_tunings): Likewise.
7655 (aarch64_sched_issue_rate): New function.
7656 (TARGET_SCHED_ISSUE_RATE): Define.
7658 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
7660 * ira-costs.c (find_costs_and_classes): Add missed
7661 ira_init_register_move_cost_if_necessary.
7663 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
7666 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
7668 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
7671 * config/i386/i386.c (type_natural_mode): Add a bool parameter
7672 to indicate if type is used for function return value. Warn ABI
7673 change if the vector mode isn't available for function return value.
7674 (ix86_function_arg_advance): Pass false to type_natural_mode.
7675 (ix86_function_arg): Likewise.
7676 (ix86_gimplify_va_arg): Likewise.
7677 (function_arg_32): Don't warn ABI change.
7678 (ix86_function_value): Pass true to type_natural_mode.
7679 (ix86_return_in_memory): Likewise.
7680 (ix86_struct_value_rtx): Removed.
7681 (TARGET_STRUCT_VALUE_RTX): Likewise.
7683 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7685 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
7686 converting a conditional jump into a conditional return.
7688 2014-01-14 Richard Biener <rguenther@suse.de>
7690 PR tree-optimization/58921
7691 PR tree-optimization/59006
7692 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
7693 hoisting invariant stmts.
7694 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
7695 invariant loads on the preheader edge if possible.
7697 2014-01-14 Joey Ye <joey.ye@arm.com>
7699 * doc/plugin.texi (Building GCC plugins): Update to C++.
7701 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
7703 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
7704 (_mm_rcp28_round_ss): Ditto.
7705 (_mm_rsqrt28_round_sd): Ditto.
7706 (_mm_rsqrt28_round_ss): Ditto.
7707 (_mm_rcp28_sd): Ditto.
7708 (_mm_rcp28_ss): Ditto.
7709 (_mm_rsqrt28_sd): Ditto.
7710 (_mm_rsqrt28_ss): Ditto.
7711 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
7712 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
7713 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
7714 (IX86_BUILTIN_RCP28SD): Ditto.
7715 (IX86_BUILTIN_RCP28SS): Ditto.
7716 (IX86_BUILTIN_RSQRT28SD): Ditto.
7717 (IX86_BUILTIN_RSQRT28SS): Ditto.
7718 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
7719 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
7720 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
7721 (ix86_expand_special_args_builtin): Expand new FTYPE.
7722 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
7723 (srcp14<mode>): Make insn unary.
7724 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
7725 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
7726 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
7727 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
7728 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
7729 Fix rounding: make it SAE only.
7730 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
7732 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
7734 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
7735 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
7736 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
7737 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
7738 (round_saeonly_mask_scalar_operand4): Ditto.
7739 (round_saeonly_mask_scalar_op3): Ditto.
7740 (round_saeonly_mask_scalar_op4): Ditto.
7742 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7744 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
7745 Implement -maltivec=be for vec_insert and vec_extract.
7747 2014-01-10 DJ Delorie <dj@redhat.com>
7749 * config/msp430/msp430.md (call_internal): Don't allow memory
7750 references with SP as the base register.
7751 (call_value_internal): Likewise.
7752 * config/msp430/constraints.md (Yc): New. For memory references
7753 that don't use SP as a base register.
7755 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
7756 "an integer without a # prefix"
7757 * config/msp430/msp430.md (epilogue_helper): Use it.
7759 2014-01-13 Jakub Jelinek <jakub@redhat.com>
7762 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
7763 AVX512F gather builtins.
7764 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
7765 on gather decls with INTEGER_TYPE masktype.
7766 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
7767 directly into the builtin rather than hoisting it before loop.
7769 PR tree-optimization/59387
7770 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
7771 (scev_const_prop): If folded_casts and type has undefined overflow,
7772 use force_gimple_operand instead of force_gimple_operand_gsi and
7773 for each added stmt if it is assign with
7774 arith_code_with_undefined_signed_overflow, call
7775 rewrite_to_defined_overflow.
7776 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
7777 gimple-fold.h instead.
7778 (arith_code_with_undefined_signed_overflow,
7779 rewrite_to_defined_overflow): Moved to ...
7780 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
7781 rewrite_to_defined_overflow): ... here. No longer static.
7782 Include gimplify-me.h.
7783 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
7784 rewrite_to_defined_overflow): New prototypes.
7786 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7788 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
7790 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
7792 * builtins.c (get_object_alignment_2): Minor tweak.
7793 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
7795 2014-01-13 Christian Bruel <christian.bruel@st.com>
7797 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
7798 optimized non constant lengths.
7800 2014-01-13 Jakub Jelinek <jakub@redhat.com>
7803 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
7804 load as __atomic_load_N if possible.
7806 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
7808 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
7810 (rs6000_expand_builtin): Adjust call.
7812 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
7815 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
7816 * config/rs6000/rs6000.c: Include target-globals.h.
7817 (rs6000_set_current_function): Instead of doing target_reinit
7818 unconditionally, use save_target_globals_default_opts and
7819 restore_target_globals.
7821 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
7823 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
7824 (rs6000_expand_builtin): Handle mffs and mtfsf.
7825 (rs6000_init_builtins): Define mffs and mtfsf.
7826 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
7827 (rs6000_mffs): New pattern.
7828 (rs6000_mtfsf): New pattern.
7830 2014-01-11 Bin Cheng <bin.cheng@arm.com>
7832 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
7833 Start narrowing with START. Apply candidate-use pair
7834 and check overall cost in narrowing.
7835 (iv_ca_prune): Pass new argument.
7837 2014-01-10 Jeff Law <law@redhat.com>
7840 * ree.c (combine_reaching_defs): Ensure the defining statement
7841 occurs before the extension when optimizing extensions with
7842 different source and destination hard registers.
7844 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
7847 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
7848 vtables into the type inheritance graph.
7850 2014-01-10 Jakub Jelinek <jakub@redhat.com>
7852 PR rtl-optimization/59754
7853 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
7854 modes in the REGNO != REGNO case.
7856 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7858 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
7860 2014-01-10 Jakub Jelinek <jakub@redhat.com>
7862 PR tree-optimization/59745
7863 * tree-predcom.c (tree_predictive_commoning_loop): Call
7864 free_affine_expand_cache if giving up because components is NULL.
7866 * target-globals.c (save_target_globals): Allocate < 4KB structs using
7867 GC in payload of target_globals struct instead of allocating them on
7868 the heap and the larger structs separately using GC.
7869 * target-globals.h (struct target_globals): Make regs, hard_regs,
7870 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
7871 of GTY((skip)) and change type to void *.
7872 (reset_target_globals): Cast loads from those fields to corresponding
7875 2014-01-10 Steve Ellcey <sellcey@mips.com>
7878 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
7879 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
7880 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
7882 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
7885 * aarch64-modes.def (CC_Zmode): New flags mode.
7886 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
7887 represents an equality.
7888 (aarch64_get_condition_code): Handle CC_Zmode.
7889 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
7891 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7893 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
7894 extraction in good case.
7896 2014-01-10 Richard Biener <rguenther@suse.de>
7898 PR tree-optimization/59374
7899 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
7900 checking after SLP discovery. Mark stmts not participating
7901 in any SLP instance properly.
7903 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7905 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
7906 when handling a SET rtx.
7908 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7910 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
7911 (cortex-a57): Likewise.
7912 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
7914 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7916 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
7917 non-iwmmxt builtins.
7919 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
7923 * ipa-devirt.c record_target_from_binfo): Take as argument
7924 stack of binfos and lookup matching one for virtual inheritance.
7925 (possible_polymorphic_call_targets_1): Update.
7927 2014-01-10 Huacai Chen <chenhc@lemote.com>
7929 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
7930 kernel strings for Loongson-2E/2F/3A.
7932 2014-01-10 Jakub Jelinek <jakub@redhat.com>
7935 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
7936 is_gimple_call before calling gimple_call_internal_p.
7938 2014-01-09 Steve Ellcey <sellcey@mips.com>
7940 * Makefile.in (TREE_FLOW_H): Remove.
7941 (TREE_SSA_H): Add file names from tree-flow.h.
7942 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
7943 * tree.h: Remove tree-flow.h reference.
7944 * hash-table.h: Remove tree-flow.h reference.
7945 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
7946 reference with tree-ssa-loop.h.
7948 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7950 * doc/invoke.texi: Add -maltivec={be,le} options, and document
7951 default element-order behavior for -maltivec.
7952 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
7953 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
7954 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
7955 when targeting big endian, at least for now.
7956 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
7958 2014-01-09 Jakub Jelinek <jakub@redhat.com>
7961 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
7962 var satisfies use_register_for_decl, just take into account type
7963 alignment, rather than decl alignment.
7965 PR tree-optimization/59622
7966 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
7967 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
7968 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
7969 Don't devirtualize for inplace at all. For targets.length () == 1,
7970 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
7972 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
7974 * config/i386/i386.md (cpu): Remove the unused btver1.
7976 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
7978 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
7980 2014-01-09 Jakub Jelinek <jakub@redhat.com>
7983 * tree-core.h (struct target_globals): New forward declaration.
7984 (struct tree_target_option): Add globals field.
7985 * tree.h (TREE_TARGET_GLOBALS): Define.
7986 (prepare_target_option_nodes_for_pch): New prototype.
7987 * target-globals.h (struct target_globals): Define even if
7989 * tree.c (prepare_target_option_node_for_pch,
7990 prepare_target_option_nodes_for_pch): New functions.
7991 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
7992 * config/i386/i386.c: Include target-globals.h.
7993 (ix86_set_current_function): Instead of doing target_reinit
7994 unconditionally, use save_target_globals_default_opts and
7995 restore_target_globals.
7997 2014-01-09 Richard Biener <rguenther@suse.de>
7999 PR tree-optimization/59715
8000 * tree-cfg.h (split_critical_edges): Declare.
8001 * tree-cfg.c (split_critical_edges): Export.
8002 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
8004 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
8006 * cfgexpand.c (expand_stack_vars): Optionally disable
8007 asan stack protection.
8008 (expand_used_vars): Likewise.
8009 (partition_stack_vars): Likewise.
8010 * asan.c (asan_emit_stack_protection): Optionally disable
8011 after return stack usage.
8012 (instrument_derefs): Optionally disable memory access instrumentation.
8013 (instrument_builtin_call): Likewise.
8014 (instrument_strlen_call): Likewise.
8015 (asan_protect_global): Optionally disable global variables protection.
8016 * doc/invoke.texi: Added doc for new options.
8017 * params.def: Added new options.
8018 * params.h: Likewise.
8020 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8022 PR rtl-optimization/59724
8023 * ifcvt.c (cond_exec_process_if_block): Don't call
8024 flow_find_head_matching_sequence with 0 longest_match.
8025 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
8026 non-active insns if !stop_after.
8027 (try_head_merge_bb): Revert 2014-01-07 changes.
8029 2014-01-08 Jeff Law <law@redhat.com>
8031 * ree.c (get_sub_rtx): New function, extracted from...
8032 (merge_def_and_ext): Here.
8033 (combine_reaching_defs): Use get_sub_rtx.
8035 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
8037 * cgraph.h (varpool_variable_node): Do not choke on null node.
8039 2014-01-08 Catherine Moore <clm@codesourcery.com>
8041 * config/mips/mips.md (simple_return): Attempt to use JRC
8043 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
8045 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
8047 PR rtl-optimization/59137
8048 * reorg.c (steal_delay_list_from_target): Call update_block for
8050 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
8052 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8054 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
8055 two duplicate entries.
8057 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
8060 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
8062 * config/mips/mips.c (mips_truncated_op_cost): New function.
8063 (mips_rtx_costs): Adjust test for BADDU.
8064 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
8066 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
8068 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
8069 (*baddu_si): ...this new pattern.
8071 2014-01-08 Jakub Jelinek <jakub@redhat.com>
8074 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
8076 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
8079 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
8081 (expand_expr, expand_normal): Adjust.
8082 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
8083 inner_reference_p. Use inner_reference_p to expand inner references.
8084 (store_expr): Adjust.
8085 * cfgexpand.c (expand_call_stmt): Adjust.
8087 2014-01-08 Rong Xu <xur@google.com>
8089 * gcov-io.c (gcov_var): Move from gcov-io.h.
8090 (gcov_position): Ditto.
8091 (gcov_is_error): Ditto.
8092 (gcov_rewrite): Ditto.
8093 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
8094 only part to libgcc/libgcov.h.
8096 2014-01-08 Marek Polacek <polacek@redhat.com>
8099 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
8101 2014-01-08 Marek Polacek <polacek@redhat.com>
8104 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
8106 2014-01-08 Jakub Jelinek <jakub@redhat.com>
8108 PR rtl-optimization/59649
8109 * stor-layout.c (get_mode_bounds): For BImode return
8110 0 and STORE_FLAG_VALUE.
8112 2014-01-08 Richard Biener <rguenther@suse.de>
8115 * gimple.h (is_gimple_builtin_call): Remove.
8116 (gimple_builtin_call_types_compatible_p): New.
8117 (gimple_call_builtin_p): New overload.
8118 * gimple.c (is_gimple_builtin_call): Remove.
8119 (validate_call): Rename to ...
8120 (gimple_builtin_call_types_compatible_p): ... this and export. Also
8122 (validate_type): New static function.
8123 (gimple_call_builtin_p): New overload and adjust.
8124 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
8125 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
8126 (gimple_fold_stmt_to_constant_1): Likewise.
8127 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
8129 2014-01-08 Richard Biener <rguenther@suse.de>
8132 * gimplify.c (gimplify_expr): Gimplify register-register type
8133 VIEW_CONVERT_EXPRs to separate stmts.
8135 2014-01-07 Jeff Law <law@redhat.com>
8138 * ree.c (combine_set_extension): Handle case where source
8139 and destination registers in an extension insn are different.
8140 (combine_reaching_defs): Allow source and destination registers
8141 in extension to be different under limited circumstances.
8142 (add_removable_extension): Remove restriction that the
8143 source and destination registers in the extension are the same.
8144 (find_and_remove_re): Emit a copy from the extension's
8145 destination to its source after the defining insn if
8146 the source and destination registers are different.
8149 * ifcvt.c (merge_if_block): If we are merging a block with more than
8150 one successor with a block with no successors, remove any BARRIER
8151 after the second block.
8153 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
8155 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
8157 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
8160 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
8161 for 14-bit register offsets when INT14_OK_STRICT is false.
8163 2014-01-07 Roland Stigge <stigge@antcom.de>
8164 Michael Meissner <meissner@linux.vnet.ibm.com>
8167 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
8168 Only check TFmode for SPE constants. Don't check TImode or TDmode.
8170 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
8172 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
8175 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
8177 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
8178 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
8179 rtx is const0_rtx or not.
8181 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
8184 * target-globals.c (save_target_globals): Remove this_fn_optab
8186 * toplev.c: Include optabs.h.
8187 (target_reinit): Temporarily restore the global options if another
8188 set of options are in force.
8190 2014-01-07 Jakub Jelinek <jakub@redhat.com>
8192 PR rtl-optimization/58668
8193 * cfgcleanup.c (flow_find_cross_jump): Don't count
8194 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
8195 to determine what is counted.
8196 (flow_find_head_matching_sequence): Use active_insn_p to determine
8198 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
8200 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
8201 determine what is counted.
8203 PR tree-optimization/59643
8204 * tree-predcom.c (split_data_refs_to_components): If one dr is
8205 read and one write, determine_offset fails and the write isn't
8206 in the bad component, just put the read into the bad component.
8208 2014-01-07 Mike Stump <mikestump@comcast.net>
8209 Jakub Jelinek <jakub@redhat.com>
8212 * tree-core.h (struct tree_optimization_option): Change optabs
8213 type from unsigned char * to void *.
8214 * optabs.c (init_tree_optimization_optabs): Adjust
8215 TREE_OPTIMIZATION_OPTABS initialization.
8217 2014-01-06 Jakub Jelinek <jakub@redhat.com>
8220 * config/i386/i386.h (struct machine_function): Add
8221 no_drap_save_restore field.
8222 * config/i386/i386.c (ix86_save_reg): Use
8223 !cfun->machine->no_drap_save_restore instead of
8224 crtl->stack_realign_needed.
8225 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
8226 this function clears frame_pointer_needed. Set
8227 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
8228 and DRAP reg is needed.
8230 2014-01-06 Marek Polacek <polacek@redhat.com>
8233 * doc/implement-c.texi: Mention that other integer types are
8234 permitted as bit-field types in strictly conforming mode.
8236 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
8238 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
8241 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
8243 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
8245 2014-01-06 Martin Jambor <mjambor@suse.cz>
8248 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
8250 * ipa-prop.c (ipa_print_node_params): Fix indentation.
8252 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
8256 * var-tracking.c (add_stores): Preserve the value of the source even if
8257 we don't record the store.
8259 2014-01-06 Terry Guo <terry.guo@arm.com>
8261 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
8263 2014-01-05 Iain Sandoe <iain@codesourcery.com>
8266 * config/darwin.c (darwin_function_section): Adjust return values to
8267 correspond to optimisation changes made in r206070.
8269 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
8271 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
8272 from prefetch_block tune setting.
8273 (nocona_cost): Correct size of prefetch block to 64.
8275 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
8277 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
8278 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
8279 used to save the static chain register in the computation of the offset
8280 from which the FP registers need to be restored.
8282 2014-01-04 Jakub Jelinek <jakub@redhat.com>
8284 PR tree-optimization/59519
8285 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
8286 ICE if get_current_def (current_new_name) is already non-NULL, as long
8287 as it is a phi result of some other phi in *new_exit_bb that has
8290 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
8291 or vmovdqu* for misaligned_operand.
8292 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
8293 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
8294 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
8295 aligned_mem for AVX512F masked aligned load and store builtins and for
8298 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
8300 PR tree-optimization/59651
8301 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
8302 Address range for negative step should be added by TYPE_SIZE_UNIT.
8304 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
8306 * config/m68k/m68k.c (handle_move_double): Handle pushes with
8307 overlapping registers also for registers other than the stack pointer.
8309 2014-01-03 Marek Polacek <polacek@redhat.com>
8312 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
8315 2014-01-03 Jakub Jelinek <jakub@redhat.com>
8318 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
8321 * config/i386/i386.md (MODE_SIZE): New mode attribute.
8322 (push splitter): Use <P:MODE_SIZE> instead of
8323 GET_MODE_SIZE (<P:MODE>mode).
8324 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
8325 (mov -1, reg peephole2): Likewise.
8326 * config/i386/sse.md (*mov<mode>_internal,
8327 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
8328 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
8329 *<code><mode>3, *andnot<mode>3<mask_name>,
8330 <mask_codefor><code><mode>3<mask_name>): Likewise.
8331 * config/i386/subst.md (mask_mode512bit_condition,
8332 sd_mask_mode512bit_condition): Likewise.
8334 2014-01-02 Xinliang David Li <davidxl@google.com>
8336 PR tree-optimization/59303
8337 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
8338 (dump_predicates): Better output format.
8339 (pred_equal_p): New function.
8340 (is_neq_relop_p): Ditto.
8341 (is_neq_zero_form_p): Ditto.
8342 (pred_expr_equal_p): Ditto.
8343 (pred_neg_p): Ditto.
8344 (simplify_pred): Ditto.
8345 (simplify_preds_2): Ditto.
8346 (simplify_preds_3): Ditto.
8347 (simplify_preds_4): Ditto.
8348 (simplify_preds): Ditto.
8350 (push_to_worklist): Ditto.
8351 (get_pred_info_from_cmp): Ditto.
8352 (is_degenerated_phi): Ditto.
8353 (normalize_one_pred_1): Ditto.
8354 (normalize_one_pred): Ditto.
8355 (normalize_one_pred_chain): Ditto.
8356 (normalize_preds): Ditto.
8357 (normalize_cond_1): Remove function.
8358 (normalize_cond): Ditto.
8359 (is_gcond_subset_of): Ditto.
8360 (is_subset_of_any): Ditto.
8361 (is_or_set_subset_of): Ditto.
8362 (is_and_set_subset_of): Ditto.
8363 (is_norm_cond_subset_of): Ditto.
8364 (pred_chain_length_cmp): Ditto.
8365 (convert_control_dep_chain_into_preds): Type change.
8366 (find_predicates): Ditto.
8367 (find_def_preds): Ditto.
8368 (destroy_predicates_vecs): Ditto.
8369 (find_matching_predicates_in_rest_chains): Ditto.
8370 (use_pred_not_overlap_with_undef_path_pred): Ditto.
8371 (is_pred_expr_subset): Ditto.
8372 (is_pred_chain_subset_of): Ditto.
8373 (is_included_in): Ditto.
8374 (is_superset_of): Ditto.
8376 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8378 Update copyright years.
8380 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8382 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
8383 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
8384 config/arc/arc.md, config/arc/arc.opt,
8385 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
8386 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
8387 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
8388 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
8389 config/linux-protos.h, config/linux.c, config/winnt-c.c,
8390 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
8391 vtable-verify.c, vtable-verify.h: Use the standard form for the
8394 2014-01-02 Tobias Burnus <burnus@net-b.de>
8396 * gcc.c (process_command): Update copyright notice dates.
8397 * gcov-dump.c: Ditto.
8399 * doc/cpp.texi: Bump @copying's copyright year.
8400 * doc/cppinternals.texi: Ditto.
8401 * doc/gcc.texi: Ditto.
8402 * doc/gccint.texi: Ditto.
8403 * doc/gcov.texi: Ditto.
8404 * doc/install.texi: Ditto.
8405 * doc/invoke.texi: Ditto.
8407 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8409 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
8411 2014-01-01 Jakub Jelinek <jakub@redhat.com>
8413 * config/i386/sse.md (*mov<mode>_internal): Guard
8414 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
8416 PR rtl-optimization/59647
8417 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
8418 new_rtx into UNSIGNED_FLOAT rtxes.
8420 Copyright (C) 2014 Free Software Foundation, Inc.
8422 Copying and distribution of this file, with or without modification,
8423 are permitted in any medium without royalty provided the copyright
8424 notice and this notice are preserved.