1 2017-01-30 Dominik Vogt <vogt@linux.vnet.ibm.com>
4 * config/s390/s390.md ("*r<noxa>sbg_<mode>_srl_bitmask")
5 ("*r<noxa>sbg_<mode>_sll_bitmask")
6 ("*extzv_<mode>_srl<clobbercc_or_nocc>")
7 ("*extzv_<mode>_sll<clobbercc_or_nocc>"):
8 Use contiguous_bitmask_nowrap_operand.
10 2017-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13 * config/rs6000/altivec.h (vec_xl): Revise #define.
16 2017-01-27 Uros Bizjak <ubizjak@gmail.com>
18 * config/i386/i386.c (print_reg): Use REGNO instead of true_regnum.
20 2017-01-27 Bernd Schmidt <bschmidt@redhat.com>
22 PR rtl-optimization/79194
23 * cprop.c (one_cprop_pass): Move deletion of code after unconditional
24 traps before call to bypass_conditional_jumps.
26 2017-01-27 Vladimir Makarov <vmakarov@redhat.com>
28 PR tree-optimization/71374
29 * lra-constraints.c (check_conflict_input_operands): New.
30 (match_reload): Use it.
32 2017-01-27 Vladimir Makarov <vmakarov@redhat.com>
35 * lra-assigns.c (find_hard_regno_for_1): Take endianess for into
36 account to calculate conflict_set.
38 2017-01-27 Bin Cheng <bin.cheng@arm.com>
40 PR rtl-optimization/78559
41 * combine.c (try_combine): Discard REG_EQUAL and REG_EQUIV for
42 other_insn in combine.
44 2017-01-27 Pekka Jääskeläinen <pekka.jaaskelainen@parmance.com>
46 * builtin-types.def: Use unsigned_char_type_node for BT_UINT8. Use
47 uint16_type_node for BT_UINT16.
49 2017-01-27 David Malcolm <dmalcolm@redhat.com>
51 * doc/sourcebuild.texi (Testsuites): Add "GIMPLE Tests" and
53 (GIMPLE Tests): New node.
54 (RTL Tests): New node.
56 2017-01-27 Richard Biener <rguenther@suse.de>
58 PR tree-optimization/79245
59 * tree-loop-distribution.c (distribute_loop): Apply cost
60 modeling also to detected patterns.
62 2017-01-27 Richard Biener <rguenther@suse.de>
64 PR tree-optimization/71433
65 * tree-vrp.c (register_new_assert_for): Revert earlier changes.
66 (compare_assert_loc): New function.
67 (process_assert_insertions): Sort and optimize assert locations
68 to remove duplicates and push down identical assertions on
69 edges to their destination block.
71 2017-01-27 Richard Biener <rguenther@suse.de>
73 PR tree-optimization/79244
74 * tree-vrp.c (remove_range_assertions): Forcefully propagate
75 out SSA names even if abnormal.
77 2017-01-27 Jakub Jelinek <jakub@redhat.com>
79 * realmpfr.h: Poison MPFR_RND{N,Z,U,D}.
80 * gimple-ssa-sprintf.c (format_floating_max): Use GMP_RNDN
83 2017-01-27 Richard Earnshaw <rearnsha@arm.com>
86 * arm.c (arm_option_override): Don't call build_target_option_node
87 until after doing all option overrides.
88 (arm_valid_target_attribute_tree): Likewise.
90 2017-01-27 Martin Liska <mliska@suse.cz>
92 * doc/invoke.texi (-fprofile-arcs): Document profiling support
93 for {cd}tors and C++ {cd}tors.
95 2017-01-27 Dominik Vogt <vogt@linux.vnet.ibm.com>
97 * config/s390/s390.md ("*setmem_long_and")
98 ("*setmem_long_and_31z"): Use zero_extend instead of and.
100 2017-01-26 Martin Sebor <msebor@redhat.com>
102 * gimple-ssa-sprintf.c (format_floating): Simplify the computation
105 2017-01-26 Martin Sebor <msebor@redhat.com>
107 * gimple-ssa-sprintf.c (format_floating): Test HAVE_XFmode and
108 HAVE_DFmode before using XFmode or DFmode.
109 (parse_directive): Avoid using the z length modifier to avoid
110 the ISO C++98 does not support the ‘z’ gnu_printf length modifier.
113 * gimple-ssa-sprintf.c (adjust_for_width_or_precision): Change
114 to accept adjustment as an array.
115 (get_int_range): New function.
116 (struct directive): Make width and prec arrays.
117 (directive::set_width, directive::set_precision): Call get_int_range.
118 (format_integer, format_floating): Handle width and precision ranges.
119 (format_string, parse_directive): Same.
121 2017-01-26 Jakub Jelinek <jakub@redhat.com>
124 * dwarf2out.c (generate_skeleton_bottom_up): For children with
125 comdat_type_p set, just clone them, but keep the children in the
129 * dwarf2out.c (prune_unused_types): Mark all functions with DIEs
130 which have direct callers with -fvar-tracking-assignments enabled
132 (resolve_addr): Avoid adding skeleton DIEs for DW_AT_call_origin
133 inside of type units.
135 2017-01-26 Martin Sebor <msebor@redhat.com>
138 * gimple-ssa-sprintf.c (struct result_range): Add likely and
140 (struct format_result): Replace number_chars, number_chars_min,
141 and number_chars_max with a single member of struct result_range.
143 (format_result::operator+=): Adjust.
144 (struct fmtresult): Remove bounded. Handle likely and unlikely
146 (fmtresult::adjust_for_width_or_precision): New function.
147 (fmtresult:type_max_digits): New function.
148 (bytes_remaining): Handle likely and unlikely counters.
149 (min_bytes_remaining): Remove.
150 (format_percent): Simplify.
151 (format_integer, format_floating): Set likely and unlikely counters.
152 (get_string_length, format_character, format_string): Same.
153 (format_plain, should_warn_p): New function.
154 (maybe_warn): Call should_warn_p. Update diagnostic messages
155 and handle those for all directives, including plain strings.
156 (format_directive): Handle likely and unlikely counters.
157 Remove unnecessary quoting from diagnostics. Add an informational
160 (pass_sprintf_length::compute_format_length): Simplify.
161 (try_substitute_return_value): Handle likely and unlikely counters.
163 2017-01-26 Carl Love <cel@us.ibm.com>
165 * config/rs6000/rs6000-c (altivec_overloaded_builtins): Remove
166 bogus entries for the P8V_BUILTIN_VEC_VGBBD built-ins
168 2017-01-26 Vladimir Makarov <vmakarov@redhat.com>
171 * lra-assigns.c (setup_live_pseudos_and_spill_after_risky): Take
172 endianess for subregs into account.
173 * lra-constraints.c (lra_constraints): Do risky transformations
174 always on the first iteration.
175 * lra-lives.c (check_pseudos_live_through_calls): Add arg
176 last_call_used_reg_set.
177 (process_bb_lives): Define and use last_call_used_reg_set.
178 * lra.c (lra): Always continue after lra_constraints on the first
181 2017-01-26 Kirill Yukhin <kirill.yukhin@gmail.com>
183 * gcc.target/i386/avx512bw-kshiftlq-2.c: Use unsigned long long
185 * gcc.target/i386/avx512bw-kshiftrq-2.c: Ditto.
187 2017-01-26 Jakub Jelinek <jakub@redhat.com>
189 * config/i386/avx512fintrin.h (_ktest_mask16_u8,
190 _ktestz_mask16_u8, _ktestc_mask16_u8, _kadd_mask16): Move to ...
191 * config/i386/avx512dqintrin.h (_ktest_mask16_u8,
192 _ktestz_mask16_u8, _ktestc_mask16_u8, _kadd_mask16): ... here.
193 * config/i386/i386-builtin.def (__builtin_ia32_ktestchi,
194 __builtin_ia32_ktestzhi, __builtin_ia32_kaddhi): Use
195 OPTION_MASK_ISA_AVX512DQ instead of OPTION_MASK_ISA_AVX512F.
196 * config/i386/sse.md (SWI1248_AVX512BWDQ2): New mode iterator.
197 (kadd<mode>, ktest<mode>): Use it instead of SWI1248_AVX512BWDQ.
199 2017-01-26 Marek Polacek <polacek@redhat.com>
202 * fold-const.c (operand_equal_p) [COND_EXPR]: Use OP_SAME_WITH_NULL
203 for the third operand.
205 2017-01-26 Jakub Jelinek <jakub@redhat.com>
208 * omp-low.c (struct omp_context): Add simt_stmt field.
209 (scan_omp_for): Return omp_context *.
210 (scan_omp_simd): Set simt_stmt on the non-_simt_ SIMD
211 context to the _simt_ SIMD stmt.
212 (lower_omp_for): For combined SIMD with sibling _simt_
213 SIMD, make sure to use the same decls in _looptemp_
214 clauses as in the sibling.
216 2017-01-26 David Sherwood <david.sherwood@arm.com>
219 * gimplify.c (omp_notice_variable): Add GOVD_SEEN flag to variables in
222 2017-01-26 Jakub Jelinek <jakub@redhat.com>
225 * reg-stack.c (emit_swap_insn): Instead of fld a; fld b; fxchg %st(1);
226 emit fld b; fld a; if possible.
228 * brig-builtins.def: Update copyright years.
229 * config/arm/arm_acle_builtins.def: Update copyright years.
231 2017-01-25 Michael Meissner <meissner@linux.vnet.ibm.com>
234 * config/rs6000/vsx.md (vsx_extract_<mode>_store): Use wY
235 constraint instead of o for the stxsd instruction.
237 2017-01-25 Carl Love <cel@us.ibm.com>
239 * config/rs6000/rs6000-c (altivec_overloaded_builtins): Fix order
240 of entries for ALTIVEC_BUILTIN_VEC_PACKS and P8V_BUILTIN_VEC_VGBBD.
242 2017-01-25 Jonathan Wakely <jwakely@redhat.com>
244 * doc/invoke.texi (C++ Dialect Options): Fix typo.
246 2017-01-25 Richard Biener <rguenther@suse.de>
248 PR tree-optimization/69264
249 * target.def (vector_alignment_reachable): Improve documentation.
250 * doc/tm.texi: Regenerate.
251 * targhooks.c (default_builtin_vector_alignment_reachable): Simplify
253 * tree-vect-data-refs.c (vect_supportable_dr_alignment): Revert
254 earlier changes with respect to TYPE_USER_ALIGN.
255 (vector_alignment_reachable_p): Likewise. Improve dumping.
257 2016-01-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
260 * config/arm/arm.md (xordi3): Force constant operand into a register
263 2016-01-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
265 * doc/invoke.texi (-fstore-merging): Correct default optimization
266 levels at which it is enabled.
267 (-O): Move -fstore-merging from list to...
270 2017-01-25 Richard Biener <rguenther@suse.de>
273 * omp-expand.c: Include debug.h.
274 (expand_omp_taskreg): Make sure to generate early debug before
275 outlining anything from a function.
276 (expand_omp_target): Likewise.
277 (grid_expand_target_grid_body): Likewise.
279 2017-01-25 Maxim Ostapenko <m.ostapenko@samsung.com>
282 * asan.c (get_translation_unit_decl): New function.
283 (asan_add_global): Extract modules file name from globals
284 TRANSLATION_UNIT_DECL name.
286 2017-01-24 Eric Botcazou <ebotcazou@adacore.com>
289 * config/arm/arm.c (arm_function_ok_for_sibcall): Add back restriction
290 for long calls with APCS frame and VFP.
292 2017-01-24 David Malcolm <dmalcolm@redhat.com>
294 * cfg.c (original_copy_tables_initialized_p): New function.
295 * cfg.h (original_copy_tables_initialized_p): New decl.
296 * cfgrtl.c (relink_block_chain): Guard the call to
297 free_original_copy_tables with a call to
298 original_copy_tables_initialized_p.
299 * cgraph.h (symtab_node::native_rtl_p): New decl.
300 * cgraphunit.c (symtab_node::native_rtl_p): New function.
301 (symtab_node::needed_p): Don't assert for early assembly output
303 (cgraph_node::finalize_function): Set "force_output" for __RTL
305 (cgraph_node::analyze): Bail out early for __RTL functions.
306 (analyze_functions): Update assertion to support __RTL functions.
307 (cgraph_node::expand): Bail out early for __RTL functions.
308 * final.c (rest_of_clean_state): Don't call delete_tree_ssa for
310 * function.h (struct function): Update comment for field
312 * gimple-expr.c: Include "tree-pass.h".
313 (gimple_has_body_p): Return false for __RTL functions.
314 * Makefile.in (OBJS): Add run-rtl-passes.o.
315 * pass_manager.h (gcc::pass_manager::get_rest_of_compilation): New
317 (gcc::pass_manager::get_clean_slate): New accessor.
318 * passes.c: Include "insn-addr.h".
319 (should_skip_pass_p): Add logging. Update logic for running
320 "expand" to be compatible with both __GIMPLE and __RTL. Guard
321 property-provider override so it is only done for gimple passes.
323 (skip_pass): New function.
324 (execute_one_pass): Call skip_pass when skipping passes.
325 * read-md.c (md_reader::read_char): Support filtering
326 the input to a subset of line numbers.
327 (md_reader::md_reader): Initialize fields
328 m_first_line and m_last_line.
329 (md_reader::read_file_fragment): New function.
330 * read-md.h (md_reader::read_file_fragment): New decl.
331 (md_reader::m_first_line): New field.
332 (md_reader::m_last_line): New field.
333 * read-rtl-function.c (function_reader::create_function): Only
334 create cfun if it doesn't already exist. Set PROP_rtl on cfun's
335 curr_properties. Set DECL_INITIAL to a dummy block.
336 (read_rtl_function_body_from_file_range): New function.
337 * read-rtl-function.h (read_rtl_function_body_from_file_range):
339 * run-rtl-passes.c: New file.
340 * run-rtl-passes.h: New file.
342 2017-01-24 Jeff Law <law@redhat.com>
344 * config/microblaze/microblaze.h (ASM_FORMAT_PRIVATE_NAME): Increase
347 2017-01-24 Bin Cheng <bin.cheng@arm.com>
349 PR tree-optimization/79159
350 * tree-ssa-loop-niter.c (get_cst_init_from_scev): New function.
351 (record_nonwrapping_iv): Improve boundary using above function if no
352 value range information.
354 2017-01-24 Pekka Jääskeläinen <pekka@parmance.com>
355 Martin Jambor <mjambor@suse.cz>
357 * brig-builtins.def: New file.
358 * builtins.def (DEF_HSAIL_BUILTIN): New macro.
359 (DEF_HSAIL_ATOMIC_BUILTIN): Likewise.
360 (DEF_HSAIL_SAT_BUILTIN): Likewise.
361 (DEF_HSAIL_INTR_BUILTIN): Likewise.
362 (DEF_HSAIL_CVT_ZEROI_SAT_BUILTIN): Likewise.
363 * builtin-types.def (BT_INT8): New.
364 (BT_INT16): Likewise.
365 (BT_UINT8): Likewise.
366 (BT_UINT16): Likewise.
367 (BT_FN_ULONG): Likewise.
368 (BT_FN_UINT_INT): Likewise.
369 (BT_FN_UINT_ULONG): Likewise.
370 (BT_FN_UINT_LONG): Likewise.
371 (BT_FN_UINT_PTR): Likewise.
372 (BT_FN_ULONG_PTR): Likewise.
373 (BT_FN_INT8_FLOAT): Likewise.
374 (BT_FN_INT16_FLOAT): Likewise.
375 (BT_FN_UINT32_FLOAT): Likewise.
376 (BT_FN_UINT16_FLOAT): Likewise.
377 (BT_FN_UINT8_FLOAT): Likewise.
378 (BT_FN_UINT64_FLOAT): Likewise.
379 (BT_FN_UINT16_UINT32): Likewise.
380 (BT_FN_UINT32_UINT16): Likewise.
381 (BT_FN_UINT16_UINT16_UINT16): Likewise.
382 (BT_FN_INT_PTR_INT): Likewise.
383 (BT_FN_UINT_PTR_UINT): Likewise.
384 (BT_FN_LONG_PTR_LONG): Likewise.
385 (BT_FN_ULONG_PTR_ULONG): Likewise.
386 (BT_FN_VOID_UINT64_UINT64): Likewise.
387 (BT_FN_UINT8_UINT8_UINT8): Likewise.
388 (BT_FN_INT8_INT8_INT8): Likewise.
389 (BT_FN_INT16_INT16_INT16): Likewise.
390 (BT_FN_INT_INT_INT): Likewise.
391 (BT_FN_UINT_FLOAT_UINT): Likewise.
392 (BT_FN_FLOAT_UINT_UINT): Likewise.
393 (BT_FN_ULONG_UINT_UINT): Likewise.
394 (BT_FN_ULONG_UINT_PTR): Likewise.
395 (BT_FN_ULONG_ULONG_ULONG): Likewise.
396 (BT_FN_UINT_UINT_UINT): Likewise.
397 (BT_FN_VOID_UINT_PTR): Likewise.
398 (BT_FN_UINT_UINT_PTR: Likewise.
399 (BT_FN_UINT32_UINT64_PTR): Likewise.
400 (BT_FN_INT_INT_UINT_UINT): Likewise.
401 (BT_FN_UINT_UINT_UINT_UINT): Likewise.
402 (BT_FN_UINT_UINT_UINT_PTR): Likewise.
403 (BT_FN_UINT_ULONG_ULONG_UINT): Likewise.
404 (BT_FN_ULONG_ULONG_ULONG_ULONG): Likewise.
405 (BT_FN_LONG_LONG_UINT_UINT): Likewise.
406 (BT_FN_ULONG_ULONG_UINT_UINT): Likewise.
407 (BT_FN_VOID_UINT32_UINT64_PTR): Likewise.
408 (BT_FN_VOID_UINT32_UINT32_PTR): Likewise.
409 (BT_FN_UINT_UINT_UINT_UINT_UINT): Likewise.
410 (BT_FN_UINT_FLOAT_FLOAT_FLOAT_FLOAT): Likewise.
411 (BT_FN_ULONG_ULONG_ULONG_UINT_UINT): Likewise.
412 * doc/frontends.texi: List BRIG FE.
413 * doc/install.texi (Testing): Add BRIG tesring requirements.
414 * doc/invoke.texi (Overall Options): Mention BRIG.
415 * doc/standards.texi (Standards): Doucment BRIG HSA version.
417 2017-01-24 Richard Biener <rguenther@suse.de>
420 * ipa-devirt.c (odr_types_equivalent_p): Fix typo in diagnostic.
422 2017-01-24 Martin Jambor <mjambor@suse.cz>
425 * ipa-prop.c (ipa_free_all_node_params): Call summary destructor.
426 * ipa-prop.c (ipa_node_params_t::insert): Initialize fields known_csts
429 2017-01-24 Aldy Hernandez <aldyh@redhat.com>
432 * gimple-ssa-warn-alloca.c (alloca_call_type): Make sure
433 casts from signed to unsigned really don't have a range.
435 2017-01-24 Markus Trippelsdorf <markus@trippelsdorf.de>
437 * gimple-ssa-sprintf.c (format_floating): Change MPFR_RNDx to
438 GMP_RNDx for compatiblity.
440 2017-01-24 Martin Liska <mliska@suse.cz>
443 * tree-ssa-reassoc.c (rewrite_expr_tree_parallel): Insert assert
444 that would prevent us to call alloca with -1 as argument.
446 2017-01-24 Jakub Jelinek <jakub@redhat.com>
448 * dwarf2out.c (output_compilation_unit_header, output_file_names):
449 Avoid -Wformat-security warning.
451 2017-01-23 Andrew Pinski <apinski@cavium.com>
453 * config/aarch64/aarch64.c (thunderx2t99_addrcost_table): Improve
456 2017-01-23 Martin Sebor <msebor@redhat.com>
459 * gimple-ssa-sprintf.c (warn_level): New global.
460 (format_integer): Use it here and throughout the rest of the file.
461 Use the same switch to compute sign as base.
462 (maybe_warn): New function.
463 (format_directive): Factor out warnings into maybe_warn.
464 Add debugging output. Use warn_level.
465 (add_bytes): Use warn_level.
466 (pass_sprintf_length::compute_format_length): Add debugging output.
467 (try_substitute_return_value): Same.
468 (pass_sprintf_length::handle_gimple_call): Set and use warn_level.
471 * gimple-ssa-sprintf.c (struct format_result): Remove constant member.
472 (struct fmtresult, format_integer, format_floating): Adjust.
473 (fmtresult::fmtresult): Set max correctly in two argument ctor.
474 (get_string_length, format_string,format_directive): Same.
475 (pass_sprintf_length::compute_format_length): Same.
476 (try_substitute_return_value): Simplify slightly.
479 * gimple-ssa-sprintf.c (pass_sprintf_length::gate): Adjust formatting.
480 (fmtresult::operator+=): Outlined.
481 (struct fmtresult): Add ctors.
482 (struct conversion_spec): Rename...
483 (struct directive): ...to this. Add and remove data members.
484 (directive::set_width, directive::set_precision): New functions.
485 (format_percent): Use fmtresult ctor.
486 (get_width_and_precision): Remove.
487 (format_integer): Make naming changes. Avoid computing width and
489 (format_floating): Same. Adjust indentation.
490 (format_character, format_none): New functions.
491 (format_string): Moved character handling to format_character.
492 (format_directive): Remove arguments, change return type.
493 (parse_directive): New function.
494 (pass_sprintf_length::compute_format_length): Move directive
495 parsing to parse_directive.
497 2017-01-23 Jakub Jelinek <jakub@redhat.com>
499 * tree.h (assign_assembler_name_if_neeeded): Rename to ...
500 (assign_assembler_name_if_needed): ... this.
501 * tree.c (assign_assembler_name_if_neeeded): Rename to ...
502 (assign_assembler_name_if_needed): ... this.
503 (free_lang_data_in_cgraph): Adjust callers.
504 * cgraphunit.c (cgraph_node::analyze): Likewise.
505 * omp-expand.c (expand_omp_taskreg, expand_omp_target):
508 2017-01-23 Richard Biener <rguenther@suse.de>
510 PR tree-optimization/79088
511 PR tree-optimization/79188
512 * tree-ssa-threadupdate.c (mark_threaded_blocks): Move code
513 resetting loop bounds after last path deletion. Reset loop
514 bounds of the target loop, make code match the comments.
515 * tree-ssa-threadbackwards.c (pass_early_thread_jumps::execute):
516 Make sure loops need no fixups.
518 2017-01-23 Kelvin Nilsen <kelvin@gcc.gnu.org>
520 * config/rs6000/rs6000-builtin.def (VSIEDPF): Add scalar insert
521 exponent support with double type for first argument.
522 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Changed
523 type returned by __builtin_vec_extract_sig,
524 __builtin_vec_extract_sig_sp, and __builtin_vec_extract_sig_dp
525 functions from "vector int" to "vector unsigned int" or from
526 "vector long long int" to "vector unsigned long long int".
527 Changed type returned by __builtin_vec_extract_exp,
528 __builtin_vec_extract_exp_sp, and __builtin_vec_extract_exp_dp
529 functions from "vector int" to "vector unsigned int" or from
530 "vector long long int" to "vector unsigned long long int".
531 Changed return type of __builtin_vec_test_data_class,
532 __builtin_vec_test_data_class_sp, and
533 __builtin_vec_test_data_class_dp from "vector int" to
534 "vector bool int" or from "vector long long int" to "vector bool
535 long long int" and changed second argument type from "unsigned
536 int" to "int". Added new overloaded function forms "vector float
537 __builtin_vec_insert_exp (vector float, vector unsigned int)" and
538 "vector float __builtin_vec_insert_exp_sp (vector float, vector
539 unsigned int)" and "vector double __builtin_vec_insert_exp (vector
540 double, vector unsigned long long int)" and "vector double
541 __builtin_vec_insert_exp_dp (vector double, vector unsigned long
542 long int)". Changed return type of
543 __builtin_scalar_test_data_class and
544 __builtin_scalar_test_data_class_sp and
545 __builtin_scalar_test_data_class_dp from "unsigned int" to "bool
546 int" and changed second argument from "unsigned int" to "int".
547 Changed type returned by __builtin_scalar_test_neg,
548 __builtin_scalar_test_neg_sp, and __builtin_scalar_test_neg_dp
549 from "int" to "bool int". Added new overloaded function form
550 "double __builtin_scalar_insert_exp (double, unsigned long long int)".
551 * config/rs6000/vsx.md (xsiexpdpf): New insn for scalar insert
552 exponent double-precision with floating point first argument.
553 * doc/extend.texi (PowerPC AltiVec Built-in Functions): Adjust
554 documentation of scalar_test_data_class, scalar_test_neg,
555 scalar_extract_sig, scalar_extract_exp, scalar_insert_exp,
556 vector_extract_exp, vec_extract_sig, vec_insert_exp, and
557 vec_test_data_class built-in functions to reflect refinements in
558 their type signatures.
560 2017-01-23 Andreas Tobler <andreast@gcc.gnu.org>
562 * config/aarch64/aarch64.c (aarch64_elf_asm_constructor): Increase
564 (aarch64_elf_asm_destructor): Likewise.
566 2017-01-23 Bernd Schmidt <bschmidt@redhat.com>
568 PR rtl-optimization/78634
569 * config/i386/i386.c (ix86_max_noce_ifcvt_seq_cost): New function.
570 (TARGET_MAX_NOCE_IFCVT_SEQ_COST): Define.
571 * ifcvt.c (noce_try_cmove): Add missing cost check.
573 PR rtl-optimization/71724
574 * combine.c (if_then_else_cond): Look for situations where it is
575 beneficial to undo the work of one of the recursive calls.
577 2017-01-23 Bin Cheng <bin.cheng@arm.com>
579 PR tree-optimization/70754
580 * tree-predcom.c (stmt_combining_refs): New parameter INSERT_BEFORE.
581 (reassociate_to_the_same_stmt): New parameter INSERT_BEFORE. Insert
582 combined stmt before it if not NULL.
583 (combine_chains): Process refs reversely and compute dominance point
586 2017-01-23 Martin Liska <mliska@suse.cz>
588 PR tree-optimization/79196
589 * tree-ssa-strlen.c (fold_strstr_to_memcmp): Rename to ...
590 (fold_strstr_to_strncmp): ... this. Fold the pattern to strncmp
592 (strlen_optimize_stmt): Call the renamed function.
594 2017-01-23 Michael Matz <matz@suse.de>
596 PR tree-optimization/78384
597 * tree-ssa-loop-split.c (patch_loop_exit): Use correct edge.
599 2017-01-23 Richard Biener <rguenther@suse.de>
601 PR tree-optimization/79186
602 * tree-vrp.c (register_new_assert_for): Make sure we've seen
603 both incoming edges before moving an assert.
605 2017-01-23 Martin Jambor <mjambor@suse.cz>
607 * ipa-prop.c (load_from_param_1): Removed.
608 (load_from_unmodified_param): Bits from load_from_param_1 put back
610 (load_from_param): Removed.
611 (compute_complex_assign_jump_func): Removed stmt2 and just replaced it
612 with stmt. Reverted back to use of load_from_unmodified_param.
614 2017-01-23 Martin Jambor <mjambor@suse.cz>
617 * ipa-prop.h (ipa_param_descriptor): Anotate with with GTY(()).
618 (ipa_node_params): Annotate with GTY((for_user)). Make descriptors
619 field a pointer to garbage collected vector, mark lattices and
620 ipcp_orig_node with GTY((skip)).
621 (ipa_get_param_count): Adjust to descriptors being a pointer.
622 (ipa_get_param): Likewise.
623 (ipa_get_type): Likewise.
624 (ipa_get_param_move_cost): Likewise.
625 (ipa_set_param_used): Likewise.
626 (ipa_get_controlled_uses): Likewise.
627 (ipa_set_controlled_uses): Likewise.
628 (ipa_is_param_used): Likewise.
629 (ipa_node_params_t): Move into garbage collector. New methods insert
631 (ipa_node_params_sum): Annotate wth GTY(()).
632 (ipa_check_create_node_params): Adjust to ipa_node_params_sum being
634 (ipa_load_from_parm_agg): Adjust declaration.
635 * ipa-icf.c (param_used_p): Adjust to descriptors being a pointer.
636 * ipa-profile.c (ipa_profile): Likewise.
637 * ipa-prop.c (ipa_get_param_decl_index_1): Likewise.
638 (ipa_populate_param_decls): Make descriptors parameter garbage
640 (ipa_dump_param): Adjust to descriptors being a pointer.
641 (ipa_alloc_node_params): Likewise.
642 (ipa_initialize_node_params): Likewise.
643 (load_from_param_1): Make descriptors parameter garbage collected.
644 (load_from_unmodified_param): Likewise.
645 (load_from_param): Likewise.
646 (ipa_load_from_parm_agg): Likewise.
647 (ipa_node_params::~ipa_node_params): Removed.
648 (ipa_free_all_node_params): Remove call to delete operator.
649 (ipa_node_params_t::insert): New.
650 (ipa_node_params_t::remove): Likewise.
651 (ipa_node_params_t::duplicate): Adjust to descriptors being a pointer,
652 copy known_csts and known_contexts vectors.
653 (ipa_read_node_info): Adjust to descriptors being a pointer.
654 (ipcp_modif_dom_walker): Make m_descriptors field garbage
656 (ipcp_transform_function): Make descriptors variable garbage
659 2017-01-23 Andrew Senkevich <andrew.senkevich@intel.com>
661 * config/i386/avx512bwintrin.h: Add k-mask test, kortest intrinsics.
662 * config/i386/avx512dqintrin.h: Ditto.
663 * config/i386/avx512fintrin.h: Ditto.
664 * gcc/config/i386/i386.c: Handle new builtins.
665 * config/i386/i386-builtin.def: Add new builtins.
666 * config/i386/sse.md (ktest<mode>, kortest<mode>): New.
667 (UNSPEC_KORTEST, UNSPEC_KTEST): New.
669 2017-01-23 Jakub Jelinek <jakub@redhat.com>
670 Martin Liska <mliska@suse.cz>
672 * asan.h: Define ASAN_USE_AFTER_SCOPE_ATTRIBUTE.
673 * asan.c (asan_expand_poison_ifn): Support stores and use
674 appropriate ASAN report function.
675 * internal-fn.c (expand_ASAN_POISON_USE): New function.
676 * internal-fn.def (ASAN_POISON_USE): Declare.
677 * tree-into-ssa.c (maybe_add_asan_poison_write): New function.
678 (maybe_register_def): Create ASAN_POISON_USE when sanitizing.
679 * tree-ssa-dce.c (eliminate_unnecessary_stmts): Remove
680 ASAN_POISON calls w/o LHS.
681 * tree-ssa.c (execute_update_addresses_taken): Create clobber
682 for ASAN_MARK (UNPOISON, &x, ...) in order to prevent usage of a LHS
683 from ASAN_MARK (POISON, &x, ...) coming to a PHI node.
684 * gimplify.c (asan_poison_variables): Add attribute
685 use_after_scope_memory to variables that really needs to live
687 * tree-ssa.c (is_asan_mark_p): Do not rewrite into SSA when
688 having the attribute.
690 2017-01-23 Martin Liska <mliska@suse.cz>
692 * asan.c (create_asan_shadow_var): New function.
693 (asan_expand_poison_ifn): Likewise.
694 * asan.h (asan_expand_poison_ifn): New declaration.
695 * internal-fn.c (expand_ASAN_POISON): Likewise.
696 * internal-fn.def (ASAN_POISON): New builtin.
697 * sanopt.c (pass_sanopt::execute): Expand
698 asan_expand_poison_ifn.
699 * tree-inline.c (copy_decl_for_dup_finish): Make function
701 * tree-inline.h (copy_decl_for_dup_finish): Likewise.
702 * tree-ssa.c (is_asan_mark_p): New function.
703 (execute_update_addresses_taken): Rewrite local variables
704 (identified just by use-after-scope as addressable) into SSA.
706 2017-01-22 Gerald Pfeifer <gerald@pfeifer.com>
708 * doc/install.texi (Specific): opensource.apple.com uses https
709 now. Remove trailing slash.
711 2017-01-22 Gerald Pfeifer <gerald@pfeifer.com>
713 * README.Portability: Remove note on an Irix compatibility issue.
715 2017-01-22 Dimitry Andric <dim@FreeBSD.org>
717 * gcov.c (INCLUDE_ALGORITHM): Define.
718 (INCLUDE_VECTOR): Define.
719 No longer include <vector> and <algorithm> directly.
721 2017-01-21 Gerald Pfeifer <gerald@pfeifer.com>
723 * doc/extend.texi (Thread-Local): Change www.akkadia.org reference
725 * doc/invoke.texi (Code Gen Options): Ditto.
727 2017-01-21 Jan Hubicka <hubicka@ucw.cz>
730 * cfg.c (update_bb_profile_for_threading): Fix updating of probablity.
732 2017-01-21 Bernd Schmidt <bschmidt@redhat.com>
734 rtl-optimization/79125
735 * cprop.c (local_cprop_pass): Handle cases where we make an
738 2017-01-20 Segher Boessenkool <segher@kernel.crashing.org>
742 * config/rs6000/rs6000.c (rs6000_gimplify_va_arg): Adjust address to
743 read from, for big endian.
745 2017-01-20 Jiong Wang <jiong.wang@arm.com>
747 * config/aarch64/aarch64-builtins.c (aarch64_init_builtins): Register
748 register pauth builtins for LP64 only.
750 2017-01-20 Marek Polacek <polacek@redhat.com>
753 * gimplify.c (should_warn_for_implicit_fallthrough): Handle consecutive
756 2017-01-20 Alexander Monakov <amonakov@ispras.ru>
758 * omp-expand.c (expand_omp_simd): Clear PROP_gimple_lomp_dev regardless
760 * omp-offload.c (pass_omp_device_lower::gate): Use PROP_gimple_lomp_dev.
761 * passes.c (dump_properties): Handle PROP_gimple_lomp_dev.
762 * tree-inline.c (expand_call_inline): Propagate PROP_gimple_lomp_dev.
764 2017-01-20 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
767 * config/arm/arm.c (neon_valid_immediate): Reject vector constants
768 in big-endian mode when they are not a single duplicated value.
770 2017-01-20 Richard Biener <rguenther@suse.de>
772 * BASE-VER: Bump to 7.0.1.
774 2017-01-20 Alexander Monakov <amonakov@ispras.ru>
776 * omp-low.c (omplow_simd_context): New struct. Use it...
777 (lower_rec_simd_input_clauses): ...here and...
778 (lower_rec_input_clauses): ...here to hold common data. Adjust all
779 references to idx, lane, max_vf, is_simt.
781 2017-01-20 Graham Markall <graham.markall@embecosm.com>
783 * config/arc/arc.h (LINK_SPEC): Use arclinux_nps emulation when
786 2017-01-20 Martin Jambor <mjambor@suse.cz>
788 * hsa.h: Renaed to hsa-common.h. Adjusted a comment.
789 * hsa.c: Renaed to hsa-common.c. Change include of gt-hsa.h to
791 * Makefile.in (OBJS): Rename hsa.o to hsa-common.o.
792 (GTFILES): Rename hsa.c to hsa-common.c.
793 * hsa-brig.c: Change include of hsa.h to hsa-common.h.
794 * hsa-dump.c: Likewise.
795 * hsa-gen.c: Likewise.
796 * hsa-regalloc.c: Likewise.
797 * ipa-hsa.c: Likewise.
798 * omp-expand.c: Likewise.
799 * omp-low.c: Likewise.
800 * toplev.c: Likewise.
802 2017-01-20 Marek Polacek <polacek@redhat.com>
805 * doc/invoke.texi: Document -Wduplicated-branches.
806 * fold-const.c (operand_equal_p): Handle MODIFY_EXPR, INIT_EXPR,
807 COMPOUND_EXPR, PREDECREMENT_EXPR, PREINCREMENT_EXPR,
808 POSTDECREMENT_EXPR, POSTINCREMENT_EXPR, CLEANUP_POINT_EXPR, EXPR_STMT,
809 STATEMENT_LIST, and RETURN_EXPR. For non-pure non-const functions
810 return 0 only when not OEP_LEXICOGRAPHIC.
811 (fold_build_cleanup_point_expr): Use the expression
812 location when building CLEANUP_POINT_EXPR.
813 * tree-core.h (enum operand_equal_flag): Add OEP_LEXICOGRAPHIC.
814 * tree.c (add_expr): Handle error_mark_node.
816 2017-01-20 Martin Liska <mliska@suse.cz>
819 * tree-profile.c (init_ic_make_global_vars): Do not call
821 (gimple_init_gcov_profiler): Likewise.
823 2017-01-20 Martin Liska <mliska@suse.cz>
826 * cgraph.h (maybe_create_reference): Remove argument and
828 * cgraphclones.c (cgraph_node::create_virtual_clone): Remove one
830 * ipa-cp.c (create_specialized_node): Likewise.
831 * symtab.c (symtab_node::maybe_create_reference): Handle
832 VAR_DECLs and ADDR_EXPRs and select ipa_ref_use type.
834 2017-01-20 Martin Liska <mliska@suse.cz>
836 * read-rtl-function.c (function_reader::create_function): Use
837 build_decl instread of build_decl_stat.
839 2017-01-20 Andrew Senkevich <andrew.senkevich@intel.com>
841 * config/i386/avx512bwintrin.h: Add k-mask registers shift intrinsics.
842 * config/i386/avx512dqintrin.h: Ditto.
843 * config/i386/avx512fintrin.h: Ditto.
844 * config/i386/i386-builtin-types.def: Add new types.
845 * gcc/config/i386/i386.c: Handle new types.
846 * config/i386/i386-builtin.def (__builtin_ia32_kshiftliqi)
847 (__builtin_ia32_kshiftlihi, __builtin_ia32_kshiftlisi)
848 (__builtin_ia32_kshiftlidi, __builtin_ia32_kshiftriqi)
849 (__builtin_ia32_kshiftrihi, __builtin_ia32_kshiftrisi)
850 (__builtin_ia32_kshiftridi): New.
851 * config/i386/sse.md (k<code><mode>): Rename *k<code><mode>.
853 2017-01-19 Segher Boessenkool <segher@kernel.crashing.org>
857 * config/rs6000/rs6000.c (TARGET_STACK_PROTECT_GUARD): Unconditionally
858 define to rs6000_init_stack_protect_guard.
859 (rs6000_init_stack_protect_guard): New function.
861 2017-01-19 Matthew Fortune <matthew.fortune@imgtec.com>
862 Yunqiang Su <yunqiang.su@imgtec.com>
864 * config.gcc (supported_defaults): Add madd4.
865 (with_madd4): Add validation.
866 (all_defaults): Add madd4.
867 * config/mips/mips.opt (mmadd4): New option.
868 * gcc/config/mips/mips.h (OPTION_DEFAULT_SPECS): Add a default for
870 (TARGET_CPU_CPP_BUILTINS): Add builtin_define for
872 (ISA_HAS_UNFUSED_MADD4): Gate with mips_madd4.
873 (ISA_HAS_FUSED_MADD4): Likewise.
874 * gcc/doc/invoke.texi (-mmadd4): Document the new option.
875 * gcc/doc/install.texi (--with-madd4): Document the new option.
877 2017-01-19 Jiong Wang <jiong.wang@arm.com>
879 * config/aarch64/aarch64-builtins.c (enum aarch64_builtins): New
880 entries for AARCH64_PAUTH_BUILTIN_XPACLRI,
881 AARCH64_PAUTH_BUILTIN_PACIA1716, AARCH64_PAUTH_BUILTIN_AUTIA1716.
882 (aarch64_init_pauth_hint_builtins): New.
883 (aarch64_init_builtins): Call aarch64_init_pauth_hint_builtins.
884 (aarch64_expand_builtin): Expand new builtins.
886 2017-01-19 Jiong Wang <jiong.wang@arm.com>
888 * reg-notes.def (CFA_TOGGLE_RA_MANGLE): New reg-note.
889 * combine-stack-adj.c (no_unhandled_cfa): Handle
890 REG_CFA_TOGGLE_RA_MANGLE.
891 * dwarf2cfi.c (dwarf2out_frame_debug): Handle REG_CFA_TOGGLE_RA_MANGLE.
892 * config/aarch64/aarch64.c (aarch64_expand_prologue): Generates DWARF
893 info for return address signing.
894 (aarch64_expand_epilogue): Likewise.
896 2017-01-19 Jiong Wang <jiong.wang@arm.com>
898 * config/aarch64/aarch64-opts.h (aarch64_function_type): New enum.
899 * config/aarch64/aarch64-protos.h
900 (aarch64_return_address_signing_enabled): New declaration.
901 * config/aarch64/aarch64.c (aarch64_return_address_signing_enabled):
903 (aarch64_expand_prologue): Sign return address before it's pushed onto
905 (aarch64_expand_epilogue): Authenticate return address fetched from
907 (aarch64_override_options): Sanity check for ILP32 and ISA level.
908 (aarch64_attributes): New function attributes for "sign-return-address".
909 * config/aarch64/aarch64.md (UNSPEC_AUTI1716, UNSPEC_AUTISP,
910 UNSPEC_PACI1716, UNSPEC_PACISP, UNSPEC_XPACLRI): New unspecs.
911 ("*do_return"): Generate combined instructions according to key index.
912 ("<pauth_mnem_prefix>sp", "<pauth_mnem_prefix1716", "xpaclri"): New.
913 * config/aarch64/iterators.md (PAUTH_LR_SP, PAUTH_17_16): New integer
915 (pauth_mnem_prefix, pauth_hint_num_a): New integer attributes.
916 * config/aarch64/aarch64.opt (msign-return-address=): New.
917 * doc/extend.texi (AArch64 Function Attributes): Documents
918 "sign-return-address=".
919 * doc/invoke.texi (AArch64 Options): Documents "-msign-return-address=".
921 2017-01-19 Matthew Fortune <matthew.fortune@imgtec.com>
923 * doc/invoke.texi: Add missing -mlxc1-sxc1 options to
924 overall option summary.
926 2017-01-19 Jiong Wang <jiong.wang@arm.com>
928 * config/aarch64/aarch64-arches.def: New entry for "armv8.3-a".
929 * config/aarch64/aarch64.h (AARCH64_FL_V8_3, AARCH64_FL_FOR_ARCH8_3,
930 AARCH64_ISA_V8_3, TARGET_ARMV8_3): New.
931 * doc/invoke.texi (AArch64 Options): Document "armv8.3-a".
933 2017-01-19 Michael Meissner <meissner@linux.vnet.ibm.com>
935 * config/rs6000/rs6000-cpus.def (ISA_3_0_MASKS_SERVER): Enable
936 -mpower9-minmax by default for -mcpu=power9.
937 (ISA_3_MASKS_IEEE): Require -mvsx-small-integer to enable IEEE
938 128-bit floating point.
940 2017-01-20 Alan Modra <amodra@gmail.com>
942 * config/rs6000/rs6000.md (cmpstrnsi, cmpstrsi): Fail if
945 2017-01-20 Alan Modra <amodra@gmail.com>
948 * config/rs6000/rs6000.c (expand_strn_compare): Get the asm name
949 for strcmp and strncmp from corresponding builtin decl.
951 2017-01-19 Uros Bizjak <ubizjak@gmail.com>
953 * config.gcc (x86_64-*-rtems*): Use i386/rtemself.h
954 instead of i386/rtems-64.h.
955 * config/i386/rtems-64.h: Remove.
957 2017-01-19 Uros Bizjak <ubizjak@gmail.com>
961 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
963 * config/i386/rtemself.h (LONG_DOUBLE_TYPE_SIZE): New define.
965 2017-01-19 Tamar Christina <tamar.christina@arm.com>
967 * config/aarch64/aarch64.c (aarch64_simd_gen_const_vector_dup):
968 Change int to HOST_WIDE_INT.
969 * config/aarch64/aarch64-protos.h
970 (aarch64_simd_gen_const_vector_dup): Likewise.
971 * config/aarch64/aarch64-simd.md: Add copysign<mode>3.
973 2017-01-19 David Malcolm <dmalcolm@redhat.com>
975 * langhooks-def.h (lhd_type_for_size): New decl.
976 (LANG_HOOKS_TYPE_FOR_SIZE): Define as lhd_type_for_size.
977 * langhooks.c (lhd_type_for_size): New function, taken from
980 2017-01-19 Pat Haugen <pthaugen@us.ibm.com>
982 * config/rs6000/power9.md (power9-alu): Remove 'cmp' type and add
983 define_bypass for CR latency.
984 (power9-cracked-alu): Update bypass latency and remove power9-branch.
985 (power9-alu2): Add define_bypass for CR latency.
987 (power9-mul): Update insn latency.
988 (power9-mul-compare): Update insn latency, bypass latency and remove
991 2016-01-19 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
993 * config/aarch64/aarch64-protos.h (aarch64_nopcrelative_literal_loads):
995 * config/aarch64/aarch64.md
996 (aarch64_reload_movcp<GPF_TF:mode><P:mode>): Delete reference to
997 aarch64_nopcrelative_literal_loads.
998 (aarch64_reload_movcp<VALL:mode><P:mode>): Likewise.
1000 2017-01-19 Chenghua Xu <paul.hua.gm@gmail.com>
1002 * config/mips/mips.h (ISA_HAS_FUSED_MADD4): Enable for
1004 (ISA_HAS_UNFUSED_MADD4): Exclude TARGET_LOONGSON_3A.
1006 2017-01-19 Doug Gilmore <doug.gilmore@imgtec.com>
1009 * config.gcc (supported_defaults): Add lxc1-sxc1.
1010 (with_lxc1_sxc1): Add validation.
1011 (all_defaults): Add lxc1-sxc1.
1012 * config/mips/mips.opt (mlxc1-sxc1): New option.
1013 * gcc/config/mips/mips.h (OPTION_DEFAULT_SPECS): Add a default for
1015 (TARGET_CPU_CPP_BUILTINS): Add builtin_define for
1016 __mips_no_lxc1_sxc1.
1017 (ISA_HAS_LXC1_SXC1): Gate with mips_lxc1_sxc1.
1018 * gcc/doc/invoke.texi (-mlxc1-sxc1): Document the new option.
1019 * doc/install.texi (--with-lxc1-sxc1): Document the new option.
1021 2017-01-19 Richard Biener <rguenther@suse.de>
1023 PR tree-optimization/72488
1024 * tree-ssa-sccvn.c (run_scc_vn): When we abort the VN make
1025 sure to restore SSA info.
1026 * tree-ssa.c (verify_ssa): Verify SSA info is not shared.
1028 2017-01-19 Richard Earnshaw <rearnsha@arm.com>
1030 PR rtl-optimization/79121
1031 * expr.c (expand_expr_real_2, case LSHIFT_EXPR): Look at the signedness
1032 of the inner type when shifting an extended value.
1034 2017-01-17 Jan Hubicka <hubicka@ucw.cz>
1037 * symtab.c (symtab_node::equal_address_to): Fix comparing of
1038 interposable aliases.
1040 2017-01-18 Peter Bergner <bergner@vnet.ibm.com>
1043 * config/rs6000/spe.md (mov_si<mode>_e500_subreg0): Fix constraints.
1044 Use the evmergelohi instruction.
1045 (mov_si<mode>_e500_subreg4_2_le): Likewise.
1046 (mov_sitf_e500_subreg8_2_be): Likewise.
1047 (mov_sitf_e500_subreg12_2_le): Likewise.
1048 (mov_si<mode>_e500_subreg0_2_le): Fix constraints.
1049 (mov_si<mode>_e500_subreg4_2_be): Likewise.
1050 (mov_sitf_e500_subreg8_2_le): Likewise.
1051 (mov_sitf_e500_subreg12_2_be): Likewise.
1053 2017-01-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1055 * config/rs6000/altivec.md (altivec_vbpermq): Change "type"
1056 attribute from vecsimple to vecperm.
1057 (altivec_vbpermq2): Likewise.
1059 2017-01-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1062 * config/rs6000/altivec.h: Fix typo of vec_cntlz to vec_cnttz.
1064 2017-01-18 Aaron Sawdey <acsawdey@linux.vnet.ibm.com>
1065 * config/rs6000/rs6000-protos.h (expand_strn_compare): Add arg.
1066 * config/rs6000/rs6000.c (expand_strn_compare): Add ability to expand
1067 strcmp. Fix bug where comparison didn't stop with zero byte. Fix
1068 case where N arg is SIZE_MAX.
1069 * config/rs6000/rs6000.md (cmpstrnsi): Args to expand_strn_compare.
1070 (cmpstrsi): Add pattern.
1072 2017-01-18 Michael Meissner <meissner@linux.vnet.ibm.com>
1074 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
1075 __builtin_vec_revb builtins.
1076 * config/rs6000/rs6000-builtins.def (P9V_BUILTIN_XXBRQ_V16QI): Add
1077 built-in functions to support generation of the ISA 3.0 XXBR<x>
1078 vector byte reverse instructions.
1079 (P9V_BUILTIN_XXBRQ_V1TI): Likewise.
1080 (P9V_BUILTIN_XXBRD_V2DI): Likewise.
1081 (P9V_BUILTIN_XXBRD_V2DF): Likewise.
1082 (P9V_BUILTIN_XXBGW_V4SI): Likewise.
1083 (P9V_BUILTIN_XXBGW_V4SF): Likewise.
1084 (P9V_BUILTIN_XXBGH_V8HI): Likewise.
1085 (P9V_BUILTIN_VEC_REVB): Likewise.
1086 * config/rs6000/vsx.md (p9_xxbrq_v1ti): New insns/expanders to
1087 generate the ISA 3.0 XXBR<x> vector byte reverse instructions.
1088 (p9_xxbrq_v16qi): Likewise.
1089 (p9_xxbrd_<mode>, VSX_D iterator): Likewise.
1090 (p9_xxbrw_<mode>, VSX_W iterator): Likewise.
1091 (p9_xxbrh_v8hi): Likewise.
1092 * config/rs6000/altivec.h (vec_revb): Define if ISA 3.0.
1093 * doc/extend.texi (RS/6000 Altivec Built-ins): Document the
1094 vec_revb built-in functions.
1096 2017-01-18 Uros Bizjak <ubizjak@gmail.com>
1098 PR rtl-optimization/78952
1099 * config/i386/i386.md (any_extract): New code iterator.
1100 (*insvqi_2): Use any_extract for source operand.
1101 (*insvqi_3): Use any_shiftrt for source operand.
1103 2017-01-18 Wilco Dijkstra <wdijkstr@arm.com>
1105 * config/aarch64/aarch64.c (aarch64_sched_adjust_priority)
1107 (TARGET_SCHED_ADJUST_PRIORITY): Define target hook.
1109 2017-01-18 Matthias Klose <doko@ubuntu.com>
1111 * doc/install.texi: Allow default for --with-target-bdw-gc-include.
1113 2016-01-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1115 * config/rs6000/altivec.h (vec_bperm): Change #define.
1116 * config/rs6000/altivec.md (UNSPEC_VBPERMD): New enum constant.
1117 (altivec_vbpermq2): New define_insn.
1118 (altivec_vbpermd): Likewise.
1119 * config/rs6000/rs6000-builtin.def (VBPERMQ2): New monomorphic
1121 (VBPERMD): Likewise.
1122 (VBPERM): New polymorphic function interface.
1123 * config/rs6000/r6000-c.c (altivec_overloaded_builtins_table):
1124 Add entries for P9V_BUILTIN_VEC_VBPERM.
1125 * doc/extend.texi: Add interfaces for vec_bperm.
1127 2017-01-18 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
1129 * config/s390/s390-c.c (s390_expand_overloaded_builtin): Downcase
1130 first letter of error messages.
1131 (s390_resolve_overloaded_builtin): Likewise.
1132 * config/s390/s390.c (s390_expand_builtin): Likewise.
1133 (s390_invalid_arg_for_unprototyped_fn): Likewise.
1134 (s390_valid_target_attribute_inner_p): Likewise.
1135 * config/s390/s390.md ("tabort"): Likewise.
1137 2017-01-18 Toma Tabacu <toma.tabacu@imgtec.com>
1139 * config/mips/mips.h (ISA_HAS_DIV3): Remove unused macro.
1140 (ISA_AVOID_DIV_HILO): New macro.
1141 (ISA_HAS_DIV): Use new ISA_AVOID_DIV_HILO macro.
1142 (ISA_HAS_DDIV): Likewise.
1144 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
1146 * doc/invoke.texi (fabi-version): Correct number of occurrences.
1148 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
1150 * doc/invoke.texi (fabi-version): Spelling fix.
1152 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
1155 * doc/invoke.texi (fabi-version): Mention mangling fix for
1158 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
1161 * doc/invoke.texi (fabi-version): Document discriminator mangling.
1163 2017-01-17 Segher Boessenkool <segher@kernel.crashing.org>
1166 * config/rs6000/rs6000-opts.h (stack_protector_guard): New enum.
1167 * config/rs6000/rs6000.c (rs6000_option_override_internal): Handle
1169 * config/rs6000/rs6000.md (stack_protect_set): Handle the new more
1171 (stack_protect_test): Ditto.
1172 * config/rs6000/rs6000.opt (mstack-protector-guard=,
1173 mstack-protector-guard-reg=, mstack-protector-guard-offset=): New
1175 * doc/invoke.texi (Option Summary) [RS/6000 and PowerPC Options]:
1176 Add -mstack-protector-guard=, -mstack-protector-guard-reg=, and
1177 -mstack-protector-guard-offset=.
1178 (RS/6000 and PowerPC Options): Ditto.
1180 2017-01-17 Uros Bizjak <ubizjak@gmail.com>
1182 * config/i386/i386.h (MASK_CLASS_P): New define.
1183 * config/i386/i386.c (inline_secondary_memory_needed): Ensure that
1184 there are no registers from different register sets also when
1185 mask registers are used. Update function comment.
1186 * config/i386/i386.md (*movsi_internal): Split (*k/*krm) alternative
1187 to (*k/*r) and (*k/*km) alternatives.
1189 2017-01-17 Wilco Dijkstra <wdijkstr@arm.com>
1191 * config/aarch64/aarch64.md (eh_return): Remove pattern and splitter.
1192 * config/aarch64/aarch64.h (AARCH64_EH_STACKADJ_REGNUM): Remove.
1193 (EH_RETURN_HANDLER_RTX): New define.
1194 * config/aarch64/aarch64.c (aarch64_frame_pointer_required):
1195 Force frame pointer in EH return functions.
1196 (aarch64_expand_epilogue): Add barrier for eh_return.
1197 (aarch64_final_eh_return_addr): Remove.
1198 (aarch64_eh_return_handler_rtx): New function.
1199 * config/aarch64/aarch64-protos.h (aarch64_final_eh_return_addr):
1201 (aarch64_eh_return_handler_rtx): New prototype.
1203 2017-01-17 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1205 * config/rs6000/altivec.h (vec_rlmi): New #define.
1206 (vec_vrlnm): Likewise.
1207 (vec_rlnm): Likewise.
1208 * config/rs6000/altivec.md (UNSPEC_VRLMI): New UNSPEC enum value.
1209 (UNSPEC_VRLNM): Likewise.
1210 (VIlong): New mode iterator.
1211 (altivec_vrl<VI_char>mi): New define_insn.
1212 (altivec_vrl<VI_char>nm): Likewise.
1213 * config/rs6000/rs6000-builtin.def (VRLWNM): New monomorphic
1216 (RLNM): New polymorphic function entry.
1217 (VRLWMI): New monomorphic function entry.
1219 (RLMI): New polymorphic function entry.
1220 * config/rs6000/r6000-c.c (altivec_overloaded_builtin_table): Add
1221 new entries for P9V_BUILTIN_VEC_RLMI and P9V_BUILTIN_VEC_RLNM.
1222 * doc/extend.texi: Add description of vec_rlmi, vec_rlnm, and
1225 2017-01-17 Jakub Jelinek <jakub@redhat.com>
1228 * dwarf2out.c (field_byte_offset): Restore the
1229 PCC_BITFIELD_TYPE_MATTERS behavior for INTEGER_CST DECL_FIELD_OFFSET
1230 and DECL_FIELD_BIT_OFFSET. Use fold_build2 instead of build2 + fold.
1231 (analyze_variants_discr, gen_variant_part): Use fold_build2 instead
1234 2017-01-17 Eric Botcazou <ebotcazou@adacore.com>
1237 * config/aarch64/aarch64.c (TARGET_CUSTOM_FUNCTION_DESCRIPTORS): Define
1239 2017-01-17 Jakub Jelinek <jakub@redhat.com>
1242 * dwarf2out.c (add_data_member_location_attribute): For constant
1243 offset bitfield emit for -gdwarf-5 DW_AT_data_bit_offset attribute
1244 instead of DW_AT_data_member_location, DW_AT_bit_offset and
1245 DW_AT_byte_size attributes.
1247 2017-01-17 Eric Botcazou <ebotcazou@adacore.com>
1249 * config/rs6000/rs6000.c (rs6000_emit_move): Also use a TOC reference
1250 after forcing to constant memory when the code model is medium.
1252 2017-01-17 Julia Koval <julia.koval@intel.com>
1255 * config/i386/avx512fintrin.h
1256 (_mm512_i32gather_ps): Change __addr type to void const*.
1257 (_mm512_mask_i32gather_ps): Ditto.
1258 (_mm512_i32gather_pd): Ditto.
1259 (_mm512_mask_i32gather_pd): Ditto.
1260 (_mm512_i64gather_ps): Ditto.
1261 (_mm512_mask_i64gather_ps): Ditto.
1262 (_mm512_i64gather_pd): Ditto.
1263 (_mm512_mask_i64gather_pd): Ditto.
1264 (_mm512_i32gather_epi32): Ditto.
1265 (_mm512_mask_i32gather_epi32): Ditto.
1266 (_mm512_i32gather_epi64): Ditto.
1267 (_mm512_mask_i32gather_epi64): Ditto.
1268 (_mm512_i64gather_epi32): Ditto.
1269 (_mm512_mask_i64gather_epi32): Ditto.
1270 (_mm512_i64gather_epi64): Ditto.
1271 (_mm512_mask_i64gather_epi64): Ditto.
1272 (_mm512_i32scatter_ps): Change __addr type to void*.
1273 (_mm512_mask_i32scatter_ps): Ditto.
1274 (_mm512_i32scatter_pd): Ditto.
1275 (_mm512_mask_i32scatter_pd): Ditto.
1276 (_mm512_i64scatter_ps): Ditto.
1277 (_mm512_mask_i64scatter_ps): Ditto.
1278 (_mm512_i64scatter_pd): Ditto.
1279 (_mm512_mask_i64scatter_pd): Ditto.
1280 (_mm512_i32scatter_epi32): Ditto.
1281 (_mm512_mask_i32scatter_epi32): Ditto.
1282 (_mm512_i32scatter_epi64): Ditto.
1283 (_mm512_mask_i32scatter_epi64): Ditto.
1284 (_mm512_i64scatter_epi32): Ditto.
1285 (_mm512_mask_i64scatter_epi32): Ditto.
1286 (_mm512_i64scatter_epi64): Ditto.
1287 (_mm512_mask_i64scatter_epi64): Ditto.
1288 * config/i386/avx512pfintrin.h
1289 (_mm512_mask_prefetch_i32gather_pd): Change __addr type to void const*.
1290 (_mm512_mask_prefetch_i32gather_ps): Ditto.
1291 (_mm512_mask_prefetch_i64gather_pd): Ditto.
1292 (_mm512_mask_prefetch_i64gather_ps): Ditto.
1293 (_mm512_prefetch_i32scatter_pd): Change __addr type to void*.
1294 (_mm512_prefetch_i32scatter_ps): Ditto.
1295 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
1296 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
1297 (_mm512_prefetch_i64scatter_pd): Ditto.
1298 (_mm512_prefetch_i64scatter_ps): Ditto.
1299 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
1300 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
1301 * config/i386/avx512vlintrin.h
1302 (_mm256_mmask_i32gather_ps): Change __addr type to void const*.
1303 (_mm_mmask_i32gather_ps): Ditto.
1304 (_mm256_mmask_i32gather_pd): Ditto.
1305 (_mm_mmask_i32gather_pd): Ditto.
1306 (_mm256_mmask_i64gather_ps): Ditto.
1307 (_mm_mmask_i64gather_ps): Ditto.
1308 (_mm256_mmask_i64gather_pd): Ditto.
1309 (_mm_mmask_i64gather_pd): Ditto.
1310 (_mm256_mmask_i32gather_epi32): Ditto.
1311 (_mm_mmask_i32gather_epi32): Ditto.
1312 (_mm256_mmask_i32gather_epi64): Ditto.
1313 (_mm_mmask_i32gather_epi64): Ditto.
1314 (_mm256_mmask_i64gather_epi32): Ditto.
1315 (_mm_mmask_i64gather_epi32): Ditto.
1316 (_mm256_mmask_i64gather_epi64): Ditto.
1317 (_mm_mmask_i64gather_epi64): Ditto.
1318 (_mm256_i32scatter_ps): Change __addr type to void*.
1319 (_mm256_mask_i32scatter_ps): Ditto.
1320 (_mm_i32scatter_ps): Ditto.
1321 (_mm_mask_i32scatter_ps): Ditto.
1322 (_mm256_i32scatter_pd): Ditto.
1323 (_mm256_mask_i32scatter_pd): Ditto.
1324 (_mm_i32scatter_pd): Ditto.
1325 (_mm_mask_i32scatter_pd): Ditto.
1326 (_mm256_i64scatter_ps): Ditto.
1327 (_mm256_mask_i64scatter_ps): Ditto.
1328 (_mm_i64scatter_ps): Ditto.
1329 (_mm_mask_i64scatter_ps): Ditto.
1330 (_mm256_i64scatter_pd): Ditto.
1331 (_mm256_mask_i64scatter_pd): Ditto.
1332 (_mm_i64scatter_pd): Ditto.
1333 (_mm_mask_i64scatter_pd): Ditto.
1334 (_mm256_i32scatter_epi32): Ditto.
1335 (_mm256_mask_i32scatter_epi32): Ditto.
1336 (_mm_i32scatter_epi32): Ditto.
1337 (_mm_mask_i32scatter_epi32): Ditto.
1338 (_mm256_i32scatter_epi64): Ditto.
1339 (_mm256_mask_i32scatter_epi64): Ditto.
1340 (_mm_i32scatter_epi64): Ditto.
1341 (_mm_mask_i32scatter_epi64): Ditto.
1342 (_mm256_i64scatter_epi32): Ditto.
1343 (_mm256_mask_i64scatter_epi32): Ditto.
1344 (_mm_i64scatter_epi32): Ditto.
1345 (_mm_mask_i64scatter_epi32): Ditto.
1346 (_mm256_i64scatter_epi64): Ditto.
1347 (_mm256_mask_i64scatter_epi64): Ditto.
1348 (_mm_i64scatter_epi64): Ditto.
1349 (_mm_mask_i64scatter_epi64): Ditto.
1350 * config/i386/i386-builtin-types.def (V16SF_V16SF_PCFLOAT_V16SI_HI_INT)
1351 (V8DF_V8DF_PCDOUBLE_V8SI_QI_INT, V8SF_V8SF_PCFLOAT_V8DI_QI_INT)
1352 (V8DF_V8DF_PCDOUBLE_V8DI_QI_INT, V16SI_V16SI_PCINT_V16SI_HI_INT)
1353 (V8DI_V8DI_PCINT64_V8SI_QI_INT, V8SI_V8SI_PCINT_V8DI_QI_INT)
1354 (V8DI_V8DI_PCINT64_V8DI_QI_INT, V2DF_V2DF_PCDOUBLE_V4SI_QI_INT)
1355 (V4DF_V4DF_PCDOUBLE_V4SI_QI_INT, V2DF_V2DF_PCDOUBLE_V2DI_QI_INT)
1356 (V4DF_V4DF_PCDOUBLE_V4DI_QI_INT, V4SF_V4SF_PCFLOAT_V4SI_QI_INT)
1357 (V8SF_V8SF_PCFLOAT_V8SI_QI_INT, V4SF_V4SF_PCFLOAT_V2DI_QI_INT)
1358 (V4SF_V4SF_PCFLOAT_V4DI_QI_INT, V2DI_V2DI_PCINT64_V4SI_QI_INT)
1359 (V4DI_V4DI_PCINT64_V4SI_QI_INT, V2DI_V2DI_PCINT64_V2DI_QI_INT)
1360 (V4DI_V4DI_PCINT64_V4DI_QI_INT, V4SI_V4SI_PCINT_V4SI_QI_INT)
1361 (V8SI_V8SI_PCINT_V8SI_QI_INT, V4SI_V4SI_PCINT_V2DI_QI_INT)
1362 (V4SI_V4SI_PCINT_V4DI_QI_INT, VOID_PFLOAT_HI_V16SI_V16SF_INT)
1363 (VOID_PFLOAT_QI_V8SI_V8SF_INT, VOID_PFLOAT_QI_V4SI_V4SF_INT)
1364 (VOID_PDOUBLE_QI_V8SI_V8DF_INT, VOID_PDOUBLE_QI_V4SI_V4DF_INT)
1365 (VOID_PDOUBLE_QI_V4SI_V2DF_INT, VOID_PFLOAT_QI_V8DI_V8SF_INT)
1366 (VOID_PFLOAT_QI_V4DI_V4SF_INT, VOID_PFLOAT_QI_V2DI_V4SF_INT)
1367 (VOID_PDOUBLE_QI_V8DI_V8DF_INT, VOID_PDOUBLE_QI_V4DI_V4DF_INT)
1368 (VOID_PDOUBLE_QI_V2DI_V2DF_INT, VOID_PINT_HI_V16SI_V16SI_INT)
1369 (VOID_PINT_QI_V8SI_V8SI_INT, VOID_PINT_QI_V4SI_V4SI_INT)
1370 (VOID_PLONGLONG_QI_V8SI_V8DI_INT, VOID_PLONGLONG_QI_V4SI_V4DI_INT)
1371 (VOID_PLONGLONG_QI_V4SI_V2DI_INT, VOID_PINT_QI_V8DI_V8SI_INT)
1372 (VOID_PINT_QI_V4DI_V4SI_INT, VOID_PINT_QI_V2DI_V4SI_INT)
1373 (VOID_PLONGLONG_QI_V8DI_V8DI_INT, VOID_QI_V8SI_PCINT64_INT_INT)
1374 (VOID_PLONGLONG_QI_V4DI_V4DI_INT, VOID_PLONGLONG_QI_V2DI_V2DI_INT)
1375 (VOID_HI_V16SI_PCINT_INT_INT, VOID_QI_V8DI_PCINT64_INT_INT)
1376 (VOID_QI_V8DI_PCINT_INT_INT): Remove.
1377 (V16SF_V16SF_PCVOID_V16SI_HI_INT, V8DF_V8DF_PCVOID_V8SI_QI_INT)
1378 (V8SF_V8SF_PCVOID_V8DI_QI_INT, V8DF_V8DF_PCVOID_V8DI_QI_INT)
1379 (V16SI_V16SI_PCVOID_V16SI_HI_INT, V8DI_V8DI_PCVOID_V8SI_QI_INT)
1380 (V8SI_V8SI_PCVOID_V8DI_QI_INT, V8DI_V8DI_PCVOID_V8DI_QI_INT)
1381 (VOID_PVOID_HI_V16SI_V16SF_INT, VOID_PVOID_QI_V8SI_V8DF_INT)
1382 (VOID_PVOID_QI_V8DI_V8SF_INT, VOID_PVOID_QI_V8DI_V8DF_INT)
1383 (VOID_PVOID_HI_V16SI_V16SI_INT, VOID_PVOID_QI_V8SI_V8DI_INT)
1384 (VOID_PVOID_QI_V8DI_V8SI_INT, VOID_PVOID_QI_V8DI_V8DI_INT)
1385 (V2DF_V2DF_PCVOID_V4SI_QI_INT, V4DF_V4DF_PCVOID_V4SI_QI_INT)
1386 (V2DF_V2DF_PCVOID_V2DI_QI_INT, V4DF_V4DF_PCVOID_V4DI_QI_INT
1387 (V4SF_V4SF_PCVOID_V4SI_QI_INT, V8SF_V8SF_PCVOID_V8SI_QI_INT)
1388 (V4SF_V4SF_PCVOID_V2DI_QI_INT, V4SF_V4SF_PCVOID_V4DI_QI_INT)
1389 (V2DI_V2DI_PCVOID_V4SI_QI_INT, V4DI_V4DI_PCVOID_V4SI_QI_INT)
1390 (V2DI_V2DI_PCVOID_V2DI_QI_INT, V4DI_V4DI_PCVOID_V4DI_QI_INT)
1391 (V4SI_V4SI_PCVOID_V4SI_QI_INT, V8SI_V8SI_PCVOID_V8SI_QI_INT)
1392 (V4SI_V4SI_PCVOID_V2DI_QI_INT, V4SI_V4SI_PCVOID_V4DI_QI_INT)
1393 (VOID_PVOID_QI_V8SI_V8SF_INT, VOID_PVOID_QI_V4SI_V4SF_INT)
1394 (VOID_PVOID_QI_V4SI_V4DF_INT, VOID_PVOID_QI_V4SI_V2DF_INT)
1395 (VOID_PVOID_QI_V4DI_V4SF_INT, VOID_PVOID_QI_V2DI_V4SF_INT)
1396 (VOID_PVOID_QI_V4DI_V4DF_INT, VOID_PVOID_QI_V2DI_V2DF_INT)
1397 (VOID_PVOID_QI_V8SI_V8SI_INT, VOID_PVOID_QI_V4SI_V4SI_INT)
1398 (VOID_PVOID_QI_V4SI_V4DI_INT, VOID_PVOID_QI_V4SI_V2DI_INT)
1399 (VOID_PVOID_QI_V4DI_V4SI_INT, VOID_PVOID_QI_V2DI_V4SI_INT)
1400 (VOID_PVOID_QI_V4DI_V4DI_INT, VOID_PVOID_QI_V2DI_V2DI_INT)
1401 (VOID_QI_V8SI_PCVOID_INT_INT, VOID_HI_V16SI_PCVOID_INT_INT)
1402 (VOID_QI_V8DI_PCVOID_INT_INT): Add.
1403 * config/i386/i386.c (ix86_init_mmx_sse_builtins): Adjust builtin
1404 definitions accordingly.
1406 2017-01-17 Kito Cheng <kito.cheng@gmail.com>
1407 Kuan-Lin Chen <kuanlinchentw@gmail.com>
1410 * internal-fn.c (expand_mul_overflow): Use convert_modes instead of
1413 2017-01-17 Vladimir Makarov <vmakarov@redhat.com>
1416 * ira-conflicts.c (ira_build_conflicts): Update total conflict
1417 hard regs for inner regno.
1419 2017-01-17 Martin Liska <mliska@suse.cz>
1422 * ipa-polymorphic-call.c (contains_type_p): Fix wrong
1423 assumption and add comment.
1425 2017-01-17 Nathan Sidwell <nathan@acm.org>
1427 * ipa-visibility.c (localize_node): New function, broken out of ...
1428 (function_and_variable_visibility): ... here. Call it.
1430 2017-01-17 Jan Hubicka <hubicka@ucw.cz>
1433 * tree-ssa-threadupdate.c (remove_ctrl_stmt_and_useless_edges):
1434 correctly set frequency of oudgoing edge.
1435 (duplicate_thread_path): Fix profile updating.
1437 2017-01-17 Jakub Jelinek <jakub@redhat.com>
1440 * configure.ac: Add GCC_BASE_VER.
1441 * Makefile.in (version): Use @get_gcc_base_ver@ instead of cat to get
1442 version from BASE-VER file.
1443 (CFLAGS-gcc.o): Add -DBASEVER=$(BASEVER_s).
1444 (gcc.o): Depend on $(BASEVER).
1445 * common.opt (dumpfullversion): New option.
1446 * gcc.c (driver_handle_option): Handle OPT_dumpfullversion.
1447 * doc/invoke.texi: Document -dumpfullversion.
1448 * doc/install.texi: Document --with-gcc-major-version-only.
1449 * configure: Regenerated.
1451 2017-01-17 Richard Biener <rguenther@suse.de>
1453 PR tree-optimization/71433
1454 * tree-vrp.c (register_new_assert_for): Merge same asserts
1455 on all incoming edges.
1456 (process_assert_insertions_for): Handle insertions at the
1459 2017-01-17 Gerald Pfeifer <gerald@pfeifer.com>
1461 * config/i386/cygwin.h (LIBGCJ_SONAME): Remove.
1462 * config/i386/mingw32.h (LIBGCJ_SONAME): Remove.
1464 2017-01-17 Kaz Kojima <kkojima@gcc.gnu.org>
1467 * config/sh/sh.md (cmpeqsi_t+1): Call copy_rtx to avoid invalid
1470 2017-01-17 Alan Modra <amodra@gmail.com>
1473 * config/rs6000/rs6000.md (elf_high, elf_low): Disable when pic.
1474 * config/rs6000/rs6000.c (rs6000_emit_allocate_stack): Don't allow
1475 symbolic stack limit when pic.
1477 2017-01-16 Martin Sebor <msebor@redhat.com>
1479 PR tree-optimization/78608
1480 * gimple-ssa-sprintf.c (tree_digits): Avoid negating TYPE_MIN.
1482 2017-01-16 Jeff Law <law@redhat.com>
1485 2016-12-02 Tadek Kijkowski <tkijkowski@gmail.com>
1486 * Makefile.in (PREPROCESSOR_DEFINES): Add a level of indirection
1487 for several include directories that may be relative to sysroot.
1488 * config/i386/x-mingw32 (gplus_includedir): Define.
1489 (gplus_tool_includedir, gplus_backward_include_dir): Likewise.
1490 (native_system_includedir): Likewise.
1491 * config/i386/mingw32.h (STANDARD_STARTFILE_PREFIX_1): Do not
1492 override if TARGET_SYSTEM_ROOT is defined.
1493 (NATIVE_SYSTEM_HEADER_DIR): Likewise.
1495 PR tree-optimization/79090
1496 PR tree-optimization/33562
1497 PR tree-optimization/61912
1498 PR tree-optimization/77485
1499 * tree-ssa-dse.c (compute_trims): Accept STMT argument. Dump STMT
1500 and computed trims into the dump file.
1502 2017-01-17 Uros Bizjak <ubizjak@gmail.com>
1504 * config/i386/i386.h (LIMIT_RELOAD_CLASS): Remove.
1506 2017-01-16 Jakub Jelinek <jakub@redhat.com>
1509 * gimplify.c (gimplify_init_constructor): If want_value and
1510 object == lhs, unshare lhs to avoid invalid tree sharing. Formatting
1514 * loop-doloop.c (doloop_modify): Call unshare_all_rtl_in_chain on
1515 sequence. Formatting fixes.
1516 (doloop_optimize): Formatting fixes.
1519 * gcc.c (debug_level_greater_than_spec_func): New function.
1520 (static_spec_functions): Add debug-level-gt spec function.
1521 (ASM_DEBUG_SPEC, cpp_options): Use %:debug-level-gt(0) instead of
1523 * config/darwin.h (DSYMUTIL_SPEC, ASM_DEBUG_SPEC): Likewise.
1524 * config/darwin9.h (DSYMUTIL_SPEC, ASM_DEBUG_SPEC): Likewise.
1525 * common.opt (g, gcoff, gdwarf, gdwarf-, ggdb, gno-pubnames,
1526 gpubnames, ggnu-pubnames, gno-record-gcc-switches,
1527 grecord-gcc-switches, gno-strict-dwarf, gstrict-dwarf, gstabs,
1528 gstabs+, gtoggle, gvms, gxcoff, gxcoff+): Add Driver flag.
1530 2017-01-16 Uros Bizjak <ubizjak@gmail.com>
1532 * config/i386/i386.h (HARD_REGNO_CALLER_SAVE_MODE): Apply HImode and
1533 QImode fixups to general and mask registers only.
1535 2017-01-16 Carl Love <cel@us.ibm.com>
1537 * config/rs6000/rs6000-c (altivec_overloaded_builtins): Add support
1538 for built-in functions
1539 vector signed char vec_nabs (vector signed char)
1540 vector signed short vec_nabs (vector signed short)
1541 vector signed int vec_nabs (vector signed int)
1542 vector signed long long vec_nabs (vector signed long long)
1543 vector float vec_nabs (vector float)
1544 vector double vec_nabs (vector double)
1545 * config/rs6000/rs6000-builtin.def: Add definitions for NABS functions
1547 * config/rs6000/altivec.md: New define_expand nabs<mode>2 types
1548 * config/rs6000/altivec.h: New define for vec_nabs built-in function.
1549 * doc/extend.texi: Update the documentation file for the new built-in
1552 2017-01-16 Martin Sebor <msebor@redhat.com>
1554 * gimple-ssa-sprintf.c (format_directive): Correct a typo in a warning
1557 2017-01-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1559 * config/rs6000/rs6000.c (rtx_is_swappable_p): Change
1560 UNSPEC_VSX__XXSPLTD to require special splat handling.
1562 2017-01-16 David Malcolm <dmalcolm@redhat.com>
1565 * system.h: Poison strndup.
1567 2017-01-16 Alan Modra <amodra@gmail.com>
1570 * config/rs6000/rs6000.c (rs6000_legitimate_combined_insn): Don't
1573 2017-01-16 Georg-Johann Lay <avr@gjlay.de>
1575 * config/avr/avr.h (BRANCH_COST) [reload_completed]: Increase by 4.
1577 2017-01-15 Uros Bizjak <ubizjak@gmail.com>
1579 * config/i386/i386.c (ix86_legitimate_combined_insn): Do not
1580 call recog here. Assert that INSN_CODE (insn) is non-negative.
1582 2017-01-15 Segher Boessenkool <segher@kernel.crashing.org>
1585 * cfgrtl.c (rtl_split_edge): Also patch jump insns that jump to the
1587 * haifa-sched.c (dump_insn_stream): Don't crash if there is a label
1588 in the currently scheduled RTL fragment.
1590 2017-01-15 Segher Boessenkool <segher@kernel.crashing.org>
1592 PR rtl-optimization/78751
1593 * ifcvt.c (find_cond_trap): If we generated a non-existing insn,
1596 2017-01-14 Jeff Law <law@redhat.com>
1598 PR tree-optimization/79090
1599 * tree-ssa-dse.c (valid_ao_ref_for_dse): Reject zero length and
1600 variable length stores.
1601 (compute_trims): Delete dead assignment to *trim_tail.
1602 (dse_dom_walker::dse_optimize_stmt): Optimize mem* calls with
1605 2017-01-14 Bernd Schmidt <bschmidt@redhat.com>
1607 PR rtl-optimization/78626
1608 PR rtl-optimization/78727
1609 * cprop.c (one_cprop_pass): Collect unconditional traps in the middle
1610 of a block, and split such blocks after everything else is finished.
1612 2017-01-14 Alan Modra <amodra@gmail.com>
1615 * combine.c (recog_for_combine_1): Set INSN_CODE before calling
1616 target legitimate_combined_insn.
1617 * config/rs6000/rs6000.c (TARGET_LEGITIMATE_COMBINED_INSN): Define.
1618 (rs6000_legitimate_combined_insn): New function.
1619 * config/rs6000/rs6000.md (UNSPEC_DOLOOP): Delete, and remove
1621 (ctr<mode>_internal3): Rename from *ctr<mode>_internal5.
1622 (ctr<mode>_internal4): Rename from *ctr<mode>_internal6.
1623 (ctr<mode>_internal1, ctr<mode>_internal2): Remove '*' from name.
1625 2017-01-14 Gerald Pfeifer <gerald@pfeifer.com>
1627 * doc/frontends.texi (G++ and GCC): Remove references to Java.
1629 2017-01-13 Jeff Law <law@redhat.com>
1631 PR tree-optimization/33562
1632 PR tree-optimization/61912
1633 PR tree-optimization/77485
1634 * tree-ssa-dse.c (delete_dead_call): Accept gsi rather than
1636 (delete_dead_assignment): Likewise.
1637 (dse_dom_walker::dse_optimize_stmt): Pass in the gsi rather than
1638 statement to delete_dead_call and delete_dead_assignment.
1640 2017-01-13 David Malcolm <dmalcolm@redhat.com>
1643 * substring-locations.c (format_warning_va): Strengthen case 1 so
1644 that both endpoints of the substring must be within the format
1645 range for just the substring to be printed.
1647 2017-01-13 Uros Bizjak <ubizjak@gmail.com>
1649 * config/i386/i386.opt (msgx): Use ix86_isa_flags2 variable.
1650 * config/i386/i386.c (ix86_target_string): Add missing options
1651 to isa_opts and reorder options by implied ISAs. Rename isa_opts2 to
1652 isa2_opts, ix86_flag_opts to flag2_opts, ix86_target_other to
1653 flags_other and ix86_target_other to flags2_other. Display unknown
1655 (ix86_valid_target_attribute_inner_p): Add missing options and
1656 reorder options by implied ISAs, as in ix86_target_string.
1658 2017-01-13 Richard Sandiford <richard.sandiford@arm.com>
1660 * hash-table.h (hash_table::too_empty_p): New function.
1661 (hash_table::expand): Use it.
1662 (hash_table::traverse): Likewise.
1663 (hash_table::empty_slot): Use sizeof (value_type) instead of
1664 sizeof (PTR) to convert bytes to elements. Shrink the table
1665 if the current size is excessive for the current number of
1668 2017-01-13 Richard Sandiford <richard.sandiford@arm.com>
1670 * ira-costs.c (record_reg_classes): Break from the inner loop
1671 early once alt_fail is known to be true. Update outer loop
1672 handling accordingly.
1674 2017-01-13 Jeff Law <law@redhat.com>
1676 * tree-ssa-dse.c (decrement_count): New function.
1677 (increment_start_addr, maybe_trim_memstar_call): Likewise.
1678 (dse_dom_walker::optimize_stmt): Call maybe_trim_memstar_call directly
1679 when we know the partially dead statement is a mem* function.
1681 PR tree-optimization/61912
1682 PR tree-optimization/77485
1683 * tree-ssa-dse.c: Include expr.h.
1684 (maybe_trim_constructor_store): New function.
1685 (maybe_trim_partially_dead_store): Call maybe_trim_constructor_store.
1687 PR tree-optimization/33562
1688 PR tree-optimization/61912
1689 PR tree-optimization/77485
1690 * doc/invoke.texi: Document new dse-max-object-size param.
1691 * params.def (PARM_DSE_MAX_OBJECT_SIZE): New PARAM.
1692 * tree-ssa-dse.c: Include params.h.
1693 (dse_store_status): New enum.
1694 (initialize_ao_ref_for_dse): New, partially extracted from
1696 (valid_ao_ref_for_dse, normalize_ref): New.
1697 (setup_live_bytes_from_ref, compute_trims): Likewise.
1698 (clear_bytes_written_by, maybe_trim_complex_store): Likewise.
1699 (maybe_trim_partially_dead_store): Likewise.
1700 (maybe_trim_complex_store): Likewise.
1701 (dse_classify_store): Renamed from dse_possibly_dead_store_p.
1702 Track what bytes live from the original store. Return tri-state
1703 for dead, partially dead or live.
1704 (dse_dom_walker): Add constructor, destructor and new private members.
1705 (delete_dead_call, delete_dead_assignment): New extracted from
1707 (dse_optimize_stmt): Make a member of dse_dom_walker.
1708 Use initialize_ao_ref_for_dse.
1710 PR tree-optimization/33562
1711 PR tree-optimization/61912
1712 PR tree-optimization/77485
1713 * sbitmap.h (bitmap_count_bits): Prototype.
1714 (bitmap_clear_range, bitmap_set_range): Likewise.
1715 * sbitmap.c (bitmap_clear_range): New function.
1716 (bitmap_set_range, sbitmap_popcount, bitmap_count_bits): Likewise.
1718 2017-01-13 Martin Liska <mliska@suse.cz>
1721 * function.c (set_cfun): Add new argument force.
1722 * function.h (set_cfun): Likewise.
1723 * ipa-inline-transform.c (inline_call): Use the function when
1724 strict alising from is dropped for function we inline to.
1726 2017-01-13 Richard Biener <rguenther@suse.de>
1728 * tree-pretty-print.c (dump_generic_node): Fix inverted condition
1729 for dumping GIMPLE INTEGER_CSTs.
1731 2017-01-13 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1733 * config/sol2.h (TARGET_OS_CPP_BUILTINS): Define __STDC_VERSION__
1734 to 201112L since C++17.
1736 2017-01-13 Maxim Ostapenko <m.ostapenko@samsung.com>
1739 * asan.c (asan_needs_odr_indicator_p): Don't emit ODR indicators
1740 if -fsanitize=kernel-address is present.
1742 2017-01-13 Richard Biener <rguenther@suse.de>
1744 * tree-pretty-print.c (dump_generic_node): Dump INTEGER_CSTs
1745 as _Literal ( type ) number in case usual suffixes do not
1746 preserve all information.
1748 2017-01-13 Richard Biener <rguenther@suse.de>
1750 PR tree-optimization/77283
1751 * gimple-ssa-split-paths.c: Include gimple-ssa.h, tree-phinodes.h
1752 and ssa-iterators.h.
1753 (is_feasible_trace): Implement a cost model based on joiner
1756 2017-01-12 Michael Meissner <meissner@linux.vnet.ibm.com>
1759 * config/rs6000/rs6000.md (FP_ISA3): Do not optimize converting
1760 char or short to __float128/_Float128 directly.
1762 2017-01-12 Martin Sebor <msebor@redhat.com>
1764 to -Wformat-overflow.
1765 * gimple-ssa-sprintf.c (pass_sprintf_length::gate): Adjust.
1766 (min_bytes_remaining): Same.
1767 (get_string_length): Same.
1768 (format_string): Same.
1769 (format_directive): Same.
1771 (pass_sprintf_length::handle_gimple_call): Same.
1773 2017-01-12 Jakub Jelinek <jakub@redhat.com>
1775 * gimple-ssa-sprintf.c (try_substitute_return_value): Remove
1776 info.nowrite calls with no lhs that can't throw. Return bool
1777 whether gsi_remove has been called or not.
1778 (pass_sprintf_length::handle_gimple_call): Return bool whether
1779 try_substitute_return_value called gsi_remove. Formatting fix.
1780 (pass_sprintf_length::execute): Don't use gsi_remove if
1781 handle_gimple_call returned true.
1784 * cfgrtl.c (rtl_tidy_fallthru_edge): For any_uncondjump_p that can't
1785 be removed due to side-effects, don't remove following barrier nor
1786 turn the successor edge into fallthru edge.
1788 2017-01-12 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1791 * config/rs6000/rs6000.c (insn_is_swappable_p): Mark
1792 element-reversing loads and stores as not swappable.
1794 2017-01-12 Nathan Sidwell <nathan@acm.org>
1795 Nicolai Stange <nicstange@gmail.com>
1797 * combine.c (try_combine): Don't ignore result of overlap checking
1798 loop. Combine overlap & asm check into single loop.
1800 2017-01-12 Richard Biener <rguenther@suse.de>
1802 * tree-pretty-print.c (dump_generic_node): Provide -gimple
1803 variant for MEM_REF. Sanitize INTEGER_CST for -gimple.
1805 2017-01-12 Richard Biener <rguenther@suse.de>
1807 * tree.c (initialize_tree_contains_struct): Make TS_OPTIMIZATION
1808 and TS_TARGET_OPTION directly derive from TS_BASE.
1809 * tree-core.h (tree_optimization_option): Derive from tree_base.
1810 (tree_target_option): Likewise.
1812 2017-01-11 Uros Bizjak <ubizjak@gmail.com>
1814 * config/i386/i386.c (memory_address_length): Increase len
1815 only when rip_relative_addr_p returns false.
1817 2017-01-11 Julia Koval <julia.koval@intel.com>
1819 * common/config/i386/i386-common.c (OPTION_MASK_ISA_SGX_UNSET): New.
1820 (OPTION_MASK_ISA_SGX_SET): New.
1821 (ix86_handle_option): Handle OPT_msgx.
1822 * config.gcc: Added sgxintrin.h.
1823 * config/i386/driver-i386.c (host_detect_local_cpu): Detect sgx.
1824 * config/i386/i386-c.c (ix86_target_macros_internal): Define __SGX__.
1825 * config/i386/i386.c (ix86_target_string): Add -msgx.
1827 (ix86_option_override_internal): Handle new options.
1828 (ix86_valid_target_attribute_inner_p): Add sgx.
1829 * config/i386/i386.h (TARGET_SGX, TARGET_SGX_P): New.
1830 * config/i386/i386.opt: Add msgx.
1831 * config/i386/sgxintrin.h: New file.
1832 * config/i386/x86intrin.h: Add sgxintrin.h.
1834 2017-01-11 Jakub Jelinek <jakub@redhat.com>
1837 * fold-const.c (maybe_nonzero_address): Return 1 for function
1839 (tree_single_nonzero_warnv_p): Don't handle function local objects
1843 * gcc.c (default_compilers): Don't add -o %g.s for -S -save-temps
1846 2017-01-11 David Malcolm <dmalcolm@redhat.com>
1849 * opts.c: Include "spellcheck.h"
1850 (struct string_fragment): New struct.
1851 (struct edit_distance_traits<const string_fragment &>): New
1853 (get_closest_sanitizer_option): New function.
1854 (parse_sanitizer_options): Offer suggestions for unrecognized arguments.
1856 2017-01-11 Jakub Jelinek <jakub@redhat.com>
1858 * dwarf2out.c (DWARF_COMPILE_UNIT_HEADER_SIZE): For DWARF5 decrease
1860 (DWARF_COMDAT_TYPE_UNIT_HEADER_SIZE): Always
1861 DWARF_COMPILE_UNIT_HEADER_SIZE plus 12.
1862 (DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE): Define.
1863 (calc_base_type_die_sizes): Use DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE
1864 for initial die_offset if dwarf_split_debug_info.
1865 (output_comp_unit): Use DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE for
1866 initial next_die_offset if dwo_id is non-NULL. Don't emit padding
1868 (output_skeleton_debug_sections): Formatting fix. Use
1869 DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE instead of
1870 DWARF_COMPILE_UNIT_HEADER_SIZE. Don't emit padding.
1872 2017-01-11 Wilco Dijkstra <wdijkstr@arm.com>
1874 * config/arm/cortex-a53.md: Add bypasses for
1876 (cortex_a53_r2f): Only use for transfers.
1877 (cortex_a53_f2r): Likewise.
1878 (cortex_a53_r2f_cvt): Add reservation for conversions.
1879 (cortex_a53_f2r_cvt): Likewise.
1881 2017-01-11 Tamar Christina <tamar.christina@arm.com>
1883 * config/arm/arm_neon.h: Add __artificial__ and gnu_inline
1884 to all inlined functions, change static to extern.
1886 2017-01-11 Christophe Lyon <christophe.lyon@linaro.org>
1889 * config/arm/arm.c (legitimize_pic_address): Handle reference to
1891 (arm_assemble_integer): Likewise.
1893 2017-01-11 Richard Earnshaw <rearnsha@arm.com>
1895 * config.gcc: Use new awk script to check CPU, FPU and architecture
1896 parameters for --with-... options.
1897 * config/arm/parsecpu.awk: New file
1898 * config/arm/arm-cpus.in: New file.
1899 * config/arm/arm-opts.h: Include arm-cpu.h instead of processing .def
1901 * config/arm/arm.c: Include arm-cpu-data.h instead of processing .def
1903 * config/arm/t-arm: Update dependency rules.
1904 * common/config/arm/arm-common.c: Include arm-cpu-cdata.h instead
1905 of processing .def files.
1906 * config/arm/genopt.sh: Deleted.
1907 * config/arm/gentune.sh: Deleted.
1908 * config/arm/arm-cores.def: Deleted.
1909 * config/arm/arm-arches.def: Deleted.
1910 * config/arm/arm-fpus.def: Deleted.
1911 * config/arm/arm-tune.md: Regenerated.
1912 * config/arm/arm-tables.opt: Regenerated.
1913 * config/arm/arm-cpu.h: New generated file.
1914 * config/arm/arm-cpu-data.h: New generated file.
1915 * config/arm/arm-cpu-cdata.h: New generated file.
1917 2017-01-11 Maxim Ostapenko <m.ostapenko@samsung.com>
1920 * lto-cgraph.c (lto_output_varpool_node): Pack dynamically_initialized
1922 (input_varpool_node): Unpack dynamically_initialized bit.
1924 2017-01-11 Eric Botcazou <ebotcazou@adacore.com>
1926 PR rtl-optimization/79032
1927 * lra-constraints.c (simplify_operand_subreg): In the MEM case, test
1928 the alignment of the adjusted memory reference against that of MODE,
1929 instead of the alignment of the original memory reference.
1931 2017-01-11 Martin Jambor <mjambor@suse.cz>
1933 * hsa.c (hsa_callable_function_p): Revert addition of DECL_ARTIFICIAL
1935 * ipa-hsa.c (process_hsa_functions): Only duplicate non-artificial
1936 decorated functions.
1938 2017-01-11 Richard Biener <rguenther@suse.de>
1940 * tree-vrp.c (evrp_dom_walker::before_dom_children): Also
1941 set range/nonnull info for PHI results. Do not set it on
1942 stmts marked for removal.
1944 2017-01-10 Eric Botcazou <ebotcazou@adacore.com>
1946 * expr.c (store_field): In the bitfield case, fetch the return value
1947 from the registers before applying a single big-endian adjustment.
1948 Always do a final load for a BLKmode value not larger than a word.
1950 2017-01-10 David Malcolm <dmalcolm@redhat.com>
1953 * input.c (selftest::test_accessing_ordinary_linemaps): Verify
1954 that we correctly handle column numbers greater than
1955 LINE_MAP_MAX_COLUMN_NUMBER.
1957 2017-01-10 Martin Sebor <msebor@redhat.com>
1960 * gimple-ssa-sprintf.c (get_destination_size): Call
1961 {init,fini}object_sizes.
1962 * tree-object-size.c (addr_object_size): Adjust.
1963 (pass_through_call): Adjust.
1964 (pass_object_sizes::execute): Adjust.
1965 * tree-object-size.h (fini_object_sizes): Declare.
1967 2017-01-10 Martin Sebor <msebor@redhat.com>
1969 PR tree-optimization/78775
1970 * builtins.c (get_size_range): Move...
1971 * calls.c: ...to here.
1972 (alloc_max_size): Accept zero argument.
1973 (operand_signed_p): Remove.
1974 (maybe_warn_alloc_args_overflow): Call get_size_range.
1975 * calls.h (get_size_range): Declare.
1977 2017-01-10 Joe Seymour <joe.s@somniumtech.com>
1979 * config/msp430/driver-msp430.c (msp430_mcu_data): Sync with data
1980 from TI's devices.csv file as of September 2016.
1981 * config/msp430/msp430.c (msp430_mcu_data): Likewise.
1983 2017-01-10 Sandra Loosemore <sandra@codesourcery.com>
1985 * doc/extend.texi: Tweak formatting to fix overfull hbox warnings.
1986 * doc/invoke.texi: Likewise.
1987 * doc/md.texi: Likewise.
1988 * doc/objc.texi: Likewise.
1990 2017-01-10 Joshua Conner <joshconner@google.com>
1992 * config/arm/fuchsia-elf.h: New file.
1993 * config/fuchsia.h: New file.
1994 * config.gcc (*-*-fuchsia*): Set native_system_header_dir.
1995 (aarch64*-*-fuchsia*, arm*-*-fuchsia*, x86_64-*-fuchsia*): Add to
1997 * config.host: (aarch64*-*-fuchsia*, arm*-*-fuchsia*): Add to hosts.
1999 2016-01-10 Richard Biener <rguenther@suse.de>
2001 PR tree-optimization/79034
2002 * tree-call-cdce.c (shrink_wrap_one_built_in_call_with_conds):
2003 Propagate out degenerate PHIs in the joiner.
2005 2017-01-10 Martin Liska <mliska@suse.cz>
2007 * ipa-icf.c (sort_sem_items_by_decl_uid): New function.
2008 (sort_congruence_classes_by_decl_uid): Likewise.
2009 (sort_congruence_class_groups_by_decl_uid): Likewise.
2010 (sem_item_optimizer::merge_classes): Sort class, groups in these
2011 classes and members in the groups by DECL_UID of declarations.
2012 This would make merge operations stable.
2014 2017-01-10 Martin Liska <mliska@suse.cz>
2016 * ipa-icf.c (sem_item_optimizer::sem_item_optimizer): Remove
2017 usage of m_classes_vec.
2018 (sem_item_optimizer::~sem_item_optimizer): Likewise.
2019 (sem_item_optimizer::get_group_by_hash): Likewise.
2020 (sem_item_optimizer::subdivide_classes_by_equality): Likewise.
2021 (sem_item_optimizer::subdivide_classes_by_sensitive_refs): Likewise.
2022 (sem_item_optimizer::verify_classes): Likewise.
2023 (sem_item_optimizer::process_cong_reduction): Likewise.
2024 (sem_item_optimizer::dump_cong_classes): Likewise.
2025 (sem_item_optimizer::merge_classes): Likewise.
2026 * ipa-icf.h (congruence_class_hash): Rename from
2027 congruence_class_group_hash. Remove declaration of m_classes_vec.
2029 2017-01-10 Andrew Senkevich <andrew.senkevich@intel.com>
2031 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512VPOPCNTDQ_SET,
2032 OPTION_MASK_ISA_AVX512VPOPCNTDQ_UNSET): New.
2033 * config.gcc: Add avx512vpopcntdqintrin.h.
2034 * config/i386/avx512vpopcntdqintrin.h: New.
2035 * config/i386/cpuid.h (bit_AVX512VPOPCNTDQ): New.
2036 * config/i386/i386-builtin-types.def: Add new types.
2037 * config/i386/i386-builtin.def (__builtin_ia32_vpopcountd_v16si,
2038 __builtin_ia32_vpopcountd_v16si_mask, __builtin_ia32_vpopcountq_v8di,
2039 __builtin_ia32_vpopcountq_v8di_mask): New.
2040 * config/i386/i386-c.c (ix86_target_macros_internal): Define
2041 __AVX512VPOPCNTDQ__.
2042 * config/i386/i386.c (ix86_target_string): Add -mavx512vpopcntdq.
2043 (PTA_AVX512VPOPCNTDQ): Define.
2044 * config/i386/i386.h (TARGET_AVX512VPOPCNTDQ,
2045 TARGET_AVX512VPOPCNTDQ_P): Define.
2046 * config/i386/i386.opt: Add mavx512vpopcntdq.
2047 * config/i386/immintrin.h: Include avx512vpopcntdqintrin.h.
2048 * config/i386/sse.md (define_insn "vpopcount<mode><mask_name>"): New.
2050 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
2053 * predict.def (PRED_CALL): Set to 67.
2055 2017-01-09 Eric Botcazou <ebotcazou@adacore.com>
2057 * expr.c (store_field): In the bitfield case, if the value comes from
2058 a function call and is of an aggregate type returned in registers, do
2059 not modify the field mode; extract the value in all cases if the mode
2060 is BLKmode and the size is not larger than a word.
2062 2017-01-09 Dominique d'Humieres <dominiq@lps.ens.fr>
2065 * config/i386/cpuid.h: Fix undefined behavior.
2067 2017-01-04 Jeff Law <law@redhat.com>
2069 PR tree-optimization/79007
2070 PR tree-optimization/67955
2071 * tree-ssa-alias.c (same_addr_size_stores_p): Only need to be
2072 conservative for pt.null when flag_non_call_exceptions is on.
2074 2017-01-09 Jakub Jelinek <jakub@redhat.com>
2076 PR translation/79019
2077 PR translation/79020
2078 * params.def (PARAM_INLINE_MIN_SPEEDUP,
2079 PARAM_IPA_CP_SINGLE_CALL_PENALTY,
2080 PARAM_USE_AFTER_SCOPE_DIRECT_EMISSION_THRESHOLD): Fix typos
2082 * config/avr/avr.opt (maccumulate-args): Likewise.
2083 * config/msp430/msp430.opt (mwarn-mcu): Likewise.
2084 * common.opt (freport-bug): Likewise.
2085 * cif-code.def (CIF_FINAL_ERROR): Likewise.
2086 * doc/invoke.texi (ipa-cp-single-call-penalty): Likewise.
2087 * config/s390/s390.c (s390_invalid_binary_op): Fix spelling in
2088 translatable string.
2089 * config/i386/i386.c (function_value_32): Likewise.
2090 * config/nios2/nios2.c (nios2_valid_target_attribute_rec): Likewise.
2091 * config/msp430/msp430.c (msp430_option_override, msp430_attr):
2093 * config/msp430/driver-msp430.c (msp430_select_hwmult_lib): Likewise.
2094 * common/config/msp430/msp430-common.c (msp430_handle_option):
2096 * symtab.c (symtab_node::verify_base): Likewise.
2097 * opts.c (set_debug_level): Likewise.
2098 * tree.c (verify_type_variant): Likewise. Fix typo in comment.
2099 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin): Add
2100 missing whitespace to translatable strings.
2101 * config/avr/avr.md (bswapsi2): Fix typo in comment.
2102 * config/sh/superh.h: Likewise.
2103 * config/i386/xopintrin.h: Likewise.
2104 * config/i386/znver1.md: Likewise.
2105 * config/rs6000/rs6000.c (struct rs6000_opt_mask): Likewise.
2106 * ipa-inline-analysis.c (compute_inline_parameters): Likewise.
2107 * double-int.h (struct double_int): Likewise.
2108 * double-int.c (div_and_round_double): Likewise.
2109 * wide-int.cc: Likewise.
2110 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
2111 * tree-ssa-sccvn.c (vn_reference_lookup_3): Likewise.
2112 * cfgcleanup.c (crossjumps_occured): Renamed to ...
2113 (crossjumps_occurred): ... this.
2114 (try_crossjump_bb, try_head_merge_bb, try_optimize_cfg, cleanup_cfg):
2117 PR tree-optimization/78899
2118 * tree-if-conv.c (version_loop_for_if_conversion): Instead of
2119 returning bool return struct loop *, NULL for failure and the new
2121 (versionable_outer_loop_p): Don't version outer loop if it has
2122 dont_vectorized bit set.
2123 (tree_if_conversion): When versioning outer loop, ensure
2124 tree_if_conversion is performed also on the inner loop of the
2125 non-vectorizable outer loop copy.
2126 * tree-vectorizer.c (set_uid_loop_bbs): Formatting fix. Fold
2127 LOOP_VECTORIZED in inner loop of the scalar outer loop and
2128 prevent vectorization of it.
2129 (vectorize_loops): For outer + inner LOOP_VECTORIZED, ensure
2130 the outer loop vectorization of the non-scalar version is attempted
2131 before vectorization of the inner loop in scalar version. If
2132 outer LOOP_VECTORIZED guarded loop is not vectorized, prevent
2133 vectorization of its inner loop.
2134 * tree-vect-loop-manip.c (rename_variables_in_bb): If outer_loop
2135 has 2 inner loops, rename also on edges from bb whose single pred
2136 is outer_loop->header. Fix typo in function comment.
2138 2017-01-09 Martin Sebor <msebor@redhat.com>
2141 * asan.c (asan_emit_stack_protection): Increase local buffer size
2142 to avoid snprintf truncation warning.
2144 2017-01-09 Andrew Pinski <apinski@cavium.com>
2146 * config/aarch64/aarch64-cores.def: Add thunderx2t99. Change vulcan
2147 to reference thunderx2t99 for the tuning structure
2148 * config/aarch64/aarch64-cost-tables.h (vulcan_extra_costs):
2150 (thunderx2t99_extra_costs): This.
2151 * config/aarch64/aarch64-tune.md: Regenerate.
2152 * config/aarch64/aarch64.c (vulcan_addrcost_table): Rename to ...
2153 (vulcan_addrcost_table): This.
2154 (vulcan_regmove_cost): Rename to ...
2155 (thunderx2t99_regmove_cost): This.
2156 (vulcan_vector_cost): Rename to ...
2157 (thunderx2t99_vector_cost): this.
2158 (vulcan_branch_cost): Rename to ...
2159 (thunderx2t99_branch_cost): This.
2160 (vulcan_tunings): Rename to ...
2161 (thunderx2t99_tunings): This and s/vulcan/thunderx2t99 .
2162 * doc/invoke.texi (AARCH64/mtune): Add thunderx2t99.
2164 2017-01-09 Martin Jambor <mjambor@suse.cz>
2168 * ipa-prop.h (ipa_jump_func): Swap positions of vr_known and m_vr.
2169 * ipa-cp.c (ipa_vr_operation_and_type_effects): New function.
2170 (propagate_vr_accross_jump_function): Use the above function for all
2171 value range computations for pass-through jump functions and type
2172 converasion from explicit value range values.
2173 (ipcp_propagate_stage): Do not attempt to deduce types of formal
2174 parameters from TYPE_ARG_TYPES.
2175 * ipa-prop.c (ipa_write_jump_function): Remove trailing whitespace.
2176 (ipa_write_node_info): Stream type of the actual argument.
2177 (ipa_read_node_info): Likewise. Also remove trailing whitespace.
2179 2017-01-09 Martin Liska <mliska@suse.cz>
2182 * gcc.c (driver_handle_option): Handle OPT_E and set have_E.
2183 (lookup_compiler): Do not show error message with have_E.
2185 2017-01-09 Jakub Jelinek <jakub@redhat.com>
2187 PR tree-optimization/78938
2188 * tree-vect-stmts.c (vectorizable_condition): For non-masked COND_EXPR
2189 where comp_vectype is VECTOR_BOOLEAN_TYPE_P, use
2190 BIT_{NOT,XOR,AND,IOR}_EXPR on the comparison operands instead of
2191 {EQ,NE,GE,GT,LE,LT}_EXPR directly inside of VEC_COND_EXPR. Formatting
2194 2017-01-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2196 * tree-ssa-address.c (gen_addr_rtx): Don't handle index if it
2199 2017-01-09 Richard Biener <rguenther@suse.de>
2201 PR tree-optimization/78997
2202 * tree-vect-slp.c (vect_mask_constant_operand_p): Handle SSA
2203 name condition properly.
2205 2017-01-09 Richard Biener <rguenther@suse.de>
2208 * dwarf2out.c (is_cxx): New overload with context.
2209 (is_naming_typedef_decl): Use it.
2211 2017-01-08 Sandra Loosemore <sandra@codesourcery.com>
2213 * invoke.texi (Option Summary): Correct spacing in option lists
2214 and add line breaks to fix over-long lines.
2216 2017-01-08 Sandra Loosemore <sandra@codesourcery.com>
2220 * extend.texi (Common Variable Attributes): Add xref to GCC
2221 Internals manual to explain mode attribute keywords.
2223 2017-01-08 Sandra Loosemore <sandra@codesourcery.com>
2226 * doc/invoke.texi (Option Summary): Move -pthread to Linker Options
2227 and Preprocessor Options.
2228 (Options for Linking): Document -pthread here....
2229 (RS/6000 and PowerPC Options): ...not here.
2230 (Solaris 2 Options): ...or here.
2231 * doc/cppopts.texi: Document -pthread.
2233 2017-01-08 Martin Sebor <msebor@redhat.com>
2236 * doc/invoke.texi (Warning Options): Document -Wformat-truncation.
2237 * gimple-ssa-sprintf.c (call_info::reval_used, call_info::warnopt):
2238 New member functions.
2239 (format_directive): Used them.
2241 (pass_sprintf_length::handle_gimple_call): Same.
2242 * graphite-sese-to-poly.c (tree_int_to_gmp): Increase buffer size
2243 to avoid truncation for any argument.
2244 (extract_affine_mul): Same.
2245 * tree.c (get_file_function_name): Same.
2247 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
2250 * predict.def (PRED_INDIR_CALL): Set to 86.
2252 2017-01-07 Sandra Loosemore <sandra@codesourcery.com>
2254 PR preprocessor/54124
2255 * doc/cppopts.texi: Reformat -d subtable to list the full name
2256 of the options. Add cross-reference to the docs for the general
2257 compiler -d options.
2258 * doc/invoke.texi (Developer Options): Add cross-reference to the
2259 preprocessor-specific -d option documentation.
2261 2017-01-07 Sandra Loosemore <sandra@codesourcery.com>
2263 PR preprocessor/13498
2264 * doc/cpp.texi (Search Path): Rewrite to remove obsolete and
2265 redudant material, and reflect new command-line options.
2266 (System Headers): Likewise.
2268 2017-01-07 Sandra Loosemore <sandra@codesourcery.com>
2270 * doc/cppdiropts.texi: Merge documentation of -I, -iquote,
2271 -isystem, and -idirafter. Copy-edit.
2272 * doc/cppopts.texi: Copy-edit. Remove contradiction about
2273 default for -ftrack-macro-expansion. Delete obsolete and
2274 badly-formatted implementation details about -fdebug-cpp output.
2275 * doc/cppwarnopts.texi: Copy-edit.
2277 2017-01-07 David Malcolm <dmalcolm@redhat.com>
2280 * input.c (selftest::test_accessing_ordinary_linemaps): Verify
2281 that the transition from a max line width >= 1<<10 to narrower
2282 lines works correctly.
2284 2017-01-07 Alexandre Oliva <aoliva@redhat.com>
2286 * doc/options.texi (PerFunction): New.
2287 * opt-functions.awk (switch_flags): Map both Optimization and
2288 PerFunction to CL_OPTIMIZATION.
2289 * opth-gen.awk: Test for PerFunction flag along with
2291 * optc-save-gen.awk: Likewise. Introduce var_opt_hash and set
2292 it only when the latter is present. Skip those that don't in
2293 the hash function generator.
2294 * common.opt (fvar-tracking): Mark as PerFunction instead of
2296 (fvar-tracking-assignments): Likewise.
2297 (fvar-tracking-assignments-toggle): Likewise.
2298 (fvar-tracking-uninit): Likewise.
2300 2017-01-07 Jakub Jelinek <jakub@redhat.com>
2302 PR translation/79018
2303 * params.def (PARAM_MAX_STORES_TO_MERGE): Add missing space between
2306 2017-01-06 Mikael Pettersson <mikpelinux@gmail.com>
2309 * config/m68k/m68k.opt (LONG_JUMP_TABLE_OFFSETS): New option.
2310 * config/m68k/linux.h (ASM_RETURN_CASE_JUMP): Handle
2311 TARGET_LONG_JUMP_TABLE_OFFSETS.
2312 * config/m68k/m68kelf.h (ASM_RETURN_CASE_JUMP): Likewise.
2313 * config/m68k/netbsd-elf.h (ASM_RETURN_CASE_JUMP): Likewise.
2314 * config/m68k/m68k.h (CASE_VECTOR_MODE): Likewise.
2315 (ASM_OUTPUT_ADDR_DIFF_ELF): Likewise.
2316 * config/m68k/m68k.md (tablejump expander): Likewise.
2317 (*tablejump_pcrel_hi): Renamed from unnamed insn, reject
2318 TARGET_LONG_JUMP_TABLE_OFFSETS.
2319 (*tablejump_pcrel_si): New insn, handle TARGET_LONG_JUMP_TABLE_OFFSETS.
2320 * doc/invoke.texi (M68K options): Add -mlong-jump-table-offsets.
2322 2017-01-06 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
2323 David Holsgrove <david.holsgrove@xilinx.com>
2325 * common/config/microblaze/microblaze-common.c
2326 (TARGET_EXCEPT_UNWIND_INFO): Remove.
2327 * config/microblaze/microblaze-protos.h (microblaze_eh_return):
2329 * config/microblaze/microblaze.c (microblaze_must_save_register)
2330 (microblaze_expand_epilogue, microblaze_return_addr): Handle
2332 (microblaze_eh_return): New function.
2333 * config/microblaze/microblaze.h (RETURN_ADDR_OFFSET)
2334 (EH_RETURN_DATA_REGNO, MB_EH_STACKADJ_REGNUM)
2335 (EH_RETURN_STACKADJ_RTX, ASM_PREFERRED_EH_DATA_FORMAT): New macros.
2336 * config/microblaze/microblaze.md (eh_return): New pattern.
2338 2017-01-06 Jakub Jelinek <jakub@redhat.com>
2340 * system.h (GCC_DIAGNOSTIC_PUSH_IGNORED, GCC_DIAGNOSTIC_POP,
2341 GCC_DIAGNOSTIC_STRINGIFY): Define.
2343 * read-rtl.c (rtx_reader::read_rtx_code): Avoid -Wsign-compare warning.
2345 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2347 * config/arm/arm.md (<mcrr>): New.
2349 * config/arm/arm.c (arm_arch5te): New.
2350 (arm_option_override): Set arm_arch5te.
2351 (arm_coproc_builtin_available): Add support for mcrr, mcrr2, mrrc
2353 * config/arm/arm-builtins.c (MCRR_QUALIFIERS): Define to...
2354 (arm_mcrr_qualifiers): ... this. New.
2355 (MRRC_QUALIFIERS): Define to...
2356 (arm_mrrc_qualifiers): ... this. New.
2357 * config/arm/arm_acle.h (__arm_mcrr, __arm_mcrr2, __arm_mrrc,
2359 * config/arm/arm_acle_builtins.def (mcrr, mcrr2, mrrc, mrrc2): New.
2360 * config/arm/iterators.md (MCRRI, mcrr, MCRR): New.
2361 (MRRCI, mrrc, MRRC): New.
2362 * config/arm/unspecs.md (VUNSPEC_MCRR, VUNSPEC_MCRR2, VUNSPEC_MRRC,
2363 VUNSPEC_MRRC2): New.
2365 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2367 * config/arm/arm.md (<mcr>): New.
2369 * config/arm/arm.c (arm_coproc_builtin_available): Add
2370 support for mcr, mrc, mcr2 and mrc2.
2371 * config/arm/arm-builtins.c (MCR_QUALIFIERS): Define to...
2372 (arm_mcr_qualifiers): ... this. New.
2373 (MRC_QUALIFIERS): Define to ...
2374 (arm_mrc_qualifiers): ... this. New.
2375 (MCR_QUALIFIERS): Define to ...
2376 (arm_mcr_qualifiers): ... this. New.
2377 * config/arm/arm_acle.h (__arm_mcr, __arm_mrc, __arm_mcr2,
2379 * config/arm/arm_acle_builtins.def (mcr, mcr2, mrc, mrc2): New.
2380 * config/arm/iterators.md (MCRI, mcr, MCR, MRCI, mrc, MRC): New.
2381 * config/arm/unspecs.md (VUNSPEC_MCR, VUNSPEC_MCR2, VUNSPEC_MRC,
2384 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2386 * config/arm/arm.md (*ldc): New.
2390 * config/arm/arm.c (arm_coproc_builtin_available): Add
2391 support for ldc,ldcl,stc,stcl,ldc2,ldc2l,stc2 and stc2l.
2392 (arm_coproc_ldc_stc_legitimate_address): New.
2393 * config/arm/arm-builtins.c (arm_type_qualifiers): Add
2394 'qualifier_const_pointer'.
2395 (LDC_QUALIFIERS): Define to...
2396 (arm_ldc_qualifiers): ... this. New.
2397 (STC_QUALIFIERS): Define to...
2398 (arm_stc_qualifiers): ... this. New.
2399 * config/arm/arm-protos.h
2400 (arm_coproc_ldc_stc_legitimate_address): New.
2401 * config/arm/arm_acle.h (__arm_ldc, __arm_ldcl, __arm_stc,
2402 __arm_stcl, __arm_ldc2, __arm_ldc2l, __arm_stc2, __arm_stc2l): New.
2403 * config/arm/arm_acle_builtins.def (ldc, ldc2, ldcl, ldc2l, stc,
2404 stc2, stcl, stc2l): New.
2405 * config/arm/constraints.md (Uz): New.
2406 * config/arm/iterators.md (LDCI, STCI, ldc, stc, LDC STC): New.
2407 * config/arm/unspecs.md (VUNSPEC_LDC, VUNSPEC_LDC2, VUNSPEC_LDCL,
2408 VUNSPEC_LDC2L, VUNSPEC_STC, VUNSPEC_STC2, VUNSPEC_STCL,
2409 VUNSPEC_STC2L): New.
2411 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2413 * config/arm/arm.md (<cdp>): New.
2414 * config/arm/arm.c (neon_const_bounds): Rename this ...
2415 (arm_const_bounds): ... this.
2416 (arm_coproc_builtin_available): New.
2417 * config/arm/arm-builtins.c (SIMD_MAX_BUILTIN_ARGS): Increase.
2418 (arm_type_qualifiers): Add 'qualifier_unsigned_immediate'.
2419 (CDP_QUALIFIERS): Define to...
2420 (arm_cdp_qualifiers): ... this. New.
2422 (arm_expand_builtin_args): Add case for 6 arguments.
2423 * config/arm/arm-protos.h (neon_const_bounds): Rename this ...
2424 (arm_const_bounds): ... this.
2425 (arm_coproc_builtin_available): New.
2426 * config/arm/arm_acle.h (__arm_cdp): New.
2428 * config/arm/arm_acle_builtins.def (cdp): New.
2430 * config/arm/iterators.md (CDPI,CDP,cdp): New.
2431 * config/arm/neon.md: Rename all 'neon_const_bounds' to
2433 * config/arm/types.md (coproc): New.
2434 * config/arm/unspecs.md (VUNSPEC_CDP, VUNSPEC_CDP2): New.
2435 * gcc/doc/extend.texi (ACLE): Add a mention of Coprocessor intrinsics.
2436 * gcc/doc/sourcebuild.texi (arm_coproc1_ok, arm_coproc2_ok,
2437 arm_coproc3_ok, arm_coproc4_ok): Document new effective targets.
2439 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2441 * config/arm/arm-builtins.c (arm_unsigned_binop_qualifiers): New.
2442 (UBINOP_QUALIFIERS): New.
2444 (acle_builtin_data): New. Change comment.
2445 (arm_builtins): Remove ARM_BUILTIN_CRC32B, ARM_BUILTIN_CRC32H,
2446 ARM_BUILTIN_CRC32W, ARM_BUILTIN_CRC32CB, ARM_BUILTIN_CRC32CH,
2447 ARM_BUILTIN_CRC32CW. Add ARM_BUILTIN_ACLE_BASE and include
2448 arm_acle_builtins.def.
2449 (ARM_BUILTIN_ACLE_PATTERN_START): Define.
2450 (arm_init_acle_builtins): New.
2451 (CRC32_BUILTIN): Remove.
2452 (bdesc_2arg): Remove entries for crc32b, crc32h, crc32w,
2453 crc32cb, crc32ch and crc32cw.
2454 (arm_init_crc32_builtins): Remove.
2455 (arm_init_builtins): Use arm_init_acle_builtins rather
2456 than arm_init_crc32_builtins.
2457 (arm_expand_acle_builtin): New.
2458 (arm_expand_builtin): Use 'arm_expand_acle_builtin'.
2459 * config/arm/arm_acle_builtins.def: New.
2461 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2463 * config/arm/arm-builtins.c (neon_builtin_datum): Rename to ..
2464 (arm_builtin_datum): ... this.
2465 (arm_init_neon_builtin): Rename to ...
2466 (arm_init_builtin): ... this. Add a new parameters PREFIX
2467 and USE_SIG_IN_NAME.
2468 (arm_init_neon_builtins): Replace 'arm_init_neon_builtin' with
2469 'arm_init_builtin'. Replace type 'neon_builtin_datum' with
2470 'arm_builtin_datum'.
2471 (arm_init_vfp_builtins): Likewise.
2472 (builtin_arg): Rename enum's replacing 'NEON_ARG' with
2473 'ARG_BUILTIN' and add a 'ARG_BUILTIN_NEON_MEMORY.
2474 (arm_expand_neon_args): Rename to ...
2475 (arm_expand_builtin_args): ... this. Rename builtin_arg
2476 enum values and differentiate between ARG_BUILTIN_MEMORY
2477 and ARG_BUILTIN_NEON_MEMORY.
2478 (arm_expand_neon_builtin_1): Rename to ...
2479 (arm_expand_builtin_1): ... this. Rename builtin_arg enum
2480 values, arm_expand_builtin_args and add bool parameter NEON.
2481 (arm_expand_neon_builtin): Use arm_expand_builtin_1.
2482 (arm_expand_vfp_builtin): Likewise.
2483 (NEON_MAX_BUILTIN_ARGS): Remove, it was unused.
2485 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
2488 * predict.def (PRED_POLYMORPHIC_CALL): Set to 59.
2489 * predict.c (tree_estimate_probability_bb): Reverse direction of
2490 polymorphic call predictor.
2492 2017-01-06 David Malcolm <dmalcolm@redhat.com>
2494 * passes.c (execute_one_pass): Split out pass-skipping logic into...
2495 (determine_pass_name_match): ...this new function and...
2496 (should_skip_pass_p): ...this new function.
2498 2017-01-06 Nathan Sidwell <nathan@acm.org>
2500 * ipa-visibility.c (function_and_variable_visibility): Reformat
2501 comments and long lines. Remove extrneous if.
2502 * symtab.c (symtab_node::make_decl_local): Fix code format.
2503 (symtab_node::set_section_for_node): Fix comment typo.
2505 2017-01-06 Martin Liska <mliska@suse.cz>
2508 * lra-constraints.c: Rename invariant to lra_invariant.
2509 * predict.c (set_even_probabilities): Initialize e to NULL.
2511 2017-01-05 Martin Sebor <msebor@redhat.com>
2513 PR tree-optimization/78910
2514 * gimple-ssa-sprintf.c (tree_digits): Add an argument.
2515 (format_integer): Correct off-by-one error in the handling
2516 of precision with negative numbers in signed conversions..
2518 2017-01-05 Eric Botcazou <ebotcazou@adacore.com>
2520 * doc/invoke.texi (C Dialect Options): Adjust -fsso-struct entry.
2522 2017-01-05 Jakub Jelinek <jakub@redhat.com>
2524 PR tree-optimization/71016
2525 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Pass cond_stmt to
2526 factor_out_conditional_conversion. Formatting fix.
2527 (factor_out_conditional_conversion): Add cond_stmt argument.
2528 If arg1 is INTEGER_CST, punt if new_arg0 is not any operand of
2529 cond_stmt and if arg0_def_stmt is not the only stmt in its bb.
2532 2017-01-05 David Malcolm <dmalcolm@redhat.com>
2534 * Makefile.in (OBJS): Add read-md.o, read-rtl.o,
2535 read-rtl-function.o, and selftest-rtl.o.
2536 * config/aarch64/aarch64.c: Include selftest.h and selftest-rtl.h.
2537 (selftest::aarch64_test_loading_full_dump): New function.
2538 (selftest::aarch64_run_selftests): New function.
2539 (TARGET_RUN_TARGET_SELFTESTS): Wire it up to
2540 selftest::aarch64_run_selftests.
2541 * config/i386/i386.c
2542 (selftest::ix86_test_loading_dump_fragment_1): New function.
2543 (selftest::ix86_test_loading_call_insn): New function.
2544 (selftest::ix86_test_loading_full_dump): New function.
2545 (selftest::ix86_test_loading_unspec): New function.
2546 (selftest::ix86_run_selftests): Call the new functions.
2547 * emit-rtl.c (maybe_set_max_label_num): New function.
2548 * emit-rtl.h (maybe_set_max_label_num): New decl.
2549 * function.c (instantiate_decls): Guard call to
2550 instantiate_decls_1 with if (DECL_INITIAL (fndecl)).
2551 * function-tests.c (selftest::verify_three_block_rtl_cfg): Remove
2553 * gensupport.c (gen_reader::gen_reader): Pass "false"
2554 for new "compact" param of rtx_reader.
2555 * print-rtl.c (rtx_writer::print_rtx_operand): Print "(nil)"
2556 rather than an empty string for NULL strings.
2557 * read-md.c: Potentially include config.h rather than bconfig.h.
2558 Wrap include of errors.h with #ifdef GENERATOR_FILE.
2559 (have_error): New global, copied from errors.c.
2560 (md_reader::read_name): Rename to...
2561 (md_reader::read_name_1): ...this, adding "out_loc" param,
2562 and converting "missing name or number" to returning false, rather
2564 (md_reader::read_name): Reimplement in terms of read_name_1.
2565 (md_reader::read_name_or_nil): New function.
2566 (md_reader::read_string): Handle "(nil)" by returning NULL.
2567 (md_reader::md_reader): Add new param "compact".
2568 (md_reader::read_md_files): Wrap with #ifdef GENERATOR_FILE.
2569 (md_reader::read_file): New method.
2570 * read-md.h (md_reader::md_reader): Add new param "compact".
2571 (md_reader::read_file): New method.
2572 (md_reader::is_compact): New accessor.
2573 (md_reader::read_name): Convert return type from void to file_location.
2574 (md_reader::read_name_or_nil): New decl.
2575 (md_reader::read_name_1): New decl.
2576 (md_reader::m_compact): New field.
2577 (noop_reader::noop_reader): Pass "false" for new "compact" param
2579 (rtx_reader::rtx_reader): Add new "compact" param.
2580 (rtx_reader::read_rtx_operand): Make virtual and convert return
2581 type from void to rtx.
2582 (rtx_reader::read_until): New decl.
2583 (rtx_reader::handle_any_trailing_information): New virtual function.
2584 (rtx_reader::postprocess): New virtual function.
2585 (rtx_reader::finalize_string): New virtual function.
2586 (rtx_reader::m_in_call_function_usage): New field.
2587 (rtx_reader::m_reuse_rtx_by_id): New field.
2588 * read-rtl-function.c: New file.
2589 * selftest-rtl.c (selftest::assert_rtx_ptr_eq_at): New function.
2590 * selftest-rtl.h (ASSERT_RTX_PTR_EQ): New macro.
2591 (selftest::verify_three_block_rtl_cfg): New decl.
2592 * read-rtl-function.h: New file.
2593 * read-rtl.c: Potentially include config.h rather than bconfig.h.
2594 For host, include function.h, memmodel.h, and emit-rtl.h.
2595 (one_time_initialization): New function.
2596 (struct compact_insn_name): New struct.
2597 (compact_insn_names): New array.
2598 (find_code): Handle insn codes in compact dumps.
2599 (apply_subst_iterator): Wrap with #ifdef GENERATOR_FILE.
2600 (bind_subst_iter_and_attr): Likewise.
2601 (add_condition_to_string): Likewise.
2602 (add_condition_to_rtx): Likewise.
2603 (apply_attribute_uses): Likewise.
2604 (add_current_iterators): Likewise.
2605 (apply_iterators): Likewise.
2606 (initialize_iterators): Guard usage of apply_subst_iterator with
2607 #ifdef GENERATOR_FILE.
2608 (read_conditions): Wrap with #ifdef GENERATOR_FILE.
2609 (md_reader::read_mapping): Likewise.
2610 (add_define_attr_for_define_subst): Likewise.
2611 (add_define_subst_attr): Likewise.
2612 (read_subst_mapping): Likewise.
2613 (check_code_iterator): Likewise.
2614 (rtx_reader::read_rtx): Likewise. Move one-time initialization
2616 (one_time_initialization): New function.
2617 (rtx_reader::read_until): New method.
2618 (read_flags): New function.
2619 (parse_reg_note_name): New function.
2620 (rtx_reader::read_rtx_code): Initialize "iterator" to NULL.
2621 Handle reuse_rtx ids.
2622 Wrap iterator lookup within #ifdef GENERATOR_FILE.
2623 Add parsing support for RTL dumps, mirroring the special-cases in
2624 print_rtx, by calling read_flags, reading REG_NOTE names, INSN_UID
2625 values, and calling handle_any_trailing_information.
2626 (rtx_reader::read_rtx_operand): Convert return type from void
2627 to rtx, returning return_rtx. Handle case 'e'. Call
2628 finalize_string on XSTR and XTMPL fields.
2629 (rtx_reader::read_nested_rtx): Handle dumps in which trailing
2630 "(nil)" values were omitted. Call the postprocess vfunc on the
2632 (rtx_reader::rtx_reader): Add new "compact" param and pass to base
2633 class ctor. Initialize m_in_call_function_usage. Call
2634 one_time_initialization.
2635 * rtl-tests.c (selftest::test_uncond_jump): Call
2636 set_new_first_and_last_insn.
2637 * rtl.h (read_rtx): Wrap decl with #ifdef GENERATOR_FILE.
2638 * selftest-rtl.c: New file.
2639 * selftest-rtl.h (class selftest::rtl_dump_test): New class.
2640 (selftest::get_insn_by_uid): New decl.
2641 * selftest-run-tests.c (selftest::run_tests): Call
2642 read_rtl_function_c_tests.
2643 * selftest.h (selftest::read_rtl_function_c_tests): New decl.
2644 * tree-dfa.c (ssa_default_def): Return NULL_TREE for rtl function
2647 2017-01-05 Uros Bizjak <ubizjak@gmail.com>
2649 * config/i386/i386.md (*testqi_ext_3): No need to handle memory
2650 operands in a special way. Assert that pos+len <= mode precision.
2652 2017-01-05 Jakub Jelinek <jakub@redhat.com>
2654 * common.opt (fvect-cost-model): Remove RejectNegative flag, use
2655 3 argument Alias with unlimited for the negative form.
2656 (fno-vect-cost-model): Removed.
2658 2017-01-05 Martin Liska <mliska@suse.cz>
2660 * hsa-gen.c (gen_hsa_divmod): New function.
2661 (gen_hsa_insn_for_internal_fn_call): Use the function for IFN_DIVMOD.
2663 2017-01-05 Martin Liska <mliska@suse.cz>
2666 * gcc.c (lookup_compiler): Reject '-' filename for a precompiled
2669 2017-01-05 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
2671 * config/s390/s390.c (s390_expand_setmem): Unroll the loop for
2672 small constant length operands.
2674 2017-01-05 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
2676 * config/s390/s390.c (s390_expand_setmem): Avoid overlapping bytes
2677 between loop iterations.
2679 2017-01-05 Martin Liska <mliska@suse.cz>
2682 * gimplify.c (gimplify_decl_expr): Compare to
2683 asan_poisoned_variables instread of checking flags.
2684 (gimplify_target_expr): Likewise.
2685 (gimplify_expr): Likewise.
2686 (gimplify_function_tree): Conditionally initialize
2687 asan_poisoned_variables.
2689 2017-01-04 Jeff Law <law@redhat.com>
2691 PR tree-optimizatin/78812
2692 * rtl.h (contains_mem_rtx_p): Prototype.
2693 * ifcvt.c (containts_mem_rtx_p): Move from here to...
2694 * rtlanal.c (contains_mem_rtx_p): Here and remove static linkage.
2695 * gcse.c (prune_expressions): Use contains_mem_rtx_p to discover
2696 and prune MEMs that are not at the toplevel of a SET_SRC rtx. Look
2697 through ZERO_EXTEND and SIGN_EXTEND when trying to avoid pruning MEMs.
2699 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
2701 * input.c (assert_char_at_range): Default-initialize actual_range.
2703 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
2705 * df-scan.c (df_ref_create_structure): Make regno unsigned,
2706 to match the caller.
2708 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
2710 * cfgexpand.c (expand_gimple_basic_block): Disregard debug
2711 insns after final jump in test to emit dummy move.
2713 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
2715 * gimple-iterator.h (gsi_one_nondebug_before_end_p): New.
2716 * tree-eh.c (cleanup_empty_eh): Skip more debug stmts.
2718 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
2720 * multiple_target.c (create_dispatcher_calls): Init e_next.
2721 * tree-ssa-loop-split.c (split_loop): Init border.
2722 * tree-vect-loop.c (vect_determine_vectorization_factor): Init
2725 2017-01-04 Michael Meissner <meissner@linux.vnet.ibm.com>
2730 * config/rs6000/predicates.md (sf_subreg_operand): New predicate.
2731 (altivec_register_operand): Do not return true if the operand
2732 contains a SUBREG mixing SImode and SFmode.
2733 (vsx_register_operand): Likewise.
2734 (vsx_reg_sfsubreg_ok): New predicate.
2735 (vfloat_operand): Do not return true if the operand contains a
2736 SUBREG mixing SImode and SFmode.
2737 (vint_operand): Likewise.
2738 (vlogical_operand): Likewise.
2739 (gpc_reg_operand): Likewise.
2740 (int_reg_operand): Likewise.
2741 * config/rs6000/rs6000-protos.h (valid_sf_si_move): Add declaration.
2742 * config/rs6000/rs6000.c (valid_sf_si_move): New function to
2743 determine if a MOVSI or MOVSF operation contains SUBREGs that mix
2745 (rs6000_emit_move_si_sf_subreg): New helper function.
2746 (rs6000_emit_move): Call rs6000_emit_move_si_sf_subreg to possbily
2747 fixup SUBREGs involving SImode and SFmode.
2748 * config/rs6000/vsx.md (SFBOOL_*): New constants that are operand
2749 numbers for the new peephole2 optimization.
2750 (peephole2 for SFmode unions): New peephole2 to optimize cases in
2751 the GLIBC math library that do AND/IOR/XOR operations on single
2752 precision floating point.
2753 * config/rs6000/rs6000.h (TARGET_NO_SF_SUBREG): New internal
2754 target macros to say whether we need to avoid SUBREGs mixing
2756 (TARGET_ALLOW_SF_SUBREG): Likewise.
2757 * config/rs6000/rs6000.md (UNSPEC_SF_FROM_SI): New unspecs.
2758 (UNSPEC_SI_FROM_SF): Likewise.
2759 (iorxor): Change spacing.
2760 (and_ior_xor): New iterator for AND, IOR, and XOR.
2761 (movsi_from_sf): New insns for SImode/SFmode SUBREG support.
2762 (movdi_from_sf_zero_ext): Likewise.
2763 (mov<mode>_hardfloat, FMOVE32 iterator): Use register_operand
2764 instead of gpc_reg_operand. Add SImode/SFmode SUBREG support.
2765 (movsf_from_si): New insn for SImode/SFmode SUBREG support.
2766 (fma<mode>4): Use gpc_reg_operand instead of register_operand.
2767 (fms<mode>4): Likewise.
2768 (fnma<mode>4): Likewise.
2769 (fnms<mode>4): Likewise.
2770 (nfma<mode>4): Likewise.
2771 (nfms<mode>4): Likewise.
2773 2017-01-04 Marek Polacek <polacek@redhat.com>
2776 * doc/invoke.texi: Document -Wpointer-compare.
2778 2017-01-04 Jakub Jelinek <jakub@redhat.com>
2780 * optc-gen.awk: Emit #error for -W*/-f*/-m* Enum without
2783 * dwarf2out.c (output_loc_list): Don't throw away 64K+ location
2784 descriptions for -gdwarf-5 and emit them as uleb128 instead of
2787 2017-01-04 Kelvin Nilsen <kelvin@gcc.gnu.org>
2790 * doc/sourcebuild.texi (PowerPC-specific attributes): Add
2791 documentation of the powerpc_popcntb_ok attribute.
2792 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
2793 code to issue warning messages if a requested CPU configuration is
2794 not supported by the binary (assembler and loader) toolchain.
2795 (spe_init_builtins): Add two assertions to prevent ICE if attempt is
2796 made to define a built-in function that has been disabled.
2797 (paired_init_builtins): Add assertion to prevent ICE if attempt is
2798 made to define a built-in function that has been disabled.
2799 (altivec_init_builtins): Add comment explaining why definition
2800 of the DST built-in functions is not preceded by an assertion
2801 check. Add assertions to prevent ICE if attempts are made to
2802 define an altivec predicate or an abs* built-in function that has
2804 (htm_init_builtins): Add comment explaining why definition of the
2805 htm built-in functions is not preceded by an assertion check.
2807 2017-01-04 Jeff Law <law@redhat.com>
2809 PR tree-optimizatin/67955
2810 * tree-ssa-alias.c (same_addr_size_stores_p): Check offsets first.
2811 Allow any SSA_VAR_P as the base objects. Use integer_zerop. Verify
2812 the points-to solution does not include pt_null. Use DECL_PT_UID
2815 2017-01-04 Uros Bizjak <ubizjak@gmail.com>
2817 * config/i386/i386.md (HI/SImode test with imm to QImode splitters):
2818 Use gen_int_mode instead of gen_lopwart for const_int operands.
2820 2017-01-04 Jakub Jelinek <jakub@redhat.com>
2822 PR tree-optimization/71563
2823 * match.pd: Simplify X << Y into X if Y is known to be 0 or
2824 out of range value - has low bits known to be zero.
2826 2017-01-04 Alan Modra <amodra@gmail.com>
2828 * Makefile.in (aclocal_deps): Update and order as per aclocal.m4.
2829 * configure: Regenerate.
2830 * config.in: Regenerate.
2832 2017-01-04 Jakub Jelinek <jakub@redhat.com>
2835 * input.c (ebcdic_execution_charset::on_error): Don't use strstr for
2836 a substring of the message, but strcmp with the whole message. Ifdef
2837 ENABLE_NLS, translate the message first using dgettext.
2839 2017-01-03 Jeff Law <law@redhat.com>
2841 PR tree-optimizatin/78856
2842 * tree-ssa-threadupdate.c: Include tree-vectorizer.h.
2843 (mark_threaded_blocks): Remove code to truncate thread paths that
2844 cross multiple loop headers. Instead invalidate the cached loop
2845 iteration information and handle case of a thread path walking
2846 into an irreducible region.
2848 2017-01-03 Michael Meissner <meissner@linux.vnet.ibm.com>
2851 * config/rs6000/rs6000.c (rs6000_split_signbit): Change some
2852 assertions. Add support for doing the signbit if the IEEE 128-bit
2853 floating point value is in a GPR.
2854 * config/rs6000/rs6000.md (Fsignbit): Delete.
2855 (signbit<mode>2_dm): Delete using <Fsignbit> and just use "wa".
2856 Update the length attribute if the value is in a GPR.
2857 (signbit<mode>2_dm_<su>ext): Add combiner pattern to eliminate
2858 the sign or zero extension instruction, since the value is always 0/1.
2859 (signbit<mode>2_dm2): Delete using <Fsignbit>.
2862 * config/rs6000/vsx.md (vsx_extract_<mode>_store_p9): If we are
2863 extracting SImode to a GPR register so that we can generate a
2864 store, limit the vector to be in a traditional Altivec register
2865 for the vextuwrx instruction.
2867 2017-01-03 Ian Lance Taylor <iant@google.com>
2869 * godump.c (go_format_type): Treat ENUMERAL_TYPE like INTEGER_TYPE.
2871 2017-01-03 Martin Sebor <msebor@redhat.com>
2873 PR tree-optimization/78696
2874 * gimple-ssa-sprintf.c (format_floating): Correct handling of
2875 precision. Use MPFR for %f for greater fidelity. Correct handling
2877 (pass_sprintf_length::compute_format_length): Set width and precision
2878 specified by asrerisk to void_node for vararg functions.
2879 (try_substitute_return_value): Adjust dump output.
2881 2017-01-03 David Edelsohn <dje.gcc@gmail.com>
2883 * doc/invoke.texi (RS6000 options): LRA is enabled by default.
2885 2017-01-03 Eric Botcazou <ebotcazou@adacore.com>
2887 * doc/invoke.texi (SPARC options): Document -mlra as the default.
2888 * config/sparc/sparc.c (sparc_option_override): Force LRA unless
2889 -mlra/-mno-lra was passed to the compiler.
2891 2017-01-03 James Cowgill <James.Cowgill@imgtec.com>
2893 PR rtl-optimization/65618
2894 * emit-rtl.c (try_split): Move initialization of "before" and
2895 "after" to just before the call to emit_insn_after_setloc.
2897 2017-01-03 Gerald Pfeifer <gerald@pfeifer.com>
2899 * doc/md.texi (Standard Names): Remove reference to Java frontend.
2901 2017-01-03 Pierre-Marie de Rodat <derodat@adacore.com>
2903 * dwarf2out.c (gen_enumeration_type_die): When
2904 -gno-strict-dwarf, add a DW_AT_encoding attribute.
2906 2017-01-03 Jakub Jelinek <jakub@redhat.com>
2908 PR tree-optimization/78965
2909 * gimple-ssa-sprintf.c (pass_sprintf_length::compute_format_length):
2910 Change first argument from const call_info & to call_info &. For %n
2911 set info.nowrite to false.
2914 * gimple-ssa-sprintf.c (try_substitute_return_value): Don't change
2915 possibly throwing calls.
2917 * genmatch.c (dt_node::gen_kids_1): If generic_exprs include SSA_NAME
2918 and exprs_len || fns_len, emit the code for SSA_NAME next to the exprs
2919 and fns handling, rather than in a separate case SSA_NAME.
2921 2017-01-02 Jeff Law <law@redhat.com>
2923 * config/darwin-driver.c (darwin_driver_init): Const-correctness
2924 fixes for first_period and second_period variables.
2926 2017-01-02 Uros Bizjak <ubizjak@gmail.com>
2929 * config/i386/i386.md (UNSPEC_NOREX_MEM): New unspec.
2930 (*insvqi_1): New insn pattern.
2931 (*insvqi_1_mem_rex64): Ditto.
2933 (*insvqi_3): Rename from *insvqi.
2935 (*extzvqi_mem_rex64): Add UNSPEC_NOREX_MEM tag.
2937 2017-01-02 Gerald Pfeifer <gerald@pfeifer.com>
2939 * doc/cfg.texi (Edges): Remove reference to Java.
2940 (Maintaining the CFG): Ditto.
2942 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
2945 * symtab.c (symtab_node::binds_to_current_def_p): Fix handling of
2946 transparent aliases.
2948 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
2951 * predict.def (PRED_CALL): Update hitrate.
2952 (PRED_INDIR_CALL, PRED_POLYMORPHIC_CALL): New predictors.
2953 * predict.c (tree_estimate_probability_bb): Split CALL predictor
2954 into direct/indirect/polymorphic variants.
2956 2017-01-01 Jakub Jelinek <jakub@redhat.com>
2958 Update copyright years.
2960 * gcc.c (process_command): Update copyright notice dates.
2961 * gcov-dump.c (print_version): Ditto.
2962 * gcov.c (print_version): Ditto.
2963 * gcov-tool.c (print_version): Ditto.
2964 * gengtype.c (create_file): Ditto.
2965 * doc/cpp.texi: Bump @copying's copyright year.
2966 * doc/cppinternals.texi: Ditto.
2967 * doc/gcc.texi: Ditto.
2968 * doc/gccint.texi: Ditto.
2969 * doc/gcov.texi: Ditto.
2970 * doc/install.texi: Ditto.
2971 * doc/invoke.texi: Ditto.
2973 Copyright (C) 2017 Free Software Foundation, Inc.
2975 Copying and distribution of this file, with or without modification,
2976 are permitted in any medium without royalty provided the copyright
2977 notice and this notice are preserved.