1 2017-01-27 Bernd Schmidt <bschmidt@redhat.com>
3 PR rtl-optimization/79194
4 * cprop.c (one_cprop_pass): Move deletion of code after unconditional
5 traps before call to bypass_conditional_jumps.
7 2017-01-27 Vladimir Makarov <vmakarov@redhat.com>
9 PR tree-optimization/71374
10 * lra-constraints.c (check_conflict_input_operands): New.
11 (match_reload): Use it.
13 2017-01-27 Vladimir Makarov <vmakarov@redhat.com>
16 * lra-assigns.c (find_hard_regno_for_1): Take endianess for into
17 account to calculate conflict_set.
19 2017-01-27 Bin Cheng <bin.cheng@arm.com>
21 PR rtl-optimization/78559
22 * combine.c (try_combine): Discard REG_EQUAL and REG_EQUIV for
23 other_insn in combine.
25 2017-01-27 Pekka Jääskeläinen <pekka.jaaskelainen@parmance.com>
27 * builtin-types.def: Use unsigned_char_type_node for BT_UINT8. Use
28 uint16_type_node for BT_UINT16.
30 2017-01-27 David Malcolm <dmalcolm@redhat.com>
32 * doc/sourcebuild.texi (Testsuites): Add "GIMPLE Tests" and
34 (GIMPLE Tests): New node.
35 (RTL Tests): New node.
37 2017-01-27 Richard Biener <rguenther@suse.de>
39 PR tree-optimization/79245
40 * tree-loop-distribution.c (distribute_loop): Apply cost
41 modeling also to detected patterns.
43 2017-01-27 Richard Biener <rguenther@suse.de>
45 PR tree-optimization/71433
46 * tree-vrp.c (register_new_assert_for): Revert earlier changes.
47 (compare_assert_loc): New function.
48 (process_assert_insertions): Sort and optimize assert locations
49 to remove duplicates and push down identical assertions on
50 edges to their destination block.
52 2017-01-27 Richard Biener <rguenther@suse.de>
54 PR tree-optimization/79244
55 * tree-vrp.c (remove_range_assertions): Forcefully propagate
56 out SSA names even if abnormal.
58 2017-01-27 Jakub Jelinek <jakub@redhat.com>
60 * realmpfr.h: Poison MPFR_RND{N,Z,U,D}.
61 * gimple-ssa-sprintf.c (format_floating_max): Use GMP_RNDN
64 2017-01-27 Richard Earnshaw <rearnsha@arm.com>
67 * arm.c (arm_option_override): Don't call build_target_option_node
68 until after doing all option overrides.
69 (arm_valid_target_attribute_tree): Likewise.
71 2017-01-27 Martin Liska <mliska@suse.cz>
73 * doc/invoke.texi (-fprofile-arcs): Document profiling support
74 for {cd}tors and C++ {cd}tors.
76 2017-01-27 Dominik Vogt <vogt@linux.vnet.ibm.com>
78 * config/s390/s390.md ("*setmem_long_and")
79 ("*setmem_long_and_31z"): Use zero_extend instead of and.
81 2017-01-26 Martin Sebor <msebor@redhat.com>
83 * gimple-ssa-sprintf.c (format_floating): Simplify the computation
86 2017-01-26 Martin Sebor <msebor@redhat.com>
88 * gimple-ssa-sprintf.c (format_floating): Test HAVE_XFmode and
89 HAVE_DFmode before using XFmode or DFmode.
90 (parse_directive): Avoid using the z length modifier to avoid
91 the ISO C++98 does not support the ‘z’ gnu_printf length modifier.
94 * gimple-ssa-sprintf.c (adjust_for_width_or_precision): Change
95 to accept adjustment as an array.
96 (get_int_range): New function.
97 (struct directive): Make width and prec arrays.
98 (directive::set_width, directive::set_precision): Call get_int_range.
99 (format_integer, format_floating): Handle width and precision ranges.
100 (format_string, parse_directive): Same.
102 2017-01-26 Jakub Jelinek <jakub@redhat.com>
105 * dwarf2out.c (generate_skeleton_bottom_up): For children with
106 comdat_type_p set, just clone them, but keep the children in the
110 * dwarf2out.c (prune_unused_types): Mark all functions with DIEs
111 which have direct callers with -fvar-tracking-assignments enabled
113 (resolve_addr): Avoid adding skeleton DIEs for DW_AT_call_origin
114 inside of type units.
116 2017-01-26 Martin Sebor <msebor@redhat.com>
119 * gimple-ssa-sprintf.c (struct result_range): Add likely and
121 (struct format_result): Replace number_chars, number_chars_min,
122 and number_chars_max with a single member of struct result_range.
124 (format_result::operator+=): Adjust.
125 (struct fmtresult): Remove bounded. Handle likely and unlikely
127 (fmtresult::adjust_for_width_or_precision): New function.
128 (fmtresult:type_max_digits): New function.
129 (bytes_remaining): Handle likely and unlikely counters.
130 (min_bytes_remaining): Remove.
131 (format_percent): Simplify.
132 (format_integer, format_floating): Set likely and unlikely counters.
133 (get_string_length, format_character, format_string): Same.
134 (format_plain, should_warn_p): New function.
135 (maybe_warn): Call should_warn_p. Update diagnostic messages
136 and handle those for all directives, including plain strings.
137 (format_directive): Handle likely and unlikely counters.
138 Remove unnecessary quoting from diagnostics. Add an informational
141 (pass_sprintf_length::compute_format_length): Simplify.
142 (try_substitute_return_value): Handle likely and unlikely counters.
144 2017-01-26 Carl Love <cel@us.ibm.com>
146 * config/rs6000/rs6000-c (altivec_overloaded_builtins): Remove
147 bogus entries for the P8V_BUILTIN_VEC_VGBBD built-ins
149 2017-01-26 Vladimir Makarov <vmakarov@redhat.com>
152 * lra-assigns.c (setup_live_pseudos_and_spill_after_risky): Take
153 endianess for subregs into account.
154 * lra-constraints.c (lra_constraints): Do risky transformations
155 always on the first iteration.
156 * lra-lives.c (check_pseudos_live_through_calls): Add arg
157 last_call_used_reg_set.
158 (process_bb_lives): Define and use last_call_used_reg_set.
159 * lra.c (lra): Always continue after lra_constraints on the first
162 2017-01-26 Kirill Yukhin <kirill.yukhin@gmail.com>
164 * gcc.target/i386/avx512bw-kshiftlq-2.c: Use unsigned long long
166 * gcc.target/i386/avx512bw-kshiftrq-2.c: Ditto.
168 2017-01-26 Jakub Jelinek <jakub@redhat.com>
170 * config/i386/avx512fintrin.h (_ktest_mask16_u8,
171 _ktestz_mask16_u8, _ktestc_mask16_u8, _kadd_mask16): Move to ...
172 * config/i386/avx512dqintrin.h (_ktest_mask16_u8,
173 _ktestz_mask16_u8, _ktestc_mask16_u8, _kadd_mask16): ... here.
174 * config/i386/i386-builtin.def (__builtin_ia32_ktestchi,
175 __builtin_ia32_ktestzhi, __builtin_ia32_kaddhi): Use
176 OPTION_MASK_ISA_AVX512DQ instead of OPTION_MASK_ISA_AVX512F.
177 * config/i386/sse.md (SWI1248_AVX512BWDQ2): New mode iterator.
178 (kadd<mode>, ktest<mode>): Use it instead of SWI1248_AVX512BWDQ.
180 2017-01-26 Marek Polacek <polacek@redhat.com>
183 * fold-const.c (operand_equal_p) [COND_EXPR]: Use OP_SAME_WITH_NULL
184 for the third operand.
186 2017-01-26 Jakub Jelinek <jakub@redhat.com>
189 * omp-low.c (struct omp_context): Add simt_stmt field.
190 (scan_omp_for): Return omp_context *.
191 (scan_omp_simd): Set simt_stmt on the non-_simt_ SIMD
192 context to the _simt_ SIMD stmt.
193 (lower_omp_for): For combined SIMD with sibling _simt_
194 SIMD, make sure to use the same decls in _looptemp_
195 clauses as in the sibling.
197 2017-01-26 David Sherwood <david.sherwood@arm.com>
200 * gimplify.c (omp_notice_variable): Add GOVD_SEEN flag to variables in
203 2017-01-26 Jakub Jelinek <jakub@redhat.com>
206 * reg-stack.c (emit_swap_insn): Instead of fld a; fld b; fxchg %st(1);
207 emit fld b; fld a; if possible.
209 * brig-builtins.def: Update copyright years.
210 * config/arm/arm_acle_builtins.def: Update copyright years.
212 2017-01-25 Michael Meissner <meissner@linux.vnet.ibm.com>
215 * config/rs6000/vsx.md (vsx_extract_<mode>_store): Use wY
216 constraint instead of o for the stxsd instruction.
218 2017-01-25 Carl Love <cel@us.ibm.com>
220 * config/rs6000/rs6000-c (altivec_overloaded_builtins): Fix order
221 of entries for ALTIVEC_BUILTIN_VEC_PACKS and P8V_BUILTIN_VEC_VGBBD.
223 2017-01-25 Jonathan Wakely <jwakely@redhat.com>
225 * doc/invoke.texi (C++ Dialect Options): Fix typo.
227 2017-01-25 Richard Biener <rguenther@suse.de>
229 PR tree-optimization/69264
230 * target.def (vector_alignment_reachable): Improve documentation.
231 * doc/tm.texi: Regenerate.
232 * targhooks.c (default_builtin_vector_alignment_reachable): Simplify
234 * tree-vect-data-refs.c (vect_supportable_dr_alignment): Revert
235 earlier changes with respect to TYPE_USER_ALIGN.
236 (vector_alignment_reachable_p): Likewise. Improve dumping.
238 2016-01-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
241 * config/arm/arm.md (xordi3): Force constant operand into a register
244 2016-01-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
246 * doc/invoke.texi (-fstore-merging): Correct default optimization
247 levels at which it is enabled.
248 (-O): Move -fstore-merging from list to...
251 2017-01-25 Richard Biener <rguenther@suse.de>
254 * omp-expand.c: Include debug.h.
255 (expand_omp_taskreg): Make sure to generate early debug before
256 outlining anything from a function.
257 (expand_omp_target): Likewise.
258 (grid_expand_target_grid_body): Likewise.
260 2017-01-25 Maxim Ostapenko <m.ostapenko@samsung.com>
263 * asan.c (get_translation_unit_decl): New function.
264 (asan_add_global): Extract modules file name from globals
265 TRANSLATION_UNIT_DECL name.
267 2017-01-24 Eric Botcazou <ebotcazou@adacore.com>
270 * config/arm/arm.c (arm_function_ok_for_sibcall): Add back restriction
271 for long calls with APCS frame and VFP.
273 2017-01-24 David Malcolm <dmalcolm@redhat.com>
275 * cfg.c (original_copy_tables_initialized_p): New function.
276 * cfg.h (original_copy_tables_initialized_p): New decl.
277 * cfgrtl.c (relink_block_chain): Guard the call to
278 free_original_copy_tables with a call to
279 original_copy_tables_initialized_p.
280 * cgraph.h (symtab_node::native_rtl_p): New decl.
281 * cgraphunit.c (symtab_node::native_rtl_p): New function.
282 (symtab_node::needed_p): Don't assert for early assembly output
284 (cgraph_node::finalize_function): Set "force_output" for __RTL
286 (cgraph_node::analyze): Bail out early for __RTL functions.
287 (analyze_functions): Update assertion to support __RTL functions.
288 (cgraph_node::expand): Bail out early for __RTL functions.
289 * final.c (rest_of_clean_state): Don't call delete_tree_ssa for
291 * function.h (struct function): Update comment for field
293 * gimple-expr.c: Include "tree-pass.h".
294 (gimple_has_body_p): Return false for __RTL functions.
295 * Makefile.in (OBJS): Add run-rtl-passes.o.
296 * pass_manager.h (gcc::pass_manager::get_rest_of_compilation): New
298 (gcc::pass_manager::get_clean_slate): New accessor.
299 * passes.c: Include "insn-addr.h".
300 (should_skip_pass_p): Add logging. Update logic for running
301 "expand" to be compatible with both __GIMPLE and __RTL. Guard
302 property-provider override so it is only done for gimple passes.
304 (skip_pass): New function.
305 (execute_one_pass): Call skip_pass when skipping passes.
306 * read-md.c (md_reader::read_char): Support filtering
307 the input to a subset of line numbers.
308 (md_reader::md_reader): Initialize fields
309 m_first_line and m_last_line.
310 (md_reader::read_file_fragment): New function.
311 * read-md.h (md_reader::read_file_fragment): New decl.
312 (md_reader::m_first_line): New field.
313 (md_reader::m_last_line): New field.
314 * read-rtl-function.c (function_reader::create_function): Only
315 create cfun if it doesn't already exist. Set PROP_rtl on cfun's
316 curr_properties. Set DECL_INITIAL to a dummy block.
317 (read_rtl_function_body_from_file_range): New function.
318 * read-rtl-function.h (read_rtl_function_body_from_file_range):
320 * run-rtl-passes.c: New file.
321 * run-rtl-passes.h: New file.
323 2017-01-24 Jeff Law <law@redhat.com>
325 * config/microblaze/microblaze.h (ASM_FORMAT_PRIVATE_NAME): Increase
328 2017-01-24 Bin Cheng <bin.cheng@arm.com>
330 PR tree-optimization/79159
331 * tree-ssa-loop-niter.c (get_cst_init_from_scev): New function.
332 (record_nonwrapping_iv): Improve boundary using above function if no
333 value range information.
335 2017-01-24 Pekka Jääskeläinen <pekka@parmance.com>
336 Martin Jambor <mjambor@suse.cz>
338 * brig-builtins.def: New file.
339 * builtins.def (DEF_HSAIL_BUILTIN): New macro.
340 (DEF_HSAIL_ATOMIC_BUILTIN): Likewise.
341 (DEF_HSAIL_SAT_BUILTIN): Likewise.
342 (DEF_HSAIL_INTR_BUILTIN): Likewise.
343 (DEF_HSAIL_CVT_ZEROI_SAT_BUILTIN): Likewise.
344 * builtin-types.def (BT_INT8): New.
345 (BT_INT16): Likewise.
346 (BT_UINT8): Likewise.
347 (BT_UINT16): Likewise.
348 (BT_FN_ULONG): Likewise.
349 (BT_FN_UINT_INT): Likewise.
350 (BT_FN_UINT_ULONG): Likewise.
351 (BT_FN_UINT_LONG): Likewise.
352 (BT_FN_UINT_PTR): Likewise.
353 (BT_FN_ULONG_PTR): Likewise.
354 (BT_FN_INT8_FLOAT): Likewise.
355 (BT_FN_INT16_FLOAT): Likewise.
356 (BT_FN_UINT32_FLOAT): Likewise.
357 (BT_FN_UINT16_FLOAT): Likewise.
358 (BT_FN_UINT8_FLOAT): Likewise.
359 (BT_FN_UINT64_FLOAT): Likewise.
360 (BT_FN_UINT16_UINT32): Likewise.
361 (BT_FN_UINT32_UINT16): Likewise.
362 (BT_FN_UINT16_UINT16_UINT16): Likewise.
363 (BT_FN_INT_PTR_INT): Likewise.
364 (BT_FN_UINT_PTR_UINT): Likewise.
365 (BT_FN_LONG_PTR_LONG): Likewise.
366 (BT_FN_ULONG_PTR_ULONG): Likewise.
367 (BT_FN_VOID_UINT64_UINT64): Likewise.
368 (BT_FN_UINT8_UINT8_UINT8): Likewise.
369 (BT_FN_INT8_INT8_INT8): Likewise.
370 (BT_FN_INT16_INT16_INT16): Likewise.
371 (BT_FN_INT_INT_INT): Likewise.
372 (BT_FN_UINT_FLOAT_UINT): Likewise.
373 (BT_FN_FLOAT_UINT_UINT): Likewise.
374 (BT_FN_ULONG_UINT_UINT): Likewise.
375 (BT_FN_ULONG_UINT_PTR): Likewise.
376 (BT_FN_ULONG_ULONG_ULONG): Likewise.
377 (BT_FN_UINT_UINT_UINT): Likewise.
378 (BT_FN_VOID_UINT_PTR): Likewise.
379 (BT_FN_UINT_UINT_PTR: Likewise.
380 (BT_FN_UINT32_UINT64_PTR): Likewise.
381 (BT_FN_INT_INT_UINT_UINT): Likewise.
382 (BT_FN_UINT_UINT_UINT_UINT): Likewise.
383 (BT_FN_UINT_UINT_UINT_PTR): Likewise.
384 (BT_FN_UINT_ULONG_ULONG_UINT): Likewise.
385 (BT_FN_ULONG_ULONG_ULONG_ULONG): Likewise.
386 (BT_FN_LONG_LONG_UINT_UINT): Likewise.
387 (BT_FN_ULONG_ULONG_UINT_UINT): Likewise.
388 (BT_FN_VOID_UINT32_UINT64_PTR): Likewise.
389 (BT_FN_VOID_UINT32_UINT32_PTR): Likewise.
390 (BT_FN_UINT_UINT_UINT_UINT_UINT): Likewise.
391 (BT_FN_UINT_FLOAT_FLOAT_FLOAT_FLOAT): Likewise.
392 (BT_FN_ULONG_ULONG_ULONG_UINT_UINT): Likewise.
393 * doc/frontends.texi: List BRIG FE.
394 * doc/install.texi (Testing): Add BRIG tesring requirements.
395 * doc/invoke.texi (Overall Options): Mention BRIG.
396 * doc/standards.texi (Standards): Doucment BRIG HSA version.
398 2017-01-24 Richard Biener <rguenther@suse.de>
401 * ipa-devirt.c (odr_types_equivalent_p): Fix typo in diagnostic.
403 2017-01-24 Martin Jambor <mjambor@suse.cz>
406 * ipa-prop.c (ipa_free_all_node_params): Call summary destructor.
407 * ipa-prop.c (ipa_node_params_t::insert): Initialize fields known_csts
410 2017-01-24 Aldy Hernandez <aldyh@redhat.com>
413 * gimple-ssa-warn-alloca.c (alloca_call_type): Make sure
414 casts from signed to unsigned really don't have a range.
416 2017-01-24 Markus Trippelsdorf <markus@trippelsdorf.de>
418 * gimple-ssa-sprintf.c (format_floating): Change MPFR_RNDx to
419 GMP_RNDx for compatiblity.
421 2017-01-24 Martin Liska <mliska@suse.cz>
424 * tree-ssa-reassoc.c (rewrite_expr_tree_parallel): Insert assert
425 that would prevent us to call alloca with -1 as argument.
427 2017-01-24 Jakub Jelinek <jakub@redhat.com>
429 * dwarf2out.c (output_compilation_unit_header, output_file_names):
430 Avoid -Wformat-security warning.
432 2017-01-23 Andrew Pinski <apinski@cavium.com>
434 * config/aarch64/aarch64.c (thunderx2t99_addrcost_table): Improve
437 2017-01-23 Martin Sebor <msebor@redhat.com>
440 * gimple-ssa-sprintf.c (warn_level): New global.
441 (format_integer): Use it here and throughout the rest of the file.
442 Use the same switch to compute sign as base.
443 (maybe_warn): New function.
444 (format_directive): Factor out warnings into maybe_warn.
445 Add debugging output. Use warn_level.
446 (add_bytes): Use warn_level.
447 (pass_sprintf_length::compute_format_length): Add debugging output.
448 (try_substitute_return_value): Same.
449 (pass_sprintf_length::handle_gimple_call): Set and use warn_level.
452 * gimple-ssa-sprintf.c (struct format_result): Remove constant member.
453 (struct fmtresult, format_integer, format_floating): Adjust.
454 (fmtresult::fmtresult): Set max correctly in two argument ctor.
455 (get_string_length, format_string,format_directive): Same.
456 (pass_sprintf_length::compute_format_length): Same.
457 (try_substitute_return_value): Simplify slightly.
460 * gimple-ssa-sprintf.c (pass_sprintf_length::gate): Adjust formatting.
461 (fmtresult::operator+=): Outlined.
462 (struct fmtresult): Add ctors.
463 (struct conversion_spec): Rename...
464 (struct directive): ...to this. Add and remove data members.
465 (directive::set_width, directive::set_precision): New functions.
466 (format_percent): Use fmtresult ctor.
467 (get_width_and_precision): Remove.
468 (format_integer): Make naming changes. Avoid computing width and
470 (format_floating): Same. Adjust indentation.
471 (format_character, format_none): New functions.
472 (format_string): Moved character handling to format_character.
473 (format_directive): Remove arguments, change return type.
474 (parse_directive): New function.
475 (pass_sprintf_length::compute_format_length): Move directive
476 parsing to parse_directive.
478 2017-01-23 Jakub Jelinek <jakub@redhat.com>
480 * tree.h (assign_assembler_name_if_neeeded): Rename to ...
481 (assign_assembler_name_if_needed): ... this.
482 * tree.c (assign_assembler_name_if_neeeded): Rename to ...
483 (assign_assembler_name_if_needed): ... this.
484 (free_lang_data_in_cgraph): Adjust callers.
485 * cgraphunit.c (cgraph_node::analyze): Likewise.
486 * omp-expand.c (expand_omp_taskreg, expand_omp_target):
489 2017-01-23 Richard Biener <rguenther@suse.de>
491 PR tree-optimization/79088
492 PR tree-optimization/79188
493 * tree-ssa-threadupdate.c (mark_threaded_blocks): Move code
494 resetting loop bounds after last path deletion. Reset loop
495 bounds of the target loop, make code match the comments.
496 * tree-ssa-threadbackwards.c (pass_early_thread_jumps::execute):
497 Make sure loops need no fixups.
499 2017-01-23 Kelvin Nilsen <kelvin@gcc.gnu.org>
501 * config/rs6000/rs6000-builtin.def (VSIEDPF): Add scalar insert
502 exponent support with double type for first argument.
503 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Changed
504 type returned by __builtin_vec_extract_sig,
505 __builtin_vec_extract_sig_sp, and __builtin_vec_extract_sig_dp
506 functions from "vector int" to "vector unsigned int" or from
507 "vector long long int" to "vector unsigned long long int".
508 Changed type returned by __builtin_vec_extract_exp,
509 __builtin_vec_extract_exp_sp, and __builtin_vec_extract_exp_dp
510 functions from "vector int" to "vector unsigned int" or from
511 "vector long long int" to "vector unsigned long long int".
512 Changed return type of __builtin_vec_test_data_class,
513 __builtin_vec_test_data_class_sp, and
514 __builtin_vec_test_data_class_dp from "vector int" to
515 "vector bool int" or from "vector long long int" to "vector bool
516 long long int" and changed second argument type from "unsigned
517 int" to "int". Added new overloaded function forms "vector float
518 __builtin_vec_insert_exp (vector float, vector unsigned int)" and
519 "vector float __builtin_vec_insert_exp_sp (vector float, vector
520 unsigned int)" and "vector double __builtin_vec_insert_exp (vector
521 double, vector unsigned long long int)" and "vector double
522 __builtin_vec_insert_exp_dp (vector double, vector unsigned long
523 long int)". Changed return type of
524 __builtin_scalar_test_data_class and
525 __builtin_scalar_test_data_class_sp and
526 __builtin_scalar_test_data_class_dp from "unsigned int" to "bool
527 int" and changed second argument from "unsigned int" to "int".
528 Changed type returned by __builtin_scalar_test_neg,
529 __builtin_scalar_test_neg_sp, and __builtin_scalar_test_neg_dp
530 from "int" to "bool int". Added new overloaded function form
531 "double __builtin_scalar_insert_exp (double, unsigned long long int)".
532 * config/rs6000/vsx.md (xsiexpdpf): New insn for scalar insert
533 exponent double-precision with floating point first argument.
534 * doc/extend.texi (PowerPC AltiVec Built-in Functions): Adjust
535 documentation of scalar_test_data_class, scalar_test_neg,
536 scalar_extract_sig, scalar_extract_exp, scalar_insert_exp,
537 vector_extract_exp, vec_extract_sig, vec_insert_exp, and
538 vec_test_data_class built-in functions to reflect refinements in
539 their type signatures.
541 2017-01-23 Andreas Tobler <andreast@gcc.gnu.org>
543 * config/aarch64/aarch64.c (aarch64_elf_asm_constructor): Increase
545 (aarch64_elf_asm_destructor): Likewise.
547 2017-01-23 Bernd Schmidt <bschmidt@redhat.com>
549 PR rtl-optimization/78634
550 * config/i386/i386.c (ix86_max_noce_ifcvt_seq_cost): New function.
551 (TARGET_MAX_NOCE_IFCVT_SEQ_COST): Define.
552 * ifcvt.c (noce_try_cmove): Add missing cost check.
554 PR rtl-optimization/71724
555 * combine.c (if_then_else_cond): Look for situations where it is
556 beneficial to undo the work of one of the recursive calls.
558 2017-01-23 Bin Cheng <bin.cheng@arm.com>
560 PR tree-optimization/70754
561 * tree-predcom.c (stmt_combining_refs): New parameter INSERT_BEFORE.
562 (reassociate_to_the_same_stmt): New parameter INSERT_BEFORE. Insert
563 combined stmt before it if not NULL.
564 (combine_chains): Process refs reversely and compute dominance point
567 2017-01-23 Martin Liska <mliska@suse.cz>
569 PR tree-optimization/79196
570 * tree-ssa-strlen.c (fold_strstr_to_memcmp): Rename to ...
571 (fold_strstr_to_strncmp): ... this. Fold the pattern to strncmp
573 (strlen_optimize_stmt): Call the renamed function.
575 2017-01-23 Michael Matz <matz@suse.de>
577 PR tree-optimization/78384
578 * tree-ssa-loop-split.c (patch_loop_exit): Use correct edge.
580 2017-01-23 Richard Biener <rguenther@suse.de>
582 PR tree-optimization/79186
583 * tree-vrp.c (register_new_assert_for): Make sure we've seen
584 both incoming edges before moving an assert.
586 2017-01-23 Martin Jambor <mjambor@suse.cz>
588 * ipa-prop.c (load_from_param_1): Removed.
589 (load_from_unmodified_param): Bits from load_from_param_1 put back
591 (load_from_param): Removed.
592 (compute_complex_assign_jump_func): Removed stmt2 and just replaced it
593 with stmt. Reverted back to use of load_from_unmodified_param.
595 2017-01-23 Martin Jambor <mjambor@suse.cz>
598 * ipa-prop.h (ipa_param_descriptor): Anotate with with GTY(()).
599 (ipa_node_params): Annotate with GTY((for_user)). Make descriptors
600 field a pointer to garbage collected vector, mark lattices and
601 ipcp_orig_node with GTY((skip)).
602 (ipa_get_param_count): Adjust to descriptors being a pointer.
603 (ipa_get_param): Likewise.
604 (ipa_get_type): Likewise.
605 (ipa_get_param_move_cost): Likewise.
606 (ipa_set_param_used): Likewise.
607 (ipa_get_controlled_uses): Likewise.
608 (ipa_set_controlled_uses): Likewise.
609 (ipa_is_param_used): Likewise.
610 (ipa_node_params_t): Move into garbage collector. New methods insert
612 (ipa_node_params_sum): Annotate wth GTY(()).
613 (ipa_check_create_node_params): Adjust to ipa_node_params_sum being
615 (ipa_load_from_parm_agg): Adjust declaration.
616 * ipa-icf.c (param_used_p): Adjust to descriptors being a pointer.
617 * ipa-profile.c (ipa_profile): Likewise.
618 * ipa-prop.c (ipa_get_param_decl_index_1): Likewise.
619 (ipa_populate_param_decls): Make descriptors parameter garbage
621 (ipa_dump_param): Adjust to descriptors being a pointer.
622 (ipa_alloc_node_params): Likewise.
623 (ipa_initialize_node_params): Likewise.
624 (load_from_param_1): Make descriptors parameter garbage collected.
625 (load_from_unmodified_param): Likewise.
626 (load_from_param): Likewise.
627 (ipa_load_from_parm_agg): Likewise.
628 (ipa_node_params::~ipa_node_params): Removed.
629 (ipa_free_all_node_params): Remove call to delete operator.
630 (ipa_node_params_t::insert): New.
631 (ipa_node_params_t::remove): Likewise.
632 (ipa_node_params_t::duplicate): Adjust to descriptors being a pointer,
633 copy known_csts and known_contexts vectors.
634 (ipa_read_node_info): Adjust to descriptors being a pointer.
635 (ipcp_modif_dom_walker): Make m_descriptors field garbage
637 (ipcp_transform_function): Make descriptors variable garbage
640 2017-01-23 Andrew Senkevich <andrew.senkevich@intel.com>
642 * config/i386/avx512bwintrin.h: Add k-mask test, kortest intrinsics.
643 * config/i386/avx512dqintrin.h: Ditto.
644 * config/i386/avx512fintrin.h: Ditto.
645 * gcc/config/i386/i386.c: Handle new builtins.
646 * config/i386/i386-builtin.def: Add new builtins.
647 * config/i386/sse.md (ktest<mode>, kortest<mode>): New.
648 (UNSPEC_KORTEST, UNSPEC_KTEST): New.
650 2017-01-23 Jakub Jelinek <jakub@redhat.com>
651 Martin Liska <mliska@suse.cz>
653 * asan.h: Define ASAN_USE_AFTER_SCOPE_ATTRIBUTE.
654 * asan.c (asan_expand_poison_ifn): Support stores and use
655 appropriate ASAN report function.
656 * internal-fn.c (expand_ASAN_POISON_USE): New function.
657 * internal-fn.def (ASAN_POISON_USE): Declare.
658 * tree-into-ssa.c (maybe_add_asan_poison_write): New function.
659 (maybe_register_def): Create ASAN_POISON_USE when sanitizing.
660 * tree-ssa-dce.c (eliminate_unnecessary_stmts): Remove
661 ASAN_POISON calls w/o LHS.
662 * tree-ssa.c (execute_update_addresses_taken): Create clobber
663 for ASAN_MARK (UNPOISON, &x, ...) in order to prevent usage of a LHS
664 from ASAN_MARK (POISON, &x, ...) coming to a PHI node.
665 * gimplify.c (asan_poison_variables): Add attribute
666 use_after_scope_memory to variables that really needs to live
668 * tree-ssa.c (is_asan_mark_p): Do not rewrite into SSA when
669 having the attribute.
671 2017-01-23 Martin Liska <mliska@suse.cz>
673 * asan.c (create_asan_shadow_var): New function.
674 (asan_expand_poison_ifn): Likewise.
675 * asan.h (asan_expand_poison_ifn): New declaration.
676 * internal-fn.c (expand_ASAN_POISON): Likewise.
677 * internal-fn.def (ASAN_POISON): New builtin.
678 * sanopt.c (pass_sanopt::execute): Expand
679 asan_expand_poison_ifn.
680 * tree-inline.c (copy_decl_for_dup_finish): Make function
682 * tree-inline.h (copy_decl_for_dup_finish): Likewise.
683 * tree-ssa.c (is_asan_mark_p): New function.
684 (execute_update_addresses_taken): Rewrite local variables
685 (identified just by use-after-scope as addressable) into SSA.
687 2017-01-22 Gerald Pfeifer <gerald@pfeifer.com>
689 * doc/install.texi (Specific): opensource.apple.com uses https
690 now. Remove trailing slash.
692 2017-01-22 Gerald Pfeifer <gerald@pfeifer.com>
694 * README.Portability: Remove note on an Irix compatibility issue.
696 2017-01-22 Dimitry Andric <dim@FreeBSD.org>
698 * gcov.c (INCLUDE_ALGORITHM): Define.
699 (INCLUDE_VECTOR): Define.
700 No longer include <vector> and <algorithm> directly.
702 2017-01-21 Gerald Pfeifer <gerald@pfeifer.com>
704 * doc/extend.texi (Thread-Local): Change www.akkadia.org reference
706 * doc/invoke.texi (Code Gen Options): Ditto.
708 2017-01-21 Jan Hubicka <hubicka@ucw.cz>
711 * cfg.c (update_bb_profile_for_threading): Fix updating of probablity.
713 2017-01-21 Bernd Schmidt <bschmidt@redhat.com>
715 rtl-optimization/79125
716 * cprop.c (local_cprop_pass): Handle cases where we make an
719 2017-01-20 Segher Boessenkool <segher@kernel.crashing.org>
723 * config/rs6000/rs6000.c (rs6000_gimplify_va_arg): Adjust address to
724 read from, for big endian.
726 2017-01-20 Jiong Wang <jiong.wang@arm.com>
728 * config/aarch64/aarch64-builtins.c (aarch64_init_builtins): Register
729 register pauth builtins for LP64 only.
731 2017-01-20 Marek Polacek <polacek@redhat.com>
734 * gimplify.c (should_warn_for_implicit_fallthrough): Handle consecutive
737 2017-01-20 Alexander Monakov <amonakov@ispras.ru>
739 * omp-expand.c (expand_omp_simd): Clear PROP_gimple_lomp_dev regardless
741 * omp-offload.c (pass_omp_device_lower::gate): Use PROP_gimple_lomp_dev.
742 * passes.c (dump_properties): Handle PROP_gimple_lomp_dev.
743 * tree-inline.c (expand_call_inline): Propagate PROP_gimple_lomp_dev.
745 2017-01-20 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
748 * config/arm/arm.c (neon_valid_immediate): Reject vector constants
749 in big-endian mode when they are not a single duplicated value.
751 2017-01-20 Richard Biener <rguenther@suse.de>
753 * BASE-VER: Bump to 7.0.1.
755 2017-01-20 Alexander Monakov <amonakov@ispras.ru>
757 * omp-low.c (omplow_simd_context): New struct. Use it...
758 (lower_rec_simd_input_clauses): ...here and...
759 (lower_rec_input_clauses): ...here to hold common data. Adjust all
760 references to idx, lane, max_vf, is_simt.
762 2017-01-20 Graham Markall <graham.markall@embecosm.com>
764 * config/arc/arc.h (LINK_SPEC): Use arclinux_nps emulation when
767 2017-01-20 Martin Jambor <mjambor@suse.cz>
769 * hsa.h: Renaed to hsa-common.h. Adjusted a comment.
770 * hsa.c: Renaed to hsa-common.c. Change include of gt-hsa.h to
772 * Makefile.in (OBJS): Rename hsa.o to hsa-common.o.
773 (GTFILES): Rename hsa.c to hsa-common.c.
774 * hsa-brig.c: Change include of hsa.h to hsa-common.h.
775 * hsa-dump.c: Likewise.
776 * hsa-gen.c: Likewise.
777 * hsa-regalloc.c: Likewise.
778 * ipa-hsa.c: Likewise.
779 * omp-expand.c: Likewise.
780 * omp-low.c: Likewise.
781 * toplev.c: Likewise.
783 2017-01-20 Marek Polacek <polacek@redhat.com>
786 * doc/invoke.texi: Document -Wduplicated-branches.
787 * fold-const.c (operand_equal_p): Handle MODIFY_EXPR, INIT_EXPR,
788 COMPOUND_EXPR, PREDECREMENT_EXPR, PREINCREMENT_EXPR,
789 POSTDECREMENT_EXPR, POSTINCREMENT_EXPR, CLEANUP_POINT_EXPR, EXPR_STMT,
790 STATEMENT_LIST, and RETURN_EXPR. For non-pure non-const functions
791 return 0 only when not OEP_LEXICOGRAPHIC.
792 (fold_build_cleanup_point_expr): Use the expression
793 location when building CLEANUP_POINT_EXPR.
794 * tree-core.h (enum operand_equal_flag): Add OEP_LEXICOGRAPHIC.
795 * tree.c (add_expr): Handle error_mark_node.
797 2017-01-20 Martin Liska <mliska@suse.cz>
800 * tree-profile.c (init_ic_make_global_vars): Do not call
802 (gimple_init_gcov_profiler): Likewise.
804 2017-01-20 Martin Liska <mliska@suse.cz>
807 * cgraph.h (maybe_create_reference): Remove argument and
809 * cgraphclones.c (cgraph_node::create_virtual_clone): Remove one
811 * ipa-cp.c (create_specialized_node): Likewise.
812 * symtab.c (symtab_node::maybe_create_reference): Handle
813 VAR_DECLs and ADDR_EXPRs and select ipa_ref_use type.
815 2017-01-20 Martin Liska <mliska@suse.cz>
817 * read-rtl-function.c (function_reader::create_function): Use
818 build_decl instread of build_decl_stat.
820 2017-01-20 Andrew Senkevich <andrew.senkevich@intel.com>
822 * config/i386/avx512bwintrin.h: Add k-mask registers shift intrinsics.
823 * config/i386/avx512dqintrin.h: Ditto.
824 * config/i386/avx512fintrin.h: Ditto.
825 * config/i386/i386-builtin-types.def: Add new types.
826 * gcc/config/i386/i386.c: Handle new types.
827 * config/i386/i386-builtin.def (__builtin_ia32_kshiftliqi)
828 (__builtin_ia32_kshiftlihi, __builtin_ia32_kshiftlisi)
829 (__builtin_ia32_kshiftlidi, __builtin_ia32_kshiftriqi)
830 (__builtin_ia32_kshiftrihi, __builtin_ia32_kshiftrisi)
831 (__builtin_ia32_kshiftridi): New.
832 * config/i386/sse.md (k<code><mode>): Rename *k<code><mode>.
834 2017-01-19 Segher Boessenkool <segher@kernel.crashing.org>
838 * config/rs6000/rs6000.c (TARGET_STACK_PROTECT_GUARD): Unconditionally
839 define to rs6000_init_stack_protect_guard.
840 (rs6000_init_stack_protect_guard): New function.
842 2017-01-19 Matthew Fortune <matthew.fortune@imgtec.com>
843 Yunqiang Su <yunqiang.su@imgtec.com>
845 * config.gcc (supported_defaults): Add madd4.
846 (with_madd4): Add validation.
847 (all_defaults): Add madd4.
848 * config/mips/mips.opt (mmadd4): New option.
849 * gcc/config/mips/mips.h (OPTION_DEFAULT_SPECS): Add a default for
851 (TARGET_CPU_CPP_BUILTINS): Add builtin_define for
853 (ISA_HAS_UNFUSED_MADD4): Gate with mips_madd4.
854 (ISA_HAS_FUSED_MADD4): Likewise.
855 * gcc/doc/invoke.texi (-mmadd4): Document the new option.
856 * gcc/doc/install.texi (--with-madd4): Document the new option.
858 2017-01-19 Jiong Wang <jiong.wang@arm.com>
860 * config/aarch64/aarch64-builtins.c (enum aarch64_builtins): New
861 entries for AARCH64_PAUTH_BUILTIN_XPACLRI,
862 AARCH64_PAUTH_BUILTIN_PACIA1716, AARCH64_PAUTH_BUILTIN_AUTIA1716.
863 (aarch64_init_pauth_hint_builtins): New.
864 (aarch64_init_builtins): Call aarch64_init_pauth_hint_builtins.
865 (aarch64_expand_builtin): Expand new builtins.
867 2017-01-19 Jiong Wang <jiong.wang@arm.com>
869 * reg-notes.def (CFA_TOGGLE_RA_MANGLE): New reg-note.
870 * combine-stack-adj.c (no_unhandled_cfa): Handle
871 REG_CFA_TOGGLE_RA_MANGLE.
872 * dwarf2cfi.c (dwarf2out_frame_debug): Handle REG_CFA_TOGGLE_RA_MANGLE.
873 * config/aarch64/aarch64.c (aarch64_expand_prologue): Generates DWARF
874 info for return address signing.
875 (aarch64_expand_epilogue): Likewise.
877 2017-01-19 Jiong Wang <jiong.wang@arm.com>
879 * config/aarch64/aarch64-opts.h (aarch64_function_type): New enum.
880 * config/aarch64/aarch64-protos.h
881 (aarch64_return_address_signing_enabled): New declaration.
882 * config/aarch64/aarch64.c (aarch64_return_address_signing_enabled):
884 (aarch64_expand_prologue): Sign return address before it's pushed onto
886 (aarch64_expand_epilogue): Authenticate return address fetched from
888 (aarch64_override_options): Sanity check for ILP32 and ISA level.
889 (aarch64_attributes): New function attributes for "sign-return-address".
890 * config/aarch64/aarch64.md (UNSPEC_AUTI1716, UNSPEC_AUTISP,
891 UNSPEC_PACI1716, UNSPEC_PACISP, UNSPEC_XPACLRI): New unspecs.
892 ("*do_return"): Generate combined instructions according to key index.
893 ("<pauth_mnem_prefix>sp", "<pauth_mnem_prefix1716", "xpaclri"): New.
894 * config/aarch64/iterators.md (PAUTH_LR_SP, PAUTH_17_16): New integer
896 (pauth_mnem_prefix, pauth_hint_num_a): New integer attributes.
897 * config/aarch64/aarch64.opt (msign-return-address=): New.
898 * doc/extend.texi (AArch64 Function Attributes): Documents
899 "sign-return-address=".
900 * doc/invoke.texi (AArch64 Options): Documents "-msign-return-address=".
902 2017-01-19 Matthew Fortune <matthew.fortune@imgtec.com>
904 * doc/invoke.texi: Add missing -mlxc1-sxc1 options to
905 overall option summary.
907 2017-01-19 Jiong Wang <jiong.wang@arm.com>
909 * config/aarch64/aarch64-arches.def: New entry for "armv8.3-a".
910 * config/aarch64/aarch64.h (AARCH64_FL_V8_3, AARCH64_FL_FOR_ARCH8_3,
911 AARCH64_ISA_V8_3, TARGET_ARMV8_3): New.
912 * doc/invoke.texi (AArch64 Options): Document "armv8.3-a".
914 2017-01-19 Michael Meissner <meissner@linux.vnet.ibm.com>
916 * config/rs6000/rs6000-cpus.def (ISA_3_0_MASKS_SERVER): Enable
917 -mpower9-minmax by default for -mcpu=power9.
918 (ISA_3_MASKS_IEEE): Require -mvsx-small-integer to enable IEEE
919 128-bit floating point.
921 2017-01-20 Alan Modra <amodra@gmail.com>
923 * config/rs6000/rs6000.md (cmpstrnsi, cmpstrsi): Fail if
926 2017-01-20 Alan Modra <amodra@gmail.com>
929 * config/rs6000/rs6000.c (expand_strn_compare): Get the asm name
930 for strcmp and strncmp from corresponding builtin decl.
932 2017-01-19 Uros Bizjak <ubizjak@gmail.com>
934 * config.gcc (x86_64-*-rtems*): Use i386/rtemself.h
935 instead of i386/rtems-64.h.
936 * config/i386/rtems-64.h: Remove.
938 2017-01-19 Uros Bizjak <ubizjak@gmail.com>
942 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
944 * config/i386/rtemself.h (LONG_DOUBLE_TYPE_SIZE): New define.
946 2017-01-19 Tamar Christina <tamar.christina@arm.com>
948 * config/aarch64/aarch64.c (aarch64_simd_gen_const_vector_dup):
949 Change int to HOST_WIDE_INT.
950 * config/aarch64/aarch64-protos.h
951 (aarch64_simd_gen_const_vector_dup): Likewise.
952 * config/aarch64/aarch64-simd.md: Add copysign<mode>3.
954 2017-01-19 David Malcolm <dmalcolm@redhat.com>
956 * langhooks-def.h (lhd_type_for_size): New decl.
957 (LANG_HOOKS_TYPE_FOR_SIZE): Define as lhd_type_for_size.
958 * langhooks.c (lhd_type_for_size): New function, taken from
961 2017-01-19 Pat Haugen <pthaugen@us.ibm.com>
963 * config/rs6000/power9.md (power9-alu): Remove 'cmp' type and add
964 define_bypass for CR latency.
965 (power9-cracked-alu): Update bypass latency and remove power9-branch.
966 (power9-alu2): Add define_bypass for CR latency.
968 (power9-mul): Update insn latency.
969 (power9-mul-compare): Update insn latency, bypass latency and remove
972 2016-01-19 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
974 * config/aarch64/aarch64-protos.h (aarch64_nopcrelative_literal_loads):
976 * config/aarch64/aarch64.md
977 (aarch64_reload_movcp<GPF_TF:mode><P:mode>): Delete reference to
978 aarch64_nopcrelative_literal_loads.
979 (aarch64_reload_movcp<VALL:mode><P:mode>): Likewise.
981 2017-01-19 Chenghua Xu <paul.hua.gm@gmail.com>
983 * config/mips/mips.h (ISA_HAS_FUSED_MADD4): Enable for
985 (ISA_HAS_UNFUSED_MADD4): Exclude TARGET_LOONGSON_3A.
987 2017-01-19 Doug Gilmore <doug.gilmore@imgtec.com>
990 * config.gcc (supported_defaults): Add lxc1-sxc1.
991 (with_lxc1_sxc1): Add validation.
992 (all_defaults): Add lxc1-sxc1.
993 * config/mips/mips.opt (mlxc1-sxc1): New option.
994 * gcc/config/mips/mips.h (OPTION_DEFAULT_SPECS): Add a default for
996 (TARGET_CPU_CPP_BUILTINS): Add builtin_define for
998 (ISA_HAS_LXC1_SXC1): Gate with mips_lxc1_sxc1.
999 * gcc/doc/invoke.texi (-mlxc1-sxc1): Document the new option.
1000 * doc/install.texi (--with-lxc1-sxc1): Document the new option.
1002 2017-01-19 Richard Biener <rguenther@suse.de>
1004 PR tree-optimization/72488
1005 * tree-ssa-sccvn.c (run_scc_vn): When we abort the VN make
1006 sure to restore SSA info.
1007 * tree-ssa.c (verify_ssa): Verify SSA info is not shared.
1009 2017-01-19 Richard Earnshaw <rearnsha@arm.com>
1011 PR rtl-optimization/79121
1012 * expr.c (expand_expr_real_2, case LSHIFT_EXPR): Look at the signedness
1013 of the inner type when shifting an extended value.
1015 2017-01-17 Jan Hubicka <hubicka@ucw.cz>
1018 * symtab.c (symtab_node::equal_address_to): Fix comparing of
1019 interposable aliases.
1021 2017-01-18 Peter Bergner <bergner@vnet.ibm.com>
1024 * config/rs6000/spe.md (mov_si<mode>_e500_subreg0): Fix constraints.
1025 Use the evmergelohi instruction.
1026 (mov_si<mode>_e500_subreg4_2_le): Likewise.
1027 (mov_sitf_e500_subreg8_2_be): Likewise.
1028 (mov_sitf_e500_subreg12_2_le): Likewise.
1029 (mov_si<mode>_e500_subreg0_2_le): Fix constraints.
1030 (mov_si<mode>_e500_subreg4_2_be): Likewise.
1031 (mov_sitf_e500_subreg8_2_le): Likewise.
1032 (mov_sitf_e500_subreg12_2_be): Likewise.
1034 2017-01-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1036 * config/rs6000/altivec.md (altivec_vbpermq): Change "type"
1037 attribute from vecsimple to vecperm.
1038 (altivec_vbpermq2): Likewise.
1040 2017-01-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1043 * config/rs6000/altivec.h: Fix typo of vec_cntlz to vec_cnttz.
1045 2017-01-18 Aaron Sawdey <acsawdey@linux.vnet.ibm.com>
1046 * config/rs6000/rs6000-protos.h (expand_strn_compare): Add arg.
1047 * config/rs6000/rs6000.c (expand_strn_compare): Add ability to expand
1048 strcmp. Fix bug where comparison didn't stop with zero byte. Fix
1049 case where N arg is SIZE_MAX.
1050 * config/rs6000/rs6000.md (cmpstrnsi): Args to expand_strn_compare.
1051 (cmpstrsi): Add pattern.
1053 2017-01-18 Michael Meissner <meissner@linux.vnet.ibm.com>
1055 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
1056 __builtin_vec_revb builtins.
1057 * config/rs6000/rs6000-builtins.def (P9V_BUILTIN_XXBRQ_V16QI): Add
1058 built-in functions to support generation of the ISA 3.0 XXBR<x>
1059 vector byte reverse instructions.
1060 (P9V_BUILTIN_XXBRQ_V1TI): Likewise.
1061 (P9V_BUILTIN_XXBRD_V2DI): Likewise.
1062 (P9V_BUILTIN_XXBRD_V2DF): Likewise.
1063 (P9V_BUILTIN_XXBGW_V4SI): Likewise.
1064 (P9V_BUILTIN_XXBGW_V4SF): Likewise.
1065 (P9V_BUILTIN_XXBGH_V8HI): Likewise.
1066 (P9V_BUILTIN_VEC_REVB): Likewise.
1067 * config/rs6000/vsx.md (p9_xxbrq_v1ti): New insns/expanders to
1068 generate the ISA 3.0 XXBR<x> vector byte reverse instructions.
1069 (p9_xxbrq_v16qi): Likewise.
1070 (p9_xxbrd_<mode>, VSX_D iterator): Likewise.
1071 (p9_xxbrw_<mode>, VSX_W iterator): Likewise.
1072 (p9_xxbrh_v8hi): Likewise.
1073 * config/rs6000/altivec.h (vec_revb): Define if ISA 3.0.
1074 * doc/extend.texi (RS/6000 Altivec Built-ins): Document the
1075 vec_revb built-in functions.
1077 2017-01-18 Uros Bizjak <ubizjak@gmail.com>
1079 PR rtl-optimization/78952
1080 * config/i386/i386.md (any_extract): New code iterator.
1081 (*insvqi_2): Use any_extract for source operand.
1082 (*insvqi_3): Use any_shiftrt for source operand.
1084 2017-01-18 Wilco Dijkstra <wdijkstr@arm.com>
1086 * config/aarch64/aarch64.c (aarch64_sched_adjust_priority)
1088 (TARGET_SCHED_ADJUST_PRIORITY): Define target hook.
1090 2017-01-18 Matthias Klose <doko@ubuntu.com>
1092 * doc/install.texi: Allow default for --with-target-bdw-gc-include.
1094 2016-01-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1096 * config/rs6000/altivec.h (vec_bperm): Change #define.
1097 * config/rs6000/altivec.md (UNSPEC_VBPERMD): New enum constant.
1098 (altivec_vbpermq2): New define_insn.
1099 (altivec_vbpermd): Likewise.
1100 * config/rs6000/rs6000-builtin.def (VBPERMQ2): New monomorphic
1102 (VBPERMD): Likewise.
1103 (VBPERM): New polymorphic function interface.
1104 * config/rs6000/r6000-c.c (altivec_overloaded_builtins_table):
1105 Add entries for P9V_BUILTIN_VEC_VBPERM.
1106 * doc/extend.texi: Add interfaces for vec_bperm.
1108 2017-01-18 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
1110 * config/s390/s390-c.c (s390_expand_overloaded_builtin): Downcase
1111 first letter of error messages.
1112 (s390_resolve_overloaded_builtin): Likewise.
1113 * config/s390/s390.c (s390_expand_builtin): Likewise.
1114 (s390_invalid_arg_for_unprototyped_fn): Likewise.
1115 (s390_valid_target_attribute_inner_p): Likewise.
1116 * config/s390/s390.md ("tabort"): Likewise.
1118 2017-01-18 Toma Tabacu <toma.tabacu@imgtec.com>
1120 * config/mips/mips.h (ISA_HAS_DIV3): Remove unused macro.
1121 (ISA_AVOID_DIV_HILO): New macro.
1122 (ISA_HAS_DIV): Use new ISA_AVOID_DIV_HILO macro.
1123 (ISA_HAS_DDIV): Likewise.
1125 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
1127 * doc/invoke.texi (fabi-version): Correct number of occurrences.
1129 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
1131 * doc/invoke.texi (fabi-version): Spelling fix.
1133 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
1136 * doc/invoke.texi (fabi-version): Mention mangling fix for
1139 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
1142 * doc/invoke.texi (fabi-version): Document discriminator mangling.
1144 2017-01-17 Segher Boessenkool <segher@kernel.crashing.org>
1147 * config/rs6000/rs6000-opts.h (stack_protector_guard): New enum.
1148 * config/rs6000/rs6000.c (rs6000_option_override_internal): Handle
1150 * config/rs6000/rs6000.md (stack_protect_set): Handle the new more
1152 (stack_protect_test): Ditto.
1153 * config/rs6000/rs6000.opt (mstack-protector-guard=,
1154 mstack-protector-guard-reg=, mstack-protector-guard-offset=): New
1156 * doc/invoke.texi (Option Summary) [RS/6000 and PowerPC Options]:
1157 Add -mstack-protector-guard=, -mstack-protector-guard-reg=, and
1158 -mstack-protector-guard-offset=.
1159 (RS/6000 and PowerPC Options): Ditto.
1161 2017-01-17 Uros Bizjak <ubizjak@gmail.com>
1163 * config/i386/i386.h (MASK_CLASS_P): New define.
1164 * config/i386/i386.c (inline_secondary_memory_needed): Ensure that
1165 there are no registers from different register sets also when
1166 mask registers are used. Update function comment.
1167 * config/i386/i386.md (*movsi_internal): Split (*k/*krm) alternative
1168 to (*k/*r) and (*k/*km) alternatives.
1170 2017-01-17 Wilco Dijkstra <wdijkstr@arm.com>
1172 * config/aarch64/aarch64.md (eh_return): Remove pattern and splitter.
1173 * config/aarch64/aarch64.h (AARCH64_EH_STACKADJ_REGNUM): Remove.
1174 (EH_RETURN_HANDLER_RTX): New define.
1175 * config/aarch64/aarch64.c (aarch64_frame_pointer_required):
1176 Force frame pointer in EH return functions.
1177 (aarch64_expand_epilogue): Add barrier for eh_return.
1178 (aarch64_final_eh_return_addr): Remove.
1179 (aarch64_eh_return_handler_rtx): New function.
1180 * config/aarch64/aarch64-protos.h (aarch64_final_eh_return_addr):
1182 (aarch64_eh_return_handler_rtx): New prototype.
1184 2017-01-17 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1186 * config/rs6000/altivec.h (vec_rlmi): New #define.
1187 (vec_vrlnm): Likewise.
1188 (vec_rlnm): Likewise.
1189 * config/rs6000/altivec.md (UNSPEC_VRLMI): New UNSPEC enum value.
1190 (UNSPEC_VRLNM): Likewise.
1191 (VIlong): New mode iterator.
1192 (altivec_vrl<VI_char>mi): New define_insn.
1193 (altivec_vrl<VI_char>nm): Likewise.
1194 * config/rs6000/rs6000-builtin.def (VRLWNM): New monomorphic
1197 (RLNM): New polymorphic function entry.
1198 (VRLWMI): New monomorphic function entry.
1200 (RLMI): New polymorphic function entry.
1201 * config/rs6000/r6000-c.c (altivec_overloaded_builtin_table): Add
1202 new entries for P9V_BUILTIN_VEC_RLMI and P9V_BUILTIN_VEC_RLNM.
1203 * doc/extend.texi: Add description of vec_rlmi, vec_rlnm, and
1206 2017-01-17 Jakub Jelinek <jakub@redhat.com>
1209 * dwarf2out.c (field_byte_offset): Restore the
1210 PCC_BITFIELD_TYPE_MATTERS behavior for INTEGER_CST DECL_FIELD_OFFSET
1211 and DECL_FIELD_BIT_OFFSET. Use fold_build2 instead of build2 + fold.
1212 (analyze_variants_discr, gen_variant_part): Use fold_build2 instead
1215 2017-01-17 Eric Botcazou <ebotcazou@adacore.com>
1218 * config/aarch64/aarch64.c (TARGET_CUSTOM_FUNCTION_DESCRIPTORS): Define
1220 2017-01-17 Jakub Jelinek <jakub@redhat.com>
1223 * dwarf2out.c (add_data_member_location_attribute): For constant
1224 offset bitfield emit for -gdwarf-5 DW_AT_data_bit_offset attribute
1225 instead of DW_AT_data_member_location, DW_AT_bit_offset and
1226 DW_AT_byte_size attributes.
1228 2017-01-17 Eric Botcazou <ebotcazou@adacore.com>
1230 * config/rs6000/rs6000.c (rs6000_emit_move): Also use a TOC reference
1231 after forcing to constant memory when the code model is medium.
1233 2017-01-17 Julia Koval <julia.koval@intel.com>
1236 * config/i386/avx512fintrin.h
1237 (_mm512_i32gather_ps): Change __addr type to void const*.
1238 (_mm512_mask_i32gather_ps): Ditto.
1239 (_mm512_i32gather_pd): Ditto.
1240 (_mm512_mask_i32gather_pd): Ditto.
1241 (_mm512_i64gather_ps): Ditto.
1242 (_mm512_mask_i64gather_ps): Ditto.
1243 (_mm512_i64gather_pd): Ditto.
1244 (_mm512_mask_i64gather_pd): Ditto.
1245 (_mm512_i32gather_epi32): Ditto.
1246 (_mm512_mask_i32gather_epi32): Ditto.
1247 (_mm512_i32gather_epi64): Ditto.
1248 (_mm512_mask_i32gather_epi64): Ditto.
1249 (_mm512_i64gather_epi32): Ditto.
1250 (_mm512_mask_i64gather_epi32): Ditto.
1251 (_mm512_i64gather_epi64): Ditto.
1252 (_mm512_mask_i64gather_epi64): Ditto.
1253 (_mm512_i32scatter_ps): Change __addr type to void*.
1254 (_mm512_mask_i32scatter_ps): Ditto.
1255 (_mm512_i32scatter_pd): Ditto.
1256 (_mm512_mask_i32scatter_pd): Ditto.
1257 (_mm512_i64scatter_ps): Ditto.
1258 (_mm512_mask_i64scatter_ps): Ditto.
1259 (_mm512_i64scatter_pd): Ditto.
1260 (_mm512_mask_i64scatter_pd): Ditto.
1261 (_mm512_i32scatter_epi32): Ditto.
1262 (_mm512_mask_i32scatter_epi32): Ditto.
1263 (_mm512_i32scatter_epi64): Ditto.
1264 (_mm512_mask_i32scatter_epi64): Ditto.
1265 (_mm512_i64scatter_epi32): Ditto.
1266 (_mm512_mask_i64scatter_epi32): Ditto.
1267 (_mm512_i64scatter_epi64): Ditto.
1268 (_mm512_mask_i64scatter_epi64): Ditto.
1269 * config/i386/avx512pfintrin.h
1270 (_mm512_mask_prefetch_i32gather_pd): Change __addr type to void const*.
1271 (_mm512_mask_prefetch_i32gather_ps): Ditto.
1272 (_mm512_mask_prefetch_i64gather_pd): Ditto.
1273 (_mm512_mask_prefetch_i64gather_ps): Ditto.
1274 (_mm512_prefetch_i32scatter_pd): Change __addr type to void*.
1275 (_mm512_prefetch_i32scatter_ps): Ditto.
1276 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
1277 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
1278 (_mm512_prefetch_i64scatter_pd): Ditto.
1279 (_mm512_prefetch_i64scatter_ps): Ditto.
1280 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
1281 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
1282 * config/i386/avx512vlintrin.h
1283 (_mm256_mmask_i32gather_ps): Change __addr type to void const*.
1284 (_mm_mmask_i32gather_ps): Ditto.
1285 (_mm256_mmask_i32gather_pd): Ditto.
1286 (_mm_mmask_i32gather_pd): Ditto.
1287 (_mm256_mmask_i64gather_ps): Ditto.
1288 (_mm_mmask_i64gather_ps): Ditto.
1289 (_mm256_mmask_i64gather_pd): Ditto.
1290 (_mm_mmask_i64gather_pd): Ditto.
1291 (_mm256_mmask_i32gather_epi32): Ditto.
1292 (_mm_mmask_i32gather_epi32): Ditto.
1293 (_mm256_mmask_i32gather_epi64): Ditto.
1294 (_mm_mmask_i32gather_epi64): Ditto.
1295 (_mm256_mmask_i64gather_epi32): Ditto.
1296 (_mm_mmask_i64gather_epi32): Ditto.
1297 (_mm256_mmask_i64gather_epi64): Ditto.
1298 (_mm_mmask_i64gather_epi64): Ditto.
1299 (_mm256_i32scatter_ps): Change __addr type to void*.
1300 (_mm256_mask_i32scatter_ps): Ditto.
1301 (_mm_i32scatter_ps): Ditto.
1302 (_mm_mask_i32scatter_ps): Ditto.
1303 (_mm256_i32scatter_pd): Ditto.
1304 (_mm256_mask_i32scatter_pd): Ditto.
1305 (_mm_i32scatter_pd): Ditto.
1306 (_mm_mask_i32scatter_pd): Ditto.
1307 (_mm256_i64scatter_ps): Ditto.
1308 (_mm256_mask_i64scatter_ps): Ditto.
1309 (_mm_i64scatter_ps): Ditto.
1310 (_mm_mask_i64scatter_ps): Ditto.
1311 (_mm256_i64scatter_pd): Ditto.
1312 (_mm256_mask_i64scatter_pd): Ditto.
1313 (_mm_i64scatter_pd): Ditto.
1314 (_mm_mask_i64scatter_pd): Ditto.
1315 (_mm256_i32scatter_epi32): Ditto.
1316 (_mm256_mask_i32scatter_epi32): Ditto.
1317 (_mm_i32scatter_epi32): Ditto.
1318 (_mm_mask_i32scatter_epi32): Ditto.
1319 (_mm256_i32scatter_epi64): Ditto.
1320 (_mm256_mask_i32scatter_epi64): Ditto.
1321 (_mm_i32scatter_epi64): Ditto.
1322 (_mm_mask_i32scatter_epi64): Ditto.
1323 (_mm256_i64scatter_epi32): Ditto.
1324 (_mm256_mask_i64scatter_epi32): Ditto.
1325 (_mm_i64scatter_epi32): Ditto.
1326 (_mm_mask_i64scatter_epi32): Ditto.
1327 (_mm256_i64scatter_epi64): Ditto.
1328 (_mm256_mask_i64scatter_epi64): Ditto.
1329 (_mm_i64scatter_epi64): Ditto.
1330 (_mm_mask_i64scatter_epi64): Ditto.
1331 * config/i386/i386-builtin-types.def (V16SF_V16SF_PCFLOAT_V16SI_HI_INT)
1332 (V8DF_V8DF_PCDOUBLE_V8SI_QI_INT, V8SF_V8SF_PCFLOAT_V8DI_QI_INT)
1333 (V8DF_V8DF_PCDOUBLE_V8DI_QI_INT, V16SI_V16SI_PCINT_V16SI_HI_INT)
1334 (V8DI_V8DI_PCINT64_V8SI_QI_INT, V8SI_V8SI_PCINT_V8DI_QI_INT)
1335 (V8DI_V8DI_PCINT64_V8DI_QI_INT, V2DF_V2DF_PCDOUBLE_V4SI_QI_INT)
1336 (V4DF_V4DF_PCDOUBLE_V4SI_QI_INT, V2DF_V2DF_PCDOUBLE_V2DI_QI_INT)
1337 (V4DF_V4DF_PCDOUBLE_V4DI_QI_INT, V4SF_V4SF_PCFLOAT_V4SI_QI_INT)
1338 (V8SF_V8SF_PCFLOAT_V8SI_QI_INT, V4SF_V4SF_PCFLOAT_V2DI_QI_INT)
1339 (V4SF_V4SF_PCFLOAT_V4DI_QI_INT, V2DI_V2DI_PCINT64_V4SI_QI_INT)
1340 (V4DI_V4DI_PCINT64_V4SI_QI_INT, V2DI_V2DI_PCINT64_V2DI_QI_INT)
1341 (V4DI_V4DI_PCINT64_V4DI_QI_INT, V4SI_V4SI_PCINT_V4SI_QI_INT)
1342 (V8SI_V8SI_PCINT_V8SI_QI_INT, V4SI_V4SI_PCINT_V2DI_QI_INT)
1343 (V4SI_V4SI_PCINT_V4DI_QI_INT, VOID_PFLOAT_HI_V16SI_V16SF_INT)
1344 (VOID_PFLOAT_QI_V8SI_V8SF_INT, VOID_PFLOAT_QI_V4SI_V4SF_INT)
1345 (VOID_PDOUBLE_QI_V8SI_V8DF_INT, VOID_PDOUBLE_QI_V4SI_V4DF_INT)
1346 (VOID_PDOUBLE_QI_V4SI_V2DF_INT, VOID_PFLOAT_QI_V8DI_V8SF_INT)
1347 (VOID_PFLOAT_QI_V4DI_V4SF_INT, VOID_PFLOAT_QI_V2DI_V4SF_INT)
1348 (VOID_PDOUBLE_QI_V8DI_V8DF_INT, VOID_PDOUBLE_QI_V4DI_V4DF_INT)
1349 (VOID_PDOUBLE_QI_V2DI_V2DF_INT, VOID_PINT_HI_V16SI_V16SI_INT)
1350 (VOID_PINT_QI_V8SI_V8SI_INT, VOID_PINT_QI_V4SI_V4SI_INT)
1351 (VOID_PLONGLONG_QI_V8SI_V8DI_INT, VOID_PLONGLONG_QI_V4SI_V4DI_INT)
1352 (VOID_PLONGLONG_QI_V4SI_V2DI_INT, VOID_PINT_QI_V8DI_V8SI_INT)
1353 (VOID_PINT_QI_V4DI_V4SI_INT, VOID_PINT_QI_V2DI_V4SI_INT)
1354 (VOID_PLONGLONG_QI_V8DI_V8DI_INT, VOID_QI_V8SI_PCINT64_INT_INT)
1355 (VOID_PLONGLONG_QI_V4DI_V4DI_INT, VOID_PLONGLONG_QI_V2DI_V2DI_INT)
1356 (VOID_HI_V16SI_PCINT_INT_INT, VOID_QI_V8DI_PCINT64_INT_INT)
1357 (VOID_QI_V8DI_PCINT_INT_INT): Remove.
1358 (V16SF_V16SF_PCVOID_V16SI_HI_INT, V8DF_V8DF_PCVOID_V8SI_QI_INT)
1359 (V8SF_V8SF_PCVOID_V8DI_QI_INT, V8DF_V8DF_PCVOID_V8DI_QI_INT)
1360 (V16SI_V16SI_PCVOID_V16SI_HI_INT, V8DI_V8DI_PCVOID_V8SI_QI_INT)
1361 (V8SI_V8SI_PCVOID_V8DI_QI_INT, V8DI_V8DI_PCVOID_V8DI_QI_INT)
1362 (VOID_PVOID_HI_V16SI_V16SF_INT, VOID_PVOID_QI_V8SI_V8DF_INT)
1363 (VOID_PVOID_QI_V8DI_V8SF_INT, VOID_PVOID_QI_V8DI_V8DF_INT)
1364 (VOID_PVOID_HI_V16SI_V16SI_INT, VOID_PVOID_QI_V8SI_V8DI_INT)
1365 (VOID_PVOID_QI_V8DI_V8SI_INT, VOID_PVOID_QI_V8DI_V8DI_INT)
1366 (V2DF_V2DF_PCVOID_V4SI_QI_INT, V4DF_V4DF_PCVOID_V4SI_QI_INT)
1367 (V2DF_V2DF_PCVOID_V2DI_QI_INT, V4DF_V4DF_PCVOID_V4DI_QI_INT
1368 (V4SF_V4SF_PCVOID_V4SI_QI_INT, V8SF_V8SF_PCVOID_V8SI_QI_INT)
1369 (V4SF_V4SF_PCVOID_V2DI_QI_INT, V4SF_V4SF_PCVOID_V4DI_QI_INT)
1370 (V2DI_V2DI_PCVOID_V4SI_QI_INT, V4DI_V4DI_PCVOID_V4SI_QI_INT)
1371 (V2DI_V2DI_PCVOID_V2DI_QI_INT, V4DI_V4DI_PCVOID_V4DI_QI_INT)
1372 (V4SI_V4SI_PCVOID_V4SI_QI_INT, V8SI_V8SI_PCVOID_V8SI_QI_INT)
1373 (V4SI_V4SI_PCVOID_V2DI_QI_INT, V4SI_V4SI_PCVOID_V4DI_QI_INT)
1374 (VOID_PVOID_QI_V8SI_V8SF_INT, VOID_PVOID_QI_V4SI_V4SF_INT)
1375 (VOID_PVOID_QI_V4SI_V4DF_INT, VOID_PVOID_QI_V4SI_V2DF_INT)
1376 (VOID_PVOID_QI_V4DI_V4SF_INT, VOID_PVOID_QI_V2DI_V4SF_INT)
1377 (VOID_PVOID_QI_V4DI_V4DF_INT, VOID_PVOID_QI_V2DI_V2DF_INT)
1378 (VOID_PVOID_QI_V8SI_V8SI_INT, VOID_PVOID_QI_V4SI_V4SI_INT)
1379 (VOID_PVOID_QI_V4SI_V4DI_INT, VOID_PVOID_QI_V4SI_V2DI_INT)
1380 (VOID_PVOID_QI_V4DI_V4SI_INT, VOID_PVOID_QI_V2DI_V4SI_INT)
1381 (VOID_PVOID_QI_V4DI_V4DI_INT, VOID_PVOID_QI_V2DI_V2DI_INT)
1382 (VOID_QI_V8SI_PCVOID_INT_INT, VOID_HI_V16SI_PCVOID_INT_INT)
1383 (VOID_QI_V8DI_PCVOID_INT_INT): Add.
1384 * config/i386/i386.c (ix86_init_mmx_sse_builtins): Adjust builtin
1385 definitions accordingly.
1387 2017-01-17 Kito Cheng <kito.cheng@gmail.com>
1388 Kuan-Lin Chen <kuanlinchentw@gmail.com>
1391 * internal-fn.c (expand_mul_overflow): Use convert_modes instead of
1394 2017-01-17 Vladimir Makarov <vmakarov@redhat.com>
1397 * ira-conflicts.c (ira_build_conflicts): Update total conflict
1398 hard regs for inner regno.
1400 2017-01-17 Martin Liska <mliska@suse.cz>
1403 * ipa-polymorphic-call.c (contains_type_p): Fix wrong
1404 assumption and add comment.
1406 2017-01-17 Nathan Sidwell <nathan@acm.org>
1408 * ipa-visibility.c (localize_node): New function, broken out of ...
1409 (function_and_variable_visibility): ... here. Call it.
1411 2017-01-17 Jan Hubicka <hubicka@ucw.cz>
1414 * tree-ssa-threadupdate.c (remove_ctrl_stmt_and_useless_edges):
1415 correctly set frequency of oudgoing edge.
1416 (duplicate_thread_path): Fix profile updating.
1418 2017-01-17 Jakub Jelinek <jakub@redhat.com>
1421 * configure.ac: Add GCC_BASE_VER.
1422 * Makefile.in (version): Use @get_gcc_base_ver@ instead of cat to get
1423 version from BASE-VER file.
1424 (CFLAGS-gcc.o): Add -DBASEVER=$(BASEVER_s).
1425 (gcc.o): Depend on $(BASEVER).
1426 * common.opt (dumpfullversion): New option.
1427 * gcc.c (driver_handle_option): Handle OPT_dumpfullversion.
1428 * doc/invoke.texi: Document -dumpfullversion.
1429 * doc/install.texi: Document --with-gcc-major-version-only.
1430 * configure: Regenerated.
1432 2017-01-17 Richard Biener <rguenther@suse.de>
1434 PR tree-optimization/71433
1435 * tree-vrp.c (register_new_assert_for): Merge same asserts
1436 on all incoming edges.
1437 (process_assert_insertions_for): Handle insertions at the
1440 2017-01-17 Gerald Pfeifer <gerald@pfeifer.com>
1442 * config/i386/cygwin.h (LIBGCJ_SONAME): Remove.
1443 * config/i386/mingw32.h (LIBGCJ_SONAME): Remove.
1445 2017-01-17 Kaz Kojima <kkojima@gcc.gnu.org>
1448 * config/sh/sh.md (cmpeqsi_t+1): Call copy_rtx to avoid invalid
1451 2017-01-17 Alan Modra <amodra@gmail.com>
1454 * config/rs6000/rs6000.md (elf_high, elf_low): Disable when pic.
1455 * config/rs6000/rs6000.c (rs6000_emit_allocate_stack): Don't allow
1456 symbolic stack limit when pic.
1458 2017-01-16 Martin Sebor <msebor@redhat.com>
1460 PR tree-optimization/78608
1461 * gimple-ssa-sprintf.c (tree_digits): Avoid negating TYPE_MIN.
1463 2017-01-16 Jeff Law <law@redhat.com>
1466 2016-12-02 Tadek Kijkowski <tkijkowski@gmail.com>
1467 * Makefile.in (PREPROCESSOR_DEFINES): Add a level of indirection
1468 for several include directories that may be relative to sysroot.
1469 * config/i386/x-mingw32 (gplus_includedir): Define.
1470 (gplus_tool_includedir, gplus_backward_include_dir): Likewise.
1471 (native_system_includedir): Likewise.
1472 * config/i386/mingw32.h (STANDARD_STARTFILE_PREFIX_1): Do not
1473 override if TARGET_SYSTEM_ROOT is defined.
1474 (NATIVE_SYSTEM_HEADER_DIR): Likewise.
1476 PR tree-optimization/79090
1477 PR tree-optimization/33562
1478 PR tree-optimization/61912
1479 PR tree-optimization/77485
1480 * tree-ssa-dse.c (compute_trims): Accept STMT argument. Dump STMT
1481 and computed trims into the dump file.
1483 2017-01-17 Uros Bizjak <ubizjak@gmail.com>
1485 * config/i386/i386.h (LIMIT_RELOAD_CLASS): Remove.
1487 2017-01-16 Jakub Jelinek <jakub@redhat.com>
1490 * gimplify.c (gimplify_init_constructor): If want_value and
1491 object == lhs, unshare lhs to avoid invalid tree sharing. Formatting
1495 * loop-doloop.c (doloop_modify): Call unshare_all_rtl_in_chain on
1496 sequence. Formatting fixes.
1497 (doloop_optimize): Formatting fixes.
1500 * gcc.c (debug_level_greater_than_spec_func): New function.
1501 (static_spec_functions): Add debug-level-gt spec function.
1502 (ASM_DEBUG_SPEC, cpp_options): Use %:debug-level-gt(0) instead of
1504 * config/darwin.h (DSYMUTIL_SPEC, ASM_DEBUG_SPEC): Likewise.
1505 * config/darwin9.h (DSYMUTIL_SPEC, ASM_DEBUG_SPEC): Likewise.
1506 * common.opt (g, gcoff, gdwarf, gdwarf-, ggdb, gno-pubnames,
1507 gpubnames, ggnu-pubnames, gno-record-gcc-switches,
1508 grecord-gcc-switches, gno-strict-dwarf, gstrict-dwarf, gstabs,
1509 gstabs+, gtoggle, gvms, gxcoff, gxcoff+): Add Driver flag.
1511 2017-01-16 Uros Bizjak <ubizjak@gmail.com>
1513 * config/i386/i386.h (HARD_REGNO_CALLER_SAVE_MODE): Apply HImode and
1514 QImode fixups to general and mask registers only.
1516 2017-01-16 Carl Love <cel@us.ibm.com>
1518 * config/rs6000/rs6000-c (altivec_overloaded_builtins): Add support
1519 for built-in functions
1520 vector signed char vec_nabs (vector signed char)
1521 vector signed short vec_nabs (vector signed short)
1522 vector signed int vec_nabs (vector signed int)
1523 vector signed long long vec_nabs (vector signed long long)
1524 vector float vec_nabs (vector float)
1525 vector double vec_nabs (vector double)
1526 * config/rs6000/rs6000-builtin.def: Add definitions for NABS functions
1528 * config/rs6000/altivec.md: New define_expand nabs<mode>2 types
1529 * config/rs6000/altivec.h: New define for vec_nabs built-in function.
1530 * doc/extend.texi: Update the documentation file for the new built-in
1533 2017-01-16 Martin Sebor <msebor@redhat.com>
1535 * gimple-ssa-sprintf.c (format_directive): Correct a typo in a warning
1538 2017-01-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1540 * config/rs6000/rs6000.c (rtx_is_swappable_p): Change
1541 UNSPEC_VSX__XXSPLTD to require special splat handling.
1543 2017-01-16 David Malcolm <dmalcolm@redhat.com>
1546 * system.h: Poison strndup.
1548 2017-01-16 Alan Modra <amodra@gmail.com>
1551 * config/rs6000/rs6000.c (rs6000_legitimate_combined_insn): Don't
1554 2017-01-16 Georg-Johann Lay <avr@gjlay.de>
1556 * config/avr/avr.h (BRANCH_COST) [reload_completed]: Increase by 4.
1558 2017-01-15 Uros Bizjak <ubizjak@gmail.com>
1560 * config/i386/i386.c (ix86_legitimate_combined_insn): Do not
1561 call recog here. Assert that INSN_CODE (insn) is non-negative.
1563 2017-01-15 Segher Boessenkool <segher@kernel.crashing.org>
1566 * cfgrtl.c (rtl_split_edge): Also patch jump insns that jump to the
1568 * haifa-sched.c (dump_insn_stream): Don't crash if there is a label
1569 in the currently scheduled RTL fragment.
1571 2017-01-15 Segher Boessenkool <segher@kernel.crashing.org>
1573 PR rtl-optimization/78751
1574 * ifcvt.c (find_cond_trap): If we generated a non-existing insn,
1577 2017-01-14 Jeff Law <law@redhat.com>
1579 PR tree-optimization/79090
1580 * tree-ssa-dse.c (valid_ao_ref_for_dse): Reject zero length and
1581 variable length stores.
1582 (compute_trims): Delete dead assignment to *trim_tail.
1583 (dse_dom_walker::dse_optimize_stmt): Optimize mem* calls with
1586 2017-01-14 Bernd Schmidt <bschmidt@redhat.com>
1588 PR rtl-optimization/78626
1589 PR rtl-optimization/78727
1590 * cprop.c (one_cprop_pass): Collect unconditional traps in the middle
1591 of a block, and split such blocks after everything else is finished.
1593 2017-01-14 Alan Modra <amodra@gmail.com>
1596 * combine.c (recog_for_combine_1): Set INSN_CODE before calling
1597 target legitimate_combined_insn.
1598 * config/rs6000/rs6000.c (TARGET_LEGITIMATE_COMBINED_INSN): Define.
1599 (rs6000_legitimate_combined_insn): New function.
1600 * config/rs6000/rs6000.md (UNSPEC_DOLOOP): Delete, and remove
1602 (ctr<mode>_internal3): Rename from *ctr<mode>_internal5.
1603 (ctr<mode>_internal4): Rename from *ctr<mode>_internal6.
1604 (ctr<mode>_internal1, ctr<mode>_internal2): Remove '*' from name.
1606 2017-01-14 Gerald Pfeifer <gerald@pfeifer.com>
1608 * doc/frontends.texi (G++ and GCC): Remove references to Java.
1610 2017-01-13 Jeff Law <law@redhat.com>
1612 PR tree-optimization/33562
1613 PR tree-optimization/61912
1614 PR tree-optimization/77485
1615 * tree-ssa-dse.c (delete_dead_call): Accept gsi rather than
1617 (delete_dead_assignment): Likewise.
1618 (dse_dom_walker::dse_optimize_stmt): Pass in the gsi rather than
1619 statement to delete_dead_call and delete_dead_assignment.
1621 2017-01-13 David Malcolm <dmalcolm@redhat.com>
1624 * substring-locations.c (format_warning_va): Strengthen case 1 so
1625 that both endpoints of the substring must be within the format
1626 range for just the substring to be printed.
1628 2017-01-13 Uros Bizjak <ubizjak@gmail.com>
1630 * config/i386/i386.opt (msgx): Use ix86_isa_flags2 variable.
1631 * config/i386/i386.c (ix86_target_string): Add missing options
1632 to isa_opts and reorder options by implied ISAs. Rename isa_opts2 to
1633 isa2_opts, ix86_flag_opts to flag2_opts, ix86_target_other to
1634 flags_other and ix86_target_other to flags2_other. Display unknown
1636 (ix86_valid_target_attribute_inner_p): Add missing options and
1637 reorder options by implied ISAs, as in ix86_target_string.
1639 2017-01-13 Richard Sandiford <richard.sandiford@arm.com>
1641 * hash-table.h (hash_table::too_empty_p): New function.
1642 (hash_table::expand): Use it.
1643 (hash_table::traverse): Likewise.
1644 (hash_table::empty_slot): Use sizeof (value_type) instead of
1645 sizeof (PTR) to convert bytes to elements. Shrink the table
1646 if the current size is excessive for the current number of
1649 2017-01-13 Richard Sandiford <richard.sandiford@arm.com>
1651 * ira-costs.c (record_reg_classes): Break from the inner loop
1652 early once alt_fail is known to be true. Update outer loop
1653 handling accordingly.
1655 2017-01-13 Jeff Law <law@redhat.com>
1657 * tree-ssa-dse.c (decrement_count): New function.
1658 (increment_start_addr, maybe_trim_memstar_call): Likewise.
1659 (dse_dom_walker::optimize_stmt): Call maybe_trim_memstar_call directly
1660 when we know the partially dead statement is a mem* function.
1662 PR tree-optimization/61912
1663 PR tree-optimization/77485
1664 * tree-ssa-dse.c: Include expr.h.
1665 (maybe_trim_constructor_store): New function.
1666 (maybe_trim_partially_dead_store): Call maybe_trim_constructor_store.
1668 PR tree-optimization/33562
1669 PR tree-optimization/61912
1670 PR tree-optimization/77485
1671 * doc/invoke.texi: Document new dse-max-object-size param.
1672 * params.def (PARM_DSE_MAX_OBJECT_SIZE): New PARAM.
1673 * tree-ssa-dse.c: Include params.h.
1674 (dse_store_status): New enum.
1675 (initialize_ao_ref_for_dse): New, partially extracted from
1677 (valid_ao_ref_for_dse, normalize_ref): New.
1678 (setup_live_bytes_from_ref, compute_trims): Likewise.
1679 (clear_bytes_written_by, maybe_trim_complex_store): Likewise.
1680 (maybe_trim_partially_dead_store): Likewise.
1681 (maybe_trim_complex_store): Likewise.
1682 (dse_classify_store): Renamed from dse_possibly_dead_store_p.
1683 Track what bytes live from the original store. Return tri-state
1684 for dead, partially dead or live.
1685 (dse_dom_walker): Add constructor, destructor and new private members.
1686 (delete_dead_call, delete_dead_assignment): New extracted from
1688 (dse_optimize_stmt): Make a member of dse_dom_walker.
1689 Use initialize_ao_ref_for_dse.
1691 PR tree-optimization/33562
1692 PR tree-optimization/61912
1693 PR tree-optimization/77485
1694 * sbitmap.h (bitmap_count_bits): Prototype.
1695 (bitmap_clear_range, bitmap_set_range): Likewise.
1696 * sbitmap.c (bitmap_clear_range): New function.
1697 (bitmap_set_range, sbitmap_popcount, bitmap_count_bits): Likewise.
1699 2017-01-13 Martin Liska <mliska@suse.cz>
1702 * function.c (set_cfun): Add new argument force.
1703 * function.h (set_cfun): Likewise.
1704 * ipa-inline-transform.c (inline_call): Use the function when
1705 strict alising from is dropped for function we inline to.
1707 2017-01-13 Richard Biener <rguenther@suse.de>
1709 * tree-pretty-print.c (dump_generic_node): Fix inverted condition
1710 for dumping GIMPLE INTEGER_CSTs.
1712 2017-01-13 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1714 * config/sol2.h (TARGET_OS_CPP_BUILTINS): Define __STDC_VERSION__
1715 to 201112L since C++17.
1717 2017-01-13 Maxim Ostapenko <m.ostapenko@samsung.com>
1720 * asan.c (asan_needs_odr_indicator_p): Don't emit ODR indicators
1721 if -fsanitize=kernel-address is present.
1723 2017-01-13 Richard Biener <rguenther@suse.de>
1725 * tree-pretty-print.c (dump_generic_node): Dump INTEGER_CSTs
1726 as _Literal ( type ) number in case usual suffixes do not
1727 preserve all information.
1729 2017-01-13 Richard Biener <rguenther@suse.de>
1731 PR tree-optimization/77283
1732 * gimple-ssa-split-paths.c: Include gimple-ssa.h, tree-phinodes.h
1733 and ssa-iterators.h.
1734 (is_feasible_trace): Implement a cost model based on joiner
1737 2017-01-12 Michael Meissner <meissner@linux.vnet.ibm.com>
1740 * config/rs6000/rs6000.md (FP_ISA3): Do not optimize converting
1741 char or short to __float128/_Float128 directly.
1743 2017-01-12 Martin Sebor <msebor@redhat.com>
1745 to -Wformat-overflow.
1746 * gimple-ssa-sprintf.c (pass_sprintf_length::gate): Adjust.
1747 (min_bytes_remaining): Same.
1748 (get_string_length): Same.
1749 (format_string): Same.
1750 (format_directive): Same.
1752 (pass_sprintf_length::handle_gimple_call): Same.
1754 2017-01-12 Jakub Jelinek <jakub@redhat.com>
1756 * gimple-ssa-sprintf.c (try_substitute_return_value): Remove
1757 info.nowrite calls with no lhs that can't throw. Return bool
1758 whether gsi_remove has been called or not.
1759 (pass_sprintf_length::handle_gimple_call): Return bool whether
1760 try_substitute_return_value called gsi_remove. Formatting fix.
1761 (pass_sprintf_length::execute): Don't use gsi_remove if
1762 handle_gimple_call returned true.
1765 * cfgrtl.c (rtl_tidy_fallthru_edge): For any_uncondjump_p that can't
1766 be removed due to side-effects, don't remove following barrier nor
1767 turn the successor edge into fallthru edge.
1769 2017-01-12 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1772 * config/rs6000/rs6000.c (insn_is_swappable_p): Mark
1773 element-reversing loads and stores as not swappable.
1775 2017-01-12 Nathan Sidwell <nathan@acm.org>
1776 Nicolai Stange <nicstange@gmail.com>
1778 * combine.c (try_combine): Don't ignore result of overlap checking
1779 loop. Combine overlap & asm check into single loop.
1781 2017-01-12 Richard Biener <rguenther@suse.de>
1783 * tree-pretty-print.c (dump_generic_node): Provide -gimple
1784 variant for MEM_REF. Sanitize INTEGER_CST for -gimple.
1786 2017-01-12 Richard Biener <rguenther@suse.de>
1788 * tree.c (initialize_tree_contains_struct): Make TS_OPTIMIZATION
1789 and TS_TARGET_OPTION directly derive from TS_BASE.
1790 * tree-core.h (tree_optimization_option): Derive from tree_base.
1791 (tree_target_option): Likewise.
1793 2017-01-11 Uros Bizjak <ubizjak@gmail.com>
1795 * config/i386/i386.c (memory_address_length): Increase len
1796 only when rip_relative_addr_p returns false.
1798 2017-01-11 Julia Koval <julia.koval@intel.com>
1800 * common/config/i386/i386-common.c (OPTION_MASK_ISA_SGX_UNSET): New.
1801 (OPTION_MASK_ISA_SGX_SET): New.
1802 (ix86_handle_option): Handle OPT_msgx.
1803 * config.gcc: Added sgxintrin.h.
1804 * config/i386/driver-i386.c (host_detect_local_cpu): Detect sgx.
1805 * config/i386/i386-c.c (ix86_target_macros_internal): Define __SGX__.
1806 * config/i386/i386.c (ix86_target_string): Add -msgx.
1808 (ix86_option_override_internal): Handle new options.
1809 (ix86_valid_target_attribute_inner_p): Add sgx.
1810 * config/i386/i386.h (TARGET_SGX, TARGET_SGX_P): New.
1811 * config/i386/i386.opt: Add msgx.
1812 * config/i386/sgxintrin.h: New file.
1813 * config/i386/x86intrin.h: Add sgxintrin.h.
1815 2017-01-11 Jakub Jelinek <jakub@redhat.com>
1818 * fold-const.c (maybe_nonzero_address): Return 1 for function
1820 (tree_single_nonzero_warnv_p): Don't handle function local objects
1824 * gcc.c (default_compilers): Don't add -o %g.s for -S -save-temps
1827 2017-01-11 David Malcolm <dmalcolm@redhat.com>
1830 * opts.c: Include "spellcheck.h"
1831 (struct string_fragment): New struct.
1832 (struct edit_distance_traits<const string_fragment &>): New
1834 (get_closest_sanitizer_option): New function.
1835 (parse_sanitizer_options): Offer suggestions for unrecognized arguments.
1837 2017-01-11 Jakub Jelinek <jakub@redhat.com>
1839 * dwarf2out.c (DWARF_COMPILE_UNIT_HEADER_SIZE): For DWARF5 decrease
1841 (DWARF_COMDAT_TYPE_UNIT_HEADER_SIZE): Always
1842 DWARF_COMPILE_UNIT_HEADER_SIZE plus 12.
1843 (DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE): Define.
1844 (calc_base_type_die_sizes): Use DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE
1845 for initial die_offset if dwarf_split_debug_info.
1846 (output_comp_unit): Use DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE for
1847 initial next_die_offset if dwo_id is non-NULL. Don't emit padding
1849 (output_skeleton_debug_sections): Formatting fix. Use
1850 DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE instead of
1851 DWARF_COMPILE_UNIT_HEADER_SIZE. Don't emit padding.
1853 2017-01-11 Wilco Dijkstra <wdijkstr@arm.com>
1855 * config/arm/cortex-a53.md: Add bypasses for
1857 (cortex_a53_r2f): Only use for transfers.
1858 (cortex_a53_f2r): Likewise.
1859 (cortex_a53_r2f_cvt): Add reservation for conversions.
1860 (cortex_a53_f2r_cvt): Likewise.
1862 2017-01-11 Tamar Christina <tamar.christina@arm.com>
1864 * config/arm/arm_neon.h: Add __artificial__ and gnu_inline
1865 to all inlined functions, change static to extern.
1867 2017-01-11 Christophe Lyon <christophe.lyon@linaro.org>
1870 * config/arm/arm.c (legitimize_pic_address): Handle reference to
1872 (arm_assemble_integer): Likewise.
1874 2017-01-11 Richard Earnshaw <rearnsha@arm.com>
1876 * config.gcc: Use new awk script to check CPU, FPU and architecture
1877 parameters for --with-... options.
1878 * config/arm/parsecpu.awk: New file
1879 * config/arm/arm-cpus.in: New file.
1880 * config/arm/arm-opts.h: Include arm-cpu.h instead of processing .def
1882 * config/arm/arm.c: Include arm-cpu-data.h instead of processing .def
1884 * config/arm/t-arm: Update dependency rules.
1885 * common/config/arm/arm-common.c: Include arm-cpu-cdata.h instead
1886 of processing .def files.
1887 * config/arm/genopt.sh: Deleted.
1888 * config/arm/gentune.sh: Deleted.
1889 * config/arm/arm-cores.def: Deleted.
1890 * config/arm/arm-arches.def: Deleted.
1891 * config/arm/arm-fpus.def: Deleted.
1892 * config/arm/arm-tune.md: Regenerated.
1893 * config/arm/arm-tables.opt: Regenerated.
1894 * config/arm/arm-cpu.h: New generated file.
1895 * config/arm/arm-cpu-data.h: New generated file.
1896 * config/arm/arm-cpu-cdata.h: New generated file.
1898 2017-01-11 Maxim Ostapenko <m.ostapenko@samsung.com>
1901 * lto-cgraph.c (lto_output_varpool_node): Pack dynamically_initialized
1903 (input_varpool_node): Unpack dynamically_initialized bit.
1905 2017-01-11 Eric Botcazou <ebotcazou@adacore.com>
1907 PR rtl-optimization/79032
1908 * lra-constraints.c (simplify_operand_subreg): In the MEM case, test
1909 the alignment of the adjusted memory reference against that of MODE,
1910 instead of the alignment of the original memory reference.
1912 2017-01-11 Martin Jambor <mjambor@suse.cz>
1914 * hsa.c (hsa_callable_function_p): Revert addition of DECL_ARTIFICIAL
1916 * ipa-hsa.c (process_hsa_functions): Only duplicate non-artificial
1917 decorated functions.
1919 2017-01-11 Richard Biener <rguenther@suse.de>
1921 * tree-vrp.c (evrp_dom_walker::before_dom_children): Also
1922 set range/nonnull info for PHI results. Do not set it on
1923 stmts marked for removal.
1925 2017-01-10 Eric Botcazou <ebotcazou@adacore.com>
1927 * expr.c (store_field): In the bitfield case, fetch the return value
1928 from the registers before applying a single big-endian adjustment.
1929 Always do a final load for a BLKmode value not larger than a word.
1931 2017-01-10 David Malcolm <dmalcolm@redhat.com>
1934 * input.c (selftest::test_accessing_ordinary_linemaps): Verify
1935 that we correctly handle column numbers greater than
1936 LINE_MAP_MAX_COLUMN_NUMBER.
1938 2017-01-10 Martin Sebor <msebor@redhat.com>
1941 * gimple-ssa-sprintf.c (get_destination_size): Call
1942 {init,fini}object_sizes.
1943 * tree-object-size.c (addr_object_size): Adjust.
1944 (pass_through_call): Adjust.
1945 (pass_object_sizes::execute): Adjust.
1946 * tree-object-size.h (fini_object_sizes): Declare.
1948 2017-01-10 Martin Sebor <msebor@redhat.com>
1950 PR tree-optimization/78775
1951 * builtins.c (get_size_range): Move...
1952 * calls.c: ...to here.
1953 (alloc_max_size): Accept zero argument.
1954 (operand_signed_p): Remove.
1955 (maybe_warn_alloc_args_overflow): Call get_size_range.
1956 * calls.h (get_size_range): Declare.
1958 2017-01-10 Joe Seymour <joe.s@somniumtech.com>
1960 * config/msp430/driver-msp430.c (msp430_mcu_data): Sync with data
1961 from TI's devices.csv file as of September 2016.
1962 * config/msp430/msp430.c (msp430_mcu_data): Likewise.
1964 2017-01-10 Sandra Loosemore <sandra@codesourcery.com>
1966 * doc/extend.texi: Tweak formatting to fix overfull hbox warnings.
1967 * doc/invoke.texi: Likewise.
1968 * doc/md.texi: Likewise.
1969 * doc/objc.texi: Likewise.
1971 2017-01-10 Joshua Conner <joshconner@google.com>
1973 * config/arm/fuchsia-elf.h: New file.
1974 * config/fuchsia.h: New file.
1975 * config.gcc (*-*-fuchsia*): Set native_system_header_dir.
1976 (aarch64*-*-fuchsia*, arm*-*-fuchsia*, x86_64-*-fuchsia*): Add to
1978 * config.host: (aarch64*-*-fuchsia*, arm*-*-fuchsia*): Add to hosts.
1980 2016-01-10 Richard Biener <rguenther@suse.de>
1982 PR tree-optimization/79034
1983 * tree-call-cdce.c (shrink_wrap_one_built_in_call_with_conds):
1984 Propagate out degenerate PHIs in the joiner.
1986 2017-01-10 Martin Liska <mliska@suse.cz>
1988 * ipa-icf.c (sort_sem_items_by_decl_uid): New function.
1989 (sort_congruence_classes_by_decl_uid): Likewise.
1990 (sort_congruence_class_groups_by_decl_uid): Likewise.
1991 (sem_item_optimizer::merge_classes): Sort class, groups in these
1992 classes and members in the groups by DECL_UID of declarations.
1993 This would make merge operations stable.
1995 2017-01-10 Martin Liska <mliska@suse.cz>
1997 * ipa-icf.c (sem_item_optimizer::sem_item_optimizer): Remove
1998 usage of m_classes_vec.
1999 (sem_item_optimizer::~sem_item_optimizer): Likewise.
2000 (sem_item_optimizer::get_group_by_hash): Likewise.
2001 (sem_item_optimizer::subdivide_classes_by_equality): Likewise.
2002 (sem_item_optimizer::subdivide_classes_by_sensitive_refs): Likewise.
2003 (sem_item_optimizer::verify_classes): Likewise.
2004 (sem_item_optimizer::process_cong_reduction): Likewise.
2005 (sem_item_optimizer::dump_cong_classes): Likewise.
2006 (sem_item_optimizer::merge_classes): Likewise.
2007 * ipa-icf.h (congruence_class_hash): Rename from
2008 congruence_class_group_hash. Remove declaration of m_classes_vec.
2010 2017-01-10 Andrew Senkevich <andrew.senkevich@intel.com>
2012 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512VPOPCNTDQ_SET,
2013 OPTION_MASK_ISA_AVX512VPOPCNTDQ_UNSET): New.
2014 * config.gcc: Add avx512vpopcntdqintrin.h.
2015 * config/i386/avx512vpopcntdqintrin.h: New.
2016 * config/i386/cpuid.h (bit_AVX512VPOPCNTDQ): New.
2017 * config/i386/i386-builtin-types.def: Add new types.
2018 * config/i386/i386-builtin.def (__builtin_ia32_vpopcountd_v16si,
2019 __builtin_ia32_vpopcountd_v16si_mask, __builtin_ia32_vpopcountq_v8di,
2020 __builtin_ia32_vpopcountq_v8di_mask): New.
2021 * config/i386/i386-c.c (ix86_target_macros_internal): Define
2022 __AVX512VPOPCNTDQ__.
2023 * config/i386/i386.c (ix86_target_string): Add -mavx512vpopcntdq.
2024 (PTA_AVX512VPOPCNTDQ): Define.
2025 * config/i386/i386.h (TARGET_AVX512VPOPCNTDQ,
2026 TARGET_AVX512VPOPCNTDQ_P): Define.
2027 * config/i386/i386.opt: Add mavx512vpopcntdq.
2028 * config/i386/immintrin.h: Include avx512vpopcntdqintrin.h.
2029 * config/i386/sse.md (define_insn "vpopcount<mode><mask_name>"): New.
2031 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
2034 * predict.def (PRED_CALL): Set to 67.
2036 2017-01-09 Eric Botcazou <ebotcazou@adacore.com>
2038 * expr.c (store_field): In the bitfield case, if the value comes from
2039 a function call and is of an aggregate type returned in registers, do
2040 not modify the field mode; extract the value in all cases if the mode
2041 is BLKmode and the size is not larger than a word.
2043 2017-01-09 Dominique d'Humieres <dominiq@lps.ens.fr>
2046 * config/i386/cpuid.h: Fix undefined behavior.
2048 2017-01-04 Jeff Law <law@redhat.com>
2050 PR tree-optimization/79007
2051 PR tree-optimization/67955
2052 * tree-ssa-alias.c (same_addr_size_stores_p): Only need to be
2053 conservative for pt.null when flag_non_call_exceptions is on.
2055 2017-01-09 Jakub Jelinek <jakub@redhat.com>
2057 PR translation/79019
2058 PR translation/79020
2059 * params.def (PARAM_INLINE_MIN_SPEEDUP,
2060 PARAM_IPA_CP_SINGLE_CALL_PENALTY,
2061 PARAM_USE_AFTER_SCOPE_DIRECT_EMISSION_THRESHOLD): Fix typos
2063 * config/avr/avr.opt (maccumulate-args): Likewise.
2064 * config/msp430/msp430.opt (mwarn-mcu): Likewise.
2065 * common.opt (freport-bug): Likewise.
2066 * cif-code.def (CIF_FINAL_ERROR): Likewise.
2067 * doc/invoke.texi (ipa-cp-single-call-penalty): Likewise.
2068 * config/s390/s390.c (s390_invalid_binary_op): Fix spelling in
2069 translatable string.
2070 * config/i386/i386.c (function_value_32): Likewise.
2071 * config/nios2/nios2.c (nios2_valid_target_attribute_rec): Likewise.
2072 * config/msp430/msp430.c (msp430_option_override, msp430_attr):
2074 * config/msp430/driver-msp430.c (msp430_select_hwmult_lib): Likewise.
2075 * common/config/msp430/msp430-common.c (msp430_handle_option):
2077 * symtab.c (symtab_node::verify_base): Likewise.
2078 * opts.c (set_debug_level): Likewise.
2079 * tree.c (verify_type_variant): Likewise. Fix typo in comment.
2080 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin): Add
2081 missing whitespace to translatable strings.
2082 * config/avr/avr.md (bswapsi2): Fix typo in comment.
2083 * config/sh/superh.h: Likewise.
2084 * config/i386/xopintrin.h: Likewise.
2085 * config/i386/znver1.md: Likewise.
2086 * config/rs6000/rs6000.c (struct rs6000_opt_mask): Likewise.
2087 * ipa-inline-analysis.c (compute_inline_parameters): Likewise.
2088 * double-int.h (struct double_int): Likewise.
2089 * double-int.c (div_and_round_double): Likewise.
2090 * wide-int.cc: Likewise.
2091 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
2092 * tree-ssa-sccvn.c (vn_reference_lookup_3): Likewise.
2093 * cfgcleanup.c (crossjumps_occured): Renamed to ...
2094 (crossjumps_occurred): ... this.
2095 (try_crossjump_bb, try_head_merge_bb, try_optimize_cfg, cleanup_cfg):
2098 PR tree-optimization/78899
2099 * tree-if-conv.c (version_loop_for_if_conversion): Instead of
2100 returning bool return struct loop *, NULL for failure and the new
2102 (versionable_outer_loop_p): Don't version outer loop if it has
2103 dont_vectorized bit set.
2104 (tree_if_conversion): When versioning outer loop, ensure
2105 tree_if_conversion is performed also on the inner loop of the
2106 non-vectorizable outer loop copy.
2107 * tree-vectorizer.c (set_uid_loop_bbs): Formatting fix. Fold
2108 LOOP_VECTORIZED in inner loop of the scalar outer loop and
2109 prevent vectorization of it.
2110 (vectorize_loops): For outer + inner LOOP_VECTORIZED, ensure
2111 the outer loop vectorization of the non-scalar version is attempted
2112 before vectorization of the inner loop in scalar version. If
2113 outer LOOP_VECTORIZED guarded loop is not vectorized, prevent
2114 vectorization of its inner loop.
2115 * tree-vect-loop-manip.c (rename_variables_in_bb): If outer_loop
2116 has 2 inner loops, rename also on edges from bb whose single pred
2117 is outer_loop->header. Fix typo in function comment.
2119 2017-01-09 Martin Sebor <msebor@redhat.com>
2122 * asan.c (asan_emit_stack_protection): Increase local buffer size
2123 to avoid snprintf truncation warning.
2125 2017-01-09 Andrew Pinski <apinski@cavium.com>
2127 * config/aarch64/aarch64-cores.def: Add thunderx2t99. Change vulcan
2128 to reference thunderx2t99 for the tuning structure
2129 * config/aarch64/aarch64-cost-tables.h (vulcan_extra_costs):
2131 (thunderx2t99_extra_costs): This.
2132 * config/aarch64/aarch64-tune.md: Regenerate.
2133 * config/aarch64/aarch64.c (vulcan_addrcost_table): Rename to ...
2134 (vulcan_addrcost_table): This.
2135 (vulcan_regmove_cost): Rename to ...
2136 (thunderx2t99_regmove_cost): This.
2137 (vulcan_vector_cost): Rename to ...
2138 (thunderx2t99_vector_cost): this.
2139 (vulcan_branch_cost): Rename to ...
2140 (thunderx2t99_branch_cost): This.
2141 (vulcan_tunings): Rename to ...
2142 (thunderx2t99_tunings): This and s/vulcan/thunderx2t99 .
2143 * doc/invoke.texi (AARCH64/mtune): Add thunderx2t99.
2145 2017-01-09 Martin Jambor <mjambor@suse.cz>
2149 * ipa-prop.h (ipa_jump_func): Swap positions of vr_known and m_vr.
2150 * ipa-cp.c (ipa_vr_operation_and_type_effects): New function.
2151 (propagate_vr_accross_jump_function): Use the above function for all
2152 value range computations for pass-through jump functions and type
2153 converasion from explicit value range values.
2154 (ipcp_propagate_stage): Do not attempt to deduce types of formal
2155 parameters from TYPE_ARG_TYPES.
2156 * ipa-prop.c (ipa_write_jump_function): Remove trailing whitespace.
2157 (ipa_write_node_info): Stream type of the actual argument.
2158 (ipa_read_node_info): Likewise. Also remove trailing whitespace.
2160 2017-01-09 Martin Liska <mliska@suse.cz>
2163 * gcc.c (driver_handle_option): Handle OPT_E and set have_E.
2164 (lookup_compiler): Do not show error message with have_E.
2166 2017-01-09 Jakub Jelinek <jakub@redhat.com>
2168 PR tree-optimization/78938
2169 * tree-vect-stmts.c (vectorizable_condition): For non-masked COND_EXPR
2170 where comp_vectype is VECTOR_BOOLEAN_TYPE_P, use
2171 BIT_{NOT,XOR,AND,IOR}_EXPR on the comparison operands instead of
2172 {EQ,NE,GE,GT,LE,LT}_EXPR directly inside of VEC_COND_EXPR. Formatting
2175 2017-01-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2177 * tree-ssa-address.c (gen_addr_rtx): Don't handle index if it
2180 2017-01-09 Richard Biener <rguenther@suse.de>
2182 PR tree-optimization/78997
2183 * tree-vect-slp.c (vect_mask_constant_operand_p): Handle SSA
2184 name condition properly.
2186 2017-01-09 Richard Biener <rguenther@suse.de>
2189 * dwarf2out.c (is_cxx): New overload with context.
2190 (is_naming_typedef_decl): Use it.
2192 2017-01-08 Sandra Loosemore <sandra@codesourcery.com>
2194 * invoke.texi (Option Summary): Correct spacing in option lists
2195 and add line breaks to fix over-long lines.
2197 2017-01-08 Sandra Loosemore <sandra@codesourcery.com>
2201 * extend.texi (Common Variable Attributes): Add xref to GCC
2202 Internals manual to explain mode attribute keywords.
2204 2017-01-08 Sandra Loosemore <sandra@codesourcery.com>
2207 * doc/invoke.texi (Option Summary): Move -pthread to Linker Options
2208 and Preprocessor Options.
2209 (Options for Linking): Document -pthread here....
2210 (RS/6000 and PowerPC Options): ...not here.
2211 (Solaris 2 Options): ...or here.
2212 * doc/cppopts.texi: Document -pthread.
2214 2017-01-08 Martin Sebor <msebor@redhat.com>
2217 * doc/invoke.texi (Warning Options): Document -Wformat-truncation.
2218 * gimple-ssa-sprintf.c (call_info::reval_used, call_info::warnopt):
2219 New member functions.
2220 (format_directive): Used them.
2222 (pass_sprintf_length::handle_gimple_call): Same.
2223 * graphite-sese-to-poly.c (tree_int_to_gmp): Increase buffer size
2224 to avoid truncation for any argument.
2225 (extract_affine_mul): Same.
2226 * tree.c (get_file_function_name): Same.
2228 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
2231 * predict.def (PRED_INDIR_CALL): Set to 86.
2233 2017-01-07 Sandra Loosemore <sandra@codesourcery.com>
2235 PR preprocessor/54124
2236 * doc/cppopts.texi: Reformat -d subtable to list the full name
2237 of the options. Add cross-reference to the docs for the general
2238 compiler -d options.
2239 * doc/invoke.texi (Developer Options): Add cross-reference to the
2240 preprocessor-specific -d option documentation.
2242 2017-01-07 Sandra Loosemore <sandra@codesourcery.com>
2244 PR preprocessor/13498
2245 * doc/cpp.texi (Search Path): Rewrite to remove obsolete and
2246 redudant material, and reflect new command-line options.
2247 (System Headers): Likewise.
2249 2017-01-07 Sandra Loosemore <sandra@codesourcery.com>
2251 * doc/cppdiropts.texi: Merge documentation of -I, -iquote,
2252 -isystem, and -idirafter. Copy-edit.
2253 * doc/cppopts.texi: Copy-edit. Remove contradiction about
2254 default for -ftrack-macro-expansion. Delete obsolete and
2255 badly-formatted implementation details about -fdebug-cpp output.
2256 * doc/cppwarnopts.texi: Copy-edit.
2258 2017-01-07 David Malcolm <dmalcolm@redhat.com>
2261 * input.c (selftest::test_accessing_ordinary_linemaps): Verify
2262 that the transition from a max line width >= 1<<10 to narrower
2263 lines works correctly.
2265 2017-01-07 Alexandre Oliva <aoliva@redhat.com>
2267 * doc/options.texi (PerFunction): New.
2268 * opt-functions.awk (switch_flags): Map both Optimization and
2269 PerFunction to CL_OPTIMIZATION.
2270 * opth-gen.awk: Test for PerFunction flag along with
2272 * optc-save-gen.awk: Likewise. Introduce var_opt_hash and set
2273 it only when the latter is present. Skip those that don't in
2274 the hash function generator.
2275 * common.opt (fvar-tracking): Mark as PerFunction instead of
2277 (fvar-tracking-assignments): Likewise.
2278 (fvar-tracking-assignments-toggle): Likewise.
2279 (fvar-tracking-uninit): Likewise.
2281 2017-01-07 Jakub Jelinek <jakub@redhat.com>
2283 PR translation/79018
2284 * params.def (PARAM_MAX_STORES_TO_MERGE): Add missing space between
2287 2017-01-06 Mikael Pettersson <mikpelinux@gmail.com>
2290 * config/m68k/m68k.opt (LONG_JUMP_TABLE_OFFSETS): New option.
2291 * config/m68k/linux.h (ASM_RETURN_CASE_JUMP): Handle
2292 TARGET_LONG_JUMP_TABLE_OFFSETS.
2293 * config/m68k/m68kelf.h (ASM_RETURN_CASE_JUMP): Likewise.
2294 * config/m68k/netbsd-elf.h (ASM_RETURN_CASE_JUMP): Likewise.
2295 * config/m68k/m68k.h (CASE_VECTOR_MODE): Likewise.
2296 (ASM_OUTPUT_ADDR_DIFF_ELF): Likewise.
2297 * config/m68k/m68k.md (tablejump expander): Likewise.
2298 (*tablejump_pcrel_hi): Renamed from unnamed insn, reject
2299 TARGET_LONG_JUMP_TABLE_OFFSETS.
2300 (*tablejump_pcrel_si): New insn, handle TARGET_LONG_JUMP_TABLE_OFFSETS.
2301 * doc/invoke.texi (M68K options): Add -mlong-jump-table-offsets.
2303 2017-01-06 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
2304 David Holsgrove <david.holsgrove@xilinx.com>
2306 * common/config/microblaze/microblaze-common.c
2307 (TARGET_EXCEPT_UNWIND_INFO): Remove.
2308 * config/microblaze/microblaze-protos.h (microblaze_eh_return):
2310 * config/microblaze/microblaze.c (microblaze_must_save_register)
2311 (microblaze_expand_epilogue, microblaze_return_addr): Handle
2313 (microblaze_eh_return): New function.
2314 * config/microblaze/microblaze.h (RETURN_ADDR_OFFSET)
2315 (EH_RETURN_DATA_REGNO, MB_EH_STACKADJ_REGNUM)
2316 (EH_RETURN_STACKADJ_RTX, ASM_PREFERRED_EH_DATA_FORMAT): New macros.
2317 * config/microblaze/microblaze.md (eh_return): New pattern.
2319 2017-01-06 Jakub Jelinek <jakub@redhat.com>
2321 * system.h (GCC_DIAGNOSTIC_PUSH_IGNORED, GCC_DIAGNOSTIC_POP,
2322 GCC_DIAGNOSTIC_STRINGIFY): Define.
2324 * read-rtl.c (rtx_reader::read_rtx_code): Avoid -Wsign-compare warning.
2326 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2328 * config/arm/arm.md (<mcrr>): New.
2330 * config/arm/arm.c (arm_arch5te): New.
2331 (arm_option_override): Set arm_arch5te.
2332 (arm_coproc_builtin_available): Add support for mcrr, mcrr2, mrrc
2334 * config/arm/arm-builtins.c (MCRR_QUALIFIERS): Define to...
2335 (arm_mcrr_qualifiers): ... this. New.
2336 (MRRC_QUALIFIERS): Define to...
2337 (arm_mrrc_qualifiers): ... this. New.
2338 * config/arm/arm_acle.h (__arm_mcrr, __arm_mcrr2, __arm_mrrc,
2340 * config/arm/arm_acle_builtins.def (mcrr, mcrr2, mrrc, mrrc2): New.
2341 * config/arm/iterators.md (MCRRI, mcrr, MCRR): New.
2342 (MRRCI, mrrc, MRRC): New.
2343 * config/arm/unspecs.md (VUNSPEC_MCRR, VUNSPEC_MCRR2, VUNSPEC_MRRC,
2344 VUNSPEC_MRRC2): New.
2346 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2348 * config/arm/arm.md (<mcr>): New.
2350 * config/arm/arm.c (arm_coproc_builtin_available): Add
2351 support for mcr, mrc, mcr2 and mrc2.
2352 * config/arm/arm-builtins.c (MCR_QUALIFIERS): Define to...
2353 (arm_mcr_qualifiers): ... this. New.
2354 (MRC_QUALIFIERS): Define to ...
2355 (arm_mrc_qualifiers): ... this. New.
2356 (MCR_QUALIFIERS): Define to ...
2357 (arm_mcr_qualifiers): ... this. New.
2358 * config/arm/arm_acle.h (__arm_mcr, __arm_mrc, __arm_mcr2,
2360 * config/arm/arm_acle_builtins.def (mcr, mcr2, mrc, mrc2): New.
2361 * config/arm/iterators.md (MCRI, mcr, MCR, MRCI, mrc, MRC): New.
2362 * config/arm/unspecs.md (VUNSPEC_MCR, VUNSPEC_MCR2, VUNSPEC_MRC,
2365 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2367 * config/arm/arm.md (*ldc): New.
2371 * config/arm/arm.c (arm_coproc_builtin_available): Add
2372 support for ldc,ldcl,stc,stcl,ldc2,ldc2l,stc2 and stc2l.
2373 (arm_coproc_ldc_stc_legitimate_address): New.
2374 * config/arm/arm-builtins.c (arm_type_qualifiers): Add
2375 'qualifier_const_pointer'.
2376 (LDC_QUALIFIERS): Define to...
2377 (arm_ldc_qualifiers): ... this. New.
2378 (STC_QUALIFIERS): Define to...
2379 (arm_stc_qualifiers): ... this. New.
2380 * config/arm/arm-protos.h
2381 (arm_coproc_ldc_stc_legitimate_address): New.
2382 * config/arm/arm_acle.h (__arm_ldc, __arm_ldcl, __arm_stc,
2383 __arm_stcl, __arm_ldc2, __arm_ldc2l, __arm_stc2, __arm_stc2l): New.
2384 * config/arm/arm_acle_builtins.def (ldc, ldc2, ldcl, ldc2l, stc,
2385 stc2, stcl, stc2l): New.
2386 * config/arm/constraints.md (Uz): New.
2387 * config/arm/iterators.md (LDCI, STCI, ldc, stc, LDC STC): New.
2388 * config/arm/unspecs.md (VUNSPEC_LDC, VUNSPEC_LDC2, VUNSPEC_LDCL,
2389 VUNSPEC_LDC2L, VUNSPEC_STC, VUNSPEC_STC2, VUNSPEC_STCL,
2390 VUNSPEC_STC2L): New.
2392 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2394 * config/arm/arm.md (<cdp>): New.
2395 * config/arm/arm.c (neon_const_bounds): Rename this ...
2396 (arm_const_bounds): ... this.
2397 (arm_coproc_builtin_available): New.
2398 * config/arm/arm-builtins.c (SIMD_MAX_BUILTIN_ARGS): Increase.
2399 (arm_type_qualifiers): Add 'qualifier_unsigned_immediate'.
2400 (CDP_QUALIFIERS): Define to...
2401 (arm_cdp_qualifiers): ... this. New.
2403 (arm_expand_builtin_args): Add case for 6 arguments.
2404 * config/arm/arm-protos.h (neon_const_bounds): Rename this ...
2405 (arm_const_bounds): ... this.
2406 (arm_coproc_builtin_available): New.
2407 * config/arm/arm_acle.h (__arm_cdp): New.
2409 * config/arm/arm_acle_builtins.def (cdp): New.
2411 * config/arm/iterators.md (CDPI,CDP,cdp): New.
2412 * config/arm/neon.md: Rename all 'neon_const_bounds' to
2414 * config/arm/types.md (coproc): New.
2415 * config/arm/unspecs.md (VUNSPEC_CDP, VUNSPEC_CDP2): New.
2416 * gcc/doc/extend.texi (ACLE): Add a mention of Coprocessor intrinsics.
2417 * gcc/doc/sourcebuild.texi (arm_coproc1_ok, arm_coproc2_ok,
2418 arm_coproc3_ok, arm_coproc4_ok): Document new effective targets.
2420 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2422 * config/arm/arm-builtins.c (arm_unsigned_binop_qualifiers): New.
2423 (UBINOP_QUALIFIERS): New.
2425 (acle_builtin_data): New. Change comment.
2426 (arm_builtins): Remove ARM_BUILTIN_CRC32B, ARM_BUILTIN_CRC32H,
2427 ARM_BUILTIN_CRC32W, ARM_BUILTIN_CRC32CB, ARM_BUILTIN_CRC32CH,
2428 ARM_BUILTIN_CRC32CW. Add ARM_BUILTIN_ACLE_BASE and include
2429 arm_acle_builtins.def.
2430 (ARM_BUILTIN_ACLE_PATTERN_START): Define.
2431 (arm_init_acle_builtins): New.
2432 (CRC32_BUILTIN): Remove.
2433 (bdesc_2arg): Remove entries for crc32b, crc32h, crc32w,
2434 crc32cb, crc32ch and crc32cw.
2435 (arm_init_crc32_builtins): Remove.
2436 (arm_init_builtins): Use arm_init_acle_builtins rather
2437 than arm_init_crc32_builtins.
2438 (arm_expand_acle_builtin): New.
2439 (arm_expand_builtin): Use 'arm_expand_acle_builtin'.
2440 * config/arm/arm_acle_builtins.def: New.
2442 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2444 * config/arm/arm-builtins.c (neon_builtin_datum): Rename to ..
2445 (arm_builtin_datum): ... this.
2446 (arm_init_neon_builtin): Rename to ...
2447 (arm_init_builtin): ... this. Add a new parameters PREFIX
2448 and USE_SIG_IN_NAME.
2449 (arm_init_neon_builtins): Replace 'arm_init_neon_builtin' with
2450 'arm_init_builtin'. Replace type 'neon_builtin_datum' with
2451 'arm_builtin_datum'.
2452 (arm_init_vfp_builtins): Likewise.
2453 (builtin_arg): Rename enum's replacing 'NEON_ARG' with
2454 'ARG_BUILTIN' and add a 'ARG_BUILTIN_NEON_MEMORY.
2455 (arm_expand_neon_args): Rename to ...
2456 (arm_expand_builtin_args): ... this. Rename builtin_arg
2457 enum values and differentiate between ARG_BUILTIN_MEMORY
2458 and ARG_BUILTIN_NEON_MEMORY.
2459 (arm_expand_neon_builtin_1): Rename to ...
2460 (arm_expand_builtin_1): ... this. Rename builtin_arg enum
2461 values, arm_expand_builtin_args and add bool parameter NEON.
2462 (arm_expand_neon_builtin): Use arm_expand_builtin_1.
2463 (arm_expand_vfp_builtin): Likewise.
2464 (NEON_MAX_BUILTIN_ARGS): Remove, it was unused.
2466 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
2469 * predict.def (PRED_POLYMORPHIC_CALL): Set to 59.
2470 * predict.c (tree_estimate_probability_bb): Reverse direction of
2471 polymorphic call predictor.
2473 2017-01-06 David Malcolm <dmalcolm@redhat.com>
2475 * passes.c (execute_one_pass): Split out pass-skipping logic into...
2476 (determine_pass_name_match): ...this new function and...
2477 (should_skip_pass_p): ...this new function.
2479 2017-01-06 Nathan Sidwell <nathan@acm.org>
2481 * ipa-visibility.c (function_and_variable_visibility): Reformat
2482 comments and long lines. Remove extrneous if.
2483 * symtab.c (symtab_node::make_decl_local): Fix code format.
2484 (symtab_node::set_section_for_node): Fix comment typo.
2486 2017-01-06 Martin Liska <mliska@suse.cz>
2489 * lra-constraints.c: Rename invariant to lra_invariant.
2490 * predict.c (set_even_probabilities): Initialize e to NULL.
2492 2017-01-05 Martin Sebor <msebor@redhat.com>
2494 PR tree-optimization/78910
2495 * gimple-ssa-sprintf.c (tree_digits): Add an argument.
2496 (format_integer): Correct off-by-one error in the handling
2497 of precision with negative numbers in signed conversions..
2499 2017-01-05 Eric Botcazou <ebotcazou@adacore.com>
2501 * doc/invoke.texi (C Dialect Options): Adjust -fsso-struct entry.
2503 2017-01-05 Jakub Jelinek <jakub@redhat.com>
2505 PR tree-optimization/71016
2506 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Pass cond_stmt to
2507 factor_out_conditional_conversion. Formatting fix.
2508 (factor_out_conditional_conversion): Add cond_stmt argument.
2509 If arg1 is INTEGER_CST, punt if new_arg0 is not any operand of
2510 cond_stmt and if arg0_def_stmt is not the only stmt in its bb.
2513 2017-01-05 David Malcolm <dmalcolm@redhat.com>
2515 * Makefile.in (OBJS): Add read-md.o, read-rtl.o,
2516 read-rtl-function.o, and selftest-rtl.o.
2517 * config/aarch64/aarch64.c: Include selftest.h and selftest-rtl.h.
2518 (selftest::aarch64_test_loading_full_dump): New function.
2519 (selftest::aarch64_run_selftests): New function.
2520 (TARGET_RUN_TARGET_SELFTESTS): Wire it up to
2521 selftest::aarch64_run_selftests.
2522 * config/i386/i386.c
2523 (selftest::ix86_test_loading_dump_fragment_1): New function.
2524 (selftest::ix86_test_loading_call_insn): New function.
2525 (selftest::ix86_test_loading_full_dump): New function.
2526 (selftest::ix86_test_loading_unspec): New function.
2527 (selftest::ix86_run_selftests): Call the new functions.
2528 * emit-rtl.c (maybe_set_max_label_num): New function.
2529 * emit-rtl.h (maybe_set_max_label_num): New decl.
2530 * function.c (instantiate_decls): Guard call to
2531 instantiate_decls_1 with if (DECL_INITIAL (fndecl)).
2532 * function-tests.c (selftest::verify_three_block_rtl_cfg): Remove
2534 * gensupport.c (gen_reader::gen_reader): Pass "false"
2535 for new "compact" param of rtx_reader.
2536 * print-rtl.c (rtx_writer::print_rtx_operand): Print "(nil)"
2537 rather than an empty string for NULL strings.
2538 * read-md.c: Potentially include config.h rather than bconfig.h.
2539 Wrap include of errors.h with #ifdef GENERATOR_FILE.
2540 (have_error): New global, copied from errors.c.
2541 (md_reader::read_name): Rename to...
2542 (md_reader::read_name_1): ...this, adding "out_loc" param,
2543 and converting "missing name or number" to returning false, rather
2545 (md_reader::read_name): Reimplement in terms of read_name_1.
2546 (md_reader::read_name_or_nil): New function.
2547 (md_reader::read_string): Handle "(nil)" by returning NULL.
2548 (md_reader::md_reader): Add new param "compact".
2549 (md_reader::read_md_files): Wrap with #ifdef GENERATOR_FILE.
2550 (md_reader::read_file): New method.
2551 * read-md.h (md_reader::md_reader): Add new param "compact".
2552 (md_reader::read_file): New method.
2553 (md_reader::is_compact): New accessor.
2554 (md_reader::read_name): Convert return type from void to file_location.
2555 (md_reader::read_name_or_nil): New decl.
2556 (md_reader::read_name_1): New decl.
2557 (md_reader::m_compact): New field.
2558 (noop_reader::noop_reader): Pass "false" for new "compact" param
2560 (rtx_reader::rtx_reader): Add new "compact" param.
2561 (rtx_reader::read_rtx_operand): Make virtual and convert return
2562 type from void to rtx.
2563 (rtx_reader::read_until): New decl.
2564 (rtx_reader::handle_any_trailing_information): New virtual function.
2565 (rtx_reader::postprocess): New virtual function.
2566 (rtx_reader::finalize_string): New virtual function.
2567 (rtx_reader::m_in_call_function_usage): New field.
2568 (rtx_reader::m_reuse_rtx_by_id): New field.
2569 * read-rtl-function.c: New file.
2570 * selftest-rtl.c (selftest::assert_rtx_ptr_eq_at): New function.
2571 * selftest-rtl.h (ASSERT_RTX_PTR_EQ): New macro.
2572 (selftest::verify_three_block_rtl_cfg): New decl.
2573 * read-rtl-function.h: New file.
2574 * read-rtl.c: Potentially include config.h rather than bconfig.h.
2575 For host, include function.h, memmodel.h, and emit-rtl.h.
2576 (one_time_initialization): New function.
2577 (struct compact_insn_name): New struct.
2578 (compact_insn_names): New array.
2579 (find_code): Handle insn codes in compact dumps.
2580 (apply_subst_iterator): Wrap with #ifdef GENERATOR_FILE.
2581 (bind_subst_iter_and_attr): Likewise.
2582 (add_condition_to_string): Likewise.
2583 (add_condition_to_rtx): Likewise.
2584 (apply_attribute_uses): Likewise.
2585 (add_current_iterators): Likewise.
2586 (apply_iterators): Likewise.
2587 (initialize_iterators): Guard usage of apply_subst_iterator with
2588 #ifdef GENERATOR_FILE.
2589 (read_conditions): Wrap with #ifdef GENERATOR_FILE.
2590 (md_reader::read_mapping): Likewise.
2591 (add_define_attr_for_define_subst): Likewise.
2592 (add_define_subst_attr): Likewise.
2593 (read_subst_mapping): Likewise.
2594 (check_code_iterator): Likewise.
2595 (rtx_reader::read_rtx): Likewise. Move one-time initialization
2597 (one_time_initialization): New function.
2598 (rtx_reader::read_until): New method.
2599 (read_flags): New function.
2600 (parse_reg_note_name): New function.
2601 (rtx_reader::read_rtx_code): Initialize "iterator" to NULL.
2602 Handle reuse_rtx ids.
2603 Wrap iterator lookup within #ifdef GENERATOR_FILE.
2604 Add parsing support for RTL dumps, mirroring the special-cases in
2605 print_rtx, by calling read_flags, reading REG_NOTE names, INSN_UID
2606 values, and calling handle_any_trailing_information.
2607 (rtx_reader::read_rtx_operand): Convert return type from void
2608 to rtx, returning return_rtx. Handle case 'e'. Call
2609 finalize_string on XSTR and XTMPL fields.
2610 (rtx_reader::read_nested_rtx): Handle dumps in which trailing
2611 "(nil)" values were omitted. Call the postprocess vfunc on the
2613 (rtx_reader::rtx_reader): Add new "compact" param and pass to base
2614 class ctor. Initialize m_in_call_function_usage. Call
2615 one_time_initialization.
2616 * rtl-tests.c (selftest::test_uncond_jump): Call
2617 set_new_first_and_last_insn.
2618 * rtl.h (read_rtx): Wrap decl with #ifdef GENERATOR_FILE.
2619 * selftest-rtl.c: New file.
2620 * selftest-rtl.h (class selftest::rtl_dump_test): New class.
2621 (selftest::get_insn_by_uid): New decl.
2622 * selftest-run-tests.c (selftest::run_tests): Call
2623 read_rtl_function_c_tests.
2624 * selftest.h (selftest::read_rtl_function_c_tests): New decl.
2625 * tree-dfa.c (ssa_default_def): Return NULL_TREE for rtl function
2628 2017-01-05 Uros Bizjak <ubizjak@gmail.com>
2630 * config/i386/i386.md (*testqi_ext_3): No need to handle memory
2631 operands in a special way. Assert that pos+len <= mode precision.
2633 2017-01-05 Jakub Jelinek <jakub@redhat.com>
2635 * common.opt (fvect-cost-model): Remove RejectNegative flag, use
2636 3 argument Alias with unlimited for the negative form.
2637 (fno-vect-cost-model): Removed.
2639 2017-01-05 Martin Liska <mliska@suse.cz>
2641 * hsa-gen.c (gen_hsa_divmod): New function.
2642 (gen_hsa_insn_for_internal_fn_call): Use the function for IFN_DIVMOD.
2644 2017-01-05 Martin Liska <mliska@suse.cz>
2647 * gcc.c (lookup_compiler): Reject '-' filename for a precompiled
2650 2017-01-05 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
2652 * config/s390/s390.c (s390_expand_setmem): Unroll the loop for
2653 small constant length operands.
2655 2017-01-05 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
2657 * config/s390/s390.c (s390_expand_setmem): Avoid overlapping bytes
2658 between loop iterations.
2660 2017-01-05 Martin Liska <mliska@suse.cz>
2663 * gimplify.c (gimplify_decl_expr): Compare to
2664 asan_poisoned_variables instread of checking flags.
2665 (gimplify_target_expr): Likewise.
2666 (gimplify_expr): Likewise.
2667 (gimplify_function_tree): Conditionally initialize
2668 asan_poisoned_variables.
2670 2017-01-04 Jeff Law <law@redhat.com>
2672 PR tree-optimizatin/78812
2673 * rtl.h (contains_mem_rtx_p): Prototype.
2674 * ifcvt.c (containts_mem_rtx_p): Move from here to...
2675 * rtlanal.c (contains_mem_rtx_p): Here and remove static linkage.
2676 * gcse.c (prune_expressions): Use contains_mem_rtx_p to discover
2677 and prune MEMs that are not at the toplevel of a SET_SRC rtx. Look
2678 through ZERO_EXTEND and SIGN_EXTEND when trying to avoid pruning MEMs.
2680 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
2682 * input.c (assert_char_at_range): Default-initialize actual_range.
2684 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
2686 * df-scan.c (df_ref_create_structure): Make regno unsigned,
2687 to match the caller.
2689 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
2691 * cfgexpand.c (expand_gimple_basic_block): Disregard debug
2692 insns after final jump in test to emit dummy move.
2694 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
2696 * gimple-iterator.h (gsi_one_nondebug_before_end_p): New.
2697 * tree-eh.c (cleanup_empty_eh): Skip more debug stmts.
2699 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
2701 * multiple_target.c (create_dispatcher_calls): Init e_next.
2702 * tree-ssa-loop-split.c (split_loop): Init border.
2703 * tree-vect-loop.c (vect_determine_vectorization_factor): Init
2706 2017-01-04 Michael Meissner <meissner@linux.vnet.ibm.com>
2711 * config/rs6000/predicates.md (sf_subreg_operand): New predicate.
2712 (altivec_register_operand): Do not return true if the operand
2713 contains a SUBREG mixing SImode and SFmode.
2714 (vsx_register_operand): Likewise.
2715 (vsx_reg_sfsubreg_ok): New predicate.
2716 (vfloat_operand): Do not return true if the operand contains a
2717 SUBREG mixing SImode and SFmode.
2718 (vint_operand): Likewise.
2719 (vlogical_operand): Likewise.
2720 (gpc_reg_operand): Likewise.
2721 (int_reg_operand): Likewise.
2722 * config/rs6000/rs6000-protos.h (valid_sf_si_move): Add declaration.
2723 * config/rs6000/rs6000.c (valid_sf_si_move): New function to
2724 determine if a MOVSI or MOVSF operation contains SUBREGs that mix
2726 (rs6000_emit_move_si_sf_subreg): New helper function.
2727 (rs6000_emit_move): Call rs6000_emit_move_si_sf_subreg to possbily
2728 fixup SUBREGs involving SImode and SFmode.
2729 * config/rs6000/vsx.md (SFBOOL_*): New constants that are operand
2730 numbers for the new peephole2 optimization.
2731 (peephole2 for SFmode unions): New peephole2 to optimize cases in
2732 the GLIBC math library that do AND/IOR/XOR operations on single
2733 precision floating point.
2734 * config/rs6000/rs6000.h (TARGET_NO_SF_SUBREG): New internal
2735 target macros to say whether we need to avoid SUBREGs mixing
2737 (TARGET_ALLOW_SF_SUBREG): Likewise.
2738 * config/rs6000/rs6000.md (UNSPEC_SF_FROM_SI): New unspecs.
2739 (UNSPEC_SI_FROM_SF): Likewise.
2740 (iorxor): Change spacing.
2741 (and_ior_xor): New iterator for AND, IOR, and XOR.
2742 (movsi_from_sf): New insns for SImode/SFmode SUBREG support.
2743 (movdi_from_sf_zero_ext): Likewise.
2744 (mov<mode>_hardfloat, FMOVE32 iterator): Use register_operand
2745 instead of gpc_reg_operand. Add SImode/SFmode SUBREG support.
2746 (movsf_from_si): New insn for SImode/SFmode SUBREG support.
2747 (fma<mode>4): Use gpc_reg_operand instead of register_operand.
2748 (fms<mode>4): Likewise.
2749 (fnma<mode>4): Likewise.
2750 (fnms<mode>4): Likewise.
2751 (nfma<mode>4): Likewise.
2752 (nfms<mode>4): Likewise.
2754 2017-01-04 Marek Polacek <polacek@redhat.com>
2757 * doc/invoke.texi: Document -Wpointer-compare.
2759 2017-01-04 Jakub Jelinek <jakub@redhat.com>
2761 * optc-gen.awk: Emit #error for -W*/-f*/-m* Enum without
2764 * dwarf2out.c (output_loc_list): Don't throw away 64K+ location
2765 descriptions for -gdwarf-5 and emit them as uleb128 instead of
2768 2017-01-04 Kelvin Nilsen <kelvin@gcc.gnu.org>
2771 * doc/sourcebuild.texi (PowerPC-specific attributes): Add
2772 documentation of the powerpc_popcntb_ok attribute.
2773 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
2774 code to issue warning messages if a requested CPU configuration is
2775 not supported by the binary (assembler and loader) toolchain.
2776 (spe_init_builtins): Add two assertions to prevent ICE if attempt is
2777 made to define a built-in function that has been disabled.
2778 (paired_init_builtins): Add assertion to prevent ICE if attempt is
2779 made to define a built-in function that has been disabled.
2780 (altivec_init_builtins): Add comment explaining why definition
2781 of the DST built-in functions is not preceded by an assertion
2782 check. Add assertions to prevent ICE if attempts are made to
2783 define an altivec predicate or an abs* built-in function that has
2785 (htm_init_builtins): Add comment explaining why definition of the
2786 htm built-in functions is not preceded by an assertion check.
2788 2017-01-04 Jeff Law <law@redhat.com>
2790 PR tree-optimizatin/67955
2791 * tree-ssa-alias.c (same_addr_size_stores_p): Check offsets first.
2792 Allow any SSA_VAR_P as the base objects. Use integer_zerop. Verify
2793 the points-to solution does not include pt_null. Use DECL_PT_UID
2796 2017-01-04 Uros Bizjak <ubizjak@gmail.com>
2798 * config/i386/i386.md (HI/SImode test with imm to QImode splitters):
2799 Use gen_int_mode instead of gen_lopwart for const_int operands.
2801 2017-01-04 Jakub Jelinek <jakub@redhat.com>
2803 PR tree-optimization/71563
2804 * match.pd: Simplify X << Y into X if Y is known to be 0 or
2805 out of range value - has low bits known to be zero.
2807 2017-01-04 Alan Modra <amodra@gmail.com>
2809 * Makefile.in (aclocal_deps): Update and order as per aclocal.m4.
2810 * configure: Regenerate.
2811 * config.in: Regenerate.
2813 2017-01-04 Jakub Jelinek <jakub@redhat.com>
2816 * input.c (ebcdic_execution_charset::on_error): Don't use strstr for
2817 a substring of the message, but strcmp with the whole message. Ifdef
2818 ENABLE_NLS, translate the message first using dgettext.
2820 2017-01-03 Jeff Law <law@redhat.com>
2822 PR tree-optimizatin/78856
2823 * tree-ssa-threadupdate.c: Include tree-vectorizer.h.
2824 (mark_threaded_blocks): Remove code to truncate thread paths that
2825 cross multiple loop headers. Instead invalidate the cached loop
2826 iteration information and handle case of a thread path walking
2827 into an irreducible region.
2829 2017-01-03 Michael Meissner <meissner@linux.vnet.ibm.com>
2832 * config/rs6000/rs6000.c (rs6000_split_signbit): Change some
2833 assertions. Add support for doing the signbit if the IEEE 128-bit
2834 floating point value is in a GPR.
2835 * config/rs6000/rs6000.md (Fsignbit): Delete.
2836 (signbit<mode>2_dm): Delete using <Fsignbit> and just use "wa".
2837 Update the length attribute if the value is in a GPR.
2838 (signbit<mode>2_dm_<su>ext): Add combiner pattern to eliminate
2839 the sign or zero extension instruction, since the value is always 0/1.
2840 (signbit<mode>2_dm2): Delete using <Fsignbit>.
2843 * config/rs6000/vsx.md (vsx_extract_<mode>_store_p9): If we are
2844 extracting SImode to a GPR register so that we can generate a
2845 store, limit the vector to be in a traditional Altivec register
2846 for the vextuwrx instruction.
2848 2017-01-03 Ian Lance Taylor <iant@google.com>
2850 * godump.c (go_format_type): Treat ENUMERAL_TYPE like INTEGER_TYPE.
2852 2017-01-03 Martin Sebor <msebor@redhat.com>
2854 PR tree-optimization/78696
2855 * gimple-ssa-sprintf.c (format_floating): Correct handling of
2856 precision. Use MPFR for %f for greater fidelity. Correct handling
2858 (pass_sprintf_length::compute_format_length): Set width and precision
2859 specified by asrerisk to void_node for vararg functions.
2860 (try_substitute_return_value): Adjust dump output.
2862 2017-01-03 David Edelsohn <dje.gcc@gmail.com>
2864 * doc/invoke.texi (RS6000 options): LRA is enabled by default.
2866 2017-01-03 Eric Botcazou <ebotcazou@adacore.com>
2868 * doc/invoke.texi (SPARC options): Document -mlra as the default.
2869 * config/sparc/sparc.c (sparc_option_override): Force LRA unless
2870 -mlra/-mno-lra was passed to the compiler.
2872 2017-01-03 James Cowgill <James.Cowgill@imgtec.com>
2874 PR rtl-optimization/65618
2875 * emit-rtl.c (try_split): Move initialization of "before" and
2876 "after" to just before the call to emit_insn_after_setloc.
2878 2017-01-03 Gerald Pfeifer <gerald@pfeifer.com>
2880 * doc/md.texi (Standard Names): Remove reference to Java frontend.
2882 2017-01-03 Pierre-Marie de Rodat <derodat@adacore.com>
2884 * dwarf2out.c (gen_enumeration_type_die): When
2885 -gno-strict-dwarf, add a DW_AT_encoding attribute.
2887 2017-01-03 Jakub Jelinek <jakub@redhat.com>
2889 PR tree-optimization/78965
2890 * gimple-ssa-sprintf.c (pass_sprintf_length::compute_format_length):
2891 Change first argument from const call_info & to call_info &. For %n
2892 set info.nowrite to false.
2895 * gimple-ssa-sprintf.c (try_substitute_return_value): Don't change
2896 possibly throwing calls.
2898 * genmatch.c (dt_node::gen_kids_1): If generic_exprs include SSA_NAME
2899 and exprs_len || fns_len, emit the code for SSA_NAME next to the exprs
2900 and fns handling, rather than in a separate case SSA_NAME.
2902 2017-01-02 Jeff Law <law@redhat.com>
2904 * config/darwin-driver.c (darwin_driver_init): Const-correctness
2905 fixes for first_period and second_period variables.
2907 2017-01-02 Uros Bizjak <ubizjak@gmail.com>
2910 * config/i386/i386.md (UNSPEC_NOREX_MEM): New unspec.
2911 (*insvqi_1): New insn pattern.
2912 (*insvqi_1_mem_rex64): Ditto.
2914 (*insvqi_3): Rename from *insvqi.
2916 (*extzvqi_mem_rex64): Add UNSPEC_NOREX_MEM tag.
2918 2017-01-02 Gerald Pfeifer <gerald@pfeifer.com>
2920 * doc/cfg.texi (Edges): Remove reference to Java.
2921 (Maintaining the CFG): Ditto.
2923 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
2926 * symtab.c (symtab_node::binds_to_current_def_p): Fix handling of
2927 transparent aliases.
2929 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
2932 * predict.def (PRED_CALL): Update hitrate.
2933 (PRED_INDIR_CALL, PRED_POLYMORPHIC_CALL): New predictors.
2934 * predict.c (tree_estimate_probability_bb): Split CALL predictor
2935 into direct/indirect/polymorphic variants.
2937 2017-01-01 Jakub Jelinek <jakub@redhat.com>
2939 Update copyright years.
2941 * gcc.c (process_command): Update copyright notice dates.
2942 * gcov-dump.c (print_version): Ditto.
2943 * gcov.c (print_version): Ditto.
2944 * gcov-tool.c (print_version): Ditto.
2945 * gengtype.c (create_file): Ditto.
2946 * doc/cpp.texi: Bump @copying's copyright year.
2947 * doc/cppinternals.texi: Ditto.
2948 * doc/gcc.texi: Ditto.
2949 * doc/gccint.texi: Ditto.
2950 * doc/gcov.texi: Ditto.
2951 * doc/install.texi: Ditto.
2952 * doc/invoke.texi: Ditto.
2954 Copyright (C) 2017 Free Software Foundation, Inc.
2956 Copying and distribution of this file, with or without modification,
2957 are permitted in any medium without royalty provided the copyright
2958 notice and this notice are preserved.