1 2014-05-17 Uros Bizjak <ubizjak@gmail.com>
3 * doc/extend.texi (Size of an asm): Move node text according
4 to its @menu entry position.
6 2014-05-17 Marc Glisse <marc.glisse@inria.fr>
8 PR tree-optimization/61140
9 PR tree-optimization/61150
10 PR tree-optimization/61197
11 * tree-ssa-phiopt.c (value_replacement): Punt on multiple phis.
13 2014-05-17 Uros Bizjak <ubizjak@gmail.com>
15 * doc/invoke.texi (free): Mention Alpha. Also enabled at -Os.
17 2014-05-17 Richard Sandiford <r.sandiford@uk.ibm.com>
19 * wide-int.cc: Only include longlong.h if W_TYPE_SIZE==32 or
20 __SIZEOF_INT128__ is defined.
22 2014-05-17 Richard Sandiford <rdsandiford@googlemail.com>
24 * config/rs6000/rs6000.c (rs6000_real_tls_symbol_ref_p): New function.
25 (rs6000_delegitimize_address): Use it.
27 2014-05-17 Richard Sandiford <rdsandiford@googlemail.com>
29 * emit-rtl.h (replace_equiv_address, replace_equiv_address_nv): Add an
30 inplace argument. Store the new address in the original MEM when true.
31 * emit-rtl.c (change_address_1): Likewise.
32 (adjust_address_1, adjust_automodify_address_1, offset_address):
34 * rtl.h (plus_constant): Add an inplace argument.
35 * explow.c (plus_constant): Likewise. Try to reuse the original PLUS
36 when true. Avoid generating (plus X (const_int 0)).
37 * function.c (instantiate_virtual_regs_in_rtx): Adjust the PLUS
38 in-place. Pass true to plus_constant.
39 (instantiate_virtual_regs_in_insn): Pass true to replace_equiv_address.
41 2014-05-16 Dehao Chen <dehao@google.com>
43 * tree-cfg.c (gimple_merge_blocks): Updates bb count with max count.
45 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
48 * config/sh/predicates.md (negt_reg_shl31_operand): Match additional
50 * config/sh/sh.md (*negt_msb): Merge SH2A and non-SH2A variants.
52 2014-05-16 Dehao Chen <dehao@google.com>
54 * ira-int.h (REG_FREQ_FROM_EDGE_FREQ): Use
55 optimize_function_for_size_p.
56 * regs.h (REG_FREQ_FROM_BB): Likewise.
58 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
61 * config/sh/sh.c (sh_eval_treg_value): Handle t_reg_operand and
62 negt_reg_operand cases.
63 * config/sh/sh.md (*cset_zero): Likewise by using cbranch_treg_value
65 * config/sh/predicates.md (cbranch_treg_value): Simplify.
67 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
69 * config/sh/sh.c (sh_option_override): Set branch cost to 2 for all
72 2014-05-16 David Malcolm <dmalcolm@redhat.com>
75 2014-04-29 David Malcolm <dmalcolm@redhat.com>
77 * tree-cfg.c (dump_function_to_file): Dump the return type of
78 functions, in a line to itself before the function body, mimicking
79 the layout of a C function.
81 2014-05-16 Dehao Chen <dehao@google.com>
83 * cfghooks.c (make_forwarder_block): Use direct computation to
84 get fall-through edge's count and frequency.
86 2014-05-16 Benno Schulenberg <bensberg@justemail.net>
88 * config/arc/arc.c (arc_init): Fix typo in error message.
89 * config/i386/i386.c (ix86_expand_builtin): Likewise.
90 (split_stack_prologue_scratch_regno): Likewise.
91 * fortran/check.c (gfc_check_fn_rc2008): Remove duplicate
92 word from error message.
94 2014-05-16 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
96 * ira-costs.c: Fix typo in comment.
98 2014-05-16 David Wohlferd <dw@LimeGreenSocks.com>
100 * doc/extend.texi: (Visibility Pragmas) Fix misplaced @xref
102 2014-05-16 Jan Hubicka <hubicka@ucw.cz>
104 * varpool.c (dump_varpool_node): Dump write-only flag.
105 * lto-cgraph.c (lto_output_varpool_node, input_varpool_node): Stream
107 * tree-cfg.c (execute_fixup_cfg): Remove statements setting
108 write-only variables.
109 * ipa.c (process_references): New function.
110 (set_readonly_bit): New function.
111 (set_writeonly_bit): New function.
112 (clear_addressable_bit): New function.
113 (ipa_discover_readonly_nonaddressable_var): Mark write only variables;
114 fix handling of aliases.
115 * cgraph.h (struct varpool_node): Add writeonly flag.
117 2014-05-16 Vladimir Makarov <vmakarov@redhat.com>
119 PR rtl-optimization/60969
120 * ira-costs.c (record_reg_classes): Allow only memory for pseudo.
121 Calculate costs for this case.
123 2014-05-16 Eric Botcazou <ebotcazou@adacore.com>
125 * fold-const (fold_unary_loc) <NON_LVALUE_EXPR>: New case.
126 <CASE_CONVERT>: Pass arg0 instead of op0 to fold_convert_const.
128 2014-05-16 Richard Biener <rguenther@suse.de>
130 PR tree-optimization/61194
131 * tree-vect-patterns.c (adjust_bool_pattern): Also handle
132 bool patterns ending in a COND_EXPR.
134 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
136 * config/aarch64/aarch64.c (aarch64_rtx_mult_cost): Fix FNMUL case.
138 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
140 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle the case
141 where we were unable to cost an RTX.
143 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
145 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost SYMBOL_REF,
148 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
149 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
151 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost TRUNCATE.
153 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
154 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
156 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost FMA,
157 FLOAT_EXTEND, FLOAT_TRUNCATE, ABS, SMAX, and SMIN.
159 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
160 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
162 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost comparison
165 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
166 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
168 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costs for
171 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
172 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
174 * config/aarch64/aarch64.c (aarch64_rtx_arith_op_extract_p): New.
175 (aarch64_rtx_costs): Improve costs for SIGN/ZERO_EXTRACT.
177 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
178 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
180 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costs for
183 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
184 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
186 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost
187 ZERO_EXTEND and SIGN_EXTEND better.
189 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
190 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
192 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve cost for
195 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
196 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
198 * config/aarch64/aarch64.c (aarch64_rtx_costs): Use address
199 costs when costing loads and stores to memory.
201 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
202 Philip Tomsich <philipp.tomsich@theobroma-systems.com>
204 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costing
207 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
209 * config/aarch64/aarch64.c (aarch64_rtx_costs): Set default costs.
211 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
212 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
214 * config/aarch64/aarch64.c (aarch64_strip_shift_or_extend): Rename
216 (aarch64_strip_extend): ...this, don't strip shifts, check RTX is
218 (aarch64_rtx_mult_cost): New.
219 (aarch64_rtx_costs): Use it, refactor as appropriate.
221 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
222 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
224 * config/aarch64/aarch64.c (aarch64_build_constant): Conditionally
225 emit instructions, return number of instructions which would
227 (aarch64_add_constant): Update call to aarch64_build_constant.
228 (aarch64_output_mi_thunk): Likewise.
229 (aarch64_rtx_costs): Estimate cost of a CONST_INT, cost of
232 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
234 * config/aarch64/aarch64.c (aarch64_rtx_costs_wrapper): New.
235 (TARGET_RTX_COSTS): Call it.
237 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
239 * config/aarch64/aarch64.c (cortexa57_addrcost_table): New.
240 (cortexa57_vector_cost): Likewise.
241 (cortexa57_tunings): Use them.
243 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
245 * config/aarch64/aarch64-protos.h (scale_addr_mode_cost): New.
246 (cpu_addrcost_table): Use it.
247 * config/aarch64/aarch64.c (generic_addrcost_table): Initialize it.
248 (aarch64_address_cost): Rewrite using aarch64_classify_address,
251 2014-05-16 Richard Biener <rguenther@suse.de>
253 * tree-ssa-sccvn.c: Include tree-cfg.h and domwalk.h.
254 (set_ssa_val_to): Handle unexpected sets to VN_TOP.
255 (visit_phi): Ignore edges marked as not executable.
256 (class cond_dom_walker): New.
257 (cond_dom_walker::before_dom_children): Value-number
258 control statements and mark successor edges as not
259 executable if possible.
260 (run_scc_vn): First walk all control statements in
261 dominator order, marking edges as not executable.
262 * tree-inline.c (copy_edges_for_bb): Be not confused
263 about random edge flags.
265 2014-05-16 Richard Biener <rguenther@suse.de>
267 * tree-ssa-sccvn.c (visit_use): Also constant-fold calls.
269 2014-05-15 Peter Bergner <bergner@vnet.ibm.com>
272 * config/rs6000/htmxlintrin.h (_HTM_TBEGIN_STARTED): New define.
273 (__TM_simple_begin): Use it.
274 (__TM_begin): Likewise.
276 2014-05-15 Martin Jambor <mjambor@suse.cz>
279 * ipa-prop.c (update_indirect_edges_after_inlining): Check
280 type_preserved flag when the indirect edge is polymorphic.
282 2014-05-15 Martin Jambor <mjambor@suse.cz>
284 PR tree-optimization/61090
285 * tree-sra.c (sra_modify_expr): Pass the current gsi to
288 2014-05-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
290 * config/arm/arm.c (arm_option_override): Use the SCHED_PRESSURE_MODEL
291 enum name for PARAM_SCHED_PRESSURE_ALGORITHM.
293 2014-05-15 Jakub Jelinek <jakub@redhat.com>
295 PR tree-optimization/61158
296 * fold-const.c (fold_binary_loc): If X is zero-extended and
297 shiftc >= prec, make sure zerobits is all ones instead of
298 invoking undefined behavior.
300 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
302 * regcprop.h: New file.
303 * regcprop.c (skip_debug_insn_p): New decl.
304 (replace_oldest_value_reg): Check skip_debug_insn_p.
305 (copyprop_hardreg_forward_bb_without_debug_insn): New function.
306 * shrink-wrap.c: Include regcprop.h.
307 (prepare_shrink_wrap): Call
308 copyprop_hardreg_forward_bb_without_debug_insn.
310 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
312 * shrink-wrap.h: Update comment.
313 * shrink-wrap.c: Update comment.
314 (next_block_for_reg): Rename to live_edge_for_reg.
315 (live_edge_for_reg): Allow live_edge->dest has two predecessors.
316 (move_insn_for_shrink_wrap): Split live_edge.
317 (prepre_shrink_wrap): One more parameter for move_insn_for_shrink_wrap.
319 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
321 * config/sparc/sparc-protos.h (sparc_absnegfloat_split_legitimate):
323 * config/sparc/sparc.c (sparc_absnegfloat_split_legitimate): Likewise.
324 * config/sparc/sparc.md (fptype_ut699): New attribute.
325 (in_branch_delay): Return false if -mfix-ut699 is specified and
326 fptype_ut699 is set to single.
327 (truncdfsf2): Add fptype_ut699 attribute.
328 (fix_truncdfsi2): Likewise.
329 (floatsisf2): Change fptype attribute.
330 (fix_truncsfsi2): Likewise.
331 (negtf2_notv9): Delete.
332 (negtf2_v9): Likewise.
333 (negtf2_hq): New instruction.
334 (negtf2): New instruction and splitter.
335 (negdf2_notv9): Rewrite.
336 (abstf2_notv9): Delete.
337 (abstf2_hq_v9): Likewise.
338 (abstf2_v9): Likewise.
339 (abstf2_hq): New instruction.
340 (abstf2): New instruction and splitter.
341 (absdf2_notv9): Rewrite.
343 2014-05-14 Cary Coutant <ccoutant@google.com>
346 * opts.c (common_handle_option): Don't special-case "-g".
347 (set_debug_level): Default to at least level 2 with "-g".
349 2014-05-14 DJ Delorie <dj@redhat.com>
351 * config/msp430/msp430.c (msp430_builtin): Add
352 MSP430_BUILTIN_DELAY_CYCLES.
353 (msp430_init_builtins): Register void __delay_cycles(long long).
354 (msp430_builtin_decl): Add it.
355 (cg_magic_constant): New.
356 (msp430_expand_delay_cycles): New.
357 (msp430_expand_builtin): Call it.
358 (msp430_print_operand_raw): Change integer printing from "int" to
360 * config/msp430/msp430.md (define_constants): Add delay_cycles tags.
361 (delay_cycles_start): New.
362 (delay_cycles_end): New.
363 (delay_cycles_32): New.
364 (delay_cycles_32x): New.
365 (delay_cycles_16): New.
366 (delay_cycles_16x): New.
367 (delay_cycles_2): New.
368 (delay_cycles_1): New.
369 * doc/extend.texi: Document __delay_cycles().
371 2014-05-14 Sandra Loosemore <sandra@codesourcery.com>
373 * config/nios2/nios2.md (nios2_cbranch): Fix paste-o in
374 length attribute computation.
376 2014-05-14 Richard Sandiford <rdsandiford@googlemail.com>
379 * print-rtl.c (print_rtx): Suppress uids if flag_dump_unnumbered.
381 2014-05-14 Richard Sandiford <r.sandiford@uk.ibm.com>
384 * config/sparc/sparc.md: Fix types of low and high in DI constant
385 splitter. Use gen_int_mode in some other splitters.
387 2014-05-14 Martin Jambor <mjambor@suse.cz>
390 * ipa-prop.c (ipa_modify_formal_parameters): Reset DECL_LANG_SPECIFIC.
392 2014-05-14 James Norris <jnorris@codesourcery.com>
394 * omp-low.c (expand_parallel_call): Remove shadow variable.
395 (expand_omp_taskreg): Likewise.
397 2014-05-14 Ilya Tocar <ilya.tocar@intel.com>
399 * common/config/i386/i386-common.c
400 (OPTION_MASK_ISA_CLFLUSHOPT_SET): Define.
401 (OPTION_MASK_ISA_XSAVES_SET): Ditto.
402 (OPTION_MASK_ISA_XSAVEC_SET): Ditto.
403 (OPTION_MASK_ISA_CLFLUSHOPT_UNSET): Ditto.
404 (OPTION_MASK_ISA_XSAVES_UNSET): Ditto.
405 (OPTION_MASK_ISA_XSAVEC_UNSET): Ditto.
406 (ix86_handle_option): Handle OPT_mxsavec, OPT_mxsaves, OPT_mclflushopt.
407 * config.gcc (i[34567]86-*-*): Add clflushoptintrin.h,
408 xsavecintrin.h, xsavesintrin.h.
410 * config/i386/clflushoptintrin.h: New.
411 * config/i386/xsavecintrin.h: Ditto.
412 * config/i386/xsavesintrin.h: Ditto.
413 * config/i386/cpuid.h (bit_CLFLUSHOPT): Define.
416 * config/i386/driver-i386.c (host_detect_local_cpu): Handle
417 -mclflushopt, -mxsavec, -mxsaves, -mno-xsaves, -mno-xsavec,
419 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
420 OPTION_MASK_ISA_CLFLUSHOPT, OPTION_MASK_ISA_XSAVEC,
421 OPTION_MASK_ISA_XSAVES.
422 * config/i386/i386.c (ix86_target_string): Handle -mclflushopt,
424 (PTA_CLFLUSHOPT) Define.
427 (ix86_option_override_internal): Handle new options.
428 (ix86_valid_target_attribute_inner_p): Ditto.
429 (ix86_builtins): Add IX86_BUILTIN_XSAVEC, IX86_BUILTIN_XSAVEC64,
430 IX86_BUILTIN_XSAVES, IX86_BUILTIN_XRSTORS, IX86_BUILTIN_XSAVES64,
431 IX86_BUILTIN_XRSTORS64, IX86_BUILTIN_CLFLUSHOPT.
432 (bdesc_special_args): Add __builtin_ia32_xsaves,
433 __builtin_ia32_xrstors, __builtin_ia32_xsavec, __builtin_ia32_xsaves64,
434 __builtin_ia32_xrstors64, __builtin_ia32_xsavec64.
435 (ix86_init_mmx_sse_builtins): Add __builtin_ia32_clflushopt.
436 (ix86_expand_builtin): Handle new builtins.
437 * config/i386/i386.h (TARGET_CLFLUSHOPT) Define.
438 (TARGET_CLFLUSHOPT_P): Ditto.
439 (TARGET_XSAVEC): Ditto.
440 (TARGET_XSAVEC_P): Ditto.
441 (TARGET_XSAVES): Ditto.
442 (TARGET_XSAVES_P): Ditto.
443 * config/i386/i386.md (ANY_XSAVE): Add UNSPECV_XSAVEC, UNSPECV_XSAVES.
444 (ANY_XSAVE64)" Add UNSPECV_XSAVEC64, UNSPECV_XSAVES64.
445 (attr xsave): Add xsavec, xsavec64, xsaves, xsaves64.
447 (ANY_XRSTOR64): Ditto.
449 (xrstor): Change into <xrstor>.
450 (xrstor_rex64): Change into <xrstor>_rex64.
451 (xrstor64): Change into <xrstor>64
453 * config/i386/i386.opt (mclflushopt): New.
456 * config/i386/x86intrin.h: Add clflushoptintrin.h, xsavesintrin.h,
458 * doc/invoke.texi: Document new options.
460 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
462 PR rtl-optimization/60866
463 * sel-sched-ir (sel_init_new_insn): New parameter old_seqno.
464 Default it to -1. Pass it down to init_simplejump_data.
465 (init_simplejump_data): New parameter old_seqno. Pass it down
466 to get_seqno_for_a_jump.
467 (get_seqno_for_a_jump): New parameter old_seqno. Use it for
468 initializing new jump seqno as a last resort. Add comment.
469 (sel_redirect_edge_and_branch): Save old seqno of the conditional
470 jump and pass it down to sel_init_new_insn.
471 (sel_redirect_edge_and_branch_force): Likewise.
473 2014-05-14 Georg-Johann Lay <avr@gjlay.de>
475 * config/avr/avr.h (REG_CLASS_CONTENTS): Use unsigned suffix for
476 shifted values to avoid build warning.
478 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
480 * cfgcleanup.c (try_forward_edges): Use location_t for locations.
481 * cfgrtl.c (rtl_merge_blocks): Fix comment.
482 (cfg_layout_merge_blocks): Likewise.
483 * except.c (emit_to_new_bb_before): Remove prev_bb local variable.
485 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
487 PR rtl-optimization/60901
488 * config/i386/i386.c (ix86_dependencies_evaluation_hook): Check that
489 bb predecessor belongs to the same scheduling region. Adjust comment.
491 2014-05-13 Peter Bergner <bergner@vnet.ibm.com>
493 * doc/sourcebuild.texi: (dfp_hw): Document.
494 (p8vector_hw): Likewise.
495 (powerpc_eabi_ok): Likewise.
496 (powerpc_elfv2): Likewise.
497 (powerpc_htm_ok): Likewise.
498 (ppc_recip_hw): Likewise.
501 2014-05-13 Cary Coutant <ccoutant@google.com>
503 * opts.c (finish_options): Use -ggnu-pubnames with -gsplit-dwarf.
505 2014-05-13 David Malcolm <dmalcolm@redhat.com>
507 * gengtype-parse.c (require3): Eliminate in favor of...
509 (require_template_declaration): Update to support optional single *
512 * gengtype.c (get_ultimate_base_class): Add a non-const overload.
513 (create_user_defined_type): Handle a single level of explicit
514 pointerness within template arguments.
515 (struct write_types_data): Add field "kind".
516 (filter_type_name): Handle "*" character.
517 (write_user_func_for_structure_ptr): Require a write_types_data
518 rather than just a prefix string, so that we can look up the kind
519 of the wtd and use it as an index into wrote_user_func_for_ptr,
520 ensuring that such functions are written at most once. Support
521 subclasses by invoking the marking function of the ultimate base class.
522 (write_user_func_for_structure_body): Require a write_types_data
523 rather than just a prefix string, so that we can pass this to
524 write_user_func_for_structure_ptr.
525 (write_func_for_structure): Likewise.
526 (ggc_wtd): Add initializer of new "kind" field.
529 * gengtype.h (enum write_types_kinds): New.
530 (struct type): Add field wrote_user_func_for_ptr to the "s"
533 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
535 * fold-const.c (optimize_bit_field_compare): Use wi:: operations
536 instead of const_binop.
537 (fold_binary_loc): Likewise.
539 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
541 * tree-dfa.h (get_addr_base_and_unit_offset_1): Update array index
542 calculation to match get_ref_base_and_extent.
544 2014-05-13 Catherine Moore <clm@codesourcery.com>
545 Sandra Loosemore <sandra@codesourcery.com>
547 * configure.ac: Fix assembly for explicit JALR relocation check.
548 * configure: Regenerate.
550 2014-05-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
552 * config/arm/arm.c (neon_itype): Remove NEON_RESULTPAIR.
553 (arm_init_neon_builtins): Remove handling of NEON_RESULTPAIR.
554 Remove associated type declarations and initialisations.
555 (arm_expand_neon_builtin): Likewise.
556 (neon_emit_pair_result_insn): Delete.
557 * config/arm/arm_neon_builtins (vtrn, vzip, vuzp): Delete.
558 * config/arm/neon.md (neon_vtrn<mode>): Delete.
559 (neon_vzip<mode>): Likewise.
560 (neon_vuzp<mode>): Likewise.
562 2014-05-13 Richard Biener <rguenther@suse.de>
565 * tree-inline.c (remap_gimple_stmt): Clear tail call flag,
566 it needs revisiting whether the call still may be tail-called.
568 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
570 * rtl.def (SYMBOL_REF): Remove middle "0" field.
571 * rtl.h (block_symbol): Reduce number of fields to 2.
572 (rtx_def): Add u2.symbol_ref_flags.
573 (SYMBOL_REF_FLAGS): Use it.
574 (SYMBOL_REF_DATA, SET_SYMBOL_REF_DECL, SYMBOL_REF_DECL)
575 (SET_SYMBOL_REF_CONSTANT, SYMBOL_REF_CONSTANT): Lower index.
576 * gengtype.c (adjust_field_rtx_def): Remove SYMBOL_REF_FLAGS handling.
577 Lower index of SYMBOL_REF_DATA.
578 * print-rtl.c (print_rtx): Lower index for SYMBOL_REF_DATA.
579 Print SYMBOL_REF_FLAGS at the same time.
580 * genattrtab.c (attr_rtx_1): Only initialize 1 "0" SYMBOL_REF field.
582 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
584 * rtl.def (VAR_LOCATION): Remove "i" field.
585 * rtl.h (rtx_def): Add u2.var_location_status.
586 (PAT_VAR_LOCATION_STATUS): Use it.
587 (gen_rtx_VAR_LOCATION): Declare.
588 * gengenrtl.c (excluded_rtx): Add VAR_LOCATION.
589 * emit-rtl.c (gen_rtx_VAR_LOCATION): New function.
590 * var-tracking.c (emit_note_insn_var_location): Remove casts.
592 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
594 * rtl.def (scratch): Fix outdated comment and remove "0" field.
595 * gengtype.c (adjust_field_rtx_def): Update accordingly.
597 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
599 * rtl.def (DEBUG_INSN, INSN, JUMP_INSN, CALL_INSN, JUMP_TABLE_DATA)
600 (BARRIER, CODE_LABEL, NOTE): Remove first "i" field.
601 * rtl.h (rtx_def): Add insn_uid to u2 field.
602 (RTX_FLAG_CHECK8): Delete in favor of...
603 (RTL_INSN_CHAIN_FLAG_CHECK): ...this new macro.
604 (INSN_DELETED_P): Update accordingly.
605 (INSN_UID): Use u2.insn_uid.
606 (INSN_CHAIN_CODE_P): Define.
607 (PREV_INSN, NEXT_INSN, BLOCK_FOR_INSN, PATTERN, INSN_LOCATION)
608 (INSN_CODE, REG_NOTES, CALL_INSN_FUNCTION_USAGE, CODE_LABEL_NUMBER)
609 (NOTE_DATA, NOTE_DELETED_LABEL_NAME, NOTE_BLOCK, NOTE_EH_HANDLER)
610 (NOTE_BASIC_BLOCK, NOTE_VAR_LOCATION, NOTE_CFI, NOTE_LABEL_NUMBER)
611 (NOTE_KIND, LABEL_NAME, LABEL_NUSES, JUMP_LABEL, LABEL_REFS): Lower
613 * print-rtl.c (print_rtx): Print INSN_UIDs before the main loop.
614 Update indices for insn-chain rtxes.
615 * gengtype.c (gen_rtx_next): Adjust test for insn-chain rtxes.
616 (adjust_field_rtx_def): Lower '0' indices for all insn-chain rtxes.
617 * emit-rtl.c (gen_label_rtx): Update gen_rtx_LABEL call.
618 * caller-save.c (init_caller_save): Update gen_rtx_INSN calls.
619 * combine.c (try_combine): Likewise.
620 * ira.c (setup_prohibited_mode_move_regs): Likewise.
622 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
624 * rtl.def (REG): Remove middle field.
625 * rtl.h (rtx_def): Add orignal_regno to u2.
626 (ORIGINAL_REGNO): Use it instead of field 1.
627 (REG_ATTRS): Lower field index accordingly.
628 * gengtype.c (adjust_field_rtx_def): Remove handling of
629 ORIGINAL_REGNO. Move REG_ATTRS index down.
630 * print-rtl.c (print_rtx): Move ORIGINAL_REGNO handling to the
631 code that prints the REGNO.
633 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
635 * print-rtl.c (print_rtx): Guard whole '0' block with ifndef
638 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
640 * rtl.h (rtx_def): Mark u2 as GTY ((skip)).
642 2014-05-13 Bin Cheng <bin.cheng@arm.com>
644 * tree-ssa-loop-ivopts.c (contain_complex_addr_expr): New.
645 (alloc_iv): Lower base expressions containing ADDR_EXPR.
647 2014-05-13 Ian Bolton <ian.bolton@arm.com>
649 * config/aarch64/aarch64-protos.h
650 (aarch64_hard_regno_caller_save_mode): New prototype.
651 * config/aarch64/aarch64.c (aarch64_hard_regno_caller_save_mode):
653 * config/aarch64/aarch64.h (HARD_REGNO_CALLER_SAVE_MODE): New macro.
655 2014-05-13 Christian Bruel <christian.bruel@st.com>
657 * target.def (mode_switching): New hook vector.
658 (mode_emit, mode_needed, mode_after, mode_entry): New hooks.
659 (mode_exit, modepriority_to_mode): Likewise.
660 * mode-switching.c (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Hookify.
661 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
662 * target.h: Include tm.h and hard-reg-set.h.
663 * doc/tm.texi.in (EMIT_MODE_SET, MODE_NEEDED, MODE_AFTER, MODE_ENTRY)
664 (MODE_EXIT, MODE_PRIORITY_TO_MODE): Delete and hookify.
665 * doc/tm.texi Regenerate.
666 * config/sh/sh.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
667 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
668 * config/sh/sh.c (sh_emit_mode_set, sh_mode_priority): Hookify.
669 (sh_mode_needed, sh_mode_after, sh_mode_entry, sh_mode_exit): Likewise.
670 * config/i386/i386.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
671 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
672 * config/i386/i386-protos.h (ix86_mode_needed, ix86_mode_after)
673 (ix86_mode_entrym, ix86_emit_mode_set): Remove external declaration.
674 * config/i386/i386.c (ix86_mode_needed, ix86_mode_after,
675 (ix86_mode_exit, ix86_mode_entry, ix86_mode_priority)
676 (ix86_emit_mode_set): Hookify.
677 * config/epiphany/epiphany.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY):
679 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
680 * config/epiphany/epiphany-protos.h (epiphany_mode_needed)
681 (emit_set_fp_mode, epiphany_mode_entry_exit, epiphany_mode_after)
682 (epiphany_mode_priority_to_mode): Remove declaration.
683 * config/epiphany/epiphany.c (emit_set_fp_mode): Hookify.
684 (epiphany_mode_needed, epiphany_mode_priority_to_mode): Likewise.
685 (epiphany_mode_entry, epiphany_mode_exit, epiphany_mode_after):
687 (epiphany_mode_priority_to_mode): Change priority type. Hookify.
688 (epiphany_mode_needed, epiphany_mode_entry_exit): Hookify.
689 (epiphany_mode_after, epiphany_mode_entry, emit_set_fp_mode): Hookify.
691 2014-05-13 Jakub Jelinek <jakub@redhat.com>
694 * config/i386/i386.c (ix86_expand_set_or_movmem): If count_exp
695 is const0_rtx, return immediately. Don't test count == 0 when
698 2014-05-13 Zhenqiang Chen <zhenqiang.chen@linaro.org>
700 * Makefile.in: add shrink-wrap.o.
701 * config/i386/i386.c: include "shrink-wrap.h"
702 * function.c: Likewise.
703 (requires_stack_frame_p, next_block_for_reg,
704 move_insn_for_shrink_wrap, prepare_shrink_wrap,
705 dup_block_and_redirect): Move to shrink-wrap.c
706 (thread_prologue_and_epilogue_insns): Extract three code segments
707 as functions in shrink-wrap.c
708 * function.h: Move #ifdef HAVE_simple_return ... #endif block to
710 * shrink-wrap.c: New file.
711 * shrink-wrap.h: New file.
713 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
715 * doc/extend.texi: Reflect current numbers of pragmas. Remove
716 reference to Solaris.
718 2014-05-12 Mike Stump <mikestump@comcast.net>
721 * genattrtab.c (filename): Add.
722 (convert_set_attr_alternative): Improve error message.
723 (check_defs): Restore read_md_filename for error messages.
724 (gen_insn): Save filename.
726 2014-05-12 Dimitris Papavasiliou <dpapavas@gmail.com>
728 * doc/invoke.texi: Document new switches -Wno-shadow-ivar,
729 -fno-local-ivars and -fivar-visibility.
730 * c-family/c.opt: Make -Wshadow also implicitly enable
733 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
735 * doc/tm.texi: Remove reference to deleted macro.
736 * doc/tm.texi.in: Likewise.
738 2014-05-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
741 * config/avr/avr.c (avr_out_store_psi): Use correct constant
744 2014-05-12 Georg-Johann Lay <avr@gjlay.de>
746 * config/arm/arm.h (License): Add GCC Runtime Library Exception.
747 * config/arm/aout.h (License): Same.
748 * config/arm/bpabi.h (License): Same.
749 * config/arm/elf.h (License): Same.
750 * config/arm/linux-elf.h (License): Same.
751 * config/arm/linux-gas.h (License): Same.
752 * config/arm/netbsd-elf.h (License): Same.
753 * config/arm/uclinux-eabi.h (License): Same.
754 * config/arm/uclinux-elf.h (License): Same.
755 * config/arm/vxworks.h (License): Same.
757 2014-05-11 Jakub Jelinek <jakub@redhat.com>
759 * tree.h (OMP_CLAUSE_LINEAR_STMT): Define.
760 * tree.c (omp_clause_num_ops): Increase OMP_CLAUSE_LINEAR
761 number of operands to 3.
762 (walk_tree_1): Walk all operands of OMP_CLAUSE_LINEAR.
763 * tree-nested.c (convert_nonlocal_omp_clauses,
764 convert_local_omp_clauses): Handle OMP_CLAUSE_DEPEND.
765 * gimplify.c (gimplify_scan_omp_clauses): Handle
766 OMP_CLAUSE_LINEAR_STMT.
767 * omp-low.c (lower_rec_input_clauses): Fix typo.
768 (maybe_add_implicit_barrier_cancel, lower_omp_1): Add
769 cast between Fortran boolean_type_node and C _Bool if
772 2014-05-11 Richard Sandiford <rdsandiford@googlemail.com>
774 PR tree-optimization/61136
775 * wide-int.h (multiple_of_p): Define a version that doesn't return
777 * fold-const.c (extract_muldiv_1): Use wi::multiple_of_p instead of an
778 integer_zerop/const_binop pair.
779 (multiple_of_p): Likewise, converting both operands to widest_int
782 2014-05-09 Teresa Johnson <tejohnson@google.com>
784 * cgraphunit.c (analyze_functions): Use correct dump file.
786 2014-05-09 Florian Weimer <fweimer@redhat.com>
788 * cfgexpand.c (stack_protect_decl_p): New function, extracted from
790 (stack_protect_return_slot_p): New function.
791 (expand_used_vars): Call stack_protect_decl_p and
792 stack_protect_return_slot_p for -fstack-protector-strong.
794 2014-05-09 Kenneth Zadeck <zadeck@naturalbridge.com>
797 * fold-const.c (fold_binary_loc): Changed width of mask.
799 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
801 * config/avr/avr-fixed.md (round<mode>3): Use -1U instead of -1 in
802 unsigned int initializers for regno_in, regno_out.
804 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
807 * config/avr/avr.md (cc): Add new attribute set_vzn.
808 (addqi3, addqq3, adduqq3, subqi3, subqq3, subuqq3, negqi2) [cc]:
809 Set cc insn attribute to set_vzn instead of set_zn for alternatives
810 with INC, DEC or NEG.
811 * config/avr/avr.c (avr_notice_update_cc): Handle SET_VZN.
812 (avr_out_plus_1): ADIW sets cc0 to CC_SET_CZN.
813 INC, DEC and ADD+ADC set cc0 to CC_CLOBBER.
815 2014-05-09 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
818 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
820 * wide-int.cc (UTItype): Define.
821 (UDWtype): Define for appropriate W_TYPE_SIZE.
823 2014-05-09 Richard Biener <rguenther@suse.de>
825 * Makefile.in (GTFILES): Remove tree-ssa-propagate.c.
826 * tree-ssa-propagate.c: Do not include gt-tree-ssa-propagate.h.
827 (interesting_ssa_edges, varying_ssa_edges): Move out of GC space.
828 (add_ssa_edge, process_ssa_edge_worklist, ssa_prop_init,
829 ssa_propagate): Adjust.
831 2014-05-08 Jeff Law <law@redhat.com>
833 PR tree-optimization/61009
834 * tree-ssa-threadedge.c (thread_through_normal_block): Return a
835 tri-state rather than a boolean. When a block is too big to
836 thread through, inform caller via negative return value.
837 (thread_across_edge): If a block was too big for normal threading,
838 then it's too big for a joiner too, so remove temporary equivalences
839 and return immediately.
841 2014-05-08 Manuel López-Ibáñez <manu@gcc.gnu.org>
842 Matthias Klose <doko@ubuntu.com>
845 * optc-gen.awk: Fix option handling for -Wunused-parameter.
847 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
850 * config/i386/i386.c (PTA_HASWELL): Remove PTA_RTM.
852 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
855 * config/alpha/alpha.c: Include gimple-iterator.h.
856 (alpha_gimple_fold_builtin): New function. Move
857 ALPHA_BUILTIN_UMULH folding from ...
858 (alpha_fold_builtin): ... here.
859 (TARGET_GIMPLE_FOLD_BUILTIN): New define.
861 2014-05-08 Wei Mi <wmi@google.com>
864 * config/i386/i386.c (ix86_compute_frame_layout): Update
865 preferred_stack_boundary for call, expanded from tls descriptor.
866 * config/i386/i386.md (*tls_global_dynamic_32_gnu): Update RTX
867 to depend on SP register.
868 (*tls_local_dynamic_base_32_gnu): Ditto.
869 (*tls_local_dynamic_32_once): Ditto.
870 (tls_global_dynamic_64_<mode>): Set
871 ix86_tls_descriptor_calls_expanded_in_cfun.
872 (tls_local_dynamic_base_64_<mode>): Ditto.
873 (tls_global_dynamic_32): Set
874 ix86_tls_descriptor_calls_expanded_in_cfun. Update RTX
875 to depend on SP register.
876 (tls_local_dynamic_base_32): Ditto.
878 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
880 * config/arm/arm_neon.h: Update comment.
881 * config/arm/neon-docgen.ml: Delete.
882 * config/arm/neon-gen.ml: Delete.
883 * doc/arm-neon-intrinsics.texi: Update comment.
885 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
887 * config/arm/arm_neon_builtins.def (vadd, vsub): Only define the v2sf
889 (vand, vorr, veor, vorn, vbic): Remove.
890 * config/arm/neon.md (neon_vadd, neon_vsub, neon_vadd_unspec): Adjust
892 (neon_vsub_unspec): Likewise.
893 (neon_vorr, neon_vand, neon_vbic, neon_veor, neon_vorn): Remove.
895 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
897 * config/arm/arm_neon.h (vadd_s8): GNU C implementation
898 (vadd_s16): Likewise.
899 (vadd_s32): Likewise.
900 (vadd_f32): Likewise.
902 (vadd_u16): Likewise.
903 (vadd_u32): Likewise.
904 (vadd_s64): Likewise.
905 (vadd_u64): Likewise.
906 (vaddq_s8): Likewise.
907 (vaddq_s16): Likewise.
908 (vaddq_s32): Likewise.
909 (vaddq_s64): Likewise.
910 (vaddq_f32): Likewise.
911 (vaddq_u8): Likewise.
912 (vaddq_u16): Likewise.
913 (vaddq_u32): Likewise.
914 (vaddq_u64): Likewise.
916 (vmul_s16): Likewise.
917 (vmul_s32): Likewise.
918 (vmul_f32): Likewise.
920 (vmul_u16): Likewise.
921 (vmul_u32): Likewise.
923 (vmulq_s8): Likewise.
924 (vmulq_s16): Likewise.
925 (vmulq_s32): Likewise.
926 (vmulq_f32): Likewise.
927 (vmulq_u8): Likewise.
928 (vmulq_u16): Likewise.
929 (vmulq_u32): Likewise.
931 (vsub_s16): Likewise.
932 (vsub_s32): Likewise.
933 (vsub_f32): Likewise.
935 (vsub_u16): Likewise.
936 (vsub_u32): Likewise.
937 (vsub_s64): Likewise.
938 (vsub_u64): Likewise.
939 (vsubq_s8): Likewise.
940 (vsubq_s16): Likewise.
941 (vsubq_s32): Likewise.
942 (vsubq_s64): Likewise.
943 (vsubq_f32): Likewise.
944 (vsubq_u8): Likewise.
945 (vsubq_u16): Likewise.
946 (vsubq_u32): Likewise.
947 (vsubq_u64): Likewise.
949 (vand_s16): Likewise.
950 (vand_s32): Likewise.
952 (vand_u16): Likewise.
953 (vand_u32): Likewise.
954 (vand_s64): Likewise.
955 (vand_u64): Likewise.
956 (vandq_s8): Likewise.
957 (vandq_s16): Likewise.
958 (vandq_s32): Likewise.
959 (vandq_s64): Likewise.
960 (vandq_u8): Likewise.
961 (vandq_u16): Likewise.
962 (vandq_u32): Likewise.
963 (vandq_u64): Likewise.
965 (vorr_s16): Likewise.
966 (vorr_s32): Likewise.
968 (vorr_u16): Likewise.
969 (vorr_u32): Likewise.
970 (vorr_s64): Likewise.
971 (vorr_u64): Likewise.
972 (vorrq_s8): Likewise.
973 (vorrq_s16): Likewise.
974 (vorrq_s32): Likewise.
975 (vorrq_s64): Likewise.
976 (vorrq_u8): Likewise.
977 (vorrq_u16): Likewise.
978 (vorrq_u32): Likewise.
979 (vorrq_u64): Likewise.
981 (veor_s16): Likewise.
982 (veor_s32): Likewise.
984 (veor_u16): Likewise.
985 (veor_u32): Likewise.
986 (veor_s64): Likewise.
987 (veor_u64): Likewise.
988 (veorq_s8): Likewise.
989 (veorq_s16): Likewise.
990 (veorq_s32): Likewise.
991 (veorq_s64): Likewise.
992 (veorq_u8): Likewise.
993 (veorq_u16): Likewise.
994 (veorq_u32): Likewise.
995 (veorq_u64): Likewise.
997 (vbic_s16): Likewise.
998 (vbic_s32): Likewise.
1000 (vbic_u16): Likewise.
1001 (vbic_u32): Likewise.
1002 (vbic_s64): Likewise.
1003 (vbic_u64): Likewise.
1004 (vbicq_s8): Likewise.
1005 (vbicq_s16): Likewise.
1006 (vbicq_s32): Likewise.
1007 (vbicq_s64): Likewise.
1008 (vbicq_u8): Likewise.
1009 (vbicq_u16): Likewise.
1010 (vbicq_u32): Likewise.
1011 (vbicq_u64): Likewise.
1012 (vorn_s8): Likewise.
1013 (vorn_s16): Likewise.
1014 (vorn_s32): Likewise.
1015 (vorn_u8): Likewise.
1016 (vorn_u16): Likewise.
1017 (vorn_u32): Likewise.
1018 (vorn_s64): Likewise.
1019 (vorn_u64): Likewise.
1020 (vornq_s8): Likewise.
1021 (vornq_s16): Likewise.
1022 (vornq_s32): Likewise.
1023 (vornq_s64): Likewise.
1024 (vornq_u8): Likewise.
1025 (vornq_u16): Likewise.
1026 (vornq_u32): Likewise.
1027 (vornq_u64): Likewise.
1029 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1031 * wide-int.cc (UTItype): Define.
1032 (UDWtype): Define for appropriate W_TYPE_SIZE.
1034 2014-05-08 Marc Glisse <marc.glisse@inria.fr>
1036 PR tree-optimization/59100
1037 * tree-ssa-phiopt.c: Include tree-inline.h.
1038 (neutral_element_p, absorbing_element_p): New functions.
1039 (value_replacement): Handle conditional binary operations with a
1040 neutral or absorbing element.
1042 2014-05-08 Richard Biener <rguenther@suse.de>
1044 * tree-ssa-sccvn.c (vn_get_expr_for): Valueize operands before
1045 folding the expression.
1046 (valueize_expr): Remove.
1047 (visit_reference_op_load): Do not valueize the result of
1049 (simplify_binary_expression): Likewise.
1050 (simplify_unary_expression): Likewise.
1052 2014-05-08 Richard Biener <rguenther@suse.de>
1054 * gimplify.c (gimplify_call_expr): Use saved fnptrtype for
1055 looking at TYPE_ARG_TYPES.
1057 2014-05-08 Richard Biener <rguenther@suse.de>
1059 * gimple-fold.c (gimple_fold_stmt_to_constant_1): Remove
1060 pointer propagation special-case.
1062 2014-05-08 Bin Cheng <bin.cheng@arm.com>
1064 * tree-affine.c (tree_to_aff_combination): Handle MEM_REF for
1065 core part of address expressions.
1067 2014-05-08 Alan Modra <amodra@gmail.com>
1070 * config/rs6000/rs6000.c (expand_block_move): Allow 64-bit
1071 loads and stores when -mno-strict-align at any alignment.
1072 (expand_block_clear): Similarly. Also correct calculation of
1075 2014-05-07 Thomas Preud'homme <thomas.preudhomme@arm.com>
1078 * tree-complex.c (expand_complex_move): Keep line info when expanding
1080 * tree-ssa-uninit.c (warn_uninit): New argument. Ignore assignment
1081 of complex expression. Use new argument to display correct location
1082 for values coming from phi statement.
1083 (warn_uninitialized_vars): Adapt to new signature of warn_uninit.
1084 (warn_uninitialized_phi): Pass location of phi argument to
1086 * tree-ssa.c (ssa_undefined_value_p): For SSA_NAME initialized by a
1087 COMPLEX_EXPR, recurse on each part of the COMPLEX_EXPR.
1089 2014-05-07 Segher Boessenkool <segher@kernel.crashing.org>
1091 * config/rs6000/predicates.md (indexed_address_mem): New.
1092 * config/rs6000/rs6000.md (type): Remove load_ext, load_ext_u,
1093 load_ext_ux, load_ux, load_u, store_ux, store_u, fpload_ux, fpload_u,
1094 fpstore_ux, fpstore_u.
1095 (sign_extend, indexed, update): New.
1096 (cell_micro): Adjust.
1097 (*zero_extend<mode>di2_internal1, *zero_extendsidi2_lfiwzx,
1098 *extendsidi2_lfiwax, *extendsidi2_nocell, *extendsfdf2_fpr,
1099 *movsi_internal1, *movsi_internal1_single, *movhi_internal,
1100 *movqi_internal, *movcc_internal1, mov<mode>_hardfloat,
1101 *mov<mode>_softfloat, *mov<mode>_hardfloat32, *mov<mode>_hardfloat64,
1102 *mov<mode>_softfloat64, *movdi_internal32, *movdi_internal64,
1103 *mov<mode>_string, *ldmsi8, *ldmsi7, *ldmsi6, *ldmsi5, *ldmsi4,
1104 *ldmsi3, *stmsi8, *stmsi7, *stmsi6, *stmsi5, *stmsi4, *stmsi3,
1105 *movdi_update1, movdi_<mode>_update, movdi_<mode>_update_stack,
1106 *movsi_update1, *movsi_update2, movsi_update, movsi_update_stack,
1107 *movhi_update1, *movhi_update2, *movhi_update3, *movhi_update4,
1108 *movqi_update1, *movqi_update2, *movqi_update3, *movsf_update1,
1109 *movsf_update2, *movsf_update3, *movsf_update4, *movdf_update1,
1110 *movdf_update2, load_toc_aix_si, load_toc_aix_di, probe_stack_<mode>,
1111 *stmw, *lmw, as well as 10 anonymous patterns): Adjust.
1113 * config/rs6000/dfp.md (movsd_store, movsd_load): Adjust.
1114 * config/rs6000/vsx.md (*vsx_movti_32bit, *vsx_extract_<mode>_load,
1115 *vsx_extract_<mode>_store): Adjust.
1116 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_microcoded_insn,
1117 is_cracked_insn, insn_must_be_first_in_group,
1118 insn_must_be_last_in_group): Adjust.
1120 * config/rs6000/40x.md (ppc403-load, ppc403-store, ppc405-float):
1122 * config/rs6000/440.md (ppc440-load, ppc440-store, ppc440-fpload,
1123 ppc440-fpstore): Adjust.
1124 * config/rs6000/476.md (ppc476-load, ppc476-store, ppc476-fpload,
1125 ppc476-fpstore): Adjust.
1126 * config/rs6000/601.md (ppc601-load, ppc601-store, ppc601-fpload,
1127 ppc601-fpstore): Adjust.
1128 * config/rs6000/603.md (ppc603-load, ppc603-store, ppc603-fpload):
1130 * config/rs6000/6xx.md (ppc604-load, ppc604-store, ppc604-fpload):
1132 * config/rs6000/7450.md (ppc7450-load, ppc7450-store, ppc7450-fpload,
1133 ppc7450-fpstore): Adjust.
1134 * config/rs6000/7xx.md (ppc750-load, ppc750-store): Adjust.
1135 * config/rs6000/8540.md (ppc8540_load, ppc8540_store): Adjust.
1136 * config/rs6000/a2.md (ppca2-load, ppca2-fp-load, ppca2-fp-store):
1138 * config/rs6000/cell.md (cell-load, cell-load-ux, cell-load-ext,
1139 cell-fpload, cell-fpload-update, cell-store, cell-store-update,
1140 cell-fpstore, cell-fpstore-update): Adjust.
1141 * config/rs6000/e300c2c3.md (ppce300c3_load, ppce300c3_fpload,
1142 ppce300c3_store, ppce300c3_fpstore): Adjust.
1143 * config/rs6000/e500mc.md (e500mc_load, e500mc_fpload, e500mc_store,
1144 e500mc_fpstore): Adjust.
1145 * config/rs6000/e500mc64.md (e500mc64_load, e500mc64_fpload,
1146 e500mc64_store, e500mc64_fpstore): Adjust.
1147 * config/rs6000/e5500.md (e5500_load, e5500_fpload, e5500_store,
1148 e5500_fpstore): Adjust.
1149 * config/rs6000/e6500.md (e6500_load, e6500_fpload, e6500_store,
1150 e6500_fpstore): Adjust.
1151 * config/rs6000/mpc.md (mpccore-load, mpccore-store, mpccore-fpload):
1153 * config/rs6000/power4.md (power4-load, power4-load-ext,
1154 power4-load-ext-update, power4-load-ext-update-indexed,
1155 power4-load-update-indexed, power4-load-update, power4-fpload,
1156 power4-fpload-update, power4-store, power4-store-update,
1157 power4-store-update-indexed, power4-fpstore, power4-fpstore-update):
1159 * config/rs6000/power5.md (power5-load, power5-load-ext,
1160 power5-load-ext-update, power5-load-ext-update-indexed,
1161 power5-load-update-indexed, power5-load-update, power5-fpload,
1162 power5-fpload-update, power5-store, power5-store-update,
1163 power5-store-update-indexed, power5-fpstore, power5-fpstore-update):
1165 * config/rs6000/power6.md (power6-load, power6-load-ext,
1166 power6-load-update, power6-load-update-indexed,
1167 power6-load-ext-update, power6-load-ext-update-indexed, power6-fpload,
1168 power6-fpload-update, power6-store, power6-store-update,
1169 power6-store-update-indexed, power6-fpstore, power6-fpstore-update):
1171 * config/rs6000/power7.md (power7-load, power7-load-ext,
1172 power7-load-update, power7-load-update-indexed,
1173 power7-load-ext-update, power7-load-ext-update-indexed, power7-fpload,
1174 power7-fpload-update, power7-store, power7-store-update,
1175 power7-store-update-indexed, power7-fpstore, power7-fpstore-update):
1177 * config/rs6000/power8.md (power8-load, power8-load-update,
1178 power8-load-ext, power8-load-ext-update, power8-fpload,
1179 power8-fpload-update, power8-store, power8-store-update-indexed,
1180 power8-fpstore, power8-fpstore-update): Adjust.
1181 * config/rs6000/rs64.md (rs64a-load, rs64a-store, rs64a-fpload):
1183 * config/rs6000/titan.md (titan_lsu_load, titan_lsu_fpload,
1184 titan_lsu_store, titan_lsu_fpstore): Adjust.
1185 * config/rs6000/xfpu.md (fp-load, fp-store): Adjust.
1187 2014-05-07 Oleg Endo <olegendo@gcc.gnu.org>
1190 * config/sh/sh-mem.cc (sh_expand_strlen): Use loop when emitting
1191 unrolled byte insns. Emit address increments after move insns.
1193 2014-05-07 David Malcolm <dmalcolm@redhat.com>
1195 * gimple.h (gimple_builtin_call_types_compatible_p): Accept a
1196 const_gimple, rather than a gimple.
1197 (gimple_call_builtin_p): Likewise, for the three variants.
1199 * gimple.c (gimple_builtin_call_types_compatible_p): Likewise.
1200 (gimple_call_builtin_p): Likewise, for the three variants.
1202 2014-05-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
1204 PR tree-optimization/61095
1205 * tree-ssanames.c (get_nonzero_bits): Fix type extension in wi::shwi.
1207 2014-05-07 Richard Biener <rguenther@suse.de>
1209 PR tree-optimization/61034
1210 * tree-ssa-alias.c (call_may_clobber_ref_p_1): Export.
1211 (maybe_skip_until): Use translate to take into account
1212 lattices when trying to do disambiguations.
1213 (get_continuation_for_phi_1): Likewise.
1214 (get_continuation_for_phi): Adjust for added translate arguments.
1215 (walk_non_aliased_vuses): Likewise.
1216 * tree-ssa-alias.h (get_continuation_for_phi): Adjust prototype.
1217 (walk_non_aliased_vuses): Likewise.
1218 (call_may_clobber_ref_p_1): Declare.
1219 * tree-ssa-sccvn.c (vn_reference_lookup_3): Also disambiguate against
1220 calls. Stop early if we are only supposed to disambiguate.
1221 * tree-ssa-pre.c (translate_vuse_through_block): Adjust.
1223 2014-05-07 Joern Rennecke <joern.rennecke@embecosm.com>
1225 * config/epiphany/epiphany.c (epiphany_handle_interrupt_attribute):
1226 Emit an error when the function has arguments.
1228 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
1230 * cfgloop.h (unswitch_loops): Remove.
1231 * doc/passes.texi: Remove references to loop-unswitch.c
1232 * timevar.def (TV_LOOP_UNSWITCH): Remove.
1234 2014-05-07 Evgeny Stupachenko <evstupac@gmail.com>
1236 * tree-vect-data-refs.c (vect_grouped_load_supported): New
1237 check for loads group of length 3.
1238 (vect_permute_load_chain): New permutations for loads group of
1240 * tree-vect-stmts.c (vect_model_load_cost): Change cost
1241 of vec_perm_shuffle for the new permutations.
1243 2014-05-07 Alan Lawrence <alan.lawrence@arm.com>
1245 * config/aarch64/arm_neon.h (vtrn1_f32, vtrn1_p8, vtrn1_p16, vtrn1_s8,
1246 vtrn1_s16, vtrn1_s32, vtrn1_u8, vtrn1_u16, vtrn1_u32, vtrn1q_f32,
1247 vtrn1q_f64, vtrn1q_p8, vtrn1q_p16, vtrn1q_s8, vtrn1q_s16, vtrn1q_s32,
1248 vtrn1q_s64, vtrn1q_u8, vtrn1q_u16, vtrn1q_u32, vtrn1q_u64, vtrn2_f32,
1249 vtrn2_p8, vtrn2_p16, vtrn2_s8, vtrn2_s16, vtrn2_s32, vtrn2_u8,
1250 vtrn2_u16, vtrn2_u32, vtrn2q_f32, vtrn2q_f64, vtrn2q_p8, vtrn2q_p16,
1251 vtrn2q_s8, vtrn2q_s16, vtrn2q_s32, vtrn2q_s64, vtrn2q_u8, vtrn2q_u16,
1252 vtrn2q_u32, vtrn2q_u64): Replace temporary asm with __builtin_shuffle.
1254 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
1256 * loop-unswitch.c: Delete.
1258 2014-05-07 Richard Biener <rguenther@suse.de>
1260 * config.gcc: Always set need_64bit_hwint to yes.
1262 2014-05-07 Chung-Ju Wu <jasonwucj@gmail.com>
1264 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Have it in favor
1265 of using optimize_size.
1267 2014-05-06 Mike Stump <mikestump@comcast.net>
1269 * wide-int.h (wi::int_traits <HOST_WIDE_INT>): Always define.
1271 2014-05-06 Joseph Myers <joseph@codesourcery.com>
1273 * config/i386/sse.md (*mov<mode>_internal)
1274 (*<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>)
1275 (*<sse2_avx_avx512f>_loaddqu<mode><mask_name>)
1276 (<sse>_andnot<mode>3, <code><mode>3, *andnot<mode>3)
1277 (*<code><mode>3, *andnot<mode>3<mask_name>)
1278 (<mask_codefor><code><mode>3<mask_name>): Only consider
1279 TARGET_SSE_PACKED_SINGLE_INSN_OPTIMAL for modes of size 16.
1281 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
1284 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
1286 * lra-constraints.c (valid_address_p): Move earlier in file.
1287 Add a constraint argument to the address_info version.
1288 (satisfies_memory_constraint_p): New function.
1289 (satisfies_address_constraint_p): Likewise.
1290 (process_alt_operands, curr_insn_transform): Use them.
1291 (process_address): Pass the constraint to valid_address_p when
1292 checking address operands.
1294 2014-05-06 Richard Sandiford <r.sandiford@uk.ibm.com>
1296 * lto-cgraph.c (compute_ltrans_boundary): Make node variables local
1297 to their respective blocks. Fix inadvertent use of "node".
1299 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
1301 * emit-rtl.c (init_derived_machine_modes): New functionm, split
1303 (init_emit_once): ...here.
1304 * rtl.h (init_derived_machine_modes): Declare.
1305 * toplev.c (do_compile): Call it even if no_backend.
1307 2014-05-06 Kenneth Zadeck <zadeck@naturalbridge.com>
1308 Mike Stump <mikestump@comcast.net>
1309 Richard Sandiford <rdsandiford@googlemail.com>
1310 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1312 * alias.c (ao_ref_from_mem): Use wide-int interfaces.
1313 (rtx_equal_for_memref_p): Update comment.
1314 (adjust_offset_for_component_ref): Use wide-int interfaces.
1315 * builtins.c (get_object_alignment_2): Likewise.
1316 (c_readstr): Likewise.
1317 (target_char_cast): Add comment.
1318 (determine_block_size): Use wide-int interfaces.
1319 (expand_builtin_signbit): Likewise.
1320 (fold_builtin_int_roundingfn): Likewise.
1321 (fold_builtin_bitop): Likewise.
1322 (fold_builtin_bswap): Likewise.
1323 (fold_builtin_logarithm): Use signop.
1324 (fold_builtin_pow): Likewise.
1325 (fold_builtin_memory_op): Use wide-int interfaces.
1326 (fold_builtin_object_size): Likewise.
1327 * cfgloop.c (alloc_loop): Initialize nb_iterations_upper_bound and
1328 nb_iterations_estimate.
1329 (record_niter_bound): Use wide-int interfaces.
1330 (get_estimated_loop_iterations_int): Likewise.
1331 (get_estimated_loop_iterations): Likewise.
1332 (get_max_loop_iterations): Likewise.
1333 * cfgloop.h: Include wide-int.h.
1334 (struct nb_iter_bound): Change bound to widest_int.
1335 (struct loop): Change nb_iterations_upper_bound and
1336 nb_iterations_estimate to widest_int.
1337 (record_niter_bound): Switch to use widest_int.
1338 (get_estimated_loop_iterations): Likewise.
1339 (get_max_loop_iterations): Likewise.
1340 (gcov_type_to_double_int): Rename to gcov_type_to_wide_int and
1341 update for wide-int.
1342 * cgraph.c (cgraph_add_thunk): Use wide-int interfaces.
1343 * combine.c (try_combine): Likewise.
1344 (subst): Use CONST_SCALAR_INT_P rather than CONST_INT_P.
1345 * config/aarch64/aarch64.c (aapcs_vfp_sub_candidate): Use wide-int
1347 (aarch64_float_const_representable_p): Likewise.
1348 * config/arc/arc.c: Include wide-int.h.
1349 (arc_can_use_doloop_p): Use wide-int interfaces.
1350 * config/arm/arm.c (aapcs_vfp_sub_candidate): Likewise.
1351 (vfp3_const_double_index): Likewise.
1352 * config/avr/avr.c (avr_out_round): Likewise.
1353 (avr_fold_builtin): Likewise.
1354 * config/bfin/bfin.c (bfin_local_alignment): Likewise.
1355 (bfin_can_use_doloop_p): Likewise.
1356 * config/darwin.c (darwin_mergeable_constant_section): Likewise.
1357 (machopic_select_rtx_section): Update to handle CONST_WIDE_INT.
1358 * config/i386/i386.c: Include wide-int.h.
1359 (ix86_data_alignment): Use wide-int interfaces.
1360 (ix86_local_alignment): Likewise.
1361 (ix86_emit_swsqrtsf): Update real_from_integer.
1362 * config/msp430/msp430.c (msp430_attr): Use wide-int interfaces.
1363 * config/nds32/nds32.c (nds32_insert_attributes): Likewise.
1364 * config/rs6000/predicates.md (any_operand): Add const_wide_int.
1365 (zero_constant): Likewise.
1366 (input_operand): Likewise.
1367 (splat_input_operand): Likewise.
1368 (non_logical_cint_operand): Change const_double to const_wide_int.
1369 * config/rs6000/rs6000.c (num_insns_constant): Handle CONST_WIDE_INT.
1370 (easy_altivec_constant): Remove comment.
1371 (paired_expand_vector_init): Use CONSTANT_P.
1372 (rs6000_legitimize_address): Handle CONST_WIDE_INT.
1373 (rs6000_emit_move): Update checks.
1374 (rs6000_aggregate_candidate): Use wide-int interfaces.
1375 (rs6000_expand_ternop_builtin): Likewise.
1376 (rs6000_output_move_128bit): Handle CONST_WIDE_INT.
1377 (rs6000_assemble_integer): Likewise.
1378 (rs6000_hash_constant): Likewise.
1379 (output_toc): Likewise.
1380 (rs6000_rtx_costs): Likewise.
1381 (rs6000_emit_swrsqrt); Update call to real_from_integer.
1382 * config/rs6000/rs6000-c.c: Include wide-int.h.
1383 (altivec_resolve_overloaded_builtin): Use wide-int interfaces.
1384 * config/rs6000/rs6000.h (TARGET_SUPPORTS_WIDE_INT): New.
1385 * config/rs6000/rs6000.md: Use const_scalar_int_operand.
1386 Handle CONST_WIDE_INT.
1387 * config/sol2-c.c (solaris_pragma_align): Change low to unsigned HWI.
1388 Use tree_fits_uhwi_p.
1389 * config/sparc/sparc.c: Include wide-int.h.
1390 (sparc_fold_builtin): Use wide-int interfaces.
1391 * config/vax/vax.c: Include wide-int.h.
1392 (vax_float_literal): Use real_from_integer.
1393 * coretypes.h (struct hwivec_def): New.
1395 (const_hwivec): New.
1396 * cse.c (hash_rtx_cb): Handle CONST_WIDE_INT.
1397 (equiv_constant): Handle CONST_WIDE_INT.
1398 * cselib.c (rtx_equal_for_cselib_1): Use CASE_CONST_UNIQUE.
1399 (cselib_hash_rtx): Handle CONST_WIDE_INT.
1400 * dbxout.c (stabstr_U): Use wide-int interfaces.
1401 (dbxout_type): Update to use cst_fits_shwi_p.
1402 * defaults.h (LOG2_BITS_PER_UNIT): Define.
1403 (TARGET_SUPPORTS_WIDE_INT): Add default.
1404 * dfp.c: Include wide-int.h.
1405 (decimal_real_to_integer2): Use wide-int interfaces and rename to
1406 decimal_real_to_integer.
1407 * dfp.h (decimal_real_to_integer2): Return a wide_int and rename to
1408 decimal_real_to_integer.
1409 * doc/generic.texi (Constant expressions): Update for wide_int.
1410 * doc/rtl.texi (const_double): Likewise.
1411 (const_wide_int, CONST_WIDE_INT, CONST_WIDE_INT_VEC): New.
1412 (CONST_WIDE_INT_NUNITS, CONST_WIDE_INT_ELT): New.
1413 * doc/tm.texi.in (REAL_VALUE_TO_INT): Remove.
1414 (REAL_VALUE_FROM_INT): Remove.
1415 (TARGET_SUPPORTS_WIDE_INT): New.
1416 * doc/tm.texi: Regenerate.
1417 * dojump.c (prefer_and_bit_test): Use wide-int interfaces.
1418 * double-int.h: Include wide-int.h.
1419 (struct wi::int_traits): New.
1420 * dwarf2out.c (get_full_len): New.
1421 (dw_val_equal_p): Add case dw_val_class_wide_int.
1422 (size_of_loc_descr): Likewise.
1423 (output_loc_operands): Likewise.
1424 (insert_double): Remove.
1425 (insert_wide_int): New.
1427 (print_die): Add case dw_val_class_wide_int.
1428 (attr_checksum): Likewise.
1429 (attr_checksum_ordered): Likewise.
1430 (same_dw_val_p): Likewise.
1431 (size_of_die): Likewise.
1432 (value_format): Likewise.
1433 (output_die): Likewise.
1434 (double_int_type_size_in_bits): Rename to offset_int_type_size_in_bits.
1436 (clz_loc_descriptor): Use wide-int interfaces.
1437 (mem_loc_descriptor): Likewise. Handle CONST_WIDE_INT.
1438 (loc_descriptor): Use wide-int interfaces. Handle CONST_WIDE_INT.
1439 (round_up_to_align): Use wide-int interfaces.
1440 (field_byte_offset): Likewise.
1441 (insert_double): Rename to insert_wide_int. Use wide-int interfaces.
1442 (add_const_value_attribute): Handle CONST_WIDE_INT. Update
1443 CONST_DOUBLE handling. Use wide-int interfaces.
1444 (add_bound_info): Use tree_fits_uhwi_p. Use wide-int interfaces.
1445 (gen_enumeration_type_die): Use add_AT_wide.
1446 (hash_loc_operands): Add case dw_val_class_wide_int.
1447 (compare_loc_operands): Likewise.
1448 * dwarf2out.h: Include wide-int.h.
1449 (wide_int_ptr): New.
1450 (enum dw_val_class): Add dw_val_class_wide_int.
1451 (struct dw_val_struct): Add val_wide.
1452 * emit-rtl.c (const_wide_int_htab): New.
1453 (const_wide_int_htab_hash): New.
1454 (const_wide_int_htab_eq): New.
1455 (lookup_const_wide_int): New.
1456 (const_double_htab_hash): Use wide-int interfaces.
1457 (const_double_htab_eq): Likewise.
1458 (rtx_to_double_int): Conditionally compile for wide-int.
1459 (immed_double_int_const): Rename to immed_wide_int_const and
1460 update for wide-int.
1461 (immed_double_const): Conditionally compile for wide-int.
1462 (init_emit_once): Use wide-int interfaces.
1463 * explow.c (plus_constant): Likewise.
1464 * expmed.c (mask_rtx): Move further up file. Use wide-int interfaces.
1465 (lshift_value): Use wide-int interfaces.
1466 (expand_mult): Likewise.
1467 (choose_multiplier): Likewise.
1468 (expand_smod_pow2): Likewise.
1469 (make_tree): Likewise.
1470 * expr.c (convert_modes): Consolidate handling of constants.
1471 Use wide-int interfaces.
1472 (emit_group_load_1): Add note.
1473 (store_expr): Update comment.
1474 (get_inner_reference): Use wide-int interfaces.
1475 (expand_constructor): Update comment.
1476 (expand_expr_real_2): Use wide-int interfaces.
1477 (expand_expr_real_1): Likewise.
1478 (reduce_to_bit_field_precision): Likewise.
1479 (const_vector_from_tree): Likewise.
1480 * final.c: Include wide-int-print.h.
1481 (output_addr_const): Handle CONST_WIDE_INT. Use CONST_DOUBLE_AS_INT_P.
1482 * fixed-value.c: Include wide-int.h.
1483 (fixed_from_string): Use wide-int interfaces.
1484 (fixed_to_decimal): Likewise.
1485 (fixed_convert_from_real): Likewise.
1486 (real_convert_from_fixed): Likewise.
1487 * fold-const.h (mem_ref_offset): Return an offset_int.
1488 (div_if_zero_remainder): Remove code parameter.
1489 * fold-const.c (div_if_zero_remainder): Remove code parameter.
1490 Use wide-int interfaces.
1491 (may_negate_without_overflow_p): Use wide-int interfaces.
1492 (negate_expr_p): Likewise.
1493 (fold_negate_expr): Likewise.
1494 (int_const_binop_1): Likewise.
1495 (const_binop): Likewise.
1496 (fold_convert_const_int_from_int): Likewise.
1497 (fold_convert_const_int_from_real): Likewise.
1498 (fold_convert_const_int_from_fixed): Likewise.
1499 (fold_convert_const_fixed_from_int): Likewise.
1500 (all_ones_mask_p): Take an unsigned size. Use wide-int interfaces.
1501 (sign_bit_p): Use wide-int interfaces.
1502 (make_range_step): Likewise.
1503 (build_range_check): Likewise. Pass an integer of the correct type
1504 instead of using integer_one_node.
1505 (range_predecessor): Pass an integer of the correct type instead
1506 of using integer_one_node.
1507 (range_successor): Likewise.
1508 (merge_ranges): Likewise.
1509 (unextend): Use wide-int interfaces.
1510 (extract_muldiv_1): Likewise.
1511 (fold_div_compare): Likewise.
1512 (fold_single_bit_test): Likewise.
1513 (fold_sign_changed_comparison): Likewise.
1514 (try_move_mult_to_index): Update calls to div_if_zero_remainder.
1515 (fold_plusminus_mult_expr): Use wide-int interfaces.
1516 (native_encode_int): Likewise.
1517 (native_interpret_int): Likewise.
1518 (fold_unary_loc): Likewise.
1519 (pointer_may_wrap_p): Likewise.
1520 (size_low_cst): Likewise.
1521 (mask_with_tz): Likewise.
1522 (fold_binary_loc): Likewise.
1523 (fold_ternary_loc): Likewise.
1524 (multiple_of_p): Likewise.
1525 (tree_call_nonnegative_warnv_p): Update calls to
1526 tree_int_cst_min_precision and real_from_integer.
1527 (fold_negate_const): Use wide-int interfaces.
1528 (fold_abs_const): Likewise.
1529 (fold_relational_const): Use tree_int_cst_lt.
1530 (round_up_loc): Use wide-int interfaces.
1531 * genemit.c (gen_exp): Add CONST_WIDE_INT case.
1532 * gengenrtl.c (excluded_rtx): Add CONST_WIDE_INT case.
1533 * gengtype.c: Remove include of double-int.h.
1534 (do_typedef): Use wide-int interfaces.
1535 (open_base_files): Add wide-int.h.
1536 (main): Add offset_int and widest_int typedefs.
1537 * gengtype-lex.l: Handle "^".
1538 (CXX_KEYWORD): Add "static".
1539 * gengtype-parse.c (require3): New.
1540 (require_template_declaration): Handle constant template arguments
1541 and nested templates.
1542 * gengtype-state.c: Don't include "double-int.h".
1543 * genpreds.c (write_one_predicate_function): Update comment.
1544 (write_tm_constrs_h): Add check for hval and lval use in
1546 * genrecog.c (validate_pattern): Add CONST_WIDE_INT case.
1547 (add_to_sequence): Likewise.
1548 * gensupport.c (struct std_pred_table): Add const_scalar_int_operand
1549 and const_double_operand.
1550 * gimple.c (preprocess_case_label_vec_for_gimple): Use wide-int
1552 * gimple-fold.c (get_base_constructor): Likewise.
1553 (fold_array_ctor_reference): Likewise.
1554 (fold_nonarray_ctor_reference): Likewise.
1555 (fold_const_aggregate_ref_1): Likewise.
1556 (gimple_val_nonnegative_real_p): Likewise.
1557 (gimple_fold_indirect_ref): Likewise.
1558 * gimple-pretty-print.c (dump_ssaname_info): Likewise.
1559 * gimple-ssa-strength-reduction.c: Include wide-int-print.h.
1560 (struct slsr_cand_d): Change index to be widest_int.
1561 (struct incr_info_d): Change incr to be widest_int.
1562 (alloc_cand_and_find_basis): Use wide-int interfaces.
1563 (slsr_process_phi): Likewise.
1564 (backtrace_base_for_ref): Likewise. Return a widest_int.
1565 (restructure_reference): Take a widest_int instead of a double_int.
1566 (slsr_process_ref): Use wide-int interfaces.
1567 (create_mul_ssa_cand): Likewise.
1568 (create_mul_imm_cand): Likewise.
1569 (create_add_ssa_cand): Likewise.
1570 (create_add_imm_cand): Take a widest_int instead of a double_int.
1571 (slsr_process_add): Use wide-int interfaces.
1572 (slsr_process_cast): Likewise.
1573 (slsr_process_copy): Likewise.
1574 (dump_candidate): Likewise.
1575 (dump_incr_vec): Likewise.
1576 (replace_ref): Likewise.
1577 (cand_increment): Likewise. Return a widest_int.
1578 (cand_abs_increment): Likewise.
1579 (replace_mult_candidate): Take a widest_int instead of a double_int.
1580 (replace_unconditional_candidate): Use wide-int interfaces.
1581 (incr_vec_index): Take a widest_int instead of a double_int.
1582 (create_add_on_incoming_edge): Likewise.
1583 (create_phi_basis): Use wide-int interfaces.
1584 (replace_conditional_candidate): Likewise.
1585 (record_increment): Take a widest_int instead of a double_int.
1586 (record_phi_increments): Use wide-int interfaces.
1587 (phi_incr_cost): Take a widest_int instead of a double_int.
1588 (lowest_cost_path): Likewise.
1589 (total_savings): Likewise.
1590 (analyze_increments): Use wide-int interfaces.
1591 (ncd_with_phi): Take a widest_int instead of a double_int.
1592 (ncd_of_cand_and_phis): Likewise.
1593 (nearest_common_dominator_for_cands): Likewise.
1594 (insert_initializers): Use wide-int interfaces.
1595 (all_phi_incrs_profitable): Likewise.
1596 (replace_one_candidate): Likewise.
1597 (replace_profitable_candidates): Likewise.
1598 * godump.c: Include wide-int-print.h.
1599 (go_output_typedef): Use wide-int interfaces.
1600 * graphite-clast-to-gimple.c (gmp_cst_to_tree): Likewise.
1601 * graphite-sese-to-poly.c (tree_int_to_gmp): Likewise.
1602 (build_loop_iteration_domains): Likewise.
1603 * hooks.h: Include wide-int.h rather than double-int.h.
1604 (hook_bool_dint_dint_uint_bool_true): Delete.
1605 (hook_bool_wint_wint_uint_bool_true): Declare.
1606 * hooks.c (hook_bool_dint_dint_uint_bool_true): Removed.
1607 (hook_bool_wint_wint_uint_bool_true): New.
1608 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): Use wide-int
1610 (ubsan_expand_si_overflow_mul_check): Likewise.
1611 * ipa-devirt.c (get_polymorphic_call_info): Likewise.
1612 * ipa-prop.c (compute_complex_assign_jump_func): Likewise.
1613 (get_ancestor_addr_info): Likewise.
1614 (ipa_modify_call_arguments): Likewise.
1615 * loop-doloop.c (doloop_modify): Likewise.
1616 (doloop_optimize): Likewise.
1617 * loop-iv.c (iv_number_of_iterations): Likewise.
1618 * loop-unroll.c (decide_unroll_constant_iterations): Likewise.
1619 (unroll_loop_constant_iterations): Likewise.
1620 (decide_unroll_runtime_iterations): Likewise.
1621 (unroll_loop_runtime_iterations): Likewise.
1622 (decide_peel_simple): Likewise.
1623 (decide_unroll_stupid): Likewise.
1624 * lto-streamer-in.c (streamer_read_wi): Add.
1625 (input_cfg): Use wide-int interfaces.
1626 (lto_input_tree_1): Likewise.
1627 * lto-streamer-out.c (streamer_write_wi): Add.
1628 (hash_tree): Use wide-int interfaces.
1629 (output_cfg): Likewise.
1630 * Makefile.in (OBJS): Add wide-int.o and wide-int-print.o.
1631 (GTFILES): Add wide-int.h and signop.h.
1632 (TAGS): Look for .cc files too.
1633 * omp-low.c (scan_omp_1_op): Use wide-int interfaces.
1634 * optabs.c (expand_subword_shift): Likewise.
1635 (expand_doubleword_shift): Likewise.
1636 (expand_absneg_bit): Likewise.
1637 (expand_copysign_absneg): Likewise.
1638 (expand_copysign_bit): Likewise.
1639 * postreload.c (reload_cse_simplify_set): Likewise.
1640 * predict.c (predict_iv_comparison): Likewise.
1641 * pretty-print.h: Include wide-int-print.h.
1643 * print-rtl.c (print_rtx): Add CONST_WIDE_INT case.
1644 * print-tree.c: Include wide-int-print.h.
1645 (print_node_brief): Use wide-int interfaces.
1646 (print_node): Likewise.
1647 * read-rtl.c (validate_const_wide_int): New.
1648 (read_rtx_code): Add CONST_WIDE_INT case.
1649 * real.c: Include wide-int.h.
1650 (real_to_integer2): Delete.
1651 (real_to_integer): New function, returning a wide_int.
1652 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
1653 (ten_to_ptwo): Update call to real_from_integer.
1654 (real_digit): Likewise.
1655 * real.h: Include signop.h, wide-int.h and insn-modes.h.
1656 (real_to_integer2, REAL_VALUE_FROM_INT, REAL_VALUE_FROM_UNSIGNED_INT)
1657 (REAL_VALUE_TO_INT): Delete.
1658 (real_to_integer): Declare a wide-int form.
1659 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
1660 * recog.c (const_int_operand): Improve comment.
1661 (const_scalar_int_operand): New.
1662 (const_double_operand): Add a separate definition for CONST_WIDE_INT.
1663 * rtlanal.c (commutative_operand_precedence): Handle CONST_WIDE_INT.
1664 (split_double): Likewise.
1665 * rtl.c (DEF_RTL_EXPR): Handle CONST_WIDE_INT.
1666 (rtx_size): Likewise.
1667 (rtx_alloc_stat_v): New.
1668 (rtx_alloc_stat): Now calls rtx_alloc_stat_v.
1669 (cwi_output_hex): New.
1670 (iterative_hash_rtx): Handle CONST_WIDE_INT.
1671 (cwi_check_failed_bounds): New.
1672 * rtl.def (CONST_WIDE_INT): New.
1673 * rtl.h: Include <utility> and wide-int.h.
1674 (struct hwivec_def): New.
1675 (CWI_GET_NUM_ELEM): New.
1676 (CWI_PUT_NUM_ELEM): New.
1677 (struct rtx_def): Add num_elem and hwiv.
1678 (CASE_CONST_SCALAR_INT): Modify for TARGET_SUPPORTS_WIDE_INT.
1679 (CASE_CONST_UNIQUE): Likewise.
1680 (CASE_CONST_ANY): Likewise.
1681 (CONST_SCALAR_INT_P): Likewise.
1682 (CONST_WIDE_INT_P): New.
1684 (HWIVEC_CHECK): New.
1685 (cwi_check_failed_bounds): New.
1687 (HWIVEC_CHECK): New.
1688 (CONST_WIDE_INT_VEC) New.
1689 (CONST_WIDE_INT_NUNITS) New.
1690 (CONST_WIDE_INT_ELT) New.
1691 (rtx_mode_t): New type.
1692 (wi::int_traits <rtx_mode_t>): New.
1694 (wi::min_value): New.
1695 (wi::max_value): New.
1697 (const_wide_int_alloc): New.
1698 (immed_wide_int_const): New.
1699 * sched-vis.c (print_value): Handle CONST_WIDE_INT.
1700 * sel-sched-ir.c (lhs_and_rhs_separable_p): Update comment.
1701 * signop.h: New file.
1702 * simplify-rtx.c (mode_signbit_p): Handle CONST_WIDE_INT.
1703 (simplify_const_unary_operation): Use wide-int interfaces.
1704 (simplify_binary_operation_1): Likewise.
1705 (simplify_const_binary_operation): Likewise.
1706 (simplify_const_relational_operation): Likewise.
1707 (simplify_immed_subreg): Likewise.
1708 * stmt.c (expand_case): Likewise.
1709 * stor-layout.h (set_min_and_max_values_for_integral_type): Take a
1710 signop rather than a bool.
1711 * stor-layout.c (layout_type): Use wide-int interfaces.
1712 (initialize_sizetypes): Update calls to
1713 set_min_and_max_values_for_integral_type.
1714 (set_min_and_max_values_for_integral_type): Take a signop rather
1715 than a bool. Use wide-int interfaces.
1716 (fixup_signed_type): Update accordingly. Remove
1717 HOST_BITS_PER_DOUBLE_INT limit.
1718 (fixup_unsigned_type): Likewise.
1719 * system.h (STATIC_CONSTANT_P): New.
1720 (STATIC_ASSERT): New.
1721 * target.def (can_use_doloop_p): Take widest_ints rather than
1723 * target.h: Include wide-int.h rather than double-int.h.
1724 * targhooks.h (can_use_doloop_if_innermost): Take widest_ints rather
1726 * targhooks.c (default_cxx_get_cookie_size): Use tree_int_cst_lt
1727 rather than INT_CST_LT_UNSIGNED.
1728 (can_use_doloop_if_innermost): Take widest_ints rather than
1730 * tree-affine.c: Include wide-int-print.h.
1731 (double_int_ext_for_comb): Delete.
1732 (wide_int_ext_for_comb): New.
1733 (aff_combination_zero): Use wide-int interfaces.
1734 (aff_combination_const): Take a widest_int instead of a double_int.
1735 (aff_combination_elt): Use wide-int interfaces.
1736 (aff_combination_scale): Take a widest_int instead of a double_int.
1737 (aff_combination_add_elt): Likewise.
1738 (aff_combination_add_cst): Likewise.
1739 (aff_combination_add): Use wide-int interfaces.
1740 (aff_combination_convert): Likewise.
1741 (tree_to_aff_combination): Likewise.
1742 (add_elt_to_tree): Take a widest_int instead of a double_int.
1743 (aff_combination_to_tree): Use wide-int interfaces.
1744 (aff_combination_remove_elt): Likewise.
1745 (aff_combination_add_product): Take a widest_int instead of
1747 (aff_combination_mult): Use wide-int interfaces.
1748 (aff_combination_expand): Likewise.
1749 (double_int_constant_multiple_p): Delete.
1750 (wide_int_constant_multiple_p): New.
1751 (aff_combination_constant_multiple_p): Take a widest_int pointer
1752 instead of a double_int pointer.
1753 (print_aff): Use wide-int interfaces.
1754 (get_inner_reference_aff): Take a widest_int pointer
1755 instead of a double_int pointer.
1756 (aff_comb_cannot_overlap_p): Take widest_ints instead of double_ints.
1757 * tree-affine.h: Include wide-int.h.
1758 (struct aff_comb_elt): Change type of coef to widest_int.
1759 (struct affine_tree_combination): Change type of offset to widest_int.
1760 (double_int_ext_for_comb): Delete.
1761 (wide_int_ext_for_comb): New.
1762 (aff_combination_const): Use widest_int instead of double_int.
1763 (aff_combination_scale): Likewise.
1764 (aff_combination_add_elt): Likewise.
1765 (aff_combination_constant_multiple_p): Likewise.
1766 (get_inner_reference_aff): Likewise.
1767 (aff_comb_cannot_overlap_p): Likewise.
1768 (aff_combination_zero_p): Use wide-int interfaces.
1769 * tree.c: Include tree.h.
1770 (init_ttree): Use make_int_cst.
1771 (tree_code_size): Removed code for INTEGER_CST case.
1772 (tree_size): Add INTEGER_CST case.
1773 (make_node_stat): Update comment.
1774 (get_int_cst_ext_nunits, build_new_int_cst, build_int_cstu): New.
1775 (build_int_cst_type): Use wide-int interfaces.
1776 (double_int_to_tree): Likewise.
1777 (double_int_fits_to_tree_p): Delete.
1778 (force_fit_type_double): Delete.
1779 (force_fit_type): New.
1780 (int_cst_hash_hash): Use wide-int interfaces.
1781 (int_cst_hash_eq): Likewise.
1782 (build_int_cst_wide): Delete.
1783 (wide_int_to_tree): New.
1784 (cache_integer_cst): Use wide-int interfaces.
1785 (build_low_bits_mask): Likewise.
1786 (cst_and_fits_in_hwi): Likewise.
1787 (real_value_from_int_cst): Likewise.
1788 (make_int_cst_stat): New.
1789 (integer_zerop): Use wide_int interfaces.
1790 (integer_onep): Likewise.
1791 (integer_all_onesp): Likewise.
1792 (integer_pow2p): Likewise.
1793 (integer_nonzerop): Likewise.
1794 (tree_log2): Likewise.
1795 (tree_floor_log2): Likewise.
1796 (tree_ctz): Likewise.
1797 (int_size_in_bytes): Likewise.
1798 (mem_ref_offset): Return an offset_int rather than a double_int.
1799 (build_type_attribute_qual_variant): Use wide_int interfaces.
1800 (type_hash_eq): Likewise
1801 (tree_int_cst_equal): Likewise.
1802 (tree_int_cst_lt): Delete.
1803 (tree_int_cst_compare): Likewise.
1804 (tree_fits_shwi_p): Use wide_int interfaces.
1805 (tree_fits_uhwi_p): Likewise.
1806 (tree_int_cst_sign_bit): Likewise.
1807 (tree_int_cst_sgn): Likewise.
1808 (tree_int_cst_min_precision): Take a signop rather than a bool.
1809 (simple_cst_equal): Use wide_int interfaces.
1810 (compare_tree_int): Likewise.
1811 (iterative_hash_expr): Likewise.
1812 (int_fits_type_p): Likewise. Use tree_int_cst_lt rather than
1814 (get_type_static_bounds): Use wide_int interfaces.
1815 (tree_int_cst_elt_check_failed): New.
1816 (build_common_tree_nodes): Reordered to set prec before filling in
1818 (int_cst_value): Check cst_and_fits_in_hwi.
1819 (widest_int_cst_value): Use wide_int interfaces.
1820 (upper_bound_in_type): Likewise.
1821 (lower_bound_in_type): Likewise.
1822 (num_ending_zeros): Likewise.
1823 (drop_tree_overflow): Likewise.
1824 * tree-call-cdce.c (check_pow): Update call to real_from_integer.
1825 (gen_conditions_for_pow_cst_base): Likewise.
1826 * tree-cfg.c: Include wide-int.h and wide-int-print.h.
1827 (group_case_labels_stmt): Use wide-int interfaces.
1828 (verify_gimple_assign_binary): Likewise.
1829 (print_loop): Likewise.
1830 * tree-chrec.c (tree_fold_binomial): Likewise.
1831 * tree-core.h (struct tree_base): Add int_length.
1832 (struct tree_int_cst): Change rep of value.
1833 * tree-data-ref.c (dr_analyze_innermost): Use wide-int interfaces.
1834 (dr_may_alias_p): Likewise.
1835 (max_stmt_executions_tree): Likewise.
1836 * tree.def (INTEGER_CST): Update comment.
1837 * tree-dfa.c (get_ref_base_and_extent): Use wide-int interfaces.
1838 * tree-dfa.h (get_addr_base_and_unit_offset_1): Likewise.
1839 * tree-dump.c: Include wide-int.h and wide-int-print.h.
1840 (dequeue_and_dump): Use wide-int interfaces.
1841 * tree.h: Include wide-int.h.
1842 (NULL_TREE): Moved to earlier loc in file.
1843 (TREE_INT_CST_ELT_CHECK): New.
1844 (tree_int_cst_elt_check_failed): New.
1846 (TREE_INT_CST): Delete.
1847 (TREE_INT_CST_LOW): Use wide-int interfaces.
1848 (TREE_INT_CST_HIGH): Delete.
1849 (TREE_INT_CST_NUNITS): New.
1850 (TREE_INT_CST_EXT_NUNITS): Likewise.
1851 (TREE_INT_CST_OFFSET_NUNITS): Likewise.
1852 (TREE_INT_CST_ELT): Likewise.
1853 (INT_CST_LT): Delete.
1854 (tree_int_cst_elt_check): New (two forms).
1855 (type_code_size): Update comment.
1856 (make_int_cst_stat, make_int_cst): New.
1857 (tree_to_double_int): Delete.
1858 (double_int_fits_to_tree_p): Delete.
1859 (force_fit_type_double): Delete.
1860 (build_int_cstu): Replace with out-of-line function.
1861 (build_int_cst_wide): Delete.
1862 (tree_int_cst_lt): Define inline.
1863 (tree_int_cst_le): New.
1864 (tree_int_cst_compare): Define inline.
1865 (tree_int_cst_min_precision): Take a signop rather than a bool.
1866 (wi::int_traits <const_tree>): New.
1867 (wi::int_traits <tree>): New.
1868 (wi::extended_tree): New.
1869 (wi::int_traits <wi::extended_tree>): New.
1870 (wi::to_widest): New.
1871 (wi::to_offset): New.
1872 (wi::fits_to_tree_p): New.
1873 (wi::min_value): New.
1874 (wi::max_value): New.
1875 * tree-inline.c (remap_gimple_op_r): Use wide-int interfaces.
1876 (copy_tree_body_r): Likewise.
1877 * tree-object-size.c (compute_object_offset): Likewise.
1878 (addr_object_size): Likewise.
1879 * tree-predcom.c: Include wide-int-print.h.
1880 (struct dref_d): Change type of offset to widest_int.
1881 (dump_dref): Call wide-int printer.
1882 (aff_combination_dr_offset): Use wide-int interfaces.
1883 (determine_offset): Take a widest_int pointer rather than a
1885 (split_data_refs_to_components): Use wide-int interfaces.
1886 (suitable_component_p): Likewise.
1887 (order_drefs): Likewise.
1888 (add_ref_to_chain): Likewise.
1889 (valid_initializer_p): Likewise.
1890 (determine_roots_comp): Likewise.
1891 * tree-pretty-print.c: Include wide-int-print.h.
1892 (dump_generic_node): Use wide-int interfaces.
1893 * tree-sra.c (sra_ipa_modify_expr): Likewise.
1894 * tree-ssa-address.c (addr_for_mem_ref): Likewise.
1895 (move_fixed_address_to_symbol): Likewise.
1896 (move_hint_to_base): Likewise.
1897 (move_pointer_to_base): Likewise.
1898 (move_variant_to_index): Likewise.
1899 (most_expensive_mult_to_index): Likewise.
1900 (addr_to_parts): Likewise.
1901 (copy_ref_info): Likewise.
1902 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
1903 (indirect_refs_may_alias_p): Likewise.
1904 (stmt_kills_ref_p_1): Likewise.
1905 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
1906 * tree-ssa-ccp.c: Update comment at top of file. Include
1908 (struct prop_value_d): Change type of mask to widest_int.
1909 (extend_mask): New function.
1910 (dump_lattice_value): Use wide-int interfaces.
1911 (get_default_value): Likewise.
1912 (set_constant_value): Likewise.
1913 (set_value_varying): Likewise.
1914 (valid_lattice_transition): Likewise.
1915 (set_lattice_value): Likewise.
1916 (value_to_double_int): Delete.
1917 (value_to_wide_int): New.
1918 (get_value_from_alignment): Use wide-int interfaces.
1919 (get_value_for_expr): Likewise.
1920 (do_dbg_cnt): Likewise.
1921 (ccp_finalize): Likewise.
1922 (ccp_lattice_meet): Likewise.
1923 (bit_value_unop_1): Use widest_ints rather than double_ints.
1924 (bit_value_binop_1): Likewise.
1925 (bit_value_unop): Use wide-int interfaces.
1926 (bit_value_binop): Likewise.
1927 (bit_value_assume_aligned): Likewise.
1928 (evaluate_stmt): Likewise.
1929 (ccp_fold_stmt): Likewise.
1930 (visit_cond_stmt): Likewise.
1931 (ccp_visit_stmt): Likewise.
1932 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Likewise.
1933 (constant_pointer_difference): Likewise.
1934 (associate_pointerplus): Likewise.
1935 (combine_conversions): Likewise.
1936 * tree-ssa-loop.h: Include wide-int.h.
1937 (struct tree_niter_desc): Change type of max to widest_int.
1938 * tree-ssa-loop-im.c (mem_refs_may_alias_p): Use wide-int interfaces.
1939 * tree-ssa-loop-ivcanon.c (remove_exits_and_undefined_stmts): Likewise.
1940 (remove_redundant_iv_tests): Likewise.
1941 (canonicalize_loop_induction_variables): Likewise.
1942 * tree-ssa-loop-ivopts.c (alloc_iv): Likewise.
1943 (constant_multiple_of): Take a widest_int pointer instead of
1944 a double_int pointer.
1945 (get_computation_aff): Use wide-int interfaces.
1946 (ptr_difference_cost): Likewise.
1947 (difference_cost): Likewise.
1948 (get_loop_invariant_expr_id): Likewise.
1949 (get_computation_cost_at): Likewise.
1950 (iv_elimination_compare_lt): Likewise.
1951 (may_eliminate_iv): Likewise.
1952 * tree-ssa-loop-niter.h (estimated_loop_iterations): Use widest_int
1953 instead of double_int.
1954 (max_loop_iterations): Likewise.
1955 (max_stmt_executions): Likewise.
1956 (estimated_stmt_executions): Likewise.
1957 * tree-ssa-loop-niter.c: Include wide-int-print.h.
1958 (split_to_var_and_offset): Use wide-int interfaces.
1959 (determine_value_range): Likewise.
1960 (bound_difference_of_offsetted_base): Likewise.
1961 (bounds_add): Take a widest_int instead of a double_int.
1962 (number_of_iterations_ne_max): Use wide-int interfaces.
1963 (number_of_iterations_ne): Likewise.
1964 (number_of_iterations_lt_to_ne): Likewise.
1965 (assert_loop_rolls_lt): Likewise.
1966 (number_of_iterations_lt): Likewise.
1967 (number_of_iterations_le): Likewise.
1968 (number_of_iterations_cond): Likewise.
1969 (number_of_iterations_exit): Likewise.
1970 (finite_loop_p): Likewise.
1971 (derive_constant_upper_bound_assign): Likewise.
1972 (derive_constant_upper_bound): Return a widest_int.
1973 (derive_constant_upper_bound_ops): Likewise.
1974 (do_warn_aggressive_loop_optimizations): Use wide-int interfaces.
1975 (record_estimate): Take a widest_int rather than a double_int.
1976 (record_nonwrapping_iv): Use wide-int interfaces.
1977 (double_int_cmp): Delete.
1978 (wide_int_cmp): New.
1979 (bound_index): Take a widest_int rather than a double_int.
1980 (discover_iteration_bound_by_body_walk): Use wide-int interfaces.
1981 (maybe_lower_iteration_bound): Likewise.
1982 (estimate_numbers_of_iterations_loop): Likewise.
1983 (estimated_loop_iterations): Take a widest_int pointer than than
1984 a double_int pointer.
1985 (estimated_loop_iterations_int): Use wide-int interfaces.
1986 (max_loop_iterations): Take a widest_int pointer than than
1987 a double_int pointer.
1988 (max_loop_iterations_int): Use wide-int interfaces.
1989 (max_stmt_executions): Take a widest_int pointer than than
1990 a double_int pointer.
1991 (estimated_stmt_executions): Likewise.
1992 (n_of_executions_at_most): Use wide-int interfaces.
1993 (scev_probably_wraps_p): Likewise.
1994 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Update calls
1996 * tree-scalar-evolution.c (simplify_peeled_chrec): Use wide-int
1998 * tree-ssanames.c (set_range_info): Use wide_int_refs rather than
1999 double_ints. Adjust for trailing_wide_ints <3> representation.
2000 (set_nonzero_bits): Likewise.
2001 (get_range_info): Return wide_ints rather than double_ints.
2002 Adjust for trailing_wide_ints <3> representation.
2003 (get_nonzero_bits): Likewise.
2004 (duplicate_ssa_name_range_info): Adjust for trailing_wide_ints <3>
2006 * tree-ssanames.h (struct range_info_def): Replace min, max and
2007 nonzero_bits with a trailing_wide_ints <3>.
2008 (set_range_info): Use wide_int_refs rather than double_ints.
2009 (set_nonzero_bits): Likewise.
2010 (get_range_info): Return wide_ints rather than double_ints.
2011 (get_nonzero_bits): Likewise.
2012 * tree-ssa-phiopt.c (jump_function_from_stmt): Use wide-int interfaces.
2013 * tree-ssa-pre.c (phi_translate_1): Likewise.
2014 * tree-ssa-reassoc.c (decrement_power): Use calls to real_from_integer.
2015 (acceptable_pow_call): Likewise.
2016 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Use wide-int
2018 (vn_reference_fold_indirect): Likewise.
2019 (vn_reference_maybe_forwprop_address): Likewise.
2020 (valueize_refs_1): Likewise.
2021 * tree-ssa-structalias.c (get_constraint_for_ptr_offset): Likewise.
2022 * tree-ssa-uninit.c (is_value_included_in): Use wide-int interfaces,
2023 tree_int_cst_lt and tree_int_cst_le.
2024 * tree-streamer-in.c (unpack_ts_base_value_fields): Use wide-int
2026 (streamer_alloc_tree): Likewise.
2027 * tree-streamer-out.c (pack_ts_int_cst_value_fields): Likewise.
2028 (streamer_write_tree_header): Likewise.
2029 (streamer_write_integer_cst): Likewise.
2030 * tree-switch-conversion.c (emit_case_bit_tests): Likewise.
2031 (build_constructors): Likewise.
2032 (array_value_type): Likewise.
2033 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list): Likewise.
2034 (vect_check_gather): Likewise.
2035 * tree-vect-generic.c (build_replicated_const): Likewise.
2036 (expand_vector_divmod): Likewise.
2037 * tree-vect-loop.c (vect_transform_loop): Likewise.
2038 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
2039 (vect_do_peeling_for_alignment): Likewise.
2040 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
2041 * tree-vrp.c: Include wide-int.h.
2042 (operand_less_p): Use wide-int interfaces and tree_int_cst_lt.
2043 (extract_range_from_assert): Use wide-int interfaces.
2044 (vrp_int_const_binop): Likewise.
2045 (zero_nonzero_bits_from_vr): Take wide_int pointers rather than
2046 double_int pointers.
2047 (ranges_from_anti_range): Use wide-int interfaces.
2048 (quad_int_cmp): Delete.
2049 (quad_int_pair_sort): Likewise.
2050 (extract_range_from_binary_expr_1): Use wide-int interfaces.
2051 (extract_range_from_unary_expr_1): Likewise.
2052 (adjust_range_with_scev): Likewise.
2053 (masked_increment): Take and return wide_ints rather than double_ints.
2054 (register_edge_assert_for_2): Use wide-int interfaces.
2055 (check_array_ref): Likewise.
2056 (search_for_addr_array): Likewise.
2057 (maybe_set_nonzero_bits): Likewise.
2058 (union_ranges): Pass an integer of the correct type instead of
2059 using integer_one_node.
2060 (intersect_ranges): Likewise.
2061 (simplify_truth_ops_using_ranges): Likewise.
2062 (simplify_bit_ops_using_ranges): Use wide-int interfaces.
2063 (range_fits_type_p): Likewise.
2064 (simplify_cond_using_ranges): Likewise. Take a signop rather than
2066 (simplify_conversion_using_ranges): Use wide-int interfaces.
2067 (simplify_float_conversion_using_ranges): Likewise.
2068 (vrp_finalize): Likewise.
2069 * value-prof.c (gimple_divmod_fixed_value_transform): Likewise.
2070 (gimple_stringops_transform): Likewise.
2071 * varasm.c (decode_addr_const): Likewise.
2072 (const_hash_1): Likewise.
2073 (const_rtx_hash_1): Likewise
2074 (output_constant): Likewise.
2075 (array_size_for_constructor): Likewise.
2076 (output_constructor_regular_field): Likewise.
2077 (output_constructor_bitfield): Likewise.
2078 * var-tracking.c (loc_cmp): Handle CONST_WIDE_INT.
2079 * mkconfig.sh: Include machmode.h to pick up BITS_PER_UNIT for
2081 * gencheck.c: Define BITS_PER_UNIT.
2084 * wide-int-print.cc: New.
2085 * wide-int-print.h: New.
2087 2014-05-06 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2089 * config/avr/avr.c (avr_can_eliminate): Mark unused argument.
2091 2014-05-06 Richard Biener <rguenther@suse.de>
2093 * tree-pass.h (TODO_verify_ssa, TODO_verify_flow,
2094 TODO_verify_stmts, TODO_verify_rtl_sharing): Remove.
2095 (TODO_verify_all): Adjust.
2096 * asan.c: Remove references to TODO_verify_ssa, TODO_verify_flow,
2097 TODO_verify_stmts and TODO_verify_rtl_sharing.
2098 * bb-reorder.c: Likewise.
2099 * cfgexpand.c: Likewise.
2100 * cprop.c: Likewise.
2102 * function.c: Likewise.
2103 * fwprop.c: Likewise.
2105 * gimple-ssa-isolate-paths.c: Likewise.
2106 * gimple-ssa-strength-reduction.c: Likewise.
2107 * ipa-split.c: Likewise.
2108 * loop-init.c: Likewise.
2109 * loop-unroll.c: Likewise.
2110 * lower-subreg.c: Likewise.
2111 * modulo-sched.c: Likewise.
2112 * postreload-gcse.c: Likewise.
2113 * predict.c: Likewise.
2114 * recog.c: Likewise.
2115 * sched-rgn.c: Likewise.
2116 * store-motion.c: Likewise.
2117 * tracer.c: Likewise.
2118 * trans-mem.c: Likewise.
2119 * tree-call-cdce.c: Likewise.
2120 * tree-cfg.c: Likewise.
2121 * tree-cfgcleanup.c: Likewise.
2122 * tree-complex.c: Likewise.
2123 * tree-eh.c: Likewise.
2124 * tree-emutls.c: Likewise.
2125 * tree-if-conv.c: Likewise.
2126 * tree-into-ssa.c: Likewise.
2127 * tree-loop-distribution.c: Likewise.
2128 * tree-object-size.c: Likewise.
2129 * tree-parloops.c: Likewise.
2130 * tree-pass.h: Likewise.
2131 * tree-sra.c: Likewise.
2132 * tree-ssa-ccp.c: Likewise.
2133 * tree-ssa-copy.c: Likewise.
2134 * tree-ssa-copyrename.c: Likewise.
2135 * tree-ssa-dce.c: Likewise.
2136 * tree-ssa-dom.c: Likewise.
2137 * tree-ssa-dse.c: Likewise.
2138 * tree-ssa-forwprop.c: Likewise.
2139 * tree-ssa-ifcombine.c: Likewise.
2140 * tree-ssa-loop-ch.c: Likewise.
2141 * tree-ssa-loop-ivcanon.c: Likewise.
2142 * tree-ssa-loop.c: Likewise.
2143 * tree-ssa-math-opts.c: Likewise.
2144 * tree-ssa-phiopt.c: Likewise.
2145 * tree-ssa-phiprop.c: Likewise.
2146 * tree-ssa-pre.c: Likewise.
2147 * tree-ssa-reassoc.c: Likewise.
2148 * tree-ssa-sink.c: Likewise.
2149 * tree-ssa-strlen.c: Likewise.
2150 * tree-ssa-tail-merge.c: Likewise.
2151 * tree-ssa-uncprop.c: Likewise.
2152 * tree-switch-conversion.c: Likewise.
2153 * tree-tailcall.c: Likewise.
2154 * tree-vect-generic.c: Likewise.
2155 * tree-vectorizer.c: Likewise.
2156 * tree-vrp.c: Likewise.
2158 * var-tracking.c: Likewise.
2159 * bt-load.c: Likewise.
2160 * cfgcleanup.c: Likewise.
2161 * combine-stack-adj.c: Likewise.
2162 * combine.c: Likewise.
2163 * compare-elim.c: Likewise.
2164 * config/epiphany/resolve-sw-modes.c: Likewise.
2165 * config/i386/i386.c: Likewise.
2166 * config/mips/mips.c: Likewise.
2167 * config/s390/s390.c: Likewise.
2168 * config/sh/sh_treg_combine.cc: Likewise.
2169 * config/sparc/sparc.c: Likewise.
2172 * final.c: Likewise.
2173 * ifcvt.c: Likewise.
2174 * mode-switching.c: Likewise.
2175 * passes.c: Likewise.
2176 * postreload.c: Likewise.
2178 * reg-stack.c: Likewise.
2179 * regcprop.c: Likewise.
2180 * regrename.c: Likewise.
2183 2014-05-06 Richard Biener <rguenther@suse.de>
2186 * bitmap.c (debug_bitmap): Dump to stderr, not stdout.
2187 * tree-ssa-structalias.c (dump_solution_for_var): Likewise.
2189 2014-05-05 Jan Hubicka <hubicka@ucw.cz>
2192 * ipa-devirt.c (get_class_context): Allow POD to change to non-POD.
2194 2014-05-05 Radovan Obradovic <robradovic@mips.com>
2195 Tom de Vries <tom@codesourcery.com>
2197 * target.def (call_fusage_contains_non_callee_clobbers): New
2199 * doc/tm.texi.in (@node Stack and Calling): Add Miscellaneous Register
2201 (@node Miscellaneous Register Hooks): New node.
2202 (@hook TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): New hook.
2203 * doc/tm.texi: Regenerate.
2205 2014-05-05 Marek Polacek <polacek@redhat.com>
2208 * opts.c (common_handle_option): Call error_at instead of warning_at.
2210 2014-05-05 Richard Biener <rguenther@suse.de>
2212 * passes.c (execute_function_todo): Don't reset TODO_verify_ssa
2213 from last_verified if update_ssa ran. Move TODO_verify_rtl_sharing
2214 under the TODO_verify_il umbrella.
2216 2014-05-05 Richard Biener <rguenther@suse.de>
2218 * passes.c (execute_function_todo): Move TODO_verify_flow under
2219 the TODO_verify_ul umbrella.
2221 2014-05-05 Richard Biener <rguenther@suse.de>
2224 * fold-const.c (fold_binary_loc): Consistently avoid canonicalizing
2225 X & CST away from a CST that is the mask of a mode.
2227 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2229 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
2230 int argument to enum machine_mode.
2231 (picochip_class_max_nregs): Ditto.
2232 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
2233 (picochip_class_max_nregs): Ditto.
2235 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2237 * target.def: Add new target hook.
2238 * doc/tm.texi: Regenerate.
2239 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
2240 * targhooks.c (default_keep_leaf_when_profiled): New function.
2242 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
2243 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
2245 2014-05-05 Bin Cheng <bin.cheng@arm.com>
2247 PR tree-optimization/60363
2248 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
2249 (copy_phi_args): New parameters. Call get_value_locus_in_path.
2250 (update_destination_phis): New parameter.
2251 (create_edge_and_update_destination_phis): Ditto.
2252 (ssa_fix_duplicate_block_edges): Pass new arguments.
2253 (thread_single_edge): Ditto.
2255 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
2257 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
2258 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
2259 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
2260 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
2261 Use RS6000_BTM_HARD_FLOAT.
2262 (BU_MISC_2): Likewise.
2263 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
2264 RS6000_BTM_HARD_FLOAT.
2265 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
2267 (rs6000_invalid_builtin): Add hard floating builtin support.
2268 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
2269 hard float builtins.
2270 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
2272 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
2274 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
2275 Add missing function* argument.
2277 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
2279 * lra-constraints.c (valid_address_p): Move earlier in file.
2280 Add a constraint argument to the address_info version.
2281 (satisfies_memory_constraint_p): New function.
2282 (satisfies_address_constraint_p): Likewise.
2283 (process_alt_operands, curr_insn_transform): Use them.
2284 (process_address): Pass the constraint to valid_address_p when
2285 checking address operands.
2287 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
2289 * config/mips/mips.c (mips_isa_rev): New variable.
2290 (mips_set_architecture): Set it.
2291 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
2293 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
2294 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
2295 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
2296 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
2297 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
2298 conditions in terms of mips_isa_rev.
2299 (mips_isa_rev): Declare.
2301 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
2303 * config/sh/sh-mem.cc: Use tabs instead of spaces.
2304 (prob_unlikely, prob_likely): Make variables const.
2306 2014-05-03 Denis Chertykov <chertykov@gmail.com>
2308 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
2310 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
2312 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
2314 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
2316 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
2317 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
2318 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
2320 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
2321 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
2323 Replace usage of ROUND_REG with sh_round_reg.
2324 Use CEIL instead of ROUND_ADVANCE.
2326 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
2329 * config/sh/sh.c: Include stdlib headers before everything else.
2331 2014-05-02 Jakub Jelinek <jakub@redhat.com>
2333 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
2334 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
2335 (gimplify_adjust_omp_clauses): Simd region is never
2336 directly nested in combined parallel. Instead, for linear
2337 with copyin/copyout, if in combined for simd loop, make decl
2338 firstprivate/lastprivate on OMP_FOR.
2339 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
2340 expand_omp_for_static_chunk): When setting endvar, also set
2341 fd->loop.v to the same value.
2343 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
2345 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
2347 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
2349 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
2352 2014-05-02 Marek Polacek <polacek@redhat.com>
2354 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
2356 2014-05-02 Kito Cheng <kito@0xlab.org>
2358 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
2359 to a C expression marco.
2360 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
2361 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
2362 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
2363 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
2364 HONOR_REG_ALLOC_ORDER.
2365 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
2367 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2369 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
2371 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2373 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
2375 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
2377 * tree-if-conv.c (is_cond_scalar_reduction): New function.
2378 (convert_scalar_cond_reduction): Likewise.
2379 (predicate_scalar_phi): Add recognition and transformation
2380 of simple conditioanl reduction to be vectorizable.
2382 2014-05-01 Marek Polacek <polacek@redhat.com>
2385 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
2387 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
2389 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
2390 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
2391 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
2392 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
2393 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
2394 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
2395 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
2396 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
2398 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
2400 * config/arc/arc.opt (mlra): Move comment above option name
2401 to avoid mis-parsing as language options.
2403 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2405 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
2406 * config/sol2.h: ... here.
2407 * config/sol2-10.h: Remove.
2409 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
2410 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
2411 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
2412 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
2413 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
2414 * config/sol2.h: ... here.
2415 (SECTION_NAME_FORMAT): Don't redefine.
2416 (STARTFILE_ARCH32_SPEC): Rename to ...
2417 (STARTFILE_ARCH_SPEC): ... this.
2418 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
2419 * config/sparc/sol2.h: ... here.
2420 (SECTION_NAME_FORMAT): Don't undef.
2421 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
2422 (SUBTARGET_EXTRA_SPECS): Remove.
2423 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
2425 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
2426 (MD_STARTFILE_PREFIX): Remove.
2427 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
2428 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
2429 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
2430 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
2431 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
2432 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
2433 * config/i386/sol2.h: ... here.
2434 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
2435 * config/i386/sol2-bi.h: Remove.
2436 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
2437 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
2439 * config/i386/t-sol2-64: Rename to ...
2440 * config/i386/t-sol2: ... this.
2441 * config/sparc/t-sol2-64: Rename to ...
2442 * config/sparc/t-sol2: ... this.
2444 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
2445 sol2_tm_file_head, sol2_tm_file_tail.
2446 Include ${cpu_type}/sol2.h before sol2.h.
2448 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
2449 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
2450 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
2451 Reflect i386/t-sol2-64 renaming.
2452 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
2453 Reflect sparc/t-sol2-64 renaming.
2455 2014-04-30 Richard Biener <rguenther@suse.de>
2457 * passes.c (execute_function_todo): Move TODO_verify_stmts
2458 and TODO_verify_ssa under the TODO_verify_il umbrella.
2459 * tree-ssa.h (verify_ssa): Adjust prototype.
2460 * tree-ssa.c (verify_ssa): Add parameter to tell whether
2461 we should verify SSA operands.
2462 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
2463 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
2464 whether we should verify whether not throwing stmts have EH info.
2465 * graphite-scop-detection.c (create_sese_edges): Adjust.
2466 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
2467 * tree-eh.c (lower_try_finally_switch): Do not add the
2468 default case label twice.
2470 2014-04-30 Marek Polacek <polacek@redhat.com>
2472 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
2473 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
2474 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
2475 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
2477 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
2479 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
2480 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
2481 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
2482 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
2483 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
2484 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
2485 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
2486 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
2488 2014-04-29 David Malcolm <dmalcolm@redhat.com>
2490 * tree-cfg.c (dump_function_to_file): Dump the return type of
2491 functions, in a line to itself before the function body, mimicking
2492 the layout of a C function.
2494 2014-04-29 Jakub Jelinek <jakub@redhat.com>
2496 PR tree-optimization/60971
2497 * tree-tailcall.c (process_assignment): Reject conversions which
2500 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
2502 * calls.c (initialize_argument_information): Always treat
2503 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
2504 (expand_call): Likewise.
2505 (emit_library_call_calue_1): Likewise.
2506 * expr.c (PUSH_ARGS_REVERSED): Do not define.
2507 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
2510 2014-04-29 Nick Clifton <nickc@redhat.com>
2512 * config/msp430/msp430.md (umulsidi): Fix typo.
2513 (mulhisi3): Enable even inside interrupt handlers.
2514 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
2515 bigger return address pushed in large mode.
2517 2014-04-29 Nick Clifton <nickc@redhat.com>
2519 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
2520 (arc_init_reg_tables): Use a machine_mode enum to iterate over
2522 * config/m32r/m32r.c (init_reg_tables): Likewise.
2523 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
2524 enum to hold the modes.
2526 2014-04-29 Richard Biener <rguenther@suse.de>
2528 * dominance.c (free_dominance_info): Add overload with
2530 (dom_info_state): Likewise.
2531 (dom_info_available_p): Likewise.
2532 * basic-block.h (free_dominance_info, dom_info_state,
2533 dom_info_available_p): Declare overloads.
2534 * passes.c (execute_function_todo): Verify that verifiers
2535 don't change dominator info state. Drop dominator info
2536 for IPA pass invocations.
2537 * cgraph.c (release_function_body): Restore asserts that
2538 dominator information is released.
2540 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
2542 * doc/invoke.texi: Fix typo.
2543 * tree-vrp.c: Fix typos.
2544 * gimple.c (infer_nonnull_range): Reorder operands of an && condition.
2546 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2548 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
2550 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
2552 * config/aarch64/aarch64-builtins.c
2553 (aarch64_types_storestruct_lane_qualifiers): New.
2554 (TYPES_STORESTRUCT_LANE): Likewise.
2555 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
2556 (st3_lane): Likewise.
2557 (st4_lane): Likewise.
2558 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
2559 (vec_store_lanesci_lane<mode>): Likewise.
2560 (vec_store_lanesxi_lane<mode>): Likewise.
2561 (aarch64_st2_lane<VQ:mode>): Likewise.
2562 (aarch64_st3_lane<VQ:mode>): Likewise.
2563 (aarch64_st4_lane<VQ:mode>): Likewise.
2564 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
2565 * config/aarch64/arm_neon.h
2566 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
2567 use new macro arguments.
2568 (__ST3_LANE_FUNC): Likewise.
2569 (__ST4_LANE_FUNC): Likewise.
2570 * config/aarch64/iterators.md (V_TWO_ELEM): New.
2571 (V_THREE_ELEM): Likewise.
2572 (V_FOUR_ELEM): Likewise.
2574 2014-04-28 David Malcolm <dmalcolm@redhat.com>
2576 * doc/gimple.texi: Replace the description of the now-defunct
2577 union gimple_statement_d with a diagram showing the
2578 gimple_statement_base class hierarchy and its relationships to
2579 the GSS_ and GIMPLE_ enums.
2581 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
2583 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
2584 * config/aarch64/aarch64.c
2585 (aarch64_cannot_change_mode_class): Weaken conditions.
2586 (aarch64_modes_tieable_p): New.
2587 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
2589 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
2591 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2592 (loadsync_<mode>): Change mode.
2593 (load_quadpti, store_quadpti): New.
2594 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2595 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2597 2014-04-28 Martin Jambor <mjambor@suse.cz>
2599 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
2600 same alias type as the original statement.
2601 (subreplacement_assignment_data): New type.
2602 (handle_unscalarized_data_in_subtree): New type of parameter,
2603 generate new memory accesses with same alias type as the original
2605 (load_assign_lhs_subreplacements): Likewise.
2606 (sra_modify_constructor_assign): Generate new memory accesses with
2607 same alias type as the original statement.
2609 2014-04-28 Richard Biener <rguenther@suse.de>
2611 * tree-pass.h (TODO_verify_il): Define.
2612 (TODO_verify_all): Complete properly.
2613 * passes.c (execute_function_todo): Move existing loop-closed
2614 SSA verification under TODO_verify_il.
2615 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
2616 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
2617 Fix tree sharing issue.
2619 2014-04-28 Richard Biener <rguenther@suse.de>
2622 * builtins.def (DEF_C11_BUILTIN): Add.
2623 (BUILT_IN_ALIGNED_ALLOC): Likewise.
2624 * coretypes.h (enum function_class): Add function_c11_misc.
2625 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
2626 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
2627 (call_may_clobber_ref_p_1): Likewise.
2628 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
2629 (mark_all_reaching_defs_necessary_1): Likewise.
2630 (propagate_necessity): Likewise.
2631 (eliminate_unnecessary_stmts): Likewise.
2632 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
2634 2014-04-28 Richard Biener <rguenther@suse.de>
2636 * tree-vrp.c (vrp_var_may_overflow): Remove.
2637 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
2638 with overflow immediately bump to one before that value and
2639 let iteration figure out overflow status.
2641 2014-04-28 Richard Biener <rguenther@suse.de>
2643 * configure.ac: Do valgrind header checks unconditionally.
2644 Add --enable-valgrind-annotations.
2645 * system.h: Guard valgrind header inclusion with
2646 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
2647 * alloc-pool.c (pool_alloc, pool_free): Use
2648 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
2649 to guard possibly dead code.
2650 * config.in: Regenerated.
2651 * configure: Likewise.
2653 2014-04-28 Jeff Law <law@redhat.com>
2655 PR tree-optimization/60902
2656 * tree-ssa-threadedge.c
2657 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
2658 over real defs when invalidating outputs from statements that do not
2659 produce useful outputs for threading.
2661 2014-04-28 Richard Biener <rguenther@suse.de>
2663 PR tree-optimization/60979
2664 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
2665 SCOPs that end in a block with a successor with abnormal
2668 2014-04-28 Richard Biener <rguenther@suse.de>
2670 * tree-pass.h (execute_pass_list): Adjust prototype.
2671 * passes.c (pass_manager::execute_early_local_passes): Adjust.
2672 (do_per_function): Change callback signature, push all actual
2673 work to the callbals.
2674 (do_per_function_toporder): Likewise.
2675 (execute_function_dump): Adjust.
2676 (execute_function_todo): Likewise.
2677 (clear_last_verified): Likewise.
2678 (verify_curr_properties): Likewise.
2679 (update_properties_after_pass): Likewise.
2680 (execute_pass_list_1): Split out from ...
2681 (execute_pass_list): ... here. Adjust.
2682 (execute_ipa_pass_list): Likewise.
2683 * cgraphunit.c (cgraph_add_new_function): Adjust.
2684 (analyze_function): Likewise.
2685 (expand_function): Likewise.
2686 * cgraph.c (release_function_body): Free dominance info
2687 here instead of asserting it was magically freed elsewhere.
2689 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
2691 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
2692 * configure: Regenerate.
2693 * config/sparc/sparc.opt (muser-mode): New option.
2694 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
2696 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
2697 * doc/invoke.texi (SPARC options): Document -muser-mode.
2699 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
2701 * cselib.c (find_slot_memmode): Delete.
2702 (cselib_hasher): Change compare_type to a struct.
2703 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
2705 (preserve_constants_and_equivs): Adjust for new compare_type.
2706 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
2707 (wrap_constant): Delete.
2708 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
2710 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
2712 * doc/install.texi (Building with profile feedback): Remove
2715 2014-04-26 Tom de Vries <tom@codesourcery.com>
2717 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
2720 2014-04-25 Cary Coutant <ccoutant@google.com>
2723 * dwarf2out.c (should_move_die_to_comdat): A type definition
2724 can contain a subprogram definition, but don't move it to a
2726 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
2727 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
2729 (clone_tree_hash): Rename to...
2730 (clone_tree_partial): ...this; change callers. Copy
2731 DW_TAG_subprogram DIEs as declarations.
2732 (copy_decls_walk): Don't copy children of a declaration into a
2735 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
2738 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
2741 2014-04-25 Jiong Wang <jiong.wang@arm.com>
2743 * config/arm/predicates.md (call_insn_operand): Add long_call check.
2744 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
2746 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
2749 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2751 * config/arm/arm.c (arm_cortex_a8_tune): Initialise T16-related fields.
2753 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2755 PR tree-optimization/60930
2756 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
2757 creating a multiply candidate by folding two constant
2758 multiplicands when the result overflows.
2760 2014-04-25 Jakub Jelinek <jakub@redhat.com>
2762 PR tree-optimization/60960
2763 * tree-vect-generic.c (expand_vector_operation): Only call
2764 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
2766 2014-04-25 Tom de Vries <tom@codesourcery.com>
2768 * expr.c (clobber_reg_mode): New function.
2769 * expr.h (clobber_reg): New function.
2771 2014-04-25 Tom de Vries <tom@codesourcery.com>
2773 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
2776 2014-04-25 Radovan Obradovic <robradovic@mips.com>
2777 Tom de Vries <tom@codesourcery.com>
2779 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
2781 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
2782 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
2783 new argument to find_all_hard_reg_sets call.
2785 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2787 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
2788 Use HOST_WIDE_INT_C for mask literal.
2789 (aarch_rev16_shleft_mask_imm_p): Likewise.
2791 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
2794 * config/sparc/sparc.md (ashlsi3_extend): Delete.
2796 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
2798 PR preprocessor/56540
2799 * config/i386/i386-c.c (ix86_target_macros): Define
2800 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
2802 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2804 * configure.ac (tga_func): Remove.
2805 (LIB_TLS_SPEC): Remove.
2806 * configure: Regenerate.
2807 * config.in: Regenerate.
2808 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
2810 2014-04-25 Richard Biener <rguenther@suse.de>
2813 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
2814 call stmt use/clobber sets during stmt walk instead of
2815 walking the possibly incomplete set of caller edges.
2817 2014-04-25 Richard Biener <rguenther@suse.de>
2820 * passes.c (apply_ipa_transforms): Inline into only caller ...
2821 (execute_one_pass): ... here. Properly bring in function
2822 bodies for nodes we want to apply IPA transforms to.
2824 2014-04-24 Cong Hou <congh@google.com>
2826 PR tree-optimization/60896
2827 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
2828 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
2829 (vect_mark_pattern_stmts): Set the def type of all statements in
2830 PATTERN_DEF_SEQ as vect_internal_def.
2832 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
2834 * doc/extend.texi (PowerPC Built-in Functions): Document new
2835 powerpc extended divide, bcd, pack/unpack 128-bit, builtin functions.
2836 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
2838 * config/rs6000/predicates.md (const_0_to_3_operand): New
2839 predicate to match 0..3 integer constants.
2841 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
2842 to support adding miscellaneous builtin functions.
2843 (BU_DFP_MISC_2): Likewise.
2844 (BU_P7_MISC_1): Likewise.
2845 (BU_P7_MISC_2): Likewise.
2846 (BU_P8V_MISC_3): Likewise.
2847 (BU_MISC_1): Likewise.
2848 (BU_MISC_2): Likewise.
2849 (DIVWE): Add extended divide builtin functions.
2852 (DIVWEUO): Likewise.
2856 (DIVDEUO): Likewise.
2857 (DXEX): Add decimal floating-point builtin functions.
2860 (DDEDPDQ): Likewise.
2862 (DENBCDQ): Likewise.
2869 (CDTBCD): Add new BCD builtin functions.
2873 (BCDADD_LT): Likewise.
2874 (BCDADD_EQ): Likewise.
2875 (BCDADD_GT): Likewise.
2876 (BCDADD_OV): Likewise.
2878 (BCDSUB_LT): Likewise.
2879 (BCDSUB_EQ): Likewise.
2880 (BCDSUB_GT): Likewise.
2881 (BCDSUB_OV): Likewise.
2882 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
2883 (UNPACK_TD): Likewise.
2884 (PACK_TF): Likewise.
2885 (UNPACK_TF): Likewise.
2886 (UNPACK_TF_0): Likewise.
2887 (UNPACK_TF_1): Likewise.
2888 (PACK_V1TI): Likewise.
2889 (UNPACK_V1TI): Likewise.
2891 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
2892 support for decimal floating point builtin functions.
2893 (rs6000_expand_ternop_builtin): Add checks for the new builtin
2894 functions that take constant arguments.
2895 (rs6000_invalid_builtin): Add decimal floating point builtin support.
2896 (rs6000_init_builtins): Setup long double, _Decimal64, and
2897 _Decimal128 types for new builtin functions.
2898 (builtin_function_type): Set the unsigned flags appropriately for
2899 the new builtin functions.
2900 (rs6000_opt_masks): Add support for decimal floating point builtin
2903 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
2904 floating point builtin functions.
2905 (RS6000_BTM_COMMON): Likewise.
2906 (RS6000_BTI_long_double): Likewise.
2907 (RS6000_BTI_dfloat64): Likewise.
2908 (RS6000_BTI_dfloat128): Likewise.
2909 (long_double_type_internal_node): Likewise.
2910 (dfloat64_type_internal_node): Likewise.
2911 (dfloat128_type_internal_node): Likewise.
2913 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
2914 2.07 bcd arithmetic instructions.
2915 (UNSPEC_BCDSUB): Likewise.
2916 (UNSPEC_BCD_OVERFLOW): Likewise.
2917 (UNSPEC_BCD_ADD_SUB): Likewise.
2918 (bcd_add_sub): Likewise.
2919 (BCD_TEST): Likewise.
2920 (bcd<bcd_add_sub>): Likewise.
2921 (bcd<bcd_add_sub>_test): Likewise.
2922 (bcd<bcd_add_sub>_test2): Likewise.
2923 (bcd<bcd_add_sub>_<code>): Likewise.
2924 (peephole2 for combined bcd ops): Likewise.
2926 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
2927 decimal floating point builtin functions.
2928 (UNSPEC_DENBCD): Likewise.
2929 (UNSPEC_DXEX): Likewise.
2930 (UNSPEC_DIEX): Likewise.
2931 (UNSPEC_DSCLI): Likewise.
2932 (UNSPEC_DSCRI): Likewise.
2933 (D64_D128): Likewise.
2934 (dfp_suffix): Likewise.
2935 (dfp_ddedpd_<mode>): Likewise.
2936 (dfp_denbcd_<mode>): Likewise.
2937 (dfp_dxex_<mode>): Likewise.
2938 (dfp_diex_<mode>): Likewise.
2939 (dfp_dscli_<mode>): Likewise.
2940 (dfp_dscri_<mode>): Likewise.
2942 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
2944 (UNSPEC_CDTBCD): Likewise.
2945 (UNSPEC_CBCDTD): Likewise.
2946 (UNSPEC_DIVE): Add support for new extended divide builtin functions.
2947 (UNSPEC_DIVEO): Likewise.
2948 (UNSPEC_DIVEU): Likewise.
2949 (UNSPEC_DIVEUO): Likewise.
2950 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
2951 pack/unpack 128-bit types.
2952 (UNSPEC_PACK_128BIT): Likewise.
2953 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
2954 (udiv<mode>3): Use idiv_ldiv mode attribute.
2955 (div<mode>3): Likewise.
2956 (addg6s): Add new BCD builtin functions.
2959 (UNSPEC_DIV_EXTEND): Add support for new extended divide instructions.
2960 (div_extend): Likewise.
2961 (div<div_extend>_<mode>"): Likewise.
2962 (FP128_64): Add support for new builtin functions to pack/unpack
2964 (unpack<mode>): Likewise.
2965 (unpacktf_0): Likewise.
2966 (unpacktf_1): Likewise.
2967 (unpack<mode>_dm): Likewise.
2968 (unpack<mode>_nodm): Likewise.
2969 (pack<mode>): Likewise.
2970 (unpackv1ti): Likewise.
2971 (packv1ti): Likewise.
2973 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
2975 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
2978 2014-04-24 Jakub Jelinek <jakub@redhat.com>
2980 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
2981 * gimplify.c (omp_is_private): Change last argument's type to int.
2982 Only diagnose lastprivate if the simd argument is 1, only diagnose
2983 linear if the simd argument is 2.
2984 (gimplify_omp_for): Adjust omp_is_private callers. When adding
2985 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
2986 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
2987 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
2988 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
2989 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
2990 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
2991 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
2992 * tree-nested.c (convert_nonlocal_omp_clauses,
2993 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
2995 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
2998 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
3001 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
3003 * flag-types.h (enum ivar_visibility): Add.
3005 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
3007 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
3008 function * argument.
3010 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
3012 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
3014 2014-04-24 Radovan Obradovic <robradovic@mips.com>
3015 Tom de Vries <tom@codesourcery.com>
3017 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
3018 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
3020 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
3021 * emit-rtl.c (try_split): Same.
3023 2014-04-24 Radovan Obradovic <robradovic@mips.com>
3024 Tom de Vries <tom@codesourcery.com>
3026 * common.opt (fuse-caller-save): New option.
3028 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
3030 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of
3031 elements for big-endian.
3033 2014-04-24 Richard Biener <rguenther@suse.de>
3035 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
3036 during TER and instead use the sepops interface for expanding
3037 non-GIMPLE_SINGLE_RHS.
3039 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3041 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
3042 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
3044 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3046 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
3047 assembler 64-bit option.
3048 * configure: Regenerate.
3050 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3052 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
3053 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
3054 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
3055 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
3056 (TARGET_CRYPTO): Take TARGET_SIMD into account.
3058 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3060 * config/aarch64/aarch64-builtins.c
3061 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
3062 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
3063 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
3064 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
3066 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
3067 (Vrevsuff): New mode attribute.
3069 2014-04-24 Terry Guo <terry.guo@arm.com>
3071 * config/arm/arm.h (machine_function): Define variable
3072 after_arm_reorg here.
3073 * config/arm/arm.c (after_arm_reorg): Remove the definition.
3074 (arm_split_constant): Update the way to access variable
3077 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
3079 2014-04-23 Tom de Vries <tom@codesourcery.com>
3081 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
3083 2014-04-23 David Malcolm <dmalcolm@redhat.com>
3085 * is-a.h: Update comments to reflect the following changes to the
3086 "pointerness" of the API, making the template parameter match the
3087 return type, allowing use of is-a.h with typedefs of pointers.
3088 (is_a_helper::cast): Return a T rather then a pointer to a T, so
3089 that the return type matches the parameter to the is_a_helper.
3091 (dyn_cast): Likewise.
3093 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
3094 pointer from the is-a.h API.
3096 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
3097 (is_a_helper <cgraph_node *>::test): ...this, matching change to
3099 (is_a_helper <varpool_node>::test): Likewise, convert to...
3100 (is_a_helper <varpool_node *>::test): ...this.
3102 (varpool_first_variable): Update for removal of implicit pointer
3103 from the is-a.h API.
3104 (varpool_next_variable): Likewise.
3105 (varpool_first_static_initializer): Likewise.
3106 (varpool_next_static_initializer): Likewise.
3107 (varpool_first_defined_variable): Likewise.
3108 (varpool_next_defined_variable): Likewise.
3109 (cgraph_first_defined_function): Likewise.
3110 (cgraph_next_defined_function): Likewise.
3111 (cgraph_first_function): Likewise.
3112 (cgraph_next_function): Likewise.
3113 (cgraph_first_function_with_gimple_body): Likewise.
3114 (cgraph_next_function_with_gimple_body): Likewise.
3115 (cgraph_alias_target): Likewise.
3116 (varpool_alias_target): Likewise.
3117 (cgraph_function_or_thunk_node): Likewise.
3118 (varpool_variable_node): Likewise.
3119 (symtab_real_symbol_p): Likewise.
3120 * cgraphunit.c (referred_to_p): Likewise.
3121 (analyze_functions): Likewise.
3122 (handle_alias_pairs): Likewise.
3123 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
3124 * gimple-ssa.h (gimple_vuse_op): Likewise.
3125 (gimple_vdef_op): Likewise.
3126 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
3127 * gimple.c (gimple_build_asm_1): Likewise.
3128 (gimple_build_try): Likewise.
3129 (gimple_build_resx): Likewise.
3130 (gimple_build_eh_dispatch): Likewise.
3131 (gimple_build_omp_for): Likewise.
3132 (gimple_omp_for_set_clauses): Likewise.
3134 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
3135 (is_a_helper <gimple_statement_asm *>::test): ...this.
3136 (is_a_helper <gimple_statement_bind>::test): Convert to...
3137 (is_a_helper <gimple_statement_bind *>::test): ...this.
3138 (is_a_helper <gimple_statement_call>::test): Convert to...
3139 (is_a_helper <gimple_statement_call *>::test): ...this.
3140 (is_a_helper <gimple_statement_catch>::test): Convert to...
3141 (is_a_helper <gimple_statement_catch *>::test): ...this.
3142 (is_a_helper <gimple_statement_resx>::test): Convert to...
3143 (is_a_helper <gimple_statement_resx *>::test): ...this.
3144 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
3145 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
3146 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
3147 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
3148 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
3149 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
3150 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
3151 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
3152 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
3153 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
3154 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
3155 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
3156 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
3157 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
3158 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
3159 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
3160 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
3161 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
3162 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
3163 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
3164 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
3165 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
3166 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
3167 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
3168 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
3169 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
3170 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
3171 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
3172 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
3173 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
3174 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
3175 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
3176 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
3177 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
3178 (is_a_helper <gimple_statement_phi>::test): Convert to...
3179 (is_a_helper <gimple_statement_phi *>::test): ...this.
3180 (is_a_helper <gimple_statement_transaction>::test): Convert to...
3181 (is_a_helper <gimple_statement_transaction *>::test): ...this.
3182 (is_a_helper <gimple_statement_try>::test): Convert to...
3183 (is_a_helper <gimple_statement_try *>::test): ...this.
3184 (is_a_helper <gimple_statement_wce>::test): Convert to...
3185 (is_a_helper <gimple_statement_wce *>::test): ...this.
3186 (is_a_helper <const gimple_statement_asm>::test): Convert to...
3187 (is_a_helper <const gimple_statement_asm *>::test): ...this.
3188 (is_a_helper <const gimple_statement_bind>::test): Convert to...
3189 (is_a_helper <const gimple_statement_bind *>::test): ...this.
3190 (is_a_helper <const gimple_statement_call>::test): Convert to...
3191 (is_a_helper <const gimple_statement_call *>::test): ...this.
3192 (is_a_helper <const gimple_statement_catch>::test): Convert to...
3193 (is_a_helper <const gimple_statement_catch *>::test): ...this.
3194 (is_a_helper <const gimple_statement_resx>::test): Convert to...
3195 (is_a_helper <const gimple_statement_resx *>::test): ...this.
3196 (is_a_helper <const gimple_statement_eh_dispatch>::test): Convert to...
3197 (is_a_helper <const gimple_statement_eh_dispatch *>::test): ...this.
3198 (is_a_helper <const gimple_statement_eh_filter>::test): Convert to...
3199 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
3200 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
3202 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
3204 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
3206 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
3208 (is_a_helper <const gimple_statement_omp_return>::test): Convert to...
3209 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
3210 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
3212 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
3213 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
3215 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
3216 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
3217 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
3218 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert to...
3219 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
3220 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
3222 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
3223 (is_a_helper <const gimple_statement_omp_target>::test): Convert to...
3224 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
3225 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
3227 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
3228 (is_a_helper <const gimple_statement_omp_single>::test): Convert to...
3229 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
3230 (is_a_helper <const gimple_statement_omp_teams>::test): Convert to...
3231 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
3232 (is_a_helper <const gimple_statement_omp_task>::test): Convert to...
3233 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
3234 (is_a_helper <const gimple_statement_phi>::test): Convert to...
3235 (is_a_helper <const gimple_statement_phi *>::test): ...this.
3236 (is_a_helper <const gimple_statement_transaction>::test): Convert to...
3237 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
3238 (is_a_helper <const gimple_statement_with_ops>::test): Convert to...
3239 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
3240 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
3241 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
3242 (is_a_helper <const gimple_statement_with_memory_ops>::test): Convert
3244 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
3246 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert to...
3247 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
3249 (gimple_use_ops): Update for removal of implicit pointer from the
3251 (gimple_set_use_ops): Likewise.
3252 (gimple_vuse): Likewise.
3253 (gimple_vdef): Likewise.
3254 (gimple_vuse_ptr): Likewise.
3255 (gimple_vdef_ptr): Likewise.
3256 (gimple_set_vuse): Likewise.
3257 (gimple_set_vdef): Likewise.
3258 (gimple_omp_return_set_lhs): Likewise.
3259 (gimple_omp_return_lhs): Likewise.
3260 (gimple_omp_return_lhs_ptr): Likewise.
3261 (gimple_call_fntype): Likewise.
3262 (gimple_call_set_fntype): Likewise.
3263 (gimple_call_set_internal_fn): Likewise.
3264 (gimple_call_use_set): Likewise.
3265 (gimple_call_clobber_set): Likewise.
3266 (gimple_bind_vars): Likewise.
3267 (gimple_bind_set_vars): Likewise.
3268 (gimple_bind_body_ptr): Likewise.
3269 (gimple_bind_set_body): Likewise.
3270 (gimple_bind_add_stmt): Likewise.
3271 (gimple_bind_block): Likewise.
3272 (gimple_bind_set_block): Likewise.
3273 (gimple_asm_ninputs): Likewise.
3274 (gimple_asm_noutputs): Likewise.
3275 (gimple_asm_nclobbers): Likewise.
3276 (gimple_asm_nlabels): Likewise.
3277 (gimple_asm_input_op): Likewise.
3278 (gimple_asm_input_op_ptr): Likewise.
3279 (gimple_asm_output_op): Likewise.
3280 (gimple_asm_output_op_ptr): Likewise.
3281 (gimple_asm_set_output_op): Likewise.
3282 (gimple_asm_clobber_op): Likewise.
3283 (gimple_asm_set_clobber_op): Likewise.
3284 (gimple_asm_label_op): Likewise.
3285 (gimple_asm_set_label_op): Likewise.
3286 (gimple_asm_string): Likewise.
3287 (gimple_catch_types): Likewise.
3288 (gimple_catch_types_ptr): Likewise.
3289 (gimple_catch_handler_ptr): Likewise.
3290 (gimple_catch_set_types): Likewise.
3291 (gimple_catch_set_handler): Likewise.
3292 (gimple_eh_filter_types): Likewise.
3293 (gimple_eh_filter_types_ptr): Likewise.
3294 (gimple_eh_filter_failure_ptr): Likewise.
3295 (gimple_eh_filter_set_types): Likewise.
3296 (gimple_eh_filter_set_failure): Likewise.
3297 (gimple_eh_must_not_throw_fndecl): Likewise.
3298 (gimple_eh_must_not_throw_set_fndecl): Likewise.
3299 (gimple_eh_else_n_body_ptr): Likewise.
3300 (gimple_eh_else_e_body_ptr): Likewise.
3301 (gimple_eh_else_set_n_body): Likewise.
3302 (gimple_eh_else_set_e_body): Likewise.
3303 (gimple_try_eval_ptr): Likewise.
3304 (gimple_try_cleanup_ptr): Likewise.
3305 (gimple_try_set_eval): Likewise.
3306 (gimple_try_set_cleanup): Likewise.
3307 (gimple_wce_cleanup_ptr): Likewise.
3308 (gimple_wce_set_cleanup): Likewise.
3309 (gimple_phi_capacity): Likewise.
3310 (gimple_phi_num_args): Likewise.
3311 (gimple_phi_result): Likewise.
3312 (gimple_phi_result_ptr): Likewise.
3313 (gimple_phi_set_result): Likewise.
3314 (gimple_phi_arg): Likewise.
3315 (gimple_phi_set_arg): Likewise.
3316 (gimple_resx_region): Likewise.
3317 (gimple_resx_set_region): Likewise.
3318 (gimple_eh_dispatch_region): Likewise.
3319 (gimple_eh_dispatch_set_region): Likewise.
3320 (gimple_omp_critical_name): Likewise.
3321 (gimple_omp_critical_name_ptr): Likewise.
3322 (gimple_omp_critical_set_name): Likewise.
3323 (gimple_omp_for_clauses): Likewise.
3324 (gimple_omp_for_clauses_ptr): Likewise.
3325 (gimple_omp_for_set_clauses): Likewise.
3326 (gimple_omp_for_collapse): Likewise.
3327 (gimple_omp_for_index): Likewise.
3328 (gimple_omp_for_index_ptr): Likewise.
3329 (gimple_omp_for_set_index): Likewise.
3330 (gimple_omp_for_initial): Likewise.
3331 (gimple_omp_for_initial_ptr): Likewise.
3332 (gimple_omp_for_set_initial): Likewise.
3333 (gimple_omp_for_final): Likewise.
3334 (gimple_omp_for_final_ptr): Likewise.
3335 (gimple_omp_for_set_final): Likewise.
3336 (gimple_omp_for_incr): Likewise.
3337 (gimple_omp_for_incr_ptr): Likewise.
3338 (gimple_omp_for_set_incr): Likewise.
3339 (gimple_omp_for_pre_body_ptr): Likewise.
3340 (gimple_omp_for_set_pre_body): Likewise.
3341 (gimple_omp_parallel_clauses): Likewise.
3342 (gimple_omp_parallel_clauses_ptr): Likewise.
3343 (gimple_omp_parallel_set_clauses): Likewise.
3344 (gimple_omp_parallel_child_fn): Likewise.
3345 (gimple_omp_parallel_child_fn_ptr): Likewise.
3346 (gimple_omp_parallel_set_child_fn): Likewise.
3347 (gimple_omp_parallel_data_arg): Likewise.
3348 (gimple_omp_parallel_data_arg_ptr): Likewise.
3349 (gimple_omp_parallel_set_data_arg): Likewise.
3350 (gimple_omp_task_clauses): Likewise.
3351 (gimple_omp_task_clauses_ptr): Likewise.
3352 (gimple_omp_task_set_clauses): Likewise.
3353 (gimple_omp_task_child_fn): Likewise.
3354 (gimple_omp_task_child_fn_ptr): Likewise.
3355 (gimple_omp_task_set_child_fn): Likewise.
3356 (gimple_omp_task_data_arg): Likewise.
3357 (gimple_omp_task_data_arg_ptr): Likewise.
3358 (gimple_omp_task_set_data_arg): Likewise.
3359 (gimple_omp_taskreg_clauses): Likewise.
3360 (gimple_omp_taskreg_clauses_ptr): Likewise.
3361 (gimple_omp_taskreg_set_clauses): Likewise.
3362 (gimple_omp_taskreg_child_fn): Likewise.
3363 (gimple_omp_taskreg_child_fn_ptr): Likewise.
3364 (gimple_omp_taskreg_set_child_fn): Likewise.
3365 (gimple_omp_taskreg_data_arg): Likewise.
3366 (gimple_omp_taskreg_data_arg_ptr): Likewise.
3367 (gimple_omp_taskreg_set_data_arg): Likewise.
3368 (gimple_omp_task_copy_fn): Likewise.
3369 (gimple_omp_task_copy_fn_ptr): Likewise.
3370 (gimple_omp_task_set_copy_fn): Likewise.
3371 (gimple_omp_task_arg_size): Likewise.
3372 (gimple_omp_task_arg_size_ptr): Likewise.
3373 (gimple_omp_task_set_arg_size): Likewise.
3374 (gimple_omp_task_arg_align): Likewise.
3375 (gimple_omp_task_arg_align_ptr): Likewise.
3376 (gimple_omp_task_set_arg_align): Likewise.
3377 (gimple_omp_single_clauses): Likewise.
3378 (gimple_omp_single_clauses_ptr): Likewise.
3379 (gimple_omp_single_set_clauses): Likewise.
3380 (gimple_omp_target_clauses): Likewise.
3381 (gimple_omp_target_clauses_ptr): Likewise.
3382 (gimple_omp_target_set_clauses): Likewise.
3383 (gimple_omp_target_child_fn): Likewise.
3384 (gimple_omp_target_child_fn_ptr): Likewise.
3385 (gimple_omp_target_set_child_fn): Likewise.
3386 (gimple_omp_target_data_arg): Likewise.
3387 (gimple_omp_target_data_arg_ptr): Likewise.
3388 (gimple_omp_target_set_data_arg): Likewise.
3389 (gimple_omp_teams_clauses): Likewise.
3390 (gimple_omp_teams_clauses_ptr): Likewise.
3391 (gimple_omp_teams_set_clauses): Likewise.
3392 (gimple_omp_sections_clauses): Likewise.
3393 (gimple_omp_sections_clauses_ptr): Likewise.
3394 (gimple_omp_sections_set_clauses): Likewise.
3395 (gimple_omp_sections_control): Likewise.
3396 (gimple_omp_sections_control_ptr): Likewise.
3397 (gimple_omp_sections_set_control): Likewise.
3398 (gimple_omp_for_set_cond): Likewise.
3399 (gimple_omp_for_cond): Likewise.
3400 (gimple_omp_atomic_store_set_val): Likewise.
3401 (gimple_omp_atomic_store_val): Likewise.
3402 (gimple_omp_atomic_store_val_ptr): Likewise.
3403 (gimple_omp_atomic_load_set_lhs): Likewise.
3404 (gimple_omp_atomic_load_lhs): Likewise.
3405 (gimple_omp_atomic_load_lhs_ptr): Likewise.
3406 (gimple_omp_atomic_load_set_rhs): Likewise.
3407 (gimple_omp_atomic_load_rhs): Likewise.
3408 (gimple_omp_atomic_load_rhs_ptr): Likewise.
3409 (gimple_omp_continue_control_def): Likewise.
3410 (gimple_omp_continue_control_def_ptr): Likewise.
3411 (gimple_omp_continue_set_control_def): Likewise.
3412 (gimple_omp_continue_control_use): Likewise.
3413 (gimple_omp_continue_control_use_ptr): Likewise.
3414 (gimple_omp_continue_set_control_use): Likewise.
3415 (gimple_transaction_body_ptr): Likewise.
3416 (gimple_transaction_label): Likewise.
3417 (gimple_transaction_label_ptr): Likewise.
3418 (gimple_transaction_set_body): Likewise.
3419 (gimple_transaction_set_label): Likewise.
3421 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
3422 * ipa-inline-analysis.c (inline_write_summary): Likewise.
3423 * ipa-ref.c (ipa_record_reference): Likewise.
3424 * ipa-reference.c (analyze_function): Likewise.
3425 (ipa_reference_write_optimization_summary): Likewise.
3426 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
3427 (address_taken_from_non_vtable_p): Likewise.
3428 (comdat_can_be_unshared_p_1): Likewise.
3429 * lto-cgraph.c (lto_output_ref): Likewise.
3430 (add_references): Likewise.
3431 (compute_ltrans_boundary): Likewise.
3432 (output_symtab): Likewise.
3433 (input_ref): Likewise.
3434 (input_cgraph_1): Likewise.
3435 (output_cgraph_opt_summary): Likewise.
3436 * lto-streamer-out.c (lto_output): Likewise.
3437 (output_symbol_p): Likewise.
3438 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
3439 (lsei_start_function_in_partition): Likewise.
3440 (lsei_next_variable_in_partition): Likewise.
3441 (lsei_start_variable_in_partition): Likewise.
3442 * symtab.c (insert_to_assembler_name_hash): Likewise.
3443 (unlink_from_assembler_name_hash): Likewise.
3444 (symtab_unregister_node): Likewise.
3445 (symtab_remove_node): Likewise.
3446 (dump_symtab_node): Likewise.
3447 (verify_symtab_base): Likewise.
3448 (verify_symtab_node): Likewise.
3449 (symtab_make_decl_local): Likewise.
3450 (symtab_alias_ultimate_target): Likewise.
3451 (symtab_resolve_alias): Likewise.
3452 (symtab_get_symbol_partitioning_class): Likewise.
3453 * tree-phinodes.c (allocate_phi_node): Likewise.
3454 (reserve_phi_args_for_new_edge): Likewise.
3455 (remove_phi_args): Likewise.
3456 * varpool.c (varpool_node_for_asm): Likewise.
3457 (varpool_remove_unreferenced_decls): Likewise.
3459 2014-04-23 Jeff Law <law@redhat.com>
3461 PR tree-optimization/60902
3462 * tree-ssa-threadedge.c
3463 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
3464 invalidate outputs from statements that do not produce useful
3465 outputs for threading.
3467 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
3469 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
3470 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
3471 machine descriptions for Stack Smashing Protector.
3473 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
3475 * aarch64.md (<optab>_rol<mode>3): New pattern.
3476 (<optab>_rolsi3_uxtw): Likewise.
3477 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
3479 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
3481 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
3482 (arm_cortex_a12_tune): Likewise.
3484 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3486 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
3488 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3490 * config/arm/arm.md (arm_rev16si2): New pattern.
3491 (arm_rev16si2_alt): Likewise.
3492 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
3494 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3496 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
3497 (rev16<mode>2_alt): Likewise.
3498 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
3499 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
3500 (aarch_rev16_shleft_mask_imm_p): Likewise.
3501 (aarch_rev16_p_1): Likewise.
3502 (aarch_rev16_p): Likewise.
3503 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
3504 (aarch_rev16_shright_mask_imm_p): Likewise.
3505 (aarch_rev16_shleft_mask_imm_p): Likewise.
3507 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3509 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
3510 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
3512 (cortex_a53_extra_costs): Likewise.
3513 (cortex_a57_extra_costs): Likewise.
3514 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
3515 (cortexa7_extra_costs): Likewise.
3516 (cortexa8_extra_costs): Likewise.
3517 (cortexa12_extra_costs): Likewise.
3518 (cortexa15_extra_costs): Likewise.
3519 (v7m_extra_costs): Likewise.
3520 (arm_new_rtx_costs): Handle BSWAP.
3522 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3524 * config/arm/arm.c (cortexa8_extra_costs): New table.
3525 (arm_cortex_a8_tune): New tuning struct.
3526 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
3528 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3530 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
3532 2014-04-23 Richard Biener <rguenther@suse.de>
3534 * Makefile.in (OBJS): Remove loop-unswitch.o.
3535 * tree-pass.h (make_pass_rtl_unswitch): Remove.
3536 * passes.def (pass_rtl_unswitch): Likewise.
3537 * loop-init.c (gate_rtl_unswitch): Likewise.
3538 (rtl_unswitch): Likewise.
3539 (pass_data_rtl_unswitch): Likewise.
3540 (pass_rtl_unswitch): Likewise.
3541 (make_pass_rtl_unswitch): Likewise.
3542 * rtl.h (reversed_condition): Likewise.
3543 (compare_and_jump_seq): Likewise.
3544 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
3546 * loop-unroll.c (compare_and_jump_seq): Likewise.
3548 2014-04-23 Richard Biener <rguenther@suse.de>
3550 PR tree-optimization/60903
3551 * tree-ssa-loop-im.c (analyze_memory_references): Remove
3552 commented code block.
3553 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
3554 loop flags to newly created BBs and edges.
3556 2014-04-23 Nick Clifton <nickc@redhat.com>
3558 * config/msp430/msp430.c (msp430_handle_option): Move function
3560 (msp430_option_override): Simplify mcu and mcpu option handling.
3561 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
3562 support for -mhwmult command line option.
3563 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
3564 -mhwmult command line option.
3565 (msp430_hwmult_enabled): Delete.
3566 (msp43o_output_labelref): Add support for -mhwmult command line option.
3567 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
3568 (umulsidi3): Likewise.
3569 * config/msp430/msp430.opt (mmcu): Add Report attribute.
3570 (mcpu, mlarge, msmall): Likewise.
3571 (mhwmult): New option.
3572 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
3574 (msp430_is_f5_mcu): Remove prototype.
3575 (msp430_use_f5_series_hwmult): Add prototype.
3576 * config/msp430/msp430-opts.h: New file.
3577 * common/config/msp430: New directory.
3578 * common/config/msp430/msp430-common.c: New file.
3579 * config.gcc (msp430): Remove target_has_targetm_common.
3580 * doc/invoke.texi: Document -mhwmult command line option.
3582 2014-04-23 Nick Clifton <nickc@redhat.com>
3584 * config/i386/cygwin.h (ENDFILE_SPEC): Include
3585 default-manifest.o if it can be found in the search path.
3586 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
3588 2014-04-23 Terry Guo <terry.guo@arm.com>
3590 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
3592 2014-04-23 Richard Biener <rguenther@suse.de>
3595 * tree-inline.c (declare_return_variable): Use mark_addressable.
3597 2014-04-23 Richard Biener <rguenther@suse.de>
3600 * loop-init.c (loop_optimizer_init): Make sure to apply
3601 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
3603 2014-04-22 Jakub Jelinek <jakub@redhat.com>
3606 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
3608 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
3609 if flag_sanitize_undefined_trap_on_error.
3610 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
3611 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
3612 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
3613 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
3614 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
3615 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
3616 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
3617 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
3618 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
3619 * ubsan.c (ubsan_instrument_unreachable): Return
3620 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
3621 (ubsan_expand_null_ifn): Emit __builtin_trap ()
3622 if flag_sanitize_undefined_trap_on_error and
3623 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
3624 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
3625 instrument_bool_enum_load): Emit __builtin_trap () if
3626 flag_sanitize_undefined_trap_on_error and
3627 __builtin_handle_*_abort () if !flag_sanitize_recover.
3628 * doc/invoke.texi (-fsanitize-recover,
3629 -fsanitize-undefined-trap-on-error): Document.
3631 2014-04-22 Christian Bruel <christian.bruel@st.com>
3633 * config/sh/sh.md (mov<mode>): Replace movQIHI.
3634 Force immediates to SImode.
3636 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
3638 * config/nios2/nios2.md (UNSPEC_ROUND): New.
3640 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
3641 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
3642 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
3643 (nios2_fpu_insn): Add entry for round.
3644 (N2FPU_NO_ERRNO_P): Define.
3645 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
3647 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
3649 2014-04-22 Richard Henderson <rth@redhat.com>
3651 * config/aarch64/aarch64 (addti3, subti3): New expanders.
3652 (add<GPI>3_compare0): Remove leading * from name.
3653 (add<GPI>3_carryin): Likewise.
3654 (sub<GPI>3_compare0): Likewise.
3655 (sub<GPI>3_carryin): Likewise.
3656 (<su_optab>mulditi3): New expander.
3657 (multi3): New expander.
3658 (madd<GPI>): Remove leading * from name.
3660 2014-04-22 Martin Jambor <mjambor@suse.cz>
3662 * cgraphclones.c (cgraph_function_versioning): Copy
3663 ipa_transforms_to_apply instead of asserting it is empty.
3665 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
3668 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
3669 on count_exp to get mode.
3671 2014-04-22 Andrew Pinski <apinski@cavium.com>
3673 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
3674 Handle TLS for ILP32.
3675 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
3676 (tlsie_small_<mode>): this and handle PTR.
3677 (tlsie_small_sidi): New pattern.
3678 (tlsle_small): Change to an expand to handle ILP32.
3679 (tlsle_small_<mode>): New pattern.
3680 (tlsdesc_small): Rename to ...
3681 (tlsdesc_small_<mode>): this and handle PTR.
3683 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3685 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
3687 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3689 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
3690 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
3691 (aarch64_types_signed_poly_qualifiers): Likewise.
3692 (aarch64_types_unsigned_signed_qualifiers): Likewise.
3693 (aarch64_types_poly_signed_qualifiers): Likewise.
3694 (TYPES_REINTERP_SS): Type macro added.
3695 (TYPES_REINTERP_SU): Likewise.
3696 (TYPES_REINTERP_SP): Likewise.
3697 (TYPES_REINTERP_US): Likewise.
3698 (TYPES_REINTERP_PS): Likewise.
3699 (aarch64_fold_builtin): New expression folding added.
3700 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
3701 Declarations removed.
3702 (REINTERP_SS): Declarations added.
3703 (REINTERP_US): Likewise.
3704 (REINTERP_PS): Likewise.
3705 (REINTERP_SU): Likewise.
3706 (REINTERP_SP): Likewise.
3707 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
3708 (vreinterpretq_p8_f64): Likewise.
3709 (vreinterpret_p16_f64): Likewise.
3710 (vreinterpretq_p16_f64): Likewise.
3711 (vreinterpret_f32_f64): Likewise.
3712 (vreinterpretq_f32_f64): Likewise.
3713 (vreinterpret_f64_f32): Likewise.
3714 (vreinterpret_f64_p8): Likewise.
3715 (vreinterpret_f64_p16): Likewise.
3716 (vreinterpret_f64_s8): Likewise.
3717 (vreinterpret_f64_s16): Likewise.
3718 (vreinterpret_f64_s32): Likewise.
3719 (vreinterpret_f64_s64): Likewise.
3720 (vreinterpret_f64_u8): Likewise.
3721 (vreinterpret_f64_u16): Likewise.
3722 (vreinterpret_f64_u32): Likewise.
3723 (vreinterpret_f64_u64): Likewise.
3724 (vreinterpretq_f64_f32): Likewise.
3725 (vreinterpretq_f64_p8): Likewise.
3726 (vreinterpretq_f64_p16): Likewise.
3727 (vreinterpretq_f64_s8): Likewise.
3728 (vreinterpretq_f64_s16): Likewise.
3729 (vreinterpretq_f64_s32): Likewise.
3730 (vreinterpretq_f64_s64): Likewise.
3731 (vreinterpretq_f64_u8): Likewise.
3732 (vreinterpretq_f64_u16): Likewise.
3733 (vreinterpretq_f64_u32): Likewise.
3734 (vreinterpretq_f64_u64): Likewise.
3735 (vreinterpret_s64_f64): Likewise.
3736 (vreinterpretq_s64_f64): Likewise.
3737 (vreinterpret_u64_f64): Likewise.
3738 (vreinterpretq_u64_f64): Likewise.
3739 (vreinterpret_s8_f64): Likewise.
3740 (vreinterpretq_s8_f64): Likewise.
3741 (vreinterpret_s16_f64): Likewise.
3742 (vreinterpretq_s16_f64): Likewise.
3743 (vreinterpret_s32_f64): Likewise.
3744 (vreinterpretq_s32_f64): Likewise.
3745 (vreinterpret_u8_f64): Likewise.
3746 (vreinterpretq_u8_f64): Likewise.
3747 (vreinterpret_u16_f64): Likewise.
3748 (vreinterpretq_u16_f64): Likewise.
3749 (vreinterpret_u32_f64): Likewise.
3750 (vreinterpretq_u32_f64): Likewise.
3752 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3754 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
3755 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
3756 (vreinterpret_p8_s8): Likewise.
3757 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
3758 (vreinterpret_p8_s16): Likewise.
3759 (vreinterpret_p8_s32): Likewise.
3760 (vreinterpret_p8_s64): Likewise.
3761 (vreinterpret_p8_f32): Likewise.
3762 (vreinterpret_p8_u8): Likewise.
3763 (vreinterpret_p8_u16): Likewise.
3764 (vreinterpret_p8_u32): Likewise.
3765 (vreinterpret_p8_u64): Likewise.
3766 (vreinterpret_p8_p16): Likewise.
3767 (vreinterpretq_p8_s8): Likewise.
3768 (vreinterpretq_p8_s16): Likewise.
3769 (vreinterpretq_p8_s32): Likewise.
3770 (vreinterpretq_p8_s64): Likewise.
3771 (vreinterpretq_p8_f32): Likewise.
3772 (vreinterpretq_p8_u8): Likewise.
3773 (vreinterpretq_p8_u16): Likewise.
3774 (vreinterpretq_p8_u32): Likewise.
3775 (vreinterpretq_p8_u64): Likewise.
3776 (vreinterpretq_p8_p16): Likewise.
3777 (vreinterpret_p16_s8): Likewise.
3778 (vreinterpret_p16_s16): Likewise.
3779 (vreinterpret_p16_s32): Likewise.
3780 (vreinterpret_p16_s64): Likewise.
3781 (vreinterpret_p16_f32): Likewise.
3782 (vreinterpret_p16_u8): Likewise.
3783 (vreinterpret_p16_u16): Likewise.
3784 (vreinterpret_p16_u32): Likewise.
3785 (vreinterpret_p16_u64): Likewise.
3786 (vreinterpret_p16_p8): Likewise.
3787 (vreinterpretq_p16_s8): Likewise.
3788 (vreinterpretq_p16_s16): Likewise.
3789 (vreinterpretq_p16_s32): Likewise.
3790 (vreinterpretq_p16_s64): Likewise.
3791 (vreinterpretq_p16_f32): Likewise.
3792 (vreinterpretq_p16_u8): Likewise.
3793 (vreinterpretq_p16_u16): Likewise.
3794 (vreinterpretq_p16_u32): Likewise.
3795 (vreinterpretq_p16_u64): Likewise.
3796 (vreinterpretq_p16_p8): Likewise.
3797 (vreinterpret_f32_s8): Likewise.
3798 (vreinterpret_f32_s16): Likewise.
3799 (vreinterpret_f32_s32): Likewise.
3800 (vreinterpret_f32_s64): Likewise.
3801 (vreinterpret_f32_u8): Likewise.
3802 (vreinterpret_f32_u16): Likewise.
3803 (vreinterpret_f32_u32): Likewise.
3804 (vreinterpret_f32_u64): Likewise.
3805 (vreinterpret_f32_p8): Likewise.
3806 (vreinterpret_f32_p16): Likewise.
3807 (vreinterpretq_f32_s8): Likewise.
3808 (vreinterpretq_f32_s16): Likewise.
3809 (vreinterpretq_f32_s32): Likewise.
3810 (vreinterpretq_f32_s64): Likewise.
3811 (vreinterpretq_f32_u8): Likewise.
3812 (vreinterpretq_f32_u16): Likewise.
3813 (vreinterpretq_f32_u32): Likewise.
3814 (vreinterpretq_f32_u64): Likewise.
3815 (vreinterpretq_f32_p8): Likewise.
3816 (vreinterpretq_f32_p16): Likewise.
3817 (vreinterpret_s64_s8): Likewise.
3818 (vreinterpret_s64_s16): Likewise.
3819 (vreinterpret_s64_s32): Likewise.
3820 (vreinterpret_s64_f32): Likewise.
3821 (vreinterpret_s64_u8): Likewise.
3822 (vreinterpret_s64_u16): Likewise.
3823 (vreinterpret_s64_u32): Likewise.
3824 (vreinterpret_s64_u64): Likewise.
3825 (vreinterpret_s64_p8): Likewise.
3826 (vreinterpret_s64_p16): Likewise.
3827 (vreinterpretq_s64_s8): Likewise.
3828 (vreinterpretq_s64_s16): Likewise.
3829 (vreinterpretq_s64_s32): Likewise.
3830 (vreinterpretq_s64_f32): Likewise.
3831 (vreinterpretq_s64_u8): Likewise.
3832 (vreinterpretq_s64_u16): Likewise.
3833 (vreinterpretq_s64_u32): Likewise.
3834 (vreinterpretq_s64_u64): Likewise.
3835 (vreinterpretq_s64_p8): Likewise.
3836 (vreinterpretq_s64_p16): Likewise.
3837 (vreinterpret_u64_s8): Likewise.
3838 (vreinterpret_u64_s16): Likewise.
3839 (vreinterpret_u64_s32): Likewise.
3840 (vreinterpret_u64_s64): Likewise.
3841 (vreinterpret_u64_f32): Likewise.
3842 (vreinterpret_u64_u8): Likewise.
3843 (vreinterpret_u64_u16): Likewise.
3844 (vreinterpret_u64_u32): Likewise.
3845 (vreinterpret_u64_p8): Likewise.
3846 (vreinterpret_u64_p16): Likewise.
3847 (vreinterpretq_u64_s8): Likewise.
3848 (vreinterpretq_u64_s16): Likewise.
3849 (vreinterpretq_u64_s32): Likewise.
3850 (vreinterpretq_u64_s64): Likewise.
3851 (vreinterpretq_u64_f32): Likewise.
3852 (vreinterpretq_u64_u8): Likewise.
3853 (vreinterpretq_u64_u16): Likewise.
3854 (vreinterpretq_u64_u32): Likewise.
3855 (vreinterpretq_u64_p8): Likewise.
3856 (vreinterpretq_u64_p16): Likewise.
3857 (vreinterpret_s8_s16): Likewise.
3858 (vreinterpret_s8_s32): Likewise.
3859 (vreinterpret_s8_s64): Likewise.
3860 (vreinterpret_s8_f32): Likewise.
3861 (vreinterpret_s8_u8): Likewise.
3862 (vreinterpret_s8_u16): Likewise.
3863 (vreinterpret_s8_u32): Likewise.
3864 (vreinterpret_s8_u64): Likewise.
3865 (vreinterpret_s8_p8): Likewise.
3866 (vreinterpret_s8_p16): Likewise.
3867 (vreinterpretq_s8_s16): Likewise.
3868 (vreinterpretq_s8_s32): Likewise.
3869 (vreinterpretq_s8_s64): Likewise.
3870 (vreinterpretq_s8_f32): Likewise.
3871 (vreinterpretq_s8_u8): Likewise.
3872 (vreinterpretq_s8_u16): Likewise.
3873 (vreinterpretq_s8_u32): Likewise.
3874 (vreinterpretq_s8_u64): Likewise.
3875 (vreinterpretq_s8_p8): Likewise.
3876 (vreinterpretq_s8_p16): Likewise.
3877 (vreinterpret_s16_s8): Likewise.
3878 (vreinterpret_s16_s32): Likewise.
3879 (vreinterpret_s16_s64): Likewise.
3880 (vreinterpret_s16_f32): Likewise.
3881 (vreinterpret_s16_u8): Likewise.
3882 (vreinterpret_s16_u16): Likewise.
3883 (vreinterpret_s16_u32): Likewise.
3884 (vreinterpret_s16_u64): Likewise.
3885 (vreinterpret_s16_p8): Likewise.
3886 (vreinterpret_s16_p16): Likewise.
3887 (vreinterpretq_s16_s8): Likewise.
3888 (vreinterpretq_s16_s32): Likewise.
3889 (vreinterpretq_s16_s64): Likewise.
3890 (vreinterpretq_s16_f32): Likewise.
3891 (vreinterpretq_s16_u8): Likewise.
3892 (vreinterpretq_s16_u16): Likewise.
3893 (vreinterpretq_s16_u32): Likewise.
3894 (vreinterpretq_s16_u64): Likewise.
3895 (vreinterpretq_s16_p8): Likewise.
3896 (vreinterpretq_s16_p16): Likewise.
3897 (vreinterpret_s32_s8): Likewise.
3898 (vreinterpret_s32_s16): Likewise.
3899 (vreinterpret_s32_s64): Likewise.
3900 (vreinterpret_s32_f32): Likewise.
3901 (vreinterpret_s32_u8): Likewise.
3902 (vreinterpret_s32_u16): Likewise.
3903 (vreinterpret_s32_u32): Likewise.
3904 (vreinterpret_s32_u64): Likewise.
3905 (vreinterpret_s32_p8): Likewise.
3906 (vreinterpret_s32_p16): Likewise.
3907 (vreinterpretq_s32_s8): Likewise.
3908 (vreinterpretq_s32_s16): Likewise.
3909 (vreinterpretq_s32_s64): Likewise.
3910 (vreinterpretq_s32_f32): Likewise.
3911 (vreinterpretq_s32_u8): Likewise.
3912 (vreinterpretq_s32_u16): Likewise.
3913 (vreinterpretq_s32_u32): Likewise.
3914 (vreinterpretq_s32_u64): Likewise.
3915 (vreinterpretq_s32_p8): Likewise.
3916 (vreinterpretq_s32_p16): Likewise.
3917 (vreinterpret_u8_s8): Likewise.
3918 (vreinterpret_u8_s16): Likewise.
3919 (vreinterpret_u8_s32): Likewise.
3920 (vreinterpret_u8_s64): Likewise.
3921 (vreinterpret_u8_f32): Likewise.
3922 (vreinterpret_u8_u16): Likewise.
3923 (vreinterpret_u8_u32): Likewise.
3924 (vreinterpret_u8_u64): Likewise.
3925 (vreinterpret_u8_p8): Likewise.
3926 (vreinterpret_u8_p16): Likewise.
3927 (vreinterpretq_u8_s8): Likewise.
3928 (vreinterpretq_u8_s16): Likewise.
3929 (vreinterpretq_u8_s32): Likewise.
3930 (vreinterpretq_u8_s64): Likewise.
3931 (vreinterpretq_u8_f32): Likewise.
3932 (vreinterpretq_u8_u16): Likewise.
3933 (vreinterpretq_u8_u32): Likewise.
3934 (vreinterpretq_u8_u64): Likewise.
3935 (vreinterpretq_u8_p8): Likewise.
3936 (vreinterpretq_u8_p16): Likewise.
3937 (vreinterpret_u16_s8): Likewise.
3938 (vreinterpret_u16_s16): Likewise.
3939 (vreinterpret_u16_s32): Likewise.
3940 (vreinterpret_u16_s64): Likewise.
3941 (vreinterpret_u16_f32): Likewise.
3942 (vreinterpret_u16_u8): Likewise.
3943 (vreinterpret_u16_u32): Likewise.
3944 (vreinterpret_u16_u64): Likewise.
3945 (vreinterpret_u16_p8): Likewise.
3946 (vreinterpret_u16_p16): Likewise.
3947 (vreinterpretq_u16_s8): Likewise.
3948 (vreinterpretq_u16_s16): Likewise.
3949 (vreinterpretq_u16_s32): Likewise.
3950 (vreinterpretq_u16_s64): Likewise.
3951 (vreinterpretq_u16_f32): Likewise.
3952 (vreinterpretq_u16_u8): Likewise.
3953 (vreinterpretq_u16_u32): Likewise.
3954 (vreinterpretq_u16_u64): Likewise.
3955 (vreinterpretq_u16_p8): Likewise.
3956 (vreinterpretq_u16_p16): Likewise.
3957 (vreinterpret_u32_s8): Likewise.
3958 (vreinterpret_u32_s16): Likewise.
3959 (vreinterpret_u32_s32): Likewise.
3960 (vreinterpret_u32_s64): Likewise.
3961 (vreinterpret_u32_f32): Likewise.
3962 (vreinterpret_u32_u8): Likewise.
3963 (vreinterpret_u32_u16): Likewise.
3964 (vreinterpret_u32_u64): Likewise.
3965 (vreinterpret_u32_p8): Likewise.
3966 (vreinterpret_u32_p16): Likewise.
3967 (vreinterpretq_u32_s8): Likewise.
3968 (vreinterpretq_u32_s16): Likewise.
3969 (vreinterpretq_u32_s32): Likewise.
3970 (vreinterpretq_u32_s64): Likewise.
3971 (vreinterpretq_u32_f32): Likewise.
3972 (vreinterpretq_u32_u8): Likewise.
3973 (vreinterpretq_u32_u16): Likewise.
3974 (vreinterpretq_u32_u64): Likewise.
3975 (vreinterpretq_u32_p8): Likewise.
3976 (vreinterpretq_u32_p16): Likewise.
3978 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3980 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
3982 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator extended.
3984 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
3985 (vqnegd_s64): Likewise.
3986 (vqabs_s64): Likewise.
3987 (vqabsd_s64): Likewise.
3989 2014-04-22 Richard Henderson <rth@redhat.com>
3991 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
3992 computation to the top of the loop.
3994 2014-04-22 Renlin <renlin.li@arm.com>
3995 Jiong Wang <jiong.wang@arm.com>
3997 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
3998 * config/aarch64/aarch64.c (aarch64_layout_frame)
3999 (aarch64_initial_elimination_offset): Likewise.
4001 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
4003 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
4006 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
4008 * machmode.h (bitwise_mode_for_mode): Declare.
4009 * stor-layout.h (bitwise_type_for_mode): Likewise.
4010 * stor-layout.c (bitwise_mode_for_mode): New function.
4011 (bitwise_type_for_mode): Likewise.
4012 * builtins.c (fold_builtin_memory_op): Use it instead of
4013 int_mode_for_mode and build_nonstandard_integer_type.
4015 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4017 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
4018 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
4019 (*-*-solaris2*): Simplify.
4020 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
4021 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
4022 *-*-solaris2.9* handling.
4024 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
4026 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
4027 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
4029 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
4030 * configure: Regenerate.
4032 * config/i386/sol2-9.h: Remove.
4034 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
4035 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
4036 Remove Solaris 9 references.
4038 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
4040 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
4041 (floatuns<GPI:mode><GPF:mode>2): Remove.
4042 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
4043 and floatuns conversions.
4044 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
4045 and floatuns conversions.
4046 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
4047 (w1,w2): New mode attributes for inequal width conversions.
4049 2014-04-22 Renlin Li <Renlin.Li@arm.com>
4051 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
4052 the output asm format.
4054 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
4056 * config/aarch64/aarch64-simd.md
4057 (aarch64_cm<optab>di): Always split.
4058 (*aarch64_cm<optab>di): New.
4059 (aarch64_cmtstdi): Always split.
4060 (*aarch64_cmtstdi): New.
4062 2014-04-22 Jakub Jelinek <jakub@redhat.com>
4064 PR tree-optimization/60823
4065 * omp-low.c (ipa_simd_modify_function_body): Go through
4066 all SSA_NAMEs and for those refering to vector arguments
4067 which are going to be replaced adjust SSA_NAME_VAR and,
4068 if it is a default definition, change it into a non-default
4069 definition assigned at the beginning of function from new_decl.
4070 (ipa_simd_modify_stmt_ops): Rewritten.
4071 * tree-dfa.c (set_ssa_default_def): When removing default def,
4072 check for NULL loc instead of NULL *loc.
4074 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4076 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
4077 restrictions on core registers for DImode values in Thumb2.
4079 2014-04-22 Ian Bolton <ian.bolton@arm.com>
4081 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
4082 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
4084 2014-04-22 Ian Bolton <ian.bolton@arm.com>
4086 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
4087 (*iordi_notzesidi_di): Likewise.
4088 (*iordi_notsesidi_di): Likewise.
4090 2014-04-22 Ian Bolton <ian.bolton@arm.com>
4092 * config/arm/arm-protos.h (tune_params): New struct members.
4093 * config/arm/arm.c: Initialise tune_params per processor.
4094 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
4095 for speed, based on new tune_params.
4097 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
4099 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro added.
4100 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added macro.
4101 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment corrected.
4102 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
4103 * config/aarch64/arm_neon.h (vrnd_f64): Added.
4104 (vrnda_f64): Likewise.
4105 (vrndi_f64): Likewise.
4106 (vrndm_f64): Likewise.
4107 (vrndn_f64): Likewise.
4108 (vrndp_f64): Likewise.
4109 (vrndx_f64): Likewise.
4111 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
4113 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
4114 GET_MODE_SIZE argument is enum machine_mode.
4116 2014-04-22 Jakub Jelinek <jakub@redhat.com>
4119 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
4120 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
4122 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
4125 * asan.c (asan_emit_stack_protection): Force the base to align to
4126 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
4127 appropriate bits if STRICT_ALIGNMENT.
4128 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
4130 (expand_used_vars): Leave a space in the stack frame for alignment
4131 if STRICT_ALIGNMENT.
4133 2014-04-21 David Malcolm <dmalcolm@redhat.com>
4135 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
4137 (gimple_store_p): Likewise.
4138 (gimple_assign_load_p): Likewise.
4139 (gimple_assign_cast_p): Likewise.
4140 (gimple_clobber_p): Likewise.
4142 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
4143 rather than a gimple.
4144 (gimple_assign_cast_p): Likewise.
4146 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
4149 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
4150 If mode is DDmode and TARGET_E500_DOUBLE allow move.
4152 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
4153 more debug information for E500 if -mdebug=reg.
4155 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
4158 * config/i386/i386.c (ix86_expand_builtin)
4159 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
4160 register for target RTX.
4161 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
4163 2014-04-18 Cong Hou <congh@google.com>
4165 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
4166 the widen-mult pattern by handling two operands with different sizes,
4167 and operands whose size is smaller than half of the result type.
4169 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
4171 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
4172 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
4173 (do_estimate_edge_time): Compute it.
4174 * ipa-inline.c (want_inline_small_function_p): Bypass
4175 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
4177 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
4179 * ipa-inline.c (spec_rem): New static variable.
4180 (dump_overall_stats): New function.
4181 (dump_inline_stats): New function.
4183 2014-04-18 Richard Henderson <rth@redhat.com>
4185 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
4186 to GET_MODE_SIZE, not a reg_class_t.
4188 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4190 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
4191 (vsx_xxmrglw_<mode>): Likewise.
4193 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
4196 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
4197 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
4198 (rs6000_init_hard_regno_mode_ok): Likewise.
4200 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
4202 * ipa-inline.c (inline_small_functions): Account only non-cold
4204 * doc/invoke.texi (inline-unit-growth): Update documentation.
4206 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
4208 * config/rs6000/rs6000.md (addti3, subti3): New.
4210 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
4213 * config/i386/i386.c (ix86_expand_clear): Remove outdated
4214 comment. Check optimize_insn_for_size_p instead of
4215 optimize_insn_for_speed_p.
4217 2014-04-17 Martin Jambor <mjambor@suse.cz>
4219 * gimple-iterator.c (gsi_start_edge): New function.
4220 * gimple-iterator.h (gsi_start_edge): Declare.
4221 * tree-sra.c (single_non_eh_succ): New function.
4222 (disqualify_ops_if_throwing_stmt): Renamed to
4223 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
4224 having one non-EH successor BB.
4225 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
4226 generate loads into replacements.
4227 (sra_modify_assign): Likewise and and also use the simple path for
4229 (sra_modify_function_body): Commit statements on edges.
4231 2014-04-17 Richard Biener <rguenther@suse.de>
4234 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
4235 comparison results and add clarifying comment.
4237 2014-04-17 Jakub Jelinek <jakub@redhat.com>
4239 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
4240 (blank_mode): Initialize it.
4241 (emit_mode_size_inline, emit_mode_nunits_inline,
4242 emit_mode_inner_inline): New functions.
4243 (emit_insn_modes_h): Call them and surround their output with
4244 #if GCC_VERSION >= 4001 ... #endif.
4245 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
4246 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
4247 mode_* arrays if the argument is __builtin_constant_p.
4248 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
4249 is enum machine_mode.
4251 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
4253 * passes.c (opt_pass::execute): Adjust.
4254 (pass_manager::execute_pass_mode_switching): Likewise.
4255 (early_local_passes::execute): Likewise.
4256 (execute_one_pass): Pass cfun to the pass's execute method.
4257 * tree-pass.h (opt_pass::execute): Add function * argument.
4258 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
4259 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
4260 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
4261 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
4262 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
4263 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
4264 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
4265 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
4266 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
4267 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
4268 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
4269 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
4270 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
4271 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
4272 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
4273 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
4274 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
4275 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
4276 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
4277 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
4278 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
4279 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
4280 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
4281 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
4282 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
4283 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
4284 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
4285 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
4286 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
4289 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
4291 * passes.c (opt_pass::gate): Take function * argument.
4292 (gate_all_early_local_passes): Merge into
4293 (early_local_passes::gate): this.
4294 (gate_all_early_optimizations): Merge into
4295 (all_early_optimizations::gate): this.
4296 (gate_all_optimizations): Mege into
4297 (all_optimizations::gate): this.
4298 (gate_all_optimizations_g): Merge into
4299 (all_optimizations_g::gate): this.
4300 (gate_rest_of_compilation): Mege into
4301 (rest_of_compilation::gate): this.
4302 (gate_postreload): Merge into
4303 (postreload::gate): this.
4304 (dump_one_pass): Pass cfun to the pass's gate method.
4305 (execute_ipa_summary_passes): Likewise.
4306 (execute_one_pass): Likewise.
4307 (ipa_write_summaries_2): Likewise.
4308 (ipa_write_optimization_summaries_1): Likewise.
4309 (ipa_read_summaries_1): Likewise.
4310 (ipa_read_optimization_summaries_1): Likewise.
4311 (execute_ipa_stmt_fixups): Likewise.
4312 * tree-pass.h (opt_pass::gate): Add function * argument.
4313 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
4314 combine-stack-adj.c, combine.c, compare-elim.c,
4315 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
4316 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
4317 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
4318 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
4319 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
4320 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
4321 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
4322 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
4323 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
4324 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
4325 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
4326 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
4327 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
4328 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
4329 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
4330 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
4331 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
4332 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
4333 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
4334 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
4335 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
4336 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
4337 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
4338 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
4339 var-tracking.c, vtable-verify.c, web.c: Adjust.
4341 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
4343 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
4344 * configure: Regenerate.
4346 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
4348 * passes.c (dump_one_pass): don't check pass->has_gate.
4349 (execute_ipa_summary_passes): Likewise.
4350 (execute_one_pass): Likewise.
4351 (ipa_write_summaries_2): Likewise.
4352 (ipa_write_optimization_summaries_1): Likewise.
4353 (ipa_read_optimization_summaries_1): Likewise.
4354 (execute_ipa_stmt_fixups): Likewise.
4355 * tree-pass.h (pass_data::has_gate): Remove.
4356 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
4357 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
4358 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
4359 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
4360 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
4361 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
4362 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
4363 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
4364 gimple-low.c, gimple-ssa-isolate-paths.c,
4365 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
4366 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
4367 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
4368 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
4369 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
4370 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
4371 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
4372 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
4373 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
4374 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
4375 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
4376 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
4377 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
4378 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
4379 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
4380 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
4381 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
4382 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
4383 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
4384 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
4385 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
4386 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
4389 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
4391 * pass_manager.h (pass_manager::register_dump_files_1): Remove
4393 * passes.c (pass_manager::register_dump_files_1): Merge into
4394 (pass_manager::register_dump_files): this, and remove its handling of
4395 properties since the pass always has the properties anyway.
4396 (pass_manager::pass_manager): Adjust.
4398 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
4400 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
4401 * passes.c (pass_manager::register_dump_files_1): Remove dead code
4402 dealing with properties.
4403 (pass_manager::register_dump_files): Adjust.
4405 2014-03-20 Mark Wielaard <mjw@redhat.com>
4407 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
4408 then represent the bound as normal constant value.
4410 2014-04-17 Jakub Jelinek <jakub@redhat.com>
4413 Forward port from 4.8 branch
4414 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
4416 * config/i386/bmiintrin.h (_blsi_u32): New.
4420 (_blsmsk_u32): Ditto.
4421 (_blsmsk_u64): Ditto.
4422 (_tzcnt_u32): Ditto.
4423 (_tzcnt_u64): Ditto.
4425 2014-04-17 Kito Cheng <kito@0xlab.org>
4427 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
4429 2014-04-17 Richard Biener <rguenther@suse.de>
4432 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
4433 boolean results for comparisons.
4435 2014-04-17 Richard Biener <rguenther@suse.de>
4437 PR tree-optimization/60836
4438 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
4439 initial PHI args to be gimple values.
4441 2014-04-17 Richard Biener <rguenther@suse.de>
4443 PR tree-optimization/60841
4444 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
4445 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
4446 of stmts to SLP build.
4447 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
4448 (vect_analyze_slp): Likewise.
4449 (vect_analyze_slp_instance): Likewise.
4450 (vect_build_slp_tree): Limit overall SLP tree growth.
4451 * tree-vectorizer.h (vect_analyze_data_refs,
4452 vect_analyze_slp): Adjust prototypes.
4454 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
4456 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
4459 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
4461 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
4462 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
4463 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
4464 for TARGET_SLOW_PSHUFB
4466 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
4468 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
4469 * config/i386/i386.c (intel_cost): Ditto.
4471 2014-04-17 Joey Ye <joey.ye@arm.com>
4473 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
4475 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
4477 * opts.c (common_handle_option): Disable -fipa-reference coorectly
4480 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
4482 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
4483 (type_all_derivations_known_p): New predicate.
4484 (type_all_ctors_visible_p): New predicate.
4485 (type_possibly_instantiated_p): New predicate.
4486 (get_odr_type): Compute all_derivations_known.
4487 (dump_odr_type): Dump the flag.
4488 (maybe_record_type): Cleanup.
4489 (record_target_from_binfo): Add bases_to_consider array;
4490 record bases for types w/o instances and skip CXX destructor.
4491 (possible_polymorphic_call_targets_1): Add bases_to_consider
4492 and consider_construction parameters; check if type may have instance.
4493 (get_polymorphic_call_info): Set maybe_in_construction to true
4494 when we know nothing.
4495 (record_targets_from_bases): Skip CXX destructors; they are
4496 never called for types in construction.
4497 (possible_polymorphic_call_targets): Do not record target when
4498 type may not have instance.
4500 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
4503 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
4504 external aliases alive, too.
4506 2014-04-16 Andrew Pinski <apinski@cavium.com>
4508 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
4511 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
4513 * final.c (compute_alignments): Do not apply loop alignment to a block
4514 falling through to the exit.
4516 2014-04-16 Catherine Moore <clm@codesourcery.com>
4518 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
4519 Adjust constraints for microMIPS store patterns.
4521 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
4523 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
4525 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
4527 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
4528 (append_use): Run at -O0.
4529 (append_vdef): Likewise.
4530 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
4531 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
4533 2014-04-16 Jakub Jelinek <jakub@redhat.com>
4535 PR tree-optimization/60844
4536 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
4537 (propagate_op_to_single_use, remove_visited_stmt_chain,
4538 linearize_expr, repropagate_negates, reassociate_bb): Use it
4539 instead of gsi_remove.
4541 2014-04-16 Martin Jambor <mjambor@suse.cz>
4543 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
4544 ipa_transforms_to_apply.
4545 (cgraph_function_versioning): Assert that old_node has empty
4546 ipa_transforms_to_apply.
4547 * trans-mem.c (ipa_tm_create_version): Likewise.
4548 * tree-inline.c (tree_function_versioning): Do not duplicate
4549 ipa_transforms_to_apply.
4551 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4554 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
4556 Pass necessary as flags on 64-bit Solaris/x86.
4557 Use lowercase relocs for x86_64-*-*.
4558 * configure: Regenerate.
4560 2014-04-15 Jan Hubicka <jh@suse.cz>
4562 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
4563 (maybe_record_node, likely_target_p): Use it.
4565 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4568 Revert following patch
4570 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
4573 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
4574 software floating point or no floating point registers, do not
4575 allow any type in the FPRs. Eliminate a test for SPE SIMD types
4576 in GPRs that occurs after we tested for GPRs that would never be
4579 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
4580 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
4581 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
4582 specifically allow DDmode, since that does not use the SPE SIMD
4585 2014-03-21 Mark Wielaard <mjw@redhat.com>
4587 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
4588 as unsigned or int depending on type and value used.
4590 2014-04-15 Richard Biener <rguenther@suse.de>
4592 PR rtl-optimization/56965
4593 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
4594 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
4596 * alias.c (true_dependence_1): Do not call
4597 nonoverlapping_component_refs_p.
4598 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
4599 nonoverlapping_component_refs_p.
4600 (indirect_refs_may_alias_p): Likewise.
4602 2014-04-15 Teresa Johnson <tejohnson@google.com>
4604 * cfg.c (dump_bb_info): Fix flags check.
4605 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
4607 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4609 PR rtl-optimization/60663
4610 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
4613 2014-04-15 Richard Biener <rguenther@suse.de>
4615 * lto-streamer.h (LTO_major_version): Bump to 4.
4617 2014-04-15 Richard Biener <rguenther@suse.de>
4619 * common.opt (lto_partition_model): New enum.
4620 (flto-partition=): Merge separate options with a single with argument,
4621 add -flto-partition=one support.
4622 * flag-types.h (enum lto_partition_model): Declare.
4623 * opts.c (finish_options): Remove duplicate -flto-partition=
4625 * lto-wrapper.c (run_gcc): Adjust.
4627 2014-04-15 Richard Biener <rguenther@suse.de>
4629 * alias.c (ncr_compar): New function.
4630 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
4632 2014-04-15 Richard Biener <rguenther@suse.de>
4634 * alias.c (record_component_aliases): Do not walk BINFOs.
4636 2014-04-15 Richard Biener <rguenther@suse.de>
4638 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
4639 Add struct function argument and adjust.
4640 (find_func_aliases_for_call): Likewise.
4641 (find_func_aliases): Likewise.
4642 (find_func_clobbers): Likewise.
4643 (intra_create_variable_infos): Likewise.
4644 (compute_points_to_sets): Likewise.
4645 (ipa_pta_execute): Adjust. Do not push/pop cfun.
4647 2014-04-15 Richard Biener <rguenther@suse.de>
4649 * tree.c (iterative_hash_expr): Use enum tree_code_class
4650 to store TREE_CODE_CLASS.
4651 (tree_block): Likewise.
4652 (tree_set_block): Likewise.
4653 * tree.h (fold_build_pointer_plus_loc): Use
4654 convert_to_ptrofftype_loc.
4656 2014-04-15 Jakub Jelinek <jakub@redhat.com>
4659 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
4662 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
4664 * cfgloop.h (struct loop): Move force_vectorize down.
4665 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
4666 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
4667 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
4668 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
4669 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
4670 * tree-core.h (enum annot_expr_kind): Add new kind values.
4671 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
4672 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
4674 * tree.def (ANNOTATE_EXPR): Tweak comment.
4676 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
4678 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
4681 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
4683 * tree.h (TYPE_IDENTIFIER): Declare.
4684 * tree.c (subrange_type_for_debug_p): Use it.
4685 * godump.c (go_format_type): Likewise.
4686 * dwarf2out.c (is_cxx_auto, modified_type_die,
4687 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
4688 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
4690 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
4693 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
4695 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
4697 * config/i386/i386.c (examine_argument): Return bool. Return true if
4698 parameter should be passed in memory.
4699 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
4700 (construct_container): Update calls to examine_argument.
4701 (function_arg_advance_64): Ditto.
4702 (return_in_memory_32): Merge with ix86_return_in_memory.
4703 (return_in_memory_64): Ditto.
4704 (return_in_memory_ms_64): Ditto.
4706 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
4708 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
4709 * coverage.c (coverage_compute_profile_id): Handle externally visible
4712 2014-04-14 Martin Jambor <mjambor@suse.cz>
4714 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
4715 DECL_DISREGARD_INLINE_LIMITS functions.
4717 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
4720 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
4722 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
4725 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
4726 optimize_insn_for_speed_p instead of
4727 optimize_function_for_speed_p.
4729 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
4731 * doc/invoke.texi (free): Document AArch64.
4733 2014-04-14 Richard Biener <rguenther@suse.de>
4735 PR tree-optimization/60042
4736 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
4737 (insert_into_preds_of_block): Do not prevent PHI insertion
4738 for REFERENCE exprs here ...
4739 (eliminate_dom_walker::before_dom_children): ... but prevent
4740 their use here under similar conditions when applied to the
4741 IL after PRE optimizations.
4743 2014-04-14 Richard Biener <rguenther@suse.de>
4745 * passes.def: Move early points-to after early SRA.
4747 2014-04-14 Richard Biener <rguenther@suse.de>
4749 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
4750 check for which sign-changes we allow when forwarding
4751 a converted value into a switch.
4753 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
4755 * stor-layout.c (place_field): Finalize non-constant offset for the
4758 2014-04-14 Richard Biener <rguenther@suse.de>
4760 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
4762 (expand_switch_using_bit_tests_p): Likewise.
4763 (process_switch): Compute and pass on speed_p based on the
4765 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
4766 optimize_bb_for_speed_p.
4768 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
4770 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
4771 * function.h (struct function): Rename has_force_vect_loops into
4772 has_force_vectorize_loops.
4773 * lto-streamer-in.c (input_cfg): Adjust for renaming.
4774 (input_struct_function_base): Likewise.
4775 * lto-streamer-out.c (output_cfg): Likewise.
4776 (output_struct_function_base): Likewise.
4777 * omp-low.c (expand_omp_simd): Likewise.
4778 * tree-cfg.c (move_sese_region_to_fn): Likewise.
4779 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
4780 (version_loop_for_if_conversion): Likewise.
4781 (tree_if_conversion): Likewise.
4782 (main_tree_if_conversion): Likewise.
4783 (gate_tree_if_conversion): Likewise.
4784 * tree-inline.c (copy_loops): Likewise.
4785 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
4786 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
4787 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
4788 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
4789 * tree-vectorizer.c (vectorize_loops): Likewise.
4790 * tree-vectorizer.h (unlimited_cost_model): Likewise.
4792 2014-04-14 Richard Biener <rguenther@suse.de>
4795 * lto-streamer-out.c (wrap_refs): New function.
4796 (lto_output): Wrap symbol references in global initializes in
4797 type-preserving MEM_REFs.
4799 2014-04-14 Christian Bruel <christian.bruel@st.com>
4801 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
4803 2014-04-14 Christian Bruel <christian.bruel@st.com>
4805 * config/sh/sh.md (setmemqi): New expand pattern.
4806 * config/sh/sh.h (CLEAR_RATIO): Define.
4807 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
4808 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
4810 2014-04-14 Richard Biener <rguenther@suse.de>
4813 * fold-const.c (negate_expr_p): Don't negate directional rounding
4815 (fold_negate_expr): Likewise.
4817 2014-04-14 Richard Biener <rguenther@suse.de>
4819 PR tree-optimization/59817
4820 PR tree-optimization/60453
4821 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
4822 recursion to catch all CHRECs in the scalar evolution and restrict
4823 the predicate for the remains appropriately.
4825 2014-04-12 Catherine Moore <clm@codesourcery.com>
4827 * config/mips/constraints.md: Add new register constraint "kb".
4828 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
4829 (*movhi_internal): Likewise.
4830 (*movqi_internal): Likewise.
4831 * config/mips/mips.h (M16_STORE_REGS): New register class.
4832 (REG_CLASS_NAMES): Add M16_STORE_REGS.
4833 (REG_CLASS_CONTENTS): Likewise.
4834 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
4836 2014-04-11 Tobias Burnus <burnus@net-b.de>
4839 * doc/invoke.texi (-Wformat-signedness): Document it.
4840 (Wformat=2): Mention that this enables -Wformat-signedness.
4842 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
4844 * common/config/epiphany/epiphany-common.c
4845 (epiphany_option_optimization_table): Enable section anchors by
4846 default at -O1 or higher.
4847 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
4848 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
4849 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
4850 carries no extra cost.
4851 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
4852 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
4853 * config/epiphany/predicates.md (memclob_operand): New predicate.
4854 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
4855 Use memclob_operand predicate and X constraint for operand 3.
4857 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
4859 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
4860 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
4863 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
4865 PR rtl-optimization/60651
4866 * mode-switching.c (optimize_mode_switching): Make sure to emit
4867 sets of a lower numbered entity before sets of a higher numbered
4868 entity to a mode of the same or lower priority.
4869 When creating a seginfo for a basic block that starts with a code
4870 label, move the insertion point past the code label.
4871 (new_seginfo): Document and enforce requirement that
4872 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
4873 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
4874 * doc/tm.texi: Regenerate.
4876 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
4879 * config/arc/arc.c (arc_save_restore): Fix assert typo.
4881 2013-04-11 Jakub Jelinek <jakub@redhat.com>
4883 * BASE-VER: Set to 4.10.0.
4885 2014-04-11 Tobias Burnus <burnus@net-b.de>
4888 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
4889 * doc/gcc.texi (Service): Update description in the @menu
4890 * doc/invoke.texi (Option Summary): Remove misplaced and
4893 2014-04-11 Steve Ellcey <sellcey@mips.com>
4894 Jakub Jelinek <jakub@redhat.com>
4897 * expr.c (convert_move): Use emit_store_flag_force instead of
4898 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
4901 2014-04-11 Richard Biener <rguenther@suse.de>
4904 * varasm.c (assemble_alias): Avoid endless error reporting
4905 recursion by setting TREE_ASM_WRITTEN.
4907 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4909 * config/s390/s390.md: Add a splitter for NOT rtx.
4911 2014-04-11 Jakub Jelinek <jakub@redhat.com>
4913 PR rtl-optimization/60663
4914 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
4916 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
4917 Jakub Jelinek <jakub@redhat.com>
4920 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
4921 flag from decl_node to node.
4923 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4926 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
4927 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
4928 ameliorating the cases where it can be.
4930 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
4933 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
4935 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
4936 (loadsync_<mode>): Change mode.
4937 (load_quadpti, store_quadpti): New.
4938 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
4939 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
4940 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
4942 2014-04-09 Cong Hou <congh@google.com>
4945 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
4948 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4950 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
4951 instead of vnor to exploit possible fusion opportunity in the
4953 (altivec_expand_vec_perm_const_le): Likewise.
4955 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
4957 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
4958 (loadsync_<mode>): Change mode.
4959 (load_quadpti, store_quadpti): New.
4960 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
4961 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
4963 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
4966 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
4967 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
4968 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
4970 2014-04-08 Richard Biener <rguenther@suse.de>
4973 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
4974 a 64bit widest int print double-int similar to on HWI64 hosts.
4976 2014-04-08 Richard Biener <rguenther@suse.de>
4978 PR tree-optimization/60785
4979 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
4980 default defs properly.
4982 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
4984 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
4985 (Weffc++): Likewise.
4987 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
4989 * ipa-devirt.c (maybe_record_node): When node is not recorded,
4990 set completep to false rather than true.
4992 2014-04-07 Douglas B Rupp <rupp@adacore.com>
4995 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
4996 ARM_TARGET2_DWARF_FORMAT.
4998 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
5001 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
5002 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
5005 2014-04-07 Richard Biener <rguenther@suse.de>
5007 PR tree-optimization/60766
5008 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
5009 (may_eliminate_iv): Convert cand_value_at result to desired type.
5011 2014-04-07 Jason Merrill <jason@redhat.com>
5014 * common.opt (-fno-gnu-unique): Add.
5015 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
5017 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5019 * haifa-sched.c: Fix outdated function reference and minor
5020 grammar errors in introductory comment.
5022 2014-04-07 Richard Biener <rguenther@suse.de>
5025 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
5027 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
5029 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
5032 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
5033 size accounting for thunks.
5034 (pa_asm_output_mi_thunk): Use final_start_function() and
5035 final_end_function() to output function start and end directives.
5037 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
5039 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
5040 device specific ISA/ feature information. Remove short_sp and
5041 errata_skip ds. Add avr_device_specific_features enum to have device
5043 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
5044 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
5045 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
5046 updated device specific info.
5047 * config/avr/avr-mcus.def: Merge device specific details to
5048 dev_attribute field.
5049 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
5051 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
5052 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
5053 assembler if RMW isa supported by current device.
5054 * config/avr/genmultilib.awk: Update as device info structure changed.
5055 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
5057 2014-04-04 Cong Hou <congh@google.com>
5059 PR tree-optimization/60656
5060 * tree-vect-stmts.c (supportable_widening_operation):
5061 Fix a bug that elements in a vector with vect_used_by_reduction
5062 property are incorrectly reordered when the operation on it is not
5063 consistant with the one in reduction operation.
5065 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
5067 PR rtl-optimization/60155
5068 * gcse.c (record_set_data): New function.
5069 (single_set_gcse): New function.
5070 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
5071 (hoist_code): Likewise.
5072 (get_pressure_class_and_nregs): Likewise.
5074 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
5076 * explow.c (probe_stack_range): Emit a final optimization blockage.
5078 2014-04-04 Anthony Green <green@moxielogic.com>
5080 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
5083 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
5086 * lto-cgraph.c (input_overwrite_node): Check that partitioning
5087 flags are set only during streaming.
5088 * ipa.c (process_references, walk_polymorphic_call_targets,
5089 symtab_remove_unreachable_nodes): Drop bodies of always inline
5090 after early inlining.
5091 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
5093 2014-04-04 Jakub Jelinek <jakub@redhat.com>
5094 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5097 * dwarf2out.c (const_ok_for_output_1): Reject expressions
5100 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5103 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
5105 (cortex_a53_fdivd): Likewise.
5107 2014-04-04 Martin Jambor <mjambor@suse.cz>
5110 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
5112 * cgraph.c (clone_of_p): Also return true if thunks match.
5113 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
5114 cgraph_function_or_thunk_node and an obsolete comment.
5115 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
5117 (build_function_decl_skip_args): Likewise.
5118 (set_new_clone_decl_and_node_flags): New function.
5119 (duplicate_thunk_for_node): Likewise.
5120 (redirect_edge_duplicating_thunks): Likewise.
5121 (cgraph_clone_node): New parameter args_to_skip, pass it to
5122 redirect_edge_duplicating_thunks which is called instead of
5123 cgraph_redirect_edge_callee.
5124 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
5125 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
5127 2014-04-04 Jeff Law <law@redhat.com>
5130 * config/arm/predicates.md (const_int_I_operand): New predicate.
5131 (const_int_M_operand): Similarly.
5132 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
5134 (insv_t2, extv_reg, extzv_t2): Likewise.
5135 (load_multiple_with_writeback): Similarly for const_int_I_operand.
5136 (pop_multiple_with_writeback_and_return): Likewise.
5137 (vfp_pop_multiple_with_writeback): Likewise
5139 2014-04-04 Richard Biener <rguenther@suse.de>
5142 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
5143 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
5145 * gimplify.h (gimple_add_tmp_var_fn): Declare.
5146 * gimplify.c (gimple_add_tmp_var_fn): New function.
5147 * gimple-expr.h (create_tmp_reg_fn): Declare.
5148 * gimple-expr.c (create_tmp_reg_fn): New function.
5149 * gimple-low.c (record_vars_into): Don't change cfun.
5150 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
5151 code generation without cfun.
5153 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
5156 * Makefile.in (install-driver): Fix shell scripting.
5158 2014-04-03 Cong Hou <congh@google.com>
5160 PR tree-optimization/60505
5161 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
5162 threshold of number of iterations below which no vectorization
5164 * tree-vect-loop.c (new_loop_vec_info):
5165 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
5166 * tree-vect-loop.c (vect_analyze_loop_operations):
5167 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
5168 * tree-vect-loop.c (vect_transform_loop):
5169 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
5170 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
5171 of iterations of the loop and see if we should build the epilogue.
5173 2014-04-03 Richard Biener <rguenther@suse.de>
5175 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
5176 (streamer_tree_cache_create): Adjust.
5177 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
5178 to allow optional nodes array.
5179 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
5180 (streamer_tree_cache_append): Likewise.
5181 (streamer_tree_cache_create): Create nodes array optionally
5182 as specified by parameter.
5183 * lto-streamer-out.c (create_output_block): Avoid maintaining
5184 the node array in the writer cache.
5185 (DFS_write_tree): Remove assertion.
5186 (produce_asm_for_decls): Free the out decl state hash table early.
5187 * lto-streamer-in.c (lto_data_in_create): Adjust for
5188 streamer_tree_cache_create prototype change.
5190 2014-04-03 Richard Biener <rguenther@suse.de>
5192 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
5193 set TREE_CHAIN to NULL_TREE.
5195 2014-04-03 Richard Biener <rguenther@suse.de>
5197 PR tree-optimization/60740
5198 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
5199 over all GIMPLE_COND operands.
5201 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
5203 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
5204 (Weffc++): Remove Scott's numbering, merge lists and reference
5207 2014-04-03 Nick Clifton <nickc@redhat.com>
5209 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
5212 2014-04-03 Martin Jambor <mjambor@suse.cz>
5214 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
5215 mention gcc_unreachable before failing.
5216 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
5219 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
5222 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
5223 inconsistent code and instead mark the context inconsistent.
5224 (possible_polymorphic_call_targets): For inconsistent contexts
5225 return empty complete list.
5227 2014-04-02 Anthony Green <green@moxielogic.com>
5229 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
5230 (extendqisi2, extendhisi2): Define.
5231 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
5232 (WCHAR_TYPE): Change to unsigned int.
5234 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5236 PR tree-optimization/60733
5237 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
5238 insertion point for PHI candidates to be the end of the feeding
5239 block for the PHI argument.
5241 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
5243 PR rtl-optimization/60650
5244 * lra-constraints.c (process_alt_operands): Decrease reject for
5245 earlyclobber matching.
5247 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5249 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
5251 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5253 * config/spu/spu.c (pad_bb): Do not crash when the last
5254 insn is CODE_FOR_blockage.
5256 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5258 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
5259 lies outside the target mode.
5261 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
5264 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
5265 software floating point or no floating point registers, do not
5266 allow any type in the FPRs. Eliminate a test for SPE SIMD types
5267 in GPRs that occurs after we tested for GPRs that would never be
5270 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
5271 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
5272 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
5273 specifically allow DDmode, since that does not use the SPE SIMD
5276 2014-04-02 Richard Biener <rguenther@suse.de>
5279 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
5280 MODE_INTs. Properly use negv_optab.
5281 (expand_abs): Likewise.
5283 2014-04-02 Richard Biener <rguenther@suse.de>
5286 * Makefile.in (install-driver): Guard extra installs with special
5289 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
5291 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5294 2014-04-01 Richard Henderson <rth@redhat.com>
5297 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
5298 alternative enabled before register allocation.
5300 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
5302 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
5303 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
5305 (nios2_large_got_address): Remove unneeded 'sym' parameter.
5306 (nios2_got_address): Update nios2_large_got_address call site.
5307 (nios2_delegitimize_address): New function.
5308 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
5309 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
5310 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
5312 2014-04-01 Martin Husemann <martin@duskware.de>
5314 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
5317 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
5319 PR rtl-optimization/60604
5320 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
5321 check from register_operand.
5322 (register_operand): Redefine in terms of general_operand.
5323 (nonmemory_operand): Use register_operand for the non-constant cases.
5325 2014-04-01 Richard Biener <rguenther@suse.de>
5327 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
5329 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
5331 * doc/invoke.texi (mapp-regs): Clarify.
5333 2014-03-31 Ulrich Drepper <drepper@gmail.com>
5335 * config/i386/avx512fintrin.h (__v32hi): Define type.
5336 (__v64qi): Likewise.
5337 (_mm512_set1_epi8): Define.
5338 (_mm512_set1_epi16): Define.
5339 (_mm512_set4_epi32): Define.
5340 (_mm512_set4_epi64): Define.
5341 (_mm512_set4_pd): Define.
5342 (_mm512_set4_ps): Define.
5343 (_mm512_setr4_epi64): Define.
5344 (_mm512_setr4_epi32): Define.
5345 (_mm512_setr4_pd): Define.
5346 (_mm512_setr4_ps): Define.
5347 (_mm512_setzero_epi32): Define.
5349 2014-03-31 Martin Jambor <mjambor@suse.cz>
5352 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
5353 callsite_arguments_match_p. Updated all callers. Also check types of
5354 corresponding formal parameters and actual arguments.
5355 (not_all_callers_have_enough_arguments_p) Renamed to
5356 some_callers_have_mismatched_arguments_p.
5358 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
5360 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
5362 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
5365 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
5368 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
5370 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
5371 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
5373 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
5374 Use FMAMODE_NOVF512 mode iterator.
5375 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
5376 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
5377 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
5378 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
5380 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
5381 Use VF_128_256 mode iterator.
5382 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
5385 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
5387 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
5388 static chain if needed.
5390 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
5393 * lra-constraints.c (index_part_to_reg): New.
5394 (process_address): Use it.
5396 2014-03-27 Jeff Law <law@redhat.com>
5397 Jakub Jelinek <jakub@redhat.com>
5400 * expr.c (do_tablejump): Use simplify_gen_binary rather than
5401 gen_rtx_{PLUS,MULT} to build up the address expression.
5403 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
5404 creating non-canonical RTL.
5406 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
5409 * ipa-inline.c (want_inline_small_function_p): Short circuit large
5410 functions; reorganize to make cheap checks first.
5411 (inline_small_functions): Do not estimate growth when dumping;
5413 * ipa-inline.h (inline_summary): Add min_size.
5414 (growth_likely_positive): New function.
5415 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
5416 (set_cond_stmt_execution_predicate): Cleanup.
5417 (estimate_edge_size_and_time): Compute min_size.
5418 (estimate_calls_size_and_time): Likewise.
5419 (estimate_node_size_and_time): Likewise.
5420 (inline_update_overall_summary): Update min_size.
5421 (do_estimate_edge_time): Likewise.
5422 (do_estimate_edge_size): Update.
5423 (do_estimate_edge_hints): Update.
5424 (growth_likely_positive): New function.
5426 2014-03-28 Jakub Jelinek <jakub@redhat.com>
5429 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
5430 also if addr has VOIDmode.
5432 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5434 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
5435 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
5437 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
5438 instructions as well as AdvancedSIMD loads.
5440 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5442 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
5443 Use crypto_aese type.
5444 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
5445 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
5446 crypto_aese, crypto_aesmc. Move to types.md.
5447 * config/arm/types.md (crypto_aes): Split into crypto_aese,
5449 * config/arm/iterators.md (crypto_type): Likewise.
5451 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
5453 * cgraph.c: Include expr.h and tree-dfa.h.
5454 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
5457 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
5460 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
5461 regs from checking multi-reg pseudos.
5463 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5465 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
5467 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5469 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
5470 if it would clobber the stack pointer, even temporarily.
5472 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
5474 * mode-switching.c: Make small adjustments to the top comment.
5476 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
5478 * config/rs6000/constraints.md (wD constraint): New constraint to
5479 match the constant integer to get the top DImode/DFmode out of a
5480 vector in a VSX register.
5482 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
5483 match the constant integer to get the top DImode/DFmode out of a
5484 vector in a VSX register.
5486 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
5489 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
5492 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
5493 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
5495 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
5496 Optimize vec_extract of 64-bit values, where the value being
5497 extracted is in the top word, where we can use scalar
5498 instructions. Add direct move and store support. Combine the big
5499 endian/little endian vector select load support into a single insn.
5500 (vsx_extract_<mode>_internal1): Likewise.
5501 (vsx_extract_<mode>_internal2): Likewise.
5502 (vsx_extract_<mode>_load): Likewise.
5503 (vsx_extract_<mode>_store): Likewise.
5504 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
5505 combined into vsx_extract_<mode>_load.
5506 (vsx_extract_<mode>_one_le): Likewise.
5508 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
5509 define the top 64-bit vector element.
5511 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
5514 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5515 Document vec_vbpermq builtin.
5518 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
5519 enable use of xxsldwi and xxpermdi builtin functions.
5520 (vec_xxpermdi): Likewise.
5522 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5523 Document use of vec_xxsldwi and vec_xxpermdi builtins.
5525 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
5527 PR rtl-optimization/60650
5528 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
5530 (find_spills_for): New.
5531 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
5532 Spill all pseudos on the second iteration.
5534 2014-03-27 Marek Polacek <polacek@redhat.com>
5537 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
5540 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5542 * config/s390/s390.c (s390_can_use_return_insn): Check for
5543 call-saved FPRs on 31 bit.
5545 2014-03-27 Jakub Jelinek <jakub@redhat.com>
5548 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
5549 if they need regimplification, just drop them instead of
5550 calling gimple_regimplify_operands on them.
5552 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
5555 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
5556 (aarch64_frame_pointer_required): Adjust logic.
5557 (aarch64_can_eliminate): Adjust logic.
5558 (aarch64_override_options_after_change): Adjust logic.
5560 2014-03-27 Dehao Chen <dehao@google.com>
5562 * ipa-inline.c (early_inliner): Update node's inline info.
5564 2014-03-26 Dehao Chen <dehao@google.com>
5566 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
5567 compiler inserted conditional jumps for NAN float check.
5569 2014-03-26 Jakub Jelinek <jakub@redhat.com>
5571 * ubsan.h (ubsan_create_data): Change second argument's type
5572 to const location_t *.
5573 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
5575 (ubsan_create_data): Change second argument to const location_t *PLOC.
5576 Create Loc field whenever PLOC is non-NULL.
5577 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
5578 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
5582 * real.c (real_to_integer2): Change type of low to UHWI.
5584 2014-03-26 Tobias Burnus <burnus@net-b.de>
5586 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
5587 (CILK_SELF_SPECS): New define.
5588 (driver_self_specs): Use it.
5590 2014-03-26 Richard Biener <rguenther@suse.de>
5592 * tree-pretty-print.c (percent_K_format): Implement special
5593 case for LTO and its stripped down BLOCK tree.
5595 2014-03-26 Jakub Jelinek <jakub@redhat.com>
5598 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
5600 * tree-vrp.c (simplify_internal_call_using_ranges): If only
5601 one range is range_int_cst_p, but not both, at least optimize
5602 addition/subtraction of 0 and multiplication by 0 or 1.
5603 * gimple-fold.c (gimple_fold_call): Fold
5604 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
5605 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
5606 INTEGER_CSTs, try to fold at least x * 0 and y - y.
5608 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
5610 PR rtl-optimization/60452
5611 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
5612 <case REG>: Return 1 for invalid offsets from the frame pointer.
5614 2014-03-26 Marek Polacek <polacek@redhat.com>
5617 * doc/extend.texi (C Extensions): Mention variable-length arrays in
5620 2014-03-26 Marek Polacek <polacek@redhat.com>
5623 * doc/extend.texi (Designated Inits): Describe what happens to omitted
5626 2014-03-26 Marek Polacek <polacek@redhat.com>
5629 * ira-color.c (update_conflict_hard_regno_costs): Perform the
5630 multiplication in unsigned type.
5632 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
5634 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
5636 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
5638 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
5640 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
5643 * cif-code.def (UNREACHABLE) New code.
5644 * ipa-inline.c (inline_small_functions): Skip edges to
5645 __builtlin_unreachable.
5646 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
5647 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
5648 predicate to __bulitin_unreachable.
5649 (set_cond_stmt_execution_predicate): Fix issue when
5650 invert_tree_comparison returns ERROR_MARK.
5651 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
5652 propagate to inline clones.
5653 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
5655 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
5656 * cgraphclones.c (cgraph_clone_node): If call destination is already
5657 ureachable, do not redirect it back.
5658 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
5661 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
5663 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
5664 Do not modify inline clones.
5666 2014-03-25 Jakub Jelinek <jakub@redhat.com>
5668 * config/i386/i386.md (general_sext_operand): New mode attr.
5669 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
5670 don't generate (sign_extend (const_int)).
5671 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
5672 operands[2]. Use We constraint instead of <i> and
5673 <general_sext_operand> predicate instead of <general_operand>.
5674 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
5675 * config/i386/constraints.md (We): New constraint.
5676 * config/i386/predicates.md (x86_64_sext_operand,
5677 sext_operand): New predicates.
5679 2014-03-25 Martin Jambor <mjambor@suse.cz>
5682 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
5683 inconsistent devirtualizations to __builtin_unreachable.
5685 2014-03-25 Marek Polacek <polacek@redhat.com>
5688 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
5690 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
5692 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
5693 order of elements for big-endian.
5695 2014-03-25 Richard Biener <rguenther@suse.de>
5698 * gimplify-me.c (gimple_regimplify_operands): Update the
5701 2014-03-25 Martin Jambor <mjambor@suse.cz>
5704 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
5705 (lto_output_varpool_node): Likewise.
5706 (input_overwrite_node): Likewise.
5707 (input_varpool_node): Likewise.
5709 2014-03-25 Richard Biener <rguenther@suse.de>
5711 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
5712 (run_gcc): Likewise.
5714 2014-03-25 Jakub Jelinek <jakub@redhat.com>
5716 * combine.c (simplify_compare_const): Add MODE argument.
5717 Handle mode_width 0 as very large mode_width.
5718 (try_combine, simplify_comparison): Adjust callers.
5720 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
5721 type to avoid signed integer overflow.
5722 * explow.c (plus_constant): Likewise.
5724 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
5726 * doc/generic.texi: Correct typos.
5728 2014-03-24 Tobias Burnus <burnus@net-b.de>
5730 * doc/invoke.texi (-flto): Expand section about
5731 using static libraries with LTO.
5733 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5735 PR rtl-optimization/60501
5736 * optabs.def (addptr3_optab): New optab.
5737 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
5738 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
5739 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
5741 * lra.c (emit_add3_insn): Use the addptr pattern if available.
5743 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
5745 2014-03-24 Ulrich Drepper <drepper@gmail.com>
5747 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
5750 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
5751 (_mm256_undefined_ps): Define.
5752 (_mm256_undefined_pd): Define.
5753 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
5754 (_mm_undefined_pd): Define.
5755 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
5756 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
5757 (_mm512_undefined_ps): Define.
5758 (_mm512_undefined_pd): Define.
5759 Use _mm*_undefined_*.
5760 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
5762 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
5764 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
5765 (lshr_simd): DI mode added.
5766 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
5767 (aarch64_ushr_simddi): Likewise.
5768 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
5769 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
5770 (vshrd_n_u64): Likewise.
5772 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5774 * Makefile.in (s-macro_list): Depend on cc1.
5776 2014-03-23 Teresa Johnson <tejohnson@google.com>
5778 * ipa-utils.c (ipa_print_order): Use specified dump file.
5780 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
5782 PR rtl-optimization/60601
5783 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
5785 * gcc.c (eval_spec_function): Initialize save_growing_value.
5787 2014-03-22 Jakub Jelinek <jakub@redhat.com>
5790 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
5791 code == MINUS_EXPR, never swap op0 with op1.
5793 * toplev.c (init_local_tick): Avoid signed integer multiplication
5795 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
5796 shift by first operand's bitsize.
5798 2014-03-21 Jakub Jelinek <jakub@redhat.com>
5801 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
5803 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
5804 TARGET_ISA_64BIT_P(x).
5806 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5808 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
5809 pattern for vector nor instead of subtract from splat(-1).
5810 (altivec_expand_vec_perm_const_le): Likewise.
5812 2014-03-21 Richard Henderson <rth@twiddle.net>
5815 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
5816 related insns after epilogue_completed.
5818 2014-03-21 Martin Jambor <mjambor@suse.cz>
5821 * cgraph.h (symtab_node): New flag body_removed.
5822 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
5823 when removing bodies.
5824 * symtab.c (dump_symtab_base): Dump body_removed flag.
5825 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
5826 had their bodies removed.
5828 2014-03-21 Martin Jambor <mjambor@suse.cz>
5831 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
5834 2014-03-21 Richard Biener <rguenther@suse.de>
5836 PR tree-optimization/60577
5837 * tree-core.h (struct tree_base): Document nothrow_flag use
5838 in VAR_DECL_NONALIASED.
5839 * tree.h (VAR_DECL_NONALIASED): New.
5840 (may_be_aliased): Adjust.
5841 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
5843 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
5845 * expr.c (expand_expr_real_1): Remove outdated comment.
5847 2014-03-20 Jakub Jelinek <jakub@redhat.com>
5850 * ira.c (adjust_cleared_regs): Call copy_rtx on
5851 *reg_equiv[REGNO (loc)].src_p before passing it to
5852 simplify_replace_fn_rtx.
5855 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
5856 into CONST, put pic register as first operand of PLUS. Use
5857 gen_const_mem for both 32-bit and 64-bit PIC got loads.
5859 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5861 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
5863 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
5865 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
5866 around for store forwarding issue in the FPU on the UT699.
5867 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
5868 loads and operations if -mfix-ut699 is specified.
5869 (divtf3_hq): Tweak attribute.
5870 (sqrttf2_hq): Likewise.
5872 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
5874 * calls.c (store_one_arg): Remove incorrect const qualification on the
5875 type of the temporary.
5876 * cfgexpand.c (expand_return): Likewise.
5877 * expr.c (expand_constructor): Likewise.
5878 (expand_expr_real_1): Likewise.
5880 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
5882 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
5885 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
5888 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
5890 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
5892 * config/arm/aarch-common-protos.h
5893 (alu_cost_table): Fix spelling of "extend".
5894 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
5896 2014-03-19 Richard Biener <rguenther@suse.de>
5899 * tree-core.h (tree_type_common): Re-order pointer members
5900 to reduce recursion depth during GC walks.
5902 2014-03-19 Marek Polacek <polacek@redhat.com>
5905 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
5906 before accessing it.
5908 2014-03-19 Richard Biener <rguenther@suse.de>
5911 * lto-streamer-in.c (input_function): In WPA stage do not drop
5914 2014-03-19 Jakub Jelinek <jakub@redhat.com>
5916 PR tree-optimization/60559
5917 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
5918 with build_zero_cst assignment.
5920 2014-03-18 Kai Tietz <ktietz@redhat.com>
5922 PR rtl-optimization/56356
5923 * sdbout.c (sdbout_parms): Verify that parms'
5924 incoming argument is valid.
5925 (sdbout_reg_parms): Likewise.
5927 2014-03-18 Richard Henderson <rth@redhat.com>
5930 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
5931 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
5932 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
5934 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
5936 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
5937 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
5938 Italicize plugin event names in description. Explain that
5939 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
5940 Remind that no GCC functions should be called after PLUGIN_FINISH.
5941 Explain what pragmas with expansion are.
5943 2014-03-18 Martin Liska <mliska@suse.cz>
5945 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
5946 gimple call statement is update.
5947 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
5948 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
5950 2014-03-18 Jakub Jelinek <jakub@redhat.com>
5953 * ubsan.c (ubsan_instrument_unreachable): Call
5954 initialize_sanitizer_builtins.
5955 (ubsan_pass): Likewise.
5958 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
5959 varpool_finalize_decl instead of rest_of_decl_compilation.
5961 2014-03-18 Richard Biener <rguenther@suse.de>
5963 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
5964 by using bitmap_and_compl instead of bitmap_and_compl_into.
5965 (df_rd_transfer_function): Likewise.
5967 2014-03-18 Richard Biener <rguenther@suse.de>
5969 * doc/lto.texi (fresolution): Fix typo.
5971 2014-03-18 Richard Biener <rguenther@suse.de>
5973 * doc/invoke.texi (flto): Update for changes in 4.9.
5975 2014-03-18 Richard Biener <rguenther@suse.de>
5977 * doc/loop.texi: Remove section on the removed lambda framework.
5978 Update loop docs with recent changes in preserving loop structure.
5980 2014-03-18 Richard Biener <rguenther@suse.de>
5982 * doc/lto.texi (-fresolution): Document.
5984 2014-03-18 Richard Biener <rguenther@suse.de>
5986 * doc/contrib.texi: Adjust my name.
5988 2014-03-18 Jakub Jelinek <jakub@redhat.com>
5991 * internal-fn.c: Include diagnostic-core.h.
5992 (expand_BUILTIN_EXPECT): New function.
5993 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
5994 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
5995 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
5996 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
5998 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
5999 Revert 3 argument __builtin_expect code.
6000 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
6001 * gimple-fold.c (gimple_fold_call): Likewise.
6002 * tree.h (fold_builtin_expect): New prototype.
6003 * builtins.c (build_builtin_expect_predicate): Add predictor
6004 argument, if non-NULL, create 3 argument __builtin_expect.
6005 (fold_builtin_expect): No longer static. Add ARG2 argument,
6006 pass it through to build_builtin_expect_predicate.
6007 (fold_builtin_2): Adjust caller.
6008 (fold_builtin_3): Handle BUILT_IN_EXPECT.
6009 * internal-fn.def (BUILTIN_EXPECT): New.
6011 2014-03-18 Tobias Burnus <burnus@net-b.de>
6014 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
6015 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
6016 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
6018 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
6021 * predict.c (combine_predictions_for_bb): Fix up formatting.
6022 (expr_expected_value_1, expr_expected_value): Add predictor argument,
6023 fill what it points to if non-NULL.
6024 (tree_predict_by_opcode): Adjust caller, use the predictor.
6025 * predict.def (PRED_COMPARE_AND_SWAP): Add.
6027 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
6029 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
6030 proper constant for the store mode.
6032 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
6034 * symtab.c (change_decl_assembler_name): Fix transparent alias
6037 2014-03-16 Renlin Li <Renlin.Li@arm.com>
6039 * config/aarch64/aarch64.c: Correct the comments about the
6040 aarch64 stack layout.
6042 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
6044 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
6045 check for GF_OMP_FOR_KIND_FOR.
6047 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
6049 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
6050 ymm and zmm register names.
6052 2014-03-17 Jakub Jelinek <jakub@redhat.com>
6055 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
6056 note creation for the 2010-08-31 changes.
6058 2014-03-17 Marek Polacek <polacek@redhat.com>
6061 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
6062 as -fno-tree-loop-vectorize.
6063 (expand_omp_simd): Likewise.
6065 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
6067 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
6068 (eligible_for_call_delay): New prototype.
6069 * config/sparc/sparc.c (tls_call_delay): Rename into...
6070 (eligible_for_call_delay): ...this. Return false if the instruction
6071 cannot be put in the delay slot of a branch.
6072 (eligible_for_restore_insn): Simplify.
6073 (eligible_for_return_delay): Return false if the instruction cannot be
6074 put in the delay slot of a branch and simplify.
6075 (eligible_for_sibcall_delay): Return false if the instruction cannot be
6076 put in the delay slot of a branch.
6077 * config/sparc/sparc.md (fix_ut699): New attribute.
6078 (tls_call_delay): Delete.
6079 (in_call_delay): Reimplement.
6080 (eligible_for_sibcall_delay): Rename into...
6081 (in_sibcall_delay): ...this.
6082 (eligible_for_return_delay): Rename into...
6083 (in_return_delay): ...this.
6084 (in_branch_delay): Reimplement.
6085 (in_uncond_branch_delay): Delete.
6086 (in_annul_branch_delay): Delete.
6088 2014-03-14 Richard Henderson <rth@redhat.com>
6091 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
6092 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
6093 (*floathi<X87MODEF>2_i387_with_temp): Remove.
6094 (floathi splitters): Remove.
6095 (float<SWI48x>xf2): New pattern.
6096 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
6097 code that tried to handle DImode for 32-bit, but which was excluded
6098 by the pattern's condition. Drop allocation of stack temporary.
6099 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
6100 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
6101 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
6102 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
6103 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
6104 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
6105 (*float<SWI48><MODEF>2_sse_interunit): Remove.
6106 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
6107 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
6108 (*float<SWI48x><X87MODEF>2_i387): Remove.
6109 (all float _with_temp splitters): Remove.
6110 (*float<SWI48x><MODEF>2_i387): New pattern.
6111 (*float<SWI48><MODEF>2_sse): New pattern.
6112 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
6113 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
6115 2014-03-14 Jakub Jelinek <jakub@redhat.com>
6116 Marek Polacek <polacek@redhat.com>
6119 * common.opt (dump_base_name_prefixed): New Variable.
6120 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
6121 if x_dump_base_name_prefixed is already set, set it at the end.
6123 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
6125 PR rtl-optimization/60508
6126 * lra-constraints.c (get_reload_reg): Add new parameter
6128 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
6129 Pass the new parameter values.
6131 2014-03-14 Richard Biener <rguenther@suse.de>
6133 * common.opt: Revert unintented changes from r205065.
6136 2014-03-14 Richard Biener <rguenther@suse.de>
6139 * cfghooks.c (split_block): Properly adjust all loops the
6140 block was a latch of.
6142 2014-03-14 Martin Jambor <mjambor@suse.cz>
6145 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
6148 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
6151 * config/avr/avr.c (avr_set_current_function): Pass function name
6152 through default_strip_name_encoding before sanity checking instead
6153 of skipping the first char of the assembler name.
6155 2014-03-13 Richard Henderson <rth@redhat.com>
6158 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
6159 (ix86_force_to_memory, ix86_free_from_memory): Remove.
6160 * config/i386/i386-protos.h: Likewise.
6161 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
6162 in the expander instead of a splitter.
6163 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
6164 any possibility of requiring a memory.
6165 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
6166 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
6167 (fp branch splitters): Update for ix86_split_fp_branch.
6168 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
6169 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
6170 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
6171 (*fop_<MODEF>_2_i387): Remove f/r alternative.
6172 (*fop_<MODEF>_3_i387): Likewise.
6173 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
6174 (splitters for the fop_* register patterns): Remove.
6175 (fscalexf4_i387): Rename from *fscalexf4_i387.
6176 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
6178 2014-03-13 Jakub Jelinek <jakub@redhat.com>
6180 PR tree-optimization/59779
6181 * tree-dfa.c (get_ref_base_and_extent): Use double_int
6182 type for bitsize and maxsize instead of HOST_WIDE_INT.
6184 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
6186 PR rtl-optimization/57320
6187 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
6188 the CFG after thread_prologue_and_epilogue_insns.
6190 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
6192 PR rtl-optimization/57189
6193 * lra-constraints.c (process_alt_operands): Disfavor spilling
6196 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
6198 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
6200 2014-03-13 Jakub Jelinek <jakub@redhat.com>
6202 PR tree-optimization/59025
6204 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
6205 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
6207 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
6210 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
6211 calls of avr_out_plus_1.
6213 2014-03-13 Bin Cheng <bin.cheng@arm.com>
6215 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
6216 BB's single pred and update the father loop's latch info later.
6218 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
6220 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
6224 (VEC_base): Likewise.
6225 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
6226 registers, we need to swap double words in little endian mode.
6228 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
6229 to be a container mode for 128-bit integer operations added in ISA
6230 2.07. Unlike TImode and PTImode, the preferred register set is
6231 the Altivec/VMX registers for the 128-bit operations.
6233 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
6235 (rs6000_split_128bit_ok_p): Likewise.
6237 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
6238 macros for creating ISA 2.07 normal and overloaded builtin
6239 functions with 3 arguments.
6240 (BU_P8V_OVERLOAD_3): Likewise.
6241 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
6242 for use as overloaded functions.
6243 (VPERM_1TI_UNS): Likewise.
6244 (VSEL_1TI): Likewise.
6245 (VSEL_1TI_UNS): Likewise.
6246 (ST_INTERNAL_1ti): Likewise.
6247 (LD_INTERNAL_1ti): Likewise.
6248 (XXSEL_1TI): Likewise.
6249 (XXSEL_1TI_UNS): Likewise.
6250 (VPERM_1TI): Likewise.
6251 (VPERM_1TI_UNS): Likewise.
6252 (XXPERMDI_1TI): Likewise.
6253 (SET_1TI): Likewise.
6254 (LXVD2X_V1TI): Likewise.
6255 (STXVD2X_V1TI): Likewise.
6256 (VEC_INIT_V1TI): Likewise.
6257 (VEC_SET_V1TI): Likewise.
6258 (VEC_EXT_V1TI): Likewise.
6259 (EQV_V1TI): Likewise.
6260 (NAND_V1TI): Likewise.
6261 (ORC_V1TI): Likewise.
6262 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
6263 added in ISA 2.07. Add both normal 'altivec' builtins, and the
6265 (VADDUQM): Likewise.
6266 (VSUBCUQ): Likewise.
6267 (VADDEUQM): Likewise.
6268 (VADDECUQ): Likewise.
6269 (VSUBEUQM): Likewise.
6270 (VSUBECUQ): Likewise.
6272 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
6273 __int128_t and __uint128_t types.
6274 (__uint128_type): Likewise.
6275 (altivec_categorize_keyword): Add support for vector __int128_t,
6276 vector __uint128_t, vector __int128, and vector unsigned __int128
6277 as a container type for TImode operations that need to be done in
6278 VSX/Altivec registers.
6279 (rs6000_macro_to_expand): Likewise.
6280 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
6281 to support 128-bit integer instructions vaddcuq, vadduqm,
6282 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
6283 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
6285 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
6286 for V1TImode, and set up preferences to use VSX/Altivec registers.
6287 Setup VSX reload handlers.
6288 (rs6000_debug_reg_global): Likewise.
6289 (rs6000_init_hard_regno_mode_ok): Likewise.
6290 (rs6000_preferred_simd_mode): Likewise.
6291 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
6292 (easy_altivec_constant): Likewise.
6293 (output_vec_const_move): Likewise.
6294 (rs6000_expand_vector_set): Convert V1TImode set and extract to
6296 (rs6000_expand_vector_extract): Likewise.
6297 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
6299 (rs6000_const_vec): Add support for V1TImode.
6300 (rs6000_emit_le_vsx_load): Swap double words when loading or
6301 storing TImode/V1TImode.
6302 (rs6000_emit_le_vsx_store): Likewise.
6303 (rs6000_emit_le_vsx_move): Likewise.
6304 (rs6000_emit_move): Add support for V1TImode.
6305 (altivec_expand_ld_builtin): Likewise.
6306 (altivec_expand_st_builtin): Likewise.
6307 (altivec_expand_vec_init_builtin): Likewise.
6308 (altivec_expand_builtin): Likewise.
6309 (rs6000_init_builtins): Add support for V1TImode type. Add
6310 support for ISA 2.07 128-bit integer builtins. Define type names
6311 for the VSX/Altivec vector types.
6312 (altivec_init_builtins): Add support for overloaded vector
6313 functions with V1TImode type.
6314 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
6315 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
6317 (rs6000_split_128bit_ok_p): Likewise.
6318 (rs6000_handle_altivec_attribute): Create V1TImode from vector
6319 __int128_t and vector __uint128_t.
6321 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
6322 and mode attributes.
6329 (VS_scalar): Likewise.
6330 (VS_double): Likewise.
6331 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
6333 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
6334 we support the ISA 2.07 128-bit integer arithmetic instructions.
6335 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
6336 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
6337 and TImode types for use with the builtin functions.
6338 (V1TI_type_node): Likewise.
6339 (unsigned_V1TI_type_node): Likewise.
6340 (intTI_type_internal_node): Likewise.
6341 (uintTI_type_internal_node): Likewise.
6343 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
6344 128-bit builtin functions.
6345 (UNSPEC_VADDEUQM): Likewise.
6346 (UNSPEC_VADDECUQ): Likewise.
6347 (UNSPEC_VSUBCUQ): Likewise.
6348 (UNSPEC_VSUBEUQM): Likewise.
6349 (UNSPEC_VSUBECUQ): Likewise.
6350 (VM): Add V1TImode to vector mode iterators.
6352 (VI_unit): Likewise.
6353 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
6354 (altivec_vaddcuq): Likewise.
6355 (altivec_vsubuqm): Likewise.
6356 (altivec_vsubcuq): Likewise.
6357 (altivec_vaddeuqm): Likewise.
6358 (altivec_vaddecuq): Likewise.
6359 (altivec_vsubeuqm): Likewise.
6360 (altivec_vsubecuq): Likewise.
6362 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
6364 (BOOL_128): Likewise.
6365 (BOOL_REGS_OUTPUT): Likewise.
6366 (BOOL_REGS_OP1): Likewise.
6367 (BOOL_REGS_OP2): Likewise.
6368 (BOOL_REGS_UNARY): Likewise.
6369 (BOOL_REGS_AND_CR0): Likewise.
6371 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
6372 128-bit integer builtin support.
6373 (vec_vadduqm): Likewise.
6374 (vec_vaddecuq): Likewise.
6375 (vec_vaddeuqm): Likewise.
6376 (vec_vsubecuq): Likewise.
6377 (vec_vsubeuqm): Likewise.
6378 (vec_vsubcuq): Likewise.
6379 (vec_vsubuqm): Likewise.
6381 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
6382 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
6383 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
6384 128-bit integer add/subtract to ISA 2.07.
6386 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
6388 * config/arc/arc.c (arc_predicate_delay_insns):
6389 Fix third argument passed to conditionalize_nonjump.
6391 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
6393 * config/aarch64/aarch64-builtins.c
6394 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
6395 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
6396 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
6397 instead of __builtin_lfloor.
6398 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
6400 2014-03-12 Jakub Jelinek <jakub@redhat.com>
6402 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
6403 (tree_ssa_ifcombine_bb_1): New function.
6404 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
6405 is an empty forwarder block to then_bb or vice versa and then_bb
6406 and else_bb are effectively swapped.
6408 2014-03-12 Christian Bruel <christian.bruel@st.com>
6411 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
6412 REG_CFA_DEF_CFA note.
6413 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
6414 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
6416 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
6418 PR tree-optimization/60454
6419 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
6421 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6423 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
6424 Do not define target_cpu_default2 to generic.
6425 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
6426 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
6427 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
6429 2014-03-12 Jakub Jelinek <jakub@redhat.com>
6430 Marc Glisse <marc.glisse@inria.fr>
6432 PR tree-optimization/60502
6433 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
6434 instead of build_low_bits_mask.
6436 2014-03-12 Jakub Jelinek <jakub@redhat.com>
6439 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
6440 if there are multiple uses, but op doesn't live on E edge.
6441 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
6442 clobber stmts before __builtin_unreachable.
6444 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
6446 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
6447 hard_frame_pointer_rtx.
6448 * cse.c (cse_insn): Remove volatile check.
6449 * cselib.c (cselib_process_insn): Likewise.
6450 * dse.c (scan_insn): Likewise.
6452 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
6454 * config/arc/arc.c (conditionalize_nonjump): New function,
6456 (arc_ifcvt): ... this.
6457 (arc_predicate_delay_insns): Use it.
6459 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
6461 * config/arc/predicates.md (extend_operand): During/after reload,
6462 allow const_int_operand.
6463 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
6464 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
6465 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
6467 (umulsi3_highpart_i): Likewise.
6469 2014-03-11 Richard Biener <rguenther@suse.de>
6471 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
6472 Add asserts to guard possible wrong-code bugs.
6474 2014-03-11 Richard Biener <rguenther@suse.de>
6476 PR tree-optimization/60429
6477 PR tree-optimization/60485
6478 * tree-ssa-structalias.c (set_union_with_increment): Properly
6479 take into account all fields that overlap the shifted vars.
6480 (do_sd_constraint): Likewise.
6481 (do_ds_constraint): Likewise.
6482 (get_constraint_for_ptr_offset): Likewise.
6484 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
6486 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
6487 (nios2_compute_frame_layout):
6488 Add calculation of cfun->machine->fp_save_offset.
6489 (nios2_expand_prologue): Correct setting of frame pointer register
6491 (nios2_expand_epilogue): Update recovery of stack pointer from
6492 frame pointer accordingly.
6493 (nios2_initial_elimination_offset): Update calculation of offset
6494 for eliminating to HARD_FRAME_POINTER_REGNUM.
6496 2014-03-10 Jakub Jelinek <jakub@redhat.com>
6499 * ipa.c (symtab_remove_unreachable_nodes): Don't call
6500 cgraph_get_create_node on VAR_DECLs.
6502 2014-03-10 Richard Biener <rguenther@suse.de>
6505 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
6507 2014-03-08 Douglas B Rupp <rupp@gnat.com>
6509 * config/vms/vms.opt (vms_float_format): New variable.
6511 2014-03-08 Tobias Burnus <burnus@net-b.de>
6513 * doc/invoke.texi (-fcilkplus): Update implementation status.
6515 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
6516 Richard Biener <rguenther@suse.de>
6518 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
6519 consistently accross all TUs.
6520 (run_gcc): Enable -fshort-double automatically at link at link-time
6521 and disallow override.
6523 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
6526 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
6527 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
6528 if they can't be used.
6530 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6532 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
6533 for Solaris 11/x86 ld.
6534 * configure: Regenerate.
6536 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6538 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
6539 (LIB_TLS_SPEC): Save as ld_tls_libs.
6540 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
6541 (HAVE_AS_IX86_TLSLDM): New test.
6542 * configure, config.in: Regenerate.
6543 * config/i386/i386.c (legitimize_tls_address): Fall back to
6544 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
6545 cannot support TLS_MODEL_LOCAL_DYNAMIC.
6546 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
6547 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
6549 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
6551 * common.opt (fira-loop-pressure): Mark as optimization.
6553 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
6555 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
6556 an OpenMP mappable type.
6558 2014-03-06 Matthias Klose <doko@ubuntu.com>
6560 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
6561 MULTILIB_OSDIRNAMES is not defined.
6563 2014-03-06 Jakub Jelinek <jakub@redhat.com>
6564 Meador Inge <meadori@codesourcery.com>
6567 * config/arm/arm.c (arm_tls_symbol_p): Remove.
6568 (arm_legitimize_address): Call legitimize_tls_address for any
6569 arm_tls_referenced_p expression, handle constant addend. Call it
6570 before testing for !TARGET_ARM.
6571 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
6573 2014-03-06 Richard Biener <rguenther@suse.de>
6579 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
6581 * tree-streamer.c (record_common_node): Assert we don't record
6582 nodes with type double.
6583 (preload_common_node): Skip type double, complex double and double
6584 pointer since it is now frontend dependent due to fshort-double option.
6586 2014-03-06 Richard Biener <rguenther@suse.de>
6588 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
6589 or -fno-lto is specified and the linker has full plugin support.
6590 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
6591 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
6592 * lto-wrapper.c (merge_and_complain): Merge compile-time
6593 optimization levels.
6594 (run_gcc): And pass it through to the link options.
6596 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
6600 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
6602 * cselib.c (remove_useless_values): Skip to avoid quadratic
6603 behavior if the condition moved from...
6604 (cselib_process_insn): ... here holds.
6606 2014-03-05 Jakub Jelinek <jakub@redhat.com>
6609 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
6610 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
6613 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
6614 (TM_H): Add x86-tune.def.
6616 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6618 * config/aarch64/aarch64.c (generic_tunings):
6619 Use cortexa57_extra_costs.
6621 2014-03-05 Jakub Jelinek <jakub@redhat.com>
6624 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
6625 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
6626 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
6629 2014-03-04 Heiher <r@hev.cc>
6631 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
6632 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
6634 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
6636 * config/i386/predicates.md (const2356_operand): Change to ...
6637 (const2367_operand): ... this.
6638 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
6640 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6641 (*avx512pf_scatterpf<mode>sf): Ditto.
6642 (avx512pf_scatterpf<mode>df): Ditto.
6643 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6644 (*avx512pf_scatterpf<mode>df): Ditto.
6645 * config/i386/i386.c (ix86_expand_builtin): Update
6646 incorrect hint operand error message.
6648 2014-03-04 Richard Biener <rguenther@suse.de>
6650 * lto-section-in.c (lto_get_section_data): Fix const cast.
6652 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
6654 * tree-streamer.c (record_common_node): Assert we don't record
6655 nodes with type double.
6656 (preload_common_node): Skip type double, complex double and double
6657 pointer since it is now frontend dependent due to fshort-double option.
6659 2014-03-04 Richard Biener <rguenther@suse.de>
6662 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
6663 (lto_input_toplevel_asms): Likewise.
6664 * lto-section-in.c (lto_get_section_data): Instead do it here
6667 2014-03-04 Richard Biener <rguenther@suse.de>
6669 PR tree-optimization/60382
6670 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
6671 dead PHIs a reduction.
6673 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
6675 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
6677 (_mm_prefetch): Move out of GCC target("sse") pragma.
6678 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
6679 GCC target("prfchw") pragma.
6680 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
6682 * config/i386/i386.c (ix86_option_override_internal): Enable
6683 -mprfchw with -mprefetchwt1.
6685 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
6687 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
6690 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
6692 * opts.h (CL_PCH_IGNORE): Define.
6693 * targhooks.c (option_affects_pch_p):
6694 Return false for options that have CL_PCH_IGNORE set.
6695 * opt-functions.awk: Process PchIgnore.
6696 * doc/options.texi: Document PchIgnore.
6698 * config/arc/arc.opt (misize): Add PchIgnore property.
6700 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6702 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
6703 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
6704 constraint on constants to permit them being loaded into
6705 GENERAL_REGS or BASE_REGS.
6707 2014-03-03 Nick Clifton <nickc@redhat.com>
6709 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
6710 anti-cacnonical alternatives.
6711 (negandhi3_real): New pattern.
6712 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
6714 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
6716 * config/avr/avr-mcus.def: Remove atxmega16x1.
6717 * config/avr/avr-tables.opt: Regenerate.
6718 * config/avr/t-multilib: Regenerate.
6719 * doc/avr-mmcu.texi: Regenerate.
6721 2014-03-03 Tobias Grosser <tobias@grosser.es>
6722 Mircea Namolaru <mircea.namolaru@inria.fr>
6724 PR tree-optimization/58028
6725 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
6728 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6730 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
6731 not handled by recognizers.
6733 2014-03-03 Jakub Jelinek <jakub@redhat.com>
6736 * function.c (expand_function_end): Don't emit
6737 clobber_return_register sequence if clobber_after is a BARRIER.
6738 * cfgexpand.c (construct_exit_block): Append instructions before
6739 return_label to prev_bb.
6741 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6743 * config/rs6000/constraints.md: Document reserved use of "wc".
6745 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
6748 * ipa.c (function_and_variable_visibility): When dissolving comdat
6749 group, also set all symbols to local.
6751 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
6756 2013-12-14 Jan Hubicka <jh@suse.cz>
6758 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
6760 2014-03-02 Jon Beniston <jon@beniston.com>
6766 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
6767 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
6768 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
6769 (simple_return, *simple_return): New patterns
6770 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
6771 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
6773 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
6775 * dwarf2out.c (gen_subprogram_die): Tidy.
6777 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
6780 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
6781 (*mov_t_msb_neg_negc): ... this new insn.
6783 2014-02-28 Jason Merrill <jason@redhat.com>
6786 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
6789 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
6792 * dwarf2out.c (decltype_auto_die): New static.
6793 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
6794 (gen_type_die_with_usage): Handle 'decltype(auto)'.
6795 (is_cxx_auto): Likewise.
6797 2014-02-28 Ian Bolton <ian.bolton@arm.com>
6799 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
6800 we are not using general regs only.
6802 2014-02-28 Richard Biener <rguenther@suse.de>
6805 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
6806 previous fix and only allow to remove trivial pre-headers
6807 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
6808 (remove_forwarder_block): Properly update the latch of a loop.
6810 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
6813 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
6814 (cselib_preserved_hash_table): New.
6815 (preserve_constants_and_equivs): Move preserved vals to it.
6816 (cselib_find_slot): Look it up first.
6817 (cselib_init): Initialize it.
6818 (cselib_finish): Release it.
6819 (dump_cselib_table): Dump it.
6821 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
6824 * cselib.c (remove_useless_values): Skip to avoid quadratic
6825 behavior if the condition moved from...
6826 (cselib_process_insn): ... here holds.
6828 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
6831 * var-tracking.c (vt_initialize): Apply the same condition to
6832 preserve the CFA base value.
6834 2014-02-28 Joey Ye <joey.ye@arm.com>
6837 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
6838 if reload in progress or completed.
6840 2014-02-28 Tobias Burnus <burnus@net-b.de>
6843 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
6846 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
6848 * doc/tm.texi.in (Condition Code Status): Update documention for
6849 relative locations of cc0-setter and cc0-user.
6851 2014-02-27 Jeff Law <law@redhat.com>
6853 PR rtl-optimization/52714
6854 * combine.c (try_combine): When splitting an unrecognized PARALLEL
6855 into two independent simple sets, if I3 is a jump, ensure the
6856 pattern we place into I3 is a (set (pc) ...).
6858 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
6859 Jeff Law <law@redhat.com>
6861 PR rtl-optimization/49847
6862 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
6863 are in different blocks.
6864 * doc/tm.texi (Condition Code Status): Update documention for
6865 relative locations of cc0-setter and cc0-user.
6867 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
6870 * lra.c (lra_emit_add): Check SUBREG too.
6872 2014-02-27 Andreas Schwab <schwab@suse.de>
6874 * config/m68k/m68k.c (m68k_option_override): Disable
6875 -flive-range-shrinkage for classic m68k.
6876 (m68k_override_options_after_change): Likewise.
6878 2014-02-27 Marek Polacek <polacek@redhat.com>
6881 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
6882 -Wmaybe-uninitialized.
6884 2014-02-27 Alan Modra <amodra@gmail.com>
6887 * reload1.c (emit_input_reload_insns): When reload_override_in,
6888 set old to rl->in_reg when rl->in_reg is a subreg.
6890 2014-02-26 Richard Biener <rguenther@suse.de>
6893 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
6895 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
6897 * common/config/i386/predicates.md (const1256_operand): Remove.
6898 (const2356_operand): New.
6899 (const_1_to_2_operand): Remove.
6900 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
6901 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6902 (*avx512pf_gatherpf<mode>sf): Ditto.
6903 (avx512pf_gatherpf<mode>df): Ditto.
6904 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6905 (*avx512pf_gatherpf<mode>df): Ditto.
6906 (avx512pf_scatterpf<mode>sf): Ditto.
6907 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6908 (*avx512pf_scatterpf<mode>sf): Ditto.
6909 (avx512pf_scatterpf<mode>df): Ditto.
6910 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6911 (*avx512pf_scatterpf<mode>df): Ditto.
6912 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
6914 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
6916 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
6917 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
6918 (_mm512_mask_testn_epi64_mask): Move to ...
6919 * config/i386/avx512cdintrin.h: Here.
6920 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
6921 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
6922 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
6923 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
6924 TARGET_AVX512F from TARGET_AVX512CD.
6926 2014-02-26 Richard Biener <rguenther@suse.de>
6929 * ipa.c (walk_polymorphic_call_targets): Properly guard
6930 call to inline_update_overall_summary.
6932 2014-02-26 Bin Cheng <bin.cheng@arm.com>
6935 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
6936 and latches only if requested. Fix latch if it is removed.
6937 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
6938 LOOPS_HAVE_PREHEADERS.
6940 2014-02-25 Andrew Pinski <apinski@cavium.com>
6942 * builtins.c (expand_builtin_thread_pointer): Create a new target
6943 when the target is NULL.
6945 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
6947 PR rtl-optimization/60317
6948 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
6949 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
6950 * lra-assigns.c: Include params.h.
6951 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
6952 other reload pseudos considerations.
6954 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6956 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
6957 to use canonical form for nor<mode>3.
6959 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6962 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
6965 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
6967 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
6968 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
6969 (ix86_handle_option): Handle OPT_mprefetchwt1.
6970 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
6971 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
6973 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
6974 OPTION_MASK_ISA_PREFETCHWT1.
6975 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
6976 (PTA_PREFETCHWT1): New.
6977 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
6978 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
6979 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
6980 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
6981 (*prefetch_avx512pf_<mode>_: Change into ...
6982 (*prefetch_prefetchwt1_<mode>: This.
6983 * config/i386/i386.opt (mprefetchwt1): New.
6984 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
6985 (_mm_prefetch): Handle intent to write.
6986 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
6988 2014-02-25 Richard Biener <rguenther@suse.de>
6991 * emit-rtl.c (mem_attrs_htab): Remove.
6992 (mem_attrs_htab_hash): Likewise.
6993 (mem_attrs_htab_eq): Likewise.
6994 (set_mem_attrs): Always allocate new mem-attrs when something changed.
6995 (init_emit_once): Do not allocate mem_attrs_htab.
6997 2014-02-25 Richard Biener <rguenther@suse.de>
7000 * lto-opts.c (lto_write_options): Output non-explicit conservative
7001 -fwrapv, -fno-trapv and -fno-strict-overflow.
7002 * lto-wrapper.c (merge_and_complain): Handle merging those options.
7003 (run_gcc): And pass them through.
7005 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
7007 * sel-sched.c (calculate_new_fences): New parameter ptime.
7008 Calculate it as a maximum over all fence cycles.
7009 (sel_sched_region_2): Adjust the call to calculate_new_fences.
7010 Print the final schedule timing when sched_verbose.
7012 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
7014 PR rtl-optimization/60292
7015 * sel-sched.c (fill_vec_av_set): Do not reset target availability
7016 bit fot the fence instruction.
7018 2014-02-24 Alangi Derick <alangiderick@gmail.com>
7020 * calls.h: Fix typo in comment.
7022 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
7024 * config/pa/pa.c (pa_output_move_double): Don't valididate when
7025 adjusting offsetable addresses.
7027 2014-02-24 Guozhi Wei <carrot@google.com>
7029 * sparseset.h (sparseset_pop): Fix the wrong index.
7031 2014-02-24 Walter Lee <walt@tilera.com>
7033 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
7034 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
7036 * common/config/tilegx/tilegx-common.c
7037 (TARGET_DEFAULT_TARGET_FLAGS): Define.
7038 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
7040 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
7041 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
7042 (tilegx_gimplify_va_arg_expr): Handle big endian.
7043 (tilegx_expand_unaligned_load): Ditto.
7044 (tilegx_expand_unaligned_store): Ditto.
7045 (TARGET_RETURN_IN_MSB): New.
7046 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
7047 (TARGET_ENDIAN_DEFAULT): New.
7048 (TARGET_BIG_ENDIAN): Handle big endian.
7049 (BYTES_BIG_ENDIAN): Ditto.
7050 (WORDS_BIG_ENDIAN): Ditto.
7051 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
7054 * config/tilegx/tilegx.md (extv): Handle big endian.
7056 (insn_st<n>): Ditto.
7057 (insn_st<n>_add<bitsuffix>): Ditto.
7058 (insn_stnt<n>): Ditto.
7059 (insn_stnt<n>_add<bitsuffix>):Ditto.
7060 (vec_interleave_highv8qi): Handle big endian.
7061 (vec_interleave_highv8qi_be): New.
7062 (vec_interleave_highv8qi_le): New.
7063 (insn_v1int_h): Handle big endian.
7064 (vec_interleave_lowv8qi): Handle big endian.
7065 (vec_interleave_lowv8qi_be): New.
7066 (vec_interleave_lowv8qi_le): New.
7067 (insn_v1int_l): Handle big endian.
7068 (vec_interleave_highv4hi): Handle big endian.
7069 (vec_interleave_highv4hi_be): New.
7070 (vec_interleave_highv4hi_le): New.
7071 (insn_v2int_h): Handle big endian.
7072 (vec_interleave_lowv4hi): Handle big endian.
7073 (vec_interleave_lowv4hi_be): New.
7074 (vec_interleave_lowv4hi_le): New.
7075 (insn_v2int_l): Handle big endian.
7076 (vec_interleave_highv2si): Handle big endian.
7077 (vec_interleave_highv2si_be): New.
7078 (vec_interleave_highv2si_le): New.
7079 (insn_v4int_h): Handle big endian.
7080 (vec_interleave_lowv2si): Handle big endian.
7081 (vec_interleave_lowv2si_be): New.
7082 (vec_interleave_lowv2si_le): New.
7083 (insn_v4int_l): Handle big endian.
7084 * config/tilegx/tilegx.opt (mbig-endian): New option.
7085 (mlittle-endian): New option.
7086 * doc/install.texi: Document tilegxbe-linux.
7087 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
7089 2014-02-24 Martin Jambor <mjambor@suse.cz>
7092 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
7093 there are no parameter descriptors.
7095 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
7097 PR rtl-optimization/60268
7098 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
7099 initialization to ...
7100 (sched_rgn_init): ... here.
7101 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
7103 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
7105 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
7108 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
7110 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
7113 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
7115 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
7116 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
7118 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
7120 * config/microblaze/predicates.md: Add cmp_op predicate.
7121 * config/microblaze/microblaze.md: Add branch_compare instruction
7122 which uses cmp_op predicate and emits cmp insn before branch.
7123 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
7124 to microblaze_expand_conditional_branch and consolidate logic.
7125 (microblaze_expand_conditional_branch): emit branch_compare
7126 insn instead of handling cmp op separate from branch insn.
7128 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7130 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
7133 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7135 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
7136 define_insn with define_expand and new define_insn
7137 *altivec_lve<VI_char>x_internal.
7138 (altivec_stve<VI_char>x): Replace define_insn with define_expand
7139 and new define_insn *altivec_stve<VI_char>x_internal.
7140 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
7142 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
7144 (altivec_expand_stvex_be): New function.
7146 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
7148 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
7149 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
7150 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
7151 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
7153 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
7156 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
7157 instead of emit_move_insn.
7159 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7161 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
7163 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
7164 gen_altivec_vsumsws.
7166 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7168 * config/rs6000/altivec.md (altivec_lvxl): Rename as
7169 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
7170 (altivec_lvxl_<mode>): New define_expand incorporating
7171 -maltivec=be semantics where needed.
7172 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
7173 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
7174 semantics where needed.
7175 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
7176 (altivec_stvx_<mode>): New define_expand incorporating
7177 -maltivec=be semantics where needed.
7178 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
7179 VM2 iterator instead of V4SI.
7180 (altivec_stvxl_<mode>): New define_expand incorporating
7181 -maltivec=be semantics where needed.
7182 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
7183 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
7184 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
7185 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
7186 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
7187 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
7188 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
7189 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
7190 ALTIVEC_BUILTIN_STVXL.
7191 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
7192 (altivec_expand_stvx_be): Likewise.
7193 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
7194 (altivec_expand_lvx_be): Likewise.
7195 (altivec_expand_stvx_be): Likewise.
7196 (altivec_expand_builtin): Add cases for
7197 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
7198 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
7199 (altivec_init_builtins): Add definitions for
7200 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
7201 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
7203 2014-02-21 Catherine Moore <clm@codesourcery.com>
7205 * doc/invoke.texi (mvirt, mno-virt): Document.
7206 * config/mips/mips.opt (mvirt): New option.
7207 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
7209 2014-02-21 Richard Biener <rguenther@suse.de>
7211 PR tree-optimization/60276
7212 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
7213 (STMT_VINFO_MIN_NEG_DIST): New macro.
7214 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
7215 STMT_VINFO_MIN_NEG_DIST.
7216 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
7217 made for negative dependence distances still hold.
7219 2014-02-21 Richard Biener <rguenther@suse.de>
7222 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
7223 DECL_INITIAL for globals not in the current function context.
7225 2014-02-21 Jakub Jelinek <jakub@redhat.com>
7227 PR tree-optimization/56490
7228 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
7229 * tree-ssa-uninit.c: Include params.h.
7230 (compute_control_dep_chain): Add num_calls argument, return false
7231 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
7232 num_calls to recursive call.
7233 (find_predicates): Change dep_chain into normal array,
7234 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
7235 variable and adjust compute_control_dep_chain caller.
7236 (find_def_preds): Likewise.
7238 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
7240 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
7241 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
7243 2014-02-21 Nick Clifton <nickc@redhat.com>
7245 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
7246 (pushhi1): Likewise.
7247 (popqi1): Add mode to pre_dec.
7250 2014-02-21 Jakub Jelinek <jakub@redhat.com>
7252 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
7253 mode for mask of V8SFmode permutation.
7255 2014-02-20 Richard Henderson <rth@redhat.com>
7258 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
7259 a new pseudo for OLDVAL.
7261 2014-02-20 Jakub Jelinek <jakub@redhat.com>
7264 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
7265 gen_reg_rtx if d->testing_p.
7266 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
7267 if d->testing_p and we will certainly return true.
7268 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
7271 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
7273 * emit-rtl.c (gen_reg_rtx): Assert that
7274 crtl->emit.regno_pointer_align_length is non-zero.
7276 2014-02-20 Richard Henderson <rth@redhat.com>
7279 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
7280 on failure the store back into EXPECT.
7282 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
7283 Sandra Loosemore <sandra@codesourcery.com>
7285 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
7286 * config/nios2/nios2.c (nios2_function_profiler): Add
7287 -fPIC (flag_pic == 2) support.
7288 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
7289 (nios2_large_offset_p): New function.
7290 (nios2_unspec_reloc_p): Move up position, update to use
7291 nios2_large_offset_p.
7292 (nios2_unspec_address): Remove function.
7293 (nios2_unspec_offset): New function.
7294 (nios2_large_got_address): New function.
7295 (nios2_got_address): Add large offset support.
7296 (nios2_legitimize_tls_address): Update usage of removed and new
7298 (nios2_symbol_binds_local_p): New function.
7299 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
7300 (nios2_legitimize_address): Update to use nios2_large_offset_p.
7301 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
7302 (nios2_print_operand): Merge H/L processing, add hiadj/lo
7303 processing for (const (unspec ...)).
7304 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
7306 2014-02-20 Richard Biener <rguenther@suse.de>
7308 * tree-cfg.c (replace_uses_by): Mark altered BBs before
7309 doing the substitution.
7310 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
7312 2014-02-20 Martin Jambor <mjambor@suse.cz>
7315 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
7316 info when checking whether lattices are bottom.
7318 2014-02-20 Richard Biener <rguenther@suse.de>
7321 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
7324 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
7327 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
7328 parameter specifying the scaling.
7329 (inline_call): Update.
7330 (want_inline_recursively): Guard division by zero.
7331 (recursive_inlining): Update.
7332 * ipa-inline.h (clone_inlined_nodes): Update.
7334 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
7337 * config/i386/i386.c (classify_argument): Pass structures of size
7338 64 bytes or less in register.
7340 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
7341 Kirill Yukhin <kirill.yukhin@intel.com>
7343 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
7344 (_mm_rcp28_round_ss): Ditto.
7345 (_mm_rsqrt28_round_sd): Ditto.
7346 (_mm_rsqrt28_round_ss): Ditto.
7347 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
7348 (_mm_rcp14_round_ss): Ditto.
7349 (_mm_rsqrt14_round_sd): Ditto.
7350 (_mm_rsqrt14_round_ss): Ditto.
7351 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
7352 the first input operand, get rid of match_dup.
7353 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
7355 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
7357 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
7358 operand as the first input operand, set type attribute.
7359 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
7361 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
7362 operand as the first input operand, set type attribute.
7364 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7366 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
7369 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
7372 * config/i386/i386.c (construct_container): Remove TFmode check
7373 for X86_64_INTEGER_CLASS.
7375 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
7378 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
7379 only when -Wpsabi is enabled.
7381 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
7384 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
7385 passing arrays in registers are the same as for structs, so remove the
7386 special case for them.
7388 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
7390 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
7391 destination type, extract only the valid bits if the source type is not
7392 integral and has a different mode.
7394 2014-02-19 Richard Biener <rguenther@suse.de>
7397 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
7400 2014-02-19 Richard Biener <rguenther@suse.de>
7403 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
7404 (ipa_modify_call_arguments): Emit an argument load explicitely and
7405 preserve virtual SSA form there and for the replacement call.
7406 Do not update SSA form nor free dominance info.
7408 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
7410 * ipa.c (function_and_variable_visibility): Also clear WEAK
7411 flag when disolving COMDAT_GROUP.
7413 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
7415 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
7416 * ipa-prop.c (ipa_set_jf_known_type): Return early when
7418 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
7419 do more sanity checks.
7420 (detect_type_change): Return true when giving up early.
7421 (compute_complex_assign_jump_func): Fix type parameter of
7422 ipa_set_ancestor_jf.
7423 (compute_complex_ancestor_jump_func): Likewise.
7424 (update_jump_functions_after_inlining): Fix updating of
7426 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
7428 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
7430 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
7431 inline clones when edge disappears.
7433 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
7436 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
7437 Split 64-bit moves into 2 patterns. Do not allow the use of
7438 direct move for TDmode in little endian, since the decimal value
7439 has little endian bytes within a word, but the 64-bit pieces are
7440 ordered in a big endian fashion, and normal subreg's of TDmode are
7442 (mov<mode>_64bit_dm): Likewise.
7443 (movtd_64bit_nodm): Likewise.
7445 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
7447 PR tree-optimization/60174
7448 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
7449 statement of an SSA_NAME that occurs in an abnormal PHI node.
7451 2014-02-18 Jakub Jelinek <jakub@redhat.com>
7454 * final.c (SEEN_BB): Remove.
7455 (SEEN_NOTE, SEEN_EMITTED): Renumber.
7456 (final_scan_insn): Don't force_source_line on second
7457 NOTE_INSN_BASIC_BLOCK.
7459 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
7462 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
7463 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
7464 (type_natural_mode): Warn ABI change when %zmm register is not
7465 available for AVX512F vector value passing.
7467 2014-02-18 Kai Tietz <ktietz@redhat.com>
7470 * config/i386/i386.c (ix86_expand_prologue): Use value in
7471 rax register as displacement when restoring %r10 or %rax.
7472 Fix wrong offset when restoring both registers.
7474 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
7476 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
7477 assertion with conditional return.
7479 2014-02-18 Jakub Jelinek <jakub@redhat.com>
7480 Uros Bizjak <ubizjak@gmail.com>
7483 * config/i386/driver-i386.c (host_detect_local_cpu): If
7484 YMM state is not saved by the OS, also clear has_f16c. Move
7485 CPUID 0x80000001 handling before YMM state saving checking.
7487 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
7489 PR rtl-optimization/58960
7490 * haifa-sched.c (alloc_global_sched_pressure_data): New,
7491 factored out from ...
7492 (sched_init): ... here.
7493 (free_global_sched_pressure_data): New, factored out from ...
7494 (sched_finish): ... here.
7495 * sched-int.h (free_global_sched_pressure_data): Declare.
7496 * sched-rgn.c (nr_regions_initial): New static global.
7497 (haifa_find_rgns): Initialize it.
7498 (schedule_region): Disable sched-pressure for the newly
7501 2014-02-17 Richard Biener <rguenther@suse.de>
7503 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
7504 release SSA defs of pattern stmts.
7506 2014-02-17 Richard Biener <rguenther@suse.de>
7508 * tree-inline.c (expand_call_inline): Release the virtual
7509 operand defined by the call we are about to inline.
7511 2014-02-17 Richard Biener <rguenther@suse.de>
7513 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
7515 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
7516 Ilya Tocar <ilya.tocar@intel.com>
7518 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
7519 arguments order in builtin.
7520 (_mm512_permutexvar_epi64): Ditto.
7521 (_mm512_mask_permutexvar_epi64): Ditto
7522 (_mm512_maskz_permutexvar_epi32): Ditto
7523 (_mm512_permutexvar_epi32): Ditto
7524 (_mm512_mask_permutexvar_epi32): Ditto
7526 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7528 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
7529 (p8_vmrgow): Likewise.
7531 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7533 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
7536 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
7539 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
7540 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
7541 into 64-bit and 32-bit moves. On 64-bit moves, add support for
7542 using direct move instructions on ISA 2.07. Also adjust
7543 instruction length for 64-bit.
7544 (mov<mode>_64bit, TFmode/TDmode): Likewise.
7545 (mov<mode>_32bit, TFmode/TDmode): Likewise.
7547 2014-02-15 Alan Modra <amodra@gmail.com>
7551 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
7552 find_replacement on parts of insn rtl that might be reloaded.
7554 2014-02-15 Richard Biener <rguenther@suse.de>
7556 PR tree-optimization/60183
7557 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
7558 (tree_ssa_phiprop): Calculate and free post-dominators.
7560 2014-02-14 Jeff Law <law@redhat.com>
7562 PR rtl-optimization/60131
7563 * ree.c (get_extended_src_reg): New function.
7564 (combine_reaching_defs): Use it rather than assuming location of REG.
7565 (find_and_remove_re): Verify first operand of extension is
7566 a REG before adding the insns to the copy list.
7568 2014-02-14 Roland McGrath <mcgrathr@google.com>
7570 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
7571 * configure: Regenerated.
7572 * config.in: Regenerated.
7573 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
7574 instead of ASM_SHORT.
7576 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
7577 Richard Earnshaw <rearnsha@arm.com>
7579 PR rtl-optimization/59535
7580 * lra-constraints.c (process_alt_operands): Encourage alternative
7581 when unassigned pseudo class is superset of the alternative class.
7582 (inherit_reload_reg): Don't inherit when optimizing for code size.
7583 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
7584 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
7585 modes not less than 4 for Thumb1.
7587 2014-02-14 Kyle McMartin <kyle@redhat.com>
7590 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
7592 2014-02-14 Richard Biener <rguenther@suse.de>
7594 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
7595 (get_frame_arg): Drop the assert with langhook types_compatible_p.
7596 Do not strip INDIRECT_REFs.
7598 2014-02-14 Richard Biener <rguenther@suse.de>
7601 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
7602 DECL_FUNCTION_SPECIFIC_TARGET.
7603 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
7604 * tree-streamer-out.c (pack_ts_target_option): Remove.
7605 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
7606 (write_ts_function_decl_tree_pointers): Do not stream
7607 DECL_FUNCTION_SPECIFIC_TARGET.
7608 * tree-streamer-in.c (unpack_ts_target_option): Remove.
7609 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
7610 (lto_input_ts_function_decl_tree_pointers): Do not stream
7611 DECL_FUNCTION_SPECIFIC_TARGET.
7613 2014-02-14 Jakub Jelinek <jakub@redhat.com>
7615 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
7616 (get_initial_def_for_induction, vectorizable_induction): Ignore
7617 debug stmts when looking for exit_phi.
7618 (vectorizable_live_operation): Fix up condition.
7620 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
7622 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
7623 nreverse() because it changes the content of original tree list.
7625 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
7627 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
7628 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
7630 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
7632 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
7633 GNU coding standards.
7635 2014-02-13 Jakub Jelinek <jakub@redhat.com>
7638 * dwarf2out.c (gen_subprogram_die): Don't call
7639 add_calling_convention_attribute if subr_die is old_die.
7641 2014-02-13 Sharad Singhai <singhai@google.com>
7643 * doc/optinfo.texi: Fix order of nodes.
7645 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
7647 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
7648 operands[2], not operands[3].
7650 2014-02-13 Richard Biener <rguenther@suse.de>
7653 * doc/install.texi (ISL): Update recommended version to 0.12.2,
7654 mention the possibility of an in-tree build.
7655 (CLooG): Update recommended version to 0.18.1, mention the
7656 possibility of an in-tree build and clarify that the ISL
7657 bundled with CLooG does not work.
7659 2014-02-13 Jakub Jelinek <jakub@redhat.com>
7662 * expr.c (compress_float_constant): If x is a hard register,
7663 extend into a pseudo and then move to x.
7665 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
7667 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
7668 caused by bad second argument to warning_at() with -mhotpatch and
7669 nested functions (e.g. with gfortran).
7671 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
7673 * opts.c (option_name): Remove "enabled by default" rider.
7675 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
7677 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
7679 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
7680 Uros Bizjak <ubizjak@gmail.com>
7683 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
7684 * configure: Regenerated.
7686 2014-02-12 Richard Biener <rguenther@suse.de>
7688 * vec.c (vec_prefix::calculate_allocation): Move as
7689 inline variant to vec.h.
7690 (vec_prefix::calculate_allocation_1): New out-of-line version.
7691 * vec.h (vec_prefix::calculate_allocation_1): Declare.
7692 (vec_prefix::m_has_auto_buf): Rename to ...
7693 (vec_prefix::m_using_auto_storage): ... this.
7694 (vec_prefix::calculate_allocation): Inline the easy cases
7695 and dispatch to calculate_allocation_1 which doesn't need the
7697 (va_heap::reserve): Use gcc_checking_assert.
7698 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
7699 m_using_auto_storage.
7700 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
7702 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
7703 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
7704 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
7706 2014-02-12 Richard Biener <rguenther@suse.de>
7708 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
7709 when we found a dependence.
7711 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
7713 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
7715 (maybe_fold_stmt): ... into this new function.
7716 * omp-low.c (lower_omp): Update comment.
7718 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
7721 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
7724 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
7726 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
7727 identifiers in comments.
7728 (cortexa53_extra_costs): Likewise.
7729 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
7730 (cortexa7_extra_costs): Likewise.
7731 (cortexa12_extra_costs): Likewise.
7732 (cortexa15_extra_costs): Likewise.
7733 (v7m_extra_costs): Likewise.
7735 2014-02-12 Richard Biener <rguenther@suse.de>
7738 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
7739 of posix_memalign being successful.
7740 (lower_stmt): Restrict lowering of posix_memalign to when
7741 -ftree-bit-ccp is enabled.
7743 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
7745 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
7747 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
7749 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
7751 PR rtl-optimization/60116
7752 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
7753 other_insn once the combination has been validated.
7755 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
7758 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
7760 * ipa-devirt.c: Include demangle.h
7761 (odr_violation_reported): New static variable.
7762 (add_type_duplicate): Update odr_violations.
7763 (maybe_record_node): Add completep parameter; update it.
7764 (record_target_from_binfo): Add COMPLETEP parameter;
7765 update it as needed.
7766 (possible_polymorphic_call_targets_1): Likewise.
7767 (struct polymorphic_call_target_d): Add nonconstruction_targets;
7768 rename FINAL to COMPLETE.
7769 (record_targets_from_bases): Sanity check we found the binfo;
7770 fix COMPLETEP updating.
7771 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
7772 parameter, fix computing of COMPLETEP.
7773 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
7774 at LTO time do demangling.
7775 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
7776 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
7778 (gimple_get_virt_method_for_binfo): Likewise.
7779 * gimple-fold.h (gimple_get_virt_method_for_binfo,
7780 gimple_get_virt_method_for_vtable): Update prototypes.
7782 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
7785 * genautomata.c (add_presence_absence): Fix typo with
7786 {final_}presence_list.
7788 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
7791 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
7792 for VSX/Altivec vectors that land in GPR registers.
7794 2014-02-11 Richard Henderson <rth@redhat.com>
7795 Jakub Jelinek <jakub@redhat.com>
7798 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
7799 around drhs if type conversion to lacc->type is not useless.
7801 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7803 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
7805 (cortex-a57.cortex-a53): Likewise.
7806 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
7808 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7810 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
7813 2014-02-11 Renlin Li <Renlin.Li@arm.com>
7815 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
7816 add_options_for_arm_vfp3.
7818 2014-02-11 Jeff Law <law@redhat.com>
7821 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
7822 object with an undesirable mode.
7824 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7827 * config/i386/sol2-9.h: New file.
7828 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
7829 *-*-solaris2.9*): Use it.
7831 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
7833 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
7834 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
7836 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
7838 * config/microblaze/microblaze.c: Extend mcpu version format
7840 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
7842 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
7844 2014-02-10 Richard Henderson <rth@redhat.com>
7847 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
7848 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
7849 ms-abi vs -mno-accumulate-outgoing-args.
7850 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
7851 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
7854 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
7857 * cfgexpand.c (expand_asm_operands): Attach source location to
7858 ASM_INPUT rtx objects.
7859 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
7861 2014-02-10 Nick Clifton <nickc@redhat.com>
7863 * config/mn10300/mn10300.c (popcount): New function.
7864 (mn10300_expand_prologue): Include saved registers in stack usage
7867 2014-02-10 Jeff Law <law@redhat.com>
7870 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
7871 when changing the SET_DEST of a prior insn to avoid an input reload.
7873 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
7875 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
7876 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
7877 -mcall-openbsd, or -mcall-linux.
7878 (CC1_ENDIAN_BIG_SPEC): Remove.
7879 (CC1_ENDIAN_LITTLE_SPEC): Remove.
7880 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
7881 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
7882 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
7883 and %cc1_endian_default.
7884 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
7886 2014-02-10 Richard Biener <rguenther@suse.de>
7888 PR tree-optimization/60115
7889 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
7890 MEM_REF handling. Properly verify that the accesses are not
7891 out of the objects bound.
7893 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7895 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
7898 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
7900 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
7901 proper constants and fix formatting.
7902 (possible_polymorphic_call_targets): Fix formatting.
7904 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
7905 Ilya Tocar <ilya.tocar@intel.com>
7907 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
7908 (_mm512_loadu_epi32): Renamed into...
7909 (_mm512_loadu_si512): This.
7910 (_mm512_storeu_epi32): Renamed into...
7911 (_mm512_storeu_si512): This.
7912 (_mm512_maskz_ceil_ps): Removed.
7913 (_mm512_maskz_ceil_pd): Ditto.
7914 (_mm512_maskz_floor_ps): Ditto.
7915 (_mm512_maskz_floor_pd): Ditto.
7916 (_mm512_floor_round_ps): Ditto.
7917 (_mm512_floor_round_pd): Ditto.
7918 (_mm512_ceil_round_ps): Ditto.
7919 (_mm512_ceil_round_pd): Ditto.
7920 (_mm512_mask_floor_round_ps): Ditto.
7921 (_mm512_mask_floor_round_pd): Ditto.
7922 (_mm512_mask_ceil_round_ps): Ditto.
7923 (_mm512_mask_ceil_round_pd): Ditto.
7924 (_mm512_maskz_floor_round_ps): Ditto.
7925 (_mm512_maskz_floor_round_pd): Ditto.
7926 (_mm512_maskz_ceil_round_ps): Ditto.
7927 (_mm512_maskz_ceil_round_pd): Ditto.
7928 (_mm512_expand_pd): Ditto.
7929 (_mm512_expand_ps): Ditto.
7930 * config/i386/i386.c (ix86_builtins): Remove
7931 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
7932 (bdesc_args): Ditto.
7933 * config/i386/predicates.md (const1256_operand): New.
7934 (const_1_to_2_operand): Ditto.
7935 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
7936 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
7937 (*avx512pf_gatherpf<mode>sf): Ditto.
7938 (avx512pf_gatherpf<mode>df): Ditto.
7939 (*avx512pf_gatherpf<mode>df_mask): Ditto.
7940 (*avx512pf_gatherpf<mode>df): Ditto.
7941 (avx512pf_scatterpf<mode>sf): Ditto.
7942 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
7943 (*avx512pf_scatterpf<mode>sf): Ditto.
7944 (avx512pf_scatterpf<mode>df): Ditto.
7945 (*avx512pf_scatterpf<mode>df_mask): Ditto.
7946 (*avx512pf_scatterpf<mode>df): Ditto.
7947 (avx512f_expand<mode>): Removed.
7948 (<shift_insn><mode>3<mask_name>): Change predicate type.
7950 2014-02-08 Jakub Jelinek <jakub@redhat.com>
7952 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
7953 not at the end of datarefs vector use ordered_remove to avoid
7954 reordering datarefs vector.
7957 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
7958 mark local addressable non-static vars as GOVD_PRIVATE
7959 instead of GOVD_LOCAL.
7960 * omp-low.c (lower_omp_for): Move gimple_bind_vars
7961 and BLOCK_VARS of gimple_bind_block to new_stmt rather
7965 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
7966 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
7967 assume_aligned or alloc_align attributes.
7968 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
7969 arguments. Handle also assume_aligned and alloc_align attributes.
7970 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
7971 calls to functions with assume_aligned or alloc_align attributes.
7972 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
7974 2014-02-08 Terry Guo <terry.guo@arm.com>
7976 * doc/invoke.texi: Document ARM -march=armv7e-m.
7978 2014-02-08 Jakub Jelinek <jakub@redhat.com>
7980 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
7981 flag on __cilkrts_rethrow builtin.
7984 * ipa-cp.c (determine_versionability): Fail at -O0
7985 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
7986 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
7989 2014-02-04 Jakub Jelinek <jakub@redhat.com>
7992 * tree-inline.c (copy_forbidden): Fail for
7993 __attribute__((optimize (0))) functions.
7995 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
7997 * varpool.c: Include pointer-set.h.
7998 (varpool_remove_unreferenced_decls): Variables in other partitions
7999 will not be output; be however careful to not lose information
8002 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
8004 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
8005 lookup in the vtable constructor.
8007 2014-02-07 Jeff Law <law@redhat.com>
8010 * config/m68k/m68k.md (ashldi_extsi): Turn into a
8011 define_insn_and_split.
8013 * ipa-inline.c (inline_small_functions): Fix typos.
8015 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
8017 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
8018 (s390_can_use_return_insn): Declare.
8019 * config/s390/s390.h (EPILOGUE_USES): Define.
8020 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
8022 (s390_chunkify_start): Handle return JUMP_LABELs.
8023 (s390_early_mach): Emit a main_pool instruction on the entry edge.
8024 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
8025 (s390_can_use_return_insn): New functions.
8026 (s390_fix_long_loop_prediction): Handle conditional returns.
8027 (TARGET_SET_UP_BY_PROLOGUE): Define.
8028 * config/s390/s390.md (ANY_RETURN): New code iterator.
8029 (*creturn, *csimple_return, return, simple_return): New patterns.
8031 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
8033 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
8034 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
8035 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
8036 REG_CFA_RESTORE list when deciding not to restore a register.
8038 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
8040 * config/s390/s390.c: Include tree-pass.h and context.h.
8041 (s390_early_mach): New function, split out from...
8042 (s390_emit_prologue): ...here.
8043 (pass_data_s390_early_mach): New pass structure.
8044 (pass_s390_early_mach): New class.
8045 (s390_option_override): Create and register early_mach pass.
8046 Move to end of file.
8048 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
8050 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
8051 to match for the exit block.
8053 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8055 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
8056 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
8057 Reject misaligned operands.
8059 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8061 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
8063 2014-02-07 Richard Biener <rguenther@suse.de>
8066 * gimple-low.c (lower_builtin_posix_memalign): New function.
8067 (lower_stmt): Call it to lower posix_memalign in a way
8068 to make alignment info accessible.
8070 2014-02-07 Jakub Jelinek <jakub@redhat.com>
8073 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
8074 __builtin_setjmp_receiver.
8076 2014-02-07 Richard Biener <rguenther@suse.de>
8079 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
8080 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
8081 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
8082 Handle BUILT_IN_POSIX_MEMALIGN.
8083 (find_func_clobbers): Likewise.
8084 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
8085 (call_may_clobber_ref_p_1): Likewise.
8087 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
8090 * ipa-devirt.c (record_target_from_binfo): Remove overactive
8093 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
8096 * lto-cgraph.c (lto_output_node): Use
8097 symtab_get_symbol_partitioning_class.
8098 (lto_output_varpool_node): likewise.
8099 (symtab_get_symbol_partitioning_class): Move here from
8101 * cgraph.h (symbol_partitioning_class): Likewise.
8102 (symtab_get_symbol_partitioning_class): Declare.
8104 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
8106 * ggc.h (ggc_internal_cleared_alloc): New macro.
8107 * vec.h (vec_safe_copy): Handle memory stats.
8108 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
8109 * target-globals.c (save_target_globals): Likewise.
8111 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
8114 * expr.c (emit_move_resolve_push): Export; be bit more selective
8115 on when to clear alias set.
8116 * expr.h (emit_move_resolve_push): Declare.
8117 * function.h (struct function): Add tail_call_marked.
8118 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
8119 * config/i386/i386-protos.h (ix86_expand_push): Remove.
8120 * config/i386/i386.md (TImode move expander): De not call
8122 (FP push expanders): Preserve memory attributes.
8123 * config/i386/sse.md (push<mode>1): Remove.
8124 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
8125 (ix86_expand_push): Remove.
8126 * config/i386/mmx.md (push<mode>1): Remove.
8128 2014-02-06 Jakub Jelinek <jakub@redhat.com>
8130 PR rtl-optimization/60030
8131 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
8132 lopart with paradoxical subreg before shifting it up by hprec.
8134 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8136 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
8137 Remove extra newline at end of file.
8138 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
8139 (arm_issue_rate): Handle cortexa57.
8140 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
8141 (cortex-a57.cortex-a53): Likewise.
8143 2014-02-06 Jakub Jelinek <jakub@redhat.com>
8146 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
8147 don't record in REG_FRAME_RELATED_EXPR registers not set in that
8149 (arm_expand_prologue): Adjust all callers.
8150 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
8151 info, registers also at the lowest numbered registers side. Use
8152 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
8156 * var-tracking.c (adjust_mems): Before adding a SET to
8157 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
8159 2014-02-06 Alan Modra <amodra@gmail.com>
8162 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
8163 change SDmode to DDmode when lra_in_progress.
8165 2014-02-06 Jakub Jelinek <jakub@redhat.com>
8168 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
8169 free_data_ref on the dr first, and before goto again also set dr
8170 to the next dr. For simd_lane_access, free old datarefs[i] before
8171 overwriting it. For get_vectype_for_scalar_type failure, don't
8172 free_data_ref if simd_lane_access.
8174 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
8177 * tree.h (opts_for_fn): New inline function.
8178 (opt_for_fn): Define.
8179 * config/i386/i386.c (ix86_function_regparm): Use
8180 opt_for_fn (decl, optimize) instead of optimize.
8182 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
8184 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
8185 for SYMBOL_REF in large memory model.
8187 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8189 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
8191 (cortex-a57): Likewise.
8192 (cortex-a57.cortex-a53): Likewise.
8194 2014-02-06 Yury Gribov <y.gribov@samsung.com>
8195 Kugan Vivekanandarajah <kuganv@linaro.org>
8197 * config/arm/arm.c (arm_vector_alignment_reachable): Check
8199 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
8201 2014-02-06 Richard Biener <rguenther@suse.de>
8203 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
8204 set_loop_copy and initialize_original_copy_tables.
8206 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
8208 * config/aarch64/aarch64-simd.md
8209 (aarch64_ashr_simddi): Change QI to SI.
8211 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
8212 Jakub Jelinek <jakub@redhat.com>
8215 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
8218 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8220 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
8221 CODE_FOR_altivec_vpku[hw]um to
8222 CODE_FOR_altivec_vpku[hw]um_direct.
8223 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
8224 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
8225 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
8226 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
8228 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8230 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
8231 generation for -maltivec=be.
8232 (altivec_vsumsws): Simplify redundant test.
8234 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8236 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
8237 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
8238 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
8239 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
8240 gen_altivec_vpkuwum.
8241 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
8243 (altivec_vpks<VI_char>ss): Likewise.
8244 (altivec_vpks<VI_char>us): Likewise.
8245 (altivec_vpku<VI_char>us): Likewise.
8246 (altivec_vpku<VI_char>um): Likewise.
8247 (altivec_vpku<VI_char>um_direct): New (copy of
8248 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
8250 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
8251 target is little endian and -maltivec=be is not specified.
8252 (*altivec_vupkhs<VU_char>_direct): New (copy of
8253 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
8254 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
8255 target is little endian and -maltivec=be is not specified.
8256 (*altivec_vupkls<VU_char>_direct): New (copy of
8257 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
8258 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
8259 little endian and -maltivec=be is not specified.
8260 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
8261 little endian and -maltivec=be is not specified.
8263 2014-02-05 Richard Henderson <rth@redhat.com>
8266 * combine-stack-adj.c: Revert r206943.
8267 * sched-int.h (struct deps_desc): Add last_args_size.
8268 * sched-deps.c (init_deps): Initialize it.
8269 (sched_analyze_insn): Add OUTPUT dependencies between insns that
8270 contain REG_ARGS_SIZE notes.
8272 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
8274 * lto-cgraph.c (asm_nodes_output): Make global.
8275 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
8276 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
8277 (driver_handle_option): Handle OPT_fwpa.
8279 2014-02-05 Jakub Jelinek <jakub@redhat.com>
8282 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
8283 a comment typo and formatting issue. If odr_hash hasn't been
8284 created, return vNULL and set *completep to false.
8287 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
8288 bb with no successors.
8290 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
8293 * doc/invoke.texi (-march): Clarify documentation for ARM.
8297 2014-02-05 Richard Biener <rguenther@suse.de>
8299 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
8300 when not vectorizing because of too many alias checks.
8301 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
8302 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
8304 2014-02-05 Nick Clifton <nickc@redhat.com>
8306 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
8307 accept extended registers in any mode when compiling for the MN10300.
8309 2014-02-05 Yury Gribov <y.gribov@samsung.com>
8311 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
8312 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
8313 sanitization attributes.
8314 (can_inline_edge_p): Likewise.
8315 (sanitize_attrs_match_for_inline_p): New function.
8317 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
8319 * ipa-prop.c (detect_type_change): Shor circuit testing of
8320 type changes on THIS pointer.
8322 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
8325 * config/pa/pa.c (legitimize_tls_address): Return original address
8326 if not passed a SYMBOL_REF rtx.
8327 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
8329 (pa_emit_move_sequence): Simplify TLS source operands.
8330 (pa_legitimate_constant_p): Reject all TLS constants.
8331 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
8332 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
8334 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
8336 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
8337 groups when we know they are controlled by LTO.
8338 * varasm.c (default_binds_local_p_1): If object is in other partition,
8339 it will be resolved locally.
8341 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
8343 * config/host-linux.c (linux_gt_pch_use_address): Don't
8344 use SSIZE_MAX because it is not always defined.
8346 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
8349 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
8350 threshold for pseudo splitting.
8351 (update_ebb_live_info): Process call argument hard registers and
8352 hard registers from insn definition too.
8353 (max_small_class_regs_num): New constant.
8354 (inherit_in_ebb): Update live hard regs through EBBs. Update
8355 reloads_num only for small register classes. Don't split for
8358 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
8361 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
8364 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
8366 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
8369 2014-02-04 Marek Polacek <polacek@redhat.com>
8371 * gdbinit.in (pel): Define.
8373 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
8375 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
8378 2014-02-04 Richard Biener <rguenther@suse.de>
8381 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
8382 in function context local.
8383 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
8384 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
8385 similar to LTO_imported_decl_ref.
8387 2014-02-04 Jakub Jelinek <jakub@redhat.com>
8389 PR tree-optimization/60002
8390 * cgraphclones.c (build_function_decl_skip_args): Clear
8393 PR tree-optimization/60023
8394 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
8395 false to gsi_replace.
8396 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
8397 has been in some EH region and vec_stmt could throw, add
8398 vec_stmt into the same EH region.
8399 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
8400 has no lhs, ignore it.
8401 * internal-fn.c (expand_MASK_LOAD): Likewise.
8404 * tree-inline.c (copy_forbidden): Fail for
8405 __attribute__((optimize (0))) functions.
8408 * omp-low.c (simd_clone_struct_copy): If from->inbranch
8409 is set, copy one less argument.
8410 (expand_simd_clones): Don't subtract clone_info->inbranch
8411 from simd_clone_struct_alloc argument.
8413 PR rtl-optimization/57915
8414 * recog.c (simplify_while_replacing): If all unary/binary/relational
8415 operation arguments are constant, attempt to simplify those.
8418 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
8419 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
8421 2014-02-04 Richard Biener <rguenther@suse.de>
8423 PR tree-optimization/60012
8424 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
8425 TBAA disambiguation to all DDRs.
8427 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8430 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
8431 (LINK_SPEC): Use it for -shared, -shared-libgcc.
8433 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
8436 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
8438 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
8440 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
8441 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
8443 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
8446 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
8447 to figure out targets of polymorphic calls with known decl.
8448 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
8449 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
8450 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
8451 (get_polymorphic_call_info): ... here.
8452 (get_polymorphic_call_info_from_invariant): New function.
8454 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
8456 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
8457 lookup via vtable pointer; check for type consistency
8458 and turn inconsitent facts into UNREACHABLE.
8459 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
8460 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
8461 type inconsistent querries; return UNREACHABLE instead.
8463 2014-02-03 Richard Henderson <rth@twiddle.net>
8466 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
8467 already processed this node.
8468 (normalize_one_pred_1): Pass along mark_set.
8469 (normalize_one_pred): Create and destroy a pointer_set_t.
8470 (normalize_one_pred_chain): Likewise.
8472 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
8474 PR gcov-profile/58602
8475 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
8477 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
8480 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
8481 -fno-devirtualize; try to devirtualize by the knowledge of
8482 virtual table pointer given by aggregate propagation.
8483 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
8484 (ipa_print_node_jump_functions): Dump also offset that
8485 is relevant for polymorphic calls.
8486 (determine_known_aggregate_parts): Add arg_type parameter; use it
8487 instead of determining the type from pointer type.
8488 (ipa_compute_jump_functions_for_edge): Update call of
8489 determine_known_aggregate_parts.
8490 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
8491 (gimple_get_virt_method_for_binfo): ... here; simplify using
8492 vtable_pointer_value_to_vtable.
8493 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
8494 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
8495 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
8496 (vtable_pointer_value_to_vtable): Break out from ...; handle also
8498 (vtable_pointer_value_to_binfo): ... here.
8499 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
8501 2014-02-03 Teresa Johnson <tejohnson@google.com>
8503 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
8504 redef of outer loop index variable.
8506 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
8510 * doc/extend.texi (Function Attributes): Typo.
8512 2014-02-03 Cong Hou <congh@google.com>
8514 PR tree-optimization/60000
8515 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
8516 if the vectorized statement is a store. A store statement can only
8517 appear at the end of pattern statements.
8519 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
8521 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
8522 (ix86_option_override_internal): Default long double to 64-bit for
8523 32-bit Bionic and to 128-bit for 64-bit Bionic.
8525 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
8526 TARGET_LONG_DOUBLE_128 is true.
8527 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
8529 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
8530 (mlong-double-64): Negate -mlong-double-128.
8531 (mlong-double-128): New option.
8533 * config/i386/i386-c.c (ix86_target_macros): Define
8534 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
8536 * doc/invoke.texi: Document -mlong-double-128.
8538 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
8540 PR rtl-optimization/60024
8541 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
8543 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
8545 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
8547 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
8549 PR rtl-optimization/57662
8550 * sel-sched.c (code_motion_path_driver): Do not mark already not
8551 existing blocks in the visiting bitmap.
8553 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
8555 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
8556 on the insn being emitted.
8558 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
8559 Will Deacon <will.deacon@arm.com>
8561 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
8563 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8565 * config/arm/arm-tables.opt: Regenerate.
8567 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8569 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
8570 for vector types other than V16QImode.
8571 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
8572 define_expand, and call altivec_expand_vec_perm_le when producing
8573 code with little endian element order.
8574 (*altivec_vperm_<mode>_internal): New insn having previous
8575 behavior of altivec_vperm_<mode>.
8576 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
8577 altivec_expand_vec_perm_le when producing code with little endian
8579 (*altivec_vperm_<mode>_uns_internal): New insn having previous
8580 behavior of altivec_vperm_<mode>_uns.
8582 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8584 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
8585 (altivec_vsumsws): Add handling for -maltivec=be with a little
8587 (altivec_vsumsws_direct): New.
8588 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
8589 gen_altivec_vsumsws.
8591 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
8593 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
8594 vtable_pointer_value_to_binfo): New functions.
8595 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
8596 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
8598 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
8600 * config/nios2/nios2.md (load_got_register): Initialize GOT
8601 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
8602 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
8604 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
8606 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
8607 preserverd by passthrough, do not propagate the type.
8609 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
8611 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
8612 (mips_atomic_assign_expand_fenv): New function.
8613 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
8615 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
8617 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
8618 (__builtin_mips_set_fcsr): Likewise.
8619 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
8620 MIPS_USI_FTYPE_VOID.
8621 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
8622 (mips16_expand_set_fcsr): Likewise.
8623 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
8624 (mips16_set_fcsr_stub): Likewise.
8625 (mips16_get_fcsr_one_only_stub): New class.
8626 (mips16_set_fcsr_one_only_stub): Likewise.
8627 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
8628 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
8629 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
8630 (hard_float): New availability predicate.
8631 (mips_builtins): Add get_fcsr and set_fcsr.
8632 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
8633 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
8634 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
8635 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
8636 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
8639 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
8641 * config/mips/mips.c (mips_one_only_stub): New class.
8642 (mips_need_mips16_rdhwr_p): Replace with...
8643 (mips16_rdhwr_stub): ...this new variable.
8644 (mips16_stub_call_address): New function.
8645 (mips16_rdhwr_one_only_stub): New class.
8646 (mips_expand_thread_pointer): Use mips16_stub_call_address.
8647 (mips_output_mips16_rdhwr): Delete.
8648 (mips_finish_stub): New function.
8649 (mips_code_end): Use it to handle rdhwr stubs.
8651 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
8654 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
8655 when calculating size of integer atomic types.
8657 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
8659 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
8661 2014-02-01 Jakub Jelinek <jakub@redhat.com>
8663 PR tree-optimization/60003
8664 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
8665 * profile.c (branch_prob): Use gimple_call_builtin_p
8666 to check for BUILT_IN_SETJMP_RECEIVER.
8667 * tree-inline.c (copy_bb): Call notice_special_calls.
8669 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
8672 * lra-constraints.c (process_alt_operands): Update reload_sum only
8675 2014-01-31 Richard Henderson <rth@redhat.com>
8678 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
8679 until after else_eh is processed.
8681 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
8683 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
8684 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
8685 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
8686 in smmintrin.h, remove them.
8687 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
8688 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
8689 * config/i386/i386.md (ROUND_SAE): Fix value.
8690 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
8691 (const48_operand): New.
8692 * config/i386/subst.md (round), (round_expand): Use
8693 const_4_or_8_to_11_operand.
8694 (round_saeonly), (round_saeonly_expand): Use const48_operand.
8696 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
8698 * config/i386/constraints.md (Yk): Swap meaning with k.
8699 * config/i386/i386.md (movhi_internal): Change Yk to k.
8700 (movqi_internal): Ditto.
8701 (*k<logic><mode>): Ditto.
8704 (kandn<mode>): Ditto.
8705 (*<code>hi_1): Ditto.
8706 (*<code>qi_1): Ditto.
8707 (kxnor<mode>): Ditto.
8708 (kortestzhi): Ditto.
8709 (kortestchi): Ditto.
8711 (*one_cmplhi2_1): Ditto.
8712 (*one_cmplqi2_1): Ditto.
8713 * config/i386/sse.md (): Change k to Yk.
8714 (avx512f_load<mode>_mask): Ditto.
8715 (avx512f_blendm<mode>): Ditto.
8716 (avx512f_store<mode>_mask): Ditto.
8717 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
8718 (avx512f_storedqu<mode>_mask): Ditto.
8719 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
8721 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
8722 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
8723 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
8724 (avx512f_maskcmp<mode>3): Ditto.
8725 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
8726 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
8727 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
8728 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
8729 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
8730 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
8731 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
8732 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
8733 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
8734 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
8735 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
8736 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
8737 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
8738 (vec_extract_lo_<mode>_maskm): Ditto.
8739 (vec_extract_hi_<mode>_maskm): Ditto.
8740 (avx512f_vternlog<mode>_mask): Ditto.
8741 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
8742 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
8743 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
8744 (avx512f_<code>v8div16qi2_mask): Ditto.
8745 (avx512f_<code>v8div16qi2_mask_store): Ditto.
8746 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
8747 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
8748 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
8749 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
8750 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
8751 (*avx512pf_gatherpf<mode>df_mask): Ditto.
8752 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
8753 (*avx512pf_scatterpf<mode>df_mask): Ditto.
8754 (avx512cd_maskb_vec_dupv8di): Ditto.
8755 (avx512cd_maskw_vec_dupv16si): Ditto.
8756 (avx512f_vpermi2var<mode>3_maskz): Ditto.
8757 (avx512f_vpermi2var<mode>3_mask): Ditto.
8758 (avx512f_vpermi2var<mode>3_mask): Ditto.
8759 (avx512f_vpermt2var<mode>3_maskz): Ditto.
8760 (*avx512f_gathersi<mode>): Ditto.
8761 (*avx512f_gathersi<mode>_2): Ditto.
8762 (*avx512f_gatherdi<mode>): Ditto.
8763 (*avx512f_gatherdi<mode>_2): Ditto.
8764 (*avx512f_scattersi<mode>): Ditto.
8765 (*avx512f_scatterdi<mode>): Ditto.
8766 (avx512f_compress<mode>_mask): Ditto.
8767 (avx512f_compressstore<mode>_mask): Ditto.
8768 (avx512f_expand<mode>_mask): Ditto.
8769 * config/i386/subst.md (mask): Change k to Yk.
8770 (mask_scalar_merge): Ditto.
8773 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
8775 * doc/extend.texi (Vector Extensions): Document ?: in C++.
8777 2014-01-31 Richard Biener <rguenther@suse.de>
8780 * builtins.c (fold_builtin_memory_op): Make sure to not
8781 use a floating-point mode or a boolean or enumeral type for
8784 2014-01-30 DJ Delorie <dj@redhat.com>
8786 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
8787 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
8788 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
8789 whenever main() has an epilogue.
8791 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8793 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
8794 unused variable "field".
8795 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
8796 (vsx_mergeh_<mode>): Likewise.
8797 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
8798 (altivec_vmrghh): Likewise.
8799 (altivec_vmrghw): Likewise.
8800 (altivec_vmrglb): Likewise.
8801 (altivec_vmrglh): Likewise.
8802 (altivec_vmrglw): Likewise.
8803 (altivec_vspltb): Add missing uses.
8804 (altivec_vsplth): Likewise.
8805 (altivec_vspltw): Likewise.
8806 (altivec_vspltsf): Likewise.
8808 2014-01-30 Jakub Jelinek <jakub@redhat.com>
8811 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
8812 frame related instructions.
8814 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
8816 PR rtl-optimization/59959
8817 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
8818 any reload of register whose subreg is invalid.
8820 2014-01-30 Jakub Jelinek <jakub@redhat.com>
8822 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
8823 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
8824 Add missing return type - void.
8826 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8828 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
8829 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
8830 remove element index adjustment for endian (now handled in vsx.md
8832 (altivec_expand_vec_perm_const): Use
8833 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
8834 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
8835 (vsx_xxspltw_<mode>): Adjust element index for little endian.
8836 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
8837 define_expand and a new define_insn *altivec_vspltb_internal;
8838 adjust for -maltivec=be on a little endian target.
8839 (altivec_vspltb_direct): New.
8840 (altivec_vsplth): Divide into a define_expand and a new
8841 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
8842 little endian target.
8843 (altivec_vsplth_direct): New.
8844 (altivec_vspltw): Divide into a define_expand and a new
8845 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
8846 little endian target.
8847 (altivec_vspltw_direct): New.
8848 (altivec_vspltsf): Divide into a define_expand and a new
8849 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
8850 a little endian target.
8852 2014-01-30 Richard Biener <rguenther@suse.de>
8854 PR tree-optimization/59993
8855 * tree-ssa-forwprop.c (associate_pointerplus): Check we
8856 can propagate form the earlier stmt and avoid the transform
8857 when the intermediate result is needed.
8859 2014-01-30 Alangi Derick <alangiderick@gmail.com>
8861 * README.Portability: Fix typo.
8863 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
8865 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
8866 comparison_operator with ordered_comparison_operator.
8868 2014-01-30 Nick Clifton <nickc@redhat.com>
8870 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
8871 Rename to mn10300_store_multiple_regs.
8872 * config/mn10300/mn10300.c: Likewise.
8873 * config/mn10300/mn10300.md (store_movm): Fix typo: call
8874 store_multiple_regs.
8875 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
8876 Call mn10300_store_multiple_regs.
8878 2014-01-30 Nick Clifton <nickc@redhat.com>
8879 DJ Delorie <dj@redhat.com>
8881 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
8882 %fp 2 to keep registers after it properly word-aligned.
8883 (rl78_alloc_physical_registers_umul): Handle the case where both
8884 input operands are the same.
8886 2014-01-30 Richard Biener <rguenther@suse.de>
8888 PR tree-optimization/59903
8889 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
8892 2014-01-30 Jason Merrill <jason@redhat.com>
8895 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
8898 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
8900 2014-01-30 Richard Biener <rguenther@suse.de>
8902 PR tree-optimization/59951
8903 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
8905 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
8908 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
8909 SFmode to DFmode case.
8911 2014-01-29 DJ Delorie <dj@redhat.com>
8913 * config/msp430/msp430.opt (-minrt): New.
8914 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
8916 (ENDFILE_SPEC): Likewise.
8918 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
8920 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
8921 (estimate_function_body_sizes): Use it.
8923 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
8926 * dwarf2out.c (is_cxx_auto): New.
8927 (is_base_type): Use it.
8928 (gen_type_die_with_usage): Likewise.
8930 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8932 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
8933 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
8934 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
8935 -maltivec=be with LE targets.
8936 (vsx_mergeh_<mode>): Likewise.
8937 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
8938 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
8939 (altivec_vmrghb): Replace with define_expand and new
8940 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
8941 (altivec_vmrghb_direct): New define_insn.
8942 (altivec_vmrghh): Replace with define_expand and new
8943 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
8944 (altivec_vmrghh_direct): New define_insn.
8945 (altivec_vmrghw): Replace with define_expand and new
8946 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
8947 (altivec_vmrghw_direct): New define_insn.
8948 (*altivec_vmrghsf): Adjust for endianness.
8949 (altivec_vmrglb): Replace with define_expand and new
8950 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
8951 (altivec_vmrglb_direct): New define_insn.
8952 (altivec_vmrglh): Replace with define_expand and new
8953 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
8954 (altivec_vmrglh_direct): New define_insn.
8955 (altivec_vmrglw): Replace with define_expand and new
8956 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
8957 (altivec_vmrglw_direct): New define_insn.
8958 (*altivec_vmrglsf): Adjust for endianness.
8959 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
8960 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
8961 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
8962 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
8963 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
8964 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
8965 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
8966 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
8968 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
8970 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
8971 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
8974 2014-01-29 Richard Biener <rguenther@suse.de>
8976 PR tree-optimization/58742
8977 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
8978 associate_pointerplus_align.
8979 (associate_pointerplus_diff): New function.
8980 (associate_pointerplus): Likewise. Call associate_pointerplus_align
8981 and associate_pointerplus_diff.
8983 2014-01-29 Richard Biener <rguenther@suse.de>
8985 * lto-streamer.h (LTO_major_version): Bump to 3.
8986 (LTO_minor_version): Reset to 0.
8988 2014-01-29 Renlin Li <Renlin.Li@arm.com>
8990 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
8991 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
8992 (arm_file_start): Generate correct asm header for armv7ve.
8993 * config/arm/bpabi.h: Add multilib support for armv7ve.
8994 * config/arm/driver-arm.c: Change the architectures of cortex-a7
8995 and cortex-a15 to armv7ve.
8996 * config/arm/t-aprofile: Add multilib support for armv7ve.
8997 * doc/invoke.texi: Document -march=armv7ve.
8999 2014-01-29 Richard Biener <rguenther@suse.de>
9001 PR tree-optimization/58742
9002 * tree-ssa-forwprop.c (associate_plusminus): Return true
9003 if we changed sth, defer EH cleanup to ...
9004 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
9005 (simplify_mult): New function.
9007 2014-01-29 Jakub Jelinek <jakub@redhat.com>
9010 PR tree-optimization/59920
9011 * tree.c (build_common_builtin_nodes): Remove
9012 __builtin_setjmp_dispatcher initialization.
9013 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
9014 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
9015 instead of gsi_after_labels + manually skipping debug stmts.
9016 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
9017 ignore bbs with IFN_ABNORMAL_DISPATCHER.
9018 * tree-inline.c (copy_edges_for_bb): Remove
9019 can_make_abnormal_goto argument, instead add abnormal_goto_dest
9020 argument. Ignore computed_goto_p stmts. Don't call
9021 make_abnormal_goto_edges. If a call might need abnormal edges
9022 for non-local gotos, see if it already has an edge to
9023 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
9024 with true argument, don't do anything then, otherwise add
9025 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
9026 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
9028 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
9029 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
9030 (lower_stmt): Don't set data->calls_builtin_setjmp.
9031 (lower_builtin_setjmp): Adjust comment.
9032 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
9033 * tree-cfg.c (found_computed_goto): Remove.
9034 (factor_computed_gotos): Remove.
9035 (make_goto_expr_edges): Return bool, true for computed gotos.
9036 Don't call make_abnormal_goto_edges.
9037 (build_gimple_cfg): Don't set found_computed_goto, don't call
9038 factor_computed_gotos.
9039 (computed_goto_p): No longer static.
9040 (make_blocks): Don't set found_computed_goto.
9041 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
9042 (make_edges): If make_goto_expr_edges returns true, push bb
9043 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
9044 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
9045 vector. Record mapping between bbs and OpenMP regions if there
9046 are any, adjust make_gimple_omp_edges caller. Call
9047 handle_abnormal_edges.
9048 (make_abnormal_goto_edges): Remove.
9049 * tree-cfg.h (make_abnormal_goto_edges): Remove.
9050 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
9051 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
9052 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
9053 * internal-fn.def (ABNORMAL_DISPATCHER): New.
9054 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
9055 filling *region also set *region_idx to (*region)->entry->index.
9058 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
9059 For REGs set ORIGINAL_REGNO.
9061 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
9063 * doc/md.texi: Mention that a target shouldn't implement
9064 vec_widen_(s|u)mul_even/odd pair if it is less efficient
9067 2014-01-29 Jakub Jelinek <jakub@redhat.com>
9069 PR tree-optimization/59594
9070 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
9071 a copy of the datarefs vector rather than the vector itself.
9073 2014-01-28 Jason Merrill <jason@redhat.com>
9076 * dwarf2out.c (auto_die): New static.
9077 (gen_type_die_with_usage): Handle C++1y 'auto'.
9078 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
9081 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
9084 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
9085 (SPEC_X32): Likewise.
9086 (SPEC_64): Likewise.
9087 * config/i386/i386.c (ix86_option_override_internal): Turn off
9088 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
9090 (x86_file_start): Output .code16gcc for TARGET_16BIT.
9091 * config/i386/i386.h (TARGET_16BIT): New macro.
9092 (TARGET_16BIT_P): Likewise.
9093 * config/i386/i386.opt: Add m16.
9094 * doc/invoke.texi: Document -m16.
9096 2014-01-28 Jakub Jelinek <jakub@redhat.com>
9098 PR preprocessor/59935
9099 * input.c (location_get_source_line): Bail out on when line number
9100 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
9102 2014-01-28 Richard Biener <rguenther@suse.de>
9104 PR tree-optimization/58742
9105 * tree-ssa-forwprop.c (associate_plusminus): Handle
9106 pointer subtraction of the form (T)(P + A) - (T)P.
9108 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9110 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
9113 2014-01-28 Richard Biener <rguenther@suse.de>
9116 2014-01-28 Richard Biener <rguenther@suse.de>
9118 PR rtl-optimization/45364
9119 PR rtl-optimization/59890
9120 * var-tracking.c (local_get_addr_clear_given_value): Handle
9121 already cleared slot.
9122 (val_reset): Handle not allocated local_get_addr_cache.
9123 (vt_find_locations): Use post-order on the inverted CFG.
9125 2014-01-28 Richard Biener <rguenther@suse.de>
9127 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
9129 2014-01-28 Richard Biener <rguenther@suse.de>
9131 PR rtl-optimization/45364
9132 PR rtl-optimization/59890
9133 * var-tracking.c (local_get_addr_clear_given_value): Handle
9134 already cleared slot.
9135 (val_reset): Handle not allocated local_get_addr_cache.
9136 (vt_find_locations): Use post-order on the inverted CFG.
9138 2014-01-28 Alan Modra <amodra@gmail.com>
9140 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
9141 * configure.ac <recursive call for build != host>: Define
9142 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
9143 and LD_FOR_BUILD too.
9144 * configure: Regenerate.
9146 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
9148 * config/i386/i386.c (get_builtin_code_for_version): Separate
9149 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
9150 Broadwell from Haswell.
9152 2014-01-27 Steve Ellcey <sellcey@mips.com>
9154 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
9155 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
9156 * config/mips/mips.c (mips_option_override): Change setting
9158 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
9159 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
9160 Change from Mask to Var.
9162 2014-01-27 Jeff Law <law@redhat.com>
9164 * ipa-inline.c (inline_small_functions): Fix typo.
9166 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
9168 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
9169 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
9170 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
9171 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
9172 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
9173 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
9174 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
9175 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
9176 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
9177 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
9178 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
9179 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
9180 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
9181 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
9182 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
9183 (_mm512_storeu_epi64): Ditto.
9184 (_mm512_cmpge_epi32_mask): Ditto.
9185 (_mm512_cmpge_epu32_mask): Ditto.
9186 (_mm512_cmpge_epi64_mask): Ditto.
9187 (_mm512_cmpge_epu64_mask): Ditto.
9188 (_mm512_cmple_epi32_mask): Ditto.
9189 (_mm512_cmple_epu32_mask): Ditto.
9190 (_mm512_cmple_epi64_mask): Ditto.
9191 (_mm512_cmple_epu64_mask): Ditto.
9192 (_mm512_cmplt_epi32_mask): Ditto.
9193 (_mm512_cmplt_epu32_mask): Ditto.
9194 (_mm512_cmplt_epi64_mask): Ditto.
9195 (_mm512_cmplt_epu64_mask): Ditto.
9196 (_mm512_cmpneq_epi32_mask): Ditto.
9197 (_mm512_cmpneq_epu32_mask): Ditto.
9198 (_mm512_cmpneq_epi64_mask): Ditto.
9199 (_mm512_cmpneq_epu64_mask): Ditto.
9200 (_mm512_expand_pd): Ditto.
9201 (_mm512_expand_ps): Ditto.
9202 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
9203 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
9204 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
9205 * config/i386/i386.c (ix86_builtins): Add
9206 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
9207 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
9208 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
9209 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
9210 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
9211 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
9212 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
9213 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
9214 IX86_BUILTIN_PMOVUSQW512_MEM.
9215 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
9216 __builtin_ia32_pmovsqd512mem_mask,
9217 __builtin_ia32_pmovqd512mem_mask,
9218 __builtin_ia32_pmovusqw512mem_mask,
9219 __builtin_ia32_pmovsqw512mem_mask,
9220 __builtin_ia32_pmovqw512mem_mask,
9221 __builtin_ia32_pmovusdw512mem_mask,
9222 __builtin_ia32_pmovsdw512mem_mask,
9223 __builtin_ia32_pmovdw512mem_mask,
9224 __builtin_ia32_pmovqb512mem_mask,
9225 __builtin_ia32_pmovusqb512mem_mask,
9226 __builtin_ia32_pmovsqb512mem_mask,
9227 __builtin_ia32_pmovusdb512mem_mask,
9228 __builtin_ia32_pmovsdb512mem_mask,
9229 __builtin_ia32_pmovdb512mem_mask.
9230 (bdesc_args): Add __builtin_ia32_expanddf512,
9231 __builtin_ia32_expandsf512.
9232 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
9233 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
9234 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
9235 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
9236 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
9237 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
9238 (avx512f_<code>v8div16qi2_mask_store): This.
9239 (avx512f_expand<mode>): New.
9241 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
9243 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
9245 (_mm512_mask_prefetch_i64gather_pd): Ditto.
9246 (_mm512_prefetch_i32scatter_pd): Ditto.
9247 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
9248 (_mm512_prefetch_i64scatter_pd): Ditto.
9249 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
9250 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
9251 (_mm512_mask_prefetch_i64gather_ps): Ditto.
9252 (_mm512_prefetch_i32scatter_ps): Ditto.
9253 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
9254 (_mm512_prefetch_i64scatter_ps): Ditto.
9255 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
9256 * config/i386/i386-builtin-types.def: Define
9257 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
9258 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
9259 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
9260 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
9261 IX86_BUILTIN_SCATTERPFQPD.
9262 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
9263 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
9264 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
9265 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
9266 __builtin_ia32_scatterpfqps.
9267 (ix86_expand_builtin): Expand new built-ins.
9268 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
9269 fix memory access data type.
9270 (*avx512pf_gatherpf<mode>_mask): Ditto.
9271 (*avx512pf_gatherpf<mode>): Ditto.
9272 (avx512pf_scatterpf<mode>): Ditto.
9273 (*avx512pf_scatterpf<mode>_mask): Ditto.
9274 (*avx512pf_scatterpf<mode>): Ditto.
9275 (GATHER_SCATTER_SF_MEM_MODE): New.
9276 (avx512pf_gatherpf<mode>df): Ditto.
9277 (*avx512pf_gatherpf<mode>df_mask): Ditto.
9278 (*avx512pf_scatterpf<mode>df): Ditto.
9280 2014-01-27 Jakub Jelinek <jakub@redhat.com>
9283 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
9284 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
9287 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
9289 * common/config/arm/arm-common.c
9290 (arm_rewrite_mcpu): Handle multiple names.
9292 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
9294 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
9296 * gimple-builder.h (create_gimple_tmp): Delete.
9298 2014-01-27 Christian Bruel <christian.bruel@st.com>
9300 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
9303 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
9305 * config/pa/pa.md (call): Generate indirect long calls to non-local
9306 functions when outputing 32-bit code.
9307 (call_value): Likewise except for special call to buggy powf function.
9309 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
9310 portable runtime and PIC indirect calls.
9311 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
9312 and PIC call sequences. Use ldo instead of blr to set return register
9313 in PIC call sequence.
9315 2014-01-25 Walter Lee <walt@tilera.com>
9317 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
9318 avoid clobbering a live register.
9320 2014-01-25 Walter Lee <walt@tilera.com>
9322 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
9323 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
9324 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
9325 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
9327 2014-01-25 Walter Lee <walt@tilera.com>
9329 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
9330 arguments on even registers.
9331 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
9333 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
9334 (BIGGEST_ALIGNMENT): Ditto.
9335 (BIGGEST_FIELD_ALIGNMENT): Ditto.
9337 2014-01-25 Walter Lee <walt@tilera.com>
9339 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
9340 insns before bundling.
9341 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
9343 2014-01-25 Walter Lee <walt@tilera.com>
9345 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
9346 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
9347 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
9349 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
9351 * config/mips/constraints.md (kl): Delete.
9352 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
9353 define expands, using...
9354 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
9355 instructions for MIPS16.
9356 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
9357 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
9359 2014-01-25 Walter Lee <walt@tilera.com>
9361 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
9365 2014-01-25 Walter Lee <walt@tilera.com>
9367 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
9368 (TARGET_EXPAND_TO_RTL_HOOK): Define.
9370 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
9372 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
9375 2014-01-25 Jakub Jelinek <jakub@redhat.com>
9377 * print-rtl.c (in_call_function_usage): New var.
9378 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
9379 EXPR_LIST mode as mode and not as reg note name.
9382 * cfgloopmanip.c (copy_loop_info): If
9383 loop->warned_aggressive_loop_optimizations, make sure
9384 the flag is set in target loop too.
9386 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
9388 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
9390 * builtins.def: Likewise.
9391 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
9392 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
9393 * ira.c (ira_setup_eliminable_regset): Likewise.
9394 * omp-low.c (gate_expand_omp): Likewise.
9395 (execute_lower_omp): Likewise.
9396 (diagnose_sb_0): Likewise.
9397 (gate_diagnose_omp_blocks): Likewise.
9398 (simd_clone_clauses_extract): Likewise.
9401 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9403 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
9404 correction for little endian...
9405 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
9408 2014-01-24 Jeff Law <law@redhat.com>
9410 PR tree-optimization/59919
9411 * tree-vrp.c (find_assert_locations_1): Do not register asserts
9412 for non-returning calls.
9414 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
9416 * common/config/aarch64/aarch64-common.c
9417 (aarch64_rewrite_mcpu): Handle multiple names.
9418 * config/aarch64/aarch64.h
9419 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
9421 2014-01-24 Dodji Seketeli <dodji@redhat.com>
9423 * input.c (add_file_to_cache_tab): Handle the case where fopen
9426 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
9429 * config/i386/i386.md (pushsf splitter): Get stack adjustment
9430 from push operand if code of push isn't PRE_DEC.
9432 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
9435 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
9436 -mquad-memory-atomic. Update -mquad-memory documentation to say
9437 it is only used for non-atomic loads/stores.
9439 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
9440 -mquad-memory or -mquad-memory-atomic switches.
9442 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
9443 -mquad-memory-atomic to ISA 2.07 support.
9445 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
9446 to separate support of normal quad word memory operations (ldq, stq)
9447 from the atomic quad word memory operations.
9449 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
9450 support to separate non-atomic quad word operations from atomic
9451 quad word operations. Disable non-atomic quad word operations in
9452 little endian mode so that we don't have to swap words after the
9453 load and before the store.
9454 (quad_load_store_p): Add comment about atomic quad word support.
9455 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
9456 options printed with -mdebug=reg.
9458 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
9459 -mquad-memory-atomic as the test for whether we have quad word
9460 atomic instructions.
9461 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
9462 or -mp8-vector are used, allow byte/half-word atomic operations.
9464 * config/rs6000/sync.md (load_lockedti): Insure that the address
9465 is a proper indexed or indirect address for the lqarx instruction.
9466 On little endian systems, swap the hi/lo registers after the lqarx
9468 (load_lockedpti): Use indexed_or_indirect_operand predicate to
9469 insure the address is valid for the lqarx instruction.
9470 (store_conditionalti): Insure that the address is a proper indexed
9471 or indirect address for the stqcrx. instruction. On little endian
9472 systems, swap the hi/lo registers before doing the stqcrx.
9474 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
9475 insure the address is valid for the stqcrx. instruction.
9477 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
9478 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
9479 type of quad memory support is available.
9481 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
9484 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
9485 there is a danger of looping.
9487 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
9489 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
9490 force flag_ira_loop_pressure if set via command line.
9492 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
9494 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
9495 (ashr_simd): New builtin handling DI mode.
9496 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
9497 (aarch64_sshr_simddi): New match pattern.
9498 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
9499 (vshrd_n_s64): Likewise.
9500 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
9502 2014-01-23 Nick Clifton <nickc@redhat.com>
9504 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
9505 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
9506 favour of mcu specific scripts.
9507 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
9510 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
9511 Alex Velenko <Alex.Velenko@arm.com>
9513 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
9514 (vaddv_s16): Likewise.
9515 (vaddv_s32): Likewise.
9516 (vaddv_u8): Likewise.
9517 (vaddv_u16): Likewise.
9518 (vaddv_u32): Likewise.
9519 (vaddvq_s8): Likewise.
9520 (vaddvq_s16): Likewise.
9521 (vaddvq_s32): Likewise.
9522 (vaddvq_s64): Likewise.
9523 (vaddvq_u8): Likewise.
9524 (vaddvq_u16): Likewise.
9525 (vaddvq_u32): Likewise.
9526 (vaddvq_u64): Likewise.
9527 (vaddv_f32): Likewise.
9528 (vaddvq_f32): Likewise.
9529 (vaddvq_f64): Likewise.
9530 (vmaxv_f32): Likewise.
9531 (vmaxv_s8): Likewise.
9532 (vmaxv_s16): Likewise.
9533 (vmaxv_s32): Likewise.
9534 (vmaxv_u8): Likewise.
9535 (vmaxv_u16): Likewise.
9536 (vmaxv_u32): Likewise.
9537 (vmaxvq_f32): Likewise.
9538 (vmaxvq_f64): Likewise.
9539 (vmaxvq_s8): Likewise.
9540 (vmaxvq_s16): Likewise.
9541 (vmaxvq_s32): Likewise.
9542 (vmaxvq_u8): Likewise.
9543 (vmaxvq_u16): Likewise.
9544 (vmaxvq_u32): Likewise.
9545 (vmaxnmv_f32): Likewise.
9546 (vmaxnmvq_f32): Likewise.
9547 (vmaxnmvq_f64): Likewise.
9548 (vminv_f32): Likewise.
9549 (vminv_s8): Likewise.
9550 (vminv_s16): Likewise.
9551 (vminv_s32): Likewise.
9552 (vminv_u8): Likewise.
9553 (vminv_u16): Likewise.
9554 (vminv_u32): Likewise.
9555 (vminvq_f32): Likewise.
9556 (vminvq_f64): Likewise.
9557 (vminvq_s8): Likewise.
9558 (vminvq_s16): Likewise.
9559 (vminvq_s32): Likewise.
9560 (vminvq_u8): Likewise.
9561 (vminvq_u16): Likewise.
9562 (vminvq_u32): Likewise.
9563 (vminnmv_f32): Likewise.
9564 (vminnmvq_f32): Likewise.
9565 (vminnmvq_f64): Likewise.
9567 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
9569 * config/aarch64/aarch64-simd.md
9570 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
9571 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
9572 (*aarch64_mul3_elt<mode>): Likewise.
9573 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
9574 (*aarch64_mul3_elt_to_64v2df): Likewise.
9575 (*aarch64_mla_elt<mode>): Likewise.
9576 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
9577 (*aarch64_mls_elt<mode>): Likewise.
9578 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
9579 (*aarch64_fma4_elt<mode>): Likewise.
9580 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
9581 (*aarch64_fma4_elt_to_64v2df): Likewise.
9582 (*aarch64_fnma4_elt<mode>): Likewise.
9583 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
9584 (*aarch64_fnma4_elt_to_64v2df): Likewise.
9585 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
9586 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
9587 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
9588 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
9589 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
9590 (aarch64_sqdmull_lane<mode>_internal): Likewise.
9591 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
9593 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
9595 * config/aarch64/aarch64-simd.md
9596 (aarch64_be_checked_get_lane<mode>): New define_expand.
9597 * config/aarch64/aarch64-simd-builtins.def
9598 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
9599 New builtin definition.
9600 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
9601 Use new safe be builtin.
9603 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
9605 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
9607 (aarch64_be_st1<mode>): Likewise.
9608 (aarch_ld1<VALL:mode>): Define_expand modified.
9609 (aarch_st1<VALL:mode>): Likewise.
9610 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
9611 (UNSPEC_ST1): Likewise.
9613 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
9615 * config/microblaze/microblaze.md: Add trap insn and attribute
9617 2014-01-23 Dodji Seketeli <dodji@redhat.com>
9619 PR preprocessor/58580
9620 * input.h (location_get_source_line): Take an additional line_size
9622 (void diagnostics_file_cache_fini): Declare new function.
9623 * input.c (struct fcache): New type.
9624 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
9625 New static constants.
9626 (diagnostic_file_cache_init, total_lines_num)
9627 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
9628 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
9629 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
9630 (get_next_line, read_next_line, goto_next_line, read_line_num):
9631 New static function definitions.
9632 (diagnostic_file_cache_fini): New function.
9633 (location_get_source_line): Take an additional output line_len
9634 parameter. Re-write using lookup_or_add_file_to_cache_tab and
9636 * diagnostic.c (diagnostic_finish): Call
9637 diagnostic_file_cache_fini.
9638 (adjust_line): Take an additional input parameter for the length
9639 of the line, rather than calculating it with strlen.
9640 (diagnostic_show_locus): Adjust the use of
9641 location_get_source_line and adjust_line with respect to their new
9642 signature. While displaying a line now, do not stop at the first
9643 null byte. Rather, display the zero byte as a space and keep
9644 going until we reach the size of the line.
9645 * Makefile.in: Add vec.o to OBJS-libcommon
9647 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
9648 Ilya Tocar <ilya.tocar@intel.com>
9650 * config/i386/avx512fintrin.h (_mm512_kmov): New.
9651 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
9652 (__builtin_ia32_kmov16): Ditto.
9653 * config/i386/i386.md (UNSPEC_KMOV): New.
9656 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
9658 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
9659 (_mm512_storeu_si512): Ditto.
9661 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
9664 * rtl.h (get_referenced_operands): Declare.
9665 * recog.c (get_referenced_operands): New function.
9666 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
9667 operands have been referenced when recording LO_SUM references.
9669 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
9671 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
9673 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
9675 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
9676 Enable for generic and recent AMD targets.
9678 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
9680 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
9681 ARG_SIZE note when adjustment was eliminated.
9683 2014-01-22 Jeff Law <law@redhat.com>
9685 PR tree-optimization/59597
9686 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
9687 in file. Accept new argument REGISTERING and use it to modify
9688 dump output appropriately.
9689 (register_jump_thread): Corresponding changes.
9690 (mark_threaded_blocks): Reinstate code to cancel unprofitable
9691 thread paths involving joiner blocks. Add code to dump cancelled
9692 jump threading paths.
9694 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
9696 PR rtl-optimization/59477
9697 * lra-constraints.c (inherit_in_ebb): Process call for living hard
9698 regs. Update reloads_num and potential_reload_hard_regs for all insns.
9700 2014-01-22 Tom Tromey <tromey@redhat.com>
9702 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
9704 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
9706 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
9708 PR rtl-optimization/59896
9709 * lra-constraints.c (process_alt_operands): Check unused note for
9710 matched operands of insn with no output reloads.
9712 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
9714 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
9715 (mips_move_from_gpr_cost): Likewise.
9717 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
9719 PR rtl-optimization/59858
9720 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
9721 ira_class_hard_regs_num.
9722 (process_alt_operands): Increase reject for dying matched operand.
9724 2014-01-21 Jakub Jelinek <jakub@redhat.com>
9727 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
9728 smaller than size, perform several stores or loads and stores
9729 at dst + count - size to store or copy all of size bytes, rather
9730 than just last modesize bytes.
9732 2014-01-20 DJ Delorie <dj@redhat.com>
9734 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
9735 that CLOBBERs are REGs before propogating their values.
9737 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
9740 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
9741 (cgraph_inline_failed_type): New function.
9742 * cgraph.h (DEFCIFCODE): Add type.
9743 (cgraph_inline_failed_type_t): New enum.
9744 (cgraph_inline_failed_type): New prototype.
9745 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
9746 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
9747 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
9748 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
9749 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
9750 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
9751 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
9752 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
9753 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
9754 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
9755 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
9756 OPTIMIZATION_MISMATCH.
9757 * tree-inline.c (expand_call_inline): Emit errors during
9758 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
9760 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
9763 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
9764 mode attribute in insn output.
9766 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
9768 * output.h (output_constant): Delete.
9769 * varasm.c (output_constant): Make private.
9771 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
9773 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
9775 2014-01-20 Jakub Jelinek <jakub@redhat.com>
9778 * tree.h (fold_builtin_strcat): New prototype.
9779 * builtins.c (fold_builtin_strcat): No longer static. Add len
9780 argument, if non-NULL, don't call c_strlen. Optimize
9781 directly into __builtin_memcpy instead of __builtin_strcpy.
9782 (fold_builtin_2): Adjust fold_builtin_strcat caller.
9783 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
9785 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
9787 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
9788 for SImode_address_operand operands, having only a REG argument.
9790 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
9792 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
9793 loader name using mbig-endian.
9794 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
9796 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
9798 * doc/invoke.texi (-march): Clarify documentation for AArch64.
9802 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
9804 * config/aarch64/aarch64-protos.h
9805 (aarch64_cannot_change_mode_class_ptr): Declare.
9806 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
9807 aarch64_cannot_change_mode_class_ptr): New.
9808 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
9809 backend hook aarch64_cannot_change_mode_class.
9811 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
9813 * common/config/aarch64/aarch64-common.c
9814 (aarch64_handle_option): Don't handle any option order logic here.
9815 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
9816 selected_cpu, warn on architecture version mismatch.
9817 (aarch64_override_options): Fix parsing order for option strings.
9819 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9820 Iain Sandoe <iain@codesourcery.com>
9823 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
9824 warning. Amend comment to reflect current functionality.
9826 2014-01-20 Richard Biener <rguenther@suse.de>
9829 * builtins.c (fold_builtin_strcat): Remove case better handled
9830 by tree-ssa-strlen.c.
9832 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
9834 * config/aarch64/aarch64.opt
9835 (mcpu, march, mtune): Make case-insensitive.
9837 2014-01-20 Jakub Jelinek <jakub@redhat.com>
9840 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
9841 if operands[1] is a REG or ZERO_EXTEND of a REG.
9843 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
9845 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
9847 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
9849 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
9850 long non-pic millicode calls.
9852 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9854 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
9856 2014-01-19 Kito Cheng <kito@0xlab.org>
9858 * builtins.c (expand_movstr): Check movstr expand done or fail.
9860 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
9861 H.J. Lu <hongjiu.lu@intel.com>
9864 * config/i386/i386.md (*lea<mode>): Zero-extend return register
9865 to DImode for zero-extended addresses.
9867 2014-01-19 Jakub Jelinek <jakub@redhat.com>
9869 PR rtl-optimization/57763
9870 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
9871 on the new indirect jump_insn and increment LABEL_NUSES (label).
9873 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
9877 * config.gcc (x86_archs): New variable.
9878 (x86_64_archs): Likewise.
9879 (x86_cpus): Likewise.
9880 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
9881 --with-arch/--with-cpu= options.
9882 Support --with-arch=/--with-cpu={nehalem,westmere,
9883 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
9885 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
9887 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
9888 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
9890 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
9892 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
9894 2014-01-18 Jakub Jelinek <jakub@redhat.com>
9897 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
9898 clear cpp_get_options (parse_in)->warn_unused_macros for
9899 ix86_target_macros_internal with cpp_define.
9901 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
9903 * jump.c (delete_related_insns): Keep (use (insn))s.
9904 * reorg.c (redundant_insn): Check for barriers too.
9906 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
9908 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
9910 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
9912 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
9913 call to $$dyncall when TARGET_LONG_CALLS is true.
9915 2014-01-17 Jeff Law <law@redhat.com>
9917 * ree.c (combine_set_extension): Temporarily disable test for
9918 changing number of hard registers.
9920 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
9923 * ipa-inline-analysis.c (inline_free_summary):
9924 Do not free summary of aliases.
9926 2014-01-17 Jakub Jelinek <jakub@redhat.com>
9929 * gimplify.c (gimplify_expr): Use create_tmp_var
9930 instead of create_tmp_var_raw. If cond doesn't have
9931 integral type, don't add the IFN_ANNOTATE builtin at all.
9933 2014-01-17 Martin Jambor <mjambor@suse.cz>
9936 * ipa-cp.c (prev_edge_clone): New variable.
9937 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
9938 Also resize prev_edge_clone vector.
9939 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
9940 (ipcp_edge_removal_hook): New function.
9941 (ipcp_driver): Register ipcp_edge_removal_hook.
9943 2014-01-17 Andrew Pinski <apinski@cavium.com>
9944 Steve Ellcey <sellcey@mips.com>
9947 * config/mips/mips.c (mips_print_operand): Check operand mode instead
9950 2014-01-17 Jeff Law <law@redhat.com>
9953 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
9954 so that pass_ccp runs first.
9956 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
9958 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
9959 (ix86_adjust_cost): Use !TARGET_XXX.
9960 (do_reorder_for_imul): Likewise.
9961 (swap_top_of_ready_list): Likewise.
9962 (ix86_sched_reorder): Likewise.
9964 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
9966 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
9967 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
9968 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
9969 (intel_memset): New. Duplicate slm_memset.
9970 (intel_cost): New. Duplicate slm_cost.
9971 (m_INTEL): New macro.
9972 (processor_target_table): Add "intel".
9973 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
9974 with PROCESSOR_INTEL for "intel".
9975 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
9976 PROCESSOR_SILVERMONT.
9977 (ix86_issue_rate): Likewise.
9978 (ix86_adjust_cost): Likewise.
9979 (ia32_multipass_dfa_lookahead): Likewise.
9980 (swap_top_of_ready_list): Likewise.
9981 (ix86_sched_reorder): Likewise.
9982 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
9983 instead of TARGET_OPT_AGU.
9984 * config/i386/i386.h (TARGET_INTEL): New.
9985 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
9986 (processor_type): Add PROCESSOR_INTEL.
9987 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
9988 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
9990 2014-01-17 Marek Polacek <polacek@redhat.com>
9993 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
9996 2014-01-17 Richard Biener <rguenther@suse.de>
9998 PR tree-optimization/46590
9999 * opts.c (default_options_table): Add entries for
10000 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
10001 all enabled at -O1 but not for -Og.
10002 * common.opt (fbranch-count-reg): Remove Init(1).
10003 (fmove-loop-invariants): Likewise.
10004 (ftree-pta): Likewise.
10006 2014-01-17 Jakub Jelinek <jakub@redhat.com>
10008 * config/i386/i386.c (ix86_data_alignment): For compatibility with
10009 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
10010 decls to at least the GCC 4.8 used alignments.
10013 * tree-nested.c (convert_nonlocal_reference_stmt,
10014 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
10015 of GIMPLE_BIND stmts, adjust associated decls.
10017 2014-01-17 Richard Biener <rguenther@suse.de>
10019 PR tree-optimization/46590
10020 * vec.h (vec<>::bseach): New member function implementing
10021 binary search according to C89 bsearch.
10022 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
10023 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
10024 bitmap pointer again. Make accesses_in_loop a flat array.
10025 (mem_ref_obstack): New global.
10026 (outermost_indep_loop): Adjust for mem_ref->stored changes.
10027 (mark_ref_stored): Likewise.
10028 (ref_indep_loop_p_2): Likewise.
10029 (set_ref_stored_in_loop): New helper function.
10030 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
10031 (memref_free): Adjust.
10032 (record_mem_ref_loc): Simplify.
10033 (gather_mem_refs_stmt): Adjust.
10034 (sort_locs_in_loop_postorder_cmp): New function.
10035 (analyze_memory_references): Sort accesses_in_loop after
10036 loop postorder number.
10037 (find_ref_loc_in_loop_cmp): New function.
10038 (for_all_locs_in_loop): Find relevant cluster of locs in
10039 accesses_in_loop and iterate without recursion.
10040 (execute_sm): Avoid uninit warning.
10041 (struct ref_always_accessed): Simplify.
10042 (ref_always_accessed::operator ()): Likewise.
10043 (ref_always_accessed_p): Likewise.
10044 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
10045 loop postorder numbers here.
10046 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
10049 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
10052 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
10053 on decls for which assemble_alias has been called.
10055 2014-01-17 Nick Clifton <nickc@redhat.com>
10057 * config/msp430/msp430.opt: (mcpu): New option.
10058 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
10059 (msp430_option_override): Parse target_cpu. If the MCU name
10060 matches a generic string, clear target_mcu.
10061 (msp430_attr): Allow numeric interrupt values up to 63.
10062 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
10063 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
10065 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
10067 * config/msp430/msp430.md (popm): Use %J rather than %I.
10068 (addsi3): Use msp430_nonimmediate_operand for operand 2.
10069 (addhi_cy_i): Use immediate_operand for operand 2.
10070 * doc/invoke.texi: Document -mcpu option.
10072 2014-01-17 Richard Biener <rguenther@suse.de>
10074 PR rtl-optimization/38518
10075 * df.h (df_analyze_loop): Declare.
10076 * df-core.c: Include cfgloop.h.
10077 (df_analyze_1): Split out main part of df_analyze.
10078 (df_analyze): Adjust.
10079 (loop_inverted_post_order_compute): New function.
10080 (loop_post_order_compute): Likewise.
10081 (df_analyze_loop): New function avoiding whole-function
10082 postorder computes.
10083 * loop-invariant.c (find_defs): Use df_analyze_loop.
10084 (find_invariants): Adjust.
10085 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
10087 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
10089 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
10090 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
10092 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
10094 * ipa-ref.c (ipa_remove_stmt_references): Fix references
10095 traversal when removing references.
10097 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
10100 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
10102 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
10104 PR middle-end/56791
10105 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
10106 pushing a reload for an autoinc when we had previously reloaded an
10107 inner part of the address.
10109 2014-01-16 Jakub Jelinek <jakub@redhat.com>
10111 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
10113 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
10114 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
10115 when not giving up or versioning for alias only because of
10117 (vect_analyze_data_ref_dependences): Set to true.
10118 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
10120 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
10121 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
10124 PR middle-end/58344
10125 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
10128 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
10129 operand 0 predicate for gathers, use a new pseudo as subtarget.
10131 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
10133 PR middle-end/59609
10134 * lra-constraints.c (process_alt_operands): Add printing debug info.
10135 Check absence of input/output reloads for matched operands too.
10137 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
10139 PR rtl-optimization/59835
10140 * ira.c (ira_init_register_move_cost): Increase cost for
10143 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
10145 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
10147 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
10150 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
10151 non-register objects. Use gen_(high/low)part more consistently.
10154 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
10157 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
10158 endian support, remove tests for WORDS_BIG_ENDIAN.
10159 (p8_mfvsrd_3_<mode>): Likewise.
10160 (reload_gpr_from_vsx<mode>): Likewise.
10161 (reload_gpr_from_vsxsf): Likewise.
10162 (p8_mfvsrd_4_disf): Likewise.
10164 2014-01-16 Richard Biener <rguenther@suse.de>
10166 PR rtl-optimization/46590
10167 * lcm.c (compute_antinout_edge): Use postorder iteration.
10168 (compute_laterin): Use inverted postorder iteration.
10170 2014-01-16 Nick Clifton <nickc@redhat.com>
10172 PR middle-end/28865
10173 * varasm.c (output_constant): Return the number of bytes actually
10175 (output_constructor_array_range): Update the field size with the
10176 number of bytes emitted by output_constant.
10177 (output_constructor_regular_field): Likewise. Also do not
10178 complain if the total number of bytes emitted is now greater
10179 than the expected fieldpos.
10180 * output.h (output_constant): Update prototype and descriptive comment.
10182 2014-01-16 Marek Polacek <polacek@redhat.com>
10184 PR middle-end/59827
10185 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
10186 it is error_mark_node.
10188 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
10190 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
10191 VALID_AVX256_REG_OR_OI_MODE.
10193 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
10195 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
10196 current procedure should be profiled.
10198 2014-01-15 Andrew Pinski <apinski@cavium.com>
10200 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
10201 of moving from/to the STACK_REG register class.
10203 2014-01-15 Richard Henderson <rth@redhat.com>
10206 * reginfo.c (global_regs_decl): Globalize.
10207 * rtl.h (global_regs_decl): Declare.
10208 * ira.c (do_reload): Diagnose frame_pointer_needed and it
10209 reserved via global_regs.
10211 2014-01-15 Teresa Johnson <tejohnson@google.com>
10213 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
10215 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
10217 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
10218 and vmulosh rather than call gen_vec_widen_smult_*.
10219 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
10220 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
10221 (vec_widen_smult_even_v16qi): Likewise.
10222 (vec_widen_umult_even_v8hi): Likewise.
10223 (vec_widen_smult_even_v8hi): Likewise.
10224 (vec_widen_umult_odd_v16qi): Likewise.
10225 (vec_widen_smult_odd_v16qi): Likewise.
10226 (vec_widen_umult_odd_v8hi): Likewise.
10227 (vec_widen_smult_odd_v8hi): Likewise.
10228 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
10229 vmuloub rather than call gen_vec_widen_umult_*.
10230 (vec_widen_umult_lo_v16qi): Likewise.
10231 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
10232 vmulosb rather than call gen_vec_widen_smult_*.
10233 (vec_widen_smult_lo_v16qi): Likewise.
10234 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
10235 rather than call gen_vec_widen_umult_*.
10236 (vec_widen_umult_lo_v8hi): Likewise.
10237 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
10238 rather than call gen_vec_widen_smult_*.
10239 (vec_widen_smult_lo_v8hi): Likewise.
10241 2014-01-15 Jeff Law <law@redhat.com>
10243 PR tree-optimization/59747
10244 * ree.c (find_and_remove_re): Properly handle case where a second
10245 eliminated extension requires widening a copy created for elimination
10246 of a prior extension.
10247 (combine_set_extension): Ensure that the number of hard regs needed
10248 for a destination register does not change when we widen it.
10250 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
10252 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
10253 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
10254 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
10255 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
10256 (avr-*-rtems*): Likewise.
10257 (bfin*-rtems*): Likewise.
10258 (moxie-*-rtems*): Likewise.
10259 (h8300-*-rtems*): Likewise.
10260 (i[34567]86-*-rtems*): Likewise.
10261 (lm32-*-rtems*): Likewise.
10262 (m32r-*-rtems*): Likewise.
10263 (m68k-*-rtems*): Likewise.
10264 (microblaze*-*-rtems*): Likewise.
10265 (mips*-*-rtems*): Likewise.
10266 (powerpc-*-rtems*): Likewise.
10267 (sh-*-rtems*): Likewise.
10268 (sparc-*-rtems*): Likewise.
10269 (sparc64-*-rtems*): Likewise.
10270 (v850-*-rtems*): Likewise.
10271 (m32c-*-rtems*): Likewise.
10273 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
10275 PR rtl-optimization/59511
10276 * ira.c (ira_init_register_move_cost): Use memory costs for some
10277 cases of register move cost calculations.
10278 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
10279 instead of BB frequency.
10280 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
10281 * lra-assigns.c (find_hard_regno_for): Ditto.
10283 2014-01-15 Richard Biener <rguenther@suse.de>
10285 PR tree-optimization/59822
10286 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
10287 (vectorizable_load): Use it to hoist defs of uses of invariant
10288 loads out of the loop.
10290 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
10291 Kugan Vivekanandarajah <kuganv@linaro.org>
10294 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
10297 2014-01-15 Richard Biener <rguenther@suse.de>
10299 PR rtl-optimization/59802
10300 * lcm.c (compute_available): Use inverted postorder to seed
10301 the initial worklist.
10303 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10306 * config/s390/s390.c (s390_preferred_reload_class): Don't return
10307 ADDR_REGS for invalid symrefs in non-PIC code.
10309 2014-01-15 Jakub Jelinek <jakub@redhat.com>
10312 * builtins.c (determine_block_size): Initialize *probable_max_size
10313 even if len_rtx is CONST_INT.
10315 2014-01-14 Andrew Pinski <apinski@cavium.com>
10317 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
10318 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
10319 (cortexa53_tunings): Likewise.
10320 (aarch64_sched_issue_rate): New function.
10321 (TARGET_SCHED_ISSUE_RATE): Define.
10323 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
10325 * ira-costs.c (find_costs_and_classes): Add missed
10326 ira_init_register_move_cost_if_necessary.
10328 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
10331 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
10333 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
10336 * config/i386/i386.c (type_natural_mode): Add a bool parameter
10337 to indicate if type is used for function return value. Warn ABI
10338 change if the vector mode isn't available for function return value.
10339 (ix86_function_arg_advance): Pass false to type_natural_mode.
10340 (ix86_function_arg): Likewise.
10341 (ix86_gimplify_va_arg): Likewise.
10342 (function_arg_32): Don't warn ABI change.
10343 (ix86_function_value): Pass true to type_natural_mode.
10344 (ix86_return_in_memory): Likewise.
10345 (ix86_struct_value_rtx): Removed.
10346 (TARGET_STRUCT_VALUE_RTX): Likewise.
10348 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
10350 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
10351 converting a conditional jump into a conditional return.
10353 2014-01-14 Richard Biener <rguenther@suse.de>
10355 PR tree-optimization/58921
10356 PR tree-optimization/59006
10357 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
10358 hoisting invariant stmts.
10359 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
10360 invariant loads on the preheader edge if possible.
10362 2014-01-14 Joey Ye <joey.ye@arm.com>
10364 * doc/plugin.texi (Building GCC plugins): Update to C++.
10366 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
10368 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
10369 (_mm_rcp28_round_ss): Ditto.
10370 (_mm_rsqrt28_round_sd): Ditto.
10371 (_mm_rsqrt28_round_ss): Ditto.
10372 (_mm_rcp28_sd): Ditto.
10373 (_mm_rcp28_ss): Ditto.
10374 (_mm_rsqrt28_sd): Ditto.
10375 (_mm_rsqrt28_ss): Ditto.
10376 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
10377 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
10378 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
10379 (IX86_BUILTIN_RCP28SD): Ditto.
10380 (IX86_BUILTIN_RCP28SS): Ditto.
10381 (IX86_BUILTIN_RSQRT28SD): Ditto.
10382 (IX86_BUILTIN_RSQRT28SS): Ditto.
10383 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
10384 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
10385 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
10386 (ix86_expand_special_args_builtin): Expand new FTYPE.
10387 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
10388 (srcp14<mode>): Make insn unary.
10389 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
10390 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
10391 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
10392 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
10393 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
10394 Fix rounding: make it SAE only.
10395 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
10397 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
10399 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
10400 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
10401 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
10402 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
10403 (round_saeonly_mask_scalar_operand4): Ditto.
10404 (round_saeonly_mask_scalar_op3): Ditto.
10405 (round_saeonly_mask_scalar_op4): Ditto.
10407 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10409 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
10410 Implement -maltivec=be for vec_insert and vec_extract.
10412 2014-01-10 DJ Delorie <dj@redhat.com>
10414 * config/msp430/msp430.md (call_internal): Don't allow memory
10415 references with SP as the base register.
10416 (call_value_internal): Likewise.
10417 * config/msp430/constraints.md (Yc): New. For memory references
10418 that don't use SP as a base register.
10420 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
10421 "an integer without a # prefix"
10422 * config/msp430/msp430.md (epilogue_helper): Use it.
10424 2014-01-13 Jakub Jelinek <jakub@redhat.com>
10427 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
10428 AVX512F gather builtins.
10429 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
10430 on gather decls with INTEGER_TYPE masktype.
10431 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
10432 directly into the builtin rather than hoisting it before loop.
10434 PR tree-optimization/59387
10435 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
10436 (scev_const_prop): If folded_casts and type has undefined overflow,
10437 use force_gimple_operand instead of force_gimple_operand_gsi and
10438 for each added stmt if it is assign with
10439 arith_code_with_undefined_signed_overflow, call
10440 rewrite_to_defined_overflow.
10441 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
10442 gimple-fold.h instead.
10443 (arith_code_with_undefined_signed_overflow,
10444 rewrite_to_defined_overflow): Moved to ...
10445 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
10446 rewrite_to_defined_overflow): ... here. No longer static.
10447 Include gimplify-me.h.
10448 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
10449 rewrite_to_defined_overflow): New prototypes.
10451 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10453 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
10455 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
10457 * builtins.c (get_object_alignment_2): Minor tweak.
10458 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
10460 2014-01-13 Christian Bruel <christian.bruel@st.com>
10462 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
10463 optimized non constant lengths.
10465 2014-01-13 Jakub Jelinek <jakub@redhat.com>
10468 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
10469 load as __atomic_load_N if possible.
10471 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
10473 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
10475 (rs6000_expand_builtin): Adjust call.
10477 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
10480 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
10481 * config/rs6000/rs6000.c: Include target-globals.h.
10482 (rs6000_set_current_function): Instead of doing target_reinit
10483 unconditionally, use save_target_globals_default_opts and
10484 restore_target_globals.
10486 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
10488 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
10489 (rs6000_expand_builtin): Handle mffs and mtfsf.
10490 (rs6000_init_builtins): Define mffs and mtfsf.
10491 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
10492 (rs6000_mffs): New pattern.
10493 (rs6000_mtfsf): New pattern.
10495 2014-01-11 Bin Cheng <bin.cheng@arm.com>
10497 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
10498 Start narrowing with START. Apply candidate-use pair
10499 and check overall cost in narrowing.
10500 (iv_ca_prune): Pass new argument.
10502 2014-01-10 Jeff Law <law@redhat.com>
10504 PR middle-end/59743
10505 * ree.c (combine_reaching_defs): Ensure the defining statement
10506 occurs before the extension when optimizing extensions with
10507 different source and destination hard registers.
10509 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
10512 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
10513 vtables into the type inheritance graph.
10515 2014-01-10 Jakub Jelinek <jakub@redhat.com>
10517 PR rtl-optimization/59754
10518 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
10519 modes in the REGNO != REGNO case.
10521 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10523 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
10525 2014-01-10 Jakub Jelinek <jakub@redhat.com>
10527 PR tree-optimization/59745
10528 * tree-predcom.c (tree_predictive_commoning_loop): Call
10529 free_affine_expand_cache if giving up because components is NULL.
10531 * target-globals.c (save_target_globals): Allocate < 4KB structs using
10532 GC in payload of target_globals struct instead of allocating them on
10533 the heap and the larger structs separately using GC.
10534 * target-globals.h (struct target_globals): Make regs, hard_regs,
10535 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
10536 of GTY((skip)) and change type to void *.
10537 (reset_target_globals): Cast loads from those fields to corresponding
10540 2014-01-10 Steve Ellcey <sellcey@mips.com>
10543 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
10544 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
10545 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
10547 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
10550 * aarch64-modes.def (CC_Zmode): New flags mode.
10551 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
10552 represents an equality.
10553 (aarch64_get_condition_code): Handle CC_Zmode.
10554 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
10556 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10558 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
10559 extraction in good case.
10561 2014-01-10 Richard Biener <rguenther@suse.de>
10563 PR tree-optimization/59374
10564 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
10565 checking after SLP discovery. Mark stmts not participating
10566 in any SLP instance properly.
10568 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10570 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
10571 when handling a SET rtx.
10573 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10575 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
10576 (cortex-a57): Likewise.
10577 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
10579 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10581 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
10582 non-iwmmxt builtins.
10584 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
10588 * ipa-devirt.c record_target_from_binfo): Take as argument
10589 stack of binfos and lookup matching one for virtual inheritance.
10590 (possible_polymorphic_call_targets_1): Update.
10592 2014-01-10 Huacai Chen <chenhc@lemote.com>
10594 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
10595 kernel strings for Loongson-2E/2F/3A.
10597 2014-01-10 Jakub Jelinek <jakub@redhat.com>
10599 PR middle-end/59670
10600 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
10601 is_gimple_call before calling gimple_call_internal_p.
10603 2014-01-09 Steve Ellcey <sellcey@mips.com>
10605 * Makefile.in (TREE_FLOW_H): Remove.
10606 (TREE_SSA_H): Add file names from tree-flow.h.
10607 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
10608 * tree.h: Remove tree-flow.h reference.
10609 * hash-table.h: Remove tree-flow.h reference.
10610 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
10611 reference with tree-ssa-loop.h.
10613 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10615 * doc/invoke.texi: Add -maltivec={be,le} options, and document
10616 default element-order behavior for -maltivec.
10617 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
10618 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
10619 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
10620 when targeting big endian, at least for now.
10621 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
10623 2014-01-09 Jakub Jelinek <jakub@redhat.com>
10625 PR middle-end/47735
10626 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
10627 var satisfies use_register_for_decl, just take into account type
10628 alignment, rather than decl alignment.
10630 PR tree-optimization/59622
10631 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
10632 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
10633 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
10634 Don't devirtualize for inplace at all. For targets.length () == 1,
10635 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
10637 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
10639 * config/i386/i386.md (cpu): Remove the unused btver1.
10641 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
10643 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
10645 2014-01-09 Jakub Jelinek <jakub@redhat.com>
10648 * tree-core.h (struct target_globals): New forward declaration.
10649 (struct tree_target_option): Add globals field.
10650 * tree.h (TREE_TARGET_GLOBALS): Define.
10651 (prepare_target_option_nodes_for_pch): New prototype.
10652 * target-globals.h (struct target_globals): Define even if
10653 !SWITCHABLE_TARGET.
10654 * tree.c (prepare_target_option_node_for_pch,
10655 prepare_target_option_nodes_for_pch): New functions.
10656 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
10657 * config/i386/i386.c: Include target-globals.h.
10658 (ix86_set_current_function): Instead of doing target_reinit
10659 unconditionally, use save_target_globals_default_opts and
10660 restore_target_globals.
10662 2014-01-09 Richard Biener <rguenther@suse.de>
10664 PR tree-optimization/59715
10665 * tree-cfg.h (split_critical_edges): Declare.
10666 * tree-cfg.c (split_critical_edges): Export.
10667 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
10669 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
10671 * cfgexpand.c (expand_stack_vars): Optionally disable
10672 asan stack protection.
10673 (expand_used_vars): Likewise.
10674 (partition_stack_vars): Likewise.
10675 * asan.c (asan_emit_stack_protection): Optionally disable
10676 after return stack usage.
10677 (instrument_derefs): Optionally disable memory access instrumentation.
10678 (instrument_builtin_call): Likewise.
10679 (instrument_strlen_call): Likewise.
10680 (asan_protect_global): Optionally disable global variables protection.
10681 * doc/invoke.texi: Added doc for new options.
10682 * params.def: Added new options.
10683 * params.h: Likewise.
10685 2014-01-09 Jakub Jelinek <jakub@redhat.com>
10687 PR rtl-optimization/59724
10688 * ifcvt.c (cond_exec_process_if_block): Don't call
10689 flow_find_head_matching_sequence with 0 longest_match.
10690 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
10691 non-active insns if !stop_after.
10692 (try_head_merge_bb): Revert 2014-01-07 changes.
10694 2014-01-08 Jeff Law <law@redhat.com>
10696 * ree.c (get_sub_rtx): New function, extracted from...
10697 (merge_def_and_ext): Here.
10698 (combine_reaching_defs): Use get_sub_rtx.
10700 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
10702 * cgraph.h (varpool_variable_node): Do not choke on null node.
10704 2014-01-08 Catherine Moore <clm@codesourcery.com>
10706 * config/mips/mips.md (simple_return): Attempt to use JRC
10708 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
10710 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
10712 PR rtl-optimization/59137
10713 * reorg.c (steal_delay_list_from_target): Call update_block for
10715 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
10717 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10719 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
10720 two duplicate entries.
10722 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
10725 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
10727 * config/mips/mips.c (mips_truncated_op_cost): New function.
10728 (mips_rtx_costs): Adjust test for BADDU.
10729 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
10731 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
10733 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
10734 (*baddu_si): ...this new pattern.
10736 2014-01-08 Jakub Jelinek <jakub@redhat.com>
10739 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
10741 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
10743 PR middle-end/57748
10744 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
10746 (expand_expr, expand_normal): Adjust.
10747 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
10748 inner_reference_p. Use inner_reference_p to expand inner references.
10749 (store_expr): Adjust.
10750 * cfgexpand.c (expand_call_stmt): Adjust.
10752 2014-01-08 Rong Xu <xur@google.com>
10754 * gcov-io.c (gcov_var): Move from gcov-io.h.
10755 (gcov_position): Ditto.
10756 (gcov_is_error): Ditto.
10757 (gcov_rewrite): Ditto.
10758 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
10759 only part to libgcc/libgcov.h.
10761 2014-01-08 Marek Polacek <polacek@redhat.com>
10763 PR middle-end/59669
10764 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
10766 2014-01-08 Marek Polacek <polacek@redhat.com>
10769 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
10771 2014-01-08 Jakub Jelinek <jakub@redhat.com>
10773 PR rtl-optimization/59649
10774 * stor-layout.c (get_mode_bounds): For BImode return
10775 0 and STORE_FLAG_VALUE.
10777 2014-01-08 Richard Biener <rguenther@suse.de>
10779 PR middle-end/59630
10780 * gimple.h (is_gimple_builtin_call): Remove.
10781 (gimple_builtin_call_types_compatible_p): New.
10782 (gimple_call_builtin_p): New overload.
10783 * gimple.c (is_gimple_builtin_call): Remove.
10784 (validate_call): Rename to ...
10785 (gimple_builtin_call_types_compatible_p): ... this and export. Also
10786 check return types.
10787 (validate_type): New static function.
10788 (gimple_call_builtin_p): New overload and adjust.
10789 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
10790 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
10791 (gimple_fold_stmt_to_constant_1): Likewise.
10792 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
10794 2014-01-08 Richard Biener <rguenther@suse.de>
10796 PR middle-end/59471
10797 * gimplify.c (gimplify_expr): Gimplify register-register type
10798 VIEW_CONVERT_EXPRs to separate stmts.
10800 2014-01-07 Jeff Law <law@redhat.com>
10802 PR middle-end/53623
10803 * ree.c (combine_set_extension): Handle case where source
10804 and destination registers in an extension insn are different.
10805 (combine_reaching_defs): Allow source and destination registers
10806 in extension to be different under limited circumstances.
10807 (add_removable_extension): Remove restriction that the
10808 source and destination registers in the extension are the same.
10809 (find_and_remove_re): Emit a copy from the extension's
10810 destination to its source after the defining insn if
10811 the source and destination registers are different.
10813 PR middle-end/59285
10814 * ifcvt.c (merge_if_block): If we are merging a block with more than
10815 one successor with a block with no successors, remove any BARRIER
10816 after the second block.
10818 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
10820 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
10822 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
10825 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
10826 for 14-bit register offsets when INT14_OK_STRICT is false.
10828 2014-01-07 Roland Stigge <stigge@antcom.de>
10829 Michael Meissner <meissner@linux.vnet.ibm.com>
10832 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
10833 Only check TFmode for SPE constants. Don't check TImode or TDmode.
10835 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
10837 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
10840 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
10842 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
10843 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
10844 rtx is const0_rtx or not.
10846 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
10849 * target-globals.c (save_target_globals): Remove this_fn_optab
10851 * toplev.c: Include optabs.h.
10852 (target_reinit): Temporarily restore the global options if another
10853 set of options are in force.
10855 2014-01-07 Jakub Jelinek <jakub@redhat.com>
10857 PR rtl-optimization/58668
10858 * cfgcleanup.c (flow_find_cross_jump): Don't count
10859 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
10860 to determine what is counted.
10861 (flow_find_head_matching_sequence): Use active_insn_p to determine
10863 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
10865 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
10866 determine what is counted.
10868 PR tree-optimization/59643
10869 * tree-predcom.c (split_data_refs_to_components): If one dr is
10870 read and one write, determine_offset fails and the write isn't
10871 in the bad component, just put the read into the bad component.
10873 2014-01-07 Mike Stump <mikestump@comcast.net>
10874 Jakub Jelinek <jakub@redhat.com>
10877 * tree-core.h (struct tree_optimization_option): Change optabs
10878 type from unsigned char * to void *.
10879 * optabs.c (init_tree_optimization_optabs): Adjust
10880 TREE_OPTIMIZATION_OPTABS initialization.
10882 2014-01-06 Jakub Jelinek <jakub@redhat.com>
10885 * config/i386/i386.h (struct machine_function): Add
10886 no_drap_save_restore field.
10887 * config/i386/i386.c (ix86_save_reg): Use
10888 !cfun->machine->no_drap_save_restore instead of
10889 crtl->stack_realign_needed.
10890 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
10891 this function clears frame_pointer_needed. Set
10892 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
10893 and DRAP reg is needed.
10895 2014-01-06 Marek Polacek <polacek@redhat.com>
10898 * doc/implement-c.texi: Mention that other integer types are
10899 permitted as bit-field types in strictly conforming mode.
10901 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
10903 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
10904 is newly allocated.
10906 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
10908 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
10910 2014-01-06 Martin Jambor <mjambor@suse.cz>
10913 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
10915 * ipa-prop.c (ipa_print_node_params): Fix indentation.
10917 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
10921 * var-tracking.c (add_stores): Preserve the value of the source even if
10922 we don't record the store.
10924 2014-01-06 Terry Guo <terry.guo@arm.com>
10926 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
10928 2014-01-05 Iain Sandoe <iain@codesourcery.com>
10931 * config/darwin.c (darwin_function_section): Adjust return values to
10932 correspond to optimisation changes made in r206070.
10934 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
10936 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
10937 from prefetch_block tune setting.
10938 (nocona_cost): Correct size of prefetch block to 64.
10940 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
10942 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
10943 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
10944 used to save the static chain register in the computation of the offset
10945 from which the FP registers need to be restored.
10947 2014-01-04 Jakub Jelinek <jakub@redhat.com>
10949 PR tree-optimization/59519
10950 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
10951 ICE if get_current_def (current_new_name) is already non-NULL, as long
10952 as it is a phi result of some other phi in *new_exit_bb that has
10955 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
10956 or vmovdqu* for misaligned_operand.
10957 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
10958 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
10959 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
10960 aligned_mem for AVX512F masked aligned load and store builtins and for
10961 non-temporal moves.
10963 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
10965 PR tree-optimization/59651
10966 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
10967 Address range for negative step should be added by TYPE_SIZE_UNIT.
10969 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
10971 * config/m68k/m68k.c (handle_move_double): Handle pushes with
10972 overlapping registers also for registers other than the stack pointer.
10974 2014-01-03 Marek Polacek <polacek@redhat.com>
10977 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
10980 2014-01-03 Jakub Jelinek <jakub@redhat.com>
10983 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
10986 * config/i386/i386.md (MODE_SIZE): New mode attribute.
10987 (push splitter): Use <P:MODE_SIZE> instead of
10988 GET_MODE_SIZE (<P:MODE>mode).
10989 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
10990 (mov -1, reg peephole2): Likewise.
10991 * config/i386/sse.md (*mov<mode>_internal,
10992 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
10993 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
10994 *<code><mode>3, *andnot<mode>3<mask_name>,
10995 <mask_codefor><code><mode>3<mask_name>): Likewise.
10996 * config/i386/subst.md (mask_mode512bit_condition,
10997 sd_mask_mode512bit_condition): Likewise.
10999 2014-01-02 Xinliang David Li <davidxl@google.com>
11001 PR tree-optimization/59303
11002 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
11003 (dump_predicates): Better output format.
11004 (pred_equal_p): New function.
11005 (is_neq_relop_p): Ditto.
11006 (is_neq_zero_form_p): Ditto.
11007 (pred_expr_equal_p): Ditto.
11008 (pred_neg_p): Ditto.
11009 (simplify_pred): Ditto.
11010 (simplify_preds_2): Ditto.
11011 (simplify_preds_3): Ditto.
11012 (simplify_preds_4): Ditto.
11013 (simplify_preds): Ditto.
11014 (push_pred): Ditto.
11015 (push_to_worklist): Ditto.
11016 (get_pred_info_from_cmp): Ditto.
11017 (is_degenerated_phi): Ditto.
11018 (normalize_one_pred_1): Ditto.
11019 (normalize_one_pred): Ditto.
11020 (normalize_one_pred_chain): Ditto.
11021 (normalize_preds): Ditto.
11022 (normalize_cond_1): Remove function.
11023 (normalize_cond): Ditto.
11024 (is_gcond_subset_of): Ditto.
11025 (is_subset_of_any): Ditto.
11026 (is_or_set_subset_of): Ditto.
11027 (is_and_set_subset_of): Ditto.
11028 (is_norm_cond_subset_of): Ditto.
11029 (pred_chain_length_cmp): Ditto.
11030 (convert_control_dep_chain_into_preds): Type change.
11031 (find_predicates): Ditto.
11032 (find_def_preds): Ditto.
11033 (destroy_predicates_vecs): Ditto.
11034 (find_matching_predicates_in_rest_chains): Ditto.
11035 (use_pred_not_overlap_with_undef_path_pred): Ditto.
11036 (is_pred_expr_subset): Ditto.
11037 (is_pred_chain_subset_of): Ditto.
11038 (is_included_in): Ditto.
11039 (is_superset_of): Ditto.
11041 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
11043 Update copyright years.
11045 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
11047 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
11048 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
11049 config/arc/arc.md, config/arc/arc.opt,
11050 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
11051 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
11052 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
11053 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
11054 config/linux-protos.h, config/linux.c, config/winnt-c.c,
11055 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
11056 vtable-verify.c, vtable-verify.h: Use the standard form for the
11059 2014-01-02 Tobias Burnus <burnus@net-b.de>
11061 * gcc.c (process_command): Update copyright notice dates.
11062 * gcov-dump.c: Ditto.
11064 * doc/cpp.texi: Bump @copying's copyright year.
11065 * doc/cppinternals.texi: Ditto.
11066 * doc/gcc.texi: Ditto.
11067 * doc/gccint.texi: Ditto.
11068 * doc/gcov.texi: Ditto.
11069 * doc/install.texi: Ditto.
11070 * doc/invoke.texi: Ditto.
11072 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
11074 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
11076 2014-01-01 Jakub Jelinek <jakub@redhat.com>
11078 * config/i386/sse.md (*mov<mode>_internal): Guard
11079 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
11081 PR rtl-optimization/59647
11082 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
11083 new_rtx into UNSIGNED_FLOAT rtxes.
11085 Copyright (C) 2014 Free Software Foundation, Inc.
11087 Copying and distribution of this file, with or without modification,
11088 are permitted in any medium without royalty provided the copyright
11089 notice and this notice are preserved.