| blob | 8308c260bcafb4752961a3c30f79d1de762b168c |
1 2014-05-14 Georg-Johann Lay <avr@gjlay.de>
3 * config/avr/avr.h (REG_CLASS_CONTENTS): Use unsigned suffix for
4 shifted values to avoid build warning.
6 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
8 * cfgcleanup.c (try_forward_edges): Use location_t for locations.
9 * cfgrtl.c (rtl_merge_blocks): Fix comment.
10 (cfg_layout_merge_blocks): Likewise.
11 * except.c (emit_to_new_bb_before): Remove prev_bb local variable.
13 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
15 PR rtl-optimization/60901
16 * config/i386/i386.c (ix86_dependencies_evaluation_hook): Check that
17 bb predecessor belongs to the same scheduling region. Adjust comment.
19 2014-05-13 Peter Bergner <bergner@vnet.ibm.com>
21 * doc/sourcebuild.texi: (dfp_hw): Document.
22 (p8vector_hw): Likewise.
23 (powerpc_eabi_ok): Likewise.
24 (powerpc_elfv2): Likewise.
25 (powerpc_htm_ok): Likewise.
26 (ppc_recip_hw): Likewise.
27 (vsx_hw): Likewise.
29 2014-05-13 Cary Coutant <ccoutant@google.com>
31 * opts.c (finish_options): Use -ggnu-pubnames with -gsplit-dwarf.
33 2014-05-13 David Malcolm <dmalcolm@redhat.com>
35 * gengtype-parse.c (require3): Eliminate in favor of...
36 (require4): New.
37 (require_template_declaration): Update to support optional single *
38 on a type.
40 * gengtype.c (get_ultimate_base_class): Add a non-const overload.
41 (create_user_defined_type): Handle a single level of explicit
42 pointerness within template arguments.
43 (struct write_types_data): Add field "kind".
44 (filter_type_name): Handle "*" character.
45 (write_user_func_for_structure_ptr): Require a write_types_data
46 rather than just a prefix string, so that we can look up the kind
47 of the wtd and use it as an index into wrote_user_func_for_ptr,
48 ensuring that such functions are written at most once. Support
49 subclasses by invoking the marking function of the ultimate base
50 class.
51 (write_user_func_for_structure_body): Require a write_types_data
52 rather than just a prefix string, so that we can pass this to
53 write_user_func_for_structure_ptr.
54 (write_func_for_structure): Likewise.
55 (ggc_wtd): Add initializer of new "kind" field.
56 (pch_wtd): Likewise.
58 * gengtype.h (enum write_types_kinds): New.
59 (struct type): Add field wrote_user_func_for_ptr to the "s"
60 union member.
62 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
64 * fold-const.c (optimize_bit_field_compare): Use wi:: operations
65 instead of const_binop.
66 (fold_binary_loc): Likewise.
68 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
70 * tree-dfa.h (get_addr_base_and_unit_offset_1): Update array index
71 calculation to match get_ref_base_and_extent.
73 2014-05-13 Catherine Moore <clm@codesourcery.com>
74 Sandra Loosemore <sandra@codesourcery.com>
76 * configure.ac: Fix assembly for explicit JALR relocation check.
77 * configure: Regenerate.
79 2014-05-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
81 * config/arm/arm.c (neon_itype): Remove NEON_RESULTPAIR.
82 (arm_init_neon_builtins): Remove handling of NEON_RESULTPAIR.
83 Remove associated type declarations and initialisations.
84 (arm_expand_neon_builtin): Likewise.
85 (neon_emit_pair_result_insn): Delete.
86 * config/arm/arm_neon_builtins (vtrn, vzip, vuzp): Delete.
87 * config/arm/neon.md (neon_vtrn<mode>): Delete.
88 (neon_vzip<mode>): Likewise.
89 (neon_vuzp<mode>): Likewise.
91 2014-05-13 Richard Biener <rguenther@suse.de>
93 PR ipa/60973
94 * tree-inline.c (remap_gimple_stmt): Clear tail call flag,
95 it needs revisiting whether the call still may be tail-called.
97 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
99 * rtl.def (SYMBOL_REF): Remove middle "0" field.
100 * rtl.h (block_symbol): Reduce number of fields to 2.
101 (rtx_def): Add u2.symbol_ref_flags.
102 (SYMBOL_REF_FLAGS): Use it.
103 (SYMBOL_REF_DATA, SET_SYMBOL_REF_DECL, SYMBOL_REF_DECL)
104 (SET_SYMBOL_REF_CONSTANT, SYMBOL_REF_CONSTANT): Lower index.
105 * gengtype.c (adjust_field_rtx_def): Remove SYMBOL_REF_FLAGS handling.
106 Lower index of SYMBOL_REF_DATA.
107 * print-rtl.c (print_rtx): Lower index for SYMBOL_REF_DATA.
108 Print SYMBOL_REF_FLAGS at the same time.
109 * genattrtab.c (attr_rtx_1): Only initialize 1 "0" SYMBOL_REF field.
111 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
113 * rtl.def (VAR_LOCATION): Remove "i" field.
114 * rtl.h (rtx_def): Add u2.var_location_status.
115 (PAT_VAR_LOCATION_STATUS): Use it.
116 (gen_rtx_VAR_LOCATION): Declare.
117 * gengenrtl.c (excluded_rtx): Add VAR_LOCATION.
118 * emit-rtl.c (gen_rtx_VAR_LOCATION): New function.
119 * var-tracking.c (emit_note_insn_var_location): Remove casts.
121 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
123 * rtl.def (scratch): Fix outdated comment and remove "0" field.
124 * gengtype.c (adjust_field_rtx_def): Update accordingly.
126 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
128 * rtl.def (DEBUG_INSN, INSN, JUMP_INSN, CALL_INSN, JUMP_TABLE_DATA)
129 (BARRIER, CODE_LABEL, NOTE): Remove first "i" field.
130 * rtl.h (rtx_def): Add insn_uid to u2 field.
131 (RTX_FLAG_CHECK8): Delete in favor of...
132 (RTL_INSN_CHAIN_FLAG_CHECK): ...this new macro.
133 (INSN_DELETED_P): Update accordingly.
134 (INSN_UID): Use u2.insn_uid.
135 (INSN_CHAIN_CODE_P): Define.
136 (PREV_INSN, NEXT_INSN, BLOCK_FOR_INSN, PATTERN, INSN_LOCATION)
137 (INSN_CODE, REG_NOTES, CALL_INSN_FUNCTION_USAGE, CODE_LABEL_NUMBER)
138 (NOTE_DATA, NOTE_DELETED_LABEL_NAME, NOTE_BLOCK, NOTE_EH_HANDLER)
139 (NOTE_BASIC_BLOCK, NOTE_VAR_LOCATION, NOTE_CFI, NOTE_LABEL_NUMBER)
140 (NOTE_KIND, LABEL_NAME, LABEL_NUSES, JUMP_LABEL, LABEL_REFS): Lower
141 indices accordingly.
142 * print-rtl.c (print_rtx): Print INSN_UIDs before the main loop.
143 Update indices for insn-chain rtxes.
144 * gengtype.c (gen_rtx_next): Adjust test for insn-chain rtxes.
145 (adjust_field_rtx_def): Lower '0' indices for all insn-chain rtxes.
146 * emit-rtl.c (gen_label_rtx): Update gen_rtx_LABEL call.
147 * caller-save.c (init_caller_save): Update gen_rtx_INSN calls.
148 * combine.c (try_combine): Likewise.
149 * ira.c (setup_prohibited_mode_move_regs): Likewise.
151 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
153 * rtl.def (REG): Remove middle field.
154 * rtl.h (rtx_def): Add orignal_regno to u2.
155 (ORIGINAL_REGNO): Use it instead of field 1.
156 (REG_ATTRS): Lower field index accordingly.
157 * gengtype.c (adjust_field_rtx_def): Remove handling of
158 ORIGINAL_REGNO. Move REG_ATTRS index down.
159 * print-rtl.c (print_rtx): Move ORIGINAL_REGNO handling to the
160 code that prints the REGNO.
162 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
164 * print-rtl.c (print_rtx): Guard whole '0' block with ifndef
165 GENERATOR_FILE.
167 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
169 * rtl.h (rtx_def): Mark u2 as GTY ((skip)).
171 2014-05-13 Bin Cheng <bin.cheng@arm.com>
173 * tree-ssa-loop-ivopts.c (contain_complex_addr_expr): New.
174 (alloc_iv): Lower base expressions containing ADDR_EXPR.
176 2014-05-13 Ian Bolton <ian.bolton@arm.com>
178 * config/aarch64/aarch64-protos.h
179 (aarch64_hard_regno_caller_save_mode): New prototype.
180 * config/aarch64/aarch64.c (aarch64_hard_regno_caller_save_mode):
181 New function.
182 * config/aarch64/aarch64.h (HARD_REGNO_CALLER_SAVE_MODE): New macro.
184 2014-05-13 Christian Bruel <christian.bruel@st.com>
186 * target.def (mode_switching): New hook vector.
187 (mode_emit, mode_needed, mode_after, mode_entry): New hooks.
188 (mode_exit, modepriority_to_mode): Likewise.
189 * mode-switching.c (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Hookify.
190 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
191 (default_priority_to_mode): Define.
192 * targhooks.h (default_priority_to_mode): Declare.
193 * target.h: Include tm.h and hard-reg-set.h.
194 * doc/tm.texi.in (EMIT_MODE_SET, MODE_NEEDED, MODE_AFTER, MODE_ENTRY)
195 (MODE_EXIT, MODE_PRIORITY_TO_MODE): Delete and hookify.
196 * doc/tm.texi Regenerate.
197 * config/sh/sh.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
198 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
199 * config/sh/sh.c (sh_emit_mode_set, sh_mode_priority): Hookify.
200 (sh_mode_needed, sh_mode_after, sh_mode_entry, sh_mode_exit): Likewise.
201 * config/i386/i386.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
202 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
203 * config/i386/i386-protos.h (ix86_mode_needed, ix86_mode_after)
204 (ix86_mode_entrym, ix86_emit_mode_set): Remove external declaration.
205 * config/i386/i386.c (ix86_mode_needed, ix86_mode_after, ix86_mode_exit,
206 (ix86_mode_entry, ix86_mode_priority, ix86_emit_mode_set): Hookify.
207 * config/epiphany/epiphany.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY):
208 Delete
209 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
210 * config/sh/sh.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
211 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
212 * config/sh/sh.c (sh4_emit_mode_set, sh4_mode_needed): Hookify.
213 (sh4_mode_after, sh4_mode_entry, sh4_mode_exit): Likewise.
214 * config/epiphany/epiphany-protos.h (epiphany_mode_needed)
215 (emit_set_fp_mode, epiphany_mode_entry_exit, epiphany_mode_after)
216 (epiphany_mode_priority_to_mode): Remove declaration.
217 * config/epiphany/epiphany.c (emit_set_fp_mode): Hookify.
218 (epiphany_mode_needed, epiphany_mode_priority_to_mode): Likewise.
219 (epiphany_mode_entry, epiphany_mode_exit, epiphany_mode_after):
220 Likewise.
221 (epiphany_mode_priority_to_mode): Change priority type. Hookify.
222 (epiphany_mode_needed, epiphany_mode_entry_exit): Hookify.
223 (epiphany_mode_after, epiphany_mode_entry, emit_set_fp_mode): Hookify.
225 2014-05-13 Jakub Jelinek <jakub@redhat.com>
227 PR target/61060
228 * config/i386/i386.c (ix86_expand_set_or_movmem): If count_exp
229 is const0_rtx, return immediately. Don't test count == 0 when
230 it is always true.
232 2014-05-13 Zhenqiang Chen <zhenqiang.chen@linaro.org>
234 * Makefile.in: add shrink-wrap.o.
235 * config/i386/i386.c: include "shrink-wrap.h"
236 * function.c: Likewise.
237 (requires_stack_frame_p, next_block_for_reg,
238 move_insn_for_shrink_wrap, prepare_shrink_wrap,
239 dup_block_and_redirect): Move to shrink-wrap.c
240 (thread_prologue_and_epilogue_insns): Extract three code segments
241 as functions in shrink-wrap.c
242 * function.h: Move #ifdef HAVE_simple_return ... #endif block to
243 shrink-wrap.h
244 * shrink-wrap.c: New file.
245 * shrink-wrap.h: New file.
247 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
249 * doc/extend.texi: Reflect current numbers of pragmas. Remove
250 reference to Solaris.
252 2014-05-12 Mike Stump <mikestump@comcast.net>
254 PR other/31778
255 * genattrtab.c (filename): Add.
256 (convert_set_attr_alternative): Improve error message.
257 (check_defs): Restore read_md_filename for error messages.
258 (gen_insn): Save filename.
260 2014-05-12 Dimitris Papavasiliou <dpapavas@gmail.com>
262 * doc/invoke.texi: Document new switches -Wno-shadow-ivar,
263 -fno-local-ivars and -fivar-visibility.
264 * c-family/c.opt: Make -Wshadow also implicitly enable
265 -Wshadow-ivar.
267 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
269 * doc/tm.texi: Remove reference to deleted macro.
270 * doc/tm.texi.in: Likewise.
272 2014-05-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
274 PR target/60991
275 * config/avr/avr.c (avr_out_store_psi): Use correct constant
276 to restore Y.
278 2014-05-12 Georg-Johann Lay <avr@gjlay.de>
280 * config/arm/arm.h (License): Add GCC Runtime Library Exception.
281 * config/arm/aout.h (License): Same.
282 * config/arm/bpabi.h (License): Same.
283 * config/arm/elf.h (License): Same.
284 * config/arm/linux-elf.h (License): Same.
285 * config/arm/linux-gas.h (License): Same.
286 * config/arm/netbsd-elf.h (License): Same.
287 * config/arm/uclinux-eabi.h (License): Same.
288 * config/arm/uclinux-elf.h (License): Same.
289 * config/arm/vxworks.h (License): Same.
291 2014-05-11 Jakub Jelinek <jakub@redhat.com>
293 * tree.h (OMP_CLAUSE_LINEAR_STMT): Define.
294 * tree.c (omp_clause_num_ops): Increase OMP_CLAUSE_LINEAR
295 number of operands to 3.
296 (walk_tree_1): Walk all operands of OMP_CLAUSE_LINEAR.
297 * tree-nested.c (convert_nonlocal_omp_clauses,
298 convert_local_omp_clauses): Handle OMP_CLAUSE_DEPEND.
299 * gimplify.c (gimplify_scan_omp_clauses): Handle
300 OMP_CLAUSE_LINEAR_STMT.
301 * omp-low.c (lower_rec_input_clauses): Fix typo.
302 (maybe_add_implicit_barrier_cancel, lower_omp_1): Add
303 cast between Fortran boolean_type_node and C _Bool if
304 needed.
306 2014-05-11 Richard Sandiford <rdsandiford@googlemail.com>
308 PR tree-optimization/61136
309 * wide-int.h (multiple_of_p): Define a version that doesn't return
310 the quotient.
311 * fold-const.c (extract_muldiv_1): Use wi::multiple_of_p instead of an
312 integer_zerop/const_binop pair.
313 (multiple_of_p): Likewise, converting both operands to widest_int
314 precision.
316 2014-05-09 Teresa Johnson <tejohnson@google.com>
318 * cgraphunit.c (analyze_functions): Use correct dump file.
320 2014-05-09 Florian Weimer <fweimer@redhat.com>
322 * cfgexpand.c (stack_protect_decl_p): New function, extracted from
323 expand_used_vars.
324 (stack_protect_return_slot_p): New function.
325 (expand_used_vars): Call stack_protect_decl_p and
326 stack_protect_return_slot_p for -fstack-protector-strong.
328 2014-05-09 Kenneth Zadeck <zadeck@naturalbridge.com>
330 PR middle-end/61111
331 * fold-const.c (fold_binary_loc): Changed width of mask.
333 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
335 * config/avr/avr-fixed.md (round<mode>3): Use -1U instead of -1 in
336 unsigned int initializers for regno_in, regno_out.
338 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
340 PR target/61055
341 * config/avr/avr.md (cc): Add new attribute set_vzn.
342 (addqi3, addqq3, adduqq3, subqi3, subqq3, subuqq3, negqi2) [cc]:
343 Set cc insn attribute to set_vzn instead of set_zn for alternatives
344 with INC, DEC or NEG.
345 * config/avr/avr.c (avr_notice_update_cc): Handle SET_VZN.
346 (avr_out_plus_1): ADIW sets cc0 to CC_SET_CZN.
347 INC, DEC and ADD+ADC set cc0 to CC_CLOBBER.
349 2014-05-09 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
351 Revert:
352 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
354 * wide-int.cc (UTItype): Define.
355 (UDWtype): Define for appropriate W_TYPE_SIZE.
357 2014-05-09 Richard Biener <rguenther@suse.de>
359 * Makefile.in (GTFILES): Remove tree-ssa-propagate.c.
360 * tree-ssa-propagate.c: Do not include gt-tree-ssa-propagate.h.
361 (interesting_ssa_edges, varying_ssa_edges): Move out of GC space.
362 (add_ssa_edge, process_ssa_edge_worklist, ssa_prop_init,
363 ssa_propagate): Adjust.
365 2014-05-08 Jeff Law <law@redhat.com>
367 PR tree-optimization/61009
368 * tree-ssa-threadedge.c (thread_through_normal_block): Return a
369 tri-state rather than a boolean. When a block is too big to
370 thread through, inform caller via negative return value.
371 (thread_across_edge): If a block was too big for normal threading,
372 then it's too big for a joiner too, so remove temporary equivalences
373 and return immediately.
375 2014-05-08 Manuel López-Ibáñez <manu@gcc.gnu.org>
376 Matthias Klose <doko@ubuntu.com>
378 PR driver/61106
379 * optc-gen.awk: Fix option handling for -Wunused-parameter.
381 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
383 PR target/59952
384 * config/i386/i386.c (PTA_HASWELL): Remove PTA_RTM.
386 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
388 PR target/61092
389 * config/alpha/alpha.c: Include gimple-iterator.h.
390 (alpha_gimple_fold_builtin): New function. Move
391 ALPHA_BUILTIN_UMULH folding from ...
392 (alpha_fold_builtin): ... here.
393 (TARGET_GIMPLE_FOLD_BUILTIN): New define.
395 2014-05-08 Wei Mi <wmi@google.com>
397 PR target/58066
398 * config/i386/i386.c (ix86_compute_frame_layout): Update
399 preferred_stack_boundary for call, expanded from tls descriptor.
400 * config/i386/i386.md (*tls_global_dynamic_32_gnu): Update RTX
401 to depend on SP register.
402 (*tls_local_dynamic_base_32_gnu): Ditto.
403 (*tls_local_dynamic_32_once): Ditto.
404 (tls_global_dynamic_64_<mode>): Set
405 ix86_tls_descriptor_calls_expanded_in_cfun.
406 (tls_local_dynamic_base_64_<mode>): Ditto.
407 (tls_global_dynamic_32): Set
408 ix86_tls_descriptor_calls_expanded_in_cfun. Update RTX
409 to depend on SP register.
410 (tls_local_dynamic_base_32): Ditto.
412 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
414 * config/arm/arm_neon.h: Update comment.
415 * config/arm/neon-docgen.ml: Delete.
416 * config/arm/neon-gen.ml: Delete.
417 * doc/arm-neon-intrinsics.texi: Update comment.
419 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
421 * config/arm/arm_neon_builtins.def (vadd, vsub): Only define the v2sf
422 and v4sf versions.
423 (vand, vorr, veor, vorn, vbic): Remove.
424 * config/arm/neon.md (neon_vadd, neon_vsub, neon_vadd_unspec): Adjust
425 iterator.
426 (neon_vsub_unspec): Likewise.
427 (neon_vorr, neon_vand, neon_vbic, neon_veor, neon_vorn): Remove.
429 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
431 * config/arm/arm_neon.h (vadd_s8): GNU C implementation
432 (vadd_s16): Likewise.
433 (vadd_s32): Likewise.
434 (vadd_f32): Likewise.
435 (vadd_u8): Likewise.
436 (vadd_u16): Likewise.
437 (vadd_u32): Likewise.
438 (vadd_s64): Likewise.
439 (vadd_u64): Likewise.
440 (vaddq_s8): Likewise.
441 (vaddq_s16): Likewise.
442 (vaddq_s32): Likewise.
443 (vaddq_s64): Likewise.
444 (vaddq_f32): Likewise.
445 (vaddq_u8): Likewise.
446 (vaddq_u16): Likewise.
447 (vaddq_u32): Likewise.
448 (vaddq_u64): Likewise.
449 (vmul_s8): Likewise.
450 (vmul_s16): Likewise.
451 (vmul_s32): Likewise.
452 (vmul_f32): Likewise.
453 (vmul_u8): Likewise.
454 (vmul_u16): Likewise.
455 (vmul_u32): Likewise.
456 (vmul_p8): Likewise.
457 (vmulq_s8): Likewise.
458 (vmulq_s16): Likewise.
459 (vmulq_s32): Likewise.
460 (vmulq_f32): Likewise.
461 (vmulq_u8): Likewise.
462 (vmulq_u16): Likewise.
463 (vmulq_u32): Likewise.
464 (vsub_s8): Likewise.
465 (vsub_s16): Likewise.
466 (vsub_s32): Likewise.
467 (vsub_f32): Likewise.
468 (vsub_u8): Likewise.
469 (vsub_u16): Likewise.
470 (vsub_u32): Likewise.
471 (vsub_s64): Likewise.
472 (vsub_u64): Likewise.
473 (vsubq_s8): Likewise.
474 (vsubq_s16): Likewise.
475 (vsubq_s32): Likewise.
476 (vsubq_s64): Likewise.
477 (vsubq_f32): Likewise.
478 (vsubq_u8): Likewise.
479 (vsubq_u16): Likewise.
480 (vsubq_u32): Likewise.
481 (vsubq_u64): Likewise.
482 (vand_s8): Likewise.
483 (vand_s16): Likewise.
484 (vand_s32): Likewise.
485 (vand_u8): Likewise.
486 (vand_u16): Likewise.
487 (vand_u32): Likewise.
488 (vand_s64): Likewise.
489 (vand_u64): Likewise.
490 (vandq_s8): Likewise.
491 (vandq_s16): Likewise.
492 (vandq_s32): Likewise.
493 (vandq_s64): Likewise.
494 (vandq_u8): Likewise.
495 (vandq_u16): Likewise.
496 (vandq_u32): Likewise.
497 (vandq_u64): Likewise.
498 (vorr_s8): Likewise.
499 (vorr_s16): Likewise.
500 (vorr_s32): Likewise.
501 (vorr_u8): Likewise.
502 (vorr_u16): Likewise.
503 (vorr_u32): Likewise.
504 (vorr_s64): Likewise.
505 (vorr_u64): Likewise.
506 (vorrq_s8): Likewise.
507 (vorrq_s16): Likewise.
508 (vorrq_s32): Likewise.
509 (vorrq_s64): Likewise.
510 (vorrq_u8): Likewise.
511 (vorrq_u16): Likewise.
512 (vorrq_u32): Likewise.
513 (vorrq_u64): Likewise.
514 (veor_s8): Likewise.
515 (veor_s16): Likewise.
516 (veor_s32): Likewise.
517 (veor_u8): Likewise.
518 (veor_u16): Likewise.
519 (veor_u32): Likewise.
520 (veor_s64): Likewise.
521 (veor_u64): Likewise.
522 (veorq_s8): Likewise.
523 (veorq_s16): Likewise.
524 (veorq_s32): Likewise.
525 (veorq_s64): Likewise.
526 (veorq_u8): Likewise.
527 (veorq_u16): Likewise.
528 (veorq_u32): Likewise.
529 (veorq_u64): Likewise.
530 (vbic_s8): Likewise.
531 (vbic_s16): Likewise.
532 (vbic_s32): Likewise.
533 (vbic_u8): Likewise.
534 (vbic_u16): Likewise.
535 (vbic_u32): Likewise.
536 (vbic_s64): Likewise.
537 (vbic_u64): Likewise.
538 (vbicq_s8): Likewise.
539 (vbicq_s16): Likewise.
540 (vbicq_s32): Likewise.
541 (vbicq_s64): Likewise.
542 (vbicq_u8): Likewise.
543 (vbicq_u16): Likewise.
544 (vbicq_u32): Likewise.
545 (vbicq_u64): Likewise.
546 (vorn_s8): Likewise.
547 (vorn_s16): Likewise.
548 (vorn_s32): Likewise.
549 (vorn_u8): Likewise.
550 (vorn_u16): Likewise.
551 (vorn_u32): Likewise.
552 (vorn_s64): Likewise.
553 (vorn_u64): Likewise.
554 (vornq_s8): Likewise.
555 (vornq_s16): Likewise.
556 (vornq_s32): Likewise.
557 (vornq_s64): Likewise.
558 (vornq_u8): Likewise.
559 (vornq_u16): Likewise.
560 (vornq_u32): Likewise.
561 (vornq_u64): Likewise.
563 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
565 * wide-int.cc (UTItype): Define.
566 (UDWtype): Define for appropriate W_TYPE_SIZE.
568 2014-05-08 Marc Glisse <marc.glisse@inria.fr>
570 PR tree-optimization/59100
571 * tree-ssa-phiopt.c: Include tree-inline.h.
572 (neutral_element_p, absorbing_element_p): New functions.
573 (value_replacement): Handle conditional binary operations with a
574 neutral or absorbing element.
576 2014-05-08 Richard Biener <rguenther@suse.de>
578 * tree-ssa-sccvn.c (vn_get_expr_for): Valueize operands before
579 folding the expression.
580 (valueize_expr): Remove.
581 (visit_reference_op_load): Do not valueize the result of
582 vn_get_expr_for.
583 (simplify_binary_expression): Likewise.
584 (simplify_unary_expression): Likewise.
586 2014-05-08 Richard Biener <rguenther@suse.de>
588 * gimplify.c (gimplify_call_expr): Use saved fnptrtype for
589 looking at TYPE_ARG_TYPES.
591 2014-05-08 Richard Biener <rguenther@suse.de>
593 * gimple-fold.c (gimple_fold_stmt_to_constant_1): Remove
594 pointer propagation special-case.
596 2014-05-08 Bin Cheng <bin.cheng@arm.com>
598 * tree-affine.c (tree_to_aff_combination): Handle MEM_REF for
599 core part of address expressions.
601 2014-05-08 Alan Modra <amodra@gmail.com>
603 PR target/60737
604 * config/rs6000/rs6000.c (expand_block_move): Allow 64-bit
605 loads and stores when -mno-strict-align at any alignment.
606 (expand_block_clear): Similarly. Also correct calculation of
607 instruction count.
609 2014-05-07 Thomas Preud'homme <thomas.preudhomme@arm.com>
611 PR middle-end/39246
612 * tree-complex.c (expand_complex_move): Keep line info when expanding
613 complex move.
614 * tree-ssa-uninit.c (warn_uninit): New argument. Ignore assignment
615 of complex expression. Use new argument to display correct location
616 for values coming from phi statement.
617 (warn_uninitialized_vars): Adapt to new signature of warn_uninit.
618 (warn_uninitialized_phi): Pass location of phi argument to
619 warn_uninit.
620 * tree-ssa.c (ssa_undefined_value_p): For SSA_NAME initialized by a
621 COMPLEX_EXPR, recurse on each part of the COMPLEX_EXPR.
623 2014-05-07 Segher Boessenkool <segher@kernel.crashing.org>
625 * config/rs6000/predicates.md (indexed_address_mem): New.
626 * config/rs6000/rs6000.md (type): Remove load_ext, load_ext_u,
627 load_ext_ux, load_ux, load_u, store_ux, store_u, fpload_ux, fpload_u,
628 fpstore_ux, fpstore_u.
629 (sign_extend, indexed, update): New.
630 (cell_micro): Adjust.
631 (*zero_extend<mode>di2_internal1, *zero_extendsidi2_lfiwzx,
632 *extendsidi2_lfiwax, *extendsidi2_nocell, *extendsfdf2_fpr,
633 *movsi_internal1, *movsi_internal1_single, *movhi_internal,
634 *movqi_internal, *movcc_internal1, mov<mode>_hardfloat,
635 *mov<mode>_softfloat, *mov<mode>_hardfloat32, *mov<mode>_hardfloat64,
636 *mov<mode>_softfloat64, *movdi_internal32, *movdi_internal64,
637 *mov<mode>_string, *ldmsi8, *ldmsi7, *ldmsi6, *ldmsi5, *ldmsi4,
638 *ldmsi3, *stmsi8, *stmsi7, *stmsi6, *stmsi5, *stmsi4, *stmsi3,
639 *movdi_update1, movdi_<mode>_update, movdi_<mode>_update_stack,
640 *movsi_update1, *movsi_update2, movsi_update, movsi_update_stack,
641 *movhi_update1, *movhi_update2, *movhi_update3, *movhi_update4,
642 *movqi_update1, *movqi_update2, *movqi_update3, *movsf_update1,
643 *movsf_update2, *movsf_update3, *movsf_update4, *movdf_update1,
644 *movdf_update2, load_toc_aix_si, load_toc_aix_di, probe_stack_<mode>,
645 *stmw, *lmw, as well as 10 anonymous patterns): Adjust.
647 * config/rs6000/dfp.md (movsd_store, movsd_load): Adjust.
648 * config/rs6000/vsx.md (*vsx_movti_32bit, *vsx_extract_<mode>_load,
649 *vsx_extract_<mode>_store): Adjust.
650 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_microcoded_insn,
651 is_cracked_insn, insn_must_be_first_in_group,
652 insn_must_be_last_in_group): Adjust.
654 * config/rs6000/40x.md (ppc403-load, ppc403-store, ppc405-float):
655 Adjust.
656 * config/rs6000/440.md (ppc440-load, ppc440-store, ppc440-fpload,
657 ppc440-fpstore): Adjust.
658 * config/rs6000/476.md (ppc476-load, ppc476-store, ppc476-fpload,
659 ppc476-fpstore): Adjust.
660 * config/rs6000/601.md (ppc601-load, ppc601-store, ppc601-fpload,
661 ppc601-fpstore): Adjust.
662 * config/rs6000/603.md (ppc603-load, ppc603-store, ppc603-fpload):
663 Adjust.
664 * config/rs6000/6xx.md (ppc604-load, ppc604-store, ppc604-fpload):
665 Adjust.
666 * config/rs6000/7450.md (ppc7450-load, ppc7450-store, ppc7450-fpload,
667 ppc7450-fpstore): Adjust.
668 * config/rs6000/7xx.md (ppc750-load, ppc750-store): Adjust.
669 * config/rs6000/8540.md (ppc8540_load, ppc8540_store): Adjust.
670 * config/rs6000/a2.md (ppca2-load, ppca2-fp-load, ppca2-fp-store):
671 Adjust.
672 * config/rs6000/cell.md (cell-load, cell-load-ux, cell-load-ext,
673 cell-fpload, cell-fpload-update, cell-store, cell-store-update,
674 cell-fpstore, cell-fpstore-update): Adjust.
675 * config/rs6000/e300c2c3.md (ppce300c3_load, ppce300c3_fpload,
676 ppce300c3_store, ppce300c3_fpstore): Adjust.
677 * config/rs6000/e500mc.md (e500mc_load, e500mc_fpload, e500mc_store,
678 e500mc_fpstore): Adjust.
679 * config/rs6000/e500mc64.md (e500mc64_load, e500mc64_fpload,
680 e500mc64_store, e500mc64_fpstore): Adjust.
681 * config/rs6000/e5500.md (e5500_load, e5500_fpload, e5500_store,
682 e5500_fpstore): Adjust.
683 * config/rs6000/e6500.md (e6500_load, e6500_fpload, e6500_store,
684 e6500_fpstore): Adjust.
685 * config/rs6000/mpc.md (mpccore-load, mpccore-store, mpccore-fpload):
686 Adjust.
687 * config/rs6000/power4.md (power4-load, power4-load-ext,
688 power4-load-ext-update, power4-load-ext-update-indexed,
689 power4-load-update-indexed, power4-load-update, power4-fpload,
690 power4-fpload-update, power4-store, power4-store-update,
691 power4-store-update-indexed, power4-fpstore, power4-fpstore-update):
692 Adjust.
693 * config/rs6000/power5.md (power5-load, power5-load-ext,
694 power5-load-ext-update, power5-load-ext-update-indexed,
695 power5-load-update-indexed, power5-load-update, power5-fpload,
696 power5-fpload-update, power5-store, power5-store-update,
697 power5-store-update-indexed, power5-fpstore, power5-fpstore-update):
698 Adjust.
699 * config/rs6000/power6.md (power6-load, power6-load-ext,
700 power6-load-update, power6-load-update-indexed,
701 power6-load-ext-update, power6-load-ext-update-indexed, power6-fpload,
702 power6-fpload-update, power6-store, power6-store-update,
703 power6-store-update-indexed, power6-fpstore, power6-fpstore-update):
704 Adjust.
705 * config/rs6000/power7.md (power7-load, power7-load-ext,
706 power7-load-update, power7-load-update-indexed,
707 power7-load-ext-update, power7-load-ext-update-indexed, power7-fpload,
708 power7-fpload-update, power7-store, power7-store-update,
709 power7-store-update-indexed, power7-fpstore, power7-fpstore-update):
710 Adjust.
711 * config/rs6000/power8.md (power8-load, power8-load-update,
712 power8-load-ext, power8-load-ext-update, power8-fpload,
713 power8-fpload-update, power8-store, power8-store-update-indexed,
714 power8-fpstore, power8-fpstore-update): Adjust.
715 * config/rs6000/rs64.md (rs64a-load, rs64a-store, rs64a-fpload):
716 Adjust.
717 * config/rs6000/titan.md (titan_lsu_load, titan_lsu_fpload,
718 titan_lsu_store, titan_lsu_fpstore): Adjust.
719 * config/rs6000/xfpu.md (fp-load, fp-store): Adjust.
721 2014-05-07 Oleg Endo <olegendo@gcc.gnu.org>
723 PR target/60884
724 * config/sh/sh-mem.cc (sh_expand_strlen): Use loop when emitting
725 unrolled byte insns. Emit address increments after move insns.
727 2014-05-07 David Malcolm <dmalcolm@redhat.com>
729 * gimple.h (gimple_builtin_call_types_compatible_p): Accept a
730 const_gimple, rather than a gimple.
731 (gimple_call_builtin_p): Likewise, for the three variants.
733 * gimple.c (gimple_builtin_call_types_compatible_p): Likewise.
734 (gimple_call_builtin_p): Likewise, for the three variants.
736 2014-05-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
738 PR tree-optimization/61095
739 * tree-ssanames.c (get_nonzero_bits): Fix type extension in wi::shwi.
741 2014-05-07 Richard Biener <rguenther@suse.de>
743 PR tree-optimization/61034
744 * tree-ssa-alias.c (call_may_clobber_ref_p_1): Export.
745 (maybe_skip_until): Use translate to take into account
746 lattices when trying to do disambiguations.
747 (get_continuation_for_phi_1): Likewise.
748 (get_continuation_for_phi): Adjust for added translate arguments.
749 (walk_non_aliased_vuses): Likewise.
750 * tree-ssa-alias.h (get_continuation_for_phi): Adjust prototype.
751 (walk_non_aliased_vuses): Likewise.
752 (call_may_clobber_ref_p_1): Declare.
753 * tree-ssa-sccvn.c (vn_reference_lookup_3): Also disambiguate against
754 calls. Stop early if we are only supposed to disambiguate.
755 * tree-ssa-pre.c (translate_vuse_through_block): Adjust.
757 2014-05-07 Joern Rennecke <joern.rennecke@embecosm.com>
759 * config/epiphany/epiphany.c (epiphany_handle_interrupt_attribute):
760 Emit an error when the function has arguments.
762 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
764 * cfgloop.h (unswitch_loops): Remove.
765 * doc/passes.texi: Remove references to loop-unswitch.c
766 * timevar.def (TV_LOOP_UNSWITCH): Remove.
768 2014-05-07 Evgeny Stupachenko <evstupac@gmail.com>
770 * tree-vect-data-refs.c (vect_grouped_load_supported): New
771 check for loads group of length 3.
772 (vect_permute_load_chain): New permutations for loads group of
773 length 3.
774 * tree-vect-stmts.c (vect_model_load_cost): Change cost
775 of vec_perm_shuffle for the new permutations.
777 2014-05-07 Alan Lawrence <alan.lawrence@arm.com>
779 * config/aarch64/arm_neon.h (vtrn1_f32, vtrn1_p8, vtrn1_p16, vtrn1_s8,
780 vtrn1_s16, vtrn1_s32, vtrn1_u8, vtrn1_u16, vtrn1_u32, vtrn1q_f32,
781 vtrn1q_f64, vtrn1q_p8, vtrn1q_p16, vtrn1q_s8, vtrn1q_s16, vtrn1q_s32,
782 vtrn1q_s64, vtrn1q_u8, vtrn1q_u16, vtrn1q_u32, vtrn1q_u64, vtrn2_f32,
783 vtrn2_p8, vtrn2_p16, vtrn2_s8, vtrn2_s16, vtrn2_s32, vtrn2_u8,
784 vtrn2_u16, vtrn2_u32, vtrn2q_f32, vtrn2q_f64, vtrn2q_p8, vtrn2q_p16,
785 vtrn2q_s8, vtrn2q_s16, vtrn2q_s32, vtrn2q_s64, vtrn2q_u8, vtrn2q_u16,
786 vtrn2q_u32, vtrn2q_u64): Replace temporary asm with __builtin_shuffle.
788 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
790 * loop-unswitch.c: Delete.
792 2014-05-07 Richard Biener <rguenther@suse.de>
794 * config.gcc: Always set need_64bit_hwint to yes.
796 2014-05-07 Chung-Ju Wu <jasonwucj@gmail.com>
798 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Have it in favor
799 of using optimize_size.
801 2014-05-06 Mike Stump <mikestump@comcast.net>
803 * wide-int.h (wi::int_traits <HOST_WIDE_INT>): Always define.
805 2014-05-06 Joseph Myers <joseph@codesourcery.com>
807 * config/i386/sse.md (*mov<mode>_internal)
808 (*<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>)
809 (*<sse2_avx_avx512f>_loaddqu<mode><mask_name>)
810 (<sse>_andnot<mode>3, <code><mode>3, *andnot<mode>3)
811 (*<code><mode>3, *andnot<mode>3<mask_name>)
812 (<mask_codefor><code><mode>3<mask_name>): Only consider
813 TARGET_SSE_PACKED_SINGLE_INSN_OPTIMAL for modes of size 16.
815 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
817 Revert:
818 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
820 * lra-constraints.c (valid_address_p): Move earlier in file.
821 Add a constraint argument to the address_info version.
822 (satisfies_memory_constraint_p): New function.
823 (satisfies_address_constraint_p): Likewise.
824 (process_alt_operands, curr_insn_transform): Use them.
825 (process_address): Pass the constraint to valid_address_p when
826 checking address operands.
828 2014-05-06 Richard Sandiford <r.sandiford@uk.ibm.com>
830 * lto-cgraph.c (compute_ltrans_boundary): Make node variables local
831 to their respective blocks. Fix inadvertent use of "node".
833 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
835 * emit-rtl.c (init_derived_machine_modes): New functionm, split
836 out from...
837 (init_emit_once): ...here.
838 * rtl.h (init_derived_machine_modes): Declare.
839 * toplev.c (do_compile): Call it even if no_backend.
841 2014-05-06 Kenneth Zadeck <zadeck@naturalbridge.com>
842 Mike Stump <mikestump@comcast.net>
843 Richard Sandiford <rdsandiford@googlemail.com>
844 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
846 * alias.c (ao_ref_from_mem): Use wide-int interfaces.
847 (rtx_equal_for_memref_p): Update comment.
848 (adjust_offset_for_component_ref): Use wide-int interfaces.
849 * builtins.c (get_object_alignment_2): Likewise.
850 (c_readstr): Likewise.
851 (target_char_cast): Add comment.
852 (determine_block_size): Use wide-int interfaces.
853 (expand_builtin_signbit): Likewise.
854 (fold_builtin_int_roundingfn): Likewise.
855 (fold_builtin_bitop): Likewise.
856 (fold_builtin_bswap): Likewise.
857 (fold_builtin_logarithm): Use signop.
858 (fold_builtin_pow): Likewise.
859 (fold_builtin_memory_op): Use wide-int interfaces.
860 (fold_builtin_object_size): Likewise.
861 * cfgloop.c (alloc_loop): Initialize nb_iterations_upper_bound and
862 nb_iterations_estimate.
863 (record_niter_bound): Use wide-int interfaces.
864 (get_estimated_loop_iterations_int): Likewise.
865 (get_estimated_loop_iterations): Likewise.
866 (get_max_loop_iterations): Likewise.
867 * cfgloop.h: Include wide-int.h.
868 (struct nb_iter_bound): Change bound to widest_int.
869 (struct loop): Change nb_iterations_upper_bound and
870 nb_iterations_estimate to widest_int.
871 (record_niter_bound): Switch to use widest_int.
872 (get_estimated_loop_iterations): Likewise.
873 (get_max_loop_iterations): Likewise.
874 (gcov_type_to_double_int): Rename to gcov_type_to_wide_int and
875 update for wide-int.
876 * cgraph.c (cgraph_add_thunk): Use wide-int interfaces.
877 * combine.c (try_combine): Likewise.
878 (subst): Use CONST_SCALAR_INT_P rather than CONST_INT_P.
879 * config/aarch64/aarch64.c (aapcs_vfp_sub_candidate): Use wide-int
880 interfaces.
881 (aarch64_float_const_representable_p): Likewise.
882 * config/arc/arc.c: Include wide-int.h.
883 (arc_can_use_doloop_p): Use wide-int interfaces.
884 * config/arm/arm.c (aapcs_vfp_sub_candidate): Likewise.
885 (vfp3_const_double_index): Likewise.
886 * config/avr/avr.c (avr_out_round): Likewise.
887 (avr_fold_builtin): Likewise.
888 * config/bfin/bfin.c (bfin_local_alignment): Likewise.
889 (bfin_can_use_doloop_p): Likewise.
890 * config/darwin.c (darwin_mergeable_constant_section): Likewise.
891 (machopic_select_rtx_section): Update to handle CONST_WIDE_INT.
892 * config/i386/i386.c: Include wide-int.h.
893 (ix86_data_alignment): Use wide-int interfaces.
894 (ix86_local_alignment): Likewise.
895 (ix86_emit_swsqrtsf): Update real_from_integer.
896 * config/msp430/msp430.c (msp430_attr): Use wide-int interfaces.
897 * config/nds32/nds32.c (nds32_insert_attributes): Likewise.
898 * config/rs6000/predicates.md (any_operand): Add const_wide_int.
899 (zero_constant): Likewise.
900 (input_operand): Likewise.
901 (splat_input_operand): Likewise.
902 (non_logical_cint_operand): Change const_double to const_wide_int.
903 * config/rs6000/rs6000.c (num_insns_constant): Handle CONST_WIDE_INT.
904 (easy_altivec_constant): Remove comment.
905 (paired_expand_vector_init): Use CONSTANT_P.
906 (rs6000_legitimize_address): Handle CONST_WIDE_INT.
907 (rs6000_emit_move): Update checks.
908 (rs6000_aggregate_candidate): Use wide-int interfaces.
909 (rs6000_expand_ternop_builtin): Likewise.
910 (rs6000_output_move_128bit): Handle CONST_WIDE_INT.
911 (rs6000_assemble_integer): Likewise.
912 (rs6000_hash_constant): Likewise.
913 (output_toc): Likewise.
914 (rs6000_rtx_costs): Likewise.
915 (rs6000_emit_swrsqrt); Update call to real_from_integer.
916 * config/rs6000/rs6000-c.c: Include wide-int.h.
917 (altivec_resolve_overloaded_builtin): Use wide-int interfaces.
918 * config/rs6000/rs6000.h (TARGET_SUPPORTS_WIDE_INT): New.
919 * config/rs6000/rs6000.md: Use const_scalar_int_operand.
920 Handle CONST_WIDE_INT.
921 * config/sol2-c.c (solaris_pragma_align): Change low to unsigned HWI.
922 Use tree_fits_uhwi_p.
923 * config/sparc/sparc.c: Include wide-int.h.
924 (sparc_fold_builtin): Use wide-int interfaces.
925 * config/vax/vax.c: Include wide-int.h.
926 (vax_float_literal): Use real_from_integer.
927 * coretypes.h (struct hwivec_def): New.
928 (hwivec): New.
929 (const_hwivec): New.
930 * cse.c (hash_rtx_cb): Handle CONST_WIDE_INT.
931 (equiv_constant): Handle CONST_WIDE_INT.
932 * cselib.c (rtx_equal_for_cselib_1): Use CASE_CONST_UNIQUE.
933 (cselib_hash_rtx): Handle CONST_WIDE_INT.
934 * dbxout.c (stabstr_U): Use wide-int interfaces.
935 (dbxout_type): Update to use cst_fits_shwi_p.
936 * defaults.h (LOG2_BITS_PER_UNIT): Define.
937 (TARGET_SUPPORTS_WIDE_INT): Add default.
938 * dfp.c: Include wide-int.h.
939 (decimal_real_to_integer2): Use wide-int interfaces and rename to
940 decimal_real_to_integer.
941 * dfp.h (decimal_real_to_integer2): Return a wide_int and rename to
942 decimal_real_to_integer.
943 * doc/generic.texi (Constant expressions): Update for wide_int.
944 * doc/rtl.texi (const_double): Likewise.
945 (const_wide_int, CONST_WIDE_INT, CONST_WIDE_INT_VEC): New.
946 (CONST_WIDE_INT_NUNITS, CONST_WIDE_INT_ELT): New.
947 * doc/tm.texi.in (REAL_VALUE_TO_INT): Remove.
948 (REAL_VALUE_FROM_INT): Remove.
949 (TARGET_SUPPORTS_WIDE_INT): New.
950 * doc/tm.texi: Regenerate.
951 * dojump.c (prefer_and_bit_test): Use wide-int interfaces.
952 * double-int.h: Include wide-int.h.
953 (struct wi::int_traits): New.
954 * dwarf2out.c (get_full_len): New.
955 (dw_val_equal_p): Add case dw_val_class_wide_int.
956 (size_of_loc_descr): Likewise.
957 (output_loc_operands): Likewise.
958 (insert_double): Remove.
959 (insert_wide_int): New.
960 (add_AT_wide): New.
961 (print_die): Add case dw_val_class_wide_int.
962 (attr_checksum): Likewise.
963 (attr_checksum_ordered): Likewise.
964 (same_dw_val_p): Likewise.
965 (size_of_die): Likewise.
966 (value_format): Likewise.
967 (output_die): Likewise.
968 (double_int_type_size_in_bits): Rename to offset_int_type_size_in_bits.
969 Use wide-int.
970 (clz_loc_descriptor): Use wide-int interfaces.
971 (mem_loc_descriptor): Likewise. Handle CONST_WIDE_INT.
972 (loc_descriptor): Use wide-int interfaces. Handle CONST_WIDE_INT.
973 (round_up_to_align): Use wide-int interfaces.
974 (field_byte_offset): Likewise.
975 (insert_double): Rename to insert_wide_int. Use wide-int interfaces.
976 (add_const_value_attribute): Handle CONST_WIDE_INT. Update
977 CONST_DOUBLE handling. Use wide-int interfaces.
978 (add_bound_info): Use tree_fits_uhwi_p. Use wide-int interfaces.
979 (gen_enumeration_type_die): Use add_AT_wide.
980 (hash_loc_operands): Add case dw_val_class_wide_int.
981 (compare_loc_operands): Likewise.
982 * dwarf2out.h: Include wide-int.h.
983 (wide_int_ptr): New.
984 (enum dw_val_class): Add dw_val_class_wide_int.
985 (struct dw_val_struct): Add val_wide.
986 * emit-rtl.c (const_wide_int_htab): New.
987 (const_wide_int_htab_hash): New.
988 (const_wide_int_htab_eq): New.
989 (lookup_const_wide_int): New.
990 (const_double_htab_hash): Use wide-int interfaces.
991 (const_double_htab_eq): Likewise.
992 (rtx_to_double_int): Conditionally compile for wide-int.
993 (immed_double_int_const): Rename to immed_wide_int_const and
994 update for wide-int.
995 (immed_double_const): Conditionally compile for wide-int.
996 (init_emit_once): Use wide-int interfaces.
997 * explow.c (plus_constant): Likewise.
998 * expmed.c (mask_rtx): Move further up file. Use wide-int interfaces.
999 (lshift_value): Use wide-int interfaces.
1000 (expand_mult): Likewise.
1001 (choose_multiplier): Likewise.
1002 (expand_smod_pow2): Likewise.
1003 (make_tree): Likewise.
1004 * expr.c (convert_modes): Consolidate handling of constants.
1005 Use wide-int interfaces.
1006 (emit_group_load_1): Add note.
1007 (store_expr): Update comment.
1008 (get_inner_reference): Use wide-int interfaces.
1009 (expand_constructor): Update comment.
1010 (expand_expr_real_2): Use wide-int interfaces.
1011 (expand_expr_real_1): Likewise.
1012 (reduce_to_bit_field_precision): Likewise.
1013 (const_vector_from_tree): Likewise.
1014 * final.c: Include wide-int-print.h.
1015 (output_addr_const): Handle CONST_WIDE_INT. Use CONST_DOUBLE_AS_INT_P.
1016 * fixed-value.c: Include wide-int.h.
1017 (fixed_from_string): Use wide-int interfaces.
1018 (fixed_to_decimal): Likewise.
1019 (fixed_convert_from_real): Likewise.
1020 (real_convert_from_fixed): Likewise.
1021 * fold-const.h (mem_ref_offset): Return an offset_int.
1022 (div_if_zero_remainder): Remove code parameter.
1023 * fold-const.c (div_if_zero_remainder): Remove code parameter.
1024 Use wide-int interfaces.
1025 (may_negate_without_overflow_p): Use wide-int interfaces.
1026 (negate_expr_p): Likewise.
1027 (fold_negate_expr): Likewise.
1028 (int_const_binop_1): Likewise.
1029 (const_binop): Likewise.
1030 (fold_convert_const_int_from_int): Likewise.
1031 (fold_convert_const_int_from_real): Likewise.
1032 (fold_convert_const_int_from_fixed): Likewise.
1033 (fold_convert_const_fixed_from_int): Likewise.
1034 (all_ones_mask_p): Take an unsigned size. Use wide-int interfaces.
1035 (sign_bit_p): Use wide-int interfaces.
1036 (make_range_step): Likewise.
1037 (build_range_check): Likewise. Pass an integer of the correct type
1038 instead of using integer_one_node.
1039 (range_predecessor): Pass an integer of the correct type instead
1040 of using integer_one_node.
1041 (range_successor): Likewise.
1042 (merge_ranges): Likewise.
1043 (unextend): Use wide-int interfaces.
1044 (extract_muldiv_1): Likewise.
1045 (fold_div_compare): Likewise.
1046 (fold_single_bit_test): Likewise.
1047 (fold_sign_changed_comparison): Likewise.
1048 (try_move_mult_to_index): Update calls to div_if_zero_remainder.
1049 (fold_plusminus_mult_expr): Use wide-int interfaces.
1050 (native_encode_int): Likewise.
1051 (native_interpret_int): Likewise.
1052 (fold_unary_loc): Likewise.
1053 (pointer_may_wrap_p): Likewise.
1054 (size_low_cst): Likewise.
1055 (mask_with_tz): Likewise.
1056 (fold_binary_loc): Likewise.
1057 (fold_ternary_loc): Likewise.
1058 (multiple_of_p): Likewise.
1059 (tree_call_nonnegative_warnv_p): Update calls to
1060 tree_int_cst_min_precision and real_from_integer.
1061 (fold_negate_const): Use wide-int interfaces.
1062 (fold_abs_const): Likewise.
1063 (fold_relational_const): Use tree_int_cst_lt.
1064 (round_up_loc): Use wide-int interfaces.
1065 * genemit.c (gen_exp): Add CONST_WIDE_INT case.
1066 * gengenrtl.c (excluded_rtx): Add CONST_WIDE_INT case.
1067 * gengtype.c: Remove include of double-int.h.
1068 (do_typedef): Use wide-int interfaces.
1069 (open_base_files): Add wide-int.h.
1070 (main): Add offset_int and widest_int typedefs.
1071 * gengtype-lex.l: Handle "^".
1072 (CXX_KEYWORD): Add "static".
1073 * gengtype-parse.c (require3): New.
1074 (require_template_declaration): Handle constant template arguments
1075 and nested templates.
1076 * gengtype-state.c: Don't include "double-int.h".
1077 * genpreds.c (write_one_predicate_function): Update comment.
1078 (write_tm_constrs_h): Add check for hval and lval use in
1079 CONST_WIDE_INT.
1080 * genrecog.c (validate_pattern): Add CONST_WIDE_INT case.
1081 (add_to_sequence): Likewise.
1082 * gensupport.c (struct std_pred_table): Add const_scalar_int_operand
1083 and const_double_operand.
1084 * gimple.c (preprocess_case_label_vec_for_gimple): Use wide-int
1085 interfaces.
1086 * gimple-fold.c (get_base_constructor): Likewise.
1087 (fold_array_ctor_reference): Likewise.
1088 (fold_nonarray_ctor_reference): Likewise.
1089 (fold_const_aggregate_ref_1): Likewise.
1090 (gimple_val_nonnegative_real_p): Likewise.
1091 (gimple_fold_indirect_ref): Likewise.
1092 * gimple-pretty-print.c (dump_ssaname_info): Likewise.
1093 * gimple-ssa-strength-reduction.c: Include wide-int-print.h.
1094 (struct slsr_cand_d): Change index to be widest_int.
1095 (struct incr_info_d): Change incr to be widest_int.
1096 (alloc_cand_and_find_basis): Use wide-int interfaces.
1097 (slsr_process_phi): Likewise.
1098 (backtrace_base_for_ref): Likewise. Return a widest_int.
1099 (restructure_reference): Take a widest_int instead of a double_int.
1100 (slsr_process_ref): Use wide-int interfaces.
1101 (create_mul_ssa_cand): Likewise.
1102 (create_mul_imm_cand): Likewise.
1103 (create_add_ssa_cand): Likewise.
1104 (create_add_imm_cand): Take a widest_int instead of a double_int.
1105 (slsr_process_add): Use wide-int interfaces.
1106 (slsr_process_cast): Likewise.
1107 (slsr_process_copy): Likewise.
1108 (dump_candidate): Likewise.
1109 (dump_incr_vec): Likewise.
1110 (replace_ref): Likewise.
1111 (cand_increment): Likewise. Return a widest_int.
1112 (cand_abs_increment): Likewise.
1113 (replace_mult_candidate): Take a widest_int instead of a double_int.
1114 (replace_unconditional_candidate): Use wide-int interfaces.
1115 (incr_vec_index): Take a widest_int instead of a double_int.
1116 (create_add_on_incoming_edge): Likewise.
1117 (create_phi_basis): Use wide-int interfaces.
1118 (replace_conditional_candidate): Likewise.
1119 (record_increment): Take a widest_int instead of a double_int.
1120 (record_phi_increments): Use wide-int interfaces.
1121 (phi_incr_cost): Take a widest_int instead of a double_int.
1122 (lowest_cost_path): Likewise.
1123 (total_savings): Likewise.
1124 (analyze_increments): Use wide-int interfaces.
1125 (ncd_with_phi): Take a widest_int instead of a double_int.
1126 (ncd_of_cand_and_phis): Likewise.
1127 (nearest_common_dominator_for_cands): Likewise.
1128 (insert_initializers): Use wide-int interfaces.
1129 (all_phi_incrs_profitable): Likewise.
1130 (replace_one_candidate): Likewise.
1131 (replace_profitable_candidates): Likewise.
1132 * godump.c: Include wide-int-print.h.
1133 (go_output_typedef): Use wide-int interfaces.
1134 * graphite-clast-to-gimple.c (gmp_cst_to_tree): Likewise.
1135 * graphite-sese-to-poly.c (tree_int_to_gmp): Likewise.
1136 (build_loop_iteration_domains): Likewise.
1137 * hooks.h: Include wide-int.h rather than double-int.h.
1138 (hook_bool_dint_dint_uint_bool_true): Delete.
1139 (hook_bool_wint_wint_uint_bool_true): Declare.
1140 * hooks.c (hook_bool_dint_dint_uint_bool_true): Removed.
1141 (hook_bool_wint_wint_uint_bool_true): New.
1142 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): Use wide-int
1143 interfaces.
1144 (ubsan_expand_si_overflow_mul_check): Likewise.
1145 * ipa-devirt.c (get_polymorphic_call_info): Likewise.
1146 * ipa-prop.c (compute_complex_assign_jump_func): Likewise.
1147 (get_ancestor_addr_info): Likewise.
1148 (ipa_modify_call_arguments): Likewise.
1149 * loop-doloop.c (doloop_modify): Likewise.
1150 (doloop_optimize): Likewise.
1151 * loop-iv.c (iv_number_of_iterations): Likewise.
1152 * loop-unroll.c (decide_unroll_constant_iterations): Likewise.
1153 (unroll_loop_constant_iterations): Likewise.
1154 (decide_unroll_runtime_iterations): Likewise.
1155 (unroll_loop_runtime_iterations): Likewise.
1156 (decide_peel_simple): Likewise.
1157 (decide_unroll_stupid): Likewise.
1158 * lto-streamer-in.c (streamer_read_wi): Add.
1159 (input_cfg): Use wide-int interfaces.
1160 (lto_input_tree_1): Likewise.
1161 * lto-streamer-out.c (streamer_write_wi): Add.
1162 (hash_tree): Use wide-int interfaces.
1163 (output_cfg): Likewise.
1164 * Makefile.in (OBJS): Add wide-int.o and wide-int-print.o.
1165 (GTFILES): Add wide-int.h and signop.h.
1166 (TAGS): Look for .cc files too.
1167 * omp-low.c (scan_omp_1_op): Use wide-int interfaces.
1168 * optabs.c (expand_subword_shift): Likewise.
1169 (expand_doubleword_shift): Likewise.
1170 (expand_absneg_bit): Likewise.
1171 (expand_copysign_absneg): Likewise.
1172 (expand_copysign_bit): Likewise.
1173 * postreload.c (reload_cse_simplify_set): Likewise.
1174 * predict.c (predict_iv_comparison): Likewise.
1175 * pretty-print.h: Include wide-int-print.h.
1176 (pp_wide_int) New.
1177 * print-rtl.c (print_rtx): Add CONST_WIDE_INT case.
1178 * print-tree.c: Include wide-int-print.h.
1179 (print_node_brief): Use wide-int interfaces.
1180 (print_node): Likewise.
1181 * read-rtl.c (validate_const_wide_int): New.
1182 (read_rtx_code): Add CONST_WIDE_INT case.
1183 * real.c: Include wide-int.h.
1184 (real_to_integer2): Delete.
1185 (real_to_integer): New function, returning a wide_int.
1186 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
1187 (ten_to_ptwo): Update call to real_from_integer.
1188 (real_digit): Likewise.
1189 * real.h: Include signop.h, wide-int.h and insn-modes.h.
1190 (real_to_integer2, REAL_VALUE_FROM_INT, REAL_VALUE_FROM_UNSIGNED_INT)
1191 (REAL_VALUE_TO_INT): Delete.
1192 (real_to_integer): Declare a wide-int form.
1193 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
1194 * recog.c (const_int_operand): Improve comment.
1195 (const_scalar_int_operand): New.
1196 (const_double_operand): Add a separate definition for CONST_WIDE_INT.
1197 * rtlanal.c (commutative_operand_precedence): Handle CONST_WIDE_INT.
1198 (split_double): Likewise.
1199 * rtl.c (DEF_RTL_EXPR): Handle CONST_WIDE_INT.
1200 (rtx_size): Likewise.
1201 (rtx_alloc_stat_v): New.
1202 (rtx_alloc_stat): Now calls rtx_alloc_stat_v.
1203 (cwi_output_hex): New.
1204 (iterative_hash_rtx): Handle CONST_WIDE_INT.
1205 (cwi_check_failed_bounds): New.
1206 * rtl.def (CONST_WIDE_INT): New.
1207 * rtl.h: Include <utility> and wide-int.h.
1208 (struct hwivec_def): New.
1209 (CWI_GET_NUM_ELEM): New.
1210 (CWI_PUT_NUM_ELEM): New.
1211 (struct rtx_def): Add num_elem and hwiv.
1212 (CASE_CONST_SCALAR_INT): Modify for TARGET_SUPPORTS_WIDE_INT.
1213 (CASE_CONST_UNIQUE): Likewise.
1214 (CASE_CONST_ANY): Likewise.
1215 (CONST_SCALAR_INT_P): Likewise.
1216 (CONST_WIDE_INT_P): New.
1217 (CWI_ELT): New.
1218 (HWIVEC_CHECK): New.
1219 (cwi_check_failed_bounds): New.
1220 (CWI_ELT): New.
1221 (HWIVEC_CHECK): New.
1222 (CONST_WIDE_INT_VEC) New.
1223 (CONST_WIDE_INT_NUNITS) New.
1224 (CONST_WIDE_INT_ELT) New.
1225 (rtx_mode_t): New type.
1226 (wi::int_traits <rtx_mode_t>): New.
1227 (wi::shwi): New.
1228 (wi::min_value): New.
1229 (wi::max_value): New.
1230 (rtx_alloc_v) New.
1231 (const_wide_int_alloc): New.
1232 (immed_wide_int_const): New.
1233 * sched-vis.c (print_value): Handle CONST_WIDE_INT.
1234 * sel-sched-ir.c (lhs_and_rhs_separable_p): Update comment.
1235 * signop.h: New file.
1236 * simplify-rtx.c (mode_signbit_p): Handle CONST_WIDE_INT.
1237 (simplify_const_unary_operation): Use wide-int interfaces.
1238 (simplify_binary_operation_1): Likewise.
1239 (simplify_const_binary_operation): Likewise.
1240 (simplify_const_relational_operation): Likewise.
1241 (simplify_immed_subreg): Likewise.
1242 * stmt.c (expand_case): Likewise.
1243 * stor-layout.h (set_min_and_max_values_for_integral_type): Take a
1244 signop rather than a bool.
1245 * stor-layout.c (layout_type): Use wide-int interfaces.
1246 (initialize_sizetypes): Update calls to
1247 set_min_and_max_values_for_integral_type.
1248 (set_min_and_max_values_for_integral_type): Take a signop rather
1249 than a bool. Use wide-int interfaces.
1250 (fixup_signed_type): Update accordingly. Remove
1251 HOST_BITS_PER_DOUBLE_INT limit.
1252 (fixup_unsigned_type): Likewise.
1253 * system.h (STATIC_CONSTANT_P): New.
1254 (STATIC_ASSERT): New.
1255 * target.def (can_use_doloop_p): Take widest_ints rather than
1256 double_ints.
1257 * target.h: Include wide-int.h rather than double-int.h.
1258 * targhooks.h (can_use_doloop_if_innermost): Take widest_ints rather
1259 than double_ints.
1260 * targhooks.c (default_cxx_get_cookie_size): Use tree_int_cst_lt
1261 rather than INT_CST_LT_UNSIGNED.
1262 (can_use_doloop_if_innermost): Take widest_ints rather than
1263 double_ints.
1264 * tree-affine.c: Include wide-int-print.h.
1265 (double_int_ext_for_comb): Delete.
1266 (wide_int_ext_for_comb): New.
1267 (aff_combination_zero): Use wide-int interfaces.
1268 (aff_combination_const): Take a widest_int instead of a double_int.
1269 (aff_combination_elt): Use wide-int interfaces.
1270 (aff_combination_scale): Take a widest_int instead of a double_int.
1271 (aff_combination_add_elt): Likewise.
1272 (aff_combination_add_cst): Likewise.
1273 (aff_combination_add): Use wide-int interfaces.
1274 (aff_combination_convert): Likewise.
1275 (tree_to_aff_combination): Likewise.
1276 (add_elt_to_tree): Take a widest_int instead of a double_int.
1277 (aff_combination_to_tree): Use wide-int interfaces.
1278 (aff_combination_remove_elt): Likewise.
1279 (aff_combination_add_product): Take a widest_int instead of
1280 a double_int.
1281 (aff_combination_mult): Use wide-int interfaces.
1282 (aff_combination_expand): Likewise.
1283 (double_int_constant_multiple_p): Delete.
1284 (wide_int_constant_multiple_p): New.
1285 (aff_combination_constant_multiple_p): Take a widest_int pointer
1286 instead of a double_int pointer.
1287 (print_aff): Use wide-int interfaces.
1288 (get_inner_reference_aff): Take a widest_int pointer
1289 instead of a double_int pointer.
1290 (aff_comb_cannot_overlap_p): Take widest_ints instead of double_ints.
1291 * tree-affine.h: Include wide-int.h.
1292 (struct aff_comb_elt): Change type of coef to widest_int.
1293 (struct affine_tree_combination): Change type of offset to widest_int.
1294 (double_int_ext_for_comb): Delete.
1295 (wide_int_ext_for_comb): New.
1296 (aff_combination_const): Use widest_int instead of double_int.
1297 (aff_combination_scale): Likewise.
1298 (aff_combination_add_elt): Likewise.
1299 (aff_combination_constant_multiple_p): Likewise.
1300 (get_inner_reference_aff): Likewise.
1301 (aff_comb_cannot_overlap_p): Likewise.
1302 (aff_combination_zero_p): Use wide-int interfaces.
1303 * tree.c: Include tree.h.
1304 (init_ttree): Use make_int_cst.
1305 (tree_code_size): Removed code for INTEGER_CST case.
1306 (tree_size): Add INTEGER_CST case.
1307 (make_node_stat): Update comment.
1308 (get_int_cst_ext_nunits, build_new_int_cst, build_int_cstu): New.
1309 (build_int_cst_type): Use wide-int interfaces.
1310 (double_int_to_tree): Likewise.
1311 (double_int_fits_to_tree_p): Delete.
1312 (force_fit_type_double): Delete.
1313 (force_fit_type): New.
1314 (int_cst_hash_hash): Use wide-int interfaces.
1315 (int_cst_hash_eq): Likewise.
1316 (build_int_cst_wide): Delete.
1317 (wide_int_to_tree): New.
1318 (cache_integer_cst): Use wide-int interfaces.
1319 (build_low_bits_mask): Likewise.
1320 (cst_and_fits_in_hwi): Likewise.
1321 (real_value_from_int_cst): Likewise.
1322 (make_int_cst_stat): New.
1323 (integer_zerop): Use wide_int interfaces.
1324 (integer_onep): Likewise.
1325 (integer_all_onesp): Likewise.
1326 (integer_pow2p): Likewise.
1327 (integer_nonzerop): Likewise.
1328 (tree_log2): Likewise.
1329 (tree_floor_log2): Likewise.
1330 (tree_ctz): Likewise.
1331 (int_size_in_bytes): Likewise.
1332 (mem_ref_offset): Return an offset_int rather than a double_int.
1333 (build_type_attribute_qual_variant): Use wide_int interfaces.
1334 (type_hash_eq): Likewise
1335 (tree_int_cst_equal): Likewise.
1336 (tree_int_cst_lt): Delete.
1337 (tree_int_cst_compare): Likewise.
1338 (tree_fits_shwi_p): Use wide_int interfaces.
1339 (tree_fits_uhwi_p): Likewise.
1340 (tree_int_cst_sign_bit): Likewise.
1341 (tree_int_cst_sgn): Likewise.
1342 (tree_int_cst_min_precision): Take a signop rather than a bool.
1343 (simple_cst_equal): Use wide_int interfaces.
1344 (compare_tree_int): Likewise.
1345 (iterative_hash_expr): Likewise.
1346 (int_fits_type_p): Likewise. Use tree_int_cst_lt rather than
1347 INT_CST_LT.
1348 (get_type_static_bounds): Use wide_int interfaces.
1349 (tree_int_cst_elt_check_failed): New.
1350 (build_common_tree_nodes): Reordered to set prec before filling in
1351 value.
1352 (int_cst_value): Check cst_and_fits_in_hwi.
1353 (widest_int_cst_value): Use wide_int interfaces.
1354 (upper_bound_in_type): Likewise.
1355 (lower_bound_in_type): Likewise.
1356 (num_ending_zeros): Likewise.
1357 (drop_tree_overflow): Likewise.
1358 * tree-call-cdce.c (check_pow): Update call to real_from_integer.
1359 (gen_conditions_for_pow_cst_base): Likewise.
1360 * tree-cfg.c: Include wide-int.h and wide-int-print.h.
1361 (group_case_labels_stmt): Use wide-int interfaces.
1362 (verify_gimple_assign_binary): Likewise.
1363 (print_loop): Likewise.
1364 * tree-chrec.c (tree_fold_binomial): Likewise.
1365 * tree-core.h (struct tree_base): Add int_length.
1366 (struct tree_int_cst): Change rep of value.
1367 * tree-data-ref.c (dr_analyze_innermost): Use wide-int interfaces.
1368 (dr_may_alias_p): Likewise.
1369 (max_stmt_executions_tree): Likewise.
1370 * tree.def (INTEGER_CST): Update comment.
1371 * tree-dfa.c (get_ref_base_and_extent): Use wide-int interfaces.
1372 * tree-dfa.h (get_addr_base_and_unit_offset_1): Likewise.
1373 * tree-dump.c: Include wide-int.h and wide-int-print.h.
1374 (dequeue_and_dump): Use wide-int interfaces.
1375 * tree.h: Include wide-int.h.
1376 (NULL_TREE): Moved to earlier loc in file.
1377 (TREE_INT_CST_ELT_CHECK): New.
1378 (tree_int_cst_elt_check_failed): New.
1379 (TYPE_SIGN): New.
1380 (TREE_INT_CST): Delete.
1381 (TREE_INT_CST_LOW): Use wide-int interfaces.
1382 (TREE_INT_CST_HIGH): Delete.
1383 (TREE_INT_CST_NUNITS): New.
1384 (TREE_INT_CST_EXT_NUNITS): Likewise.
1385 (TREE_INT_CST_OFFSET_NUNITS): Likewise.
1386 (TREE_INT_CST_ELT): Likewise.
1387 (INT_CST_LT): Delete.
1388 (tree_int_cst_elt_check): New (two forms).
1389 (type_code_size): Update comment.
1390 (make_int_cst_stat, make_int_cst): New.
1391 (tree_to_double_int): Delete.
1392 (double_int_fits_to_tree_p): Delete.
1393 (force_fit_type_double): Delete.
1394 (build_int_cstu): Replace with out-of-line function.
1395 (build_int_cst_wide): Delete.
1396 (tree_int_cst_lt): Define inline.
1397 (tree_int_cst_le): New.
1398 (tree_int_cst_compare): Define inline.
1399 (tree_int_cst_min_precision): Take a signop rather than a bool.
1400 (wi::int_traits <const_tree>): New.
1401 (wi::int_traits <tree>): New.
1402 (wi::extended_tree): New.
1403 (wi::int_traits <wi::extended_tree>): New.
1404 (wi::to_widest): New.
1405 (wi::to_offset): New.
1406 (wi::fits_to_tree_p): New.
1407 (wi::min_value): New.
1408 (wi::max_value): New.
1409 * tree-inline.c (remap_gimple_op_r): Use wide-int interfaces.
1410 (copy_tree_body_r): Likewise.
1411 * tree-object-size.c (compute_object_offset): Likewise.
1412 (addr_object_size): Likewise.
1413 * tree-predcom.c: Include wide-int-print.h.
1414 (struct dref_d): Change type of offset to widest_int.
1415 (dump_dref): Call wide-int printer.
1416 (aff_combination_dr_offset): Use wide-int interfaces.
1417 (determine_offset): Take a widest_int pointer rather than a
1418 double_int pointer.
1419 (split_data_refs_to_components): Use wide-int interfaces.
1420 (suitable_component_p): Likewise.
1421 (order_drefs): Likewise.
1422 (add_ref_to_chain): Likewise.
1423 (valid_initializer_p): Likewise.
1424 (determine_roots_comp): Likewise.
1425 * tree-pretty-print.c: Include wide-int-print.h.
1426 (dump_generic_node): Use wide-int interfaces.
1427 * tree-sra.c (sra_ipa_modify_expr): Likewise.
1428 * tree-ssa-address.c (addr_for_mem_ref): Likewise.
1429 (move_fixed_address_to_symbol): Likewise.
1430 (move_hint_to_base): Likewise.
1431 (move_pointer_to_base): Likewise.
1432 (move_variant_to_index): Likewise.
1433 (most_expensive_mult_to_index): Likewise.
1434 (addr_to_parts): Likewise.
1435 (copy_ref_info): Likewise.
1436 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
1437 (indirect_refs_may_alias_p): Likewise.
1438 (stmt_kills_ref_p_1): Likewise.
1439 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
1440 * tree-ssa-ccp.c: Update comment at top of file. Include
1441 wide-int-print.h.
1442 (struct prop_value_d): Change type of mask to widest_int.
1443 (extend_mask): New function.
1444 (dump_lattice_value): Use wide-int interfaces.
1445 (get_default_value): Likewise.
1446 (set_constant_value): Likewise.
1447 (set_value_varying): Likewise.
1448 (valid_lattice_transition): Likewise.
1449 (set_lattice_value): Likewise.
1450 (value_to_double_int): Delete.
1451 (value_to_wide_int): New.
1452 (get_value_from_alignment): Use wide-int interfaces.
1453 (get_value_for_expr): Likewise.
1454 (do_dbg_cnt): Likewise.
1455 (ccp_finalize): Likewise.
1456 (ccp_lattice_meet): Likewise.
1457 (bit_value_unop_1): Use widest_ints rather than double_ints.
1458 (bit_value_binop_1): Likewise.
1459 (bit_value_unop): Use wide-int interfaces.
1460 (bit_value_binop): Likewise.
1461 (bit_value_assume_aligned): Likewise.
1462 (evaluate_stmt): Likewise.
1463 (ccp_fold_stmt): Likewise.
1464 (visit_cond_stmt): Likewise.
1465 (ccp_visit_stmt): Likewise.
1466 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Likewise.
1467 (constant_pointer_difference): Likewise.
1468 (associate_pointerplus): Likewise.
1469 (combine_conversions): Likewise.
1470 * tree-ssa-loop.h: Include wide-int.h.
1471 (struct tree_niter_desc): Change type of max to widest_int.
1472 * tree-ssa-loop-im.c (mem_refs_may_alias_p): Use wide-int interfaces.
1473 * tree-ssa-loop-ivcanon.c (remove_exits_and_undefined_stmts): Likewise.
1474 (remove_redundant_iv_tests): Likewise.
1475 (canonicalize_loop_induction_variables): Likewise.
1476 * tree-ssa-loop-ivopts.c (alloc_iv): Likewise.
1477 (constant_multiple_of): Take a widest_int pointer instead of
1478 a double_int pointer.
1479 (get_computation_aff): Use wide-int interfaces.
1480 (ptr_difference_cost): Likewise.
1481 (difference_cost): Likewise.
1482 (get_loop_invariant_expr_id): Likewise.
1483 (get_computation_cost_at): Likewise.
1484 (iv_elimination_compare_lt): Likewise.
1485 (may_eliminate_iv): Likewise.
1486 * tree-ssa-loop-niter.h (estimated_loop_iterations): Use widest_int
1487 instead of double_int.
1488 (max_loop_iterations): Likewise.
1489 (max_stmt_executions): Likewise.
1490 (estimated_stmt_executions): Likewise.
1491 * tree-ssa-loop-niter.c: Include wide-int-print.h.
1492 (split_to_var_and_offset): Use wide-int interfaces.
1493 (determine_value_range): Likewise.
1494 (bound_difference_of_offsetted_base): Likewise.
1495 (bounds_add): Take a widest_int instead of a double_int.
1496 (number_of_iterations_ne_max): Use wide-int interfaces.
1497 (number_of_iterations_ne): Likewise.
1498 (number_of_iterations_lt_to_ne): Likewise.
1499 (assert_loop_rolls_lt): Likewise.
1500 (number_of_iterations_lt): Likewise.
1501 (number_of_iterations_le): Likewise.
1502 (number_of_iterations_cond): Likewise.
1503 (number_of_iterations_exit): Likewise.
1504 (finite_loop_p): Likewise.
1505 (derive_constant_upper_bound_assign): Likewise.
1506 (derive_constant_upper_bound): Return a widest_int.
1507 (derive_constant_upper_bound_ops): Likewise.
1508 (do_warn_aggressive_loop_optimizations): Use wide-int interfaces.
1509 (record_estimate): Take a widest_int rather than a double_int.
1510 (record_nonwrapping_iv): Use wide-int interfaces.
1511 (double_int_cmp): Delete.
1512 (wide_int_cmp): New.
1513 (bound_index): Take a widest_int rather than a double_int.
1514 (discover_iteration_bound_by_body_walk): Use wide-int interfaces.
1515 (maybe_lower_iteration_bound): Likewise.
1516 (estimate_numbers_of_iterations_loop): Likewise.
1517 (estimated_loop_iterations): Take a widest_int pointer than than
1518 a double_int pointer.
1519 (estimated_loop_iterations_int): Use wide-int interfaces.
1520 (max_loop_iterations): Take a widest_int pointer than than
1521 a double_int pointer.
1522 (max_loop_iterations_int): Use wide-int interfaces.
1523 (max_stmt_executions): Take a widest_int pointer than than
1524 a double_int pointer.
1525 (estimated_stmt_executions): Likewise.
1526 (n_of_executions_at_most): Use wide-int interfaces.
1527 (scev_probably_wraps_p): Likewise.
1528 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Update calls
1529 to real_to_integer.
1530 * tree-scalar-evolution.c (simplify_peeled_chrec): Use wide-int
1531 interfaces.
1532 * tree-ssanames.c (set_range_info): Use wide_int_refs rather than
1533 double_ints. Adjust for trailing_wide_ints <3> representation.
1534 (set_nonzero_bits): Likewise.
1535 (get_range_info): Return wide_ints rather than double_ints.
1536 Adjust for trailing_wide_ints <3> representation.
1537 (get_nonzero_bits): Likewise.
1538 (duplicate_ssa_name_range_info): Adjust for trailing_wide_ints <3>
1539 representation.
1540 * tree-ssanames.h (struct range_info_def): Replace min, max and
1541 nonzero_bits with a trailing_wide_ints <3>.
1542 (set_range_info): Use wide_int_refs rather than double_ints.
1543 (set_nonzero_bits): Likewise.
1544 (get_range_info): Return wide_ints rather than double_ints.
1545 (get_nonzero_bits): Likewise.
1546 * tree-ssa-phiopt.c (jump_function_from_stmt): Use wide-int interfaces.
1547 * tree-ssa-pre.c (phi_translate_1): Likewise.
1548 * tree-ssa-reassoc.c (decrement_power): Use calls to real_from_integer.
1549 (acceptable_pow_call): Likewise.
1550 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Use wide-int
1551 interfaces.
1552 (vn_reference_fold_indirect): Likewise.
1553 (vn_reference_maybe_forwprop_address): Likewise.
1554 (valueize_refs_1): Likewise.
1555 * tree-ssa-structalias.c (get_constraint_for_ptr_offset): Likewise.
1556 * tree-ssa-uninit.c (is_value_included_in): Use wide-int interfaces,
1557 tree_int_cst_lt and tree_int_cst_le.
1558 * tree-streamer-in.c (unpack_ts_base_value_fields): Use wide-int
1559 interfaces.
1560 (streamer_alloc_tree): Likewise.
1561 * tree-streamer-out.c (pack_ts_int_cst_value_fields): Likewise.
1562 (streamer_write_tree_header): Likewise.
1563 (streamer_write_integer_cst): Likewise.
1564 * tree-switch-conversion.c (emit_case_bit_tests): Likewise.
1565 (build_constructors): Likewise.
1566 (array_value_type): Likewise.
1567 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list): Likewise.
1568 (vect_check_gather): Likewise.
1569 * tree-vect-generic.c (build_replicated_const): Likewise.
1570 (expand_vector_divmod): Likewise.
1571 * tree-vect-loop.c (vect_transform_loop): Likewise.
1572 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
1573 (vect_do_peeling_for_alignment): Likewise.
1574 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
1575 * tree-vrp.c: Include wide-int.h.
1576 (operand_less_p): Use wide-int interfaces and tree_int_cst_lt.
1577 (extract_range_from_assert): Use wide-int interfaces.
1578 (vrp_int_const_binop): Likewise.
1579 (zero_nonzero_bits_from_vr): Take wide_int pointers rather than
1580 double_int pointers.
1581 (ranges_from_anti_range): Use wide-int interfaces.
1582 (quad_int_cmp): Delete.
1583 (quad_int_pair_sort): Likewise.
1584 (extract_range_from_binary_expr_1): Use wide-int interfaces.
1585 (extract_range_from_unary_expr_1): Likewise.
1586 (adjust_range_with_scev): Likewise.
1587 (masked_increment): Take and return wide_ints rather than double_ints.
1588 (register_edge_assert_for_2): Use wide-int interfaces.
1589 (check_array_ref): Likewise.
1590 (search_for_addr_array): Likewise.
1591 (maybe_set_nonzero_bits): Likewise.
1592 (union_ranges): Pass an integer of the correct type instead of
1593 using integer_one_node.
1594 (intersect_ranges): Likewise.
1595 (simplify_truth_ops_using_ranges): Likewise.
1596 (simplify_bit_ops_using_ranges): Use wide-int interfaces.
1597 (range_fits_type_p): Likewise.
1598 (simplify_cond_using_ranges): Likewise. Take a signop rather than
1599 a bool.
1600 (simplify_conversion_using_ranges): Use wide-int interfaces.
1601 (simplify_float_conversion_using_ranges): Likewise.
1602 (vrp_finalize): Likewise.
1603 * value-prof.c (gimple_divmod_fixed_value_transform): Likewise.
1604 (gimple_stringops_transform): Likewise.
1605 * varasm.c (decode_addr_const): Likewise.
1606 (const_hash_1): Likewise.
1607 (const_rtx_hash_1): Likewise
1608 (output_constant): Likewise.
1609 (array_size_for_constructor): Likewise.
1610 (output_constructor_regular_field): Likewise.
1611 (output_constructor_bitfield): Likewise.
1612 * var-tracking.c (loc_cmp): Handle CONST_WIDE_INT.
1613 * mkconfig.sh: Include machmode.h to pick up BITS_PER_UNIT for
1614 GENERATOR_FILEs.
1615 * gencheck.c: Define BITS_PER_UNIT.
1616 * wide-int.cc: New.
1617 * wide-int.h: New.
1618 * wide-int-print.cc: New.
1619 * wide-int-print.h: New.
1621 2014-05-06 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1623 * config/avr/avr.c (avr_can_eliminate): Mark unused argument.
1625 2014-05-06 Richard Biener <rguenther@suse.de>
1627 * tree-pass.h (TODO_verify_ssa, TODO_verify_flow,
1628 TODO_verify_stmts, TODO_verify_rtl_sharing): Remove.
1629 (TODO_verify_all): Adjust.
1630 * asan.c: Remove references to TODO_verify_ssa, TODO_verify_flow,
1631 TODO_verify_stmts and TODO_verify_rtl_sharing.
1632 * bb-reorder.c: Likewise.
1633 * cfgexpand.c: Likewise.
1634 * cprop.c: Likewise.
1635 * cse.c: Likewise.
1636 * function.c: Likewise.
1637 * fwprop.c: Likewise.
1638 * gcse.c: Likewise.
1639 * gimple-ssa-isolate-paths.c: Likewise.
1640 * gimple-ssa-strength-reduction.c: Likewise.
1641 * ipa-split.c: Likewise.
1642 * loop-init.c: Likewise.
1643 * loop-unroll.c: Likewise.
1644 * lower-subreg.c: Likewise.
1645 * modulo-sched.c: Likewise.
1646 * postreload-gcse.c: Likewise.
1647 * predict.c: Likewise.
1648 * recog.c: Likewise.
1649 * sched-rgn.c: Likewise.
1650 * store-motion.c: Likewise.
1651 * tracer.c: Likewise.
1652 * trans-mem.c: Likewise.
1653 * tree-call-cdce.c: Likewise.
1654 * tree-cfg.c: Likewise.
1655 * tree-cfgcleanup.c: Likewise.
1656 * tree-complex.c: Likewise.
1657 * tree-eh.c: Likewise.
1658 * tree-emutls.c: Likewise.
1659 * tree-if-conv.c: Likewise.
1660 * tree-into-ssa.c: Likewise.
1661 * tree-loop-distribution.c: Likewise.
1662 * tree-object-size.c: Likewise.
1663 * tree-parloops.c: Likewise.
1664 * tree-pass.h: Likewise.
1665 * tree-sra.c: Likewise.
1666 * tree-ssa-ccp.c: Likewise.
1667 * tree-ssa-copy.c: Likewise.
1668 * tree-ssa-copyrename.c: Likewise.
1669 * tree-ssa-dce.c: Likewise.
1670 * tree-ssa-dom.c: Likewise.
1671 * tree-ssa-dse.c: Likewise.
1672 * tree-ssa-forwprop.c: Likewise.
1673 * tree-ssa-ifcombine.c: Likewise.
1674 * tree-ssa-loop-ch.c: Likewise.
1675 * tree-ssa-loop-ivcanon.c: Likewise.
1676 * tree-ssa-loop.c: Likewise.
1677 * tree-ssa-math-opts.c: Likewise.
1678 * tree-ssa-phiopt.c: Likewise.
1679 * tree-ssa-phiprop.c: Likewise.
1680 * tree-ssa-pre.c: Likewise.
1681 * tree-ssa-reassoc.c: Likewise.
1682 * tree-ssa-sink.c: Likewise.
1683 * tree-ssa-strlen.c: Likewise.
1684 * tree-ssa-tail-merge.c: Likewise.
1685 * tree-ssa-uncprop.c: Likewise.
1686 * tree-switch-conversion.c: Likewise.
1687 * tree-tailcall.c: Likewise.
1688 * tree-vect-generic.c: Likewise.
1689 * tree-vectorizer.c: Likewise.
1690 * tree-vrp.c: Likewise.
1691 * tsan.c: Likewise.
1692 * var-tracking.c: Likewise.
1693 * bt-load.c: Likewise.
1694 * cfgcleanup.c: Likewise.
1695 * combine-stack-adj.c: Likewise.
1696 * combine.c: Likewise.
1697 * compare-elim.c: Likewise.
1698 * config/epiphany/resolve-sw-modes.c: Likewise.
1699 * config/i386/i386.c: Likewise.
1700 * config/mips/mips.c: Likewise.
1701 * config/s390/s390.c: Likewise.
1702 * config/sh/sh_treg_combine.cc: Likewise.
1703 * config/sparc/sparc.c: Likewise.
1704 * dce.c: Likewise.
1705 * dse.c: Likewise.
1706 * final.c: Likewise.
1707 * ifcvt.c: Likewise.
1708 * mode-switching.c: Likewise.
1709 * passes.c: Likewise.
1710 * postreload.c: Likewise.
1711 * ree.c: Likewise.
1712 * reg-stack.c: Likewise.
1713 * regcprop.c: Likewise.
1714 * regrename.c: Likewise.
1715 * web.c: Likewise.
1717 2014-05-06 Richard Biener <rguenther@suse.de>
1719 PR middle-end/61070
1720 * bitmap.c (debug_bitmap): Dump to stderr, not stdout.
1721 * tree-ssa-structalias.c (dump_solution_for_var): Likewise.
1723 2014-05-05 Jan Hubicka <hubicka@ucw.cz>
1725 PR ipa/60965
1726 * ipa-devirt.c (get_class_context): Allow POD to change to non-POD.
1728 2014-05-05 Radovan Obradovic <robradovic@mips.com>
1729 Tom de Vries <tom@codesourcery.com>
1731 * target.def (call_fusage_contains_non_callee_clobbers): New
1732 DEFHOOKPOD.
1733 * doc/tm.texi.in (@node Stack and Calling): Add Miscellaneous Register
1734 Hooks to @menu.
1735 (@node Miscellaneous Register Hooks): New node.
1736 (@hook TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): New hook.
1737 * doc/tm.texi: Regenerate.
1739 2014-05-05 Marek Polacek <polacek@redhat.com>
1741 PR driver/61065
1742 * opts.c (common_handle_option): Call error_at instead of warning_at.
1744 2014-05-05 Richard Biener <rguenther@suse.de>
1746 * passes.c (execute_function_todo): Don't reset TODO_verify_ssa
1747 from last_verified if update_ssa ran. Move TODO_verify_rtl_sharing
1748 under the TODO_verify_il umbrella.
1750 2014-05-05 Richard Biener <rguenther@suse.de>
1752 * passes.c (execute_function_todo): Move TODO_verify_flow under
1753 the TODO_verify_ul umbrella.
1755 2014-05-05 Richard Biener <rguenther@suse.de>
1757 PR middle-end/61010
1758 * fold-const.c (fold_binary_loc): Consistently avoid canonicalizing
1759 X & CST away from a CST that is the mask of a mode.
1761 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1763 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
1764 int argument to enum machine_mode.
1765 (picochip_class_max_nregs): Ditto.
1766 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
1767 (picochip_class_max_nregs): Ditto.
1769 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1771 * target.def: Add new target hook.
1772 * doc/tm.texi: Regenerate.
1773 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
1774 * targhooks.c (default_keep_leaf_when_profiled): New function.
1776 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
1777 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
1779 2014-05-05 Bin Cheng <bin.cheng@arm.com>
1781 PR tree-optimization/60363
1782 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
1783 (copy_phi_args): New parameters. Call get_value_locus_in_path.
1784 (update_destination_phis): New parameter.
1785 (create_edge_and_update_destination_phis): Ditto.
1786 (ssa_fix_duplicate_block_edges): Pass new arguments.
1787 (thread_single_edge): Ditto.
1789 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
1791 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
1792 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
1793 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
1794 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
1795 Use RS6000_BTM_HARD_FLOAT.
1796 (BU_MISC_2): Likewise.
1797 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
1798 RS6000_BTM_HARD_FLOAT.
1799 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
1800 is explicitly used.
1801 (rs6000_invalid_builtin): Add hard floating builtin support.
1802 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
1803 hard float builtins.
1804 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
1806 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1808 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
1809 Add missing function* argument.
1811 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
1813 * lra-constraints.c (valid_address_p): Move earlier in file.
1814 Add a constraint argument to the address_info version.
1815 (satisfies_memory_constraint_p): New function.
1816 (satisfies_address_constraint_p): Likewise.
1817 (process_alt_operands, curr_insn_transform): Use them.
1818 (process_address): Pass the constraint to valid_address_p when
1819 checking address operands.
1821 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
1823 * config/mips/mips.c (mips_isa_rev): New variable.
1824 (mips_set_architecture): Set it.
1825 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
1826 from mips_isa_rev.
1827 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
1828 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
1829 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
1830 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
1831 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
1832 conditions in terms of mips_isa_rev.
1833 (mips_isa_rev): Declare.
1835 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1837 * config/sh/sh-mem.cc: Use tabs instead of spaces.
1838 (prob_unlikely, prob_likely): Make variables const.
1840 2014-05-03 Denis Chertykov <chertykov@gmail.com>
1842 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
1844 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1846 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
1848 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1850 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
1851 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
1852 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
1853 functions.
1854 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
1855 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
1856 sh_pass_in_reg_p.
1857 Replace usage of ROUND_REG with sh_round_reg.
1858 Use CEIL instead of ROUND_ADVANCE.
1860 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1862 PR target/61026
1863 * config/sh/sh.c: Include stdlib headers before everything else.
1865 2014-05-02 Jakub Jelinek <jakub@redhat.com>
1867 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
1868 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
1869 (gimplify_adjust_omp_clauses): Simd region is never
1870 directly nested in combined parallel. Instead, for linear
1871 with copyin/copyout, if in combined for simd loop, make decl
1872 firstprivate/lastprivate on OMP_FOR.
1873 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
1874 expand_omp_for_static_chunk): When setting endvar, also set
1875 fd->loop.v to the same value.
1877 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
1879 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
1881 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
1883 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
1884 expression.
1886 2014-05-02 Marek Polacek <polacek@redhat.com>
1888 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
1890 2014-05-02 Kito Cheng <kito@0xlab.org>
1892 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
1893 to a C expression marco.
1894 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
1895 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
1896 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
1897 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
1898 HONOR_REG_ALLOC_ORDER.
1899 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
1901 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1903 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
1905 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1907 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
1909 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
1911 * tree-if-conv.c (is_cond_scalar_reduction): New function.
1912 (convert_scalar_cond_reduction): Likewise.
1913 (predicate_scalar_phi): Add recognition and transformation
1914 of simple conditioanl reduction to be vectorizable.
1916 2014-05-01 Marek Polacek <polacek@redhat.com>
1918 PR c/43245
1919 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
1921 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
1923 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
1924 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
1925 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
1926 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
1927 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
1928 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
1929 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
1930 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
1932 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
1934 * config/arc/arc.opt (mlra): Move comment above option name
1935 to avoid mis-parsing as language options.
1937 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1939 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
1940 * config/sol2.h: ... here.
1941 * config/sol2-10.h: Remove.
1943 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
1944 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
1945 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
1946 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
1947 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
1948 * config/sol2.h: ... here.
1949 (SECTION_NAME_FORMAT): Don't redefine.
1950 (STARTFILE_ARCH32_SPEC): Rename to ...
1951 (STARTFILE_ARCH_SPEC): ... this.
1952 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
1953 * config/sparc/sol2.h: ... here.
1954 (SECTION_NAME_FORMAT): Don't undef.
1955 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
1956 (SUBTARGET_EXTRA_SPECS): Remove.
1957 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
1959 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
1960 (MD_STARTFILE_PREFIX): Remove.
1961 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
1962 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
1963 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
1964 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
1965 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
1966 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
1967 * config/i386/sol2.h: ... here.
1968 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
1969 * config/i386/sol2-bi.h: Remove.
1970 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
1971 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
1973 * config/i386/t-sol2-64: Rename to ...
1974 * config/i386/t-sol2: ... this.
1975 * config/sparc/t-sol2-64: Rename to ...
1976 * config/sparc/t-sol2: ... this.
1978 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
1979 sol2_tm_file_head, sol2_tm_file_tail.
1980 Include ${cpu_type}/sol2.h before sol2.h.
1981 Remove sol2-10.h.
1982 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
1983 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
1984 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
1985 Reflect i386/t-sol2-64 renaming.
1986 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
1987 Reflect sparc/t-sol2-64 renaming.
1989 2014-04-30 Richard Biener <rguenther@suse.de>
1991 * passes.c (execute_function_todo): Move TODO_verify_stmts
1992 and TODO_verify_ssa under the TODO_verify_il umbrella.
1993 * tree-ssa.h (verify_ssa): Adjust prototype.
1994 * tree-ssa.c (verify_ssa): Add parameter to tell whether
1995 we should verify SSA operands.
1996 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
1997 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
1998 whether we should verify whether not throwing stmts have EH info.
1999 * graphite-scop-detection.c (create_sese_edges): Adjust.
2000 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
2001 * tree-eh.c (lower_try_finally_switch): Do not add the
2002 default case label twice.
2004 2014-04-30 Marek Polacek <polacek@redhat.com>
2006 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
2007 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
2008 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
2009 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
2011 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
2013 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
2014 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
2015 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
2016 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
2017 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
2018 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
2019 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
2020 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
2022 2014-04-29 David Malcolm <dmalcolm@redhat.com>
2024 * tree-cfg.c (dump_function_to_file): Dump the return type of
2025 functions, in a line to itself before the function body, mimicking
2026 the layout of a C function.
2028 2014-04-29 Jakub Jelinek <jakub@redhat.com>
2030 PR tree-optimization/60971
2031 * tree-tailcall.c (process_assignment): Reject conversions which
2032 reduce precision.
2034 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
2036 * calls.c (initialize_argument_information): Always treat
2037 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
2038 (expand_call): Likewise.
2039 (emit_library_call_calue_1): Likewise.
2040 * expr.c (PUSH_ARGS_REVERSED): Do not define.
2041 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
2042 code accordingly.
2044 2014-04-29 Nick Clifton <nickc@redhat.com>
2046 * config/msp430/msp430.md (umulsidi): Fix typo.
2047 (mulhisi3): Enable even inside interrupt handlers.
2048 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
2049 bigger return address pushed in large mode.
2051 2014-04-29 Nick Clifton <nickc@redhat.com>
2053 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
2054 (arc_init_reg_tables): Use a machine_mode enum to iterate over
2055 available modes.
2056 * config/m32r/m32r.c (init_reg_tables): Likewise.
2057 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
2058 enum to hold the modes.
2060 2014-04-29 Richard Biener <rguenther@suse.de>
2062 * dominance.c (free_dominance_info): Add overload with
2063 function parameter.
2064 (dom_info_state): Likewise.
2065 (dom_info_available_p): Likewise.
2066 * basic-block.h (free_dominance_info, dom_info_state,
2067 dom_info_available_p): Declare overloads.
2068 * passes.c (execute_function_todo): Verify that verifiers
2069 don't change dominator info state. Drop dominator info
2070 for IPA pass invocations.
2071 * cgraph.c (release_function_body): Restore asserts that
2072 dominator information is released.
2074 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
2076 * doc/invoke.texi: Fix typo.
2077 * tree-vrp.c: Fix typos.
2078 * gimple.c (infer_nonnull_range): Reorder operands of an && condition.
2080 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2082 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
2084 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
2086 * config/aarch64/aarch64-builtins.c
2087 (aarch64_types_storestruct_lane_qualifiers): New.
2088 (TYPES_STORESTRUCT_LANE): Likewise.
2089 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
2090 (st3_lane): Likewise.
2091 (st4_lane): Likewise.
2092 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
2093 (vec_store_lanesci_lane<mode>): Likewise.
2094 (vec_store_lanesxi_lane<mode>): Likewise.
2095 (aarch64_st2_lane<VQ:mode>): Likewise.
2096 (aarch64_st3_lane<VQ:mode>): Likewise.
2097 (aarch64_st4_lane<VQ:mode>): Likewise.
2098 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
2099 * config/aarch64/arm_neon.h
2100 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
2101 use new macro arguments.
2102 (__ST3_LANE_FUNC): Likewise.
2103 (__ST4_LANE_FUNC): Likewise.
2104 * config/aarch64/iterators.md (V_TWO_ELEM): New.
2105 (V_THREE_ELEM): Likewise.
2106 (V_FOUR_ELEM): Likewise.
2108 2014-04-28 David Malcolm <dmalcolm@redhat.com>
2110 * doc/gimple.texi: Replace the description of the now-defunct
2111 union gimple_statement_d with a diagram showing the
2112 gimple_statement_base class hierarchy and its relationships to
2113 the GSS_ and GIMPLE_ enums.
2115 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
2117 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
2118 * config/aarch64/aarch64.c
2119 (aarch64_cannot_change_mode_class): Weaken conditions.
2120 (aarch64_modes_tieable_p): New.
2121 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
2123 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
2125 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2126 (loadsync_<mode>): Change mode.
2127 (load_quadpti, store_quadpti): New.
2128 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2129 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2131 2014-04-28 Martin Jambor <mjambor@suse.cz>
2133 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
2134 same alias type as the original statement.
2135 (subreplacement_assignment_data): New type.
2136 (handle_unscalarized_data_in_subtree): New type of parameter,
2137 generate new memory accesses with same alias type as the original
2138 statement.
2139 (load_assign_lhs_subreplacements): Likewise.
2140 (sra_modify_constructor_assign): Generate new memory accesses with
2141 same alias type as the original statement.
2143 2014-04-28 Richard Biener <rguenther@suse.de>
2145 * tree-pass.h (TODO_verify_il): Define.
2146 (TODO_verify_all): Complete properly.
2147 * passes.c (execute_function_todo): Move existing loop-closed
2148 SSA verification under TODO_verify_il.
2149 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
2150 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
2151 Fix tree sharing issue.
2153 2014-04-28 Richard Biener <rguenther@suse.de>
2155 PR middle-end/60092
2156 * builtins.def (DEF_C11_BUILTIN): Add.
2157 (BUILT_IN_ALIGNED_ALLOC): Likewise.
2158 * coretypes.h (enum function_class): Add function_c11_misc.
2159 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
2160 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
2161 (call_may_clobber_ref_p_1): Likewise.
2162 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
2163 (mark_all_reaching_defs_necessary_1): Likewise.
2164 (propagate_necessity): Likewise.
2165 (eliminate_unnecessary_stmts): Likewise.
2166 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
2168 2014-04-28 Richard Biener <rguenther@suse.de>
2170 * tree-vrp.c (vrp_var_may_overflow): Remove.
2171 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
2172 with overflow immediately bump to one before that value and
2173 let iteration figure out overflow status.
2175 2014-04-28 Richard Biener <rguenther@suse.de>
2177 * configure.ac: Do valgrind header checks unconditionally.
2178 Add --enable-valgrind-annotations.
2179 * system.h: Guard valgrind header inclusion with
2180 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
2181 * alloc-pool.c (pool_alloc, pool_free): Use
2182 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
2183 to guard possibly dead code.
2184 * config.in: Regenerated.
2185 * configure: Likewise.
2187 2014-04-28 Jeff Law <law@redhat.com>
2189 PR tree-optimization/60902
2190 * tree-ssa-threadedge.c
2191 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
2192 over real defs when invalidating outputs from statements that do not
2193 produce useful outputs for threading.
2195 2014-04-28 Richard Biener <rguenther@suse.de>
2197 PR tree-optimization/60979
2198 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
2199 SCOPs that end in a block with a successor with abnormal
2200 predecessors.
2202 2014-04-28 Richard Biener <rguenther@suse.de>
2204 * tree-pass.h (execute_pass_list): Adjust prototype.
2205 * passes.c (pass_manager::execute_early_local_passes): Adjust.
2206 (do_per_function): Change callback signature, push all actual
2207 work to the callbals.
2208 (do_per_function_toporder): Likewise.
2209 (execute_function_dump): Adjust.
2210 (execute_function_todo): Likewise.
2211 (clear_last_verified): Likewise.
2212 (verify_curr_properties): Likewise.
2213 (update_properties_after_pass): Likewise.
2214 (execute_pass_list_1): Split out from ...
2215 (execute_pass_list): ... here. Adjust.
2216 (execute_ipa_pass_list): Likewise.
2217 * cgraphunit.c (cgraph_add_new_function): Adjust.
2218 (analyze_function): Likewise.
2219 (expand_function): Likewise.
2220 * cgraph.c (release_function_body): Free dominance info
2221 here instead of asserting it was magically freed elsewhere.
2223 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
2225 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
2226 * configure: Regenerate.
2227 * config/sparc/sparc.opt (muser-mode): New option.
2228 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
2229 for LEON3.
2230 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
2231 * doc/invoke.texi (SPARC options): Document -muser-mode.
2233 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
2235 * cselib.c (find_slot_memmode): Delete.
2236 (cselib_hasher): Change compare_type to a struct.
2237 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
2238 constants.
2239 (preserve_constants_and_equivs): Adjust for new compare_type.
2240 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
2241 (wrap_constant): Delete.
2242 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
2244 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
2246 * doc/install.texi (Building with profile feedback): Remove
2247 outdated sentence.
2249 2014-04-26 Tom de Vries <tom@codesourcery.com>
2251 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
2252 array accesses.
2254 2014-04-25 Cary Coutant <ccoutant@google.com>
2256 PR debug/60929
2257 * dwarf2out.c (should_move_die_to_comdat): A type definition
2258 can contain a subprogram definition, but don't move it to a
2259 comdat unit.
2260 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
2261 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
2262 from original DIE.
2263 (clone_tree_hash): Rename to...
2264 (clone_tree_partial): ...this; change callers. Copy
2265 DW_TAG_subprogram DIEs as declarations.
2266 (copy_decls_walk): Don't copy children of a declaration into a
2267 type unit.
2269 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
2271 PR target/60969
2272 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
2273 alternative 12.
2275 2014-04-25 Jiong Wang <jiong.wang@arm.com>
2277 * config/arm/predicates.md (call_insn_operand): Add long_call check.
2278 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
2279 reg for long_call.
2280 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
2281 restriction.
2283 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2285 * config/arm/arm.c (arm_cortex_a8_tune): Initialise T16-related fields.
2287 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2289 PR tree-optimization/60930
2290 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
2291 creating a multiply candidate by folding two constant
2292 multiplicands when the result overflows.
2294 2014-04-25 Jakub Jelinek <jakub@redhat.com>
2296 PR tree-optimization/60960
2297 * tree-vect-generic.c (expand_vector_operation): Only call
2298 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
2300 2014-04-25 Tom de Vries <tom@codesourcery.com>
2302 * expr.c (clobber_reg_mode): New function.
2303 * expr.h (clobber_reg): New function.
2305 2014-04-25 Tom de Vries <tom@codesourcery.com>
2307 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
2308 clobbers.
2310 2014-04-25 Radovan Obradovic <robradovic@mips.com>
2311 Tom de Vries <tom@codesourcery.com>
2313 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
2314 handle.
2315 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
2316 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
2317 new argument to find_all_hard_reg_sets call.
2319 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2321 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
2322 Use HOST_WIDE_INT_C for mask literal.
2323 (aarch_rev16_shleft_mask_imm_p): Likewise.
2325 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
2327 PR target/60941
2328 * config/sparc/sparc.md (ashlsi3_extend): Delete.
2330 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
2332 PR preprocessor/56540
2333 * config/i386/i386-c.c (ix86_target_macros): Define
2334 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
2336 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2338 * configure.ac (tga_func): Remove.
2339 (LIB_TLS_SPEC): Remove.
2340 * configure: Regenerate.
2341 * config.in: Regenerate.
2342 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
2344 2014-04-25 Richard Biener <rguenther@suse.de>
2346 PR ipa/60912
2347 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
2348 call stmt use/clobber sets during stmt walk instead of
2349 walking the possibly incomplete set of caller edges.
2351 2014-04-25 Richard Biener <rguenther@suse.de>
2353 PR ipa/60911
2354 * passes.c (apply_ipa_transforms): Inline into only caller ...
2355 (execute_one_pass): ... here. Properly bring in function
2356 bodies for nodes we want to apply IPA transforms to.
2358 2014-04-24 Cong Hou <congh@google.com>
2360 PR tree-optimization/60896
2361 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
2362 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
2363 (vect_mark_pattern_stmts): Set the def type of all statements in
2364 PATTERN_DEF_SEQ as vect_internal_def.
2366 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
2368 * doc/extend.texi (PowerPC Built-in Functions): Document new
2369 powerpc extended divide, bcd, pack/unpack 128-bit, builtin functions.
2370 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
2372 * config/rs6000/predicates.md (const_0_to_3_operand): New
2373 predicate to match 0..3 integer constants.
2375 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
2376 to support adding miscellaneous builtin functions.
2377 (BU_DFP_MISC_2): Likewise.
2378 (BU_P7_MISC_1): Likewise.
2379 (BU_P7_MISC_2): Likewise.
2380 (BU_P8V_MISC_3): Likewise.
2381 (BU_MISC_1): Likewise.
2382 (BU_MISC_2): Likewise.
2383 (DIVWE): Add extended divide builtin functions.
2384 (DIVWEO): Likewise.
2385 (DIVWEU): Likewise.
2386 (DIVWEUO): Likewise.
2387 (DIVDE): Likewise.
2388 (DIVDEO): Likewise.
2389 (DIVDEU): Likewise.
2390 (DIVDEUO): Likewise.
2391 (DXEX): Add decimal floating-point builtin functions.
2392 (DXEXQ): Likewise.
2393 (DDEDPD): Likewise.
2394 (DDEDPDQ): Likewise.
2395 (DENBCD): Likewise.
2396 (DENBCDQ): Likewise.
2397 (DIEX): Likewise.
2398 (DIEXQ): Likewise.
2399 (DSCLI): Likewise.
2400 (DSCLIQ): Likewise.
2401 (DSCRI): Likewise.
2402 (DSCRIQ): Likewise.
2403 (CDTBCD): Add new BCD builtin functions.
2404 (CBCDTD): Likewise.
2405 (ADDG6S): Likewise.
2406 (BCDADD): Likewise.
2407 (BCDADD_LT): Likewise.
2408 (BCDADD_EQ): Likewise.
2409 (BCDADD_GT): Likewise.
2410 (BCDADD_OV): Likewise.
2411 (BCDSUB): Likewise.
2412 (BCDSUB_LT): Likewise.
2413 (BCDSUB_EQ): Likewise.
2414 (BCDSUB_GT): Likewise.
2415 (BCDSUB_OV): Likewise.
2416 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
2417 (UNPACK_TD): Likewise.
2418 (PACK_TF): Likewise.
2419 (UNPACK_TF): Likewise.
2420 (UNPACK_TF_0): Likewise.
2421 (UNPACK_TF_1): Likewise.
2422 (PACK_V1TI): Likewise.
2423 (UNPACK_V1TI): Likewise.
2425 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
2426 support for decimal floating point builtin functions.
2427 (rs6000_expand_ternop_builtin): Add checks for the new builtin
2428 functions that take constant arguments.
2429 (rs6000_invalid_builtin): Add decimal floating point builtin support.
2430 (rs6000_init_builtins): Setup long double, _Decimal64, and
2431 _Decimal128 types for new builtin functions.
2432 (builtin_function_type): Set the unsigned flags appropriately for
2433 the new builtin functions.
2434 (rs6000_opt_masks): Add support for decimal floating point builtin
2435 functions.
2437 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
2438 floating point builtin functions.
2439 (RS6000_BTM_COMMON): Likewise.
2440 (RS6000_BTI_long_double): Likewise.
2441 (RS6000_BTI_dfloat64): Likewise.
2442 (RS6000_BTI_dfloat128): Likewise.
2443 (long_double_type_internal_node): Likewise.
2444 (dfloat64_type_internal_node): Likewise.
2445 (dfloat128_type_internal_node): Likewise.
2447 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
2448 2.07 bcd arithmetic instructions.
2449 (UNSPEC_BCDSUB): Likewise.
2450 (UNSPEC_BCD_OVERFLOW): Likewise.
2451 (UNSPEC_BCD_ADD_SUB): Likewise.
2452 (bcd_add_sub): Likewise.
2453 (BCD_TEST): Likewise.
2454 (bcd<bcd_add_sub>): Likewise.
2455 (bcd<bcd_add_sub>_test): Likewise.
2456 (bcd<bcd_add_sub>_test2): Likewise.
2457 (bcd<bcd_add_sub>_<code>): Likewise.
2458 (peephole2 for combined bcd ops): Likewise.
2460 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
2461 decimal floating point builtin functions.
2462 (UNSPEC_DENBCD): Likewise.
2463 (UNSPEC_DXEX): Likewise.
2464 (UNSPEC_DIEX): Likewise.
2465 (UNSPEC_DSCLI): Likewise.
2466 (UNSPEC_DSCRI): Likewise.
2467 (D64_D128): Likewise.
2468 (dfp_suffix): Likewise.
2469 (dfp_ddedpd_<mode>): Likewise.
2470 (dfp_denbcd_<mode>): Likewise.
2471 (dfp_dxex_<mode>): Likewise.
2472 (dfp_diex_<mode>): Likewise.
2473 (dfp_dscli_<mode>): Likewise.
2474 (dfp_dscri_<mode>): Likewise.
2476 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
2477 builtin functions.
2478 (UNSPEC_CDTBCD): Likewise.
2479 (UNSPEC_CBCDTD): Likewise.
2480 (UNSPEC_DIVE): Add support for new extended divide builtin functions.
2481 (UNSPEC_DIVEO): Likewise.
2482 (UNSPEC_DIVEU): Likewise.
2483 (UNSPEC_DIVEUO): Likewise.
2484 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
2485 pack/unpack 128-bit types.
2486 (UNSPEC_PACK_128BIT): Likewise.
2487 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
2488 (udiv<mode>3): Use idiv_ldiv mode attribute.
2489 (div<mode>3): Likewise.
2490 (addg6s): Add new BCD builtin functions.
2491 (cdtbcd): Likewise.
2492 (cbcdtd): Likewise.
2493 (UNSPEC_DIV_EXTEND): Add support for new extended divide instructions.
2494 (div_extend): Likewise.
2495 (div<div_extend>_<mode>"): Likewise.
2496 (FP128_64): Add support for new builtin functions to pack/unpack
2497 128-bit types.
2498 (unpack<mode>): Likewise.
2499 (unpacktf_0): Likewise.
2500 (unpacktf_1): Likewise.
2501 (unpack<mode>_dm): Likewise.
2502 (unpack<mode>_nodm): Likewise.
2503 (pack<mode>): Likewise.
2504 (unpackv1ti): Likewise.
2505 (packv1ti): Likewise.
2507 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
2509 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
2510 is disabled.
2512 2014-04-24 Jakub Jelinek <jakub@redhat.com>
2514 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
2515 * gimplify.c (omp_is_private): Change last argument's type to int.
2516 Only diagnose lastprivate if the simd argument is 1, only diagnose
2517 linear if the simd argument is 2.
2518 (gimplify_omp_for): Adjust omp_is_private callers. When adding
2519 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
2520 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
2521 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
2522 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
2523 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
2524 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
2525 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
2526 * tree-nested.c (convert_nonlocal_omp_clauses,
2527 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
2529 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
2531 PR target/60822
2532 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
2533 operand 1.
2535 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
2537 * flag-types.h (enum ivar_visibility): Add.
2539 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
2541 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
2542 function * argument.
2544 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
2546 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
2548 2014-04-24 Radovan Obradovic <robradovic@mips.com>
2549 Tom de Vries <tom@codesourcery.com>
2551 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
2552 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
2553 reg-note.
2554 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
2555 * emit-rtl.c (try_split): Same.
2557 2014-04-24 Radovan Obradovic <robradovic@mips.com>
2558 Tom de Vries <tom@codesourcery.com>
2560 * common.opt (fuse-caller-save): New option.
2562 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
2564 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of
2565 elements for big-endian.
2567 2014-04-24 Richard Biener <rguenther@suse.de>
2569 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
2570 during TER and instead use the sepops interface for expanding
2571 non-GIMPLE_SINGLE_RHS.
2573 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2575 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
2576 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
2578 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2580 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
2581 assembler 64-bit option.
2582 * configure: Regenerate.
2584 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2586 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
2587 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
2588 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
2589 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
2590 (TARGET_CRYPTO): Take TARGET_SIMD into account.
2592 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2594 * config/aarch64/aarch64-builtins.c
2595 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
2596 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
2597 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
2598 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
2599 builtins.
2600 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
2601 (Vrevsuff): New mode attribute.
2603 2014-04-24 Terry Guo <terry.guo@arm.com>
2605 * config/arm/arm.h (machine_function): Define variable
2606 after_arm_reorg here.
2607 * config/arm/arm.c (after_arm_reorg): Remove the definition.
2608 (arm_split_constant): Update the way to access variable
2609 after_arm_reorg.
2610 (arm_reorg): Ditto.
2611 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
2613 2014-04-23 Tom de Vries <tom@codesourcery.com>
2615 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
2617 2014-04-23 David Malcolm <dmalcolm@redhat.com>
2619 * is-a.h: Update comments to reflect the following changes to the
2620 "pointerness" of the API, making the template parameter match the
2621 return type, allowing use of is-a.h with typedefs of pointers.
2622 (is_a_helper::cast): Return a T rather then a pointer to a T, so
2623 that the return type matches the parameter to the is_a_helper.
2624 (as_a): Likewise.
2625 (dyn_cast): Likewise.
2627 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
2628 pointer from the is-a.h API.
2630 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
2631 (is_a_helper <cgraph_node *>::test): ...this, matching change to
2632 is-a.h API.
2633 (is_a_helper <varpool_node>::test): Likewise, convert to...
2634 (is_a_helper <varpool_node *>::test): ...this.
2636 (varpool_first_variable): Update for removal of implicit pointer
2637 from the is-a.h API.
2638 (varpool_next_variable): Likewise.
2639 (varpool_first_static_initializer): Likewise.
2640 (varpool_next_static_initializer): Likewise.
2641 (varpool_first_defined_variable): Likewise.
2642 (varpool_next_defined_variable): Likewise.
2643 (cgraph_first_defined_function): Likewise.
2644 (cgraph_next_defined_function): Likewise.
2645 (cgraph_first_function): Likewise.
2646 (cgraph_next_function): Likewise.
2647 (cgraph_first_function_with_gimple_body): Likewise.
2648 (cgraph_next_function_with_gimple_body): Likewise.
2649 (cgraph_alias_target): Likewise.
2650 (varpool_alias_target): Likewise.
2651 (cgraph_function_or_thunk_node): Likewise.
2652 (varpool_variable_node): Likewise.
2653 (symtab_real_symbol_p): Likewise.
2654 * cgraphunit.c (referred_to_p): Likewise.
2655 (analyze_functions): Likewise.
2656 (handle_alias_pairs): Likewise.
2657 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
2658 * gimple-ssa.h (gimple_vuse_op): Likewise.
2659 (gimple_vdef_op): Likewise.
2660 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
2661 * gimple.c (gimple_build_asm_1): Likewise.
2662 (gimple_build_try): Likewise.
2663 (gimple_build_resx): Likewise.
2664 (gimple_build_eh_dispatch): Likewise.
2665 (gimple_build_omp_for): Likewise.
2666 (gimple_omp_for_set_clauses): Likewise.
2668 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
2669 (is_a_helper <gimple_statement_asm *>::test): ...this.
2670 (is_a_helper <gimple_statement_bind>::test): Convert to...
2671 (is_a_helper <gimple_statement_bind *>::test): ...this.
2672 (is_a_helper <gimple_statement_call>::test): Convert to...
2673 (is_a_helper <gimple_statement_call *>::test): ...this.
2674 (is_a_helper <gimple_statement_catch>::test): Convert to...
2675 (is_a_helper <gimple_statement_catch *>::test): ...this.
2676 (is_a_helper <gimple_statement_resx>::test): Convert to...
2677 (is_a_helper <gimple_statement_resx *>::test): ...this.
2678 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
2679 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
2680 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
2681 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
2682 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
2683 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
2684 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
2685 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
2686 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
2687 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
2688 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
2689 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
2690 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
2691 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
2692 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
2693 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
2694 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
2695 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
2696 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
2697 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
2698 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
2699 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
2700 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
2701 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
2702 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
2703 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
2704 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
2705 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
2706 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
2707 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
2708 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
2709 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
2710 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
2711 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
2712 (is_a_helper <gimple_statement_phi>::test): Convert to...
2713 (is_a_helper <gimple_statement_phi *>::test): ...this.
2714 (is_a_helper <gimple_statement_transaction>::test): Convert to...
2715 (is_a_helper <gimple_statement_transaction *>::test): ...this.
2716 (is_a_helper <gimple_statement_try>::test): Convert to...
2717 (is_a_helper <gimple_statement_try *>::test): ...this.
2718 (is_a_helper <gimple_statement_wce>::test): Convert to...
2719 (is_a_helper <gimple_statement_wce *>::test): ...this.
2720 (is_a_helper <const gimple_statement_asm>::test): Convert to...
2721 (is_a_helper <const gimple_statement_asm *>::test): ...this.
2722 (is_a_helper <const gimple_statement_bind>::test): Convert to...
2723 (is_a_helper <const gimple_statement_bind *>::test): ...this.
2724 (is_a_helper <const gimple_statement_call>::test): Convert to...
2725 (is_a_helper <const gimple_statement_call *>::test): ...this.
2726 (is_a_helper <const gimple_statement_catch>::test): Convert to...
2727 (is_a_helper <const gimple_statement_catch *>::test): ...this.
2728 (is_a_helper <const gimple_statement_resx>::test): Convert to...
2729 (is_a_helper <const gimple_statement_resx *>::test): ...this.
2730 (is_a_helper <const gimple_statement_eh_dispatch>::test): Convert to...
2731 (is_a_helper <const gimple_statement_eh_dispatch *>::test): ...this.
2732 (is_a_helper <const gimple_statement_eh_filter>::test): Convert to...
2733 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
2734 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
2735 Convert to...
2736 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
2737 ...this.
2738 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
2739 Convert to...
2740 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
2741 ...this.
2742 (is_a_helper <const gimple_statement_omp_return>::test): Convert to...
2743 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
2744 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
2745 to...
2746 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
2747 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
2748 to...
2749 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
2750 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
2751 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
2752 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert to...
2753 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
2754 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
2755 to...
2756 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
2757 (is_a_helper <const gimple_statement_omp_target>::test): Convert to...
2758 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
2759 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
2760 to...
2761 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
2762 (is_a_helper <const gimple_statement_omp_single>::test): Convert to...
2763 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
2764 (is_a_helper <const gimple_statement_omp_teams>::test): Convert to...
2765 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
2766 (is_a_helper <const gimple_statement_omp_task>::test): Convert to...
2767 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
2768 (is_a_helper <const gimple_statement_phi>::test): Convert to...
2769 (is_a_helper <const gimple_statement_phi *>::test): ...this.
2770 (is_a_helper <const gimple_statement_transaction>::test): Convert to...
2771 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
2772 (is_a_helper <const gimple_statement_with_ops>::test): Convert to...
2773 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
2774 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
2775 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
2776 (is_a_helper <const gimple_statement_with_memory_ops>::test): Convert
2777 to...
2778 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
2779 ...this.
2780 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert to...
2781 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
2783 (gimple_use_ops): Update for removal of implicit pointer from the
2784 is-a.h API.
2785 (gimple_set_use_ops): Likewise.
2786 (gimple_vuse): Likewise.
2787 (gimple_vdef): Likewise.
2788 (gimple_vuse_ptr): Likewise.
2789 (gimple_vdef_ptr): Likewise.
2790 (gimple_set_vuse): Likewise.
2791 (gimple_set_vdef): Likewise.
2792 (gimple_omp_return_set_lhs): Likewise.
2793 (gimple_omp_return_lhs): Likewise.
2794 (gimple_omp_return_lhs_ptr): Likewise.
2795 (gimple_call_fntype): Likewise.
2796 (gimple_call_set_fntype): Likewise.
2797 (gimple_call_set_internal_fn): Likewise.
2798 (gimple_call_use_set): Likewise.
2799 (gimple_call_clobber_set): Likewise.
2800 (gimple_bind_vars): Likewise.
2801 (gimple_bind_set_vars): Likewise.
2802 (gimple_bind_body_ptr): Likewise.
2803 (gimple_bind_set_body): Likewise.
2804 (gimple_bind_add_stmt): Likewise.
2805 (gimple_bind_block): Likewise.
2806 (gimple_bind_set_block): Likewise.
2807 (gimple_asm_ninputs): Likewise.
2808 (gimple_asm_noutputs): Likewise.
2809 (gimple_asm_nclobbers): Likewise.
2810 (gimple_asm_nlabels): Likewise.
2811 (gimple_asm_input_op): Likewise.
2812 (gimple_asm_input_op_ptr): Likewise.
2813 (gimple_asm_output_op): Likewise.
2814 (gimple_asm_output_op_ptr): Likewise.
2815 (gimple_asm_set_output_op): Likewise.
2816 (gimple_asm_clobber_op): Likewise.
2817 (gimple_asm_set_clobber_op): Likewise.
2818 (gimple_asm_label_op): Likewise.
2819 (gimple_asm_set_label_op): Likewise.
2820 (gimple_asm_string): Likewise.
2821 (gimple_catch_types): Likewise.
2822 (gimple_catch_types_ptr): Likewise.
2823 (gimple_catch_handler_ptr): Likewise.
2824 (gimple_catch_set_types): Likewise.
2825 (gimple_catch_set_handler): Likewise.
2826 (gimple_eh_filter_types): Likewise.
2827 (gimple_eh_filter_types_ptr): Likewise.
2828 (gimple_eh_filter_failure_ptr): Likewise.
2829 (gimple_eh_filter_set_types): Likewise.
2830 (gimple_eh_filter_set_failure): Likewise.
2831 (gimple_eh_must_not_throw_fndecl): Likewise.
2832 (gimple_eh_must_not_throw_set_fndecl): Likewise.
2833 (gimple_eh_else_n_body_ptr): Likewise.
2834 (gimple_eh_else_e_body_ptr): Likewise.
2835 (gimple_eh_else_set_n_body): Likewise.
2836 (gimple_eh_else_set_e_body): Likewise.
2837 (gimple_try_eval_ptr): Likewise.
2838 (gimple_try_cleanup_ptr): Likewise.
2839 (gimple_try_set_eval): Likewise.
2840 (gimple_try_set_cleanup): Likewise.
2841 (gimple_wce_cleanup_ptr): Likewise.
2842 (gimple_wce_set_cleanup): Likewise.
2843 (gimple_phi_capacity): Likewise.
2844 (gimple_phi_num_args): Likewise.
2845 (gimple_phi_result): Likewise.
2846 (gimple_phi_result_ptr): Likewise.
2847 (gimple_phi_set_result): Likewise.
2848 (gimple_phi_arg): Likewise.
2849 (gimple_phi_set_arg): Likewise.
2850 (gimple_resx_region): Likewise.
2851 (gimple_resx_set_region): Likewise.
2852 (gimple_eh_dispatch_region): Likewise.
2853 (gimple_eh_dispatch_set_region): Likewise.
2854 (gimple_omp_critical_name): Likewise.
2855 (gimple_omp_critical_name_ptr): Likewise.
2856 (gimple_omp_critical_set_name): Likewise.
2857 (gimple_omp_for_clauses): Likewise.
2858 (gimple_omp_for_clauses_ptr): Likewise.
2859 (gimple_omp_for_set_clauses): Likewise.
2860 (gimple_omp_for_collapse): Likewise.
2861 (gimple_omp_for_index): Likewise.
2862 (gimple_omp_for_index_ptr): Likewise.
2863 (gimple_omp_for_set_index): Likewise.
2864 (gimple_omp_for_initial): Likewise.
2865 (gimple_omp_for_initial_ptr): Likewise.
2866 (gimple_omp_for_set_initial): Likewise.
2867 (gimple_omp_for_final): Likewise.
2868 (gimple_omp_for_final_ptr): Likewise.
2869 (gimple_omp_for_set_final): Likewise.
2870 (gimple_omp_for_incr): Likewise.
2871 (gimple_omp_for_incr_ptr): Likewise.
2872 (gimple_omp_for_set_incr): Likewise.
2873 (gimple_omp_for_pre_body_ptr): Likewise.
2874 (gimple_omp_for_set_pre_body): Likewise.
2875 (gimple_omp_parallel_clauses): Likewise.
2876 (gimple_omp_parallel_clauses_ptr): Likewise.
2877 (gimple_omp_parallel_set_clauses): Likewise.
2878 (gimple_omp_parallel_child_fn): Likewise.
2879 (gimple_omp_parallel_child_fn_ptr): Likewise.
2880 (gimple_omp_parallel_set_child_fn): Likewise.
2881 (gimple_omp_parallel_data_arg): Likewise.
2882 (gimple_omp_parallel_data_arg_ptr): Likewise.
2883 (gimple_omp_parallel_set_data_arg): Likewise.
2884 (gimple_omp_task_clauses): Likewise.
2885 (gimple_omp_task_clauses_ptr): Likewise.
2886 (gimple_omp_task_set_clauses): Likewise.
2887 (gimple_omp_task_child_fn): Likewise.
2888 (gimple_omp_task_child_fn_ptr): Likewise.
2889 (gimple_omp_task_set_child_fn): Likewise.
2890 (gimple_omp_task_data_arg): Likewise.
2891 (gimple_omp_task_data_arg_ptr): Likewise.
2892 (gimple_omp_task_set_data_arg): Likewise.
2893 (gimple_omp_taskreg_clauses): Likewise.
2894 (gimple_omp_taskreg_clauses_ptr): Likewise.
2895 (gimple_omp_taskreg_set_clauses): Likewise.
2896 (gimple_omp_taskreg_child_fn): Likewise.
2897 (gimple_omp_taskreg_child_fn_ptr): Likewise.
2898 (gimple_omp_taskreg_set_child_fn): Likewise.
2899 (gimple_omp_taskreg_data_arg): Likewise.
2900 (gimple_omp_taskreg_data_arg_ptr): Likewise.
2901 (gimple_omp_taskreg_set_data_arg): Likewise.
2902 (gimple_omp_task_copy_fn): Likewise.
2903 (gimple_omp_task_copy_fn_ptr): Likewise.
2904 (gimple_omp_task_set_copy_fn): Likewise.
2905 (gimple_omp_task_arg_size): Likewise.
2906 (gimple_omp_task_arg_size_ptr): Likewise.
2907 (gimple_omp_task_set_arg_size): Likewise.
2908 (gimple_omp_task_arg_align): Likewise.
2909 (gimple_omp_task_arg_align_ptr): Likewise.
2910 (gimple_omp_task_set_arg_align): Likewise.
2911 (gimple_omp_single_clauses): Likewise.
2912 (gimple_omp_single_clauses_ptr): Likewise.
2913 (gimple_omp_single_set_clauses): Likewise.
2914 (gimple_omp_target_clauses): Likewise.
2915 (gimple_omp_target_clauses_ptr): Likewise.
2916 (gimple_omp_target_set_clauses): Likewise.
2917 (gimple_omp_target_child_fn): Likewise.
2918 (gimple_omp_target_child_fn_ptr): Likewise.
2919 (gimple_omp_target_set_child_fn): Likewise.
2920 (gimple_omp_target_data_arg): Likewise.
2921 (gimple_omp_target_data_arg_ptr): Likewise.
2922 (gimple_omp_target_set_data_arg): Likewise.
2923 (gimple_omp_teams_clauses): Likewise.
2924 (gimple_omp_teams_clauses_ptr): Likewise.
2925 (gimple_omp_teams_set_clauses): Likewise.
2926 (gimple_omp_sections_clauses): Likewise.
2927 (gimple_omp_sections_clauses_ptr): Likewise.
2928 (gimple_omp_sections_set_clauses): Likewise.
2929 (gimple_omp_sections_control): Likewise.
2930 (gimple_omp_sections_control_ptr): Likewise.
2931 (gimple_omp_sections_set_control): Likewise.
2932 (gimple_omp_for_set_cond): Likewise.
2933 (gimple_omp_for_cond): Likewise.
2934 (gimple_omp_atomic_store_set_val): Likewise.
2935 (gimple_omp_atomic_store_val): Likewise.
2936 (gimple_omp_atomic_store_val_ptr): Likewise.
2937 (gimple_omp_atomic_load_set_lhs): Likewise.
2938 (gimple_omp_atomic_load_lhs): Likewise.
2939 (gimple_omp_atomic_load_lhs_ptr): Likewise.
2940 (gimple_omp_atomic_load_set_rhs): Likewise.
2941 (gimple_omp_atomic_load_rhs): Likewise.
2942 (gimple_omp_atomic_load_rhs_ptr): Likewise.
2943 (gimple_omp_continue_control_def): Likewise.
2944 (gimple_omp_continue_control_def_ptr): Likewise.
2945 (gimple_omp_continue_set_control_def): Likewise.
2946 (gimple_omp_continue_control_use): Likewise.
2947 (gimple_omp_continue_control_use_ptr): Likewise.
2948 (gimple_omp_continue_set_control_use): Likewise.
2949 (gimple_transaction_body_ptr): Likewise.
2950 (gimple_transaction_label): Likewise.
2951 (gimple_transaction_label_ptr): Likewise.
2952 (gimple_transaction_set_body): Likewise.
2953 (gimple_transaction_set_label): Likewise.
2955 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
2956 * ipa-inline-analysis.c (inline_write_summary): Likewise.
2957 * ipa-ref.c (ipa_record_reference): Likewise.
2958 * ipa-reference.c (analyze_function): Likewise.
2959 (ipa_reference_write_optimization_summary): Likewise.
2960 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
2961 (address_taken_from_non_vtable_p): Likewise.
2962 (comdat_can_be_unshared_p_1): Likewise.
2963 * lto-cgraph.c (lto_output_ref): Likewise.
2964 (add_references): Likewise.
2965 (compute_ltrans_boundary): Likewise.
2966 (output_symtab): Likewise.
2967 (input_ref): Likewise.
2968 (input_cgraph_1): Likewise.
2969 (output_cgraph_opt_summary): Likewise.
2970 * lto-streamer-out.c (lto_output): Likewise.
2971 (output_symbol_p): Likewise.
2972 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
2973 (lsei_start_function_in_partition): Likewise.
2974 (lsei_next_variable_in_partition): Likewise.
2975 (lsei_start_variable_in_partition): Likewise.
2976 * symtab.c (insert_to_assembler_name_hash): Likewise.
2977 (unlink_from_assembler_name_hash): Likewise.
2978 (symtab_unregister_node): Likewise.
2979 (symtab_remove_node): Likewise.
2980 (dump_symtab_node): Likewise.
2981 (verify_symtab_base): Likewise.
2982 (verify_symtab_node): Likewise.
2983 (symtab_make_decl_local): Likewise.
2984 (symtab_alias_ultimate_target): Likewise.
2985 (symtab_resolve_alias): Likewise.
2986 (symtab_get_symbol_partitioning_class): Likewise.
2987 * tree-phinodes.c (allocate_phi_node): Likewise.
2988 (reserve_phi_args_for_new_edge): Likewise.
2989 (remove_phi_args): Likewise.
2990 * varpool.c (varpool_node_for_asm): Likewise.
2991 (varpool_remove_unreferenced_decls): Likewise.
2993 2014-04-23 Jeff Law <law@redhat.com>
2995 PR tree-optimization/60902
2996 * tree-ssa-threadedge.c
2997 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
2998 invalidate outputs from statements that do not produce useful
2999 outputs for threading.
3001 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
3003 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
3004 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
3005 machine descriptions for Stack Smashing Protector.
3007 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
3009 * aarch64.md (<optab>_rol<mode>3): New pattern.
3010 (<optab>_rolsi3_uxtw): Likewise.
3011 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
3013 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
3015 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
3016 (arm_cortex_a12_tune): Likewise.
3018 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3020 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
3022 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3024 * config/arm/arm.md (arm_rev16si2): New pattern.
3025 (arm_rev16si2_alt): Likewise.
3026 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
3028 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3030 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
3031 (rev16<mode>2_alt): Likewise.
3032 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
3033 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
3034 (aarch_rev16_shleft_mask_imm_p): Likewise.
3035 (aarch_rev16_p_1): Likewise.
3036 (aarch_rev16_p): Likewise.
3037 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
3038 (aarch_rev16_shright_mask_imm_p): Likewise.
3039 (aarch_rev16_shleft_mask_imm_p): Likewise.
3041 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3043 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
3044 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
3045 rev cost.
3046 (cortex_a53_extra_costs): Likewise.
3047 (cortex_a57_extra_costs): Likewise.
3048 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
3049 (cortexa7_extra_costs): Likewise.
3050 (cortexa8_extra_costs): Likewise.
3051 (cortexa12_extra_costs): Likewise.
3052 (cortexa15_extra_costs): Likewise.
3053 (v7m_extra_costs): Likewise.
3054 (arm_new_rtx_costs): Handle BSWAP.
3056 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3058 * config/arm/arm.c (cortexa8_extra_costs): New table.
3059 (arm_cortex_a8_tune): New tuning struct.
3060 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
3062 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3064 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
3066 2014-04-23 Richard Biener <rguenther@suse.de>
3068 * Makefile.in (OBJS): Remove loop-unswitch.o.
3069 * tree-pass.h (make_pass_rtl_unswitch): Remove.
3070 * passes.def (pass_rtl_unswitch): Likewise.
3071 * loop-init.c (gate_rtl_unswitch): Likewise.
3072 (rtl_unswitch): Likewise.
3073 (pass_data_rtl_unswitch): Likewise.
3074 (pass_rtl_unswitch): Likewise.
3075 (make_pass_rtl_unswitch): Likewise.
3076 * rtl.h (reversed_condition): Likewise.
3077 (compare_and_jump_seq): Likewise.
3078 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
3079 and make static.
3080 * loop-unroll.c (compare_and_jump_seq): Likewise.
3082 2014-04-23 Richard Biener <rguenther@suse.de>
3084 PR tree-optimization/60903
3085 * tree-ssa-loop-im.c (analyze_memory_references): Remove
3086 commented code block.
3087 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
3088 loop flags to newly created BBs and edges.
3090 2014-04-23 Nick Clifton <nickc@redhat.com>
3092 * config/msp430/msp430.c (msp430_handle_option): Move function
3093 to msp430-common.c
3094 (msp430_option_override): Simplify mcu and mcpu option handling.
3095 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
3096 support for -mhwmult command line option.
3097 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
3098 -mhwmult command line option.
3099 (msp430_hwmult_enabled): Delete.
3100 (msp43o_output_labelref): Add support for -mhwmult command line option.
3101 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
3102 (umulsidi3): Likewise.
3103 * config/msp430/msp430.opt (mmcu): Add Report attribute.
3104 (mcpu, mlarge, msmall): Likewise.
3105 (mhwmult): New option.
3106 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
3107 prototype.
3108 (msp430_is_f5_mcu): Remove prototype.
3109 (msp430_use_f5_series_hwmult): Add prototype.
3110 * config/msp430/msp430-opts.h: New file.
3111 * common/config/msp430: New directory.
3112 * common/config/msp430/msp430-common.c: New file.
3113 * config.gcc (msp430): Remove target_has_targetm_common.
3114 * doc/invoke.texi: Document -mhwmult command line option.
3116 2014-04-23 Nick Clifton <nickc@redhat.com>
3118 * config/i386/cygwin.h (ENDFILE_SPEC): Include
3119 default-manifest.o if it can be found in the search path.
3120 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
3122 2014-04-23 Terry Guo <terry.guo@arm.com>
3124 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
3126 2014-04-23 Richard Biener <rguenther@suse.de>
3128 PR middle-end/60895
3129 * tree-inline.c (declare_return_variable): Use mark_addressable.
3131 2014-04-23 Richard Biener <rguenther@suse.de>
3133 PR middle-end/60891
3134 * loop-init.c (loop_optimizer_init): Make sure to apply
3135 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
3137 2014-04-22 Jakub Jelinek <jakub@redhat.com>
3139 PR sanitizer/60275
3140 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
3141 New options.
3142 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
3143 if flag_sanitize_undefined_trap_on_error.
3144 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
3145 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
3146 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
3147 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
3148 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
3149 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
3150 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
3151 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
3152 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
3153 * ubsan.c (ubsan_instrument_unreachable): Return
3154 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
3155 (ubsan_expand_null_ifn): Emit __builtin_trap ()
3156 if flag_sanitize_undefined_trap_on_error and
3157 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
3158 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
3159 instrument_bool_enum_load): Emit __builtin_trap () if
3160 flag_sanitize_undefined_trap_on_error and
3161 __builtin_handle_*_abort () if !flag_sanitize_recover.
3162 * doc/invoke.texi (-fsanitize-recover,
3163 -fsanitize-undefined-trap-on-error): Document.
3165 2014-04-22 Christian Bruel <christian.bruel@st.com>
3167 * config/sh/sh.md (mov<mode>): Replace movQIHI.
3168 Force immediates to SImode.
3170 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
3172 * config/nios2/nios2.md (UNSPEC_ROUND): New.
3173 (lroundsfsi2): New.
3174 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
3175 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
3176 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
3177 (nios2_fpu_insn): Add entry for round.
3178 (N2FPU_NO_ERRNO_P): Define.
3179 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
3180 flag_errno_math.
3181 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
3183 2014-04-22 Richard Henderson <rth@redhat.com>
3185 * config/aarch64/aarch64 (addti3, subti3): New expanders.
3186 (add<GPI>3_compare0): Remove leading * from name.
3187 (add<GPI>3_carryin): Likewise.
3188 (sub<GPI>3_compare0): Likewise.
3189 (sub<GPI>3_carryin): Likewise.
3190 (<su_optab>mulditi3): New expander.
3191 (multi3): New expander.
3192 (madd<GPI>): Remove leading * from name.
3194 2014-04-22 Martin Jambor <mjambor@suse.cz>
3196 * cgraphclones.c (cgraph_function_versioning): Copy
3197 ipa_transforms_to_apply instead of asserting it is empty.
3199 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
3201 PR target/60868
3202 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
3203 on count_exp to get mode.
3205 2014-04-22 Andrew Pinski <apinski@cavium.com>
3207 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
3208 Handle TLS for ILP32.
3209 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
3210 (tlsie_small_<mode>): this and handle PTR.
3211 (tlsie_small_sidi): New pattern.
3212 (tlsle_small): Change to an expand to handle ILP32.
3213 (tlsle_small_<mode>): New pattern.
3214 (tlsdesc_small): Rename to ...
3215 (tlsdesc_small_<mode>): this and handle PTR.
3217 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3219 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
3221 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3223 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
3224 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
3225 (aarch64_types_signed_poly_qualifiers): Likewise.
3226 (aarch64_types_unsigned_signed_qualifiers): Likewise.
3227 (aarch64_types_poly_signed_qualifiers): Likewise.
3228 (TYPES_REINTERP_SS): Type macro added.
3229 (TYPES_REINTERP_SU): Likewise.
3230 (TYPES_REINTERP_SP): Likewise.
3231 (TYPES_REINTERP_US): Likewise.
3232 (TYPES_REINTERP_PS): Likewise.
3233 (aarch64_fold_builtin): New expression folding added.
3234 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
3235 Declarations removed.
3236 (REINTERP_SS): Declarations added.
3237 (REINTERP_US): Likewise.
3238 (REINTERP_PS): Likewise.
3239 (REINTERP_SU): Likewise.
3240 (REINTERP_SP): Likewise.
3241 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
3242 (vreinterpretq_p8_f64): Likewise.
3243 (vreinterpret_p16_f64): Likewise.
3244 (vreinterpretq_p16_f64): Likewise.
3245 (vreinterpret_f32_f64): Likewise.
3246 (vreinterpretq_f32_f64): Likewise.
3247 (vreinterpret_f64_f32): Likewise.
3248 (vreinterpret_f64_p8): Likewise.
3249 (vreinterpret_f64_p16): Likewise.
3250 (vreinterpret_f64_s8): Likewise.
3251 (vreinterpret_f64_s16): Likewise.
3252 (vreinterpret_f64_s32): Likewise.
3253 (vreinterpret_f64_s64): Likewise.
3254 (vreinterpret_f64_u8): Likewise.
3255 (vreinterpret_f64_u16): Likewise.
3256 (vreinterpret_f64_u32): Likewise.
3257 (vreinterpret_f64_u64): Likewise.
3258 (vreinterpretq_f64_f32): Likewise.
3259 (vreinterpretq_f64_p8): Likewise.
3260 (vreinterpretq_f64_p16): Likewise.
3261 (vreinterpretq_f64_s8): Likewise.
3262 (vreinterpretq_f64_s16): Likewise.
3263 (vreinterpretq_f64_s32): Likewise.
3264 (vreinterpretq_f64_s64): Likewise.
3265 (vreinterpretq_f64_u8): Likewise.
3266 (vreinterpretq_f64_u16): Likewise.
3267 (vreinterpretq_f64_u32): Likewise.
3268 (vreinterpretq_f64_u64): Likewise.
3269 (vreinterpret_s64_f64): Likewise.
3270 (vreinterpretq_s64_f64): Likewise.
3271 (vreinterpret_u64_f64): Likewise.
3272 (vreinterpretq_u64_f64): Likewise.
3273 (vreinterpret_s8_f64): Likewise.
3274 (vreinterpretq_s8_f64): Likewise.
3275 (vreinterpret_s16_f64): Likewise.
3276 (vreinterpretq_s16_f64): Likewise.
3277 (vreinterpret_s32_f64): Likewise.
3278 (vreinterpretq_s32_f64): Likewise.
3279 (vreinterpret_u8_f64): Likewise.
3280 (vreinterpretq_u8_f64): Likewise.
3281 (vreinterpret_u16_f64): Likewise.
3282 (vreinterpretq_u16_f64): Likewise.
3283 (vreinterpret_u32_f64): Likewise.
3284 (vreinterpretq_u32_f64): Likewise.
3286 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3288 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
3289 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
3290 (vreinterpret_p8_s8): Likewise.
3291 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
3292 (vreinterpret_p8_s16): Likewise.
3293 (vreinterpret_p8_s32): Likewise.
3294 (vreinterpret_p8_s64): Likewise.
3295 (vreinterpret_p8_f32): Likewise.
3296 (vreinterpret_p8_u8): Likewise.
3297 (vreinterpret_p8_u16): Likewise.
3298 (vreinterpret_p8_u32): Likewise.
3299 (vreinterpret_p8_u64): Likewise.
3300 (vreinterpret_p8_p16): Likewise.
3301 (vreinterpretq_p8_s8): Likewise.
3302 (vreinterpretq_p8_s16): Likewise.
3303 (vreinterpretq_p8_s32): Likewise.
3304 (vreinterpretq_p8_s64): Likewise.
3305 (vreinterpretq_p8_f32): Likewise.
3306 (vreinterpretq_p8_u8): Likewise.
3307 (vreinterpretq_p8_u16): Likewise.
3308 (vreinterpretq_p8_u32): Likewise.
3309 (vreinterpretq_p8_u64): Likewise.
3310 (vreinterpretq_p8_p16): Likewise.
3311 (vreinterpret_p16_s8): Likewise.
3312 (vreinterpret_p16_s16): Likewise.
3313 (vreinterpret_p16_s32): Likewise.
3314 (vreinterpret_p16_s64): Likewise.
3315 (vreinterpret_p16_f32): Likewise.
3316 (vreinterpret_p16_u8): Likewise.
3317 (vreinterpret_p16_u16): Likewise.
3318 (vreinterpret_p16_u32): Likewise.
3319 (vreinterpret_p16_u64): Likewise.
3320 (vreinterpret_p16_p8): Likewise.
3321 (vreinterpretq_p16_s8): Likewise.
3322 (vreinterpretq_p16_s16): Likewise.
3323 (vreinterpretq_p16_s32): Likewise.
3324 (vreinterpretq_p16_s64): Likewise.
3325 (vreinterpretq_p16_f32): Likewise.
3326 (vreinterpretq_p16_u8): Likewise.
3327 (vreinterpretq_p16_u16): Likewise.
3328 (vreinterpretq_p16_u32): Likewise.
3329 (vreinterpretq_p16_u64): Likewise.
3330 (vreinterpretq_p16_p8): Likewise.
3331 (vreinterpret_f32_s8): Likewise.
3332 (vreinterpret_f32_s16): Likewise.
3333 (vreinterpret_f32_s32): Likewise.
3334 (vreinterpret_f32_s64): Likewise.
3335 (vreinterpret_f32_u8): Likewise.
3336 (vreinterpret_f32_u16): Likewise.
3337 (vreinterpret_f32_u32): Likewise.
3338 (vreinterpret_f32_u64): Likewise.
3339 (vreinterpret_f32_p8): Likewise.
3340 (vreinterpret_f32_p16): Likewise.
3341 (vreinterpretq_f32_s8): Likewise.
3342 (vreinterpretq_f32_s16): Likewise.
3343 (vreinterpretq_f32_s32): Likewise.
3344 (vreinterpretq_f32_s64): Likewise.
3345 (vreinterpretq_f32_u8): Likewise.
3346 (vreinterpretq_f32_u16): Likewise.
3347 (vreinterpretq_f32_u32): Likewise.
3348 (vreinterpretq_f32_u64): Likewise.
3349 (vreinterpretq_f32_p8): Likewise.
3350 (vreinterpretq_f32_p16): Likewise.
3351 (vreinterpret_s64_s8): Likewise.
3352 (vreinterpret_s64_s16): Likewise.
3353 (vreinterpret_s64_s32): Likewise.
3354 (vreinterpret_s64_f32): Likewise.
3355 (vreinterpret_s64_u8): Likewise.
3356 (vreinterpret_s64_u16): Likewise.
3357 (vreinterpret_s64_u32): Likewise.
3358 (vreinterpret_s64_u64): Likewise.
3359 (vreinterpret_s64_p8): Likewise.
3360 (vreinterpret_s64_p16): Likewise.
3361 (vreinterpretq_s64_s8): Likewise.
3362 (vreinterpretq_s64_s16): Likewise.
3363 (vreinterpretq_s64_s32): Likewise.
3364 (vreinterpretq_s64_f32): Likewise.
3365 (vreinterpretq_s64_u8): Likewise.
3366 (vreinterpretq_s64_u16): Likewise.
3367 (vreinterpretq_s64_u32): Likewise.
3368 (vreinterpretq_s64_u64): Likewise.
3369 (vreinterpretq_s64_p8): Likewise.
3370 (vreinterpretq_s64_p16): Likewise.
3371 (vreinterpret_u64_s8): Likewise.
3372 (vreinterpret_u64_s16): Likewise.
3373 (vreinterpret_u64_s32): Likewise.
3374 (vreinterpret_u64_s64): Likewise.
3375 (vreinterpret_u64_f32): Likewise.
3376 (vreinterpret_u64_u8): Likewise.
3377 (vreinterpret_u64_u16): Likewise.
3378 (vreinterpret_u64_u32): Likewise.
3379 (vreinterpret_u64_p8): Likewise.
3380 (vreinterpret_u64_p16): Likewise.
3381 (vreinterpretq_u64_s8): Likewise.
3382 (vreinterpretq_u64_s16): Likewise.
3383 (vreinterpretq_u64_s32): Likewise.
3384 (vreinterpretq_u64_s64): Likewise.
3385 (vreinterpretq_u64_f32): Likewise.
3386 (vreinterpretq_u64_u8): Likewise.
3387 (vreinterpretq_u64_u16): Likewise.
3388 (vreinterpretq_u64_u32): Likewise.
3389 (vreinterpretq_u64_p8): Likewise.
3390 (vreinterpretq_u64_p16): Likewise.
3391 (vreinterpret_s8_s16): Likewise.
3392 (vreinterpret_s8_s32): Likewise.
3393 (vreinterpret_s8_s64): Likewise.
3394 (vreinterpret_s8_f32): Likewise.
3395 (vreinterpret_s8_u8): Likewise.
3396 (vreinterpret_s8_u16): Likewise.
3397 (vreinterpret_s8_u32): Likewise.
3398 (vreinterpret_s8_u64): Likewise.
3399 (vreinterpret_s8_p8): Likewise.
3400 (vreinterpret_s8_p16): Likewise.
3401 (vreinterpretq_s8_s16): Likewise.
3402 (vreinterpretq_s8_s32): Likewise.
3403 (vreinterpretq_s8_s64): Likewise.
3404 (vreinterpretq_s8_f32): Likewise.
3405 (vreinterpretq_s8_u8): Likewise.
3406 (vreinterpretq_s8_u16): Likewise.
3407 (vreinterpretq_s8_u32): Likewise.
3408 (vreinterpretq_s8_u64): Likewise.
3409 (vreinterpretq_s8_p8): Likewise.
3410 (vreinterpretq_s8_p16): Likewise.
3411 (vreinterpret_s16_s8): Likewise.
3412 (vreinterpret_s16_s32): Likewise.
3413 (vreinterpret_s16_s64): Likewise.
3414 (vreinterpret_s16_f32): Likewise.
3415 (vreinterpret_s16_u8): Likewise.
3416 (vreinterpret_s16_u16): Likewise.
3417 (vreinterpret_s16_u32): Likewise.
3418 (vreinterpret_s16_u64): Likewise.
3419 (vreinterpret_s16_p8): Likewise.
3420 (vreinterpret_s16_p16): Likewise.
3421 (vreinterpretq_s16_s8): Likewise.
3422 (vreinterpretq_s16_s32): Likewise.
3423 (vreinterpretq_s16_s64): Likewise.
3424 (vreinterpretq_s16_f32): Likewise.
3425 (vreinterpretq_s16_u8): Likewise.
3426 (vreinterpretq_s16_u16): Likewise.
3427 (vreinterpretq_s16_u32): Likewise.
3428 (vreinterpretq_s16_u64): Likewise.
3429 (vreinterpretq_s16_p8): Likewise.
3430 (vreinterpretq_s16_p16): Likewise.
3431 (vreinterpret_s32_s8): Likewise.
3432 (vreinterpret_s32_s16): Likewise.
3433 (vreinterpret_s32_s64): Likewise.
3434 (vreinterpret_s32_f32): Likewise.
3435 (vreinterpret_s32_u8): Likewise.
3436 (vreinterpret_s32_u16): Likewise.
3437 (vreinterpret_s32_u32): Likewise.
3438 (vreinterpret_s32_u64): Likewise.
3439 (vreinterpret_s32_p8): Likewise.
3440 (vreinterpret_s32_p16): Likewise.
3441 (vreinterpretq_s32_s8): Likewise.
3442 (vreinterpretq_s32_s16): Likewise.
3443 (vreinterpretq_s32_s64): Likewise.
3444 (vreinterpretq_s32_f32): Likewise.
3445 (vreinterpretq_s32_u8): Likewise.
3446 (vreinterpretq_s32_u16): Likewise.
3447 (vreinterpretq_s32_u32): Likewise.
3448 (vreinterpretq_s32_u64): Likewise.
3449 (vreinterpretq_s32_p8): Likewise.
3450 (vreinterpretq_s32_p16): Likewise.
3451 (vreinterpret_u8_s8): Likewise.
3452 (vreinterpret_u8_s16): Likewise.
3453 (vreinterpret_u8_s32): Likewise.
3454 (vreinterpret_u8_s64): Likewise.
3455 (vreinterpret_u8_f32): Likewise.
3456 (vreinterpret_u8_u16): Likewise.
3457 (vreinterpret_u8_u32): Likewise.
3458 (vreinterpret_u8_u64): Likewise.
3459 (vreinterpret_u8_p8): Likewise.
3460 (vreinterpret_u8_p16): Likewise.
3461 (vreinterpretq_u8_s8): Likewise.
3462 (vreinterpretq_u8_s16): Likewise.
3463 (vreinterpretq_u8_s32): Likewise.
3464 (vreinterpretq_u8_s64): Likewise.
3465 (vreinterpretq_u8_f32): Likewise.
3466 (vreinterpretq_u8_u16): Likewise.
3467 (vreinterpretq_u8_u32): Likewise.
3468 (vreinterpretq_u8_u64): Likewise.
3469 (vreinterpretq_u8_p8): Likewise.
3470 (vreinterpretq_u8_p16): Likewise.
3471 (vreinterpret_u16_s8): Likewise.
3472 (vreinterpret_u16_s16): Likewise.
3473 (vreinterpret_u16_s32): Likewise.
3474 (vreinterpret_u16_s64): Likewise.
3475 (vreinterpret_u16_f32): Likewise.
3476 (vreinterpret_u16_u8): Likewise.
3477 (vreinterpret_u16_u32): Likewise.
3478 (vreinterpret_u16_u64): Likewise.
3479 (vreinterpret_u16_p8): Likewise.
3480 (vreinterpret_u16_p16): Likewise.
3481 (vreinterpretq_u16_s8): Likewise.
3482 (vreinterpretq_u16_s16): Likewise.
3483 (vreinterpretq_u16_s32): Likewise.
3484 (vreinterpretq_u16_s64): Likewise.
3485 (vreinterpretq_u16_f32): Likewise.
3486 (vreinterpretq_u16_u8): Likewise.
3487 (vreinterpretq_u16_u32): Likewise.
3488 (vreinterpretq_u16_u64): Likewise.
3489 (vreinterpretq_u16_p8): Likewise.
3490 (vreinterpretq_u16_p16): Likewise.
3491 (vreinterpret_u32_s8): Likewise.
3492 (vreinterpret_u32_s16): Likewise.
3493 (vreinterpret_u32_s32): Likewise.
3494 (vreinterpret_u32_s64): Likewise.
3495 (vreinterpret_u32_f32): Likewise.
3496 (vreinterpret_u32_u8): Likewise.
3497 (vreinterpret_u32_u16): Likewise.
3498 (vreinterpret_u32_u64): Likewise.
3499 (vreinterpret_u32_p8): Likewise.
3500 (vreinterpret_u32_p16): Likewise.
3501 (vreinterpretq_u32_s8): Likewise.
3502 (vreinterpretq_u32_s16): Likewise.
3503 (vreinterpretq_u32_s32): Likewise.
3504 (vreinterpretq_u32_s64): Likewise.
3505 (vreinterpretq_u32_f32): Likewise.
3506 (vreinterpretq_u32_u8): Likewise.
3507 (vreinterpretq_u32_u16): Likewise.
3508 (vreinterpretq_u32_u64): Likewise.
3509 (vreinterpretq_u32_p8): Likewise.
3510 (vreinterpretq_u32_p16): Likewise.
3512 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3514 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
3515 Pattern extended.
3516 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator extended.
3517 (sqabs): Likewise.
3518 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
3519 (vqnegd_s64): Likewise.
3520 (vqabs_s64): Likewise.
3521 (vqabsd_s64): Likewise.
3523 2014-04-22 Richard Henderson <rth@redhat.com>
3525 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
3526 computation to the top of the loop.
3528 2014-04-22 Renlin <renlin.li@arm.com>
3529 Jiong Wang <jiong.wang@arm.com>
3531 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
3532 * config/aarch64/aarch64.c (aarch64_layout_frame)
3533 (aarch64_initial_elimination_offset): Likewise.
3535 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
3537 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
3538 Fix indentation.
3540 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
3542 * machmode.h (bitwise_mode_for_mode): Declare.
3543 * stor-layout.h (bitwise_type_for_mode): Likewise.
3544 * stor-layout.c (bitwise_mode_for_mode): New function.
3545 (bitwise_type_for_mode): Likewise.
3546 * builtins.c (fold_builtin_memory_op): Use it instead of
3547 int_mode_for_mode and build_nonstandard_integer_type.
3549 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3551 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
3552 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
3553 (*-*-solaris2*): Simplify.
3554 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
3555 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
3556 *-*-solaris2.9* handling.
3558 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
3559 as bug.
3560 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
3561 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
3562 handling, simplify.
3563 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
3564 * configure: Regenerate.
3566 * config/i386/sol2-9.h: Remove.
3568 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
3569 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
3570 Remove Solaris 9 references.
3572 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
3574 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
3575 (floatuns<GPI:mode><GPF:mode>2): Remove.
3576 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
3577 and floatuns conversions.
3578 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
3579 and floatuns conversions.
3580 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
3581 (w1,w2): New mode attributes for inequal width conversions.
3583 2014-04-22 Renlin Li <Renlin.Li@arm.com>
3585 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
3586 the output asm format.
3588 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
3590 * config/aarch64/aarch64-simd.md
3591 (aarch64_cm<optab>di): Always split.
3592 (*aarch64_cm<optab>di): New.
3593 (aarch64_cmtstdi): Always split.
3594 (*aarch64_cmtstdi): New.
3596 2014-04-22 Jakub Jelinek <jakub@redhat.com>
3598 PR tree-optimization/60823
3599 * omp-low.c (ipa_simd_modify_function_body): Go through
3600 all SSA_NAMEs and for those refering to vector arguments
3601 which are going to be replaced adjust SSA_NAME_VAR and,
3602 if it is a default definition, change it into a non-default
3603 definition assigned at the beginning of function from new_decl.
3604 (ipa_simd_modify_stmt_ops): Rewritten.
3605 * tree-dfa.c (set_ssa_default_def): When removing default def,
3606 check for NULL loc instead of NULL *loc.
3608 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3610 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
3611 restrictions on core registers for DImode values in Thumb2.
3613 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3615 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
3616 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
3618 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3620 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
3621 (*iordi_notzesidi_di): Likewise.
3622 (*iordi_notsesidi_di): Likewise.
3624 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3626 * config/arm/arm-protos.h (tune_params): New struct members.
3627 * config/arm/arm.c: Initialise tune_params per processor.
3628 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
3629 for speed, based on new tune_params.
3631 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3633 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro added.
3634 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added macro.
3635 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment corrected.
3636 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
3637 * config/aarch64/arm_neon.h (vrnd_f64): Added.
3638 (vrnda_f64): Likewise.
3639 (vrndi_f64): Likewise.
3640 (vrndm_f64): Likewise.
3641 (vrndn_f64): Likewise.
3642 (vrndp_f64): Likewise.
3643 (vrndx_f64): Likewise.
3645 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3647 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
3648 GET_MODE_SIZE argument is enum machine_mode.
3650 2014-04-22 Jakub Jelinek <jakub@redhat.com>
3652 PR target/60910
3653 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
3654 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
3656 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
3658 PR middle-end/60281
3659 * asan.c (asan_emit_stack_protection): Force the base to align to
3660 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
3661 appropriate bits if STRICT_ALIGNMENT.
3662 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
3663 when asan is on.
3664 (expand_used_vars): Leave a space in the stack frame for alignment
3665 if STRICT_ALIGNMENT.
3667 2014-04-21 David Malcolm <dmalcolm@redhat.com>
3669 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
3670 than a gimple.
3671 (gimple_store_p): Likewise.
3672 (gimple_assign_load_p): Likewise.
3673 (gimple_assign_cast_p): Likewise.
3674 (gimple_clobber_p): Likewise.
3676 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
3677 rather than a gimple.
3678 (gimple_assign_cast_p): Likewise.
3680 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
3682 PR target/60735
3683 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
3684 If mode is DDmode and TARGET_E500_DOUBLE allow move.
3686 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
3687 more debug information for E500 if -mdebug=reg.
3689 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
3691 PR target/60909
3692 * config/i386/i386.c (ix86_expand_builtin)
3693 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
3694 register for target RTX.
3695 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
3697 2014-04-18 Cong Hou <congh@google.com>
3699 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
3700 the widen-mult pattern by handling two operands with different sizes,
3701 and operands whose size is smaller than half of the result type.
3703 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
3705 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
3706 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
3707 (do_estimate_edge_time): Compute it.
3708 * ipa-inline.c (want_inline_small_function_p): Bypass
3709 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
3711 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
3713 * ipa-inline.c (spec_rem): New static variable.
3714 (dump_overall_stats): New function.
3715 (dump_inline_stats): New function.
3717 2014-04-18 Richard Henderson <rth@redhat.com>
3719 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
3720 to GET_MODE_SIZE, not a reg_class_t.
3722 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3724 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
3725 (vsx_xxmrglw_<mode>): Likewise.
3727 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
3729 PR target/60876
3730 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
3731 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
3732 (rs6000_init_hard_regno_mode_ok): Likewise.
3734 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
3736 * ipa-inline.c (inline_small_functions): Account only non-cold
3737 functions.
3738 * doc/invoke.texi (inline-unit-growth): Update documentation.
3740 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
3742 * config/rs6000/rs6000.md (addti3, subti3): New.
3744 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
3746 PR target/60863
3747 * config/i386/i386.c (ix86_expand_clear): Remove outdated
3748 comment. Check optimize_insn_for_size_p instead of
3749 optimize_insn_for_speed_p.
3751 2014-04-17 Martin Jambor <mjambor@suse.cz>
3753 * gimple-iterator.c (gsi_start_edge): New function.
3754 * gimple-iterator.h (gsi_start_edge): Declare.
3755 * tree-sra.c (single_non_eh_succ): New function.
3756 (disqualify_ops_if_throwing_stmt): Renamed to
3757 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
3758 having one non-EH successor BB.
3759 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
3760 generate loads into replacements.
3761 (sra_modify_assign): Likewise and and also use the simple path for
3762 such statements.
3763 (sra_modify_function_body): Commit statements on edges.
3765 2014-04-17 Richard Biener <rguenther@suse.de>
3767 PR middle-end/60849
3768 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
3769 comparison results and add clarifying comment.
3771 2014-04-17 Jakub Jelinek <jakub@redhat.com>
3773 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
3774 (blank_mode): Initialize it.
3775 (emit_mode_size_inline, emit_mode_nunits_inline,
3776 emit_mode_inner_inline): New functions.
3777 (emit_insn_modes_h): Call them and surround their output with
3778 #if GCC_VERSION >= 4001 ... #endif.
3779 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
3780 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
3781 mode_* arrays if the argument is __builtin_constant_p.
3782 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
3783 is enum machine_mode.
3785 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3787 * passes.c (opt_pass::execute): Adjust.
3788 (pass_manager::execute_pass_mode_switching): Likewise.
3789 (early_local_passes::execute): Likewise.
3790 (execute_one_pass): Pass cfun to the pass's execute method.
3791 * tree-pass.h (opt_pass::execute): Add function * argument.
3792 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
3793 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
3794 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
3795 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3796 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
3797 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
3798 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
3799 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
3800 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
3801 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
3802 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
3803 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
3804 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
3805 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
3806 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
3807 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
3808 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
3809 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
3810 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
3811 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3812 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3813 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3814 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3815 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3816 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3817 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3818 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
3819 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
3820 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
3821 Adjust.
3823 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3825 * passes.c (opt_pass::gate): Take function * argument.
3826 (gate_all_early_local_passes): Merge into
3827 (early_local_passes::gate): this.
3828 (gate_all_early_optimizations): Merge into
3829 (all_early_optimizations::gate): this.
3830 (gate_all_optimizations): Mege into
3831 (all_optimizations::gate): this.
3832 (gate_all_optimizations_g): Merge into
3833 (all_optimizations_g::gate): this.
3834 (gate_rest_of_compilation): Mege into
3835 (rest_of_compilation::gate): this.
3836 (gate_postreload): Merge into
3837 (postreload::gate): this.
3838 (dump_one_pass): Pass cfun to the pass's gate method.
3839 (execute_ipa_summary_passes): Likewise.
3840 (execute_one_pass): Likewise.
3841 (ipa_write_summaries_2): Likewise.
3842 (ipa_write_optimization_summaries_1): Likewise.
3843 (ipa_read_summaries_1): Likewise.
3844 (ipa_read_optimization_summaries_1): Likewise.
3845 (execute_ipa_stmt_fixups): Likewise.
3846 * tree-pass.h (opt_pass::gate): Add function * argument.
3847 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
3848 combine-stack-adj.c, combine.c, compare-elim.c,
3849 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3850 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
3851 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
3852 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
3853 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
3854 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
3855 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
3856 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
3857 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
3858 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
3859 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
3860 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
3861 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
3862 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
3863 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3864 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3865 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3866 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3867 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3868 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3869 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3870 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3871 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
3872 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
3873 var-tracking.c, vtable-verify.c, web.c: Adjust.
3875 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3877 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
3878 * configure: Regenerate.
3880 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3882 * passes.c (dump_one_pass): don't check pass->has_gate.
3883 (execute_ipa_summary_passes): Likewise.
3884 (execute_one_pass): Likewise.
3885 (ipa_write_summaries_2): Likewise.
3886 (ipa_write_optimization_summaries_1): Likewise.
3887 (ipa_read_optimization_summaries_1): Likewise.
3888 (execute_ipa_stmt_fixups): Likewise.
3889 * tree-pass.h (pass_data::has_gate): Remove.
3890 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
3891 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
3892 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
3893 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3894 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
3895 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
3896 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
3897 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
3898 gimple-low.c, gimple-ssa-isolate-paths.c,
3899 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
3900 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
3901 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
3902 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
3903 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
3904 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
3905 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
3906 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
3907 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
3908 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
3909 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
3910 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3911 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3912 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3913 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3914 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3915 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3916 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3917 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3918 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
3919 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
3920 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
3921 Adjust.
3923 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3925 * pass_manager.h (pass_manager::register_dump_files_1): Remove
3926 declaration.
3927 * passes.c (pass_manager::register_dump_files_1): Merge into
3928 (pass_manager::register_dump_files): this, and remove its handling of
3929 properties since the pass always has the properties anyway.
3930 (pass_manager::pass_manager): Adjust.
3932 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3934 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
3935 * passes.c (pass_manager::register_dump_files_1): Remove dead code
3936 dealing with properties.
3937 (pass_manager::register_dump_files): Adjust.
3939 2014-03-20 Mark Wielaard <mjw@redhat.com>
3941 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
3942 then represent the bound as normal constant value.
3944 2014-04-17 Jakub Jelinek <jakub@redhat.com>
3946 PR target/60847
3947 Forward port from 4.8 branch
3948 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
3950 * config/i386/bmiintrin.h (_blsi_u32): New.
3951 (_blsi_u64): Ditto.
3952 (_blsr_u32): Ditto.
3953 (_blsr_u64): Ditto.
3954 (_blsmsk_u32): Ditto.
3955 (_blsmsk_u64): Ditto.
3956 (_tzcnt_u32): Ditto.
3957 (_tzcnt_u64): Ditto.
3959 2014-04-17 Kito Cheng <kito@0xlab.org>
3961 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
3963 2014-04-17 Richard Biener <rguenther@suse.de>
3965 PR middle-end/60849
3966 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
3967 boolean results for comparisons.
3969 2014-04-17 Richard Biener <rguenther@suse.de>
3971 PR tree-optimization/60836
3972 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
3973 initial PHI args to be gimple values.
3975 2014-04-17 Richard Biener <rguenther@suse.de>
3977 PR tree-optimization/60841
3978 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
3979 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
3980 of stmts to SLP build.
3981 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
3982 (vect_analyze_slp): Likewise.
3983 (vect_analyze_slp_instance): Likewise.
3984 (vect_build_slp_tree): Limit overall SLP tree growth.
3985 * tree-vectorizer.h (vect_analyze_data_refs,
3986 vect_analyze_slp): Adjust prototypes.
3988 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3990 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
3991 Silvermont.
3993 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3995 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
3996 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
3997 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
3998 for TARGET_SLOW_PSHUFB
4000 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
4002 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
4003 * config/i386/i386.c (intel_cost): Ditto.
4005 2014-04-17 Joey Ye <joey.ye@arm.com>
4007 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
4009 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
4011 * opts.c (common_handle_option): Disable -fipa-reference coorectly
4012 with -fuse-profile.
4014 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
4016 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
4017 (type_all_derivations_known_p): New predicate.
4018 (type_all_ctors_visible_p): New predicate.
4019 (type_possibly_instantiated_p): New predicate.
4020 (get_odr_type): Compute all_derivations_known.
4021 (dump_odr_type): Dump the flag.
4022 (maybe_record_type): Cleanup.
4023 (record_target_from_binfo): Add bases_to_consider array;
4024 record bases for types w/o instances and skip CXX destructor.
4025 (possible_polymorphic_call_targets_1): Add bases_to_consider
4026 and consider_construction parameters; check if type may have instance.
4027 (get_polymorphic_call_info): Set maybe_in_construction to true
4028 when we know nothing.
4029 (record_targets_from_bases): Skip CXX destructors; they are
4030 never called for types in construction.
4031 (possible_polymorphic_call_targets): Do not record target when
4032 type may not have instance.
4034 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
4036 PR ipa/60854
4037 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
4038 external aliases alive, too.
4040 2014-04-16 Andrew Pinski <apinski@cavium.com>
4042 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
4043 definition.
4045 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
4047 * final.c (compute_alignments): Do not apply loop alignment to a block
4048 falling through to the exit.
4050 2014-04-16 Catherine Moore <clm@codesourcery.com>
4052 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
4053 Adjust constraints for microMIPS store patterns.
4055 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
4057 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
4059 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
4061 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
4062 (append_use): Run at -O0.
4063 (append_vdef): Likewise.
4064 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
4065 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
4067 2014-04-16 Jakub Jelinek <jakub@redhat.com>
4069 PR tree-optimization/60844
4070 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
4071 (propagate_op_to_single_use, remove_visited_stmt_chain,
4072 linearize_expr, repropagate_negates, reassociate_bb): Use it
4073 instead of gsi_remove.
4075 2014-04-16 Martin Jambor <mjambor@suse.cz>
4077 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
4078 ipa_transforms_to_apply.
4079 (cgraph_function_versioning): Assert that old_node has empty
4080 ipa_transforms_to_apply.
4081 * trans-mem.c (ipa_tm_create_version): Likewise.
4082 * tree-inline.c (tree_function_versioning): Do not duplicate
4083 ipa_transforms_to_apply.
4085 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4087 PR target/60817
4088 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
4089 x86_64-*-* cases.
4090 Pass necessary as flags on 64-bit Solaris/x86.
4091 Use lowercase relocs for x86_64-*-*.
4092 * configure: Regenerate.
4094 2014-04-15 Jan Hubicka <jh@suse.cz>
4096 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
4097 (maybe_record_node, likely_target_p): Use it.
4099 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4101 PR target/60839
4102 Revert following patch
4104 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
4106 PR target/60735
4107 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
4108 software floating point or no floating point registers, do not
4109 allow any type in the FPRs. Eliminate a test for SPE SIMD types
4110 in GPRs that occurs after we tested for GPRs that would never be
4111 true.
4113 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
4114 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
4115 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
4116 specifically allow DDmode, since that does not use the SPE SIMD
4117 instructions.
4119 2014-03-21 Mark Wielaard <mjw@redhat.com>
4121 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
4122 as unsigned or int depending on type and value used.
4124 2014-04-15 Richard Biener <rguenther@suse.de>
4126 PR rtl-optimization/56965
4127 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
4128 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
4129 ... here.
4130 * alias.c (true_dependence_1): Do not call
4131 nonoverlapping_component_refs_p.
4132 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
4133 nonoverlapping_component_refs_p.
4134 (indirect_refs_may_alias_p): Likewise.
4136 2014-04-15 Teresa Johnson <tejohnson@google.com>
4138 * cfg.c (dump_bb_info): Fix flags check.
4139 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
4141 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4143 PR rtl-optimization/60663
4144 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
4145 avoid 0 cost.
4147 2014-04-15 Richard Biener <rguenther@suse.de>
4149 * lto-streamer.h (LTO_major_version): Bump to 4.
4151 2014-04-15 Richard Biener <rguenther@suse.de>
4153 * common.opt (lto_partition_model): New enum.
4154 (flto-partition=): Merge separate options with a single with argument,
4155 add -flto-partition=one support.
4156 * flag-types.h (enum lto_partition_model): Declare.
4157 * opts.c (finish_options): Remove duplicate -flto-partition=
4158 option check.
4159 * lto-wrapper.c (run_gcc): Adjust.
4161 2014-04-15 Richard Biener <rguenther@suse.de>
4163 * alias.c (ncr_compar): New function.
4164 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
4166 2014-04-15 Richard Biener <rguenther@suse.de>
4168 * alias.c (record_component_aliases): Do not walk BINFOs.
4170 2014-04-15 Richard Biener <rguenther@suse.de>
4172 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
4173 Add struct function argument and adjust.
4174 (find_func_aliases_for_call): Likewise.
4175 (find_func_aliases): Likewise.
4176 (find_func_clobbers): Likewise.
4177 (intra_create_variable_infos): Likewise.
4178 (compute_points_to_sets): Likewise.
4179 (ipa_pta_execute): Adjust. Do not push/pop cfun.
4181 2014-04-15 Richard Biener <rguenther@suse.de>
4183 * tree.c (iterative_hash_expr): Use enum tree_code_class
4184 to store TREE_CODE_CLASS.
4185 (tree_block): Likewise.
4186 (tree_set_block): Likewise.
4187 * tree.h (fold_build_pointer_plus_loc): Use
4188 convert_to_ptrofftype_loc.
4190 2014-04-15 Jakub Jelinek <jakub@redhat.com>
4192 PR plugins/59335
4193 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
4194 added in 4.9.
4196 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
4198 * cfgloop.h (struct loop): Move force_vectorize down.
4199 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
4200 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
4201 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
4202 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
4203 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
4204 * tree-core.h (enum annot_expr_kind): Add new kind values.
4205 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
4206 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
4207 kinds.
4208 * tree.def (ANNOTATE_EXPR): Tweak comment.
4210 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
4212 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
4213 cxa_pure_virtual).
4215 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
4217 * tree.h (TYPE_IDENTIFIER): Declare.
4218 * tree.c (subrange_type_for_debug_p): Use it.
4219 * godump.c (go_format_type): Likewise.
4220 * dwarf2out.c (is_cxx_auto, modified_type_die,
4221 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
4222 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
4224 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
4226 PR lto/60820
4227 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
4229 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
4231 * config/i386/i386.c (examine_argument): Return bool. Return true if
4232 parameter should be passed in memory.
4233 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
4234 (construct_container): Update calls to examine_argument.
4235 (function_arg_advance_64): Ditto.
4236 (return_in_memory_32): Merge with ix86_return_in_memory.
4237 (return_in_memory_64): Ditto.
4238 (return_in_memory_ms_64): Ditto.
4240 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
4242 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
4243 * coverage.c (coverage_compute_profile_id): Handle externally visible
4244 symbols.
4246 2014-04-14 Martin Jambor <mjambor@suse.cz>
4248 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
4249 DECL_DISREGARD_INLINE_LIMITS functions.
4251 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
4253 PR target/60827
4254 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
4256 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
4258 PR target/60827
4259 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
4260 optimize_insn_for_speed_p instead of
4261 optimize_function_for_speed_p.
4263 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
4265 * doc/invoke.texi (free): Document AArch64.
4267 2014-04-14 Richard Biener <rguenther@suse.de>
4269 PR tree-optimization/60042
4270 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
4271 (insert_into_preds_of_block): Do not prevent PHI insertion
4272 for REFERENCE exprs here ...
4273 (eliminate_dom_walker::before_dom_children): ... but prevent
4274 their use here under similar conditions when applied to the
4275 IL after PRE optimizations.
4277 2014-04-14 Richard Biener <rguenther@suse.de>
4279 * passes.def: Move early points-to after early SRA.
4281 2014-04-14 Richard Biener <rguenther@suse.de>
4283 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
4284 check for which sign-changes we allow when forwarding
4285 a converted value into a switch.
4287 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
4289 * stor-layout.c (place_field): Finalize non-constant offset for the
4290 field, if any.
4292 2014-04-14 Richard Biener <rguenther@suse.de>
4294 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
4295 as argument.
4296 (expand_switch_using_bit_tests_p): Likewise.
4297 (process_switch): Compute and pass on speed_p based on the
4298 switch stmt.
4299 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
4300 optimize_bb_for_speed_p.
4302 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
4304 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
4305 * function.h (struct function): Rename has_force_vect_loops into
4306 has_force_vectorize_loops.
4307 * lto-streamer-in.c (input_cfg): Adjust for renaming.
4308 (input_struct_function_base): Likewise.
4309 * lto-streamer-out.c (output_cfg): Likewise.
4310 (output_struct_function_base): Likewise.
4311 * omp-low.c (expand_omp_simd): Likewise.
4312 * tree-cfg.c (move_sese_region_to_fn): Likewise.
4313 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
4314 (version_loop_for_if_conversion): Likewise.
4315 (tree_if_conversion): Likewise.
4316 (main_tree_if_conversion): Likewise.
4317 (gate_tree_if_conversion): Likewise.
4318 * tree-inline.c (copy_loops): Likewise.
4319 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
4320 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
4321 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
4322 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
4323 * tree-vectorizer.c (vectorize_loops): Likewise.
4324 * tree-vectorizer.h (unlimited_cost_model): Likewise.
4326 2014-04-14 Richard Biener <rguenther@suse.de>
4328 PR lto/60720
4329 * lto-streamer-out.c (wrap_refs): New function.
4330 (lto_output): Wrap symbol references in global initializes in
4331 type-preserving MEM_REFs.
4333 2014-04-14 Christian Bruel <christian.bruel@st.com>
4335 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
4337 2014-04-14 Christian Bruel <christian.bruel@st.com>
4339 * config/sh/sh.md (setmemqi): New expand pattern.
4340 * config/sh/sh.h (CLEAR_RATIO): Define.
4341 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
4342 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
4344 2014-04-14 Richard Biener <rguenther@suse.de>
4346 PR middle-end/55022
4347 * fold-const.c (negate_expr_p): Don't negate directional rounding
4348 division.
4349 (fold_negate_expr): Likewise.
4351 2014-04-14 Richard Biener <rguenther@suse.de>
4353 PR tree-optimization/59817
4354 PR tree-optimization/60453
4355 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
4356 recursion to catch all CHRECs in the scalar evolution and restrict
4357 the predicate for the remains appropriately.
4359 2014-04-12 Catherine Moore <clm@codesourcery.com>
4361 * config/mips/constraints.md: Add new register constraint "kb".
4362 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
4363 (*movhi_internal): Likewise.
4364 (*movqi_internal): Likewise.
4365 * config/mips/mips.h (M16_STORE_REGS): New register class.
4366 (REG_CLASS_NAMES): Add M16_STORE_REGS.
4367 (REG_CLASS_CONTENTS): Likewise.
4368 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
4370 2014-04-11 Tobias Burnus <burnus@net-b.de>
4372 PR c/60194
4373 * doc/invoke.texi (-Wformat-signedness): Document it.
4374 (Wformat=2): Mention that this enables -Wformat-signedness.
4376 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
4378 * common/config/epiphany/epiphany-common.c
4379 (epiphany_option_optimization_table): Enable section anchors by
4380 default at -O1 or higher.
4381 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
4382 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
4383 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
4384 carries no extra cost.
4385 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
4386 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
4387 * config/epiphany/predicates.md (memclob_operand): New predicate.
4388 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
4389 Use memclob_operand predicate and X constraint for operand 3.
4391 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
4393 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
4394 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
4395 its operands.
4397 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
4399 PR rtl-optimization/60651
4400 * mode-switching.c (optimize_mode_switching): Make sure to emit
4401 sets of a lower numbered entity before sets of a higher numbered
4402 entity to a mode of the same or lower priority.
4403 When creating a seginfo for a basic block that starts with a code
4404 label, move the insertion point past the code label.
4405 (new_seginfo): Document and enforce requirement that
4406 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
4407 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
4408 * doc/tm.texi: Regenerate.
4410 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
4412 PR target/60811
4413 * config/arc/arc.c (arc_save_restore): Fix assert typo.
4415 2013-04-11 Jakub Jelinek <jakub@redhat.com>
4417 * BASE-VER: Set to 4.10.0.
4419 2014-04-11 Tobias Burnus <burnus@net-b.de>
4421 PR other/59055
4422 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
4423 * doc/gcc.texi (Service): Update description in the @menu
4424 * doc/invoke.texi (Option Summary): Remove misplaced and
4425 duplicated @menu.
4427 2014-04-11 Steve Ellcey <sellcey@mips.com>
4428 Jakub Jelinek <jakub@redhat.com>
4430 PR middle-end/60556
4431 * expr.c (convert_move): Use emit_store_flag_force instead of
4432 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
4433 argument to it.
4435 2014-04-11 Richard Biener <rguenther@suse.de>
4437 PR middle-end/60797
4438 * varasm.c (assemble_alias): Avoid endless error reporting
4439 recursion by setting TREE_ASM_WRITTEN.
4441 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4443 * config/s390/s390.md: Add a splitter for NOT rtx.
4445 2014-04-11 Jakub Jelinek <jakub@redhat.com>
4447 PR rtl-optimization/60663
4448 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
4450 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
4451 Jakub Jelinek <jakub@redhat.com>
4453 PR lto/60567
4454 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
4455 flag from decl_node to node.
4457 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4459 PR debug/60655
4460 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
4461 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
4462 ameliorating the cases where it can be.
4464 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
4466 Revert
4467 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
4469 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
4470 (loadsync_<mode>): Change mode.
4471 (load_quadpti, store_quadpti): New.
4472 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
4473 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
4474 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
4476 2014-04-09 Cong Hou <congh@google.com>
4478 PR testsuite/60773
4479 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
4480 documentation.
4482 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4484 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
4485 instead of vnor to exploit possible fusion opportunity in the
4486 future.
4487 (altivec_expand_vec_perm_const_le): Likewise.
4489 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
4491 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
4492 (loadsync_<mode>): Change mode.
4493 (load_quadpti, store_quadpti): New.
4494 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
4495 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
4497 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
4499 PR target/60763
4500 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
4501 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
4502 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
4504 2014-04-08 Richard Biener <rguenther@suse.de>
4506 PR middle-end/60706
4507 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
4508 a 64bit widest int print double-int similar to on HWI64 hosts.
4510 2014-04-08 Richard Biener <rguenther@suse.de>
4512 PR tree-optimization/60785
4513 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
4514 default defs properly.
4516 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
4518 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
4519 (Weffc++): Likewise.
4521 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
4523 * ipa-devirt.c (maybe_record_node): When node is not recorded,
4524 set completep to false rather than true.
4526 2014-04-07 Douglas B Rupp <rupp@adacore.com>
4528 PR target/60504
4529 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
4530 ARM_TARGET2_DWARF_FORMAT.
4532 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
4534 PR target/60609
4535 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
4536 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
4537 ADDR_DIFF_VEC.
4539 2014-04-07 Richard Biener <rguenther@suse.de>
4541 PR tree-optimization/60766
4542 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
4543 (may_eliminate_iv): Convert cand_value_at result to desired type.
4545 2014-04-07 Jason Merrill <jason@redhat.com>
4547 PR c++/60731
4548 * common.opt (-fno-gnu-unique): Add.
4549 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
4551 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4553 * haifa-sched.c: Fix outdated function reference and minor
4554 grammar errors in introductory comment.
4556 2014-04-07 Richard Biener <rguenther@suse.de>
4558 PR middle-end/60750
4559 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
4560 for noreturn calls.
4561 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
4563 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
4565 PR debug/55794
4566 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
4567 size accounting for thunks.
4568 (pa_asm_output_mi_thunk): Use final_start_function() and
4569 final_end_function() to output function start and end directives.
4571 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
4573 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
4574 device specific ISA/ feature information. Remove short_sp and
4575 errata_skip ds. Add avr_device_specific_features enum to have device
4576 specific info.
4577 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
4578 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
4579 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
4580 updated device specific info.
4581 * config/avr/avr-mcus.def: Merge device specific details to
4582 dev_attribute field.
4583 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
4584 errata_skip.
4585 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
4586 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
4587 assembler if RMW isa supported by current device.
4588 * config/avr/genmultilib.awk: Update as device info structure changed.
4589 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
4591 2014-04-04 Cong Hou <congh@google.com>
4593 PR tree-optimization/60656
4594 * tree-vect-stmts.c (supportable_widening_operation):
4595 Fix a bug that elements in a vector with vect_used_by_reduction
4596 property are incorrectly reordered when the operation on it is not
4597 consistant with the one in reduction operation.
4599 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
4601 PR rtl-optimization/60155
4602 * gcse.c (record_set_data): New function.
4603 (single_set_gcse): New function.
4604 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
4605 (hoist_code): Likewise.
4606 (get_pressure_class_and_nregs): Likewise.
4608 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
4610 * explow.c (probe_stack_range): Emit a final optimization blockage.
4612 2014-04-04 Anthony Green <green@moxielogic.com>
4614 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
4615 typos.
4617 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
4619 PR ipa/59626
4620 * lto-cgraph.c (input_overwrite_node): Check that partitioning
4621 flags are set only during streaming.
4622 * ipa.c (process_references, walk_polymorphic_call_targets,
4623 symtab_remove_unreachable_nodes): Drop bodies of always inline
4624 after early inlining.
4625 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
4627 2014-04-04 Jakub Jelinek <jakub@redhat.com>
4628 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4630 PR debug/60655
4631 * dwarf2out.c (const_ok_for_output_1): Reject expressions
4632 containing a NOT.
4634 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4636 PR bootstrap/60743
4637 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
4638 duration.
4639 (cortex_a53_fdivd): Likewise.
4641 2014-04-04 Martin Jambor <mjambor@suse.cz>
4643 PR ipa/60640
4644 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
4645 Adjust all callers.
4646 * cgraph.c (clone_of_p): Also return true if thunks match.
4647 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
4648 cgraph_function_or_thunk_node and an obsolete comment.
4649 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
4650 file.
4651 (build_function_decl_skip_args): Likewise.
4652 (set_new_clone_decl_and_node_flags): New function.
4653 (duplicate_thunk_for_node): Likewise.
4654 (redirect_edge_duplicating_thunks): Likewise.
4655 (cgraph_clone_node): New parameter args_to_skip, pass it to
4656 redirect_edge_duplicating_thunks which is called instead of
4657 cgraph_redirect_edge_callee.
4658 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
4659 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
4661 2014-04-04 Jeff Law <law@redhat.com>
4663 PR target/60657
4664 * config/arm/predicates.md (const_int_I_operand): New predicate.
4665 (const_int_M_operand): Similarly.
4666 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
4667 const_int_operand.
4668 (insv_t2, extv_reg, extzv_t2): Likewise.
4669 (load_multiple_with_writeback): Similarly for const_int_I_operand.
4670 (pop_multiple_with_writeback_and_return): Likewise.
4671 (vfp_pop_multiple_with_writeback): Likewise
4673 2014-04-04 Richard Biener <rguenther@suse.de>
4675 PR ipa/60746
4676 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
4677 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
4678 non-GIMPLE_LABELs.
4679 * gimplify.h (gimple_add_tmp_var_fn): Declare.
4680 * gimplify.c (gimple_add_tmp_var_fn): New function.
4681 * gimple-expr.h (create_tmp_reg_fn): Declare.
4682 * gimple-expr.c (create_tmp_reg_fn): New function.
4683 * gimple-low.c (record_vars_into): Don't change cfun.
4684 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
4685 code generation without cfun.
4687 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
4689 PR bootstrap/60719
4690 * Makefile.in (install-driver): Fix shell scripting.
4692 2014-04-03 Cong Hou <congh@google.com>
4694 PR tree-optimization/60505
4695 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
4696 threshold of number of iterations below which no vectorization
4697 will be done.
4698 * tree-vect-loop.c (new_loop_vec_info):
4699 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
4700 * tree-vect-loop.c (vect_analyze_loop_operations):
4701 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
4702 * tree-vect-loop.c (vect_transform_loop):
4703 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
4704 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
4705 of iterations of the loop and see if we should build the epilogue.
4707 2014-04-03 Richard Biener <rguenther@suse.de>
4709 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
4710 (streamer_tree_cache_create): Adjust.
4711 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
4712 to allow optional nodes array.
4713 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
4714 (streamer_tree_cache_append): Likewise.
4715 (streamer_tree_cache_create): Create nodes array optionally
4716 as specified by parameter.
4717 * lto-streamer-out.c (create_output_block): Avoid maintaining
4718 the node array in the writer cache.
4719 (DFS_write_tree): Remove assertion.
4720 (produce_asm_for_decls): Free the out decl state hash table early.
4721 * lto-streamer-in.c (lto_data_in_create): Adjust for
4722 streamer_tree_cache_create prototype change.
4724 2014-04-03 Richard Biener <rguenther@suse.de>
4726 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
4727 set TREE_CHAIN to NULL_TREE.
4729 2014-04-03 Richard Biener <rguenther@suse.de>
4731 PR tree-optimization/60740
4732 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
4733 over all GIMPLE_COND operands.
4735 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
4737 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
4738 (Weffc++): Remove Scott's numbering, merge lists and reference
4739 Wnon-virtual-dtor.
4741 2014-04-03 Nick Clifton <nickc@redhat.com>
4743 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
4744 properly.
4746 2014-04-03 Martin Jambor <mjambor@suse.cz>
4748 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
4749 mention gcc_unreachable before failing.
4750 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
4751 removed symbols.
4753 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
4755 PR ipa/60659
4756 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
4757 inconsistent code and instead mark the context inconsistent.
4758 (possible_polymorphic_call_targets): For inconsistent contexts
4759 return empty complete list.
4761 2014-04-02 Anthony Green <green@moxielogic.com>
4763 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
4764 (extendqisi2, extendhisi2): Define.
4765 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
4766 (WCHAR_TYPE): Change to unsigned int.
4768 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4770 PR tree-optimization/60733
4771 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
4772 insertion point for PHI candidates to be the end of the feeding
4773 block for the PHI argument.
4775 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
4777 PR rtl-optimization/60650
4778 * lra-constraints.c (process_alt_operands): Decrease reject for
4779 earlyclobber matching.
4781 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4783 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
4785 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4787 * config/spu/spu.c (pad_bb): Do not crash when the last
4788 insn is CODE_FOR_blockage.
4790 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4792 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
4793 lies outside the target mode.
4795 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
4797 PR target/60735
4798 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
4799 software floating point or no floating point registers, do not
4800 allow any type in the FPRs. Eliminate a test for SPE SIMD types
4801 in GPRs that occurs after we tested for GPRs that would never be
4802 true.
4804 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
4805 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
4806 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
4807 specifically allow DDmode, since that does not use the SPE SIMD
4808 instructions.
4810 2014-04-02 Richard Biener <rguenther@suse.de>
4812 PR middle-end/60729
4813 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
4814 MODE_INTs. Properly use negv_optab.
4815 (expand_abs): Likewise.
4817 2014-04-02 Richard Biener <rguenther@suse.de>
4819 PR bootstrap/60719
4820 * Makefile.in (install-driver): Guard extra installs with special
4821 names properly.
4823 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
4825 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4826 Document vec_vgbbd.
4828 2014-04-01 Richard Henderson <rth@redhat.com>
4830 PR target/60704
4831 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
4832 alternative enabled before register allocation.
4834 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
4836 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
4837 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
4838 typo.
4839 (nios2_large_got_address): Remove unneeded 'sym' parameter.
4840 (nios2_got_address): Update nios2_large_got_address call site.
4841 (nios2_delegitimize_address): New function.
4842 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
4843 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
4844 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
4846 2014-04-01 Martin Husemann <martin@duskware.de>
4848 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
4849 for -mabi=32.
4851 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
4853 PR rtl-optimization/60604
4854 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
4855 check from register_operand.
4856 (register_operand): Redefine in terms of general_operand.
4857 (nonmemory_operand): Use register_operand for the non-constant cases.
4859 2014-04-01 Richard Biener <rguenther@suse.de>
4861 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
4863 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
4865 * doc/invoke.texi (mapp-regs): Clarify.
4867 2014-03-31 Ulrich Drepper <drepper@gmail.com>
4869 * config/i386/avx512fintrin.h (__v32hi): Define type.
4870 (__v64qi): Likewise.
4871 (_mm512_set1_epi8): Define.
4872 (_mm512_set1_epi16): Define.
4873 (_mm512_set4_epi32): Define.
4874 (_mm512_set4_epi64): Define.
4875 (_mm512_set4_pd): Define.
4876 (_mm512_set4_ps): Define.
4877 (_mm512_setr4_epi64): Define.
4878 (_mm512_setr4_epi32): Define.
4879 (_mm512_setr4_pd): Define.
4880 (_mm512_setr4_ps): Define.
4881 (_mm512_setzero_epi32): Define.
4883 2014-03-31 Martin Jambor <mjambor@suse.cz>
4885 PR middle-end/60647
4886 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
4887 callsite_arguments_match_p. Updated all callers. Also check types of
4888 corresponding formal parameters and actual arguments.
4889 (not_all_callers_have_enough_arguments_p) Renamed to
4890 some_callers_have_mismatched_arguments_p.
4892 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
4894 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
4896 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
4898 PR target/60034
4899 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
4900 section anchor.
4902 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
4904 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
4905 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
4906 Split out
4907 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
4908 Use FMAMODE_NOVF512 mode iterator.
4909 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
4910 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
4911 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
4912 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
4913 Split out
4914 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
4915 Use VF_128_256 mode iterator.
4916 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
4917 Ditto.
4919 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4921 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
4922 static chain if needed.
4924 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
4926 PR target/60697
4927 * lra-constraints.c (index_part_to_reg): New.
4928 (process_address): Use it.
4930 2014-03-27 Jeff Law <law@redhat.com>
4931 Jakub Jelinek <jakub@redhat.com>
4933 PR target/60648
4934 * expr.c (do_tablejump): Use simplify_gen_binary rather than
4935 gen_rtx_{PLUS,MULT} to build up the address expression.
4937 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
4938 creating non-canonical RTL.
4940 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4942 PR ipa/60243
4943 * ipa-inline.c (want_inline_small_function_p): Short circuit large
4944 functions; reorganize to make cheap checks first.
4945 (inline_small_functions): Do not estimate growth when dumping;
4946 it is expensive.
4947 * ipa-inline.h (inline_summary): Add min_size.
4948 (growth_likely_positive): New function.
4949 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
4950 (set_cond_stmt_execution_predicate): Cleanup.
4951 (estimate_edge_size_and_time): Compute min_size.
4952 (estimate_calls_size_and_time): Likewise.
4953 (estimate_node_size_and_time): Likewise.
4954 (inline_update_overall_summary): Update min_size.
4955 (do_estimate_edge_time): Likewise.
4956 (do_estimate_edge_size): Update.
4957 (do_estimate_edge_hints): Update.
4958 (growth_likely_positive): New function.
4960 2014-03-28 Jakub Jelinek <jakub@redhat.com>
4962 PR target/60693
4963 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
4964 also if addr has VOIDmode.
4966 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4968 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
4969 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
4970 Declare extern.
4971 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
4972 instructions as well as AdvancedSIMD loads.
4974 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4976 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
4977 Use crypto_aese type.
4978 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
4979 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
4980 crypto_aese, crypto_aesmc. Move to types.md.
4981 * config/arm/types.md (crypto_aes): Split into crypto_aese,
4982 crypto_aesmc.
4983 * config/arm/iterators.md (crypto_type): Likewise.
4985 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4987 * cgraph.c: Include expr.h and tree-dfa.h.
4988 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
4989 remove LHS.
4991 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
4993 PR target/60675
4994 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
4995 regs from checking multi-reg pseudos.
4997 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4999 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
5001 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5003 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
5004 if it would clobber the stack pointer, even temporarily.
5006 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
5008 * mode-switching.c: Make small adjustments to the top comment.
5010 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
5012 * config/rs6000/constraints.md (wD constraint): New constraint to
5013 match the constant integer to get the top DImode/DFmode out of a
5014 vector in a VSX register.
5016 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
5017 match the constant integer to get the top DImode/DFmode out of a
5018 vector in a VSX register.
5020 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
5021 for ISA 2.07.
5023 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
5024 vbpermq builtins.
5026 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
5027 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
5029 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
5030 Optimize vec_extract of 64-bit values, where the value being
5031 extracted is in the top word, where we can use scalar
5032 instructions. Add direct move and store support. Combine the big
5033 endian/little endian vector select load support into a single insn.
5034 (vsx_extract_<mode>_internal1): Likewise.
5035 (vsx_extract_<mode>_internal2): Likewise.
5036 (vsx_extract_<mode>_load): Likewise.
5037 (vsx_extract_<mode>_store): Likewise.
5038 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
5039 combined into vsx_extract_<mode>_load.
5040 (vsx_extract_<mode>_one_le): Likewise.
5042 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
5043 define the top 64-bit vector element.
5045 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
5046 constraint.
5048 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5049 Document vec_vbpermq builtin.
5051 PR target/60672
5052 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
5053 enable use of xxsldwi and xxpermdi builtin functions.
5054 (vec_xxpermdi): Likewise.
5056 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5057 Document use of vec_xxsldwi and vec_xxpermdi builtins.
5059 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
5061 PR rtl-optimization/60650
5062 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
5063 first_p. Use it.
5064 (find_spills_for): New.
5065 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
5066 Spill all pseudos on the second iteration.
5068 2014-03-27 Marek Polacek <polacek@redhat.com>
5070 PR c/50347
5071 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
5072 types.
5074 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5076 * config/s390/s390.c (s390_can_use_return_insn): Check for
5077 call-saved FPRs on 31 bit.
5079 2014-03-27 Jakub Jelinek <jakub@redhat.com>
5081 PR middle-end/60682
5082 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
5083 if they need regimplification, just drop them instead of
5084 calling gimple_regimplify_operands on them.
5086 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
5088 PR target/60580
5089 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
5090 (aarch64_frame_pointer_required): Adjust logic.
5091 (aarch64_can_eliminate): Adjust logic.
5092 (aarch64_override_options_after_change): Adjust logic.
5094 2014-03-27 Dehao Chen <dehao@google.com>
5096 * ipa-inline.c (early_inliner): Update node's inline info.
5098 2014-03-26 Dehao Chen <dehao@google.com>
5100 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
5101 compiler inserted conditional jumps for NAN float check.
5103 2014-03-26 Jakub Jelinek <jakub@redhat.com>
5105 * ubsan.h (ubsan_create_data): Change second argument's type
5106 to const location_t *.
5107 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
5108 _("<unknown>").
5109 (ubsan_create_data): Change second argument to const location_t *PLOC.
5110 Create Loc field whenever PLOC is non-NULL.
5111 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
5112 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
5113 callers.
5115 PR other/59545
5116 * real.c (real_to_integer2): Change type of low to UHWI.
5118 2014-03-26 Tobias Burnus <burnus@net-b.de>
5120 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
5121 (CILK_SELF_SPECS): New define.
5122 (driver_self_specs): Use it.
5124 2014-03-26 Richard Biener <rguenther@suse.de>
5126 * tree-pretty-print.c (percent_K_format): Implement special
5127 case for LTO and its stripped down BLOCK tree.
5129 2014-03-26 Jakub Jelinek <jakub@redhat.com>
5131 PR sanitizer/60636
5132 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
5134 * tree-vrp.c (simplify_internal_call_using_ranges): If only
5135 one range is range_int_cst_p, but not both, at least optimize
5136 addition/subtraction of 0 and multiplication by 0 or 1.
5137 * gimple-fold.c (gimple_fold_call): Fold
5138 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
5139 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
5140 INTEGER_CSTs, try to fold at least x * 0 and y - y.
5142 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
5144 PR rtl-optimization/60452
5145 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
5146 <case REG>: Return 1 for invalid offsets from the frame pointer.
5148 2014-03-26 Marek Polacek <polacek@redhat.com>
5150 PR c/37428
5151 * doc/extend.texi (C Extensions): Mention variable-length arrays in
5152 a structure/union.
5154 2014-03-26 Marek Polacek <polacek@redhat.com>
5156 PR c/39525
5157 * doc/extend.texi (Designated Inits): Describe what happens to omitted
5158 field members.
5160 2014-03-26 Marek Polacek <polacek@redhat.com>
5162 PR other/59545
5163 * ira-color.c (update_conflict_hard_regno_costs): Perform the
5164 multiplication in unsigned type.
5166 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
5168 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
5170 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
5172 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
5174 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
5176 PR ipa/60315
5177 * cif-code.def (UNREACHABLE) New code.
5178 * ipa-inline.c (inline_small_functions): Skip edges to
5179 __builtlin_unreachable.
5180 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
5181 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
5182 predicate to __bulitin_unreachable.
5183 (set_cond_stmt_execution_predicate): Fix issue when
5184 invert_tree_comparison returns ERROR_MARK.
5185 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
5186 propagate to inline clones.
5187 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
5188 to unreachable.
5189 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
5190 * cgraphclones.c (cgraph_clone_node): If call destination is already
5191 ureachable, do not redirect it back.
5192 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
5193 unreachable.
5195 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
5197 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
5198 Do not modify inline clones.
5200 2014-03-25 Jakub Jelinek <jakub@redhat.com>
5202 * config/i386/i386.md (general_sext_operand): New mode attr.
5203 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
5204 don't generate (sign_extend (const_int)).
5205 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
5206 operands[2]. Use We constraint instead of <i> and
5207 <general_sext_operand> predicate instead of <general_operand>.
5208 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
5209 * config/i386/constraints.md (We): New constraint.
5210 * config/i386/predicates.md (x86_64_sext_operand,
5211 sext_operand): New predicates.
5213 2014-03-25 Martin Jambor <mjambor@suse.cz>
5215 PR ipa/60600
5216 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
5217 inconsistent devirtualizations to __builtin_unreachable.
5219 2014-03-25 Marek Polacek <polacek@redhat.com>
5221 PR c/35449
5222 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
5224 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
5226 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
5227 order of elements for big-endian.
5229 2014-03-25 Richard Biener <rguenther@suse.de>
5231 PR middle-end/60635
5232 * gimplify-me.c (gimple_regimplify_operands): Update the
5233 re-gimplifed stmt.
5235 2014-03-25 Martin Jambor <mjambor@suse.cz>
5237 PR ipa/59176
5238 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
5239 (lto_output_varpool_node): Likewise.
5240 (input_overwrite_node): Likewise.
5241 (input_varpool_node): Likewise.
5243 2014-03-25 Richard Biener <rguenther@suse.de>
5245 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
5246 (run_gcc): Likewise.
5248 2014-03-25 Jakub Jelinek <jakub@redhat.com>
5250 * combine.c (simplify_compare_const): Add MODE argument.
5251 Handle mode_width 0 as very large mode_width.
5252 (try_combine, simplify_comparison): Adjust callers.
5254 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
5255 type to avoid signed integer overflow.
5256 * explow.c (plus_constant): Likewise.
5258 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
5260 * doc/generic.texi: Correct typos.
5262 2014-03-24 Tobias Burnus <burnus@net-b.de>
5264 * doc/invoke.texi (-flto): Expand section about
5265 using static libraries with LTO.
5267 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5269 PR rtl-optimization/60501
5270 * optabs.def (addptr3_optab): New optab.
5271 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
5272 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
5273 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
5275 * lra.c (emit_add3_insn): Use the addptr pattern if available.
5277 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
5279 2014-03-24 Ulrich Drepper <drepper@gmail.com>
5281 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
5282 _mm512_set1_pd.
5284 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
5285 (_mm256_undefined_ps): Define.
5286 (_mm256_undefined_pd): Define.
5287 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
5288 (_mm_undefined_pd): Define.
5289 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
5290 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
5291 (_mm512_undefined_ps): Define.
5292 (_mm512_undefined_pd): Define.
5293 Use _mm*_undefined_*.
5294 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
5296 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
5298 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
5299 (lshr_simd): DI mode added.
5300 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
5301 (aarch64_ushr_simddi): Likewise.
5302 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
5303 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
5304 (vshrd_n_u64): Likewise.
5306 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5308 * Makefile.in (s-macro_list): Depend on cc1.
5310 2014-03-23 Teresa Johnson <tejohnson@google.com>
5312 * ipa-utils.c (ipa_print_order): Use specified dump file.
5314 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
5316 PR rtl-optimization/60601
5317 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
5319 * gcc.c (eval_spec_function): Initialize save_growing_value.
5321 2014-03-22 Jakub Jelinek <jakub@redhat.com>
5323 PR sanitizer/60613
5324 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
5325 code == MINUS_EXPR, never swap op0 with op1.
5327 * toplev.c (init_local_tick): Avoid signed integer multiplication
5328 overflow.
5329 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
5330 shift by first operand's bitsize.
5332 2014-03-21 Jakub Jelinek <jakub@redhat.com>
5334 PR target/60610
5335 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
5336 redefine to 1 or 0.
5337 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
5338 TARGET_ISA_64BIT_P(x).
5340 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5342 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
5343 pattern for vector nor instead of subtract from splat(-1).
5344 (altivec_expand_vec_perm_const_le): Likewise.
5346 2014-03-21 Richard Henderson <rth@twiddle.net>
5348 PR target/60598
5349 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
5350 related insns after epilogue_completed.
5352 2014-03-21 Martin Jambor <mjambor@suse.cz>
5354 PR ipa/59176
5355 * cgraph.h (symtab_node): New flag body_removed.
5356 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
5357 when removing bodies.
5358 * symtab.c (dump_symtab_base): Dump body_removed flag.
5359 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
5360 had their bodies removed.
5362 2014-03-21 Martin Jambor <mjambor@suse.cz>
5364 PR ipa/60419
5365 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
5366 in the border.
5368 2014-03-21 Richard Biener <rguenther@suse.de>
5370 PR tree-optimization/60577
5371 * tree-core.h (struct tree_base): Document nothrow_flag use
5372 in VAR_DECL_NONALIASED.
5373 * tree.h (VAR_DECL_NONALIASED): New.
5374 (may_be_aliased): Adjust.
5375 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
5377 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
5379 * expr.c (expand_expr_real_1): Remove outdated comment.
5381 2014-03-20 Jakub Jelinek <jakub@redhat.com>
5383 PR middle-end/60597
5384 * ira.c (adjust_cleared_regs): Call copy_rtx on
5385 *reg_equiv[REGNO (loc)].src_p before passing it to
5386 simplify_replace_fn_rtx.
5388 PR target/60568
5389 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
5390 into CONST, put pic register as first operand of PLUS. Use
5391 gen_const_mem for both 32-bit and 64-bit PIC got loads.
5393 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5395 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
5397 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
5399 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
5400 around for store forwarding issue in the FPU on the UT699.
5401 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
5402 loads and operations if -mfix-ut699 is specified.
5403 (divtf3_hq): Tweak attribute.
5404 (sqrttf2_hq): Likewise.
5406 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
5408 * calls.c (store_one_arg): Remove incorrect const qualification on the
5409 type of the temporary.
5410 * cfgexpand.c (expand_return): Likewise.
5411 * expr.c (expand_constructor): Likewise.
5412 (expand_expr_real_1): Likewise.
5414 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
5416 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
5417 of parts.
5419 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
5421 PR target/60039
5422 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
5424 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
5426 * config/arm/aarch-common-protos.h
5427 (alu_cost_table): Fix spelling of "extend".
5428 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
5430 2014-03-19 Richard Biener <rguenther@suse.de>
5432 PR middle-end/60553
5433 * tree-core.h (tree_type_common): Re-order pointer members
5434 to reduce recursion depth during GC walks.
5436 2014-03-19 Marek Polacek <polacek@redhat.com>
5438 PR sanitizer/60569
5439 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
5440 before accessing it.
5442 2014-03-19 Richard Biener <rguenther@suse.de>
5444 PR lto/59543
5445 * lto-streamer-in.c (input_function): In WPA stage do not drop
5446 debug stmts.
5448 2014-03-19 Jakub Jelinek <jakub@redhat.com>
5450 PR tree-optimization/60559
5451 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
5452 with build_zero_cst assignment.
5454 2014-03-18 Kai Tietz <ktietz@redhat.com>
5456 PR rtl-optimization/56356
5457 * sdbout.c (sdbout_parms): Verify that parms'
5458 incoming argument is valid.
5459 (sdbout_reg_parms): Likewise.
5461 2014-03-18 Richard Henderson <rth@redhat.com>
5463 PR target/60562
5464 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
5465 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
5466 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
5468 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
5470 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
5471 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
5472 Italicize plugin event names in description. Explain that
5473 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
5474 Remind that no GCC functions should be called after PLUGIN_FINISH.
5475 Explain what pragmas with expansion are.
5477 2014-03-18 Martin Liska <mliska@suse.cz>
5479 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
5480 gimple call statement is update.
5481 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
5482 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
5484 2014-03-18 Jakub Jelinek <jakub@redhat.com>
5486 PR sanitizer/60557
5487 * ubsan.c (ubsan_instrument_unreachable): Call
5488 initialize_sanitizer_builtins.
5489 (ubsan_pass): Likewise.
5491 PR sanitizer/60535
5492 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
5493 varpool_finalize_decl instead of rest_of_decl_compilation.
5495 2014-03-18 Richard Biener <rguenther@suse.de>
5497 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
5498 by using bitmap_and_compl instead of bitmap_and_compl_into.
5499 (df_rd_transfer_function): Likewise.
5501 2014-03-18 Richard Biener <rguenther@suse.de>
5503 * doc/lto.texi (fresolution): Fix typo.
5505 2014-03-18 Richard Biener <rguenther@suse.de>
5507 * doc/invoke.texi (flto): Update for changes in 4.9.
5509 2014-03-18 Richard Biener <rguenther@suse.de>
5511 * doc/loop.texi: Remove section on the removed lambda framework.
5512 Update loop docs with recent changes in preserving loop structure.
5514 2014-03-18 Richard Biener <rguenther@suse.de>
5516 * doc/lto.texi (-fresolution): Document.
5518 2014-03-18 Richard Biener <rguenther@suse.de>
5520 * doc/contrib.texi: Adjust my name.
5522 2014-03-18 Jakub Jelinek <jakub@redhat.com>
5524 PR ipa/58721
5525 * internal-fn.c: Include diagnostic-core.h.
5526 (expand_BUILTIN_EXPECT): New function.
5527 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
5528 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
5529 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
5530 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
5531 IFN_BUILTIN_EXPECT.
5532 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
5533 Revert 3 argument __builtin_expect code.
5534 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
5535 * gimple-fold.c (gimple_fold_call): Likewise.
5536 * tree.h (fold_builtin_expect): New prototype.
5537 * builtins.c (build_builtin_expect_predicate): Add predictor
5538 argument, if non-NULL, create 3 argument __builtin_expect.
5539 (fold_builtin_expect): No longer static. Add ARG2 argument,
5540 pass it through to build_builtin_expect_predicate.
5541 (fold_builtin_2): Adjust caller.
5542 (fold_builtin_3): Handle BUILT_IN_EXPECT.
5543 * internal-fn.def (BUILTIN_EXPECT): New.
5545 2014-03-18 Tobias Burnus <burnus@net-b.de>
5547 PR ipa/58721
5548 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
5549 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
5550 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
5552 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
5554 PR ipa/58721
5555 * predict.c (combine_predictions_for_bb): Fix up formatting.
5556 (expr_expected_value_1, expr_expected_value): Add predictor argument,
5557 fill what it points to if non-NULL.
5558 (tree_predict_by_opcode): Adjust caller, use the predictor.
5559 * predict.def (PRED_COMPARE_AND_SWAP): Add.
5561 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
5563 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
5564 proper constant for the store mode.
5566 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
5568 * symtab.c (change_decl_assembler_name): Fix transparent alias
5569 chain construction.
5571 2014-03-16 Renlin Li <Renlin.Li@arm.com>
5573 * config/aarch64/aarch64.c: Correct the comments about the
5574 aarch64 stack layout.
5576 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
5578 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
5579 check for GF_OMP_FOR_KIND_FOR.
5581 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
5583 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
5584 ymm and zmm register names.
5586 2014-03-17 Jakub Jelinek <jakub@redhat.com>
5588 PR target/60516
5589 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
5590 note creation for the 2010-08-31 changes.
5592 2014-03-17 Marek Polacek <polacek@redhat.com>
5594 PR middle-end/60534
5595 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
5596 as -fno-tree-loop-vectorize.
5597 (expand_omp_simd): Likewise.
5599 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
5601 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
5602 (eligible_for_call_delay): New prototype.
5603 * config/sparc/sparc.c (tls_call_delay): Rename into...
5604 (eligible_for_call_delay): ...this. Return false if the instruction
5605 cannot be put in the delay slot of a branch.
5606 (eligible_for_restore_insn): Simplify.
5607 (eligible_for_return_delay): Return false if the instruction cannot be
5608 put in the delay slot of a branch and simplify.
5609 (eligible_for_sibcall_delay): Return false if the instruction cannot be
5610 put in the delay slot of a branch.
5611 * config/sparc/sparc.md (fix_ut699): New attribute.
5612 (tls_call_delay): Delete.
5613 (in_call_delay): Reimplement.
5614 (eligible_for_sibcall_delay): Rename into...
5615 (in_sibcall_delay): ...this.
5616 (eligible_for_return_delay): Rename into...
5617 (in_return_delay): ...this.
5618 (in_branch_delay): Reimplement.
5619 (in_uncond_branch_delay): Delete.
5620 (in_annul_branch_delay): Delete.
5622 2014-03-14 Richard Henderson <rth@redhat.com>
5624 PR target/60525
5625 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
5626 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
5627 (*floathi<X87MODEF>2_i387_with_temp): Remove.
5628 (floathi splitters): Remove.
5629 (float<SWI48x>xf2): New pattern.
5630 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
5631 code that tried to handle DImode for 32-bit, but which was excluded
5632 by the pattern's condition. Drop allocation of stack temporary.
5633 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
5634 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
5635 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
5636 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
5637 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
5638 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
5639 (*float<SWI48><MODEF>2_sse_interunit): Remove.
5640 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
5641 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
5642 (*float<SWI48x><X87MODEF>2_i387): Remove.
5643 (all float _with_temp splitters): Remove.
5644 (*float<SWI48x><MODEF>2_i387): New pattern.
5645 (*float<SWI48><MODEF>2_sse): New pattern.
5646 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
5647 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
5649 2014-03-14 Jakub Jelinek <jakub@redhat.com>
5650 Marek Polacek <polacek@redhat.com>
5652 PR middle-end/60484
5653 * common.opt (dump_base_name_prefixed): New Variable.
5654 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
5655 if x_dump_base_name_prefixed is already set, set it at the end.
5657 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
5659 PR rtl-optimization/60508
5660 * lra-constraints.c (get_reload_reg): Add new parameter
5661 in_subreg_p.
5662 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
5663 Pass the new parameter values.
5665 2014-03-14 Richard Biener <rguenther@suse.de>
5667 * common.opt: Revert unintented changes from r205065.
5668 * opts.c: Likewise.
5670 2014-03-14 Richard Biener <rguenther@suse.de>
5672 PR middle-end/60518
5673 * cfghooks.c (split_block): Properly adjust all loops the
5674 block was a latch of.
5676 2014-03-14 Martin Jambor <mjambor@suse.cz>
5678 PR lto/60461
5679 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
5680 and simplify it.
5682 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
5684 PR target/59396
5685 * config/avr/avr.c (avr_set_current_function): Pass function name
5686 through default_strip_name_encoding before sanity checking instead
5687 of skipping the first char of the assembler name.
5689 2014-03-13 Richard Henderson <rth@redhat.com>
5691 PR debug/60438
5692 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
5693 (ix86_force_to_memory, ix86_free_from_memory): Remove.
5694 * config/i386/i386-protos.h: Likewise.
5695 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
5696 in the expander instead of a splitter.
5697 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
5698 any possibility of requiring a memory.
5699 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
5700 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
5701 (fp branch splitters): Update for ix86_split_fp_branch.
5702 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
5703 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
5704 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
5705 (*fop_<MODEF>_2_i387): Remove f/r alternative.
5706 (*fop_<MODEF>_3_i387): Likewise.
5707 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
5708 (splitters for the fop_* register patterns): Remove.
5709 (fscalexf4_i387): Rename from *fscalexf4_i387.
5710 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
5712 2014-03-13 Jakub Jelinek <jakub@redhat.com>
5714 PR tree-optimization/59779
5715 * tree-dfa.c (get_ref_base_and_extent): Use double_int
5716 type for bitsize and maxsize instead of HOST_WIDE_INT.
5718 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
5720 PR rtl-optimization/57320
5721 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
5722 the CFG after thread_prologue_and_epilogue_insns.
5724 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
5726 PR rtl-optimization/57189
5727 * lra-constraints.c (process_alt_operands): Disfavor spilling
5728 vector pseudos.
5730 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
5732 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
5734 2014-03-13 Jakub Jelinek <jakub@redhat.com>
5736 PR tree-optimization/59025
5737 PR middle-end/60418
5738 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
5739 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
5741 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
5743 PR target/60486
5744 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
5745 calls of avr_out_plus_1.
5747 2014-03-13 Bin Cheng <bin.cheng@arm.com>
5749 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
5750 BB's single pred and update the father loop's latch info later.
5752 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
5754 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
5755 (VEC_M): Likewise.
5756 (VEC_N): Likewise.
5757 (VEC_R): Likewise.
5758 (VEC_base): Likewise.
5759 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
5760 registers, we need to swap double words in little endian mode.
5762 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
5763 to be a container mode for 128-bit integer operations added in ISA
5764 2.07. Unlike TImode and PTImode, the preferred register set is
5765 the Altivec/VMX registers for the 128-bit operations.
5767 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
5768 declarations.
5769 (rs6000_split_128bit_ok_p): Likewise.
5771 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
5772 macros for creating ISA 2.07 normal and overloaded builtin
5773 functions with 3 arguments.
5774 (BU_P8V_OVERLOAD_3): Likewise.
5775 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
5776 for use as overloaded functions.
5777 (VPERM_1TI_UNS): Likewise.
5778 (VSEL_1TI): Likewise.
5779 (VSEL_1TI_UNS): Likewise.
5780 (ST_INTERNAL_1ti): Likewise.
5781 (LD_INTERNAL_1ti): Likewise.
5782 (XXSEL_1TI): Likewise.
5783 (XXSEL_1TI_UNS): Likewise.
5784 (VPERM_1TI): Likewise.
5785 (VPERM_1TI_UNS): Likewise.
5786 (XXPERMDI_1TI): Likewise.
5787 (SET_1TI): Likewise.
5788 (LXVD2X_V1TI): Likewise.
5789 (STXVD2X_V1TI): Likewise.
5790 (VEC_INIT_V1TI): Likewise.
5791 (VEC_SET_V1TI): Likewise.
5792 (VEC_EXT_V1TI): Likewise.
5793 (EQV_V1TI): Likewise.
5794 (NAND_V1TI): Likewise.
5795 (ORC_V1TI): Likewise.
5796 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
5797 added in ISA 2.07. Add both normal 'altivec' builtins, and the
5798 overloaded builtin.
5799 (VADDUQM): Likewise.
5800 (VSUBCUQ): Likewise.
5801 (VADDEUQM): Likewise.
5802 (VADDECUQ): Likewise.
5803 (VSUBEUQM): Likewise.
5804 (VSUBECUQ): Likewise.
5806 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
5807 __int128_t and __uint128_t types.
5808 (__uint128_type): Likewise.
5809 (altivec_categorize_keyword): Add support for vector __int128_t,
5810 vector __uint128_t, vector __int128, and vector unsigned __int128
5811 as a container type for TImode operations that need to be done in
5812 VSX/Altivec registers.
5813 (rs6000_macro_to_expand): Likewise.
5814 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
5815 to support 128-bit integer instructions vaddcuq, vadduqm,
5816 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
5817 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
5819 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
5820 for V1TImode, and set up preferences to use VSX/Altivec registers.
5821 Setup VSX reload handlers.
5822 (rs6000_debug_reg_global): Likewise.
5823 (rs6000_init_hard_regno_mode_ok): Likewise.
5824 (rs6000_preferred_simd_mode): Likewise.
5825 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
5826 (easy_altivec_constant): Likewise.
5827 (output_vec_const_move): Likewise.
5828 (rs6000_expand_vector_set): Convert V1TImode set and extract to
5829 simple move.
5830 (rs6000_expand_vector_extract): Likewise.
5831 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
5832 addressing.
5833 (rs6000_const_vec): Add support for V1TImode.
5834 (rs6000_emit_le_vsx_load): Swap double words when loading or
5835 storing TImode/V1TImode.
5836 (rs6000_emit_le_vsx_store): Likewise.
5837 (rs6000_emit_le_vsx_move): Likewise.
5838 (rs6000_emit_move): Add support for V1TImode.
5839 (altivec_expand_ld_builtin): Likewise.
5840 (altivec_expand_st_builtin): Likewise.
5841 (altivec_expand_vec_init_builtin): Likewise.
5842 (altivec_expand_builtin): Likewise.
5843 (rs6000_init_builtins): Add support for V1TImode type. Add
5844 support for ISA 2.07 128-bit integer builtins. Define type names
5845 for the VSX/Altivec vector types.
5846 (altivec_init_builtins): Add support for overloaded vector
5847 functions with V1TImode type.
5848 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
5849 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
5850 external function.
5851 (rs6000_split_128bit_ok_p): Likewise.
5852 (rs6000_handle_altivec_attribute): Create V1TImode from vector
5853 __int128_t and vector __uint128_t.
5855 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
5856 and mode attributes.
5857 (VSX_M): Likewise.
5858 (VSX_M2): Likewise.
5859 (VSm): Likewise.
5860 (VSs): Likewise.
5861 (VSr): Likewise.
5862 (VSv): Likewise.
5863 (VS_scalar): Likewise.
5864 (VS_double): Likewise.
5865 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
5867 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
5868 we support the ISA 2.07 128-bit integer arithmetic instructions.
5869 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
5870 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
5871 and TImode types for use with the builtin functions.
5872 (V1TI_type_node): Likewise.
5873 (unsigned_V1TI_type_node): Likewise.
5874 (intTI_type_internal_node): Likewise.
5875 (uintTI_type_internal_node): Likewise.
5877 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
5878 128-bit builtin functions.
5879 (UNSPEC_VADDEUQM): Likewise.
5880 (UNSPEC_VADDECUQ): Likewise.
5881 (UNSPEC_VSUBCUQ): Likewise.
5882 (UNSPEC_VSUBEUQM): Likewise.
5883 (UNSPEC_VSUBECUQ): Likewise.
5884 (VM): Add V1TImode to vector mode iterators.
5885 (VM2): Likewise.
5886 (VI_unit): Likewise.
5887 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
5888 (altivec_vaddcuq): Likewise.
5889 (altivec_vsubuqm): Likewise.
5890 (altivec_vsubcuq): Likewise.
5891 (altivec_vaddeuqm): Likewise.
5892 (altivec_vaddecuq): Likewise.
5893 (altivec_vsubeuqm): Likewise.
5894 (altivec_vsubecuq): Likewise.
5896 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
5897 mode iterators.
5898 (BOOL_128): Likewise.
5899 (BOOL_REGS_OUTPUT): Likewise.
5900 (BOOL_REGS_OP1): Likewise.
5901 (BOOL_REGS_OP2): Likewise.
5902 (BOOL_REGS_UNARY): Likewise.
5903 (BOOL_REGS_AND_CR0): Likewise.
5905 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
5906 128-bit integer builtin support.
5907 (vec_vadduqm): Likewise.
5908 (vec_vaddecuq): Likewise.
5909 (vec_vaddeuqm): Likewise.
5910 (vec_vsubecuq): Likewise.
5911 (vec_vsubeuqm): Likewise.
5912 (vec_vsubcuq): Likewise.
5913 (vec_vsubuqm): Likewise.
5915 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5916 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
5917 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
5918 128-bit integer add/subtract to ISA 2.07.
5920 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
5922 * config/arc/arc.c (arc_predicate_delay_insns):
5923 Fix third argument passed to conditionalize_nonjump.
5925 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
5927 * config/aarch64/aarch64-builtins.c
5928 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
5929 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
5930 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
5931 instead of __builtin_lfloor.
5932 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
5934 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5936 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
5937 (tree_ssa_ifcombine_bb_1): New function.
5938 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
5939 is an empty forwarder block to then_bb or vice versa and then_bb
5940 and else_bb are effectively swapped.
5942 2014-03-12 Christian Bruel <christian.bruel@st.com>
5944 PR target/60264
5945 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
5946 REG_CFA_DEF_CFA note.
5947 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
5948 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
5950 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
5952 PR tree-optimization/60454
5953 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
5955 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5957 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
5958 Do not define target_cpu_default2 to generic.
5959 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
5960 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
5961 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
5963 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5964 Marc Glisse <marc.glisse@inria.fr>
5966 PR tree-optimization/60502
5967 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
5968 instead of build_low_bits_mask.
5970 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5972 PR middle-end/60482
5973 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
5974 if there are multiple uses, but op doesn't live on E edge.
5975 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
5976 clobber stmts before __builtin_unreachable.
5978 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
5980 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
5981 hard_frame_pointer_rtx.
5982 * cse.c (cse_insn): Remove volatile check.
5983 * cselib.c (cselib_process_insn): Likewise.
5984 * dse.c (scan_insn): Likewise.
5986 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
5988 * config/arc/arc.c (conditionalize_nonjump): New function,
5989 broken out of ...
5990 (arc_ifcvt): ... this.
5991 (arc_predicate_delay_insns): Use it.
5993 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
5995 * config/arc/predicates.md (extend_operand): During/after reload,
5996 allow const_int_operand.
5997 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
5998 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
5999 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
6000 to "i".
6001 (umulsi3_highpart_i): Likewise.
6003 2014-03-11 Richard Biener <rguenther@suse.de>
6005 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
6006 Add asserts to guard possible wrong-code bugs.
6008 2014-03-11 Richard Biener <rguenther@suse.de>
6010 PR tree-optimization/60429
6011 PR tree-optimization/60485
6012 * tree-ssa-structalias.c (set_union_with_increment): Properly
6013 take into account all fields that overlap the shifted vars.
6014 (do_sd_constraint): Likewise.
6015 (do_ds_constraint): Likewise.
6016 (get_constraint_for_ptr_offset): Likewise.
6018 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
6020 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
6021 (nios2_compute_frame_layout):
6022 Add calculation of cfun->machine->fp_save_offset.
6023 (nios2_expand_prologue): Correct setting of frame pointer register
6024 in prologue.
6025 (nios2_expand_epilogue): Update recovery of stack pointer from
6026 frame pointer accordingly.
6027 (nios2_initial_elimination_offset): Update calculation of offset
6028 for eliminating to HARD_FRAME_POINTER_REGNUM.
6030 2014-03-10 Jakub Jelinek <jakub@redhat.com>
6032 PR ipa/60457
6033 * ipa.c (symtab_remove_unreachable_nodes): Don't call
6034 cgraph_get_create_node on VAR_DECLs.
6036 2014-03-10 Richard Biener <rguenther@suse.de>
6038 PR middle-end/60474
6039 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
6041 2014-03-08 Douglas B Rupp <rupp@gnat.com>
6043 * config/vms/vms.opt (vms_float_format): New variable.
6045 2014-03-08 Tobias Burnus <burnus@net-b.de>
6047 * doc/invoke.texi (-fcilkplus): Update implementation status.
6049 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
6050 Richard Biener <rguenther@suse.de>
6052 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
6053 consistently accross all TUs.
6054 (run_gcc): Enable -fshort-double automatically at link at link-time
6055 and disallow override.
6057 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
6059 PR target/58271
6060 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
6061 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
6062 if they can't be used.
6064 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6066 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
6067 for Solaris 11/x86 ld.
6068 * configure: Regenerate.
6070 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6072 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
6073 (LIB_TLS_SPEC): Save as ld_tls_libs.
6074 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
6075 (HAVE_AS_IX86_TLSLDM): New test.
6076 * configure, config.in: Regenerate.
6077 * config/i386/i386.c (legitimize_tls_address): Fall back to
6078 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
6079 cannot support TLS_MODEL_LOCAL_DYNAMIC.
6080 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
6081 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
6083 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
6085 * common.opt (fira-loop-pressure): Mark as optimization.
6087 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
6089 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
6090 an OpenMP mappable type.
6092 2014-03-06 Matthias Klose <doko@ubuntu.com>
6094 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
6095 MULTILIB_OSDIRNAMES is not defined.
6097 2014-03-06 Jakub Jelinek <jakub@redhat.com>
6098 Meador Inge <meadori@codesourcery.com>
6100 PR target/58595
6101 * config/arm/arm.c (arm_tls_symbol_p): Remove.
6102 (arm_legitimize_address): Call legitimize_tls_address for any
6103 arm_tls_referenced_p expression, handle constant addend. Call it
6104 before testing for !TARGET_ARM.
6105 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
6107 2014-03-06 Richard Biener <rguenther@suse.de>
6109 PR middle-end/60445
6110 PR lto/60424
6111 PR lto/60427
6112 Revert
6113 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
6115 * tree-streamer.c (record_common_node): Assert we don't record
6116 nodes with type double.
6117 (preload_common_node): Skip type double, complex double and double
6118 pointer since it is now frontend dependent due to fshort-double option.
6120 2014-03-06 Richard Biener <rguenther@suse.de>
6122 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
6123 or -fno-lto is specified and the linker has full plugin support.
6124 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
6125 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
6126 * lto-wrapper.c (merge_and_complain): Merge compile-time
6127 optimization levels.
6128 (run_gcc): And pass it through to the link options.
6130 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
6132 PR debug/60381
6133 Revert:
6134 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
6135 PR debug/59992
6136 * cselib.c (remove_useless_values): Skip to avoid quadratic
6137 behavior if the condition moved from...
6138 (cselib_process_insn): ... here holds.
6140 2014-03-05 Jakub Jelinek <jakub@redhat.com>
6142 PR plugins/59335
6143 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
6144 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
6146 PR plugins/59335
6147 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
6148 (TM_H): Add x86-tune.def.
6150 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6152 * config/aarch64/aarch64.c (generic_tunings):
6153 Use cortexa57_extra_costs.
6155 2014-03-05 Jakub Jelinek <jakub@redhat.com>
6157 PR lto/60404
6158 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
6159 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
6160 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
6161 cost for in_lto_p.
6163 2014-03-04 Heiher <r@hev.cc>
6165 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
6166 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
6168 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
6170 * config/i386/predicates.md (const2356_operand): Change to ...
6171 (const2367_operand): ... this.
6172 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
6173 const2367_operand.
6174 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6175 (*avx512pf_scatterpf<mode>sf): Ditto.
6176 (avx512pf_scatterpf<mode>df): Ditto.
6177 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6178 (*avx512pf_scatterpf<mode>df): Ditto.
6179 * config/i386/i386.c (ix86_expand_builtin): Update
6180 incorrect hint operand error message.
6182 2014-03-04 Richard Biener <rguenther@suse.de>
6184 * lto-section-in.c (lto_get_section_data): Fix const cast.
6186 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
6188 * tree-streamer.c (record_common_node): Assert we don't record
6189 nodes with type double.
6190 (preload_common_node): Skip type double, complex double and double
6191 pointer since it is now frontend dependent due to fshort-double option.
6193 2014-03-04 Richard Biener <rguenther@suse.de>
6195 PR lto/60405
6196 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
6197 (lto_input_toplevel_asms): Likewise.
6198 * lto-section-in.c (lto_get_section_data): Instead do it here
6199 for every section.
6201 2014-03-04 Richard Biener <rguenther@suse.de>
6203 PR tree-optimization/60382
6204 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
6205 dead PHIs a reduction.
6207 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
6209 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
6210 hint value.
6211 (_mm_prefetch): Move out of GCC target("sse") pragma.
6212 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
6213 GCC target("prfchw") pragma.
6214 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
6215 for locality <= 2.
6216 * config/i386/i386.c (ix86_option_override_internal): Enable
6217 -mprfchw with -mprefetchwt1.
6219 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
6221 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
6222 Mark as varying.
6224 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
6226 * opts.h (CL_PCH_IGNORE): Define.
6227 * targhooks.c (option_affects_pch_p):
6228 Return false for options that have CL_PCH_IGNORE set.
6229 * opt-functions.awk: Process PchIgnore.
6230 * doc/options.texi: Document PchIgnore.
6232 * config/arc/arc.opt (misize): Add PchIgnore property.
6234 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6236 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
6237 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
6238 constraint on constants to permit them being loaded into
6239 GENERAL_REGS or BASE_REGS.
6241 2014-03-03 Nick Clifton <nickc@redhat.com>
6243 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
6244 anti-cacnonical alternatives.
6245 (negandhi3_real): New pattern.
6246 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
6248 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
6250 * config/avr/avr-mcus.def: Remove atxmega16x1.
6251 * config/avr/avr-tables.opt: Regenerate.
6252 * config/avr/t-multilib: Regenerate.
6253 * doc/avr-mmcu.texi: Regenerate.
6255 2014-03-03 Tobias Grosser <tobias@grosser.es>
6256 Mircea Namolaru <mircea.namolaru@inria.fr>
6258 PR tree-optimization/58028
6259 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
6260 scalar dimensions.
6262 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6264 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
6265 not handled by recognizers.
6267 2014-03-03 Jakub Jelinek <jakub@redhat.com>
6269 PR middle-end/60175
6270 * function.c (expand_function_end): Don't emit
6271 clobber_return_register sequence if clobber_after is a BARRIER.
6272 * cfgexpand.c (construct_exit_block): Append instructions before
6273 return_label to prev_bb.
6275 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6277 * config/rs6000/constraints.md: Document reserved use of "wc".
6279 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
6281 PR ipa/60150
6282 * ipa.c (function_and_variable_visibility): When dissolving comdat
6283 group, also set all symbols to local.
6285 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
6287 PR ipa/60306
6289 Revert:
6290 2013-12-14 Jan Hubicka <jh@suse.cz>
6291 PR middle-end/58477
6292 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
6294 2014-03-02 Jon Beniston <jon@beniston.com>
6296 PR bootstrap/48230
6297 PR bootstrap/50927
6298 PR bootstrap/52466
6299 PR target/46898
6300 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
6301 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
6302 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
6303 (simple_return, *simple_return): New patterns
6304 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
6305 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
6307 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
6309 * dwarf2out.c (gen_subprogram_die): Tidy.
6311 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
6313 PR target/60071
6314 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
6315 (*mov_t_msb_neg_negc): ... this new insn.
6317 2014-02-28 Jason Merrill <jason@redhat.com>
6319 PR c++/58678
6320 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
6321 function.
6323 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
6325 PR c++/60314
6326 * dwarf2out.c (decltype_auto_die): New static.
6327 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
6328 (gen_type_die_with_usage): Handle 'decltype(auto)'.
6329 (is_cxx_auto): Likewise.
6331 2014-02-28 Ian Bolton <ian.bolton@arm.com>
6333 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
6334 we are not using general regs only.
6336 2014-02-28 Richard Biener <rguenther@suse.de>
6338 PR target/60280
6339 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
6340 previous fix and only allow to remove trivial pre-headers
6341 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
6342 (remove_forwarder_block): Properly update the latch of a loop.
6344 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
6346 PR debug/59992
6347 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
6348 (cselib_preserved_hash_table): New.
6349 (preserve_constants_and_equivs): Move preserved vals to it.
6350 (cselib_find_slot): Look it up first.
6351 (cselib_init): Initialize it.
6352 (cselib_finish): Release it.
6353 (dump_cselib_table): Dump it.
6355 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
6357 PR debug/59992
6358 * cselib.c (remove_useless_values): Skip to avoid quadratic
6359 behavior if the condition moved from...
6360 (cselib_process_insn): ... here holds.
6362 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
6364 PR debug/57232
6365 * var-tracking.c (vt_initialize): Apply the same condition to
6366 preserve the CFA base value.
6368 2014-02-28 Joey Ye <joey.ye@arm.com>
6370 PR target/PR60169
6371 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
6372 if reload in progress or completed.
6374 2014-02-28 Tobias Burnus <burnus@net-b.de>
6376 PR middle-end/60147
6377 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
6378 NAMELIST_DECL.
6380 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
6382 * doc/tm.texi.in (Condition Code Status): Update documention for
6383 relative locations of cc0-setter and cc0-user.
6385 2014-02-27 Jeff Law <law@redhat.com>
6387 PR rtl-optimization/52714
6388 * combine.c (try_combine): When splitting an unrecognized PARALLEL
6389 into two independent simple sets, if I3 is a jump, ensure the
6390 pattern we place into I3 is a (set (pc) ...).
6392 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
6393 Jeff Law <law@redhat.com>
6395 PR rtl-optimization/49847
6396 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
6397 are in different blocks.
6398 * doc/tm.texi (Condition Code Status): Update documention for
6399 relative locations of cc0-setter and cc0-user.
6401 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
6403 PR target/59222
6404 * lra.c (lra_emit_add): Check SUBREG too.
6406 2014-02-27 Andreas Schwab <schwab@suse.de>
6408 * config/m68k/m68k.c (m68k_option_override): Disable
6409 -flive-range-shrinkage for classic m68k.
6410 (m68k_override_options_after_change): Likewise.
6412 2014-02-27 Marek Polacek <polacek@redhat.com>
6414 PR middle-end/59223
6415 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
6416 -Wmaybe-uninitialized.
6418 2014-02-27 Alan Modra <amodra@gmail.com>
6420 PR target/57936
6421 * reload1.c (emit_input_reload_insns): When reload_override_in,
6422 set old to rl->in_reg when rl->in_reg is a subreg.
6424 2014-02-26 Richard Biener <rguenther@suse.de>
6426 PR bootstrap/60343
6427 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
6429 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
6431 * common/config/i386/predicates.md (const1256_operand): Remove.
6432 (const2356_operand): New.
6433 (const_1_to_2_operand): Remove.
6434 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
6435 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6436 (*avx512pf_gatherpf<mode>sf): Ditto.
6437 (avx512pf_gatherpf<mode>df): Ditto.
6438 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6439 (*avx512pf_gatherpf<mode>df): Ditto.
6440 (avx512pf_scatterpf<mode>sf): Ditto.
6441 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6442 (*avx512pf_scatterpf<mode>sf): Ditto.
6443 (avx512pf_scatterpf<mode>df): Ditto.
6444 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6445 (*avx512pf_scatterpf<mode>df): Ditto.
6446 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
6448 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
6450 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
6451 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
6452 (_mm512_mask_testn_epi64_mask): Move to ...
6453 * config/i386/avx512cdintrin.h: Here.
6454 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
6455 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
6456 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
6457 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
6458 TARGET_AVX512F from TARGET_AVX512CD.
6460 2014-02-26 Richard Biener <rguenther@suse.de>
6462 PR ipa/60327
6463 * ipa.c (walk_polymorphic_call_targets): Properly guard
6464 call to inline_update_overall_summary.
6466 2014-02-26 Bin Cheng <bin.cheng@arm.com>
6468 PR target/60280
6469 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
6470 and latches only if requested. Fix latch if it is removed.
6471 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
6472 LOOPS_HAVE_PREHEADERS.
6474 2014-02-25 Andrew Pinski <apinski@cavium.com>
6476 * builtins.c (expand_builtin_thread_pointer): Create a new target
6477 when the target is NULL.
6479 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
6481 PR rtl-optimization/60317
6482 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
6483 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
6484 * lra-assigns.c: Include params.h.
6485 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
6486 other reload pseudos considerations.
6488 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6490 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
6491 to use canonical form for nor<mode>3.
6493 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6495 PR target/55426
6496 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
6497 conversions.
6499 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
6501 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
6502 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
6503 (ix86_handle_option): Handle OPT_mprefetchwt1.
6504 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
6505 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
6506 PREFETCHWT1 CPUID.
6507 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
6508 OPTION_MASK_ISA_PREFETCHWT1.
6509 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
6510 (PTA_PREFETCHWT1): New.
6511 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
6512 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
6513 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
6514 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
6515 (*prefetch_avx512pf_<mode>_: Change into ...
6516 (*prefetch_prefetchwt1_<mode>: This.
6517 * config/i386/i386.opt (mprefetchwt1): New.
6518 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
6519 (_mm_prefetch): Handle intent to write.
6520 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
6522 2014-02-25 Richard Biener <rguenther@suse.de>
6524 PR middle-end/60291
6525 * emit-rtl.c (mem_attrs_htab): Remove.
6526 (mem_attrs_htab_hash): Likewise.
6527 (mem_attrs_htab_eq): Likewise.
6528 (set_mem_attrs): Always allocate new mem-attrs when something changed.
6529 (init_emit_once): Do not allocate mem_attrs_htab.
6531 2014-02-25 Richard Biener <rguenther@suse.de>
6533 PR lto/60319
6534 * lto-opts.c (lto_write_options): Output non-explicit conservative
6535 -fwrapv, -fno-trapv and -fno-strict-overflow.
6536 * lto-wrapper.c (merge_and_complain): Handle merging those options.
6537 (run_gcc): And pass them through.
6539 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
6541 * sel-sched.c (calculate_new_fences): New parameter ptime.
6542 Calculate it as a maximum over all fence cycles.
6543 (sel_sched_region_2): Adjust the call to calculate_new_fences.
6544 Print the final schedule timing when sched_verbose.
6546 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
6548 PR rtl-optimization/60292
6549 * sel-sched.c (fill_vec_av_set): Do not reset target availability
6550 bit fot the fence instruction.
6552 2014-02-24 Alangi Derick <alangiderick@gmail.com>
6554 * calls.h: Fix typo in comment.
6556 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
6558 * config/pa/pa.c (pa_output_move_double): Don't valididate when
6559 adjusting offsetable addresses.
6561 2014-02-24 Guozhi Wei <carrot@google.com>
6563 * sparseset.h (sparseset_pop): Fix the wrong index.
6565 2014-02-24 Walter Lee <walt@tilera.com>
6567 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
6568 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
6569 triplet.
6570 * common/config/tilegx/tilegx-common.c
6571 (TARGET_DEFAULT_TARGET_FLAGS): Define.
6572 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
6573 (LINK_SPEC): Ditto.
6574 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
6575 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
6576 (tilegx_gimplify_va_arg_expr): Handle big endian.
6577 (tilegx_expand_unaligned_load): Ditto.
6578 (tilegx_expand_unaligned_store): Ditto.
6579 (TARGET_RETURN_IN_MSB): New.
6580 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
6581 (TARGET_ENDIAN_DEFAULT): New.
6582 (TARGET_BIG_ENDIAN): Handle big endian.
6583 (BYTES_BIG_ENDIAN): Ditto.
6584 (WORDS_BIG_ENDIAN): Ditto.
6585 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
6586 (ENDIAN_SPEC): New.
6587 (EXTRA_SPECS): New.
6588 * config/tilegx/tilegx.md (extv): Handle big endian.
6589 (extzv): Ditto.
6590 (insn_st<n>): Ditto.
6591 (insn_st<n>_add<bitsuffix>): Ditto.
6592 (insn_stnt<n>): Ditto.
6593 (insn_stnt<n>_add<bitsuffix>):Ditto.
6594 (vec_interleave_highv8qi): Handle big endian.
6595 (vec_interleave_highv8qi_be): New.
6596 (vec_interleave_highv8qi_le): New.
6597 (insn_v1int_h): Handle big endian.
6598 (vec_interleave_lowv8qi): Handle big endian.
6599 (vec_interleave_lowv8qi_be): New.
6600 (vec_interleave_lowv8qi_le): New.
6601 (insn_v1int_l): Handle big endian.
6602 (vec_interleave_highv4hi): Handle big endian.
6603 (vec_interleave_highv4hi_be): New.
6604 (vec_interleave_highv4hi_le): New.
6605 (insn_v2int_h): Handle big endian.
6606 (vec_interleave_lowv4hi): Handle big endian.
6607 (vec_interleave_lowv4hi_be): New.
6608 (vec_interleave_lowv4hi_le): New.
6609 (insn_v2int_l): Handle big endian.
6610 (vec_interleave_highv2si): Handle big endian.
6611 (vec_interleave_highv2si_be): New.
6612 (vec_interleave_highv2si_le): New.
6613 (insn_v4int_h): Handle big endian.
6614 (vec_interleave_lowv2si): Handle big endian.
6615 (vec_interleave_lowv2si_be): New.
6616 (vec_interleave_lowv2si_le): New.
6617 (insn_v4int_l): Handle big endian.
6618 * config/tilegx/tilegx.opt (mbig-endian): New option.
6619 (mlittle-endian): New option.
6620 * doc/install.texi: Document tilegxbe-linux.
6621 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
6623 2014-02-24 Martin Jambor <mjambor@suse.cz>
6625 PR ipa/60266
6626 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
6627 there are no parameter descriptors.
6629 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
6631 PR rtl-optimization/60268
6632 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
6633 initialization to ...
6634 (sched_rgn_init): ... here.
6635 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
6637 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6639 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
6640 names.
6642 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
6644 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
6645 definition.
6647 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6649 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
6650 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
6652 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6654 * config/microblaze/predicates.md: Add cmp_op predicate.
6655 * config/microblaze/microblaze.md: Add branch_compare instruction
6656 which uses cmp_op predicate and emits cmp insn before branch.
6657 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
6658 to microblaze_expand_conditional_branch and consolidate logic.
6659 (microblaze_expand_conditional_branch): emit branch_compare
6660 insn instead of handling cmp op separate from branch insn.
6662 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6664 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
6665 to permit subregs.
6667 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6669 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
6670 define_insn with define_expand and new define_insn
6671 *altivec_lve<VI_char>x_internal.
6672 (altivec_stve<VI_char>x): Replace define_insn with define_expand
6673 and new define_insn *altivec_stve<VI_char>x_internal.
6674 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
6675 prototype.
6676 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
6677 lve*x built-ins.
6678 (altivec_expand_stvex_be): New function.
6680 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
6682 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
6683 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
6684 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
6685 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
6687 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
6689 PR target/60298
6690 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
6691 instead of emit_move_insn.
6693 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6695 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
6696 vspltw with vsldoi.
6697 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
6698 gen_altivec_vsumsws.
6700 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6702 * config/rs6000/altivec.md (altivec_lvxl): Rename as
6703 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
6704 (altivec_lvxl_<mode>): New define_expand incorporating
6705 -maltivec=be semantics where needed.
6706 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
6707 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
6708 semantics where needed.
6709 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
6710 (altivec_stvx_<mode>): New define_expand incorporating
6711 -maltivec=be semantics where needed.
6712 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
6713 VM2 iterator instead of V4SI.
6714 (altivec_stvxl_<mode>): New define_expand incorporating
6715 -maltivec=be semantics where needed.
6716 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
6717 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
6718 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
6719 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
6720 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
6721 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
6722 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
6723 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
6724 ALTIVEC_BUILTIN_STVXL.
6725 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
6726 (altivec_expand_stvx_be): Likewise.
6727 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
6728 (altivec_expand_lvx_be): Likewise.
6729 (altivec_expand_stvx_be): Likewise.
6730 (altivec_expand_builtin): Add cases for
6731 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
6732 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
6733 (altivec_init_builtins): Add definitions for
6734 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
6735 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
6737 2014-02-21 Catherine Moore <clm@codesourcery.com>
6739 * doc/invoke.texi (mvirt, mno-virt): Document.
6740 * config/mips/mips.opt (mvirt): New option.
6741 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
6743 2014-02-21 Richard Biener <rguenther@suse.de>
6745 PR tree-optimization/60276
6746 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
6747 (STMT_VINFO_MIN_NEG_DIST): New macro.
6748 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
6749 STMT_VINFO_MIN_NEG_DIST.
6750 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
6751 made for negative dependence distances still hold.
6753 2014-02-21 Richard Biener <rguenther@suse.de>
6755 PR middle-end/60291
6756 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
6757 DECL_INITIAL for globals not in the current function context.
6759 2014-02-21 Jakub Jelinek <jakub@redhat.com>
6761 PR tree-optimization/56490
6762 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
6763 * tree-ssa-uninit.c: Include params.h.
6764 (compute_control_dep_chain): Add num_calls argument, return false
6765 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
6766 num_calls to recursive call.
6767 (find_predicates): Change dep_chain into normal array,
6768 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
6769 variable and adjust compute_control_dep_chain caller.
6770 (find_def_preds): Likewise.
6772 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
6774 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
6775 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
6777 2014-02-21 Nick Clifton <nickc@redhat.com>
6779 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
6780 (pushhi1): Likewise.
6781 (popqi1): Add mode to pre_dec.
6782 (pophi1): Likewise.
6784 2014-02-21 Jakub Jelinek <jakub@redhat.com>
6786 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
6787 mode for mask of V8SFmode permutation.
6789 2014-02-20 Richard Henderson <rth@redhat.com>
6791 PR c++/60272
6792 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
6793 a new pseudo for OLDVAL.
6795 2014-02-20 Jakub Jelinek <jakub@redhat.com>
6797 PR target/57896
6798 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
6799 gen_reg_rtx if d->testing_p.
6800 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
6801 if d->testing_p and we will certainly return true.
6802 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
6803 if d->testing_p.
6805 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
6807 * emit-rtl.c (gen_reg_rtx): Assert that
6808 crtl->emit.regno_pointer_align_length is non-zero.
6810 2014-02-20 Richard Henderson <rth@redhat.com>
6812 PR c++/60272
6813 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
6814 on failure the store back into EXPECT.
6816 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
6817 Sandra Loosemore <sandra@codesourcery.com>
6819 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
6820 * config/nios2/nios2.c (nios2_function_profiler): Add
6821 -fPIC (flag_pic == 2) support.
6822 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
6823 (nios2_large_offset_p): New function.
6824 (nios2_unspec_reloc_p): Move up position, update to use
6825 nios2_large_offset_p.
6826 (nios2_unspec_address): Remove function.
6827 (nios2_unspec_offset): New function.
6828 (nios2_large_got_address): New function.
6829 (nios2_got_address): Add large offset support.
6830 (nios2_legitimize_tls_address): Update usage of removed and new
6831 functions.
6832 (nios2_symbol_binds_local_p): New function.
6833 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
6834 (nios2_legitimize_address): Update to use nios2_large_offset_p.
6835 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
6836 (nios2_print_operand): Merge H/L processing, add hiadj/lo
6837 processing for (const (unspec ...)).
6838 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
6840 2014-02-20 Richard Biener <rguenther@suse.de>
6842 * tree-cfg.c (replace_uses_by): Mark altered BBs before
6843 doing the substitution.
6844 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
6846 2014-02-20 Martin Jambor <mjambor@suse.cz>
6848 PR ipa/55260
6849 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
6850 info when checking whether lattices are bottom.
6852 2014-02-20 Richard Biener <rguenther@suse.de>
6854 PR middle-end/60221
6855 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
6856 regions at -O0.
6858 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
6860 PR ipa/58555
6861 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
6862 parameter specifying the scaling.
6863 (inline_call): Update.
6864 (want_inline_recursively): Guard division by zero.
6865 (recursive_inlining): Update.
6866 * ipa-inline.h (clone_inlined_nodes): Update.
6868 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
6870 PR target/60204
6871 * config/i386/i386.c (classify_argument): Pass structures of size
6872 64 bytes or less in register.
6874 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
6875 Kirill Yukhin <kirill.yukhin@intel.com>
6877 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
6878 (_mm_rcp28_round_ss): Ditto.
6879 (_mm_rsqrt28_round_sd): Ditto.
6880 (_mm_rsqrt28_round_ss): Ditto.
6881 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
6882 (_mm_rcp14_round_ss): Ditto.
6883 (_mm_rsqrt14_round_sd): Ditto.
6884 (_mm_rsqrt14_round_ss): Ditto.
6885 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
6886 the first input operand, get rid of match_dup.
6887 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
6888 attribute to sse.
6889 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
6890 Ditto.
6891 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
6892 operand as the first input operand, set type attribute.
6893 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
6894 Set type attribute.
6895 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
6896 operand as the first input operand, set type attribute.
6898 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6900 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
6901 bit of zero.
6903 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
6905 PR target/60207
6906 * config/i386/i386.c (construct_container): Remove TFmode check
6907 for X86_64_INTEGER_CLASS.
6909 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
6911 PR target/59794
6912 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
6913 only when -Wpsabi is enabled.
6915 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
6917 PR target/59799
6918 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
6919 passing arrays in registers are the same as for structs, so remove the
6920 special case for them.
6922 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
6924 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
6925 destination type, extract only the valid bits if the source type is not
6926 integral and has a different mode.
6928 2014-02-19 Richard Biener <rguenther@suse.de>
6930 PR ipa/60243
6931 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
6932 for all calls.
6934 2014-02-19 Richard Biener <rguenther@suse.de>
6936 PR ipa/60243
6937 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
6938 (ipa_modify_call_arguments): Emit an argument load explicitely and
6939 preserve virtual SSA form there and for the replacement call.
6940 Do not update SSA form nor free dominance info.
6942 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6944 * ipa.c (function_and_variable_visibility): Also clear WEAK
6945 flag when disolving COMDAT_GROUP.
6947 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6949 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
6950 * ipa-prop.c (ipa_set_jf_known_type): Return early when
6951 not devirtualizing.
6952 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
6953 do more sanity checks.
6954 (detect_type_change): Return true when giving up early.
6955 (compute_complex_assign_jump_func): Fix type parameter of
6956 ipa_set_ancestor_jf.
6957 (compute_complex_ancestor_jump_func): Likewise.
6958 (update_jump_functions_after_inlining): Fix updating of
6959 ancestor function.
6960 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
6962 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6964 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
6965 inline clones when edge disappears.
6967 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
6969 PR target/60203
6970 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
6971 Split 64-bit moves into 2 patterns. Do not allow the use of
6972 direct move for TDmode in little endian, since the decimal value
6973 has little endian bytes within a word, but the 64-bit pieces are
6974 ordered in a big endian fashion, and normal subreg's of TDmode are
6975 not allowed.
6976 (mov<mode>_64bit_dm): Likewise.
6977 (movtd_64bit_nodm): Likewise.
6979 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
6981 PR tree-optimization/60174
6982 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
6983 statement of an SSA_NAME that occurs in an abnormal PHI node.
6985 2014-02-18 Jakub Jelinek <jakub@redhat.com>
6987 PR sanitizer/60142
6988 * final.c (SEEN_BB): Remove.
6989 (SEEN_NOTE, SEEN_EMITTED): Renumber.
6990 (final_scan_insn): Don't force_source_line on second
6991 NOTE_INSN_BASIC_BLOCK.
6993 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
6995 PR target/60205
6996 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
6997 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
6998 (type_natural_mode): Warn ABI change when %zmm register is not
6999 available for AVX512F vector value passing.
7001 2014-02-18 Kai Tietz <ktietz@redhat.com>
7003 PR target/60193
7004 * config/i386/i386.c (ix86_expand_prologue): Use value in
7005 rax register as displacement when restoring %r10 or %rax.
7006 Fix wrong offset when restoring both registers.
7008 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
7010 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
7011 assertion with conditional return.
7013 2014-02-18 Jakub Jelinek <jakub@redhat.com>
7014 Uros Bizjak <ubizjak@gmail.com>
7016 PR driver/60233
7017 * config/i386/driver-i386.c (host_detect_local_cpu): If
7018 YMM state is not saved by the OS, also clear has_f16c. Move
7019 CPUID 0x80000001 handling before YMM state saving checking.
7021 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
7023 PR rtl-optimization/58960
7024 * haifa-sched.c (alloc_global_sched_pressure_data): New,
7025 factored out from ...
7026 (sched_init): ... here.
7027 (free_global_sched_pressure_data): New, factored out from ...
7028 (sched_finish): ... here.
7029 * sched-int.h (free_global_sched_pressure_data): Declare.
7030 * sched-rgn.c (nr_regions_initial): New static global.
7031 (haifa_find_rgns): Initialize it.
7032 (schedule_region): Disable sched-pressure for the newly
7033 generated regions.
7035 2014-02-17 Richard Biener <rguenther@suse.de>
7037 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
7038 release SSA defs of pattern stmts.
7040 2014-02-17 Richard Biener <rguenther@suse.de>
7042 * tree-inline.c (expand_call_inline): Release the virtual
7043 operand defined by the call we are about to inline.
7045 2014-02-17 Richard Biener <rguenther@suse.de>
7047 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
7049 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
7050 Ilya Tocar <ilya.tocar@intel.com>
7052 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
7053 arguments order in builtin.
7054 (_mm512_permutexvar_epi64): Ditto.
7055 (_mm512_mask_permutexvar_epi64): Ditto
7056 (_mm512_maskz_permutexvar_epi32): Ditto
7057 (_mm512_permutexvar_epi32): Ditto
7058 (_mm512_mask_permutexvar_epi32): Ditto
7060 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7062 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
7063 (p8_vmrgow): Likewise.
7065 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7067 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
7068 endian targets.
7070 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
7072 PR target/60203
7073 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
7074 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
7075 into 64-bit and 32-bit moves. On 64-bit moves, add support for
7076 using direct move instructions on ISA 2.07. Also adjust
7077 instruction length for 64-bit.
7078 (mov<mode>_64bit, TFmode/TDmode): Likewise.
7079 (mov<mode>_32bit, TFmode/TDmode): Likewise.
7081 2014-02-15 Alan Modra <amodra@gmail.com>
7083 PR target/58675
7084 PR target/57935
7085 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
7086 find_replacement on parts of insn rtl that might be reloaded.
7088 2014-02-15 Richard Biener <rguenther@suse.de>
7090 PR tree-optimization/60183
7091 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
7092 (tree_ssa_phiprop): Calculate and free post-dominators.
7094 2014-02-14 Jeff Law <law@redhat.com>
7096 PR rtl-optimization/60131
7097 * ree.c (get_extended_src_reg): New function.
7098 (combine_reaching_defs): Use it rather than assuming location of REG.
7099 (find_and_remove_re): Verify first operand of extension is
7100 a REG before adding the insns to the copy list.
7102 2014-02-14 Roland McGrath <mcgrathr@google.com>
7104 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
7105 * configure: Regenerated.
7106 * config.in: Regenerated.
7107 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
7108 instead of ASM_SHORT.
7110 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
7111 Richard Earnshaw <rearnsha@arm.com>
7113 PR rtl-optimization/59535
7114 * lra-constraints.c (process_alt_operands): Encourage alternative
7115 when unassigned pseudo class is superset of the alternative class.
7116 (inherit_reload_reg): Don't inherit when optimizing for code size.
7117 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
7118 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
7119 modes not less than 4 for Thumb1.
7121 2014-02-14 Kyle McMartin <kyle@redhat.com>
7123 PR pch/60010
7124 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
7126 2014-02-14 Richard Biener <rguenther@suse.de>
7128 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
7129 (get_frame_arg): Drop the assert with langhook types_compatible_p.
7130 Do not strip INDIRECT_REFs.
7132 2014-02-14 Richard Biener <rguenther@suse.de>
7134 PR lto/60179
7135 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
7136 DECL_FUNCTION_SPECIFIC_TARGET.
7137 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
7138 * tree-streamer-out.c (pack_ts_target_option): Remove.
7139 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
7140 (write_ts_function_decl_tree_pointers): Do not stream
7141 DECL_FUNCTION_SPECIFIC_TARGET.
7142 * tree-streamer-in.c (unpack_ts_target_option): Remove.
7143 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
7144 (lto_input_ts_function_decl_tree_pointers): Do not stream
7145 DECL_FUNCTION_SPECIFIC_TARGET.
7147 2014-02-14 Jakub Jelinek <jakub@redhat.com>
7149 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
7150 (get_initial_def_for_induction, vectorizable_induction): Ignore
7151 debug stmts when looking for exit_phi.
7152 (vectorizable_live_operation): Fix up condition.
7154 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
7156 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
7157 nreverse() because it changes the content of original tree list.
7159 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
7161 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
7162 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
7164 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
7166 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
7167 GNU coding standards.
7169 2014-02-13 Jakub Jelinek <jakub@redhat.com>
7171 PR debug/60152
7172 * dwarf2out.c (gen_subprogram_die): Don't call
7173 add_calling_convention_attribute if subr_die is old_die.
7175 2014-02-13 Sharad Singhai <singhai@google.com>
7177 * doc/optinfo.texi: Fix order of nodes.
7179 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
7181 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
7182 operands[2], not operands[3].
7184 2014-02-13 Richard Biener <rguenther@suse.de>
7186 PR bootstrap/59878
7187 * doc/install.texi (ISL): Update recommended version to 0.12.2,
7188 mention the possibility of an in-tree build.
7189 (CLooG): Update recommended version to 0.18.1, mention the
7190 possibility of an in-tree build and clarify that the ISL
7191 bundled with CLooG does not work.
7193 2014-02-13 Jakub Jelinek <jakub@redhat.com>
7195 PR target/43546
7196 * expr.c (compress_float_constant): If x is a hard register,
7197 extend into a pseudo and then move to x.
7199 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
7201 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
7202 caused by bad second argument to warning_at() with -mhotpatch and
7203 nested functions (e.g. with gfortran).
7205 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
7207 * opts.c (option_name): Remove "enabled by default" rider.
7209 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
7211 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
7213 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
7214 Uros Bizjak <ubizjak@gmail.com>
7216 PR target/60151
7217 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
7218 * configure: Regenerated.
7220 2014-02-12 Richard Biener <rguenther@suse.de>
7222 * vec.c (vec_prefix::calculate_allocation): Move as
7223 inline variant to vec.h.
7224 (vec_prefix::calculate_allocation_1): New out-of-line version.
7225 * vec.h (vec_prefix::calculate_allocation_1): Declare.
7226 (vec_prefix::m_has_auto_buf): Rename to ...
7227 (vec_prefix::m_using_auto_storage): ... this.
7228 (vec_prefix::calculate_allocation): Inline the easy cases
7229 and dispatch to calculate_allocation_1 which doesn't need the
7230 prefix address.
7231 (va_heap::reserve): Use gcc_checking_assert.
7232 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
7233 m_using_auto_storage.
7234 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
7235 member and adjust.
7236 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
7237 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
7238 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
7240 2014-02-12 Richard Biener <rguenther@suse.de>
7242 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
7243 when we found a dependence.
7245 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
7247 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
7248 common code...
7249 (maybe_fold_stmt): ... into this new function.
7250 * omp-low.c (lower_omp): Update comment.
7252 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
7253 last use.
7255 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
7256 dereference.
7258 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
7260 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
7261 identifiers in comments.
7262 (cortexa53_extra_costs): Likewise.
7263 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
7264 (cortexa7_extra_costs): Likewise.
7265 (cortexa12_extra_costs): Likewise.
7266 (cortexa15_extra_costs): Likewise.
7267 (v7m_extra_costs): Likewise.
7269 2014-02-12 Richard Biener <rguenther@suse.de>
7271 PR middle-end/60092
7272 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
7273 of posix_memalign being successful.
7274 (lower_stmt): Restrict lowering of posix_memalign to when
7275 -ftree-bit-ccp is enabled.
7277 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
7279 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
7280 arg_loc.
7281 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
7283 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
7285 PR rtl-optimization/60116
7286 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
7287 other_insn once the combination has been validated.
7289 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
7291 PR lto/59468
7292 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
7293 and wrapper.
7294 * ipa-devirt.c: Include demangle.h
7295 (odr_violation_reported): New static variable.
7296 (add_type_duplicate): Update odr_violations.
7297 (maybe_record_node): Add completep parameter; update it.
7298 (record_target_from_binfo): Add COMPLETEP parameter;
7299 update it as needed.
7300 (possible_polymorphic_call_targets_1): Likewise.
7301 (struct polymorphic_call_target_d): Add nonconstruction_targets;
7302 rename FINAL to COMPLETE.
7303 (record_targets_from_bases): Sanity check we found the binfo;
7304 fix COMPLETEP updating.
7305 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
7306 parameter, fix computing of COMPLETEP.
7307 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
7308 at LTO time do demangling.
7309 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
7310 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
7311 parameter.
7312 (gimple_get_virt_method_for_binfo): Likewise.
7313 * gimple-fold.h (gimple_get_virt_method_for_binfo,
7314 gimple_get_virt_method_for_vtable): Update prototypes.
7316 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
7318 PR target/49008
7319 * genautomata.c (add_presence_absence): Fix typo with
7320 {final_}presence_list.
7322 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
7324 PR target/60137
7325 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
7326 for VSX/Altivec vectors that land in GPR registers.
7328 2014-02-11 Richard Henderson <rth@redhat.com>
7329 Jakub Jelinek <jakub@redhat.com>
7331 PR debug/59776
7332 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
7333 around drhs if type conversion to lacc->type is not useless.
7335 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7337 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
7338 tuning struct.
7339 (cortex-a57.cortex-a53): Likewise.
7340 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
7342 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7344 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
7345 arm_restrict_it.
7347 2014-02-11 Renlin Li <Renlin.Li@arm.com>
7349 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
7350 add_options_for_arm_vfp3.
7352 2014-02-11 Jeff Law <law@redhat.com>
7354 PR middle-end/54041
7355 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
7356 object with an undesirable mode.
7358 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7360 PR libgomp/60107
7361 * config/i386/sol2-9.h: New file.
7362 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
7363 *-*-solaris2.9*): Use it.
7365 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
7367 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
7368 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
7370 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
7372 * config/microblaze/microblaze.c: Extend mcpu version format
7374 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
7376 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
7378 2014-02-10 Richard Henderson <rth@redhat.com>
7380 PR target/59927
7381 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
7382 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
7383 ms-abi vs -mno-accumulate-outgoing-args.
7384 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
7385 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
7386 respect to ms-abi.
7388 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
7390 PR middle-end/60080
7391 * cfgexpand.c (expand_asm_operands): Attach source location to
7392 ASM_INPUT rtx objects.
7393 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
7395 2014-02-10 Nick Clifton <nickc@redhat.com>
7397 * config/mn10300/mn10300.c (popcount): New function.
7398 (mn10300_expand_prologue): Include saved registers in stack usage
7399 count.
7401 2014-02-10 Jeff Law <law@redhat.com>
7403 PR middle-end/52306
7404 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
7405 when changing the SET_DEST of a prior insn to avoid an input reload.
7407 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
7409 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
7410 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
7411 -mcall-openbsd, or -mcall-linux.
7412 (CC1_ENDIAN_BIG_SPEC): Remove.
7413 (CC1_ENDIAN_LITTLE_SPEC): Remove.
7414 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
7415 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
7416 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
7417 and %cc1_endian_default.
7418 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
7420 2014-02-10 Richard Biener <rguenther@suse.de>
7422 PR tree-optimization/60115
7423 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
7424 MEM_REF handling. Properly verify that the accesses are not
7425 out of the objects bound.
7427 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7429 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
7430 coretex to cortex.
7432 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
7434 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
7435 proper constants and fix formatting.
7436 (possible_polymorphic_call_targets): Fix formatting.
7438 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
7439 Ilya Tocar <ilya.tocar@intel.com>
7441 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
7442 (_mm512_loadu_epi32): Renamed into...
7443 (_mm512_loadu_si512): This.
7444 (_mm512_storeu_epi32): Renamed into...
7445 (_mm512_storeu_si512): This.
7446 (_mm512_maskz_ceil_ps): Removed.
7447 (_mm512_maskz_ceil_pd): Ditto.
7448 (_mm512_maskz_floor_ps): Ditto.
7449 (_mm512_maskz_floor_pd): Ditto.
7450 (_mm512_floor_round_ps): Ditto.
7451 (_mm512_floor_round_pd): Ditto.
7452 (_mm512_ceil_round_ps): Ditto.
7453 (_mm512_ceil_round_pd): Ditto.
7454 (_mm512_mask_floor_round_ps): Ditto.
7455 (_mm512_mask_floor_round_pd): Ditto.
7456 (_mm512_mask_ceil_round_ps): Ditto.
7457 (_mm512_mask_ceil_round_pd): Ditto.
7458 (_mm512_maskz_floor_round_ps): Ditto.
7459 (_mm512_maskz_floor_round_pd): Ditto.
7460 (_mm512_maskz_ceil_round_ps): Ditto.
7461 (_mm512_maskz_ceil_round_pd): Ditto.
7462 (_mm512_expand_pd): Ditto.
7463 (_mm512_expand_ps): Ditto.
7464 * config/i386/i386.c (ix86_builtins): Remove
7465 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
7466 (bdesc_args): Ditto.
7467 * config/i386/predicates.md (const1256_operand): New.
7468 (const_1_to_2_operand): Ditto.
7469 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
7470 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
7471 (*avx512pf_gatherpf<mode>sf): Ditto.
7472 (avx512pf_gatherpf<mode>df): Ditto.
7473 (*avx512pf_gatherpf<mode>df_mask): Ditto.
7474 (*avx512pf_gatherpf<mode>df): Ditto.
7475 (avx512pf_scatterpf<mode>sf): Ditto.
7476 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
7477 (*avx512pf_scatterpf<mode>sf): Ditto.
7478 (avx512pf_scatterpf<mode>df): Ditto.
7479 (*avx512pf_scatterpf<mode>df_mask): Ditto.
7480 (*avx512pf_scatterpf<mode>df): Ditto.
7481 (avx512f_expand<mode>): Removed.
7482 (<shift_insn><mode>3<mask_name>): Change predicate type.
7484 2014-02-08 Jakub Jelinek <jakub@redhat.com>
7486 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
7487 not at the end of datarefs vector use ordered_remove to avoid
7488 reordering datarefs vector.
7490 PR c/59984
7491 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
7492 mark local addressable non-static vars as GOVD_PRIVATE
7493 instead of GOVD_LOCAL.
7494 * omp-low.c (lower_omp_for): Move gimple_bind_vars
7495 and BLOCK_VARS of gimple_bind_block to new_stmt rather
7496 than copying them.
7498 PR middle-end/60092
7499 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
7500 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
7501 assume_aligned or alloc_align attributes.
7502 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
7503 arguments. Handle also assume_aligned and alloc_align attributes.
7504 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
7505 calls to functions with assume_aligned or alloc_align attributes.
7506 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
7508 2014-02-08 Terry Guo <terry.guo@arm.com>
7510 * doc/invoke.texi: Document ARM -march=armv7e-m.
7512 2014-02-08 Jakub Jelinek <jakub@redhat.com>
7514 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
7515 flag on __cilkrts_rethrow builtin.
7517 PR ipa/60026
7518 * ipa-cp.c (determine_versionability): Fail at -O0
7519 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
7520 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
7522 Revert:
7523 2014-02-04 Jakub Jelinek <jakub@redhat.com>
7525 PR ipa/60026
7526 * tree-inline.c (copy_forbidden): Fail for
7527 __attribute__((optimize (0))) functions.
7529 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
7531 * varpool.c: Include pointer-set.h.
7532 (varpool_remove_unreferenced_decls): Variables in other partitions
7533 will not be output; be however careful to not lose information
7534 about partitioning.
7536 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
7538 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
7539 lookup in the vtable constructor.
7541 2014-02-07 Jeff Law <law@redhat.com>
7543 PR target/40977
7544 * config/m68k/m68k.md (ashldi_extsi): Turn into a
7545 define_insn_and_split.
7547 * ipa-inline.c (inline_small_functions): Fix typos.
7549 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7551 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
7552 (s390_can_use_return_insn): Declare.
7553 * config/s390/s390.h (EPILOGUE_USES): Define.
7554 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
7555 instructions.
7556 (s390_chunkify_start): Handle return JUMP_LABELs.
7557 (s390_early_mach): Emit a main_pool instruction on the entry edge.
7558 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
7559 (s390_can_use_return_insn): New functions.
7560 (s390_fix_long_loop_prediction): Handle conditional returns.
7561 (TARGET_SET_UP_BY_PROLOGUE): Define.
7562 * config/s390/s390.md (ANY_RETURN): New code iterator.
7563 (*creturn, *csimple_return, return, simple_return): New patterns.
7565 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7567 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
7568 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
7569 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
7570 REG_CFA_RESTORE list when deciding not to restore a register.
7572 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7574 * config/s390/s390.c: Include tree-pass.h and context.h.
7575 (s390_early_mach): New function, split out from...
7576 (s390_emit_prologue): ...here.
7577 (pass_data_s390_early_mach): New pass structure.
7578 (pass_s390_early_mach): New class.
7579 (s390_option_override): Create and register early_mach pass.
7580 Move to end of file.
7582 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7584 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
7585 to match for the exit block.
7587 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7589 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
7590 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
7591 Reject misaligned operands.
7593 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7595 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
7597 2014-02-07 Richard Biener <rguenther@suse.de>
7599 PR middle-end/60092
7600 * gimple-low.c (lower_builtin_posix_memalign): New function.
7601 (lower_stmt): Call it to lower posix_memalign in a way
7602 to make alignment info accessible.
7604 2014-02-07 Jakub Jelinek <jakub@redhat.com>
7606 PR c++/60082
7607 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
7608 __builtin_setjmp_receiver.
7610 2014-02-07 Richard Biener <rguenther@suse.de>
7612 PR middle-end/60092
7613 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
7614 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
7615 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
7616 Handle BUILT_IN_POSIX_MEMALIGN.
7617 (find_func_clobbers): Likewise.
7618 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
7619 (call_may_clobber_ref_p_1): Likewise.
7621 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7623 PR ipa/59918
7624 * ipa-devirt.c (record_target_from_binfo): Remove overactive
7625 sanity check.
7627 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7629 PR ipa/59469
7630 * lto-cgraph.c (lto_output_node): Use
7631 symtab_get_symbol_partitioning_class.
7632 (lto_output_varpool_node): likewise.
7633 (symtab_get_symbol_partitioning_class): Move here from
7634 lto/lto-partition.c
7635 * cgraph.h (symbol_partitioning_class): Likewise.
7636 (symtab_get_symbol_partitioning_class): Declare.
7638 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7640 * ggc.h (ggc_internal_cleared_alloc): New macro.
7641 * vec.h (vec_safe_copy): Handle memory stats.
7642 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
7643 * target-globals.c (save_target_globals): Likewise.
7645 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7647 PR target/60077
7648 * expr.c (emit_move_resolve_push): Export; be bit more selective
7649 on when to clear alias set.
7650 * expr.h (emit_move_resolve_push): Declare.
7651 * function.h (struct function): Add tail_call_marked.
7652 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
7653 * config/i386/i386-protos.h (ix86_expand_push): Remove.
7654 * config/i386/i386.md (TImode move expander): De not call
7655 ix86_expand_push.
7656 (FP push expanders): Preserve memory attributes.
7657 * config/i386/sse.md (push<mode>1): Remove.
7658 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
7659 (ix86_expand_push): Remove.
7660 * config/i386/mmx.md (push<mode>1): Remove.
7662 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7664 PR rtl-optimization/60030
7665 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
7666 lopart with paradoxical subreg before shifting it up by hprec.
7668 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7670 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
7671 Remove extra newline at end of file.
7672 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
7673 (arm_issue_rate): Handle cortexa57.
7674 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
7675 (cortex-a57.cortex-a53): Likewise.
7677 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7679 PR target/59575
7680 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
7681 don't record in REG_FRAME_RELATED_EXPR registers not set in that
7682 bitmask.
7683 (arm_expand_prologue): Adjust all callers.
7684 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
7685 info, registers also at the lowest numbered registers side. Use
7686 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
7687 XEXP.
7689 PR debug/59992
7690 * var-tracking.c (adjust_mems): Before adding a SET to
7691 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
7693 2014-02-06 Alan Modra <amodra@gmail.com>
7695 PR target/60032
7696 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
7697 change SDmode to DDmode when lra_in_progress.
7699 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7701 PR middle-end/59150
7702 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
7703 free_data_ref on the dr first, and before goto again also set dr
7704 to the next dr. For simd_lane_access, free old datarefs[i] before
7705 overwriting it. For get_vectype_for_scalar_type failure, don't
7706 free_data_ref if simd_lane_access.
7708 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
7710 PR target/60062
7711 * tree.h (opts_for_fn): New inline function.
7712 (opt_for_fn): Define.
7713 * config/i386/i386.c (ix86_function_regparm): Use
7714 opt_for_fn (decl, optimize) instead of optimize.
7716 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
7718 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
7719 for SYMBOL_REF in large memory model.
7721 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7723 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
7724 and crypto support.
7725 (cortex-a57): Likewise.
7726 (cortex-a57.cortex-a53): Likewise.
7728 2014-02-06 Yury Gribov <y.gribov@samsung.com>
7729 Kugan Vivekanandarajah <kuganv@linaro.org>
7731 * config/arm/arm.c (arm_vector_alignment_reachable): Check
7732 unaligned_access.
7733 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
7735 2014-02-06 Richard Biener <rguenther@suse.de>
7737 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
7738 set_loop_copy and initialize_original_copy_tables.
7740 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
7742 * config/aarch64/aarch64-simd.md
7743 (aarch64_ashr_simddi): Change QI to SI.
7745 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
7746 Jakub Jelinek <jakub@redhat.com>
7748 PR middle-end/60013
7749 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
7750 of the dataflow.
7752 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7754 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
7755 CODE_FOR_altivec_vpku[hw]um to
7756 CODE_FOR_altivec_vpku[hw]um_direct.
7757 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
7758 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
7759 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
7760 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
7762 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7764 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
7765 generation for -maltivec=be.
7766 (altivec_vsumsws): Simplify redundant test.
7768 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7770 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
7771 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
7772 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
7773 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
7774 gen_altivec_vpkuwum.
7775 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
7776 BYTES_BIG_ENDIAN.
7777 (altivec_vpks<VI_char>ss): Likewise.
7778 (altivec_vpks<VI_char>us): Likewise.
7779 (altivec_vpku<VI_char>us): Likewise.
7780 (altivec_vpku<VI_char>um): Likewise.
7781 (altivec_vpku<VI_char>um_direct): New (copy of
7782 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
7783 internal use).
7784 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
7785 target is little endian and -maltivec=be is not specified.
7786 (*altivec_vupkhs<VU_char>_direct): New (copy of
7787 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
7788 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
7789 target is little endian and -maltivec=be is not specified.
7790 (*altivec_vupkls<VU_char>_direct): New (copy of
7791 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
7792 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
7793 little endian and -maltivec=be is not specified.
7794 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
7795 little endian and -maltivec=be is not specified.
7797 2014-02-05 Richard Henderson <rth@redhat.com>
7799 PR debug/52727
7800 * combine-stack-adj.c: Revert r206943.
7801 * sched-int.h (struct deps_desc): Add last_args_size.
7802 * sched-deps.c (init_deps): Initialize it.
7803 (sched_analyze_insn): Add OUTPUT dependencies between insns that
7804 contain REG_ARGS_SIZE notes.
7806 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
7808 * lto-cgraph.c (asm_nodes_output): Make global.
7809 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
7810 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
7811 (driver_handle_option): Handle OPT_fwpa.
7813 2014-02-05 Jakub Jelinek <jakub@redhat.com>
7815 PR ipa/59947
7816 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
7817 a comment typo and formatting issue. If odr_hash hasn't been
7818 created, return vNULL and set *completep to false.
7820 PR middle-end/57499
7821 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
7822 bb with no successors.
7824 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
7826 PR target/59718
7827 * doc/invoke.texi (-march): Clarify documentation for ARM.
7828 (-mtune): Likewise.
7829 (-mcpu): Likewise.
7831 2014-02-05 Richard Biener <rguenther@suse.de>
7833 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
7834 when not vectorizing because of too many alias checks.
7835 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
7836 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
7838 2014-02-05 Nick Clifton <nickc@redhat.com>
7840 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
7841 accept extended registers in any mode when compiling for the MN10300.
7843 2014-02-05 Yury Gribov <y.gribov@samsung.com>
7845 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
7846 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
7847 sanitization attributes.
7848 (can_inline_edge_p): Likewise.
7849 (sanitize_attrs_match_for_inline_p): New function.
7851 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7853 * ipa-prop.c (detect_type_change): Shor circuit testing of
7854 type changes on THIS pointer.
7856 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
7858 PR target/59777
7859 * config/pa/pa.c (legitimize_tls_address): Return original address
7860 if not passed a SYMBOL_REF rtx.
7861 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
7862 addresses.
7863 (pa_emit_move_sequence): Simplify TLS source operands.
7864 (pa_legitimate_constant_p): Reject all TLS constants.
7865 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
7866 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
7868 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7870 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
7871 groups when we know they are controlled by LTO.
7872 * varasm.c (default_binds_local_p_1): If object is in other partition,
7873 it will be resolved locally.
7875 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
7877 * config/host-linux.c (linux_gt_pch_use_address): Don't
7878 use SSIZE_MAX because it is not always defined.
7880 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
7882 PR bootstrap/59913
7883 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
7884 threshold for pseudo splitting.
7885 (update_ebb_live_info): Process call argument hard registers and
7886 hard registers from insn definition too.
7887 (max_small_class_regs_num): New constant.
7888 (inherit_in_ebb): Update live hard regs through EBBs. Update
7889 reloads_num only for small register classes. Don't split for
7890 outputs of jumps.
7892 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
7894 PR ipa/60058
7895 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
7896 is non-null.
7898 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7900 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
7901 visibility is safe.
7903 2014-02-04 Marek Polacek <polacek@redhat.com>
7905 * gdbinit.in (pel): Define.
7907 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
7909 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
7910 behavior.
7912 2014-02-04 Richard Biener <rguenther@suse.de>
7914 PR lto/59723
7915 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
7916 in function context local.
7917 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
7918 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
7919 similar to LTO_imported_decl_ref.
7921 2014-02-04 Jakub Jelinek <jakub@redhat.com>
7923 PR tree-optimization/60002
7924 * cgraphclones.c (build_function_decl_skip_args): Clear
7925 DECL_LANG_SPECIFIC.
7927 PR tree-optimization/60023
7928 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
7929 false to gsi_replace.
7930 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
7931 has been in some EH region and vec_stmt could throw, add
7932 vec_stmt into the same EH region.
7933 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
7934 has no lhs, ignore it.
7935 * internal-fn.c (expand_MASK_LOAD): Likewise.
7937 PR ipa/60026
7938 * tree-inline.c (copy_forbidden): Fail for
7939 __attribute__((optimize (0))) functions.
7941 PR other/58712
7942 * omp-low.c (simd_clone_struct_copy): If from->inbranch
7943 is set, copy one less argument.
7944 (expand_simd_clones): Don't subtract clone_info->inbranch
7945 from simd_clone_struct_alloc argument.
7947 PR rtl-optimization/57915
7948 * recog.c (simplify_while_replacing): If all unary/binary/relational
7949 operation arguments are constant, attempt to simplify those.
7951 PR middle-end/59261
7952 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
7953 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
7955 2014-02-04 Richard Biener <rguenther@suse.de>
7957 PR tree-optimization/60012
7958 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
7959 TBAA disambiguation to all DDRs.
7961 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7963 PR target/59788
7964 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
7965 (LINK_SPEC): Use it for -shared, -shared-libgcc.
7967 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7969 PR ipa/59882
7970 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
7972 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7974 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
7975 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
7977 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7979 PR ipa/59831
7980 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
7981 to figure out targets of polymorphic calls with known decl.
7982 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7983 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
7984 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
7985 (get_polymorphic_call_info): ... here.
7986 (get_polymorphic_call_info_from_invariant): New function.
7988 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7990 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
7991 lookup via vtable pointer; check for type consistency
7992 and turn inconsitent facts into UNREACHABLE.
7993 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7994 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
7995 type inconsistent querries; return UNREACHABLE instead.
7997 2014-02-03 Richard Henderson <rth@twiddle.net>
7999 PR tree-opt/59924
8000 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
8001 already processed this node.
8002 (normalize_one_pred_1): Pass along mark_set.
8003 (normalize_one_pred): Create and destroy a pointer_set_t.
8004 (normalize_one_pred_chain): Likewise.
8006 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
8008 PR gcov-profile/58602
8009 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
8011 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
8013 PR ipa/59831
8014 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
8015 -fno-devirtualize; try to devirtualize by the knowledge of
8016 virtual table pointer given by aggregate propagation.
8017 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
8018 (ipa_print_node_jump_functions): Dump also offset that
8019 is relevant for polymorphic calls.
8020 (determine_known_aggregate_parts): Add arg_type parameter; use it
8021 instead of determining the type from pointer type.
8022 (ipa_compute_jump_functions_for_edge): Update call of
8023 determine_known_aggregate_parts.
8024 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
8025 (gimple_get_virt_method_for_binfo): ... here; simplify using
8026 vtable_pointer_value_to_vtable.
8027 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
8028 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
8029 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
8030 (vtable_pointer_value_to_vtable): Break out from ...; handle also
8031 POINTER_PLUS_EXPR.
8032 (vtable_pointer_value_to_binfo): ... here.
8033 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
8035 2014-02-03 Teresa Johnson <tejohnson@google.com>
8037 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
8038 redef of outer loop index variable.
8040 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
8042 PR c++/53017
8043 PR c++/59211
8044 * doc/extend.texi (Function Attributes): Typo.
8046 2014-02-03 Cong Hou <congh@google.com>
8048 PR tree-optimization/60000
8049 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
8050 if the vectorized statement is a store. A store statement can only
8051 appear at the end of pattern statements.
8053 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
8055 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
8056 (ix86_option_override_internal): Default long double to 64-bit for
8057 32-bit Bionic and to 128-bit for 64-bit Bionic.
8059 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
8060 TARGET_LONG_DOUBLE_128 is true.
8061 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
8063 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
8064 (mlong-double-64): Negate -mlong-double-128.
8065 (mlong-double-128): New option.
8067 * config/i386/i386-c.c (ix86_target_macros): Define
8068 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
8070 * doc/invoke.texi: Document -mlong-double-128.
8072 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
8074 PR rtl-optimization/60024
8075 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
8077 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
8079 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
8081 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
8083 PR rtl-optimization/57662
8084 * sel-sched.c (code_motion_path_driver): Do not mark already not
8085 existing blocks in the visiting bitmap.
8087 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
8089 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
8090 on the insn being emitted.
8092 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
8093 Will Deacon <will.deacon@arm.com>
8095 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
8097 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8099 * config/arm/arm-tables.opt: Regenerate.
8101 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8103 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
8104 for vector types other than V16QImode.
8105 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
8106 define_expand, and call altivec_expand_vec_perm_le when producing
8107 code with little endian element order.
8108 (*altivec_vperm_<mode>_internal): New insn having previous
8109 behavior of altivec_vperm_<mode>.
8110 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
8111 altivec_expand_vec_perm_le when producing code with little endian
8112 element order.
8113 (*altivec_vperm_<mode>_uns_internal): New insn having previous
8114 behavior of altivec_vperm_<mode>_uns.
8116 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8118 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
8119 (altivec_vsumsws): Add handling for -maltivec=be with a little
8120 endian target.
8121 (altivec_vsumsws_direct): New.
8122 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
8123 gen_altivec_vsumsws.
8125 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
8127 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
8128 vtable_pointer_value_to_binfo): New functions.
8129 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
8130 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
8132 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
8134 * config/nios2/nios2.md (load_got_register): Initialize GOT
8135 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
8136 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
8138 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
8140 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
8141 preserverd by passthrough, do not propagate the type.
8143 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
8145 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
8146 (mips_atomic_assign_expand_fenv): New function.
8147 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
8149 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
8151 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
8152 (__builtin_mips_set_fcsr): Likewise.
8153 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
8154 MIPS_USI_FTYPE_VOID.
8155 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
8156 (mips16_expand_set_fcsr): Likewise.
8157 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
8158 (mips16_set_fcsr_stub): Likewise.
8159 (mips16_get_fcsr_one_only_stub): New class.
8160 (mips16_set_fcsr_one_only_stub): Likewise.
8161 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
8162 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
8163 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
8164 (hard_float): New availability predicate.
8165 (mips_builtins): Add get_fcsr and set_fcsr.
8166 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
8167 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
8168 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
8169 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
8170 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
8171 patterns.
8173 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
8175 * config/mips/mips.c (mips_one_only_stub): New class.
8176 (mips_need_mips16_rdhwr_p): Replace with...
8177 (mips16_rdhwr_stub): ...this new variable.
8178 (mips16_stub_call_address): New function.
8179 (mips16_rdhwr_one_only_stub): New class.
8180 (mips_expand_thread_pointer): Use mips16_stub_call_address.
8181 (mips_output_mips16_rdhwr): Delete.
8182 (mips_finish_stub): New function.
8183 (mips_code_end): Use it to handle rdhwr stubs.
8185 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
8187 PR target/60017
8188 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
8189 when calculating size of integer atomic types.
8191 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
8193 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
8195 2014-02-01 Jakub Jelinek <jakub@redhat.com>
8197 PR tree-optimization/60003
8198 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
8199 * profile.c (branch_prob): Use gimple_call_builtin_p
8200 to check for BUILT_IN_SETJMP_RECEIVER.
8201 * tree-inline.c (copy_bb): Call notice_special_calls.
8203 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
8205 PR bootstrap/59985
8206 * lra-constraints.c (process_alt_operands): Update reload_sum only
8207 on the first pass.
8209 2014-01-31 Richard Henderson <rth@redhat.com>
8211 PR middle-end/60004
8212 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
8213 until after else_eh is processed.
8215 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
8217 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
8218 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
8219 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
8220 in smmintrin.h, remove them.
8221 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
8222 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
8223 * config/i386/i386.md (ROUND_SAE): Fix value.
8224 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
8225 (const48_operand): New.
8226 * config/i386/subst.md (round), (round_expand): Use
8227 const_4_or_8_to_11_operand.
8228 (round_saeonly), (round_saeonly_expand): Use const48_operand.
8230 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
8232 * config/i386/constraints.md (Yk): Swap meaning with k.
8233 * config/i386/i386.md (movhi_internal): Change Yk to k.
8234 (movqi_internal): Ditto.
8235 (*k<logic><mode>): Ditto.
8236 (*andhi_1): Ditto.
8237 (*andqi_1): Ditto.
8238 (kandn<mode>): Ditto.
8239 (*<code>hi_1): Ditto.
8240 (*<code>qi_1): Ditto.
8241 (kxnor<mode>): Ditto.
8242 (kortestzhi): Ditto.
8243 (kortestchi): Ditto.
8244 (kunpckhi): Ditto.
8245 (*one_cmplhi2_1): Ditto.
8246 (*one_cmplqi2_1): Ditto.
8247 * config/i386/sse.md (): Change k to Yk.
8248 (avx512f_load<mode>_mask): Ditto.
8249 (avx512f_blendm<mode>): Ditto.
8250 (avx512f_store<mode>_mask): Ditto.
8251 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
8252 (avx512f_storedqu<mode>_mask): Ditto.
8253 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
8254 Ditto.
8255 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
8256 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
8257 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
8258 (avx512f_maskcmp<mode>3): Ditto.
8259 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
8260 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
8261 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
8262 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
8263 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
8264 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
8265 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
8266 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
8267 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
8268 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
8269 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
8270 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
8271 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
8272 (vec_extract_lo_<mode>_maskm): Ditto.
8273 (vec_extract_hi_<mode>_maskm): Ditto.
8274 (avx512f_vternlog<mode>_mask): Ditto.
8275 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
8276 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
8277 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
8278 (avx512f_<code>v8div16qi2_mask): Ditto.
8279 (avx512f_<code>v8div16qi2_mask_store): Ditto.
8280 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
8281 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
8282 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
8283 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
8284 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
8285 (*avx512pf_gatherpf<mode>df_mask): Ditto.
8286 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
8287 (*avx512pf_scatterpf<mode>df_mask): Ditto.
8288 (avx512cd_maskb_vec_dupv8di): Ditto.
8289 (avx512cd_maskw_vec_dupv16si): Ditto.
8290 (avx512f_vpermi2var<mode>3_maskz): Ditto.
8291 (avx512f_vpermi2var<mode>3_mask): Ditto.
8292 (avx512f_vpermi2var<mode>3_mask): Ditto.
8293 (avx512f_vpermt2var<mode>3_maskz): Ditto.
8294 (*avx512f_gathersi<mode>): Ditto.
8295 (*avx512f_gathersi<mode>_2): Ditto.
8296 (*avx512f_gatherdi<mode>): Ditto.
8297 (*avx512f_gatherdi<mode>_2): Ditto.
8298 (*avx512f_scattersi<mode>): Ditto.
8299 (*avx512f_scatterdi<mode>): Ditto.
8300 (avx512f_compress<mode>_mask): Ditto.
8301 (avx512f_compressstore<mode>_mask): Ditto.
8302 (avx512f_expand<mode>_mask): Ditto.
8303 * config/i386/subst.md (mask): Change k to Yk.
8304 (mask_scalar_merge): Ditto.
8305 (sd): Ditto.
8307 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
8309 * doc/extend.texi (Vector Extensions): Document ?: in C++.
8311 2014-01-31 Richard Biener <rguenther@suse.de>
8313 PR middle-end/59990
8314 * builtins.c (fold_builtin_memory_op): Make sure to not
8315 use a floating-point mode or a boolean or enumeral type for
8316 the copy operation.
8318 2014-01-30 DJ Delorie <dj@redhat.com>
8320 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
8321 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
8322 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
8323 whenever main() has an epilogue.
8325 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8327 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
8328 unused variable "field".
8329 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
8330 (vsx_mergeh_<mode>): Likewise.
8331 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
8332 (altivec_vmrghh): Likewise.
8333 (altivec_vmrghw): Likewise.
8334 (altivec_vmrglb): Likewise.
8335 (altivec_vmrglh): Likewise.
8336 (altivec_vmrglw): Likewise.
8337 (altivec_vspltb): Add missing uses.
8338 (altivec_vsplth): Likewise.
8339 (altivec_vspltw): Likewise.
8340 (altivec_vspltsf): Likewise.
8342 2014-01-30 Jakub Jelinek <jakub@redhat.com>
8344 PR target/59923
8345 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
8346 frame related instructions.
8348 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
8350 PR rtl-optimization/59959
8351 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
8352 any reload of register whose subreg is invalid.
8354 2014-01-30 Jakub Jelinek <jakub@redhat.com>
8356 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
8357 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
8358 Add missing return type - void.
8360 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8362 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
8363 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
8364 remove element index adjustment for endian (now handled in vsx.md
8365 and altivec.md).
8366 (altivec_expand_vec_perm_const): Use
8367 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
8368 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
8369 (vsx_xxspltw_<mode>): Adjust element index for little endian.
8370 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
8371 define_expand and a new define_insn *altivec_vspltb_internal;
8372 adjust for -maltivec=be on a little endian target.
8373 (altivec_vspltb_direct): New.
8374 (altivec_vsplth): Divide into a define_expand and a new
8375 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
8376 little endian target.
8377 (altivec_vsplth_direct): New.
8378 (altivec_vspltw): Divide into a define_expand and a new
8379 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
8380 little endian target.
8381 (altivec_vspltw_direct): New.
8382 (altivec_vspltsf): Divide into a define_expand and a new
8383 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
8384 a little endian target.
8386 2014-01-30 Richard Biener <rguenther@suse.de>
8388 PR tree-optimization/59993
8389 * tree-ssa-forwprop.c (associate_pointerplus): Check we
8390 can propagate form the earlier stmt and avoid the transform
8391 when the intermediate result is needed.
8393 2014-01-30 Alangi Derick <alangiderick@gmail.com>
8395 * README.Portability: Fix typo.
8397 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
8399 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
8400 comparison_operator with ordered_comparison_operator.
8402 2014-01-30 Nick Clifton <nickc@redhat.com>
8404 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
8405 Rename to mn10300_store_multiple_regs.
8406 * config/mn10300/mn10300.c: Likewise.
8407 * config/mn10300/mn10300.md (store_movm): Fix typo: call
8408 store_multiple_regs.
8409 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
8410 Call mn10300_store_multiple_regs.
8412 2014-01-30 Nick Clifton <nickc@redhat.com>
8413 DJ Delorie <dj@redhat.com>
8415 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
8416 %fp 2 to keep registers after it properly word-aligned.
8417 (rl78_alloc_physical_registers_umul): Handle the case where both
8418 input operands are the same.
8420 2014-01-30 Richard Biener <rguenther@suse.de>
8422 PR tree-optimization/59903
8423 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
8424 check properly.
8426 2014-01-30 Jason Merrill <jason@redhat.com>
8428 PR c++/59633
8429 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
8431 PR c++/59645
8432 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
8434 2014-01-30 Richard Biener <rguenther@suse.de>
8436 PR tree-optimization/59951
8437 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
8439 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
8441 PR target/59784
8442 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
8443 SFmode to DFmode case.
8445 2014-01-29 DJ Delorie <dj@redhat.com>
8447 * config/msp430/msp430.opt (-minrt): New.
8448 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
8449 if -minrt given.
8450 (ENDFILE_SPEC): Likewise.
8452 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
8454 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
8455 (estimate_function_body_sizes): Use it.
8457 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
8459 PR c++/58561
8460 * dwarf2out.c (is_cxx_auto): New.
8461 (is_base_type): Use it.
8462 (gen_type_die_with_usage): Likewise.
8464 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8466 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
8467 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
8468 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
8469 -maltivec=be with LE targets.
8470 (vsx_mergeh_<mode>): Likewise.
8471 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
8472 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
8473 (altivec_vmrghb): Replace with define_expand and new
8474 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
8475 (altivec_vmrghb_direct): New define_insn.
8476 (altivec_vmrghh): Replace with define_expand and new
8477 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
8478 (altivec_vmrghh_direct): New define_insn.
8479 (altivec_vmrghw): Replace with define_expand and new
8480 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
8481 (altivec_vmrghw_direct): New define_insn.
8482 (*altivec_vmrghsf): Adjust for endianness.
8483 (altivec_vmrglb): Replace with define_expand and new
8484 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
8485 (altivec_vmrglb_direct): New define_insn.
8486 (altivec_vmrglh): Replace with define_expand and new
8487 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
8488 (altivec_vmrglh_direct): New define_insn.
8489 (altivec_vmrglw): Replace with define_expand and new
8490 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
8491 (altivec_vmrglw_direct): New define_insn.
8492 (*altivec_vmrglsf): Adjust for endianness.
8493 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
8494 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
8495 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
8496 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
8497 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
8498 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
8499 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
8500 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
8502 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
8504 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
8505 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
8506 whitespace.
8508 2014-01-29 Richard Biener <rguenther@suse.de>
8510 PR tree-optimization/58742
8511 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
8512 associate_pointerplus_align.
8513 (associate_pointerplus_diff): New function.
8514 (associate_pointerplus): Likewise. Call associate_pointerplus_align
8515 and associate_pointerplus_diff.
8517 2014-01-29 Richard Biener <rguenther@suse.de>
8519 * lto-streamer.h (LTO_major_version): Bump to 3.
8520 (LTO_minor_version): Reset to 0.
8522 2014-01-29 Renlin Li <Renlin.Li@arm.com>
8524 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
8525 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
8526 (arm_file_start): Generate correct asm header for armv7ve.
8527 * config/arm/bpabi.h: Add multilib support for armv7ve.
8528 * config/arm/driver-arm.c: Change the architectures of cortex-a7
8529 and cortex-a15 to armv7ve.
8530 * config/arm/t-aprofile: Add multilib support for armv7ve.
8531 * doc/invoke.texi: Document -march=armv7ve.
8533 2014-01-29 Richard Biener <rguenther@suse.de>
8535 PR tree-optimization/58742
8536 * tree-ssa-forwprop.c (associate_plusminus): Return true
8537 if we changed sth, defer EH cleanup to ...
8538 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
8539 (simplify_mult): New function.
8541 2014-01-29 Jakub Jelinek <jakub@redhat.com>
8543 PR middle-end/59917
8544 PR tree-optimization/59920
8545 * tree.c (build_common_builtin_nodes): Remove
8546 __builtin_setjmp_dispatcher initialization.
8547 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
8548 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
8549 instead of gsi_after_labels + manually skipping debug stmts.
8550 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
8551 ignore bbs with IFN_ABNORMAL_DISPATCHER.
8552 * tree-inline.c (copy_edges_for_bb): Remove
8553 can_make_abnormal_goto argument, instead add abnormal_goto_dest
8554 argument. Ignore computed_goto_p stmts. Don't call
8555 make_abnormal_goto_edges. If a call might need abnormal edges
8556 for non-local gotos, see if it already has an edge to
8557 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
8558 with true argument, don't do anything then, otherwise add
8559 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
8560 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
8561 caller.
8562 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
8563 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
8564 (lower_stmt): Don't set data->calls_builtin_setjmp.
8565 (lower_builtin_setjmp): Adjust comment.
8566 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
8567 * tree-cfg.c (found_computed_goto): Remove.
8568 (factor_computed_gotos): Remove.
8569 (make_goto_expr_edges): Return bool, true for computed gotos.
8570 Don't call make_abnormal_goto_edges.
8571 (build_gimple_cfg): Don't set found_computed_goto, don't call
8572 factor_computed_gotos.
8573 (computed_goto_p): No longer static.
8574 (make_blocks): Don't set found_computed_goto.
8575 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
8576 (make_edges): If make_goto_expr_edges returns true, push bb
8577 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
8578 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
8579 vector. Record mapping between bbs and OpenMP regions if there
8580 are any, adjust make_gimple_omp_edges caller. Call
8581 handle_abnormal_edges.
8582 (make_abnormal_goto_edges): Remove.
8583 * tree-cfg.h (make_abnormal_goto_edges): Remove.
8584 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
8585 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
8586 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
8587 * internal-fn.def (ABNORMAL_DISPATCHER): New.
8588 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
8589 filling *region also set *region_idx to (*region)->entry->index.
8591 PR other/58712
8592 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
8593 For REGs set ORIGINAL_REGNO.
8595 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
8597 * doc/md.texi: Mention that a target shouldn't implement
8598 vec_widen_(s|u)mul_even/odd pair if it is less efficient
8599 than hi/lo pair.
8601 2014-01-29 Jakub Jelinek <jakub@redhat.com>
8603 PR tree-optimization/59594
8604 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
8605 a copy of the datarefs vector rather than the vector itself.
8607 2014-01-28 Jason Merrill <jason@redhat.com>
8609 PR c++/53756
8610 * dwarf2out.c (auto_die): New static.
8611 (gen_type_die_with_usage): Handle C++1y 'auto'.
8612 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
8613 on definition.
8615 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
8617 PR target/59672
8618 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
8619 (SPEC_X32): Likewise.
8620 (SPEC_64): Likewise.
8621 * config/i386/i386.c (ix86_option_override_internal): Turn off
8622 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
8623 for TARGET_16BIT.
8624 (x86_file_start): Output .code16gcc for TARGET_16BIT.
8625 * config/i386/i386.h (TARGET_16BIT): New macro.
8626 (TARGET_16BIT_P): Likewise.
8627 * config/i386/i386.opt: Add m16.
8628 * doc/invoke.texi: Document -m16.
8630 2014-01-28 Jakub Jelinek <jakub@redhat.com>
8632 PR preprocessor/59935
8633 * input.c (location_get_source_line): Bail out on when line number
8634 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
8636 2014-01-28 Richard Biener <rguenther@suse.de>
8638 PR tree-optimization/58742
8639 * tree-ssa-forwprop.c (associate_plusminus): Handle
8640 pointer subtraction of the form (T)(P + A) - (T)P.
8642 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8644 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
8645 at const_int_cost.
8647 2014-01-28 Richard Biener <rguenther@suse.de>
8649 Revert
8650 2014-01-28 Richard Biener <rguenther@suse.de>
8652 PR rtl-optimization/45364
8653 PR rtl-optimization/59890
8654 * var-tracking.c (local_get_addr_clear_given_value): Handle
8655 already cleared slot.
8656 (val_reset): Handle not allocated local_get_addr_cache.
8657 (vt_find_locations): Use post-order on the inverted CFG.
8659 2014-01-28 Richard Biener <rguenther@suse.de>
8661 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
8663 2014-01-28 Richard Biener <rguenther@suse.de>
8665 PR rtl-optimization/45364
8666 PR rtl-optimization/59890
8667 * var-tracking.c (local_get_addr_clear_given_value): Handle
8668 already cleared slot.
8669 (val_reset): Handle not allocated local_get_addr_cache.
8670 (vt_find_locations): Use post-order on the inverted CFG.
8672 2014-01-28 Alan Modra <amodra@gmail.com>
8674 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
8675 * configure.ac <recursive call for build != host>: Define
8676 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
8677 and LD_FOR_BUILD too.
8678 * configure: Regenerate.
8680 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
8682 * config/i386/i386.c (get_builtin_code_for_version): Separate
8683 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
8684 Broadwell from Haswell.
8686 2014-01-27 Steve Ellcey <sellcey@mips.com>
8688 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
8689 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
8690 * config/mips/mips.c (mips_option_override): Change setting
8691 of TARGET_DSP.
8692 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
8693 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
8694 Change from Mask to Var.
8696 2014-01-27 Jeff Law <law@redhat.com>
8698 * ipa-inline.c (inline_small_functions): Fix typo.
8700 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
8702 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
8703 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
8704 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
8705 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
8706 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
8707 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
8708 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
8709 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
8710 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
8711 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
8712 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
8713 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
8714 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
8715 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
8716 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
8717 (_mm512_storeu_epi64): Ditto.
8718 (_mm512_cmpge_epi32_mask): Ditto.
8719 (_mm512_cmpge_epu32_mask): Ditto.
8720 (_mm512_cmpge_epi64_mask): Ditto.
8721 (_mm512_cmpge_epu64_mask): Ditto.
8722 (_mm512_cmple_epi32_mask): Ditto.
8723 (_mm512_cmple_epu32_mask): Ditto.
8724 (_mm512_cmple_epi64_mask): Ditto.
8725 (_mm512_cmple_epu64_mask): Ditto.
8726 (_mm512_cmplt_epi32_mask): Ditto.
8727 (_mm512_cmplt_epu32_mask): Ditto.
8728 (_mm512_cmplt_epi64_mask): Ditto.
8729 (_mm512_cmplt_epu64_mask): Ditto.
8730 (_mm512_cmpneq_epi32_mask): Ditto.
8731 (_mm512_cmpneq_epu32_mask): Ditto.
8732 (_mm512_cmpneq_epi64_mask): Ditto.
8733 (_mm512_cmpneq_epu64_mask): Ditto.
8734 (_mm512_expand_pd): Ditto.
8735 (_mm512_expand_ps): Ditto.
8736 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
8737 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
8738 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
8739 * config/i386/i386.c (ix86_builtins): Add
8740 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
8741 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
8742 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
8743 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
8744 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
8745 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
8746 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
8747 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
8748 IX86_BUILTIN_PMOVUSQW512_MEM.
8749 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
8750 __builtin_ia32_pmovsqd512mem_mask,
8751 __builtin_ia32_pmovqd512mem_mask,
8752 __builtin_ia32_pmovusqw512mem_mask,
8753 __builtin_ia32_pmovsqw512mem_mask,
8754 __builtin_ia32_pmovqw512mem_mask,
8755 __builtin_ia32_pmovusdw512mem_mask,
8756 __builtin_ia32_pmovsdw512mem_mask,
8757 __builtin_ia32_pmovdw512mem_mask,
8758 __builtin_ia32_pmovqb512mem_mask,
8759 __builtin_ia32_pmovusqb512mem_mask,
8760 __builtin_ia32_pmovsqb512mem_mask,
8761 __builtin_ia32_pmovusdb512mem_mask,
8762 __builtin_ia32_pmovsdb512mem_mask,
8763 __builtin_ia32_pmovdb512mem_mask.
8764 (bdesc_args): Add __builtin_ia32_expanddf512,
8765 __builtin_ia32_expandsf512.
8766 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
8767 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
8768 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
8769 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
8770 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
8771 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
8772 (avx512f_<code>v8div16qi2_mask_store): This.
8773 (avx512f_expand<mode>): New.
8775 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
8777 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
8778 New.
8779 (_mm512_mask_prefetch_i64gather_pd): Ditto.
8780 (_mm512_prefetch_i32scatter_pd): Ditto.
8781 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
8782 (_mm512_prefetch_i64scatter_pd): Ditto.
8783 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
8784 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
8785 (_mm512_mask_prefetch_i64gather_ps): Ditto.
8786 (_mm512_prefetch_i32scatter_ps): Ditto.
8787 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
8788 (_mm512_prefetch_i64scatter_ps): Ditto.
8789 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
8790 * config/i386/i386-builtin-types.def: Define
8791 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
8792 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
8793 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
8794 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
8795 IX86_BUILTIN_SCATTERPFQPD.
8796 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
8797 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
8798 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
8799 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
8800 __builtin_ia32_scatterpfqps.
8801 (ix86_expand_builtin): Expand new built-ins.
8802 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
8803 fix memory access data type.
8804 (*avx512pf_gatherpf<mode>_mask): Ditto.
8805 (*avx512pf_gatherpf<mode>): Ditto.
8806 (avx512pf_scatterpf<mode>): Ditto.
8807 (*avx512pf_scatterpf<mode>_mask): Ditto.
8808 (*avx512pf_scatterpf<mode>): Ditto.
8809 (GATHER_SCATTER_SF_MEM_MODE): New.
8810 (avx512pf_gatherpf<mode>df): Ditto.
8811 (*avx512pf_gatherpf<mode>df_mask): Ditto.
8812 (*avx512pf_scatterpf<mode>df): Ditto.
8814 2014-01-27 Jakub Jelinek <jakub@redhat.com>
8816 PR bootstrap/59934
8817 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
8818 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
8819 reached.
8821 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
8823 * common/config/arm/arm-common.c
8824 (arm_rewrite_mcpu): Handle multiple names.
8825 * config/arm/arm.h
8826 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
8828 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
8830 * gimple-builder.h (create_gimple_tmp): Delete.
8832 2014-01-27 Christian Bruel <christian.bruel@st.com>
8834 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
8835 words comparisons.
8837 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
8839 * config/pa/pa.md (call): Generate indirect long calls to non-local
8840 functions when outputing 32-bit code.
8841 (call_value): Likewise except for special call to buggy powf function.
8843 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
8844 portable runtime and PIC indirect calls.
8845 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
8846 and PIC call sequences. Use ldo instead of blr to set return register
8847 in PIC call sequence.
8849 2014-01-25 Walter Lee <walt@tilera.com>
8851 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
8852 avoid clobbering a live register.
8854 2014-01-25 Walter Lee <walt@tilera.com>
8856 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
8857 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
8858 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
8859 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
8861 2014-01-25 Walter Lee <walt@tilera.com>
8863 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
8864 arguments on even registers.
8865 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
8866 STACK_BOUNDARY.
8867 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
8868 (BIGGEST_ALIGNMENT): Ditto.
8869 (BIGGEST_FIELD_ALIGNMENT): Ditto.
8871 2014-01-25 Walter Lee <walt@tilera.com>
8873 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
8874 insns before bundling.
8875 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
8877 2014-01-25 Walter Lee <walt@tilera.com>
8879 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
8880 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
8881 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
8883 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
8885 * config/mips/constraints.md (kl): Delete.
8886 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
8887 define expands, using...
8888 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
8889 instructions for MIPS16.
8890 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
8891 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
8893 2014-01-25 Walter Lee <walt@tilera.com>
8895 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
8896 (clzdi2): Ditto.
8897 (ffsdi2): Ditto.
8899 2014-01-25 Walter Lee <walt@tilera.com>
8901 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
8902 (TARGET_EXPAND_TO_RTL_HOOK): Define.
8904 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
8906 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
8907 Handle XOR.
8909 2014-01-25 Jakub Jelinek <jakub@redhat.com>
8911 * print-rtl.c (in_call_function_usage): New var.
8912 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
8913 EXPR_LIST mode as mode and not as reg note name.
8915 PR middle-end/59561
8916 * cfgloopmanip.c (copy_loop_info): If
8917 loop->warned_aggressive_loop_optimizations, make sure
8918 the flag is set in target loop too.
8920 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
8922 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
8923 flag_cilkplus.
8924 * builtins.def: Likewise.
8925 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
8926 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
8927 * ira.c (ira_setup_eliminable_regset): Likewise.
8928 * omp-low.c (gate_expand_omp): Likewise.
8929 (execute_lower_omp): Likewise.
8930 (diagnose_sb_0): Likewise.
8931 (gate_diagnose_omp_blocks): Likewise.
8932 (simd_clone_clauses_extract): Likewise.
8933 (gate): Likewise.
8935 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8937 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
8938 correction for little endian...
8939 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
8940 here.
8942 2014-01-24 Jeff Law <law@redhat.com>
8944 PR tree-optimization/59919
8945 * tree-vrp.c (find_assert_locations_1): Do not register asserts
8946 for non-returning calls.
8948 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
8950 * common/config/aarch64/aarch64-common.c
8951 (aarch64_rewrite_mcpu): Handle multiple names.
8952 * config/aarch64/aarch64.h
8953 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
8955 2014-01-24 Dodji Seketeli <dodji@redhat.com>
8957 * input.c (add_file_to_cache_tab): Handle the case where fopen
8958 returns NULL.
8960 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
8962 PR target/59929
8963 * config/i386/i386.md (pushsf splitter): Get stack adjustment
8964 from push operand if code of push isn't PRE_DEC.
8966 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
8968 PR target/59909
8969 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
8970 -mquad-memory-atomic. Update -mquad-memory documentation to say
8971 it is only used for non-atomic loads/stores.
8973 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
8974 -mquad-memory or -mquad-memory-atomic switches.
8976 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
8977 -mquad-memory-atomic to ISA 2.07 support.
8979 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
8980 to separate support of normal quad word memory operations (ldq, stq)
8981 from the atomic quad word memory operations.
8983 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
8984 support to separate non-atomic quad word operations from atomic
8985 quad word operations. Disable non-atomic quad word operations in
8986 little endian mode so that we don't have to swap words after the
8987 load and before the store.
8988 (quad_load_store_p): Add comment about atomic quad word support.
8989 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
8990 options printed with -mdebug=reg.
8992 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
8993 -mquad-memory-atomic as the test for whether we have quad word
8994 atomic instructions.
8995 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
8996 or -mp8-vector are used, allow byte/half-word atomic operations.
8998 * config/rs6000/sync.md (load_lockedti): Insure that the address
8999 is a proper indexed or indirect address for the lqarx instruction.
9000 On little endian systems, swap the hi/lo registers after the lqarx
9001 instruction.
9002 (load_lockedpti): Use indexed_or_indirect_operand predicate to
9003 insure the address is valid for the lqarx instruction.
9004 (store_conditionalti): Insure that the address is a proper indexed
9005 or indirect address for the stqcrx. instruction. On little endian
9006 systems, swap the hi/lo registers before doing the stqcrx.
9007 instruction.
9008 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
9009 insure the address is valid for the stqcrx. instruction.
9011 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
9012 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
9013 type of quad memory support is available.
9015 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
9017 PR regression/59915
9018 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
9019 there is a danger of looping.
9021 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
9023 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
9024 force flag_ira_loop_pressure if set via command line.
9026 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
9028 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
9029 (ashr_simd): New builtin handling DI mode.
9030 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
9031 (aarch64_sshr_simddi): New match pattern.
9032 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
9033 (vshrd_n_s64): Likewise.
9034 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
9036 2014-01-23 Nick Clifton <nickc@redhat.com>
9038 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
9039 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
9040 favour of mcu specific scripts.
9041 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
9042 430x multilibs.
9044 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
9045 Alex Velenko <Alex.Velenko@arm.com>
9047 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
9048 (vaddv_s16): Likewise.
9049 (vaddv_s32): Likewise.
9050 (vaddv_u8): Likewise.
9051 (vaddv_u16): Likewise.
9052 (vaddv_u32): Likewise.
9053 (vaddvq_s8): Likewise.
9054 (vaddvq_s16): Likewise.
9055 (vaddvq_s32): Likewise.
9056 (vaddvq_s64): Likewise.
9057 (vaddvq_u8): Likewise.
9058 (vaddvq_u16): Likewise.
9059 (vaddvq_u32): Likewise.
9060 (vaddvq_u64): Likewise.
9061 (vaddv_f32): Likewise.
9062 (vaddvq_f32): Likewise.
9063 (vaddvq_f64): Likewise.
9064 (vmaxv_f32): Likewise.
9065 (vmaxv_s8): Likewise.
9066 (vmaxv_s16): Likewise.
9067 (vmaxv_s32): Likewise.
9068 (vmaxv_u8): Likewise.
9069 (vmaxv_u16): Likewise.
9070 (vmaxv_u32): Likewise.
9071 (vmaxvq_f32): Likewise.
9072 (vmaxvq_f64): Likewise.
9073 (vmaxvq_s8): Likewise.
9074 (vmaxvq_s16): Likewise.
9075 (vmaxvq_s32): Likewise.
9076 (vmaxvq_u8): Likewise.
9077 (vmaxvq_u16): Likewise.
9078 (vmaxvq_u32): Likewise.
9079 (vmaxnmv_f32): Likewise.
9080 (vmaxnmvq_f32): Likewise.
9081 (vmaxnmvq_f64): Likewise.
9082 (vminv_f32): Likewise.
9083 (vminv_s8): Likewise.
9084 (vminv_s16): Likewise.
9085 (vminv_s32): Likewise.
9086 (vminv_u8): Likewise.
9087 (vminv_u16): Likewise.
9088 (vminv_u32): Likewise.
9089 (vminvq_f32): Likewise.
9090 (vminvq_f64): Likewise.
9091 (vminvq_s8): Likewise.
9092 (vminvq_s16): Likewise.
9093 (vminvq_s32): Likewise.
9094 (vminvq_u8): Likewise.
9095 (vminvq_u16): Likewise.
9096 (vminvq_u32): Likewise.
9097 (vminnmv_f32): Likewise.
9098 (vminnmvq_f32): Likewise.
9099 (vminnmvq_f64): Likewise.
9101 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
9103 * config/aarch64/aarch64-simd.md
9104 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
9105 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
9106 (*aarch64_mul3_elt<mode>): Likewise.
9107 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
9108 (*aarch64_mul3_elt_to_64v2df): Likewise.
9109 (*aarch64_mla_elt<mode>): Likewise.
9110 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
9111 (*aarch64_mls_elt<mode>): Likewise.
9112 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
9113 (*aarch64_fma4_elt<mode>): Likewise.
9114 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
9115 (*aarch64_fma4_elt_to_64v2df): Likewise.
9116 (*aarch64_fnma4_elt<mode>): Likewise.
9117 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
9118 (*aarch64_fnma4_elt_to_64v2df): Likewise.
9119 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
9120 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
9121 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
9122 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
9123 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
9124 (aarch64_sqdmull_lane<mode>_internal): Likewise.
9125 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
9127 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
9129 * config/aarch64/aarch64-simd.md
9130 (aarch64_be_checked_get_lane<mode>): New define_expand.
9131 * config/aarch64/aarch64-simd-builtins.def
9132 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
9133 New builtin definition.
9134 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
9135 Use new safe be builtin.
9137 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
9139 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
9140 New define_insn.
9141 (aarch64_be_st1<mode>): Likewise.
9142 (aarch_ld1<VALL:mode>): Define_expand modified.
9143 (aarch_st1<VALL:mode>): Likewise.
9144 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
9145 (UNSPEC_ST1): Likewise.
9147 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
9149 * config/microblaze/microblaze.md: Add trap insn and attribute
9151 2014-01-23 Dodji Seketeli <dodji@redhat.com>
9153 PR preprocessor/58580
9154 * input.h (location_get_source_line): Take an additional line_size
9155 parameter.
9156 (void diagnostics_file_cache_fini): Declare new function.
9157 * input.c (struct fcache): New type.
9158 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
9159 New static constants.
9160 (diagnostic_file_cache_init, total_lines_num)
9161 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
9162 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
9163 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
9164 (get_next_line, read_next_line, goto_next_line, read_line_num):
9165 New static function definitions.
9166 (diagnostic_file_cache_fini): New function.
9167 (location_get_source_line): Take an additional output line_len
9168 parameter. Re-write using lookup_or_add_file_to_cache_tab and
9169 read_line_num.
9170 * diagnostic.c (diagnostic_finish): Call
9171 diagnostic_file_cache_fini.
9172 (adjust_line): Take an additional input parameter for the length
9173 of the line, rather than calculating it with strlen.
9174 (diagnostic_show_locus): Adjust the use of
9175 location_get_source_line and adjust_line with respect to their new
9176 signature. While displaying a line now, do not stop at the first
9177 null byte. Rather, display the zero byte as a space and keep
9178 going until we reach the size of the line.
9179 * Makefile.in: Add vec.o to OBJS-libcommon
9181 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
9182 Ilya Tocar <ilya.tocar@intel.com>
9184 * config/i386/avx512fintrin.h (_mm512_kmov): New.
9185 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
9186 (__builtin_ia32_kmov16): Ditto.
9187 * config/i386/i386.md (UNSPEC_KMOV): New.
9188 (kmovw): Ditto.
9190 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
9192 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
9193 (_mm512_storeu_si512): Ditto.
9195 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
9197 PR target/52125
9198 * rtl.h (get_referenced_operands): Declare.
9199 * recog.c (get_referenced_operands): New function.
9200 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
9201 operands have been referenced when recording LO_SUM references.
9203 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
9205 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
9207 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
9209 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
9210 Enable for generic and recent AMD targets.
9212 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
9214 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
9215 ARG_SIZE note when adjustment was eliminated.
9217 2014-01-22 Jeff Law <law@redhat.com>
9219 PR tree-optimization/59597
9220 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
9221 in file. Accept new argument REGISTERING and use it to modify
9222 dump output appropriately.
9223 (register_jump_thread): Corresponding changes.
9224 (mark_threaded_blocks): Reinstate code to cancel unprofitable
9225 thread paths involving joiner blocks. Add code to dump cancelled
9226 jump threading paths.
9228 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
9230 PR rtl-optimization/59477
9231 * lra-constraints.c (inherit_in_ebb): Process call for living hard
9232 regs. Update reloads_num and potential_reload_hard_regs for all insns.
9234 2014-01-22 Tom Tromey <tromey@redhat.com>
9236 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
9237 PARAMS.
9238 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
9240 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
9242 PR rtl-optimization/59896
9243 * lra-constraints.c (process_alt_operands): Check unused note for
9244 matched operands of insn with no output reloads.
9246 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
9248 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
9249 (mips_move_from_gpr_cost): Likewise.
9251 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
9253 PR rtl-optimization/59858
9254 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
9255 ira_class_hard_regs_num.
9256 (process_alt_operands): Increase reject for dying matched operand.
9258 2014-01-21 Jakub Jelinek <jakub@redhat.com>
9260 PR target/59003
9261 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
9262 smaller than size, perform several stores or loads and stores
9263 at dst + count - size to store or copy all of size bytes, rather
9264 than just last modesize bytes.
9266 2014-01-20 DJ Delorie <dj@redhat.com>
9268 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
9269 that CLOBBERs are REGs before propogating their values.
9271 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
9273 PR middle-end/59789
9274 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
9275 (cgraph_inline_failed_type): New function.
9276 * cgraph.h (DEFCIFCODE): Add type.
9277 (cgraph_inline_failed_type_t): New enum.
9278 (cgraph_inline_failed_type): New prototype.
9279 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
9280 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
9281 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
9282 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
9283 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
9284 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
9285 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
9286 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
9287 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
9288 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
9289 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
9290 OPTIMIZATION_MISMATCH.
9291 * tree-inline.c (expand_call_inline): Emit errors during
9292 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
9294 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
9296 PR target/59685
9297 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
9298 mode attribute in insn output.
9300 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
9302 * output.h (output_constant): Delete.
9303 * varasm.c (output_constant): Make private.
9305 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
9307 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
9309 2014-01-20 Jakub Jelinek <jakub@redhat.com>
9311 PR middle-end/59860
9312 * tree.h (fold_builtin_strcat): New prototype.
9313 * builtins.c (fold_builtin_strcat): No longer static. Add len
9314 argument, if non-NULL, don't call c_strlen. Optimize
9315 directly into __builtin_memcpy instead of __builtin_strcpy.
9316 (fold_builtin_2): Adjust fold_builtin_strcat caller.
9317 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
9319 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
9321 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
9322 for SImode_address_operand operands, having only a REG argument.
9324 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
9326 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
9327 loader name using mbig-endian.
9328 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
9330 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
9332 * doc/invoke.texi (-march): Clarify documentation for AArch64.
9333 (-mtune): Likewise.
9334 (-mcpu): Likewise.
9336 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
9338 * config/aarch64/aarch64-protos.h
9339 (aarch64_cannot_change_mode_class_ptr): Declare.
9340 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
9341 aarch64_cannot_change_mode_class_ptr): New.
9342 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
9343 backend hook aarch64_cannot_change_mode_class.
9345 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
9347 * common/config/aarch64/aarch64-common.c
9348 (aarch64_handle_option): Don't handle any option order logic here.
9349 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
9350 selected_cpu, warn on architecture version mismatch.
9351 (aarch64_override_options): Fix parsing order for option strings.
9353 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9354 Iain Sandoe <iain@codesourcery.com>
9356 PR bootstrap/59496
9357 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
9358 warning. Amend comment to reflect current functionality.
9360 2014-01-20 Richard Biener <rguenther@suse.de>
9362 PR middle-end/59860
9363 * builtins.c (fold_builtin_strcat): Remove case better handled
9364 by tree-ssa-strlen.c.
9366 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
9368 * config/aarch64/aarch64.opt
9369 (mcpu, march, mtune): Make case-insensitive.
9371 2014-01-20 Jakub Jelinek <jakub@redhat.com>
9373 PR target/59880
9374 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
9375 if operands[1] is a REG or ZERO_EXTEND of a REG.
9377 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
9379 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
9381 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
9383 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
9384 long non-pic millicode calls.
9386 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9388 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
9390 2014-01-19 Kito Cheng <kito@0xlab.org>
9392 * builtins.c (expand_movstr): Check movstr expand done or fail.
9394 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
9395 H.J. Lu <hongjiu.lu@intel.com>
9397 PR target/59379
9398 * config/i386/i386.md (*lea<mode>): Zero-extend return register
9399 to DImode for zero-extended addresses.
9401 2014-01-19 Jakub Jelinek <jakub@redhat.com>
9403 PR rtl-optimization/57763
9404 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
9405 on the new indirect jump_insn and increment LABEL_NUSES (label).
9407 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
9409 PR bootstrap/59580
9410 PR bootstrap/59583
9411 * config.gcc (x86_archs): New variable.
9412 (x86_64_archs): Likewise.
9413 (x86_cpus): Likewise.
9414 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
9415 --with-arch/--with-cpu= options.
9416 Support --with-arch=/--with-cpu={nehalem,westmere,
9417 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
9419 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
9421 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
9422 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
9424 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
9426 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
9428 2014-01-18 Jakub Jelinek <jakub@redhat.com>
9430 PR target/58944
9431 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
9432 clear cpp_get_options (parse_in)->warn_unused_macros for
9433 ix86_target_macros_internal with cpp_define.
9435 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
9437 * jump.c (delete_related_insns): Keep (use (insn))s.
9438 * reorg.c (redundant_insn): Check for barriers too.
9440 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
9442 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
9444 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
9446 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
9447 call to $$dyncall when TARGET_LONG_CALLS is true.
9449 2014-01-17 Jeff Law <law@redhat.com>
9451 * ree.c (combine_set_extension): Temporarily disable test for
9452 changing number of hard registers.
9454 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
9456 PR middle-end/58125
9457 * ipa-inline-analysis.c (inline_free_summary):
9458 Do not free summary of aliases.
9460 2014-01-17 Jakub Jelinek <jakub@redhat.com>
9462 PR middle-end/59706
9463 * gimplify.c (gimplify_expr): Use create_tmp_var
9464 instead of create_tmp_var_raw. If cond doesn't have
9465 integral type, don't add the IFN_ANNOTATE builtin at all.
9467 2014-01-17 Martin Jambor <mjambor@suse.cz>
9469 PR ipa/59736
9470 * ipa-cp.c (prev_edge_clone): New variable.
9471 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
9472 Also resize prev_edge_clone vector.
9473 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
9474 (ipcp_edge_removal_hook): New function.
9475 (ipcp_driver): Register ipcp_edge_removal_hook.
9477 2014-01-17 Andrew Pinski <apinski@cavium.com>
9478 Steve Ellcey <sellcey@mips.com>
9480 PR target/59462
9481 * config/mips/mips.c (mips_print_operand): Check operand mode instead
9482 of operator mode.
9484 2014-01-17 Jeff Law <law@redhat.com>
9486 PR middle-end/57904
9487 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
9488 so that pass_ccp runs first.
9490 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
9492 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
9493 (ix86_adjust_cost): Use !TARGET_XXX.
9494 (do_reorder_for_imul): Likewise.
9495 (swap_top_of_ready_list): Likewise.
9496 (ix86_sched_reorder): Likewise.
9498 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
9500 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
9501 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
9502 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
9503 (intel_memset): New. Duplicate slm_memset.
9504 (intel_cost): New. Duplicate slm_cost.
9505 (m_INTEL): New macro.
9506 (processor_target_table): Add "intel".
9507 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
9508 with PROCESSOR_INTEL for "intel".
9509 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
9510 PROCESSOR_SILVERMONT.
9511 (ix86_issue_rate): Likewise.
9512 (ix86_adjust_cost): Likewise.
9513 (ia32_multipass_dfa_lookahead): Likewise.
9514 (swap_top_of_ready_list): Likewise.
9515 (ix86_sched_reorder): Likewise.
9516 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
9517 instead of TARGET_OPT_AGU.
9518 * config/i386/i386.h (TARGET_INTEL): New.
9519 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
9520 (processor_type): Add PROCESSOR_INTEL.
9521 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
9522 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
9524 2014-01-17 Marek Polacek <polacek@redhat.com>
9526 PR c/58346
9527 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
9528 size is zero.
9530 2014-01-17 Richard Biener <rguenther@suse.de>
9532 PR tree-optimization/46590
9533 * opts.c (default_options_table): Add entries for
9534 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
9535 all enabled at -O1 but not for -Og.
9536 * common.opt (fbranch-count-reg): Remove Init(1).
9537 (fmove-loop-invariants): Likewise.
9538 (ftree-pta): Likewise.
9540 2014-01-17 Jakub Jelinek <jakub@redhat.com>
9542 * config/i386/i386.c (ix86_data_alignment): For compatibility with
9543 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
9544 decls to at least the GCC 4.8 used alignments.
9546 PR fortran/59440
9547 * tree-nested.c (convert_nonlocal_reference_stmt,
9548 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
9549 of GIMPLE_BIND stmts, adjust associated decls.
9551 2014-01-17 Richard Biener <rguenther@suse.de>
9553 PR tree-optimization/46590
9554 * vec.h (vec<>::bseach): New member function implementing
9555 binary search according to C89 bsearch.
9556 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
9557 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
9558 bitmap pointer again. Make accesses_in_loop a flat array.
9559 (mem_ref_obstack): New global.
9560 (outermost_indep_loop): Adjust for mem_ref->stored changes.
9561 (mark_ref_stored): Likewise.
9562 (ref_indep_loop_p_2): Likewise.
9563 (set_ref_stored_in_loop): New helper function.
9564 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
9565 (memref_free): Adjust.
9566 (record_mem_ref_loc): Simplify.
9567 (gather_mem_refs_stmt): Adjust.
9568 (sort_locs_in_loop_postorder_cmp): New function.
9569 (analyze_memory_references): Sort accesses_in_loop after
9570 loop postorder number.
9571 (find_ref_loc_in_loop_cmp): New function.
9572 (for_all_locs_in_loop): Find relevant cluster of locs in
9573 accesses_in_loop and iterate without recursion.
9574 (execute_sm): Avoid uninit warning.
9575 (struct ref_always_accessed): Simplify.
9576 (ref_always_accessed::operator ()): Likewise.
9577 (ref_always_accessed_p): Likewise.
9578 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
9579 loop postorder numbers here.
9580 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
9581 numbers.
9583 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
9585 PR c++/57945
9586 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
9587 on decls for which assemble_alias has been called.
9589 2014-01-17 Nick Clifton <nickc@redhat.com>
9591 * config/msp430/msp430.opt: (mcpu): New option.
9592 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
9593 (msp430_option_override): Parse target_cpu. If the MCU name
9594 matches a generic string, clear target_mcu.
9595 (msp430_attr): Allow numeric interrupt values up to 63.
9596 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
9597 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
9598 option.
9599 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
9600 Add mcpu matches.
9601 * config/msp430/msp430.md (popm): Use %J rather than %I.
9602 (addsi3): Use msp430_nonimmediate_operand for operand 2.
9603 (addhi_cy_i): Use immediate_operand for operand 2.
9604 * doc/invoke.texi: Document -mcpu option.
9606 2014-01-17 Richard Biener <rguenther@suse.de>
9608 PR rtl-optimization/38518
9609 * df.h (df_analyze_loop): Declare.
9610 * df-core.c: Include cfgloop.h.
9611 (df_analyze_1): Split out main part of df_analyze.
9612 (df_analyze): Adjust.
9613 (loop_inverted_post_order_compute): New function.
9614 (loop_post_order_compute): Likewise.
9615 (df_analyze_loop): New function avoiding whole-function
9616 postorder computes.
9617 * loop-invariant.c (find_defs): Use df_analyze_loop.
9618 (find_invariants): Adjust.
9619 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
9621 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
9623 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
9624 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
9626 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
9628 * ipa-ref.c (ipa_remove_stmt_references): Fix references
9629 traversal when removing references.
9631 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
9633 PR ipa/59775
9634 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
9636 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
9638 PR middle-end/56791
9639 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
9640 pushing a reload for an autoinc when we had previously reloaded an
9641 inner part of the address.
9643 2014-01-16 Jakub Jelinek <jakub@redhat.com>
9645 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
9646 field.
9647 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
9648 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
9649 when not giving up or versioning for alias only because of
9650 loop->safelen.
9651 (vect_analyze_data_ref_dependences): Set to true.
9652 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
9653 is a GIMPLE_PHI.
9654 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
9655 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
9656 to the condition.
9658 PR middle-end/58344
9659 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
9661 PR target/59839
9662 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
9663 operand 0 predicate for gathers, use a new pseudo as subtarget.
9665 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
9667 PR middle-end/59609
9668 * lra-constraints.c (process_alt_operands): Add printing debug info.
9669 Check absence of input/output reloads for matched operands too.
9671 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
9673 PR rtl-optimization/59835
9674 * ira.c (ira_init_register_move_cost): Increase cost for
9675 impossible modes.
9677 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
9679 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
9681 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
9683 PR target/59780
9684 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
9685 non-register objects. Use gen_(high/low)part more consistently.
9686 Fix assertions.
9688 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
9690 PR target/59844
9691 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
9692 endian support, remove tests for WORDS_BIG_ENDIAN.
9693 (p8_mfvsrd_3_<mode>): Likewise.
9694 (reload_gpr_from_vsx<mode>): Likewise.
9695 (reload_gpr_from_vsxsf): Likewise.
9696 (p8_mfvsrd_4_disf): Likewise.
9698 2014-01-16 Richard Biener <rguenther@suse.de>
9700 PR rtl-optimization/46590
9701 * lcm.c (compute_antinout_edge): Use postorder iteration.
9702 (compute_laterin): Use inverted postorder iteration.
9704 2014-01-16 Nick Clifton <nickc@redhat.com>
9706 PR middle-end/28865
9707 * varasm.c (output_constant): Return the number of bytes actually
9708 emitted.
9709 (output_constructor_array_range): Update the field size with the
9710 number of bytes emitted by output_constant.
9711 (output_constructor_regular_field): Likewise. Also do not
9712 complain if the total number of bytes emitted is now greater
9713 than the expected fieldpos.
9714 * output.h (output_constant): Update prototype and descriptive comment.
9716 2014-01-16 Marek Polacek <polacek@redhat.com>
9718 PR middle-end/59827
9719 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
9720 it is error_mark_node.
9722 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
9724 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
9725 VALID_AVX256_REG_OR_OI_MODE.
9727 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
9729 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
9730 current procedure should be profiled.
9732 2014-01-15 Andrew Pinski <apinski@cavium.com>
9734 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
9735 of moving from/to the STACK_REG register class.
9737 2014-01-15 Richard Henderson <rth@redhat.com>
9739 PR debug/54694
9740 * reginfo.c (global_regs_decl): Globalize.
9741 * rtl.h (global_regs_decl): Declare.
9742 * ira.c (do_reload): Diagnose frame_pointer_needed and it
9743 reserved via global_regs.
9745 2014-01-15 Teresa Johnson <tejohnson@google.com>
9747 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
9749 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
9751 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
9752 and vmulosh rather than call gen_vec_widen_smult_*.
9753 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
9754 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
9755 (vec_widen_smult_even_v16qi): Likewise.
9756 (vec_widen_umult_even_v8hi): Likewise.
9757 (vec_widen_smult_even_v8hi): Likewise.
9758 (vec_widen_umult_odd_v16qi): Likewise.
9759 (vec_widen_smult_odd_v16qi): Likewise.
9760 (vec_widen_umult_odd_v8hi): Likewise.
9761 (vec_widen_smult_odd_v8hi): Likewise.
9762 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
9763 vmuloub rather than call gen_vec_widen_umult_*.
9764 (vec_widen_umult_lo_v16qi): Likewise.
9765 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
9766 vmulosb rather than call gen_vec_widen_smult_*.
9767 (vec_widen_smult_lo_v16qi): Likewise.
9768 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
9769 rather than call gen_vec_widen_umult_*.
9770 (vec_widen_umult_lo_v8hi): Likewise.
9771 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
9772 rather than call gen_vec_widen_smult_*.
9773 (vec_widen_smult_lo_v8hi): Likewise.
9775 2014-01-15 Jeff Law <law@redhat.com>
9777 PR tree-optimization/59747
9778 * ree.c (find_and_remove_re): Properly handle case where a second
9779 eliminated extension requires widening a copy created for elimination
9780 of a prior extension.
9781 (combine_set_extension): Ensure that the number of hard regs needed
9782 for a destination register does not change when we widen it.
9784 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
9786 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
9787 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
9788 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
9789 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
9790 (avr-*-rtems*): Likewise.
9791 (bfin*-rtems*): Likewise.
9792 (moxie-*-rtems*): Likewise.
9793 (h8300-*-rtems*): Likewise.
9794 (i[34567]86-*-rtems*): Likewise.
9795 (lm32-*-rtems*): Likewise.
9796 (m32r-*-rtems*): Likewise.
9797 (m68k-*-rtems*): Likewise.
9798 (microblaze*-*-rtems*): Likewise.
9799 (mips*-*-rtems*): Likewise.
9800 (powerpc-*-rtems*): Likewise.
9801 (sh-*-rtems*): Likewise.
9802 (sparc-*-rtems*): Likewise.
9803 (sparc64-*-rtems*): Likewise.
9804 (v850-*-rtems*): Likewise.
9805 (m32c-*-rtems*): Likewise.
9807 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
9809 PR rtl-optimization/59511
9810 * ira.c (ira_init_register_move_cost): Use memory costs for some
9811 cases of register move cost calculations.
9812 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
9813 instead of BB frequency.
9814 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
9815 * lra-assigns.c (find_hard_regno_for): Ditto.
9817 2014-01-15 Richard Biener <rguenther@suse.de>
9819 PR tree-optimization/59822
9820 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
9821 (vectorizable_load): Use it to hoist defs of uses of invariant
9822 loads out of the loop.
9824 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
9825 Kugan Vivekanandarajah <kuganv@linaro.org>
9827 PR target/59695
9828 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
9829 truncation.
9831 2014-01-15 Richard Biener <rguenther@suse.de>
9833 PR rtl-optimization/59802
9834 * lcm.c (compute_available): Use inverted postorder to seed
9835 the initial worklist.
9837 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9839 PR target/59803
9840 * config/s390/s390.c (s390_preferred_reload_class): Don't return
9841 ADDR_REGS for invalid symrefs in non-PIC code.
9843 2014-01-15 Jakub Jelinek <jakub@redhat.com>
9845 PR other/58712
9846 * builtins.c (determine_block_size): Initialize *probable_max_size
9847 even if len_rtx is CONST_INT.
9849 2014-01-14 Andrew Pinski <apinski@cavium.com>
9851 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
9852 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
9853 (cortexa53_tunings): Likewise.
9854 (aarch64_sched_issue_rate): New function.
9855 (TARGET_SCHED_ISSUE_RATE): Define.
9857 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
9859 * ira-costs.c (find_costs_and_classes): Add missed
9860 ira_init_register_move_cost_if_necessary.
9862 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
9864 PR target/59787
9865 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
9867 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
9869 PR target/59794
9870 * config/i386/i386.c (type_natural_mode): Add a bool parameter
9871 to indicate if type is used for function return value. Warn ABI
9872 change if the vector mode isn't available for function return value.
9873 (ix86_function_arg_advance): Pass false to type_natural_mode.
9874 (ix86_function_arg): Likewise.
9875 (ix86_gimplify_va_arg): Likewise.
9876 (function_arg_32): Don't warn ABI change.
9877 (ix86_function_value): Pass true to type_natural_mode.
9878 (ix86_return_in_memory): Likewise.
9879 (ix86_struct_value_rtx): Removed.
9880 (TARGET_STRUCT_VALUE_RTX): Likewise.
9882 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9884 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
9885 converting a conditional jump into a conditional return.
9887 2014-01-14 Richard Biener <rguenther@suse.de>
9889 PR tree-optimization/58921
9890 PR tree-optimization/59006
9891 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
9892 hoisting invariant stmts.
9893 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
9894 invariant loads on the preheader edge if possible.
9896 2014-01-14 Joey Ye <joey.ye@arm.com>
9898 * doc/plugin.texi (Building GCC plugins): Update to C++.
9900 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
9902 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
9903 (_mm_rcp28_round_ss): Ditto.
9904 (_mm_rsqrt28_round_sd): Ditto.
9905 (_mm_rsqrt28_round_ss): Ditto.
9906 (_mm_rcp28_sd): Ditto.
9907 (_mm_rcp28_ss): Ditto.
9908 (_mm_rsqrt28_sd): Ditto.
9909 (_mm_rsqrt28_ss): Ditto.
9910 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
9911 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
9912 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
9913 (IX86_BUILTIN_RCP28SD): Ditto.
9914 (IX86_BUILTIN_RCP28SS): Ditto.
9915 (IX86_BUILTIN_RSQRT28SD): Ditto.
9916 (IX86_BUILTIN_RSQRT28SS): Ditto.
9917 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
9918 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
9919 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
9920 (ix86_expand_special_args_builtin): Expand new FTYPE.
9921 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
9922 (srcp14<mode>): Make insn unary.
9923 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
9924 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
9925 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
9926 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
9927 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
9928 Fix rounding: make it SAE only.
9929 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
9930 Ditto.
9931 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
9932 Ditto.
9933 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
9934 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
9935 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
9936 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
9937 (round_saeonly_mask_scalar_operand4): Ditto.
9938 (round_saeonly_mask_scalar_op3): Ditto.
9939 (round_saeonly_mask_scalar_op4): Ditto.
9941 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9943 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
9944 Implement -maltivec=be for vec_insert and vec_extract.
9946 2014-01-10 DJ Delorie <dj@redhat.com>
9948 * config/msp430/msp430.md (call_internal): Don't allow memory
9949 references with SP as the base register.
9950 (call_value_internal): Likewise.
9951 * config/msp430/constraints.md (Yc): New. For memory references
9952 that don't use SP as a base register.
9954 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
9955 "an integer without a # prefix"
9956 * config/msp430/msp430.md (epilogue_helper): Use it.
9958 2014-01-13 Jakub Jelinek <jakub@redhat.com>
9960 PR target/59617
9961 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
9962 AVX512F gather builtins.
9963 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
9964 on gather decls with INTEGER_TYPE masktype.
9965 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
9966 directly into the builtin rather than hoisting it before loop.
9968 PR tree-optimization/59387
9969 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
9970 (scev_const_prop): If folded_casts and type has undefined overflow,
9971 use force_gimple_operand instead of force_gimple_operand_gsi and
9972 for each added stmt if it is assign with
9973 arith_code_with_undefined_signed_overflow, call
9974 rewrite_to_defined_overflow.
9975 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
9976 gimple-fold.h instead.
9977 (arith_code_with_undefined_signed_overflow,
9978 rewrite_to_defined_overflow): Moved to ...
9979 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
9980 rewrite_to_defined_overflow): ... here. No longer static.
9981 Include gimplify-me.h.
9982 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
9983 rewrite_to_defined_overflow): New prototypes.
9985 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9987 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
9989 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
9991 * builtins.c (get_object_alignment_2): Minor tweak.
9992 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
9994 2014-01-13 Christian Bruel <christian.bruel@st.com>
9996 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
9997 optimized non constant lengths.
9999 2014-01-13 Jakub Jelinek <jakub@redhat.com>
10001 PR libgomp/59194
10002 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
10003 load as __atomic_load_N if possible.
10005 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
10007 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
10008 target parameter.
10009 (rs6000_expand_builtin): Adjust call.
10011 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
10013 PR target/58115
10014 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
10015 * config/rs6000/rs6000.c: Include target-globals.h.
10016 (rs6000_set_current_function): Instead of doing target_reinit
10017 unconditionally, use save_target_globals_default_opts and
10018 restore_target_globals.
10020 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
10021 FPSCR.
10022 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
10023 (rs6000_expand_builtin): Handle mffs and mtfsf.
10024 (rs6000_init_builtins): Define mffs and mtfsf.
10025 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
10026 (rs6000_mffs): New pattern.
10027 (rs6000_mtfsf): New pattern.
10029 2014-01-11 Bin Cheng <bin.cheng@arm.com>
10031 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
10032 Start narrowing with START. Apply candidate-use pair
10033 and check overall cost in narrowing.
10034 (iv_ca_prune): Pass new argument.
10036 2014-01-10 Jeff Law <law@redhat.com>
10038 PR middle-end/59743
10039 * ree.c (combine_reaching_defs): Ensure the defining statement
10040 occurs before the extension when optimizing extensions with
10041 different source and destination hard registers.
10043 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
10045 PR ipa/58585
10046 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
10047 vtables into the type inheritance graph.
10049 2014-01-10 Jakub Jelinek <jakub@redhat.com>
10051 PR rtl-optimization/59754
10052 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
10053 modes in the REGNO != REGNO case.
10055 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10057 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
10059 2014-01-10 Jakub Jelinek <jakub@redhat.com>
10061 PR tree-optimization/59745
10062 * tree-predcom.c (tree_predictive_commoning_loop): Call
10063 free_affine_expand_cache if giving up because components is NULL.
10065 * target-globals.c (save_target_globals): Allocate < 4KB structs using
10066 GC in payload of target_globals struct instead of allocating them on
10067 the heap and the larger structs separately using GC.
10068 * target-globals.h (struct target_globals): Make regs, hard_regs,
10069 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
10070 of GTY((skip)) and change type to void *.
10071 (reset_target_globals): Cast loads from those fields to corresponding
10072 types.
10074 2014-01-10 Steve Ellcey <sellcey@mips.com>
10076 PR plugins/59335
10077 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
10078 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
10079 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
10081 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
10083 PR target/59744
10084 * aarch64-modes.def (CC_Zmode): New flags mode.
10085 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
10086 represents an equality.
10087 (aarch64_get_condition_code): Handle CC_Zmode.
10088 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
10090 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10092 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
10093 extraction in good case.
10095 2014-01-10 Richard Biener <rguenther@suse.de>
10097 PR tree-optimization/59374
10098 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
10099 checking after SLP discovery. Mark stmts not participating
10100 in any SLP instance properly.
10102 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10104 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
10105 when handling a SET rtx.
10107 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10109 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
10110 (cortex-a57): Likewise.
10111 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
10113 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10115 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
10116 non-iwmmxt builtins.
10118 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
10120 PR ipa/58252
10121 PR ipa/59226
10122 * ipa-devirt.c record_target_from_binfo): Take as argument
10123 stack of binfos and lookup matching one for virtual inheritance.
10124 (possible_polymorphic_call_targets_1): Update.
10126 2014-01-10 Huacai Chen <chenhc@lemote.com>
10128 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
10129 kernel strings for Loongson-2E/2F/3A.
10131 2014-01-10 Jakub Jelinek <jakub@redhat.com>
10133 PR middle-end/59670
10134 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
10135 is_gimple_call before calling gimple_call_internal_p.
10137 2014-01-09 Steve Ellcey <sellcey@mips.com>
10139 * Makefile.in (TREE_FLOW_H): Remove.
10140 (TREE_SSA_H): Add file names from tree-flow.h.
10141 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
10142 * tree.h: Remove tree-flow.h reference.
10143 * hash-table.h: Remove tree-flow.h reference.
10144 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
10145 reference with tree-ssa-loop.h.
10147 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10149 * doc/invoke.texi: Add -maltivec={be,le} options, and document
10150 default element-order behavior for -maltivec.
10151 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
10152 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
10153 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
10154 when targeting big endian, at least for now.
10155 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
10157 2014-01-09 Jakub Jelinek <jakub@redhat.com>
10159 PR middle-end/47735
10160 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
10161 var satisfies use_register_for_decl, just take into account type
10162 alignment, rather than decl alignment.
10164 PR tree-optimization/59622
10165 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
10166 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
10167 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
10168 Don't devirtualize for inplace at all. For targets.length () == 1,
10169 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
10171 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
10173 * config/i386/i386.md (cpu): Remove the unused btver1.
10175 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
10177 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
10179 2014-01-09 Jakub Jelinek <jakub@redhat.com>
10181 PR target/58115
10182 * tree-core.h (struct target_globals): New forward declaration.
10183 (struct tree_target_option): Add globals field.
10184 * tree.h (TREE_TARGET_GLOBALS): Define.
10185 (prepare_target_option_nodes_for_pch): New prototype.
10186 * target-globals.h (struct target_globals): Define even if
10187 !SWITCHABLE_TARGET.
10188 * tree.c (prepare_target_option_node_for_pch,
10189 prepare_target_option_nodes_for_pch): New functions.
10190 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
10191 * config/i386/i386.c: Include target-globals.h.
10192 (ix86_set_current_function): Instead of doing target_reinit
10193 unconditionally, use save_target_globals_default_opts and
10194 restore_target_globals.
10196 2014-01-09 Richard Biener <rguenther@suse.de>
10198 PR tree-optimization/59715
10199 * tree-cfg.h (split_critical_edges): Declare.
10200 * tree-cfg.c (split_critical_edges): Export.
10201 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
10203 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
10205 * cfgexpand.c (expand_stack_vars): Optionally disable
10206 asan stack protection.
10207 (expand_used_vars): Likewise.
10208 (partition_stack_vars): Likewise.
10209 * asan.c (asan_emit_stack_protection): Optionally disable
10210 after return stack usage.
10211 (instrument_derefs): Optionally disable memory access instrumentation.
10212 (instrument_builtin_call): Likewise.
10213 (instrument_strlen_call): Likewise.
10214 (asan_protect_global): Optionally disable global variables protection.
10215 * doc/invoke.texi: Added doc for new options.
10216 * params.def: Added new options.
10217 * params.h: Likewise.
10219 2014-01-09 Jakub Jelinek <jakub@redhat.com>
10221 PR rtl-optimization/59724
10222 * ifcvt.c (cond_exec_process_if_block): Don't call
10223 flow_find_head_matching_sequence with 0 longest_match.
10224 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
10225 non-active insns if !stop_after.
10226 (try_head_merge_bb): Revert 2014-01-07 changes.
10228 2014-01-08 Jeff Law <law@redhat.com>
10230 * ree.c (get_sub_rtx): New function, extracted from...
10231 (merge_def_and_ext): Here.
10232 (combine_reaching_defs): Use get_sub_rtx.
10234 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
10236 * cgraph.h (varpool_variable_node): Do not choke on null node.
10238 2014-01-08 Catherine Moore <clm@codesourcery.com>
10240 * config/mips/mips.md (simple_return): Attempt to use JRC
10241 for microMIPS.
10242 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
10244 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
10246 PR rtl-optimization/59137
10247 * reorg.c (steal_delay_list_from_target): Call update_block for
10248 elided insns.
10249 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
10251 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10253 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
10254 two duplicate entries.
10256 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
10258 Revert:
10259 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
10261 * config/mips/mips.c (mips_truncated_op_cost): New function.
10262 (mips_rtx_costs): Adjust test for BADDU.
10263 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
10265 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
10267 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
10268 (*baddu_si): ...this new pattern.
10270 2014-01-08 Jakub Jelinek <jakub@redhat.com>
10272 PR ipa/59722
10273 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
10275 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
10277 PR middle-end/57748
10278 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
10279 inner_reference_p.
10280 (expand_expr, expand_normal): Adjust.
10281 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
10282 inner_reference_p. Use inner_reference_p to expand inner references.
10283 (store_expr): Adjust.
10284 * cfgexpand.c (expand_call_stmt): Adjust.
10286 2014-01-08 Rong Xu <xur@google.com>
10288 * gcov-io.c (gcov_var): Move from gcov-io.h.
10289 (gcov_position): Ditto.
10290 (gcov_is_error): Ditto.
10291 (gcov_rewrite): Ditto.
10292 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
10293 only part to libgcc/libgcov.h.
10295 2014-01-08 Marek Polacek <polacek@redhat.com>
10297 PR middle-end/59669
10298 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
10300 2014-01-08 Marek Polacek <polacek@redhat.com>
10302 PR sanitizer/59667
10303 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
10305 2014-01-08 Jakub Jelinek <jakub@redhat.com>
10307 PR rtl-optimization/59649
10308 * stor-layout.c (get_mode_bounds): For BImode return
10309 0 and STORE_FLAG_VALUE.
10311 2014-01-08 Richard Biener <rguenther@suse.de>
10313 PR middle-end/59630
10314 * gimple.h (is_gimple_builtin_call): Remove.
10315 (gimple_builtin_call_types_compatible_p): New.
10316 (gimple_call_builtin_p): New overload.
10317 * gimple.c (is_gimple_builtin_call): Remove.
10318 (validate_call): Rename to ...
10319 (gimple_builtin_call_types_compatible_p): ... this and export. Also
10320 check return types.
10321 (validate_type): New static function.
10322 (gimple_call_builtin_p): New overload and adjust.
10323 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
10324 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
10325 (gimple_fold_stmt_to_constant_1): Likewise.
10326 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
10328 2014-01-08 Richard Biener <rguenther@suse.de>
10330 PR middle-end/59471
10331 * gimplify.c (gimplify_expr): Gimplify register-register type
10332 VIEW_CONVERT_EXPRs to separate stmts.
10334 2014-01-07 Jeff Law <law@redhat.com>
10336 PR middle-end/53623
10337 * ree.c (combine_set_extension): Handle case where source
10338 and destination registers in an extension insn are different.
10339 (combine_reaching_defs): Allow source and destination registers
10340 in extension to be different under limited circumstances.
10341 (add_removable_extension): Remove restriction that the
10342 source and destination registers in the extension are the same.
10343 (find_and_remove_re): Emit a copy from the extension's
10344 destination to its source after the defining insn if
10345 the source and destination registers are different.
10347 PR middle-end/59285
10348 * ifcvt.c (merge_if_block): If we are merging a block with more than
10349 one successor with a block with no successors, remove any BARRIER
10350 after the second block.
10352 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
10354 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
10356 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
10358 PR target/59652
10359 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
10360 for 14-bit register offsets when INT14_OK_STRICT is false.
10362 2014-01-07 Roland Stigge <stigge@antcom.de>
10363 Michael Meissner <meissner@linux.vnet.ibm.com>
10365 PR 57386/target
10366 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
10367 Only check TFmode for SPE constants. Don't check TImode or TDmode.
10369 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
10371 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
10372 -mcpu.
10374 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
10376 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
10377 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
10378 rtx is const0_rtx or not.
10380 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
10382 PR target/58115
10383 * target-globals.c (save_target_globals): Remove this_fn_optab
10384 handling.
10385 * toplev.c: Include optabs.h.
10386 (target_reinit): Temporarily restore the global options if another
10387 set of options are in force.
10389 2014-01-07 Jakub Jelinek <jakub@redhat.com>
10391 PR rtl-optimization/58668
10392 * cfgcleanup.c (flow_find_cross_jump): Don't count
10393 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
10394 to determine what is counted.
10395 (flow_find_head_matching_sequence): Use active_insn_p to determine
10396 what is counted.
10397 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
10398 counting change.
10399 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
10400 determine what is counted.
10402 PR tree-optimization/59643
10403 * tree-predcom.c (split_data_refs_to_components): If one dr is
10404 read and one write, determine_offset fails and the write isn't
10405 in the bad component, just put the read into the bad component.
10407 2014-01-07 Mike Stump <mikestump@comcast.net>
10408 Jakub Jelinek <jakub@redhat.com>
10410 PR pch/59436
10411 * tree-core.h (struct tree_optimization_option): Change optabs
10412 type from unsigned char * to void *.
10413 * optabs.c (init_tree_optimization_optabs): Adjust
10414 TREE_OPTIMIZATION_OPTABS initialization.
10416 2014-01-06 Jakub Jelinek <jakub@redhat.com>
10418 PR target/59644
10419 * config/i386/i386.h (struct machine_function): Add
10420 no_drap_save_restore field.
10421 * config/i386/i386.c (ix86_save_reg): Use
10422 !cfun->machine->no_drap_save_restore instead of
10423 crtl->stack_realign_needed.
10424 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
10425 this function clears frame_pointer_needed. Set
10426 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
10427 and DRAP reg is needed.
10429 2014-01-06 Marek Polacek <polacek@redhat.com>
10431 PR c/57773
10432 * doc/implement-c.texi: Mention that other integer types are
10433 permitted as bit-field types in strictly conforming mode.
10435 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
10437 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
10438 is newly allocated.
10440 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
10442 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
10444 2014-01-06 Martin Jambor <mjambor@suse.cz>
10446 PR ipa/59008
10447 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
10448 to int.
10449 * ipa-prop.c (ipa_print_node_params): Fix indentation.
10451 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
10453 PR debug/59350
10454 PR debug/59510
10455 * var-tracking.c (add_stores): Preserve the value of the source even if
10456 we don't record the store.
10458 2014-01-06 Terry Guo <terry.guo@arm.com>
10460 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
10462 2014-01-05 Iain Sandoe <iain@codesourcery.com>
10464 PR bootstrap/59541
10465 * config/darwin.c (darwin_function_section): Adjust return values to
10466 correspond to optimisation changes made in r206070.
10468 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
10470 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
10471 from prefetch_block tune setting.
10472 (nocona_cost): Correct size of prefetch block to 64.
10474 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
10476 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
10477 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
10478 used to save the static chain register in the computation of the offset
10479 from which the FP registers need to be restored.
10481 2014-01-04 Jakub Jelinek <jakub@redhat.com>
10483 PR tree-optimization/59519
10484 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
10485 ICE if get_current_def (current_new_name) is already non-NULL, as long
10486 as it is a phi result of some other phi in *new_exit_bb that has
10487 the same argument.
10489 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
10490 or vmovdqu* for misaligned_operand.
10491 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
10492 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
10493 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
10494 aligned_mem for AVX512F masked aligned load and store builtins and for
10495 non-temporal moves.
10497 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
10499 PR tree-optimization/59651
10500 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
10501 Address range for negative step should be added by TYPE_SIZE_UNIT.
10503 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
10505 * config/m68k/m68k.c (handle_move_double): Handle pushes with
10506 overlapping registers also for registers other than the stack pointer.
10508 2014-01-03 Marek Polacek <polacek@redhat.com>
10510 PR other/59661
10511 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
10512 __builtin_FILE.
10514 2014-01-03 Jakub Jelinek <jakub@redhat.com>
10516 PR target/59625
10517 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
10518 asm goto as jump.
10520 * config/i386/i386.md (MODE_SIZE): New mode attribute.
10521 (push splitter): Use <P:MODE_SIZE> instead of
10522 GET_MODE_SIZE (<P:MODE>mode).
10523 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
10524 (mov -1, reg peephole2): Likewise.
10525 * config/i386/sse.md (*mov<mode>_internal,
10526 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
10527 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
10528 *<code><mode>3, *andnot<mode>3<mask_name>,
10529 <mask_codefor><code><mode>3<mask_name>): Likewise.
10530 * config/i386/subst.md (mask_mode512bit_condition,
10531 sd_mask_mode512bit_condition): Likewise.
10533 2014-01-02 Xinliang David Li <davidxl@google.com>
10535 PR tree-optimization/59303
10536 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
10537 (dump_predicates): Better output format.
10538 (pred_equal_p): New function.
10539 (is_neq_relop_p): Ditto.
10540 (is_neq_zero_form_p): Ditto.
10541 (pred_expr_equal_p): Ditto.
10542 (pred_neg_p): Ditto.
10543 (simplify_pred): Ditto.
10544 (simplify_preds_2): Ditto.
10545 (simplify_preds_3): Ditto.
10546 (simplify_preds_4): Ditto.
10547 (simplify_preds): Ditto.
10548 (push_pred): Ditto.
10549 (push_to_worklist): Ditto.
10550 (get_pred_info_from_cmp): Ditto.
10551 (is_degenerated_phi): Ditto.
10552 (normalize_one_pred_1): Ditto.
10553 (normalize_one_pred): Ditto.
10554 (normalize_one_pred_chain): Ditto.
10555 (normalize_preds): Ditto.
10556 (normalize_cond_1): Remove function.
10557 (normalize_cond): Ditto.
10558 (is_gcond_subset_of): Ditto.
10559 (is_subset_of_any): Ditto.
10560 (is_or_set_subset_of): Ditto.
10561 (is_and_set_subset_of): Ditto.
10562 (is_norm_cond_subset_of): Ditto.
10563 (pred_chain_length_cmp): Ditto.
10564 (convert_control_dep_chain_into_preds): Type change.
10565 (find_predicates): Ditto.
10566 (find_def_preds): Ditto.
10567 (destroy_predicates_vecs): Ditto.
10568 (find_matching_predicates_in_rest_chains): Ditto.
10569 (use_pred_not_overlap_with_undef_path_pred): Ditto.
10570 (is_pred_expr_subset): Ditto.
10571 (is_pred_chain_subset_of): Ditto.
10572 (is_included_in): Ditto.
10573 (is_superset_of): Ditto.
10575 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
10577 Update copyright years.
10579 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
10581 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
10582 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
10583 config/arc/arc.md, config/arc/arc.opt,
10584 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
10585 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
10586 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
10587 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
10588 config/linux-protos.h, config/linux.c, config/winnt-c.c,
10589 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
10590 vtable-verify.c, vtable-verify.h: Use the standard form for the
10591 copyright notice.
10593 2014-01-02 Tobias Burnus <burnus@net-b.de>
10595 * gcc.c (process_command): Update copyright notice dates.
10596 * gcov-dump.c: Ditto.
10597 * gcov.c: Ditto.
10598 * doc/cpp.texi: Bump @copying's copyright year.
10599 * doc/cppinternals.texi: Ditto.
10600 * doc/gcc.texi: Ditto.
10601 * doc/gccint.texi: Ditto.
10602 * doc/gcov.texi: Ditto.
10603 * doc/install.texi: Ditto.
10604 * doc/invoke.texi: Ditto.
10606 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
10608 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
10610 2014-01-01 Jakub Jelinek <jakub@redhat.com>
10612 * config/i386/sse.md (*mov<mode>_internal): Guard
10613 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
10615 PR rtl-optimization/59647
10616 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
10617 new_rtx into UNSIGNED_FLOAT rtxes.
10618 \f
10619 Copyright (C) 2014 Free Software Foundation, Inc.
10621 Copying and distribution of this file, with or without modification,
10622 are permitted in any medium without royalty provided the copyright
10623 notice and this notice are preserved.