1 2017-01-30 Martin Liska <mliska@suse.cz>
4 * opts.c (common_handle_option): Enable flag_ipa_bit_cp w/
7 2017-01-30 Martin Liska <mliska@suse.cz>
10 * config/aarch64/cortex-a57-fma-steering.c (func_fma_steering::analyze):
11 Initialize variables with NULL value.
13 2017-01-30 Richard Earnshaw <rearnsh@arm.com>
16 * config.gcc (arm*-*-*): Add arm/arm-flags.h and arm/arm-isa.h to
18 * arm/arm-protos.h: Don't directly include arm-flags.h and arm-isa.h.
20 2017-01-30 Richard Biener <rguenther@suse.de>
22 PR tree-optimization/79276
23 * tree-vrp.c (process_assert_insertions): Properly adjust common
24 when removing a duplicate.
26 * gcc.dg/torture/pr79276.c: New testcase.
28 2017-01-30 Richard Biener <rguenther@suse.de>
30 PR tree-optimization/79256
31 * targhooks.c (default_builtin_vector_alignment_reachable): Honor
32 BIGGEST_FIELD_ALIGNMENT and ADJUST_FIELD_ALIGN to fix up bogus
34 * tree.c (build_aligned_type): Set TYPE_USER_ALIGN.
36 2017-01-30 Dominik Vogt <vogt@linux.vnet.ibm.com>
39 * config/s390/s390.md ("*r<noxa>sbg_<mode>_srl_bitmask")
40 ("*r<noxa>sbg_<mode>_sll_bitmask")
41 ("*extzv_<mode>_srl<clobbercc_or_nocc>")
42 ("*extzv_<mode>_sll<clobbercc_or_nocc>"):
43 Use contiguous_bitmask_nowrap_operand.
45 2017-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
48 * config/rs6000/altivec.h (vec_xl): Revise #define.
51 2017-01-27 Uros Bizjak <ubizjak@gmail.com>
53 * config/i386/i386.c (print_reg): Use REGNO instead of true_regnum.
55 2017-01-27 Bernd Schmidt <bschmidt@redhat.com>
57 PR rtl-optimization/79194
58 * cprop.c (one_cprop_pass): Move deletion of code after unconditional
59 traps before call to bypass_conditional_jumps.
61 2017-01-27 Vladimir Makarov <vmakarov@redhat.com>
63 PR tree-optimization/71374
64 * lra-constraints.c (check_conflict_input_operands): New.
65 (match_reload): Use it.
67 2017-01-27 Vladimir Makarov <vmakarov@redhat.com>
70 * lra-assigns.c (find_hard_regno_for_1): Take endianess for into
71 account to calculate conflict_set.
73 2017-01-27 Bin Cheng <bin.cheng@arm.com>
75 PR rtl-optimization/78559
76 * combine.c (try_combine): Discard REG_EQUAL and REG_EQUIV for
77 other_insn in combine.
79 2017-01-27 Pekka Jääskeläinen <pekka.jaaskelainen@parmance.com>
81 * builtin-types.def: Use unsigned_char_type_node for BT_UINT8. Use
82 uint16_type_node for BT_UINT16.
84 2017-01-27 David Malcolm <dmalcolm@redhat.com>
86 * doc/sourcebuild.texi (Testsuites): Add "GIMPLE Tests" and
88 (GIMPLE Tests): New node.
89 (RTL Tests): New node.
91 2017-01-27 Richard Biener <rguenther@suse.de>
93 PR tree-optimization/79245
94 * tree-loop-distribution.c (distribute_loop): Apply cost
95 modeling also to detected patterns.
97 2017-01-27 Richard Biener <rguenther@suse.de>
99 PR tree-optimization/71433
100 * tree-vrp.c (register_new_assert_for): Revert earlier changes.
101 (compare_assert_loc): New function.
102 (process_assert_insertions): Sort and optimize assert locations
103 to remove duplicates and push down identical assertions on
104 edges to their destination block.
106 2017-01-27 Richard Biener <rguenther@suse.de>
108 PR tree-optimization/79244
109 * tree-vrp.c (remove_range_assertions): Forcefully propagate
110 out SSA names even if abnormal.
112 2017-01-27 Jakub Jelinek <jakub@redhat.com>
114 * realmpfr.h: Poison MPFR_RND{N,Z,U,D}.
115 * gimple-ssa-sprintf.c (format_floating_max): Use GMP_RNDN
116 instead of MPFR_RNDN.
118 2017-01-27 Richard Earnshaw <rearnsha@arm.com>
121 * arm.c (arm_option_override): Don't call build_target_option_node
122 until after doing all option overrides.
123 (arm_valid_target_attribute_tree): Likewise.
125 2017-01-27 Martin Liska <mliska@suse.cz>
127 * doc/invoke.texi (-fprofile-arcs): Document profiling support
128 for {cd}tors and C++ {cd}tors.
130 2017-01-27 Dominik Vogt <vogt@linux.vnet.ibm.com>
132 * config/s390/s390.md ("*setmem_long_and")
133 ("*setmem_long_and_31z"): Use zero_extend instead of and.
135 2017-01-26 Martin Sebor <msebor@redhat.com>
137 * gimple-ssa-sprintf.c (format_floating): Simplify the computation
140 2017-01-26 Martin Sebor <msebor@redhat.com>
142 * gimple-ssa-sprintf.c (format_floating): Test HAVE_XFmode and
143 HAVE_DFmode before using XFmode or DFmode.
144 (parse_directive): Avoid using the z length modifier to avoid
145 the ISO C++98 does not support the ‘z’ gnu_printf length modifier.
148 * gimple-ssa-sprintf.c (adjust_for_width_or_precision): Change
149 to accept adjustment as an array.
150 (get_int_range): New function.
151 (struct directive): Make width and prec arrays.
152 (directive::set_width, directive::set_precision): Call get_int_range.
153 (format_integer, format_floating): Handle width and precision ranges.
154 (format_string, parse_directive): Same.
156 2017-01-26 Jakub Jelinek <jakub@redhat.com>
159 * dwarf2out.c (generate_skeleton_bottom_up): For children with
160 comdat_type_p set, just clone them, but keep the children in the
164 * dwarf2out.c (prune_unused_types): Mark all functions with DIEs
165 which have direct callers with -fvar-tracking-assignments enabled
167 (resolve_addr): Avoid adding skeleton DIEs for DW_AT_call_origin
168 inside of type units.
170 2017-01-26 Martin Sebor <msebor@redhat.com>
173 * gimple-ssa-sprintf.c (struct result_range): Add likely and
175 (struct format_result): Replace number_chars, number_chars_min,
176 and number_chars_max with a single member of struct result_range.
178 (format_result::operator+=): Adjust.
179 (struct fmtresult): Remove bounded. Handle likely and unlikely
181 (fmtresult::adjust_for_width_or_precision): New function.
182 (fmtresult:type_max_digits): New function.
183 (bytes_remaining): Handle likely and unlikely counters.
184 (min_bytes_remaining): Remove.
185 (format_percent): Simplify.
186 (format_integer, format_floating): Set likely and unlikely counters.
187 (get_string_length, format_character, format_string): Same.
188 (format_plain, should_warn_p): New function.
189 (maybe_warn): Call should_warn_p. Update diagnostic messages
190 and handle those for all directives, including plain strings.
191 (format_directive): Handle likely and unlikely counters.
192 Remove unnecessary quoting from diagnostics. Add an informational
195 (pass_sprintf_length::compute_format_length): Simplify.
196 (try_substitute_return_value): Handle likely and unlikely counters.
198 2017-01-26 Carl Love <cel@us.ibm.com>
200 * config/rs6000/rs6000-c (altivec_overloaded_builtins): Remove
201 bogus entries for the P8V_BUILTIN_VEC_VGBBD built-ins
203 2017-01-26 Vladimir Makarov <vmakarov@redhat.com>
206 * lra-assigns.c (setup_live_pseudos_and_spill_after_risky): Take
207 endianess for subregs into account.
208 * lra-constraints.c (lra_constraints): Do risky transformations
209 always on the first iteration.
210 * lra-lives.c (check_pseudos_live_through_calls): Add arg
211 last_call_used_reg_set.
212 (process_bb_lives): Define and use last_call_used_reg_set.
213 * lra.c (lra): Always continue after lra_constraints on the first
216 2017-01-26 Kirill Yukhin <kirill.yukhin@gmail.com>
218 * gcc.target/i386/avx512bw-kshiftlq-2.c: Use unsigned long long
220 * gcc.target/i386/avx512bw-kshiftrq-2.c: Ditto.
222 2017-01-26 Jakub Jelinek <jakub@redhat.com>
224 * config/i386/avx512fintrin.h (_ktest_mask16_u8,
225 _ktestz_mask16_u8, _ktestc_mask16_u8, _kadd_mask16): Move to ...
226 * config/i386/avx512dqintrin.h (_ktest_mask16_u8,
227 _ktestz_mask16_u8, _ktestc_mask16_u8, _kadd_mask16): ... here.
228 * config/i386/i386-builtin.def (__builtin_ia32_ktestchi,
229 __builtin_ia32_ktestzhi, __builtin_ia32_kaddhi): Use
230 OPTION_MASK_ISA_AVX512DQ instead of OPTION_MASK_ISA_AVX512F.
231 * config/i386/sse.md (SWI1248_AVX512BWDQ2): New mode iterator.
232 (kadd<mode>, ktest<mode>): Use it instead of SWI1248_AVX512BWDQ.
234 2017-01-26 Marek Polacek <polacek@redhat.com>
237 * fold-const.c (operand_equal_p) [COND_EXPR]: Use OP_SAME_WITH_NULL
238 for the third operand.
240 2017-01-26 Jakub Jelinek <jakub@redhat.com>
243 * omp-low.c (struct omp_context): Add simt_stmt field.
244 (scan_omp_for): Return omp_context *.
245 (scan_omp_simd): Set simt_stmt on the non-_simt_ SIMD
246 context to the _simt_ SIMD stmt.
247 (lower_omp_for): For combined SIMD with sibling _simt_
248 SIMD, make sure to use the same decls in _looptemp_
249 clauses as in the sibling.
251 2017-01-26 David Sherwood <david.sherwood@arm.com>
254 * gimplify.c (omp_notice_variable): Add GOVD_SEEN flag to variables in
257 2017-01-26 Jakub Jelinek <jakub@redhat.com>
260 * reg-stack.c (emit_swap_insn): Instead of fld a; fld b; fxchg %st(1);
261 emit fld b; fld a; if possible.
263 * brig-builtins.def: Update copyright years.
264 * config/arm/arm_acle_builtins.def: Update copyright years.
266 2017-01-25 Michael Meissner <meissner@linux.vnet.ibm.com>
269 * config/rs6000/vsx.md (vsx_extract_<mode>_store): Use wY
270 constraint instead of o for the stxsd instruction.
272 2017-01-25 Carl Love <cel@us.ibm.com>
274 * config/rs6000/rs6000-c (altivec_overloaded_builtins): Fix order
275 of entries for ALTIVEC_BUILTIN_VEC_PACKS and P8V_BUILTIN_VEC_VGBBD.
277 2017-01-25 Jonathan Wakely <jwakely@redhat.com>
279 * doc/invoke.texi (C++ Dialect Options): Fix typo.
281 2017-01-25 Richard Biener <rguenther@suse.de>
283 PR tree-optimization/69264
284 * target.def (vector_alignment_reachable): Improve documentation.
285 * doc/tm.texi: Regenerate.
286 * targhooks.c (default_builtin_vector_alignment_reachable): Simplify
288 * tree-vect-data-refs.c (vect_supportable_dr_alignment): Revert
289 earlier changes with respect to TYPE_USER_ALIGN.
290 (vector_alignment_reachable_p): Likewise. Improve dumping.
292 2016-01-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
295 * config/arm/arm.md (xordi3): Force constant operand into a register
298 2016-01-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
300 * doc/invoke.texi (-fstore-merging): Correct default optimization
301 levels at which it is enabled.
302 (-O): Move -fstore-merging from list to...
305 2017-01-25 Richard Biener <rguenther@suse.de>
308 * omp-expand.c: Include debug.h.
309 (expand_omp_taskreg): Make sure to generate early debug before
310 outlining anything from a function.
311 (expand_omp_target): Likewise.
312 (grid_expand_target_grid_body): Likewise.
314 2017-01-25 Maxim Ostapenko <m.ostapenko@samsung.com>
317 * asan.c (get_translation_unit_decl): New function.
318 (asan_add_global): Extract modules file name from globals
319 TRANSLATION_UNIT_DECL name.
321 2017-01-24 Eric Botcazou <ebotcazou@adacore.com>
324 * config/arm/arm.c (arm_function_ok_for_sibcall): Add back restriction
325 for long calls with APCS frame and VFP.
327 2017-01-24 David Malcolm <dmalcolm@redhat.com>
329 * cfg.c (original_copy_tables_initialized_p): New function.
330 * cfg.h (original_copy_tables_initialized_p): New decl.
331 * cfgrtl.c (relink_block_chain): Guard the call to
332 free_original_copy_tables with a call to
333 original_copy_tables_initialized_p.
334 * cgraph.h (symtab_node::native_rtl_p): New decl.
335 * cgraphunit.c (symtab_node::native_rtl_p): New function.
336 (symtab_node::needed_p): Don't assert for early assembly output
338 (cgraph_node::finalize_function): Set "force_output" for __RTL
340 (cgraph_node::analyze): Bail out early for __RTL functions.
341 (analyze_functions): Update assertion to support __RTL functions.
342 (cgraph_node::expand): Bail out early for __RTL functions.
343 * final.c (rest_of_clean_state): Don't call delete_tree_ssa for
345 * function.h (struct function): Update comment for field
347 * gimple-expr.c: Include "tree-pass.h".
348 (gimple_has_body_p): Return false for __RTL functions.
349 * Makefile.in (OBJS): Add run-rtl-passes.o.
350 * pass_manager.h (gcc::pass_manager::get_rest_of_compilation): New
352 (gcc::pass_manager::get_clean_slate): New accessor.
353 * passes.c: Include "insn-addr.h".
354 (should_skip_pass_p): Add logging. Update logic for running
355 "expand" to be compatible with both __GIMPLE and __RTL. Guard
356 property-provider override so it is only done for gimple passes.
358 (skip_pass): New function.
359 (execute_one_pass): Call skip_pass when skipping passes.
360 * read-md.c (md_reader::read_char): Support filtering
361 the input to a subset of line numbers.
362 (md_reader::md_reader): Initialize fields
363 m_first_line and m_last_line.
364 (md_reader::read_file_fragment): New function.
365 * read-md.h (md_reader::read_file_fragment): New decl.
366 (md_reader::m_first_line): New field.
367 (md_reader::m_last_line): New field.
368 * read-rtl-function.c (function_reader::create_function): Only
369 create cfun if it doesn't already exist. Set PROP_rtl on cfun's
370 curr_properties. Set DECL_INITIAL to a dummy block.
371 (read_rtl_function_body_from_file_range): New function.
372 * read-rtl-function.h (read_rtl_function_body_from_file_range):
374 * run-rtl-passes.c: New file.
375 * run-rtl-passes.h: New file.
377 2017-01-24 Jeff Law <law@redhat.com>
379 * config/microblaze/microblaze.h (ASM_FORMAT_PRIVATE_NAME): Increase
382 2017-01-24 Bin Cheng <bin.cheng@arm.com>
384 PR tree-optimization/79159
385 * tree-ssa-loop-niter.c (get_cst_init_from_scev): New function.
386 (record_nonwrapping_iv): Improve boundary using above function if no
387 value range information.
389 2017-01-24 Pekka Jääskeläinen <pekka@parmance.com>
390 Martin Jambor <mjambor@suse.cz>
392 * brig-builtins.def: New file.
393 * builtins.def (DEF_HSAIL_BUILTIN): New macro.
394 (DEF_HSAIL_ATOMIC_BUILTIN): Likewise.
395 (DEF_HSAIL_SAT_BUILTIN): Likewise.
396 (DEF_HSAIL_INTR_BUILTIN): Likewise.
397 (DEF_HSAIL_CVT_ZEROI_SAT_BUILTIN): Likewise.
398 * builtin-types.def (BT_INT8): New.
399 (BT_INT16): Likewise.
400 (BT_UINT8): Likewise.
401 (BT_UINT16): Likewise.
402 (BT_FN_ULONG): Likewise.
403 (BT_FN_UINT_INT): Likewise.
404 (BT_FN_UINT_ULONG): Likewise.
405 (BT_FN_UINT_LONG): Likewise.
406 (BT_FN_UINT_PTR): Likewise.
407 (BT_FN_ULONG_PTR): Likewise.
408 (BT_FN_INT8_FLOAT): Likewise.
409 (BT_FN_INT16_FLOAT): Likewise.
410 (BT_FN_UINT32_FLOAT): Likewise.
411 (BT_FN_UINT16_FLOAT): Likewise.
412 (BT_FN_UINT8_FLOAT): Likewise.
413 (BT_FN_UINT64_FLOAT): Likewise.
414 (BT_FN_UINT16_UINT32): Likewise.
415 (BT_FN_UINT32_UINT16): Likewise.
416 (BT_FN_UINT16_UINT16_UINT16): Likewise.
417 (BT_FN_INT_PTR_INT): Likewise.
418 (BT_FN_UINT_PTR_UINT): Likewise.
419 (BT_FN_LONG_PTR_LONG): Likewise.
420 (BT_FN_ULONG_PTR_ULONG): Likewise.
421 (BT_FN_VOID_UINT64_UINT64): Likewise.
422 (BT_FN_UINT8_UINT8_UINT8): Likewise.
423 (BT_FN_INT8_INT8_INT8): Likewise.
424 (BT_FN_INT16_INT16_INT16): Likewise.
425 (BT_FN_INT_INT_INT): Likewise.
426 (BT_FN_UINT_FLOAT_UINT): Likewise.
427 (BT_FN_FLOAT_UINT_UINT): Likewise.
428 (BT_FN_ULONG_UINT_UINT): Likewise.
429 (BT_FN_ULONG_UINT_PTR): Likewise.
430 (BT_FN_ULONG_ULONG_ULONG): Likewise.
431 (BT_FN_UINT_UINT_UINT): Likewise.
432 (BT_FN_VOID_UINT_PTR): Likewise.
433 (BT_FN_UINT_UINT_PTR: Likewise.
434 (BT_FN_UINT32_UINT64_PTR): Likewise.
435 (BT_FN_INT_INT_UINT_UINT): Likewise.
436 (BT_FN_UINT_UINT_UINT_UINT): Likewise.
437 (BT_FN_UINT_UINT_UINT_PTR): Likewise.
438 (BT_FN_UINT_ULONG_ULONG_UINT): Likewise.
439 (BT_FN_ULONG_ULONG_ULONG_ULONG): Likewise.
440 (BT_FN_LONG_LONG_UINT_UINT): Likewise.
441 (BT_FN_ULONG_ULONG_UINT_UINT): Likewise.
442 (BT_FN_VOID_UINT32_UINT64_PTR): Likewise.
443 (BT_FN_VOID_UINT32_UINT32_PTR): Likewise.
444 (BT_FN_UINT_UINT_UINT_UINT_UINT): Likewise.
445 (BT_FN_UINT_FLOAT_FLOAT_FLOAT_FLOAT): Likewise.
446 (BT_FN_ULONG_ULONG_ULONG_UINT_UINT): Likewise.
447 * doc/frontends.texi: List BRIG FE.
448 * doc/install.texi (Testing): Add BRIG tesring requirements.
449 * doc/invoke.texi (Overall Options): Mention BRIG.
450 * doc/standards.texi (Standards): Doucment BRIG HSA version.
452 2017-01-24 Richard Biener <rguenther@suse.de>
455 * ipa-devirt.c (odr_types_equivalent_p): Fix typo in diagnostic.
457 2017-01-24 Martin Jambor <mjambor@suse.cz>
460 * ipa-prop.c (ipa_free_all_node_params): Call summary destructor.
461 * ipa-prop.c (ipa_node_params_t::insert): Initialize fields known_csts
464 2017-01-24 Aldy Hernandez <aldyh@redhat.com>
467 * gimple-ssa-warn-alloca.c (alloca_call_type): Make sure
468 casts from signed to unsigned really don't have a range.
470 2017-01-24 Markus Trippelsdorf <markus@trippelsdorf.de>
472 * gimple-ssa-sprintf.c (format_floating): Change MPFR_RNDx to
473 GMP_RNDx for compatiblity.
475 2017-01-24 Martin Liska <mliska@suse.cz>
478 * tree-ssa-reassoc.c (rewrite_expr_tree_parallel): Insert assert
479 that would prevent us to call alloca with -1 as argument.
481 2017-01-24 Jakub Jelinek <jakub@redhat.com>
483 * dwarf2out.c (output_compilation_unit_header, output_file_names):
484 Avoid -Wformat-security warning.
486 2017-01-23 Andrew Pinski <apinski@cavium.com>
488 * config/aarch64/aarch64.c (thunderx2t99_addrcost_table): Improve
491 2017-01-23 Martin Sebor <msebor@redhat.com>
494 * gimple-ssa-sprintf.c (warn_level): New global.
495 (format_integer): Use it here and throughout the rest of the file.
496 Use the same switch to compute sign as base.
497 (maybe_warn): New function.
498 (format_directive): Factor out warnings into maybe_warn.
499 Add debugging output. Use warn_level.
500 (add_bytes): Use warn_level.
501 (pass_sprintf_length::compute_format_length): Add debugging output.
502 (try_substitute_return_value): Same.
503 (pass_sprintf_length::handle_gimple_call): Set and use warn_level.
506 * gimple-ssa-sprintf.c (struct format_result): Remove constant member.
507 (struct fmtresult, format_integer, format_floating): Adjust.
508 (fmtresult::fmtresult): Set max correctly in two argument ctor.
509 (get_string_length, format_string,format_directive): Same.
510 (pass_sprintf_length::compute_format_length): Same.
511 (try_substitute_return_value): Simplify slightly.
514 * gimple-ssa-sprintf.c (pass_sprintf_length::gate): Adjust formatting.
515 (fmtresult::operator+=): Outlined.
516 (struct fmtresult): Add ctors.
517 (struct conversion_spec): Rename...
518 (struct directive): ...to this. Add and remove data members.
519 (directive::set_width, directive::set_precision): New functions.
520 (format_percent): Use fmtresult ctor.
521 (get_width_and_precision): Remove.
522 (format_integer): Make naming changes. Avoid computing width and
524 (format_floating): Same. Adjust indentation.
525 (format_character, format_none): New functions.
526 (format_string): Moved character handling to format_character.
527 (format_directive): Remove arguments, change return type.
528 (parse_directive): New function.
529 (pass_sprintf_length::compute_format_length): Move directive
530 parsing to parse_directive.
532 2017-01-23 Jakub Jelinek <jakub@redhat.com>
534 * tree.h (assign_assembler_name_if_neeeded): Rename to ...
535 (assign_assembler_name_if_needed): ... this.
536 * tree.c (assign_assembler_name_if_neeeded): Rename to ...
537 (assign_assembler_name_if_needed): ... this.
538 (free_lang_data_in_cgraph): Adjust callers.
539 * cgraphunit.c (cgraph_node::analyze): Likewise.
540 * omp-expand.c (expand_omp_taskreg, expand_omp_target):
543 2017-01-23 Richard Biener <rguenther@suse.de>
545 PR tree-optimization/79088
546 PR tree-optimization/79188
547 * tree-ssa-threadupdate.c (mark_threaded_blocks): Move code
548 resetting loop bounds after last path deletion. Reset loop
549 bounds of the target loop, make code match the comments.
550 * tree-ssa-threadbackwards.c (pass_early_thread_jumps::execute):
551 Make sure loops need no fixups.
553 2017-01-23 Kelvin Nilsen <kelvin@gcc.gnu.org>
555 * config/rs6000/rs6000-builtin.def (VSIEDPF): Add scalar insert
556 exponent support with double type for first argument.
557 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Changed
558 type returned by __builtin_vec_extract_sig,
559 __builtin_vec_extract_sig_sp, and __builtin_vec_extract_sig_dp
560 functions from "vector int" to "vector unsigned int" or from
561 "vector long long int" to "vector unsigned long long int".
562 Changed type returned by __builtin_vec_extract_exp,
563 __builtin_vec_extract_exp_sp, and __builtin_vec_extract_exp_dp
564 functions from "vector int" to "vector unsigned int" or from
565 "vector long long int" to "vector unsigned long long int".
566 Changed return type of __builtin_vec_test_data_class,
567 __builtin_vec_test_data_class_sp, and
568 __builtin_vec_test_data_class_dp from "vector int" to
569 "vector bool int" or from "vector long long int" to "vector bool
570 long long int" and changed second argument type from "unsigned
571 int" to "int". Added new overloaded function forms "vector float
572 __builtin_vec_insert_exp (vector float, vector unsigned int)" and
573 "vector float __builtin_vec_insert_exp_sp (vector float, vector
574 unsigned int)" and "vector double __builtin_vec_insert_exp (vector
575 double, vector unsigned long long int)" and "vector double
576 __builtin_vec_insert_exp_dp (vector double, vector unsigned long
577 long int)". Changed return type of
578 __builtin_scalar_test_data_class and
579 __builtin_scalar_test_data_class_sp and
580 __builtin_scalar_test_data_class_dp from "unsigned int" to "bool
581 int" and changed second argument from "unsigned int" to "int".
582 Changed type returned by __builtin_scalar_test_neg,
583 __builtin_scalar_test_neg_sp, and __builtin_scalar_test_neg_dp
584 from "int" to "bool int". Added new overloaded function form
585 "double __builtin_scalar_insert_exp (double, unsigned long long int)".
586 * config/rs6000/vsx.md (xsiexpdpf): New insn for scalar insert
587 exponent double-precision with floating point first argument.
588 * doc/extend.texi (PowerPC AltiVec Built-in Functions): Adjust
589 documentation of scalar_test_data_class, scalar_test_neg,
590 scalar_extract_sig, scalar_extract_exp, scalar_insert_exp,
591 vector_extract_exp, vec_extract_sig, vec_insert_exp, and
592 vec_test_data_class built-in functions to reflect refinements in
593 their type signatures.
595 2017-01-23 Andreas Tobler <andreast@gcc.gnu.org>
597 * config/aarch64/aarch64.c (aarch64_elf_asm_constructor): Increase
599 (aarch64_elf_asm_destructor): Likewise.
601 2017-01-23 Bernd Schmidt <bschmidt@redhat.com>
603 PR rtl-optimization/78634
604 * config/i386/i386.c (ix86_max_noce_ifcvt_seq_cost): New function.
605 (TARGET_MAX_NOCE_IFCVT_SEQ_COST): Define.
606 * ifcvt.c (noce_try_cmove): Add missing cost check.
608 PR rtl-optimization/71724
609 * combine.c (if_then_else_cond): Look for situations where it is
610 beneficial to undo the work of one of the recursive calls.
612 2017-01-23 Bin Cheng <bin.cheng@arm.com>
614 PR tree-optimization/70754
615 * tree-predcom.c (stmt_combining_refs): New parameter INSERT_BEFORE.
616 (reassociate_to_the_same_stmt): New parameter INSERT_BEFORE. Insert
617 combined stmt before it if not NULL.
618 (combine_chains): Process refs reversely and compute dominance point
621 2017-01-23 Martin Liska <mliska@suse.cz>
623 PR tree-optimization/79196
624 * tree-ssa-strlen.c (fold_strstr_to_memcmp): Rename to ...
625 (fold_strstr_to_strncmp): ... this. Fold the pattern to strncmp
627 (strlen_optimize_stmt): Call the renamed function.
629 2017-01-23 Michael Matz <matz@suse.de>
631 PR tree-optimization/78384
632 * tree-ssa-loop-split.c (patch_loop_exit): Use correct edge.
634 2017-01-23 Richard Biener <rguenther@suse.de>
636 PR tree-optimization/79186
637 * tree-vrp.c (register_new_assert_for): Make sure we've seen
638 both incoming edges before moving an assert.
640 2017-01-23 Martin Jambor <mjambor@suse.cz>
642 * ipa-prop.c (load_from_param_1): Removed.
643 (load_from_unmodified_param): Bits from load_from_param_1 put back
645 (load_from_param): Removed.
646 (compute_complex_assign_jump_func): Removed stmt2 and just replaced it
647 with stmt. Reverted back to use of load_from_unmodified_param.
649 2017-01-23 Martin Jambor <mjambor@suse.cz>
652 * ipa-prop.h (ipa_param_descriptor): Anotate with with GTY(()).
653 (ipa_node_params): Annotate with GTY((for_user)). Make descriptors
654 field a pointer to garbage collected vector, mark lattices and
655 ipcp_orig_node with GTY((skip)).
656 (ipa_get_param_count): Adjust to descriptors being a pointer.
657 (ipa_get_param): Likewise.
658 (ipa_get_type): Likewise.
659 (ipa_get_param_move_cost): Likewise.
660 (ipa_set_param_used): Likewise.
661 (ipa_get_controlled_uses): Likewise.
662 (ipa_set_controlled_uses): Likewise.
663 (ipa_is_param_used): Likewise.
664 (ipa_node_params_t): Move into garbage collector. New methods insert
666 (ipa_node_params_sum): Annotate wth GTY(()).
667 (ipa_check_create_node_params): Adjust to ipa_node_params_sum being
669 (ipa_load_from_parm_agg): Adjust declaration.
670 * ipa-icf.c (param_used_p): Adjust to descriptors being a pointer.
671 * ipa-profile.c (ipa_profile): Likewise.
672 * ipa-prop.c (ipa_get_param_decl_index_1): Likewise.
673 (ipa_populate_param_decls): Make descriptors parameter garbage
675 (ipa_dump_param): Adjust to descriptors being a pointer.
676 (ipa_alloc_node_params): Likewise.
677 (ipa_initialize_node_params): Likewise.
678 (load_from_param_1): Make descriptors parameter garbage collected.
679 (load_from_unmodified_param): Likewise.
680 (load_from_param): Likewise.
681 (ipa_load_from_parm_agg): Likewise.
682 (ipa_node_params::~ipa_node_params): Removed.
683 (ipa_free_all_node_params): Remove call to delete operator.
684 (ipa_node_params_t::insert): New.
685 (ipa_node_params_t::remove): Likewise.
686 (ipa_node_params_t::duplicate): Adjust to descriptors being a pointer,
687 copy known_csts and known_contexts vectors.
688 (ipa_read_node_info): Adjust to descriptors being a pointer.
689 (ipcp_modif_dom_walker): Make m_descriptors field garbage
691 (ipcp_transform_function): Make descriptors variable garbage
694 2017-01-23 Andrew Senkevich <andrew.senkevich@intel.com>
696 * config/i386/avx512bwintrin.h: Add k-mask test, kortest intrinsics.
697 * config/i386/avx512dqintrin.h: Ditto.
698 * config/i386/avx512fintrin.h: Ditto.
699 * gcc/config/i386/i386.c: Handle new builtins.
700 * config/i386/i386-builtin.def: Add new builtins.
701 * config/i386/sse.md (ktest<mode>, kortest<mode>): New.
702 (UNSPEC_KORTEST, UNSPEC_KTEST): New.
704 2017-01-23 Jakub Jelinek <jakub@redhat.com>
705 Martin Liska <mliska@suse.cz>
707 * asan.h: Define ASAN_USE_AFTER_SCOPE_ATTRIBUTE.
708 * asan.c (asan_expand_poison_ifn): Support stores and use
709 appropriate ASAN report function.
710 * internal-fn.c (expand_ASAN_POISON_USE): New function.
711 * internal-fn.def (ASAN_POISON_USE): Declare.
712 * tree-into-ssa.c (maybe_add_asan_poison_write): New function.
713 (maybe_register_def): Create ASAN_POISON_USE when sanitizing.
714 * tree-ssa-dce.c (eliminate_unnecessary_stmts): Remove
715 ASAN_POISON calls w/o LHS.
716 * tree-ssa.c (execute_update_addresses_taken): Create clobber
717 for ASAN_MARK (UNPOISON, &x, ...) in order to prevent usage of a LHS
718 from ASAN_MARK (POISON, &x, ...) coming to a PHI node.
719 * gimplify.c (asan_poison_variables): Add attribute
720 use_after_scope_memory to variables that really needs to live
722 * tree-ssa.c (is_asan_mark_p): Do not rewrite into SSA when
723 having the attribute.
725 2017-01-23 Martin Liska <mliska@suse.cz>
727 * asan.c (create_asan_shadow_var): New function.
728 (asan_expand_poison_ifn): Likewise.
729 * asan.h (asan_expand_poison_ifn): New declaration.
730 * internal-fn.c (expand_ASAN_POISON): Likewise.
731 * internal-fn.def (ASAN_POISON): New builtin.
732 * sanopt.c (pass_sanopt::execute): Expand
733 asan_expand_poison_ifn.
734 * tree-inline.c (copy_decl_for_dup_finish): Make function
736 * tree-inline.h (copy_decl_for_dup_finish): Likewise.
737 * tree-ssa.c (is_asan_mark_p): New function.
738 (execute_update_addresses_taken): Rewrite local variables
739 (identified just by use-after-scope as addressable) into SSA.
741 2017-01-22 Gerald Pfeifer <gerald@pfeifer.com>
743 * doc/install.texi (Specific): opensource.apple.com uses https
744 now. Remove trailing slash.
746 2017-01-22 Gerald Pfeifer <gerald@pfeifer.com>
748 * README.Portability: Remove note on an Irix compatibility issue.
750 2017-01-22 Dimitry Andric <dim@FreeBSD.org>
752 * gcov.c (INCLUDE_ALGORITHM): Define.
753 (INCLUDE_VECTOR): Define.
754 No longer include <vector> and <algorithm> directly.
756 2017-01-21 Gerald Pfeifer <gerald@pfeifer.com>
758 * doc/extend.texi (Thread-Local): Change www.akkadia.org reference
760 * doc/invoke.texi (Code Gen Options): Ditto.
762 2017-01-21 Jan Hubicka <hubicka@ucw.cz>
765 * cfg.c (update_bb_profile_for_threading): Fix updating of probablity.
767 2017-01-21 Bernd Schmidt <bschmidt@redhat.com>
769 rtl-optimization/79125
770 * cprop.c (local_cprop_pass): Handle cases where we make an
773 2017-01-20 Segher Boessenkool <segher@kernel.crashing.org>
777 * config/rs6000/rs6000.c (rs6000_gimplify_va_arg): Adjust address to
778 read from, for big endian.
780 2017-01-20 Jiong Wang <jiong.wang@arm.com>
782 * config/aarch64/aarch64-builtins.c (aarch64_init_builtins): Register
783 register pauth builtins for LP64 only.
785 2017-01-20 Marek Polacek <polacek@redhat.com>
788 * gimplify.c (should_warn_for_implicit_fallthrough): Handle consecutive
791 2017-01-20 Alexander Monakov <amonakov@ispras.ru>
793 * omp-expand.c (expand_omp_simd): Clear PROP_gimple_lomp_dev regardless
795 * omp-offload.c (pass_omp_device_lower::gate): Use PROP_gimple_lomp_dev.
796 * passes.c (dump_properties): Handle PROP_gimple_lomp_dev.
797 * tree-inline.c (expand_call_inline): Propagate PROP_gimple_lomp_dev.
799 2017-01-20 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
802 * config/arm/arm.c (neon_valid_immediate): Reject vector constants
803 in big-endian mode when they are not a single duplicated value.
805 2017-01-20 Richard Biener <rguenther@suse.de>
807 * BASE-VER: Bump to 7.0.1.
809 2017-01-20 Alexander Monakov <amonakov@ispras.ru>
811 * omp-low.c (omplow_simd_context): New struct. Use it...
812 (lower_rec_simd_input_clauses): ...here and...
813 (lower_rec_input_clauses): ...here to hold common data. Adjust all
814 references to idx, lane, max_vf, is_simt.
816 2017-01-20 Graham Markall <graham.markall@embecosm.com>
818 * config/arc/arc.h (LINK_SPEC): Use arclinux_nps emulation when
821 2017-01-20 Martin Jambor <mjambor@suse.cz>
823 * hsa.h: Renaed to hsa-common.h. Adjusted a comment.
824 * hsa.c: Renaed to hsa-common.c. Change include of gt-hsa.h to
826 * Makefile.in (OBJS): Rename hsa.o to hsa-common.o.
827 (GTFILES): Rename hsa.c to hsa-common.c.
828 * hsa-brig.c: Change include of hsa.h to hsa-common.h.
829 * hsa-dump.c: Likewise.
830 * hsa-gen.c: Likewise.
831 * hsa-regalloc.c: Likewise.
832 * ipa-hsa.c: Likewise.
833 * omp-expand.c: Likewise.
834 * omp-low.c: Likewise.
835 * toplev.c: Likewise.
837 2017-01-20 Marek Polacek <polacek@redhat.com>
840 * doc/invoke.texi: Document -Wduplicated-branches.
841 * fold-const.c (operand_equal_p): Handle MODIFY_EXPR, INIT_EXPR,
842 COMPOUND_EXPR, PREDECREMENT_EXPR, PREINCREMENT_EXPR,
843 POSTDECREMENT_EXPR, POSTINCREMENT_EXPR, CLEANUP_POINT_EXPR, EXPR_STMT,
844 STATEMENT_LIST, and RETURN_EXPR. For non-pure non-const functions
845 return 0 only when not OEP_LEXICOGRAPHIC.
846 (fold_build_cleanup_point_expr): Use the expression
847 location when building CLEANUP_POINT_EXPR.
848 * tree-core.h (enum operand_equal_flag): Add OEP_LEXICOGRAPHIC.
849 * tree.c (add_expr): Handle error_mark_node.
851 2017-01-20 Martin Liska <mliska@suse.cz>
854 * tree-profile.c (init_ic_make_global_vars): Do not call
856 (gimple_init_gcov_profiler): Likewise.
858 2017-01-20 Martin Liska <mliska@suse.cz>
861 * cgraph.h (maybe_create_reference): Remove argument and
863 * cgraphclones.c (cgraph_node::create_virtual_clone): Remove one
865 * ipa-cp.c (create_specialized_node): Likewise.
866 * symtab.c (symtab_node::maybe_create_reference): Handle
867 VAR_DECLs and ADDR_EXPRs and select ipa_ref_use type.
869 2017-01-20 Martin Liska <mliska@suse.cz>
871 * read-rtl-function.c (function_reader::create_function): Use
872 build_decl instread of build_decl_stat.
874 2017-01-20 Andrew Senkevich <andrew.senkevich@intel.com>
876 * config/i386/avx512bwintrin.h: Add k-mask registers shift intrinsics.
877 * config/i386/avx512dqintrin.h: Ditto.
878 * config/i386/avx512fintrin.h: Ditto.
879 * config/i386/i386-builtin-types.def: Add new types.
880 * gcc/config/i386/i386.c: Handle new types.
881 * config/i386/i386-builtin.def (__builtin_ia32_kshiftliqi)
882 (__builtin_ia32_kshiftlihi, __builtin_ia32_kshiftlisi)
883 (__builtin_ia32_kshiftlidi, __builtin_ia32_kshiftriqi)
884 (__builtin_ia32_kshiftrihi, __builtin_ia32_kshiftrisi)
885 (__builtin_ia32_kshiftridi): New.
886 * config/i386/sse.md (k<code><mode>): Rename *k<code><mode>.
888 2017-01-19 Segher Boessenkool <segher@kernel.crashing.org>
892 * config/rs6000/rs6000.c (TARGET_STACK_PROTECT_GUARD): Unconditionally
893 define to rs6000_init_stack_protect_guard.
894 (rs6000_init_stack_protect_guard): New function.
896 2017-01-19 Matthew Fortune <matthew.fortune@imgtec.com>
897 Yunqiang Su <yunqiang.su@imgtec.com>
899 * config.gcc (supported_defaults): Add madd4.
900 (with_madd4): Add validation.
901 (all_defaults): Add madd4.
902 * config/mips/mips.opt (mmadd4): New option.
903 * gcc/config/mips/mips.h (OPTION_DEFAULT_SPECS): Add a default for
905 (TARGET_CPU_CPP_BUILTINS): Add builtin_define for
907 (ISA_HAS_UNFUSED_MADD4): Gate with mips_madd4.
908 (ISA_HAS_FUSED_MADD4): Likewise.
909 * gcc/doc/invoke.texi (-mmadd4): Document the new option.
910 * gcc/doc/install.texi (--with-madd4): Document the new option.
912 2017-01-19 Jiong Wang <jiong.wang@arm.com>
914 * config/aarch64/aarch64-builtins.c (enum aarch64_builtins): New
915 entries for AARCH64_PAUTH_BUILTIN_XPACLRI,
916 AARCH64_PAUTH_BUILTIN_PACIA1716, AARCH64_PAUTH_BUILTIN_AUTIA1716.
917 (aarch64_init_pauth_hint_builtins): New.
918 (aarch64_init_builtins): Call aarch64_init_pauth_hint_builtins.
919 (aarch64_expand_builtin): Expand new builtins.
921 2017-01-19 Jiong Wang <jiong.wang@arm.com>
923 * reg-notes.def (CFA_TOGGLE_RA_MANGLE): New reg-note.
924 * combine-stack-adj.c (no_unhandled_cfa): Handle
925 REG_CFA_TOGGLE_RA_MANGLE.
926 * dwarf2cfi.c (dwarf2out_frame_debug): Handle REG_CFA_TOGGLE_RA_MANGLE.
927 * config/aarch64/aarch64.c (aarch64_expand_prologue): Generates DWARF
928 info for return address signing.
929 (aarch64_expand_epilogue): Likewise.
931 2017-01-19 Jiong Wang <jiong.wang@arm.com>
933 * config/aarch64/aarch64-opts.h (aarch64_function_type): New enum.
934 * config/aarch64/aarch64-protos.h
935 (aarch64_return_address_signing_enabled): New declaration.
936 * config/aarch64/aarch64.c (aarch64_return_address_signing_enabled):
938 (aarch64_expand_prologue): Sign return address before it's pushed onto
940 (aarch64_expand_epilogue): Authenticate return address fetched from
942 (aarch64_override_options): Sanity check for ILP32 and ISA level.
943 (aarch64_attributes): New function attributes for "sign-return-address".
944 * config/aarch64/aarch64.md (UNSPEC_AUTI1716, UNSPEC_AUTISP,
945 UNSPEC_PACI1716, UNSPEC_PACISP, UNSPEC_XPACLRI): New unspecs.
946 ("*do_return"): Generate combined instructions according to key index.
947 ("<pauth_mnem_prefix>sp", "<pauth_mnem_prefix1716", "xpaclri"): New.
948 * config/aarch64/iterators.md (PAUTH_LR_SP, PAUTH_17_16): New integer
950 (pauth_mnem_prefix, pauth_hint_num_a): New integer attributes.
951 * config/aarch64/aarch64.opt (msign-return-address=): New.
952 * doc/extend.texi (AArch64 Function Attributes): Documents
953 "sign-return-address=".
954 * doc/invoke.texi (AArch64 Options): Documents "-msign-return-address=".
956 2017-01-19 Matthew Fortune <matthew.fortune@imgtec.com>
958 * doc/invoke.texi: Add missing -mlxc1-sxc1 options to
959 overall option summary.
961 2017-01-19 Jiong Wang <jiong.wang@arm.com>
963 * config/aarch64/aarch64-arches.def: New entry for "armv8.3-a".
964 * config/aarch64/aarch64.h (AARCH64_FL_V8_3, AARCH64_FL_FOR_ARCH8_3,
965 AARCH64_ISA_V8_3, TARGET_ARMV8_3): New.
966 * doc/invoke.texi (AArch64 Options): Document "armv8.3-a".
968 2017-01-19 Michael Meissner <meissner@linux.vnet.ibm.com>
970 * config/rs6000/rs6000-cpus.def (ISA_3_0_MASKS_SERVER): Enable
971 -mpower9-minmax by default for -mcpu=power9.
972 (ISA_3_MASKS_IEEE): Require -mvsx-small-integer to enable IEEE
973 128-bit floating point.
975 2017-01-20 Alan Modra <amodra@gmail.com>
977 * config/rs6000/rs6000.md (cmpstrnsi, cmpstrsi): Fail if
980 2017-01-20 Alan Modra <amodra@gmail.com>
983 * config/rs6000/rs6000.c (expand_strn_compare): Get the asm name
984 for strcmp and strncmp from corresponding builtin decl.
986 2017-01-19 Uros Bizjak <ubizjak@gmail.com>
988 * config.gcc (x86_64-*-rtems*): Use i386/rtemself.h
989 instead of i386/rtems-64.h.
990 * config/i386/rtems-64.h: Remove.
992 2017-01-19 Uros Bizjak <ubizjak@gmail.com>
996 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
998 * config/i386/rtemself.h (LONG_DOUBLE_TYPE_SIZE): New define.
1000 2017-01-19 Tamar Christina <tamar.christina@arm.com>
1002 * config/aarch64/aarch64.c (aarch64_simd_gen_const_vector_dup):
1003 Change int to HOST_WIDE_INT.
1004 * config/aarch64/aarch64-protos.h
1005 (aarch64_simd_gen_const_vector_dup): Likewise.
1006 * config/aarch64/aarch64-simd.md: Add copysign<mode>3.
1008 2017-01-19 David Malcolm <dmalcolm@redhat.com>
1010 * langhooks-def.h (lhd_type_for_size): New decl.
1011 (LANG_HOOKS_TYPE_FOR_SIZE): Define as lhd_type_for_size.
1012 * langhooks.c (lhd_type_for_size): New function, taken from
1015 2017-01-19 Pat Haugen <pthaugen@us.ibm.com>
1017 * config/rs6000/power9.md (power9-alu): Remove 'cmp' type and add
1018 define_bypass for CR latency.
1019 (power9-cracked-alu): Update bypass latency and remove power9-branch.
1020 (power9-alu2): Add define_bypass for CR latency.
1022 (power9-mul): Update insn latency.
1023 (power9-mul-compare): Update insn latency, bypass latency and remove
1026 2016-01-19 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1028 * config/aarch64/aarch64-protos.h (aarch64_nopcrelative_literal_loads):
1030 * config/aarch64/aarch64.md
1031 (aarch64_reload_movcp<GPF_TF:mode><P:mode>): Delete reference to
1032 aarch64_nopcrelative_literal_loads.
1033 (aarch64_reload_movcp<VALL:mode><P:mode>): Likewise.
1035 2017-01-19 Chenghua Xu <paul.hua.gm@gmail.com>
1037 * config/mips/mips.h (ISA_HAS_FUSED_MADD4): Enable for
1039 (ISA_HAS_UNFUSED_MADD4): Exclude TARGET_LOONGSON_3A.
1041 2017-01-19 Doug Gilmore <doug.gilmore@imgtec.com>
1044 * config.gcc (supported_defaults): Add lxc1-sxc1.
1045 (with_lxc1_sxc1): Add validation.
1046 (all_defaults): Add lxc1-sxc1.
1047 * config/mips/mips.opt (mlxc1-sxc1): New option.
1048 * gcc/config/mips/mips.h (OPTION_DEFAULT_SPECS): Add a default for
1050 (TARGET_CPU_CPP_BUILTINS): Add builtin_define for
1051 __mips_no_lxc1_sxc1.
1052 (ISA_HAS_LXC1_SXC1): Gate with mips_lxc1_sxc1.
1053 * gcc/doc/invoke.texi (-mlxc1-sxc1): Document the new option.
1054 * doc/install.texi (--with-lxc1-sxc1): Document the new option.
1056 2017-01-19 Richard Biener <rguenther@suse.de>
1058 PR tree-optimization/72488
1059 * tree-ssa-sccvn.c (run_scc_vn): When we abort the VN make
1060 sure to restore SSA info.
1061 * tree-ssa.c (verify_ssa): Verify SSA info is not shared.
1063 2017-01-19 Richard Earnshaw <rearnsha@arm.com>
1065 PR rtl-optimization/79121
1066 * expr.c (expand_expr_real_2, case LSHIFT_EXPR): Look at the signedness
1067 of the inner type when shifting an extended value.
1069 2017-01-17 Jan Hubicka <hubicka@ucw.cz>
1072 * symtab.c (symtab_node::equal_address_to): Fix comparing of
1073 interposable aliases.
1075 2017-01-18 Peter Bergner <bergner@vnet.ibm.com>
1078 * config/rs6000/spe.md (mov_si<mode>_e500_subreg0): Fix constraints.
1079 Use the evmergelohi instruction.
1080 (mov_si<mode>_e500_subreg4_2_le): Likewise.
1081 (mov_sitf_e500_subreg8_2_be): Likewise.
1082 (mov_sitf_e500_subreg12_2_le): Likewise.
1083 (mov_si<mode>_e500_subreg0_2_le): Fix constraints.
1084 (mov_si<mode>_e500_subreg4_2_be): Likewise.
1085 (mov_sitf_e500_subreg8_2_le): Likewise.
1086 (mov_sitf_e500_subreg12_2_be): Likewise.
1088 2017-01-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1090 * config/rs6000/altivec.md (altivec_vbpermq): Change "type"
1091 attribute from vecsimple to vecperm.
1092 (altivec_vbpermq2): Likewise.
1094 2017-01-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1097 * config/rs6000/altivec.h: Fix typo of vec_cntlz to vec_cnttz.
1099 2017-01-18 Aaron Sawdey <acsawdey@linux.vnet.ibm.com>
1100 * config/rs6000/rs6000-protos.h (expand_strn_compare): Add arg.
1101 * config/rs6000/rs6000.c (expand_strn_compare): Add ability to expand
1102 strcmp. Fix bug where comparison didn't stop with zero byte. Fix
1103 case where N arg is SIZE_MAX.
1104 * config/rs6000/rs6000.md (cmpstrnsi): Args to expand_strn_compare.
1105 (cmpstrsi): Add pattern.
1107 2017-01-18 Michael Meissner <meissner@linux.vnet.ibm.com>
1109 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
1110 __builtin_vec_revb builtins.
1111 * config/rs6000/rs6000-builtins.def (P9V_BUILTIN_XXBRQ_V16QI): Add
1112 built-in functions to support generation of the ISA 3.0 XXBR<x>
1113 vector byte reverse instructions.
1114 (P9V_BUILTIN_XXBRQ_V1TI): Likewise.
1115 (P9V_BUILTIN_XXBRD_V2DI): Likewise.
1116 (P9V_BUILTIN_XXBRD_V2DF): Likewise.
1117 (P9V_BUILTIN_XXBGW_V4SI): Likewise.
1118 (P9V_BUILTIN_XXBGW_V4SF): Likewise.
1119 (P9V_BUILTIN_XXBGH_V8HI): Likewise.
1120 (P9V_BUILTIN_VEC_REVB): Likewise.
1121 * config/rs6000/vsx.md (p9_xxbrq_v1ti): New insns/expanders to
1122 generate the ISA 3.0 XXBR<x> vector byte reverse instructions.
1123 (p9_xxbrq_v16qi): Likewise.
1124 (p9_xxbrd_<mode>, VSX_D iterator): Likewise.
1125 (p9_xxbrw_<mode>, VSX_W iterator): Likewise.
1126 (p9_xxbrh_v8hi): Likewise.
1127 * config/rs6000/altivec.h (vec_revb): Define if ISA 3.0.
1128 * doc/extend.texi (RS/6000 Altivec Built-ins): Document the
1129 vec_revb built-in functions.
1131 2017-01-18 Uros Bizjak <ubizjak@gmail.com>
1133 PR rtl-optimization/78952
1134 * config/i386/i386.md (any_extract): New code iterator.
1135 (*insvqi_2): Use any_extract for source operand.
1136 (*insvqi_3): Use any_shiftrt for source operand.
1138 2017-01-18 Wilco Dijkstra <wdijkstr@arm.com>
1140 * config/aarch64/aarch64.c (aarch64_sched_adjust_priority)
1142 (TARGET_SCHED_ADJUST_PRIORITY): Define target hook.
1144 2017-01-18 Matthias Klose <doko@ubuntu.com>
1146 * doc/install.texi: Allow default for --with-target-bdw-gc-include.
1148 2016-01-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1150 * config/rs6000/altivec.h (vec_bperm): Change #define.
1151 * config/rs6000/altivec.md (UNSPEC_VBPERMD): New enum constant.
1152 (altivec_vbpermq2): New define_insn.
1153 (altivec_vbpermd): Likewise.
1154 * config/rs6000/rs6000-builtin.def (VBPERMQ2): New monomorphic
1156 (VBPERMD): Likewise.
1157 (VBPERM): New polymorphic function interface.
1158 * config/rs6000/r6000-c.c (altivec_overloaded_builtins_table):
1159 Add entries for P9V_BUILTIN_VEC_VBPERM.
1160 * doc/extend.texi: Add interfaces for vec_bperm.
1162 2017-01-18 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
1164 * config/s390/s390-c.c (s390_expand_overloaded_builtin): Downcase
1165 first letter of error messages.
1166 (s390_resolve_overloaded_builtin): Likewise.
1167 * config/s390/s390.c (s390_expand_builtin): Likewise.
1168 (s390_invalid_arg_for_unprototyped_fn): Likewise.
1169 (s390_valid_target_attribute_inner_p): Likewise.
1170 * config/s390/s390.md ("tabort"): Likewise.
1172 2017-01-18 Toma Tabacu <toma.tabacu@imgtec.com>
1174 * config/mips/mips.h (ISA_HAS_DIV3): Remove unused macro.
1175 (ISA_AVOID_DIV_HILO): New macro.
1176 (ISA_HAS_DIV): Use new ISA_AVOID_DIV_HILO macro.
1177 (ISA_HAS_DDIV): Likewise.
1179 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
1181 * doc/invoke.texi (fabi-version): Correct number of occurrences.
1183 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
1185 * doc/invoke.texi (fabi-version): Spelling fix.
1187 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
1190 * doc/invoke.texi (fabi-version): Mention mangling fix for
1193 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
1196 * doc/invoke.texi (fabi-version): Document discriminator mangling.
1198 2017-01-17 Segher Boessenkool <segher@kernel.crashing.org>
1201 * config/rs6000/rs6000-opts.h (stack_protector_guard): New enum.
1202 * config/rs6000/rs6000.c (rs6000_option_override_internal): Handle
1204 * config/rs6000/rs6000.md (stack_protect_set): Handle the new more
1206 (stack_protect_test): Ditto.
1207 * config/rs6000/rs6000.opt (mstack-protector-guard=,
1208 mstack-protector-guard-reg=, mstack-protector-guard-offset=): New
1210 * doc/invoke.texi (Option Summary) [RS/6000 and PowerPC Options]:
1211 Add -mstack-protector-guard=, -mstack-protector-guard-reg=, and
1212 -mstack-protector-guard-offset=.
1213 (RS/6000 and PowerPC Options): Ditto.
1215 2017-01-17 Uros Bizjak <ubizjak@gmail.com>
1217 * config/i386/i386.h (MASK_CLASS_P): New define.
1218 * config/i386/i386.c (inline_secondary_memory_needed): Ensure that
1219 there are no registers from different register sets also when
1220 mask registers are used. Update function comment.
1221 * config/i386/i386.md (*movsi_internal): Split (*k/*krm) alternative
1222 to (*k/*r) and (*k/*km) alternatives.
1224 2017-01-17 Wilco Dijkstra <wdijkstr@arm.com>
1226 * config/aarch64/aarch64.md (eh_return): Remove pattern and splitter.
1227 * config/aarch64/aarch64.h (AARCH64_EH_STACKADJ_REGNUM): Remove.
1228 (EH_RETURN_HANDLER_RTX): New define.
1229 * config/aarch64/aarch64.c (aarch64_frame_pointer_required):
1230 Force frame pointer in EH return functions.
1231 (aarch64_expand_epilogue): Add barrier for eh_return.
1232 (aarch64_final_eh_return_addr): Remove.
1233 (aarch64_eh_return_handler_rtx): New function.
1234 * config/aarch64/aarch64-protos.h (aarch64_final_eh_return_addr):
1236 (aarch64_eh_return_handler_rtx): New prototype.
1238 2017-01-17 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1240 * config/rs6000/altivec.h (vec_rlmi): New #define.
1241 (vec_vrlnm): Likewise.
1242 (vec_rlnm): Likewise.
1243 * config/rs6000/altivec.md (UNSPEC_VRLMI): New UNSPEC enum value.
1244 (UNSPEC_VRLNM): Likewise.
1245 (VIlong): New mode iterator.
1246 (altivec_vrl<VI_char>mi): New define_insn.
1247 (altivec_vrl<VI_char>nm): Likewise.
1248 * config/rs6000/rs6000-builtin.def (VRLWNM): New monomorphic
1251 (RLNM): New polymorphic function entry.
1252 (VRLWMI): New monomorphic function entry.
1254 (RLMI): New polymorphic function entry.
1255 * config/rs6000/r6000-c.c (altivec_overloaded_builtin_table): Add
1256 new entries for P9V_BUILTIN_VEC_RLMI and P9V_BUILTIN_VEC_RLNM.
1257 * doc/extend.texi: Add description of vec_rlmi, vec_rlnm, and
1260 2017-01-17 Jakub Jelinek <jakub@redhat.com>
1263 * dwarf2out.c (field_byte_offset): Restore the
1264 PCC_BITFIELD_TYPE_MATTERS behavior for INTEGER_CST DECL_FIELD_OFFSET
1265 and DECL_FIELD_BIT_OFFSET. Use fold_build2 instead of build2 + fold.
1266 (analyze_variants_discr, gen_variant_part): Use fold_build2 instead
1269 2017-01-17 Eric Botcazou <ebotcazou@adacore.com>
1272 * config/aarch64/aarch64.c (TARGET_CUSTOM_FUNCTION_DESCRIPTORS): Define
1274 2017-01-17 Jakub Jelinek <jakub@redhat.com>
1277 * dwarf2out.c (add_data_member_location_attribute): For constant
1278 offset bitfield emit for -gdwarf-5 DW_AT_data_bit_offset attribute
1279 instead of DW_AT_data_member_location, DW_AT_bit_offset and
1280 DW_AT_byte_size attributes.
1282 2017-01-17 Eric Botcazou <ebotcazou@adacore.com>
1284 * config/rs6000/rs6000.c (rs6000_emit_move): Also use a TOC reference
1285 after forcing to constant memory when the code model is medium.
1287 2017-01-17 Julia Koval <julia.koval@intel.com>
1290 * config/i386/avx512fintrin.h
1291 (_mm512_i32gather_ps): Change __addr type to void const*.
1292 (_mm512_mask_i32gather_ps): Ditto.
1293 (_mm512_i32gather_pd): Ditto.
1294 (_mm512_mask_i32gather_pd): Ditto.
1295 (_mm512_i64gather_ps): Ditto.
1296 (_mm512_mask_i64gather_ps): Ditto.
1297 (_mm512_i64gather_pd): Ditto.
1298 (_mm512_mask_i64gather_pd): Ditto.
1299 (_mm512_i32gather_epi32): Ditto.
1300 (_mm512_mask_i32gather_epi32): Ditto.
1301 (_mm512_i32gather_epi64): Ditto.
1302 (_mm512_mask_i32gather_epi64): Ditto.
1303 (_mm512_i64gather_epi32): Ditto.
1304 (_mm512_mask_i64gather_epi32): Ditto.
1305 (_mm512_i64gather_epi64): Ditto.
1306 (_mm512_mask_i64gather_epi64): Ditto.
1307 (_mm512_i32scatter_ps): Change __addr type to void*.
1308 (_mm512_mask_i32scatter_ps): Ditto.
1309 (_mm512_i32scatter_pd): Ditto.
1310 (_mm512_mask_i32scatter_pd): Ditto.
1311 (_mm512_i64scatter_ps): Ditto.
1312 (_mm512_mask_i64scatter_ps): Ditto.
1313 (_mm512_i64scatter_pd): Ditto.
1314 (_mm512_mask_i64scatter_pd): Ditto.
1315 (_mm512_i32scatter_epi32): Ditto.
1316 (_mm512_mask_i32scatter_epi32): Ditto.
1317 (_mm512_i32scatter_epi64): Ditto.
1318 (_mm512_mask_i32scatter_epi64): Ditto.
1319 (_mm512_i64scatter_epi32): Ditto.
1320 (_mm512_mask_i64scatter_epi32): Ditto.
1321 (_mm512_i64scatter_epi64): Ditto.
1322 (_mm512_mask_i64scatter_epi64): Ditto.
1323 * config/i386/avx512pfintrin.h
1324 (_mm512_mask_prefetch_i32gather_pd): Change __addr type to void const*.
1325 (_mm512_mask_prefetch_i32gather_ps): Ditto.
1326 (_mm512_mask_prefetch_i64gather_pd): Ditto.
1327 (_mm512_mask_prefetch_i64gather_ps): Ditto.
1328 (_mm512_prefetch_i32scatter_pd): Change __addr type to void*.
1329 (_mm512_prefetch_i32scatter_ps): Ditto.
1330 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
1331 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
1332 (_mm512_prefetch_i64scatter_pd): Ditto.
1333 (_mm512_prefetch_i64scatter_ps): Ditto.
1334 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
1335 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
1336 * config/i386/avx512vlintrin.h
1337 (_mm256_mmask_i32gather_ps): Change __addr type to void const*.
1338 (_mm_mmask_i32gather_ps): Ditto.
1339 (_mm256_mmask_i32gather_pd): Ditto.
1340 (_mm_mmask_i32gather_pd): Ditto.
1341 (_mm256_mmask_i64gather_ps): Ditto.
1342 (_mm_mmask_i64gather_ps): Ditto.
1343 (_mm256_mmask_i64gather_pd): Ditto.
1344 (_mm_mmask_i64gather_pd): Ditto.
1345 (_mm256_mmask_i32gather_epi32): Ditto.
1346 (_mm_mmask_i32gather_epi32): Ditto.
1347 (_mm256_mmask_i32gather_epi64): Ditto.
1348 (_mm_mmask_i32gather_epi64): Ditto.
1349 (_mm256_mmask_i64gather_epi32): Ditto.
1350 (_mm_mmask_i64gather_epi32): Ditto.
1351 (_mm256_mmask_i64gather_epi64): Ditto.
1352 (_mm_mmask_i64gather_epi64): Ditto.
1353 (_mm256_i32scatter_ps): Change __addr type to void*.
1354 (_mm256_mask_i32scatter_ps): Ditto.
1355 (_mm_i32scatter_ps): Ditto.
1356 (_mm_mask_i32scatter_ps): Ditto.
1357 (_mm256_i32scatter_pd): Ditto.
1358 (_mm256_mask_i32scatter_pd): Ditto.
1359 (_mm_i32scatter_pd): Ditto.
1360 (_mm_mask_i32scatter_pd): Ditto.
1361 (_mm256_i64scatter_ps): Ditto.
1362 (_mm256_mask_i64scatter_ps): Ditto.
1363 (_mm_i64scatter_ps): Ditto.
1364 (_mm_mask_i64scatter_ps): Ditto.
1365 (_mm256_i64scatter_pd): Ditto.
1366 (_mm256_mask_i64scatter_pd): Ditto.
1367 (_mm_i64scatter_pd): Ditto.
1368 (_mm_mask_i64scatter_pd): Ditto.
1369 (_mm256_i32scatter_epi32): Ditto.
1370 (_mm256_mask_i32scatter_epi32): Ditto.
1371 (_mm_i32scatter_epi32): Ditto.
1372 (_mm_mask_i32scatter_epi32): Ditto.
1373 (_mm256_i32scatter_epi64): Ditto.
1374 (_mm256_mask_i32scatter_epi64): Ditto.
1375 (_mm_i32scatter_epi64): Ditto.
1376 (_mm_mask_i32scatter_epi64): Ditto.
1377 (_mm256_i64scatter_epi32): Ditto.
1378 (_mm256_mask_i64scatter_epi32): Ditto.
1379 (_mm_i64scatter_epi32): Ditto.
1380 (_mm_mask_i64scatter_epi32): Ditto.
1381 (_mm256_i64scatter_epi64): Ditto.
1382 (_mm256_mask_i64scatter_epi64): Ditto.
1383 (_mm_i64scatter_epi64): Ditto.
1384 (_mm_mask_i64scatter_epi64): Ditto.
1385 * config/i386/i386-builtin-types.def (V16SF_V16SF_PCFLOAT_V16SI_HI_INT)
1386 (V8DF_V8DF_PCDOUBLE_V8SI_QI_INT, V8SF_V8SF_PCFLOAT_V8DI_QI_INT)
1387 (V8DF_V8DF_PCDOUBLE_V8DI_QI_INT, V16SI_V16SI_PCINT_V16SI_HI_INT)
1388 (V8DI_V8DI_PCINT64_V8SI_QI_INT, V8SI_V8SI_PCINT_V8DI_QI_INT)
1389 (V8DI_V8DI_PCINT64_V8DI_QI_INT, V2DF_V2DF_PCDOUBLE_V4SI_QI_INT)
1390 (V4DF_V4DF_PCDOUBLE_V4SI_QI_INT, V2DF_V2DF_PCDOUBLE_V2DI_QI_INT)
1391 (V4DF_V4DF_PCDOUBLE_V4DI_QI_INT, V4SF_V4SF_PCFLOAT_V4SI_QI_INT)
1392 (V8SF_V8SF_PCFLOAT_V8SI_QI_INT, V4SF_V4SF_PCFLOAT_V2DI_QI_INT)
1393 (V4SF_V4SF_PCFLOAT_V4DI_QI_INT, V2DI_V2DI_PCINT64_V4SI_QI_INT)
1394 (V4DI_V4DI_PCINT64_V4SI_QI_INT, V2DI_V2DI_PCINT64_V2DI_QI_INT)
1395 (V4DI_V4DI_PCINT64_V4DI_QI_INT, V4SI_V4SI_PCINT_V4SI_QI_INT)
1396 (V8SI_V8SI_PCINT_V8SI_QI_INT, V4SI_V4SI_PCINT_V2DI_QI_INT)
1397 (V4SI_V4SI_PCINT_V4DI_QI_INT, VOID_PFLOAT_HI_V16SI_V16SF_INT)
1398 (VOID_PFLOAT_QI_V8SI_V8SF_INT, VOID_PFLOAT_QI_V4SI_V4SF_INT)
1399 (VOID_PDOUBLE_QI_V8SI_V8DF_INT, VOID_PDOUBLE_QI_V4SI_V4DF_INT)
1400 (VOID_PDOUBLE_QI_V4SI_V2DF_INT, VOID_PFLOAT_QI_V8DI_V8SF_INT)
1401 (VOID_PFLOAT_QI_V4DI_V4SF_INT, VOID_PFLOAT_QI_V2DI_V4SF_INT)
1402 (VOID_PDOUBLE_QI_V8DI_V8DF_INT, VOID_PDOUBLE_QI_V4DI_V4DF_INT)
1403 (VOID_PDOUBLE_QI_V2DI_V2DF_INT, VOID_PINT_HI_V16SI_V16SI_INT)
1404 (VOID_PINT_QI_V8SI_V8SI_INT, VOID_PINT_QI_V4SI_V4SI_INT)
1405 (VOID_PLONGLONG_QI_V8SI_V8DI_INT, VOID_PLONGLONG_QI_V4SI_V4DI_INT)
1406 (VOID_PLONGLONG_QI_V4SI_V2DI_INT, VOID_PINT_QI_V8DI_V8SI_INT)
1407 (VOID_PINT_QI_V4DI_V4SI_INT, VOID_PINT_QI_V2DI_V4SI_INT)
1408 (VOID_PLONGLONG_QI_V8DI_V8DI_INT, VOID_QI_V8SI_PCINT64_INT_INT)
1409 (VOID_PLONGLONG_QI_V4DI_V4DI_INT, VOID_PLONGLONG_QI_V2DI_V2DI_INT)
1410 (VOID_HI_V16SI_PCINT_INT_INT, VOID_QI_V8DI_PCINT64_INT_INT)
1411 (VOID_QI_V8DI_PCINT_INT_INT): Remove.
1412 (V16SF_V16SF_PCVOID_V16SI_HI_INT, V8DF_V8DF_PCVOID_V8SI_QI_INT)
1413 (V8SF_V8SF_PCVOID_V8DI_QI_INT, V8DF_V8DF_PCVOID_V8DI_QI_INT)
1414 (V16SI_V16SI_PCVOID_V16SI_HI_INT, V8DI_V8DI_PCVOID_V8SI_QI_INT)
1415 (V8SI_V8SI_PCVOID_V8DI_QI_INT, V8DI_V8DI_PCVOID_V8DI_QI_INT)
1416 (VOID_PVOID_HI_V16SI_V16SF_INT, VOID_PVOID_QI_V8SI_V8DF_INT)
1417 (VOID_PVOID_QI_V8DI_V8SF_INT, VOID_PVOID_QI_V8DI_V8DF_INT)
1418 (VOID_PVOID_HI_V16SI_V16SI_INT, VOID_PVOID_QI_V8SI_V8DI_INT)
1419 (VOID_PVOID_QI_V8DI_V8SI_INT, VOID_PVOID_QI_V8DI_V8DI_INT)
1420 (V2DF_V2DF_PCVOID_V4SI_QI_INT, V4DF_V4DF_PCVOID_V4SI_QI_INT)
1421 (V2DF_V2DF_PCVOID_V2DI_QI_INT, V4DF_V4DF_PCVOID_V4DI_QI_INT
1422 (V4SF_V4SF_PCVOID_V4SI_QI_INT, V8SF_V8SF_PCVOID_V8SI_QI_INT)
1423 (V4SF_V4SF_PCVOID_V2DI_QI_INT, V4SF_V4SF_PCVOID_V4DI_QI_INT)
1424 (V2DI_V2DI_PCVOID_V4SI_QI_INT, V4DI_V4DI_PCVOID_V4SI_QI_INT)
1425 (V2DI_V2DI_PCVOID_V2DI_QI_INT, V4DI_V4DI_PCVOID_V4DI_QI_INT)
1426 (V4SI_V4SI_PCVOID_V4SI_QI_INT, V8SI_V8SI_PCVOID_V8SI_QI_INT)
1427 (V4SI_V4SI_PCVOID_V2DI_QI_INT, V4SI_V4SI_PCVOID_V4DI_QI_INT)
1428 (VOID_PVOID_QI_V8SI_V8SF_INT, VOID_PVOID_QI_V4SI_V4SF_INT)
1429 (VOID_PVOID_QI_V4SI_V4DF_INT, VOID_PVOID_QI_V4SI_V2DF_INT)
1430 (VOID_PVOID_QI_V4DI_V4SF_INT, VOID_PVOID_QI_V2DI_V4SF_INT)
1431 (VOID_PVOID_QI_V4DI_V4DF_INT, VOID_PVOID_QI_V2DI_V2DF_INT)
1432 (VOID_PVOID_QI_V8SI_V8SI_INT, VOID_PVOID_QI_V4SI_V4SI_INT)
1433 (VOID_PVOID_QI_V4SI_V4DI_INT, VOID_PVOID_QI_V4SI_V2DI_INT)
1434 (VOID_PVOID_QI_V4DI_V4SI_INT, VOID_PVOID_QI_V2DI_V4SI_INT)
1435 (VOID_PVOID_QI_V4DI_V4DI_INT, VOID_PVOID_QI_V2DI_V2DI_INT)
1436 (VOID_QI_V8SI_PCVOID_INT_INT, VOID_HI_V16SI_PCVOID_INT_INT)
1437 (VOID_QI_V8DI_PCVOID_INT_INT): Add.
1438 * config/i386/i386.c (ix86_init_mmx_sse_builtins): Adjust builtin
1439 definitions accordingly.
1441 2017-01-17 Kito Cheng <kito.cheng@gmail.com>
1442 Kuan-Lin Chen <kuanlinchentw@gmail.com>
1445 * internal-fn.c (expand_mul_overflow): Use convert_modes instead of
1448 2017-01-17 Vladimir Makarov <vmakarov@redhat.com>
1451 * ira-conflicts.c (ira_build_conflicts): Update total conflict
1452 hard regs for inner regno.
1454 2017-01-17 Martin Liska <mliska@suse.cz>
1457 * ipa-polymorphic-call.c (contains_type_p): Fix wrong
1458 assumption and add comment.
1460 2017-01-17 Nathan Sidwell <nathan@acm.org>
1462 * ipa-visibility.c (localize_node): New function, broken out of ...
1463 (function_and_variable_visibility): ... here. Call it.
1465 2017-01-17 Jan Hubicka <hubicka@ucw.cz>
1468 * tree-ssa-threadupdate.c (remove_ctrl_stmt_and_useless_edges):
1469 correctly set frequency of oudgoing edge.
1470 (duplicate_thread_path): Fix profile updating.
1472 2017-01-17 Jakub Jelinek <jakub@redhat.com>
1475 * configure.ac: Add GCC_BASE_VER.
1476 * Makefile.in (version): Use @get_gcc_base_ver@ instead of cat to get
1477 version from BASE-VER file.
1478 (CFLAGS-gcc.o): Add -DBASEVER=$(BASEVER_s).
1479 (gcc.o): Depend on $(BASEVER).
1480 * common.opt (dumpfullversion): New option.
1481 * gcc.c (driver_handle_option): Handle OPT_dumpfullversion.
1482 * doc/invoke.texi: Document -dumpfullversion.
1483 * doc/install.texi: Document --with-gcc-major-version-only.
1484 * configure: Regenerated.
1486 2017-01-17 Richard Biener <rguenther@suse.de>
1488 PR tree-optimization/71433
1489 * tree-vrp.c (register_new_assert_for): Merge same asserts
1490 on all incoming edges.
1491 (process_assert_insertions_for): Handle insertions at the
1494 2017-01-17 Gerald Pfeifer <gerald@pfeifer.com>
1496 * config/i386/cygwin.h (LIBGCJ_SONAME): Remove.
1497 * config/i386/mingw32.h (LIBGCJ_SONAME): Remove.
1499 2017-01-17 Kaz Kojima <kkojima@gcc.gnu.org>
1502 * config/sh/sh.md (cmpeqsi_t+1): Call copy_rtx to avoid invalid
1505 2017-01-17 Alan Modra <amodra@gmail.com>
1508 * config/rs6000/rs6000.md (elf_high, elf_low): Disable when pic.
1509 * config/rs6000/rs6000.c (rs6000_emit_allocate_stack): Don't allow
1510 symbolic stack limit when pic.
1512 2017-01-16 Martin Sebor <msebor@redhat.com>
1514 PR tree-optimization/78608
1515 * gimple-ssa-sprintf.c (tree_digits): Avoid negating TYPE_MIN.
1517 2017-01-16 Jeff Law <law@redhat.com>
1520 2016-12-02 Tadek Kijkowski <tkijkowski@gmail.com>
1521 * Makefile.in (PREPROCESSOR_DEFINES): Add a level of indirection
1522 for several include directories that may be relative to sysroot.
1523 * config/i386/x-mingw32 (gplus_includedir): Define.
1524 (gplus_tool_includedir, gplus_backward_include_dir): Likewise.
1525 (native_system_includedir): Likewise.
1526 * config/i386/mingw32.h (STANDARD_STARTFILE_PREFIX_1): Do not
1527 override if TARGET_SYSTEM_ROOT is defined.
1528 (NATIVE_SYSTEM_HEADER_DIR): Likewise.
1530 PR tree-optimization/79090
1531 PR tree-optimization/33562
1532 PR tree-optimization/61912
1533 PR tree-optimization/77485
1534 * tree-ssa-dse.c (compute_trims): Accept STMT argument. Dump STMT
1535 and computed trims into the dump file.
1537 2017-01-17 Uros Bizjak <ubizjak@gmail.com>
1539 * config/i386/i386.h (LIMIT_RELOAD_CLASS): Remove.
1541 2017-01-16 Jakub Jelinek <jakub@redhat.com>
1544 * gimplify.c (gimplify_init_constructor): If want_value and
1545 object == lhs, unshare lhs to avoid invalid tree sharing. Formatting
1549 * loop-doloop.c (doloop_modify): Call unshare_all_rtl_in_chain on
1550 sequence. Formatting fixes.
1551 (doloop_optimize): Formatting fixes.
1554 * gcc.c (debug_level_greater_than_spec_func): New function.
1555 (static_spec_functions): Add debug-level-gt spec function.
1556 (ASM_DEBUG_SPEC, cpp_options): Use %:debug-level-gt(0) instead of
1558 * config/darwin.h (DSYMUTIL_SPEC, ASM_DEBUG_SPEC): Likewise.
1559 * config/darwin9.h (DSYMUTIL_SPEC, ASM_DEBUG_SPEC): Likewise.
1560 * common.opt (g, gcoff, gdwarf, gdwarf-, ggdb, gno-pubnames,
1561 gpubnames, ggnu-pubnames, gno-record-gcc-switches,
1562 grecord-gcc-switches, gno-strict-dwarf, gstrict-dwarf, gstabs,
1563 gstabs+, gtoggle, gvms, gxcoff, gxcoff+): Add Driver flag.
1565 2017-01-16 Uros Bizjak <ubizjak@gmail.com>
1567 * config/i386/i386.h (HARD_REGNO_CALLER_SAVE_MODE): Apply HImode and
1568 QImode fixups to general and mask registers only.
1570 2017-01-16 Carl Love <cel@us.ibm.com>
1572 * config/rs6000/rs6000-c (altivec_overloaded_builtins): Add support
1573 for built-in functions
1574 vector signed char vec_nabs (vector signed char)
1575 vector signed short vec_nabs (vector signed short)
1576 vector signed int vec_nabs (vector signed int)
1577 vector signed long long vec_nabs (vector signed long long)
1578 vector float vec_nabs (vector float)
1579 vector double vec_nabs (vector double)
1580 * config/rs6000/rs6000-builtin.def: Add definitions for NABS functions
1582 * config/rs6000/altivec.md: New define_expand nabs<mode>2 types
1583 * config/rs6000/altivec.h: New define for vec_nabs built-in function.
1584 * doc/extend.texi: Update the documentation file for the new built-in
1587 2017-01-16 Martin Sebor <msebor@redhat.com>
1589 * gimple-ssa-sprintf.c (format_directive): Correct a typo in a warning
1592 2017-01-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1594 * config/rs6000/rs6000.c (rtx_is_swappable_p): Change
1595 UNSPEC_VSX__XXSPLTD to require special splat handling.
1597 2017-01-16 David Malcolm <dmalcolm@redhat.com>
1600 * system.h: Poison strndup.
1602 2017-01-16 Alan Modra <amodra@gmail.com>
1605 * config/rs6000/rs6000.c (rs6000_legitimate_combined_insn): Don't
1608 2017-01-16 Georg-Johann Lay <avr@gjlay.de>
1610 * config/avr/avr.h (BRANCH_COST) [reload_completed]: Increase by 4.
1612 2017-01-15 Uros Bizjak <ubizjak@gmail.com>
1614 * config/i386/i386.c (ix86_legitimate_combined_insn): Do not
1615 call recog here. Assert that INSN_CODE (insn) is non-negative.
1617 2017-01-15 Segher Boessenkool <segher@kernel.crashing.org>
1620 * cfgrtl.c (rtl_split_edge): Also patch jump insns that jump to the
1622 * haifa-sched.c (dump_insn_stream): Don't crash if there is a label
1623 in the currently scheduled RTL fragment.
1625 2017-01-15 Segher Boessenkool <segher@kernel.crashing.org>
1627 PR rtl-optimization/78751
1628 * ifcvt.c (find_cond_trap): If we generated a non-existing insn,
1631 2017-01-14 Jeff Law <law@redhat.com>
1633 PR tree-optimization/79090
1634 * tree-ssa-dse.c (valid_ao_ref_for_dse): Reject zero length and
1635 variable length stores.
1636 (compute_trims): Delete dead assignment to *trim_tail.
1637 (dse_dom_walker::dse_optimize_stmt): Optimize mem* calls with
1640 2017-01-14 Bernd Schmidt <bschmidt@redhat.com>
1642 PR rtl-optimization/78626
1643 PR rtl-optimization/78727
1644 * cprop.c (one_cprop_pass): Collect unconditional traps in the middle
1645 of a block, and split such blocks after everything else is finished.
1647 2017-01-14 Alan Modra <amodra@gmail.com>
1650 * combine.c (recog_for_combine_1): Set INSN_CODE before calling
1651 target legitimate_combined_insn.
1652 * config/rs6000/rs6000.c (TARGET_LEGITIMATE_COMBINED_INSN): Define.
1653 (rs6000_legitimate_combined_insn): New function.
1654 * config/rs6000/rs6000.md (UNSPEC_DOLOOP): Delete, and remove
1656 (ctr<mode>_internal3): Rename from *ctr<mode>_internal5.
1657 (ctr<mode>_internal4): Rename from *ctr<mode>_internal6.
1658 (ctr<mode>_internal1, ctr<mode>_internal2): Remove '*' from name.
1660 2017-01-14 Gerald Pfeifer <gerald@pfeifer.com>
1662 * doc/frontends.texi (G++ and GCC): Remove references to Java.
1664 2017-01-13 Jeff Law <law@redhat.com>
1666 PR tree-optimization/33562
1667 PR tree-optimization/61912
1668 PR tree-optimization/77485
1669 * tree-ssa-dse.c (delete_dead_call): Accept gsi rather than
1671 (delete_dead_assignment): Likewise.
1672 (dse_dom_walker::dse_optimize_stmt): Pass in the gsi rather than
1673 statement to delete_dead_call and delete_dead_assignment.
1675 2017-01-13 David Malcolm <dmalcolm@redhat.com>
1678 * substring-locations.c (format_warning_va): Strengthen case 1 so
1679 that both endpoints of the substring must be within the format
1680 range for just the substring to be printed.
1682 2017-01-13 Uros Bizjak <ubizjak@gmail.com>
1684 * config/i386/i386.opt (msgx): Use ix86_isa_flags2 variable.
1685 * config/i386/i386.c (ix86_target_string): Add missing options
1686 to isa_opts and reorder options by implied ISAs. Rename isa_opts2 to
1687 isa2_opts, ix86_flag_opts to flag2_opts, ix86_target_other to
1688 flags_other and ix86_target_other to flags2_other. Display unknown
1690 (ix86_valid_target_attribute_inner_p): Add missing options and
1691 reorder options by implied ISAs, as in ix86_target_string.
1693 2017-01-13 Richard Sandiford <richard.sandiford@arm.com>
1695 * hash-table.h (hash_table::too_empty_p): New function.
1696 (hash_table::expand): Use it.
1697 (hash_table::traverse): Likewise.
1698 (hash_table::empty_slot): Use sizeof (value_type) instead of
1699 sizeof (PTR) to convert bytes to elements. Shrink the table
1700 if the current size is excessive for the current number of
1703 2017-01-13 Richard Sandiford <richard.sandiford@arm.com>
1705 * ira-costs.c (record_reg_classes): Break from the inner loop
1706 early once alt_fail is known to be true. Update outer loop
1707 handling accordingly.
1709 2017-01-13 Jeff Law <law@redhat.com>
1711 * tree-ssa-dse.c (decrement_count): New function.
1712 (increment_start_addr, maybe_trim_memstar_call): Likewise.
1713 (dse_dom_walker::optimize_stmt): Call maybe_trim_memstar_call directly
1714 when we know the partially dead statement is a mem* function.
1716 PR tree-optimization/61912
1717 PR tree-optimization/77485
1718 * tree-ssa-dse.c: Include expr.h.
1719 (maybe_trim_constructor_store): New function.
1720 (maybe_trim_partially_dead_store): Call maybe_trim_constructor_store.
1722 PR tree-optimization/33562
1723 PR tree-optimization/61912
1724 PR tree-optimization/77485
1725 * doc/invoke.texi: Document new dse-max-object-size param.
1726 * params.def (PARM_DSE_MAX_OBJECT_SIZE): New PARAM.
1727 * tree-ssa-dse.c: Include params.h.
1728 (dse_store_status): New enum.
1729 (initialize_ao_ref_for_dse): New, partially extracted from
1731 (valid_ao_ref_for_dse, normalize_ref): New.
1732 (setup_live_bytes_from_ref, compute_trims): Likewise.
1733 (clear_bytes_written_by, maybe_trim_complex_store): Likewise.
1734 (maybe_trim_partially_dead_store): Likewise.
1735 (maybe_trim_complex_store): Likewise.
1736 (dse_classify_store): Renamed from dse_possibly_dead_store_p.
1737 Track what bytes live from the original store. Return tri-state
1738 for dead, partially dead or live.
1739 (dse_dom_walker): Add constructor, destructor and new private members.
1740 (delete_dead_call, delete_dead_assignment): New extracted from
1742 (dse_optimize_stmt): Make a member of dse_dom_walker.
1743 Use initialize_ao_ref_for_dse.
1745 PR tree-optimization/33562
1746 PR tree-optimization/61912
1747 PR tree-optimization/77485
1748 * sbitmap.h (bitmap_count_bits): Prototype.
1749 (bitmap_clear_range, bitmap_set_range): Likewise.
1750 * sbitmap.c (bitmap_clear_range): New function.
1751 (bitmap_set_range, sbitmap_popcount, bitmap_count_bits): Likewise.
1753 2017-01-13 Martin Liska <mliska@suse.cz>
1756 * function.c (set_cfun): Add new argument force.
1757 * function.h (set_cfun): Likewise.
1758 * ipa-inline-transform.c (inline_call): Use the function when
1759 strict alising from is dropped for function we inline to.
1761 2017-01-13 Richard Biener <rguenther@suse.de>
1763 * tree-pretty-print.c (dump_generic_node): Fix inverted condition
1764 for dumping GIMPLE INTEGER_CSTs.
1766 2017-01-13 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1768 * config/sol2.h (TARGET_OS_CPP_BUILTINS): Define __STDC_VERSION__
1769 to 201112L since C++17.
1771 2017-01-13 Maxim Ostapenko <m.ostapenko@samsung.com>
1774 * asan.c (asan_needs_odr_indicator_p): Don't emit ODR indicators
1775 if -fsanitize=kernel-address is present.
1777 2017-01-13 Richard Biener <rguenther@suse.de>
1779 * tree-pretty-print.c (dump_generic_node): Dump INTEGER_CSTs
1780 as _Literal ( type ) number in case usual suffixes do not
1781 preserve all information.
1783 2017-01-13 Richard Biener <rguenther@suse.de>
1785 PR tree-optimization/77283
1786 * gimple-ssa-split-paths.c: Include gimple-ssa.h, tree-phinodes.h
1787 and ssa-iterators.h.
1788 (is_feasible_trace): Implement a cost model based on joiner
1791 2017-01-12 Michael Meissner <meissner@linux.vnet.ibm.com>
1794 * config/rs6000/rs6000.md (FP_ISA3): Do not optimize converting
1795 char or short to __float128/_Float128 directly.
1797 2017-01-12 Martin Sebor <msebor@redhat.com>
1799 to -Wformat-overflow.
1800 * gimple-ssa-sprintf.c (pass_sprintf_length::gate): Adjust.
1801 (min_bytes_remaining): Same.
1802 (get_string_length): Same.
1803 (format_string): Same.
1804 (format_directive): Same.
1806 (pass_sprintf_length::handle_gimple_call): Same.
1808 2017-01-12 Jakub Jelinek <jakub@redhat.com>
1810 * gimple-ssa-sprintf.c (try_substitute_return_value): Remove
1811 info.nowrite calls with no lhs that can't throw. Return bool
1812 whether gsi_remove has been called or not.
1813 (pass_sprintf_length::handle_gimple_call): Return bool whether
1814 try_substitute_return_value called gsi_remove. Formatting fix.
1815 (pass_sprintf_length::execute): Don't use gsi_remove if
1816 handle_gimple_call returned true.
1819 * cfgrtl.c (rtl_tidy_fallthru_edge): For any_uncondjump_p that can't
1820 be removed due to side-effects, don't remove following barrier nor
1821 turn the successor edge into fallthru edge.
1823 2017-01-12 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1826 * config/rs6000/rs6000.c (insn_is_swappable_p): Mark
1827 element-reversing loads and stores as not swappable.
1829 2017-01-12 Nathan Sidwell <nathan@acm.org>
1830 Nicolai Stange <nicstange@gmail.com>
1832 * combine.c (try_combine): Don't ignore result of overlap checking
1833 loop. Combine overlap & asm check into single loop.
1835 2017-01-12 Richard Biener <rguenther@suse.de>
1837 * tree-pretty-print.c (dump_generic_node): Provide -gimple
1838 variant for MEM_REF. Sanitize INTEGER_CST for -gimple.
1840 2017-01-12 Richard Biener <rguenther@suse.de>
1842 * tree.c (initialize_tree_contains_struct): Make TS_OPTIMIZATION
1843 and TS_TARGET_OPTION directly derive from TS_BASE.
1844 * tree-core.h (tree_optimization_option): Derive from tree_base.
1845 (tree_target_option): Likewise.
1847 2017-01-11 Uros Bizjak <ubizjak@gmail.com>
1849 * config/i386/i386.c (memory_address_length): Increase len
1850 only when rip_relative_addr_p returns false.
1852 2017-01-11 Julia Koval <julia.koval@intel.com>
1854 * common/config/i386/i386-common.c (OPTION_MASK_ISA_SGX_UNSET): New.
1855 (OPTION_MASK_ISA_SGX_SET): New.
1856 (ix86_handle_option): Handle OPT_msgx.
1857 * config.gcc: Added sgxintrin.h.
1858 * config/i386/driver-i386.c (host_detect_local_cpu): Detect sgx.
1859 * config/i386/i386-c.c (ix86_target_macros_internal): Define __SGX__.
1860 * config/i386/i386.c (ix86_target_string): Add -msgx.
1862 (ix86_option_override_internal): Handle new options.
1863 (ix86_valid_target_attribute_inner_p): Add sgx.
1864 * config/i386/i386.h (TARGET_SGX, TARGET_SGX_P): New.
1865 * config/i386/i386.opt: Add msgx.
1866 * config/i386/sgxintrin.h: New file.
1867 * config/i386/x86intrin.h: Add sgxintrin.h.
1869 2017-01-11 Jakub Jelinek <jakub@redhat.com>
1872 * fold-const.c (maybe_nonzero_address): Return 1 for function
1874 (tree_single_nonzero_warnv_p): Don't handle function local objects
1878 * gcc.c (default_compilers): Don't add -o %g.s for -S -save-temps
1881 2017-01-11 David Malcolm <dmalcolm@redhat.com>
1884 * opts.c: Include "spellcheck.h"
1885 (struct string_fragment): New struct.
1886 (struct edit_distance_traits<const string_fragment &>): New
1888 (get_closest_sanitizer_option): New function.
1889 (parse_sanitizer_options): Offer suggestions for unrecognized arguments.
1891 2017-01-11 Jakub Jelinek <jakub@redhat.com>
1893 * dwarf2out.c (DWARF_COMPILE_UNIT_HEADER_SIZE): For DWARF5 decrease
1895 (DWARF_COMDAT_TYPE_UNIT_HEADER_SIZE): Always
1896 DWARF_COMPILE_UNIT_HEADER_SIZE plus 12.
1897 (DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE): Define.
1898 (calc_base_type_die_sizes): Use DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE
1899 for initial die_offset if dwarf_split_debug_info.
1900 (output_comp_unit): Use DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE for
1901 initial next_die_offset if dwo_id is non-NULL. Don't emit padding
1903 (output_skeleton_debug_sections): Formatting fix. Use
1904 DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE instead of
1905 DWARF_COMPILE_UNIT_HEADER_SIZE. Don't emit padding.
1907 2017-01-11 Wilco Dijkstra <wdijkstr@arm.com>
1909 * config/arm/cortex-a53.md: Add bypasses for
1911 (cortex_a53_r2f): Only use for transfers.
1912 (cortex_a53_f2r): Likewise.
1913 (cortex_a53_r2f_cvt): Add reservation for conversions.
1914 (cortex_a53_f2r_cvt): Likewise.
1916 2017-01-11 Tamar Christina <tamar.christina@arm.com>
1918 * config/arm/arm_neon.h: Add __artificial__ and gnu_inline
1919 to all inlined functions, change static to extern.
1921 2017-01-11 Christophe Lyon <christophe.lyon@linaro.org>
1924 * config/arm/arm.c (legitimize_pic_address): Handle reference to
1926 (arm_assemble_integer): Likewise.
1928 2017-01-11 Richard Earnshaw <rearnsha@arm.com>
1930 * config.gcc: Use new awk script to check CPU, FPU and architecture
1931 parameters for --with-... options.
1932 * config/arm/parsecpu.awk: New file
1933 * config/arm/arm-cpus.in: New file.
1934 * config/arm/arm-opts.h: Include arm-cpu.h instead of processing .def
1936 * config/arm/arm.c: Include arm-cpu-data.h instead of processing .def
1938 * config/arm/t-arm: Update dependency rules.
1939 * common/config/arm/arm-common.c: Include arm-cpu-cdata.h instead
1940 of processing .def files.
1941 * config/arm/genopt.sh: Deleted.
1942 * config/arm/gentune.sh: Deleted.
1943 * config/arm/arm-cores.def: Deleted.
1944 * config/arm/arm-arches.def: Deleted.
1945 * config/arm/arm-fpus.def: Deleted.
1946 * config/arm/arm-tune.md: Regenerated.
1947 * config/arm/arm-tables.opt: Regenerated.
1948 * config/arm/arm-cpu.h: New generated file.
1949 * config/arm/arm-cpu-data.h: New generated file.
1950 * config/arm/arm-cpu-cdata.h: New generated file.
1952 2017-01-11 Maxim Ostapenko <m.ostapenko@samsung.com>
1955 * lto-cgraph.c (lto_output_varpool_node): Pack dynamically_initialized
1957 (input_varpool_node): Unpack dynamically_initialized bit.
1959 2017-01-11 Eric Botcazou <ebotcazou@adacore.com>
1961 PR rtl-optimization/79032
1962 * lra-constraints.c (simplify_operand_subreg): In the MEM case, test
1963 the alignment of the adjusted memory reference against that of MODE,
1964 instead of the alignment of the original memory reference.
1966 2017-01-11 Martin Jambor <mjambor@suse.cz>
1968 * hsa.c (hsa_callable_function_p): Revert addition of DECL_ARTIFICIAL
1970 * ipa-hsa.c (process_hsa_functions): Only duplicate non-artificial
1971 decorated functions.
1973 2017-01-11 Richard Biener <rguenther@suse.de>
1975 * tree-vrp.c (evrp_dom_walker::before_dom_children): Also
1976 set range/nonnull info for PHI results. Do not set it on
1977 stmts marked for removal.
1979 2017-01-10 Eric Botcazou <ebotcazou@adacore.com>
1981 * expr.c (store_field): In the bitfield case, fetch the return value
1982 from the registers before applying a single big-endian adjustment.
1983 Always do a final load for a BLKmode value not larger than a word.
1985 2017-01-10 David Malcolm <dmalcolm@redhat.com>
1988 * input.c (selftest::test_accessing_ordinary_linemaps): Verify
1989 that we correctly handle column numbers greater than
1990 LINE_MAP_MAX_COLUMN_NUMBER.
1992 2017-01-10 Martin Sebor <msebor@redhat.com>
1995 * gimple-ssa-sprintf.c (get_destination_size): Call
1996 {init,fini}object_sizes.
1997 * tree-object-size.c (addr_object_size): Adjust.
1998 (pass_through_call): Adjust.
1999 (pass_object_sizes::execute): Adjust.
2000 * tree-object-size.h (fini_object_sizes): Declare.
2002 2017-01-10 Martin Sebor <msebor@redhat.com>
2004 PR tree-optimization/78775
2005 * builtins.c (get_size_range): Move...
2006 * calls.c: ...to here.
2007 (alloc_max_size): Accept zero argument.
2008 (operand_signed_p): Remove.
2009 (maybe_warn_alloc_args_overflow): Call get_size_range.
2010 * calls.h (get_size_range): Declare.
2012 2017-01-10 Joe Seymour <joe.s@somniumtech.com>
2014 * config/msp430/driver-msp430.c (msp430_mcu_data): Sync with data
2015 from TI's devices.csv file as of September 2016.
2016 * config/msp430/msp430.c (msp430_mcu_data): Likewise.
2018 2017-01-10 Sandra Loosemore <sandra@codesourcery.com>
2020 * doc/extend.texi: Tweak formatting to fix overfull hbox warnings.
2021 * doc/invoke.texi: Likewise.
2022 * doc/md.texi: Likewise.
2023 * doc/objc.texi: Likewise.
2025 2017-01-10 Joshua Conner <joshconner@google.com>
2027 * config/arm/fuchsia-elf.h: New file.
2028 * config/fuchsia.h: New file.
2029 * config.gcc (*-*-fuchsia*): Set native_system_header_dir.
2030 (aarch64*-*-fuchsia*, arm*-*-fuchsia*, x86_64-*-fuchsia*): Add to
2032 * config.host: (aarch64*-*-fuchsia*, arm*-*-fuchsia*): Add to hosts.
2034 2016-01-10 Richard Biener <rguenther@suse.de>
2036 PR tree-optimization/79034
2037 * tree-call-cdce.c (shrink_wrap_one_built_in_call_with_conds):
2038 Propagate out degenerate PHIs in the joiner.
2040 2017-01-10 Martin Liska <mliska@suse.cz>
2042 * ipa-icf.c (sort_sem_items_by_decl_uid): New function.
2043 (sort_congruence_classes_by_decl_uid): Likewise.
2044 (sort_congruence_class_groups_by_decl_uid): Likewise.
2045 (sem_item_optimizer::merge_classes): Sort class, groups in these
2046 classes and members in the groups by DECL_UID of declarations.
2047 This would make merge operations stable.
2049 2017-01-10 Martin Liska <mliska@suse.cz>
2051 * ipa-icf.c (sem_item_optimizer::sem_item_optimizer): Remove
2052 usage of m_classes_vec.
2053 (sem_item_optimizer::~sem_item_optimizer): Likewise.
2054 (sem_item_optimizer::get_group_by_hash): Likewise.
2055 (sem_item_optimizer::subdivide_classes_by_equality): Likewise.
2056 (sem_item_optimizer::subdivide_classes_by_sensitive_refs): Likewise.
2057 (sem_item_optimizer::verify_classes): Likewise.
2058 (sem_item_optimizer::process_cong_reduction): Likewise.
2059 (sem_item_optimizer::dump_cong_classes): Likewise.
2060 (sem_item_optimizer::merge_classes): Likewise.
2061 * ipa-icf.h (congruence_class_hash): Rename from
2062 congruence_class_group_hash. Remove declaration of m_classes_vec.
2064 2017-01-10 Andrew Senkevich <andrew.senkevich@intel.com>
2066 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512VPOPCNTDQ_SET,
2067 OPTION_MASK_ISA_AVX512VPOPCNTDQ_UNSET): New.
2068 * config.gcc: Add avx512vpopcntdqintrin.h.
2069 * config/i386/avx512vpopcntdqintrin.h: New.
2070 * config/i386/cpuid.h (bit_AVX512VPOPCNTDQ): New.
2071 * config/i386/i386-builtin-types.def: Add new types.
2072 * config/i386/i386-builtin.def (__builtin_ia32_vpopcountd_v16si,
2073 __builtin_ia32_vpopcountd_v16si_mask, __builtin_ia32_vpopcountq_v8di,
2074 __builtin_ia32_vpopcountq_v8di_mask): New.
2075 * config/i386/i386-c.c (ix86_target_macros_internal): Define
2076 __AVX512VPOPCNTDQ__.
2077 * config/i386/i386.c (ix86_target_string): Add -mavx512vpopcntdq.
2078 (PTA_AVX512VPOPCNTDQ): Define.
2079 * config/i386/i386.h (TARGET_AVX512VPOPCNTDQ,
2080 TARGET_AVX512VPOPCNTDQ_P): Define.
2081 * config/i386/i386.opt: Add mavx512vpopcntdq.
2082 * config/i386/immintrin.h: Include avx512vpopcntdqintrin.h.
2083 * config/i386/sse.md (define_insn "vpopcount<mode><mask_name>"): New.
2085 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
2088 * predict.def (PRED_CALL): Set to 67.
2090 2017-01-09 Eric Botcazou <ebotcazou@adacore.com>
2092 * expr.c (store_field): In the bitfield case, if the value comes from
2093 a function call and is of an aggregate type returned in registers, do
2094 not modify the field mode; extract the value in all cases if the mode
2095 is BLKmode and the size is not larger than a word.
2097 2017-01-09 Dominique d'Humieres <dominiq@lps.ens.fr>
2100 * config/i386/cpuid.h: Fix undefined behavior.
2102 2017-01-04 Jeff Law <law@redhat.com>
2104 PR tree-optimization/79007
2105 PR tree-optimization/67955
2106 * tree-ssa-alias.c (same_addr_size_stores_p): Only need to be
2107 conservative for pt.null when flag_non_call_exceptions is on.
2109 2017-01-09 Jakub Jelinek <jakub@redhat.com>
2111 PR translation/79019
2112 PR translation/79020
2113 * params.def (PARAM_INLINE_MIN_SPEEDUP,
2114 PARAM_IPA_CP_SINGLE_CALL_PENALTY,
2115 PARAM_USE_AFTER_SCOPE_DIRECT_EMISSION_THRESHOLD): Fix typos
2117 * config/avr/avr.opt (maccumulate-args): Likewise.
2118 * config/msp430/msp430.opt (mwarn-mcu): Likewise.
2119 * common.opt (freport-bug): Likewise.
2120 * cif-code.def (CIF_FINAL_ERROR): Likewise.
2121 * doc/invoke.texi (ipa-cp-single-call-penalty): Likewise.
2122 * config/s390/s390.c (s390_invalid_binary_op): Fix spelling in
2123 translatable string.
2124 * config/i386/i386.c (function_value_32): Likewise.
2125 * config/nios2/nios2.c (nios2_valid_target_attribute_rec): Likewise.
2126 * config/msp430/msp430.c (msp430_option_override, msp430_attr):
2128 * config/msp430/driver-msp430.c (msp430_select_hwmult_lib): Likewise.
2129 * common/config/msp430/msp430-common.c (msp430_handle_option):
2131 * symtab.c (symtab_node::verify_base): Likewise.
2132 * opts.c (set_debug_level): Likewise.
2133 * tree.c (verify_type_variant): Likewise. Fix typo in comment.
2134 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin): Add
2135 missing whitespace to translatable strings.
2136 * config/avr/avr.md (bswapsi2): Fix typo in comment.
2137 * config/sh/superh.h: Likewise.
2138 * config/i386/xopintrin.h: Likewise.
2139 * config/i386/znver1.md: Likewise.
2140 * config/rs6000/rs6000.c (struct rs6000_opt_mask): Likewise.
2141 * ipa-inline-analysis.c (compute_inline_parameters): Likewise.
2142 * double-int.h (struct double_int): Likewise.
2143 * double-int.c (div_and_round_double): Likewise.
2144 * wide-int.cc: Likewise.
2145 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
2146 * tree-ssa-sccvn.c (vn_reference_lookup_3): Likewise.
2147 * cfgcleanup.c (crossjumps_occured): Renamed to ...
2148 (crossjumps_occurred): ... this.
2149 (try_crossjump_bb, try_head_merge_bb, try_optimize_cfg, cleanup_cfg):
2152 PR tree-optimization/78899
2153 * tree-if-conv.c (version_loop_for_if_conversion): Instead of
2154 returning bool return struct loop *, NULL for failure and the new
2156 (versionable_outer_loop_p): Don't version outer loop if it has
2157 dont_vectorized bit set.
2158 (tree_if_conversion): When versioning outer loop, ensure
2159 tree_if_conversion is performed also on the inner loop of the
2160 non-vectorizable outer loop copy.
2161 * tree-vectorizer.c (set_uid_loop_bbs): Formatting fix. Fold
2162 LOOP_VECTORIZED in inner loop of the scalar outer loop and
2163 prevent vectorization of it.
2164 (vectorize_loops): For outer + inner LOOP_VECTORIZED, ensure
2165 the outer loop vectorization of the non-scalar version is attempted
2166 before vectorization of the inner loop in scalar version. If
2167 outer LOOP_VECTORIZED guarded loop is not vectorized, prevent
2168 vectorization of its inner loop.
2169 * tree-vect-loop-manip.c (rename_variables_in_bb): If outer_loop
2170 has 2 inner loops, rename also on edges from bb whose single pred
2171 is outer_loop->header. Fix typo in function comment.
2173 2017-01-09 Martin Sebor <msebor@redhat.com>
2176 * asan.c (asan_emit_stack_protection): Increase local buffer size
2177 to avoid snprintf truncation warning.
2179 2017-01-09 Andrew Pinski <apinski@cavium.com>
2181 * config/aarch64/aarch64-cores.def: Add thunderx2t99. Change vulcan
2182 to reference thunderx2t99 for the tuning structure
2183 * config/aarch64/aarch64-cost-tables.h (vulcan_extra_costs):
2185 (thunderx2t99_extra_costs): This.
2186 * config/aarch64/aarch64-tune.md: Regenerate.
2187 * config/aarch64/aarch64.c (vulcan_addrcost_table): Rename to ...
2188 (vulcan_addrcost_table): This.
2189 (vulcan_regmove_cost): Rename to ...
2190 (thunderx2t99_regmove_cost): This.
2191 (vulcan_vector_cost): Rename to ...
2192 (thunderx2t99_vector_cost): this.
2193 (vulcan_branch_cost): Rename to ...
2194 (thunderx2t99_branch_cost): This.
2195 (vulcan_tunings): Rename to ...
2196 (thunderx2t99_tunings): This and s/vulcan/thunderx2t99 .
2197 * doc/invoke.texi (AARCH64/mtune): Add thunderx2t99.
2199 2017-01-09 Martin Jambor <mjambor@suse.cz>
2203 * ipa-prop.h (ipa_jump_func): Swap positions of vr_known and m_vr.
2204 * ipa-cp.c (ipa_vr_operation_and_type_effects): New function.
2205 (propagate_vr_accross_jump_function): Use the above function for all
2206 value range computations for pass-through jump functions and type
2207 converasion from explicit value range values.
2208 (ipcp_propagate_stage): Do not attempt to deduce types of formal
2209 parameters from TYPE_ARG_TYPES.
2210 * ipa-prop.c (ipa_write_jump_function): Remove trailing whitespace.
2211 (ipa_write_node_info): Stream type of the actual argument.
2212 (ipa_read_node_info): Likewise. Also remove trailing whitespace.
2214 2017-01-09 Martin Liska <mliska@suse.cz>
2217 * gcc.c (driver_handle_option): Handle OPT_E and set have_E.
2218 (lookup_compiler): Do not show error message with have_E.
2220 2017-01-09 Jakub Jelinek <jakub@redhat.com>
2222 PR tree-optimization/78938
2223 * tree-vect-stmts.c (vectorizable_condition): For non-masked COND_EXPR
2224 where comp_vectype is VECTOR_BOOLEAN_TYPE_P, use
2225 BIT_{NOT,XOR,AND,IOR}_EXPR on the comparison operands instead of
2226 {EQ,NE,GE,GT,LE,LT}_EXPR directly inside of VEC_COND_EXPR. Formatting
2229 2017-01-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2231 * tree-ssa-address.c (gen_addr_rtx): Don't handle index if it
2234 2017-01-09 Richard Biener <rguenther@suse.de>
2236 PR tree-optimization/78997
2237 * tree-vect-slp.c (vect_mask_constant_operand_p): Handle SSA
2238 name condition properly.
2240 2017-01-09 Richard Biener <rguenther@suse.de>
2243 * dwarf2out.c (is_cxx): New overload with context.
2244 (is_naming_typedef_decl): Use it.
2246 2017-01-08 Sandra Loosemore <sandra@codesourcery.com>
2248 * invoke.texi (Option Summary): Correct spacing in option lists
2249 and add line breaks to fix over-long lines.
2251 2017-01-08 Sandra Loosemore <sandra@codesourcery.com>
2255 * extend.texi (Common Variable Attributes): Add xref to GCC
2256 Internals manual to explain mode attribute keywords.
2258 2017-01-08 Sandra Loosemore <sandra@codesourcery.com>
2261 * doc/invoke.texi (Option Summary): Move -pthread to Linker Options
2262 and Preprocessor Options.
2263 (Options for Linking): Document -pthread here....
2264 (RS/6000 and PowerPC Options): ...not here.
2265 (Solaris 2 Options): ...or here.
2266 * doc/cppopts.texi: Document -pthread.
2268 2017-01-08 Martin Sebor <msebor@redhat.com>
2271 * doc/invoke.texi (Warning Options): Document -Wformat-truncation.
2272 * gimple-ssa-sprintf.c (call_info::reval_used, call_info::warnopt):
2273 New member functions.
2274 (format_directive): Used them.
2276 (pass_sprintf_length::handle_gimple_call): Same.
2277 * graphite-sese-to-poly.c (tree_int_to_gmp): Increase buffer size
2278 to avoid truncation for any argument.
2279 (extract_affine_mul): Same.
2280 * tree.c (get_file_function_name): Same.
2282 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
2285 * predict.def (PRED_INDIR_CALL): Set to 86.
2287 2017-01-07 Sandra Loosemore <sandra@codesourcery.com>
2289 PR preprocessor/54124
2290 * doc/cppopts.texi: Reformat -d subtable to list the full name
2291 of the options. Add cross-reference to the docs for the general
2292 compiler -d options.
2293 * doc/invoke.texi (Developer Options): Add cross-reference to the
2294 preprocessor-specific -d option documentation.
2296 2017-01-07 Sandra Loosemore <sandra@codesourcery.com>
2298 PR preprocessor/13498
2299 * doc/cpp.texi (Search Path): Rewrite to remove obsolete and
2300 redudant material, and reflect new command-line options.
2301 (System Headers): Likewise.
2303 2017-01-07 Sandra Loosemore <sandra@codesourcery.com>
2305 * doc/cppdiropts.texi: Merge documentation of -I, -iquote,
2306 -isystem, and -idirafter. Copy-edit.
2307 * doc/cppopts.texi: Copy-edit. Remove contradiction about
2308 default for -ftrack-macro-expansion. Delete obsolete and
2309 badly-formatted implementation details about -fdebug-cpp output.
2310 * doc/cppwarnopts.texi: Copy-edit.
2312 2017-01-07 David Malcolm <dmalcolm@redhat.com>
2315 * input.c (selftest::test_accessing_ordinary_linemaps): Verify
2316 that the transition from a max line width >= 1<<10 to narrower
2317 lines works correctly.
2319 2017-01-07 Alexandre Oliva <aoliva@redhat.com>
2321 * doc/options.texi (PerFunction): New.
2322 * opt-functions.awk (switch_flags): Map both Optimization and
2323 PerFunction to CL_OPTIMIZATION.
2324 * opth-gen.awk: Test for PerFunction flag along with
2326 * optc-save-gen.awk: Likewise. Introduce var_opt_hash and set
2327 it only when the latter is present. Skip those that don't in
2328 the hash function generator.
2329 * common.opt (fvar-tracking): Mark as PerFunction instead of
2331 (fvar-tracking-assignments): Likewise.
2332 (fvar-tracking-assignments-toggle): Likewise.
2333 (fvar-tracking-uninit): Likewise.
2335 2017-01-07 Jakub Jelinek <jakub@redhat.com>
2337 PR translation/79018
2338 * params.def (PARAM_MAX_STORES_TO_MERGE): Add missing space between
2341 2017-01-06 Mikael Pettersson <mikpelinux@gmail.com>
2344 * config/m68k/m68k.opt (LONG_JUMP_TABLE_OFFSETS): New option.
2345 * config/m68k/linux.h (ASM_RETURN_CASE_JUMP): Handle
2346 TARGET_LONG_JUMP_TABLE_OFFSETS.
2347 * config/m68k/m68kelf.h (ASM_RETURN_CASE_JUMP): Likewise.
2348 * config/m68k/netbsd-elf.h (ASM_RETURN_CASE_JUMP): Likewise.
2349 * config/m68k/m68k.h (CASE_VECTOR_MODE): Likewise.
2350 (ASM_OUTPUT_ADDR_DIFF_ELF): Likewise.
2351 * config/m68k/m68k.md (tablejump expander): Likewise.
2352 (*tablejump_pcrel_hi): Renamed from unnamed insn, reject
2353 TARGET_LONG_JUMP_TABLE_OFFSETS.
2354 (*tablejump_pcrel_si): New insn, handle TARGET_LONG_JUMP_TABLE_OFFSETS.
2355 * doc/invoke.texi (M68K options): Add -mlong-jump-table-offsets.
2357 2017-01-06 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
2358 David Holsgrove <david.holsgrove@xilinx.com>
2360 * common/config/microblaze/microblaze-common.c
2361 (TARGET_EXCEPT_UNWIND_INFO): Remove.
2362 * config/microblaze/microblaze-protos.h (microblaze_eh_return):
2364 * config/microblaze/microblaze.c (microblaze_must_save_register)
2365 (microblaze_expand_epilogue, microblaze_return_addr): Handle
2367 (microblaze_eh_return): New function.
2368 * config/microblaze/microblaze.h (RETURN_ADDR_OFFSET)
2369 (EH_RETURN_DATA_REGNO, MB_EH_STACKADJ_REGNUM)
2370 (EH_RETURN_STACKADJ_RTX, ASM_PREFERRED_EH_DATA_FORMAT): New macros.
2371 * config/microblaze/microblaze.md (eh_return): New pattern.
2373 2017-01-06 Jakub Jelinek <jakub@redhat.com>
2375 * system.h (GCC_DIAGNOSTIC_PUSH_IGNORED, GCC_DIAGNOSTIC_POP,
2376 GCC_DIAGNOSTIC_STRINGIFY): Define.
2378 * read-rtl.c (rtx_reader::read_rtx_code): Avoid -Wsign-compare warning.
2380 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2382 * config/arm/arm.md (<mcrr>): New.
2384 * config/arm/arm.c (arm_arch5te): New.
2385 (arm_option_override): Set arm_arch5te.
2386 (arm_coproc_builtin_available): Add support for mcrr, mcrr2, mrrc
2388 * config/arm/arm-builtins.c (MCRR_QUALIFIERS): Define to...
2389 (arm_mcrr_qualifiers): ... this. New.
2390 (MRRC_QUALIFIERS): Define to...
2391 (arm_mrrc_qualifiers): ... this. New.
2392 * config/arm/arm_acle.h (__arm_mcrr, __arm_mcrr2, __arm_mrrc,
2394 * config/arm/arm_acle_builtins.def (mcrr, mcrr2, mrrc, mrrc2): New.
2395 * config/arm/iterators.md (MCRRI, mcrr, MCRR): New.
2396 (MRRCI, mrrc, MRRC): New.
2397 * config/arm/unspecs.md (VUNSPEC_MCRR, VUNSPEC_MCRR2, VUNSPEC_MRRC,
2398 VUNSPEC_MRRC2): New.
2400 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2402 * config/arm/arm.md (<mcr>): New.
2404 * config/arm/arm.c (arm_coproc_builtin_available): Add
2405 support for mcr, mrc, mcr2 and mrc2.
2406 * config/arm/arm-builtins.c (MCR_QUALIFIERS): Define to...
2407 (arm_mcr_qualifiers): ... this. New.
2408 (MRC_QUALIFIERS): Define to ...
2409 (arm_mrc_qualifiers): ... this. New.
2410 (MCR_QUALIFIERS): Define to ...
2411 (arm_mcr_qualifiers): ... this. New.
2412 * config/arm/arm_acle.h (__arm_mcr, __arm_mrc, __arm_mcr2,
2414 * config/arm/arm_acle_builtins.def (mcr, mcr2, mrc, mrc2): New.
2415 * config/arm/iterators.md (MCRI, mcr, MCR, MRCI, mrc, MRC): New.
2416 * config/arm/unspecs.md (VUNSPEC_MCR, VUNSPEC_MCR2, VUNSPEC_MRC,
2419 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2421 * config/arm/arm.md (*ldc): New.
2425 * config/arm/arm.c (arm_coproc_builtin_available): Add
2426 support for ldc,ldcl,stc,stcl,ldc2,ldc2l,stc2 and stc2l.
2427 (arm_coproc_ldc_stc_legitimate_address): New.
2428 * config/arm/arm-builtins.c (arm_type_qualifiers): Add
2429 'qualifier_const_pointer'.
2430 (LDC_QUALIFIERS): Define to...
2431 (arm_ldc_qualifiers): ... this. New.
2432 (STC_QUALIFIERS): Define to...
2433 (arm_stc_qualifiers): ... this. New.
2434 * config/arm/arm-protos.h
2435 (arm_coproc_ldc_stc_legitimate_address): New.
2436 * config/arm/arm_acle.h (__arm_ldc, __arm_ldcl, __arm_stc,
2437 __arm_stcl, __arm_ldc2, __arm_ldc2l, __arm_stc2, __arm_stc2l): New.
2438 * config/arm/arm_acle_builtins.def (ldc, ldc2, ldcl, ldc2l, stc,
2439 stc2, stcl, stc2l): New.
2440 * config/arm/constraints.md (Uz): New.
2441 * config/arm/iterators.md (LDCI, STCI, ldc, stc, LDC STC): New.
2442 * config/arm/unspecs.md (VUNSPEC_LDC, VUNSPEC_LDC2, VUNSPEC_LDCL,
2443 VUNSPEC_LDC2L, VUNSPEC_STC, VUNSPEC_STC2, VUNSPEC_STCL,
2444 VUNSPEC_STC2L): New.
2446 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2448 * config/arm/arm.md (<cdp>): New.
2449 * config/arm/arm.c (neon_const_bounds): Rename this ...
2450 (arm_const_bounds): ... this.
2451 (arm_coproc_builtin_available): New.
2452 * config/arm/arm-builtins.c (SIMD_MAX_BUILTIN_ARGS): Increase.
2453 (arm_type_qualifiers): Add 'qualifier_unsigned_immediate'.
2454 (CDP_QUALIFIERS): Define to...
2455 (arm_cdp_qualifiers): ... this. New.
2457 (arm_expand_builtin_args): Add case for 6 arguments.
2458 * config/arm/arm-protos.h (neon_const_bounds): Rename this ...
2459 (arm_const_bounds): ... this.
2460 (arm_coproc_builtin_available): New.
2461 * config/arm/arm_acle.h (__arm_cdp): New.
2463 * config/arm/arm_acle_builtins.def (cdp): New.
2465 * config/arm/iterators.md (CDPI,CDP,cdp): New.
2466 * config/arm/neon.md: Rename all 'neon_const_bounds' to
2468 * config/arm/types.md (coproc): New.
2469 * config/arm/unspecs.md (VUNSPEC_CDP, VUNSPEC_CDP2): New.
2470 * gcc/doc/extend.texi (ACLE): Add a mention of Coprocessor intrinsics.
2471 * gcc/doc/sourcebuild.texi (arm_coproc1_ok, arm_coproc2_ok,
2472 arm_coproc3_ok, arm_coproc4_ok): Document new effective targets.
2474 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2476 * config/arm/arm-builtins.c (arm_unsigned_binop_qualifiers): New.
2477 (UBINOP_QUALIFIERS): New.
2479 (acle_builtin_data): New. Change comment.
2480 (arm_builtins): Remove ARM_BUILTIN_CRC32B, ARM_BUILTIN_CRC32H,
2481 ARM_BUILTIN_CRC32W, ARM_BUILTIN_CRC32CB, ARM_BUILTIN_CRC32CH,
2482 ARM_BUILTIN_CRC32CW. Add ARM_BUILTIN_ACLE_BASE and include
2483 arm_acle_builtins.def.
2484 (ARM_BUILTIN_ACLE_PATTERN_START): Define.
2485 (arm_init_acle_builtins): New.
2486 (CRC32_BUILTIN): Remove.
2487 (bdesc_2arg): Remove entries for crc32b, crc32h, crc32w,
2488 crc32cb, crc32ch and crc32cw.
2489 (arm_init_crc32_builtins): Remove.
2490 (arm_init_builtins): Use arm_init_acle_builtins rather
2491 than arm_init_crc32_builtins.
2492 (arm_expand_acle_builtin): New.
2493 (arm_expand_builtin): Use 'arm_expand_acle_builtin'.
2494 * config/arm/arm_acle_builtins.def: New.
2496 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2498 * config/arm/arm-builtins.c (neon_builtin_datum): Rename to ..
2499 (arm_builtin_datum): ... this.
2500 (arm_init_neon_builtin): Rename to ...
2501 (arm_init_builtin): ... this. Add a new parameters PREFIX
2502 and USE_SIG_IN_NAME.
2503 (arm_init_neon_builtins): Replace 'arm_init_neon_builtin' with
2504 'arm_init_builtin'. Replace type 'neon_builtin_datum' with
2505 'arm_builtin_datum'.
2506 (arm_init_vfp_builtins): Likewise.
2507 (builtin_arg): Rename enum's replacing 'NEON_ARG' with
2508 'ARG_BUILTIN' and add a 'ARG_BUILTIN_NEON_MEMORY.
2509 (arm_expand_neon_args): Rename to ...
2510 (arm_expand_builtin_args): ... this. Rename builtin_arg
2511 enum values and differentiate between ARG_BUILTIN_MEMORY
2512 and ARG_BUILTIN_NEON_MEMORY.
2513 (arm_expand_neon_builtin_1): Rename to ...
2514 (arm_expand_builtin_1): ... this. Rename builtin_arg enum
2515 values, arm_expand_builtin_args and add bool parameter NEON.
2516 (arm_expand_neon_builtin): Use arm_expand_builtin_1.
2517 (arm_expand_vfp_builtin): Likewise.
2518 (NEON_MAX_BUILTIN_ARGS): Remove, it was unused.
2520 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
2523 * predict.def (PRED_POLYMORPHIC_CALL): Set to 59.
2524 * predict.c (tree_estimate_probability_bb): Reverse direction of
2525 polymorphic call predictor.
2527 2017-01-06 David Malcolm <dmalcolm@redhat.com>
2529 * passes.c (execute_one_pass): Split out pass-skipping logic into...
2530 (determine_pass_name_match): ...this new function and...
2531 (should_skip_pass_p): ...this new function.
2533 2017-01-06 Nathan Sidwell <nathan@acm.org>
2535 * ipa-visibility.c (function_and_variable_visibility): Reformat
2536 comments and long lines. Remove extrneous if.
2537 * symtab.c (symtab_node::make_decl_local): Fix code format.
2538 (symtab_node::set_section_for_node): Fix comment typo.
2540 2017-01-06 Martin Liska <mliska@suse.cz>
2543 * lra-constraints.c: Rename invariant to lra_invariant.
2544 * predict.c (set_even_probabilities): Initialize e to NULL.
2546 2017-01-05 Martin Sebor <msebor@redhat.com>
2548 PR tree-optimization/78910
2549 * gimple-ssa-sprintf.c (tree_digits): Add an argument.
2550 (format_integer): Correct off-by-one error in the handling
2551 of precision with negative numbers in signed conversions..
2553 2017-01-05 Eric Botcazou <ebotcazou@adacore.com>
2555 * doc/invoke.texi (C Dialect Options): Adjust -fsso-struct entry.
2557 2017-01-05 Jakub Jelinek <jakub@redhat.com>
2559 PR tree-optimization/71016
2560 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Pass cond_stmt to
2561 factor_out_conditional_conversion. Formatting fix.
2562 (factor_out_conditional_conversion): Add cond_stmt argument.
2563 If arg1 is INTEGER_CST, punt if new_arg0 is not any operand of
2564 cond_stmt and if arg0_def_stmt is not the only stmt in its bb.
2567 2017-01-05 David Malcolm <dmalcolm@redhat.com>
2569 * Makefile.in (OBJS): Add read-md.o, read-rtl.o,
2570 read-rtl-function.o, and selftest-rtl.o.
2571 * config/aarch64/aarch64.c: Include selftest.h and selftest-rtl.h.
2572 (selftest::aarch64_test_loading_full_dump): New function.
2573 (selftest::aarch64_run_selftests): New function.
2574 (TARGET_RUN_TARGET_SELFTESTS): Wire it up to
2575 selftest::aarch64_run_selftests.
2576 * config/i386/i386.c
2577 (selftest::ix86_test_loading_dump_fragment_1): New function.
2578 (selftest::ix86_test_loading_call_insn): New function.
2579 (selftest::ix86_test_loading_full_dump): New function.
2580 (selftest::ix86_test_loading_unspec): New function.
2581 (selftest::ix86_run_selftests): Call the new functions.
2582 * emit-rtl.c (maybe_set_max_label_num): New function.
2583 * emit-rtl.h (maybe_set_max_label_num): New decl.
2584 * function.c (instantiate_decls): Guard call to
2585 instantiate_decls_1 with if (DECL_INITIAL (fndecl)).
2586 * function-tests.c (selftest::verify_three_block_rtl_cfg): Remove
2588 * gensupport.c (gen_reader::gen_reader): Pass "false"
2589 for new "compact" param of rtx_reader.
2590 * print-rtl.c (rtx_writer::print_rtx_operand): Print "(nil)"
2591 rather than an empty string for NULL strings.
2592 * read-md.c: Potentially include config.h rather than bconfig.h.
2593 Wrap include of errors.h with #ifdef GENERATOR_FILE.
2594 (have_error): New global, copied from errors.c.
2595 (md_reader::read_name): Rename to...
2596 (md_reader::read_name_1): ...this, adding "out_loc" param,
2597 and converting "missing name or number" to returning false, rather
2599 (md_reader::read_name): Reimplement in terms of read_name_1.
2600 (md_reader::read_name_or_nil): New function.
2601 (md_reader::read_string): Handle "(nil)" by returning NULL.
2602 (md_reader::md_reader): Add new param "compact".
2603 (md_reader::read_md_files): Wrap with #ifdef GENERATOR_FILE.
2604 (md_reader::read_file): New method.
2605 * read-md.h (md_reader::md_reader): Add new param "compact".
2606 (md_reader::read_file): New method.
2607 (md_reader::is_compact): New accessor.
2608 (md_reader::read_name): Convert return type from void to file_location.
2609 (md_reader::read_name_or_nil): New decl.
2610 (md_reader::read_name_1): New decl.
2611 (md_reader::m_compact): New field.
2612 (noop_reader::noop_reader): Pass "false" for new "compact" param
2614 (rtx_reader::rtx_reader): Add new "compact" param.
2615 (rtx_reader::read_rtx_operand): Make virtual and convert return
2616 type from void to rtx.
2617 (rtx_reader::read_until): New decl.
2618 (rtx_reader::handle_any_trailing_information): New virtual function.
2619 (rtx_reader::postprocess): New virtual function.
2620 (rtx_reader::finalize_string): New virtual function.
2621 (rtx_reader::m_in_call_function_usage): New field.
2622 (rtx_reader::m_reuse_rtx_by_id): New field.
2623 * read-rtl-function.c: New file.
2624 * selftest-rtl.c (selftest::assert_rtx_ptr_eq_at): New function.
2625 * selftest-rtl.h (ASSERT_RTX_PTR_EQ): New macro.
2626 (selftest::verify_three_block_rtl_cfg): New decl.
2627 * read-rtl-function.h: New file.
2628 * read-rtl.c: Potentially include config.h rather than bconfig.h.
2629 For host, include function.h, memmodel.h, and emit-rtl.h.
2630 (one_time_initialization): New function.
2631 (struct compact_insn_name): New struct.
2632 (compact_insn_names): New array.
2633 (find_code): Handle insn codes in compact dumps.
2634 (apply_subst_iterator): Wrap with #ifdef GENERATOR_FILE.
2635 (bind_subst_iter_and_attr): Likewise.
2636 (add_condition_to_string): Likewise.
2637 (add_condition_to_rtx): Likewise.
2638 (apply_attribute_uses): Likewise.
2639 (add_current_iterators): Likewise.
2640 (apply_iterators): Likewise.
2641 (initialize_iterators): Guard usage of apply_subst_iterator with
2642 #ifdef GENERATOR_FILE.
2643 (read_conditions): Wrap with #ifdef GENERATOR_FILE.
2644 (md_reader::read_mapping): Likewise.
2645 (add_define_attr_for_define_subst): Likewise.
2646 (add_define_subst_attr): Likewise.
2647 (read_subst_mapping): Likewise.
2648 (check_code_iterator): Likewise.
2649 (rtx_reader::read_rtx): Likewise. Move one-time initialization
2651 (one_time_initialization): New function.
2652 (rtx_reader::read_until): New method.
2653 (read_flags): New function.
2654 (parse_reg_note_name): New function.
2655 (rtx_reader::read_rtx_code): Initialize "iterator" to NULL.
2656 Handle reuse_rtx ids.
2657 Wrap iterator lookup within #ifdef GENERATOR_FILE.
2658 Add parsing support for RTL dumps, mirroring the special-cases in
2659 print_rtx, by calling read_flags, reading REG_NOTE names, INSN_UID
2660 values, and calling handle_any_trailing_information.
2661 (rtx_reader::read_rtx_operand): Convert return type from void
2662 to rtx, returning return_rtx. Handle case 'e'. Call
2663 finalize_string on XSTR and XTMPL fields.
2664 (rtx_reader::read_nested_rtx): Handle dumps in which trailing
2665 "(nil)" values were omitted. Call the postprocess vfunc on the
2667 (rtx_reader::rtx_reader): Add new "compact" param and pass to base
2668 class ctor. Initialize m_in_call_function_usage. Call
2669 one_time_initialization.
2670 * rtl-tests.c (selftest::test_uncond_jump): Call
2671 set_new_first_and_last_insn.
2672 * rtl.h (read_rtx): Wrap decl with #ifdef GENERATOR_FILE.
2673 * selftest-rtl.c: New file.
2674 * selftest-rtl.h (class selftest::rtl_dump_test): New class.
2675 (selftest::get_insn_by_uid): New decl.
2676 * selftest-run-tests.c (selftest::run_tests): Call
2677 read_rtl_function_c_tests.
2678 * selftest.h (selftest::read_rtl_function_c_tests): New decl.
2679 * tree-dfa.c (ssa_default_def): Return NULL_TREE for rtl function
2682 2017-01-05 Uros Bizjak <ubizjak@gmail.com>
2684 * config/i386/i386.md (*testqi_ext_3): No need to handle memory
2685 operands in a special way. Assert that pos+len <= mode precision.
2687 2017-01-05 Jakub Jelinek <jakub@redhat.com>
2689 * common.opt (fvect-cost-model): Remove RejectNegative flag, use
2690 3 argument Alias with unlimited for the negative form.
2691 (fno-vect-cost-model): Removed.
2693 2017-01-05 Martin Liska <mliska@suse.cz>
2695 * hsa-gen.c (gen_hsa_divmod): New function.
2696 (gen_hsa_insn_for_internal_fn_call): Use the function for IFN_DIVMOD.
2698 2017-01-05 Martin Liska <mliska@suse.cz>
2701 * gcc.c (lookup_compiler): Reject '-' filename for a precompiled
2704 2017-01-05 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
2706 * config/s390/s390.c (s390_expand_setmem): Unroll the loop for
2707 small constant length operands.
2709 2017-01-05 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
2711 * config/s390/s390.c (s390_expand_setmem): Avoid overlapping bytes
2712 between loop iterations.
2714 2017-01-05 Martin Liska <mliska@suse.cz>
2717 * gimplify.c (gimplify_decl_expr): Compare to
2718 asan_poisoned_variables instread of checking flags.
2719 (gimplify_target_expr): Likewise.
2720 (gimplify_expr): Likewise.
2721 (gimplify_function_tree): Conditionally initialize
2722 asan_poisoned_variables.
2724 2017-01-04 Jeff Law <law@redhat.com>
2726 PR tree-optimizatin/78812
2727 * rtl.h (contains_mem_rtx_p): Prototype.
2728 * ifcvt.c (containts_mem_rtx_p): Move from here to...
2729 * rtlanal.c (contains_mem_rtx_p): Here and remove static linkage.
2730 * gcse.c (prune_expressions): Use contains_mem_rtx_p to discover
2731 and prune MEMs that are not at the toplevel of a SET_SRC rtx. Look
2732 through ZERO_EXTEND and SIGN_EXTEND when trying to avoid pruning MEMs.
2734 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
2736 * input.c (assert_char_at_range): Default-initialize actual_range.
2738 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
2740 * df-scan.c (df_ref_create_structure): Make regno unsigned,
2741 to match the caller.
2743 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
2745 * cfgexpand.c (expand_gimple_basic_block): Disregard debug
2746 insns after final jump in test to emit dummy move.
2748 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
2750 * gimple-iterator.h (gsi_one_nondebug_before_end_p): New.
2751 * tree-eh.c (cleanup_empty_eh): Skip more debug stmts.
2753 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
2755 * multiple_target.c (create_dispatcher_calls): Init e_next.
2756 * tree-ssa-loop-split.c (split_loop): Init border.
2757 * tree-vect-loop.c (vect_determine_vectorization_factor): Init
2760 2017-01-04 Michael Meissner <meissner@linux.vnet.ibm.com>
2765 * config/rs6000/predicates.md (sf_subreg_operand): New predicate.
2766 (altivec_register_operand): Do not return true if the operand
2767 contains a SUBREG mixing SImode and SFmode.
2768 (vsx_register_operand): Likewise.
2769 (vsx_reg_sfsubreg_ok): New predicate.
2770 (vfloat_operand): Do not return true if the operand contains a
2771 SUBREG mixing SImode and SFmode.
2772 (vint_operand): Likewise.
2773 (vlogical_operand): Likewise.
2774 (gpc_reg_operand): Likewise.
2775 (int_reg_operand): Likewise.
2776 * config/rs6000/rs6000-protos.h (valid_sf_si_move): Add declaration.
2777 * config/rs6000/rs6000.c (valid_sf_si_move): New function to
2778 determine if a MOVSI or MOVSF operation contains SUBREGs that mix
2780 (rs6000_emit_move_si_sf_subreg): New helper function.
2781 (rs6000_emit_move): Call rs6000_emit_move_si_sf_subreg to possbily
2782 fixup SUBREGs involving SImode and SFmode.
2783 * config/rs6000/vsx.md (SFBOOL_*): New constants that are operand
2784 numbers for the new peephole2 optimization.
2785 (peephole2 for SFmode unions): New peephole2 to optimize cases in
2786 the GLIBC math library that do AND/IOR/XOR operations on single
2787 precision floating point.
2788 * config/rs6000/rs6000.h (TARGET_NO_SF_SUBREG): New internal
2789 target macros to say whether we need to avoid SUBREGs mixing
2791 (TARGET_ALLOW_SF_SUBREG): Likewise.
2792 * config/rs6000/rs6000.md (UNSPEC_SF_FROM_SI): New unspecs.
2793 (UNSPEC_SI_FROM_SF): Likewise.
2794 (iorxor): Change spacing.
2795 (and_ior_xor): New iterator for AND, IOR, and XOR.
2796 (movsi_from_sf): New insns for SImode/SFmode SUBREG support.
2797 (movdi_from_sf_zero_ext): Likewise.
2798 (mov<mode>_hardfloat, FMOVE32 iterator): Use register_operand
2799 instead of gpc_reg_operand. Add SImode/SFmode SUBREG support.
2800 (movsf_from_si): New insn for SImode/SFmode SUBREG support.
2801 (fma<mode>4): Use gpc_reg_operand instead of register_operand.
2802 (fms<mode>4): Likewise.
2803 (fnma<mode>4): Likewise.
2804 (fnms<mode>4): Likewise.
2805 (nfma<mode>4): Likewise.
2806 (nfms<mode>4): Likewise.
2808 2017-01-04 Marek Polacek <polacek@redhat.com>
2811 * doc/invoke.texi: Document -Wpointer-compare.
2813 2017-01-04 Jakub Jelinek <jakub@redhat.com>
2815 * optc-gen.awk: Emit #error for -W*/-f*/-m* Enum without
2818 * dwarf2out.c (output_loc_list): Don't throw away 64K+ location
2819 descriptions for -gdwarf-5 and emit them as uleb128 instead of
2822 2017-01-04 Kelvin Nilsen <kelvin@gcc.gnu.org>
2825 * doc/sourcebuild.texi (PowerPC-specific attributes): Add
2826 documentation of the powerpc_popcntb_ok attribute.
2827 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
2828 code to issue warning messages if a requested CPU configuration is
2829 not supported by the binary (assembler and loader) toolchain.
2830 (spe_init_builtins): Add two assertions to prevent ICE if attempt is
2831 made to define a built-in function that has been disabled.
2832 (paired_init_builtins): Add assertion to prevent ICE if attempt is
2833 made to define a built-in function that has been disabled.
2834 (altivec_init_builtins): Add comment explaining why definition
2835 of the DST built-in functions is not preceded by an assertion
2836 check. Add assertions to prevent ICE if attempts are made to
2837 define an altivec predicate or an abs* built-in function that has
2839 (htm_init_builtins): Add comment explaining why definition of the
2840 htm built-in functions is not preceded by an assertion check.
2842 2017-01-04 Jeff Law <law@redhat.com>
2844 PR tree-optimizatin/67955
2845 * tree-ssa-alias.c (same_addr_size_stores_p): Check offsets first.
2846 Allow any SSA_VAR_P as the base objects. Use integer_zerop. Verify
2847 the points-to solution does not include pt_null. Use DECL_PT_UID
2850 2017-01-04 Uros Bizjak <ubizjak@gmail.com>
2852 * config/i386/i386.md (HI/SImode test with imm to QImode splitters):
2853 Use gen_int_mode instead of gen_lopwart for const_int operands.
2855 2017-01-04 Jakub Jelinek <jakub@redhat.com>
2857 PR tree-optimization/71563
2858 * match.pd: Simplify X << Y into X if Y is known to be 0 or
2859 out of range value - has low bits known to be zero.
2861 2017-01-04 Alan Modra <amodra@gmail.com>
2863 * Makefile.in (aclocal_deps): Update and order as per aclocal.m4.
2864 * configure: Regenerate.
2865 * config.in: Regenerate.
2867 2017-01-04 Jakub Jelinek <jakub@redhat.com>
2870 * input.c (ebcdic_execution_charset::on_error): Don't use strstr for
2871 a substring of the message, but strcmp with the whole message. Ifdef
2872 ENABLE_NLS, translate the message first using dgettext.
2874 2017-01-03 Jeff Law <law@redhat.com>
2876 PR tree-optimizatin/78856
2877 * tree-ssa-threadupdate.c: Include tree-vectorizer.h.
2878 (mark_threaded_blocks): Remove code to truncate thread paths that
2879 cross multiple loop headers. Instead invalidate the cached loop
2880 iteration information and handle case of a thread path walking
2881 into an irreducible region.
2883 2017-01-03 Michael Meissner <meissner@linux.vnet.ibm.com>
2886 * config/rs6000/rs6000.c (rs6000_split_signbit): Change some
2887 assertions. Add support for doing the signbit if the IEEE 128-bit
2888 floating point value is in a GPR.
2889 * config/rs6000/rs6000.md (Fsignbit): Delete.
2890 (signbit<mode>2_dm): Delete using <Fsignbit> and just use "wa".
2891 Update the length attribute if the value is in a GPR.
2892 (signbit<mode>2_dm_<su>ext): Add combiner pattern to eliminate
2893 the sign or zero extension instruction, since the value is always 0/1.
2894 (signbit<mode>2_dm2): Delete using <Fsignbit>.
2897 * config/rs6000/vsx.md (vsx_extract_<mode>_store_p9): If we are
2898 extracting SImode to a GPR register so that we can generate a
2899 store, limit the vector to be in a traditional Altivec register
2900 for the vextuwrx instruction.
2902 2017-01-03 Ian Lance Taylor <iant@google.com>
2904 * godump.c (go_format_type): Treat ENUMERAL_TYPE like INTEGER_TYPE.
2906 2017-01-03 Martin Sebor <msebor@redhat.com>
2908 PR tree-optimization/78696
2909 * gimple-ssa-sprintf.c (format_floating): Correct handling of
2910 precision. Use MPFR for %f for greater fidelity. Correct handling
2912 (pass_sprintf_length::compute_format_length): Set width and precision
2913 specified by asrerisk to void_node for vararg functions.
2914 (try_substitute_return_value): Adjust dump output.
2916 2017-01-03 David Edelsohn <dje.gcc@gmail.com>
2918 * doc/invoke.texi (RS6000 options): LRA is enabled by default.
2920 2017-01-03 Eric Botcazou <ebotcazou@adacore.com>
2922 * doc/invoke.texi (SPARC options): Document -mlra as the default.
2923 * config/sparc/sparc.c (sparc_option_override): Force LRA unless
2924 -mlra/-mno-lra was passed to the compiler.
2926 2017-01-03 James Cowgill <James.Cowgill@imgtec.com>
2928 PR rtl-optimization/65618
2929 * emit-rtl.c (try_split): Move initialization of "before" and
2930 "after" to just before the call to emit_insn_after_setloc.
2932 2017-01-03 Gerald Pfeifer <gerald@pfeifer.com>
2934 * doc/md.texi (Standard Names): Remove reference to Java frontend.
2936 2017-01-03 Pierre-Marie de Rodat <derodat@adacore.com>
2938 * dwarf2out.c (gen_enumeration_type_die): When
2939 -gno-strict-dwarf, add a DW_AT_encoding attribute.
2941 2017-01-03 Jakub Jelinek <jakub@redhat.com>
2943 PR tree-optimization/78965
2944 * gimple-ssa-sprintf.c (pass_sprintf_length::compute_format_length):
2945 Change first argument from const call_info & to call_info &. For %n
2946 set info.nowrite to false.
2949 * gimple-ssa-sprintf.c (try_substitute_return_value): Don't change
2950 possibly throwing calls.
2952 * genmatch.c (dt_node::gen_kids_1): If generic_exprs include SSA_NAME
2953 and exprs_len || fns_len, emit the code for SSA_NAME next to the exprs
2954 and fns handling, rather than in a separate case SSA_NAME.
2956 2017-01-02 Jeff Law <law@redhat.com>
2958 * config/darwin-driver.c (darwin_driver_init): Const-correctness
2959 fixes for first_period and second_period variables.
2961 2017-01-02 Uros Bizjak <ubizjak@gmail.com>
2964 * config/i386/i386.md (UNSPEC_NOREX_MEM): New unspec.
2965 (*insvqi_1): New insn pattern.
2966 (*insvqi_1_mem_rex64): Ditto.
2968 (*insvqi_3): Rename from *insvqi.
2970 (*extzvqi_mem_rex64): Add UNSPEC_NOREX_MEM tag.
2972 2017-01-02 Gerald Pfeifer <gerald@pfeifer.com>
2974 * doc/cfg.texi (Edges): Remove reference to Java.
2975 (Maintaining the CFG): Ditto.
2977 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
2980 * symtab.c (symtab_node::binds_to_current_def_p): Fix handling of
2981 transparent aliases.
2983 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
2986 * predict.def (PRED_CALL): Update hitrate.
2987 (PRED_INDIR_CALL, PRED_POLYMORPHIC_CALL): New predictors.
2988 * predict.c (tree_estimate_probability_bb): Split CALL predictor
2989 into direct/indirect/polymorphic variants.
2991 2017-01-01 Jakub Jelinek <jakub@redhat.com>
2993 Update copyright years.
2995 * gcc.c (process_command): Update copyright notice dates.
2996 * gcov-dump.c (print_version): Ditto.
2997 * gcov.c (print_version): Ditto.
2998 * gcov-tool.c (print_version): Ditto.
2999 * gengtype.c (create_file): Ditto.
3000 * doc/cpp.texi: Bump @copying's copyright year.
3001 * doc/cppinternals.texi: Ditto.
3002 * doc/gcc.texi: Ditto.
3003 * doc/gccint.texi: Ditto.
3004 * doc/gcov.texi: Ditto.
3005 * doc/install.texi: Ditto.
3006 * doc/invoke.texi: Ditto.
3008 Copyright (C) 2017 Free Software Foundation, Inc.
3010 Copying and distribution of this file, with or without modification,
3011 are permitted in any medium without royalty provided the copyright
3012 notice and this notice are preserved.