1 2014-05-17 Marc Glisse <marc.glisse@inria.fr>
3 PR tree-optimization/61140
4 PR tree-optimization/61150
5 PR tree-optimization/61197
6 * tree-ssa-phiopt.c (value_replacement): Punt on multiple phis.
8 2014-05-17 Uros Bizjak <ubizjak@gmail.com>
10 * doc/invoke.texi (free): Mention Alpha. Also enabled at -Os.
12 2014-05-17 Richard Sandiford <r.sandiford@uk.ibm.com>
14 * wide-int.cc: Only include longlong.h if W_TYPE_SIZE==32 or
15 __SIZEOF_INT128__ is defined.
17 2014-05-17 Richard Sandiford <rdsandiford@googlemail.com>
19 * config/rs6000/rs6000.c (rs6000_real_tls_symbol_ref_p): New function.
20 (rs6000_delegitimize_address): Use it.
22 2014-05-17 Richard Sandiford <rdsandiford@googlemail.com>
24 * emit-rtl.h (replace_equiv_address, replace_equiv_address_nv): Add an
25 inplace argument. Store the new address in the original MEM when true.
26 * emit-rtl.c (change_address_1): Likewise.
27 (adjust_address_1, adjust_automodify_address_1, offset_address):
29 * rtl.h (plus_constant): Add an inplace argument.
30 * explow.c (plus_constant): Likewise. Try to reuse the original PLUS
31 when true. Avoid generating (plus X (const_int 0)).
32 * function.c (instantiate_virtual_regs_in_rtx): Adjust the PLUS
33 in-place. Pass true to plus_constant.
34 (instantiate_virtual_regs_in_insn): Pass true to replace_equiv_address.
36 2014-05-16 Dehao Chen <dehao@google.com>
38 * tree-cfg.c (gimple_merge_blocks): Updates bb count with max count.
40 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
43 * config/sh/predicates.md (negt_reg_shl31_operand): Match additional
45 * config/sh/sh.md (*negt_msb): Merge SH2A and non-SH2A variants.
47 2014-05-16 Dehao Chen <dehao@google.com>
49 * ira-int.h (REG_FREQ_FROM_EDGE_FREQ): Use
50 optimize_function_for_size_p.
51 * regs.h (REG_FREQ_FROM_BB): Likewise.
53 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
56 * config/sh/sh.c (sh_eval_treg_value): Handle t_reg_operand and
57 negt_reg_operand cases.
58 * config/sh/sh.md (*cset_zero): Likewise by using cbranch_treg_value
60 * config/sh/predicates.md (cbranch_treg_value): Simplify.
62 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
64 * config/sh/sh.c (sh_option_override): Set branch cost to 2 for all
67 2014-05-16 David Malcolm <dmalcolm@redhat.com>
70 2014-04-29 David Malcolm <dmalcolm@redhat.com>
72 * tree-cfg.c (dump_function_to_file): Dump the return type of
73 functions, in a line to itself before the function body, mimicking
74 the layout of a C function.
76 2014-05-16 Dehao Chen <dehao@google.com>
78 * cfghooks.c (make_forwarder_block): Use direct computation to
79 get fall-through edge's count and frequency.
81 2014-05-16 Benno Schulenberg <bensberg@justemail.net>
83 * config/arc/arc.c (arc_init): Fix typo in error message.
84 * config/i386/i386.c (ix86_expand_builtin): Likewise.
85 (split_stack_prologue_scratch_regno): Likewise.
86 * fortran/check.c (gfc_check_fn_rc2008): Remove duplicate
87 word from error message.
89 2014-05-16 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
91 * ira-costs.c: Fix typo in comment.
93 2014-05-16 David Wohlferd <dw@LimeGreenSocks.com>
95 * doc/extend.texi: (Visibility Pragmas) Fix misplaced @xref
97 2014-05-16 Jan Hubicka <hubicka@ucw.cz>
99 * varpool.c (dump_varpool_node): Dump write-only flag.
100 * lto-cgraph.c (lto_output_varpool_node, input_varpool_node): Stream
102 * tree-cfg.c (execute_fixup_cfg): Remove statements setting
103 write-only variables.
104 * ipa.c (process_references): New function.
105 (set_readonly_bit): New function.
106 (set_writeonly_bit): New function.
107 (clear_addressable_bit): New function.
108 (ipa_discover_readonly_nonaddressable_var): Mark write only variables;
109 fix handling of aliases.
110 * cgraph.h (struct varpool_node): Add writeonly flag.
112 2014-05-16 Vladimir Makarov <vmakarov@redhat.com>
114 PR rtl-optimization/60969
115 * ira-costs.c (record_reg_classes): Allow only memory for pseudo.
116 Calculate costs for this case.
118 2014-05-16 Eric Botcazou <ebotcazou@adacore.com>
120 * fold-const (fold_unary_loc) <NON_LVALUE_EXPR>: New case.
121 <CASE_CONVERT>: Pass arg0 instead of op0 to fold_convert_const.
123 2014-05-16 Richard Biener <rguenther@suse.de>
125 PR tree-optimization/61194
126 * tree-vect-patterns.c (adjust_bool_pattern): Also handle
127 bool patterns ending in a COND_EXPR.
129 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
131 * config/aarch64/aarch64.c (aarch64_rtx_mult_cost): Fix FNMUL case.
133 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
135 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle the case
136 where we were unable to cost an RTX.
138 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
140 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost SYMBOL_REF,
143 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
144 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
146 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost TRUNCATE.
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 FMA,
152 FLOAT_EXTEND, FLOAT_TRUNCATE, ABS, SMAX, and SMIN.
154 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
155 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
157 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost comparison
160 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
161 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
163 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costs for
166 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
167 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
169 * config/aarch64/aarch64.c (aarch64_rtx_arith_op_extract_p): New.
170 (aarch64_rtx_costs): Improve costs for SIGN/ZERO_EXTRACT.
172 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
173 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
175 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costs for
178 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
179 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
181 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost
182 ZERO_EXTEND and SIGN_EXTEND better.
184 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
185 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
187 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve cost for
190 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
191 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
193 * config/aarch64/aarch64.c (aarch64_rtx_costs): Use address
194 costs when costing loads and stores to memory.
196 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
197 Philip Tomsich <philipp.tomsich@theobroma-systems.com>
199 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costing
202 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
204 * config/aarch64/aarch64.c (aarch64_rtx_costs): Set default costs.
206 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
207 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
209 * config/aarch64/aarch64.c (aarch64_strip_shift_or_extend): Rename
211 (aarch64_strip_extend): ...this, don't strip shifts, check RTX is
213 (aarch64_rtx_mult_cost): New.
214 (aarch64_rtx_costs): Use it, refactor as appropriate.
216 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
217 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
219 * config/aarch64/aarch64.c (aarch64_build_constant): Conditionally
220 emit instructions, return number of instructions which would
222 (aarch64_add_constant): Update call to aarch64_build_constant.
223 (aarch64_output_mi_thunk): Likewise.
224 (aarch64_rtx_costs): Estimate cost of a CONST_INT, cost of
227 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
229 * config/aarch64/aarch64.c (aarch64_rtx_costs_wrapper): New.
230 (TARGET_RTX_COSTS): Call it.
232 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
234 * config/aarch64/aarch64.c (cortexa57_addrcost_table): New.
235 (cortexa57_vector_cost): Likewise.
236 (cortexa57_tunings): Use them.
238 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
240 * config/aarch64/aarch64-protos.h (scale_addr_mode_cost): New.
241 (cpu_addrcost_table): Use it.
242 * config/aarch64/aarch64.c (generic_addrcost_table): Initialize it.
243 (aarch64_address_cost): Rewrite using aarch64_classify_address,
246 2014-05-16 Richard Biener <rguenther@suse.de>
248 * tree-ssa-sccvn.c: Include tree-cfg.h and domwalk.h.
249 (set_ssa_val_to): Handle unexpected sets to VN_TOP.
250 (visit_phi): Ignore edges marked as not executable.
251 (class cond_dom_walker): New.
252 (cond_dom_walker::before_dom_children): Value-number
253 control statements and mark successor edges as not
254 executable if possible.
255 (run_scc_vn): First walk all control statements in
256 dominator order, marking edges as not executable.
257 * tree-inline.c (copy_edges_for_bb): Be not confused
258 about random edge flags.
260 2014-05-16 Richard Biener <rguenther@suse.de>
262 * tree-ssa-sccvn.c (visit_use): Also constant-fold calls.
264 2014-05-15 Peter Bergner <bergner@vnet.ibm.com>
267 * config/rs6000/htmxlintrin.h (_HTM_TBEGIN_STARTED): New define.
268 (__TM_simple_begin): Use it.
269 (__TM_begin): Likewise.
271 2014-05-15 Martin Jambor <mjambor@suse.cz>
274 * ipa-prop.c (update_indirect_edges_after_inlining): Check
275 type_preserved flag when the indirect edge is polymorphic.
277 2014-05-15 Martin Jambor <mjambor@suse.cz>
279 PR tree-optimization/61090
280 * tree-sra.c (sra_modify_expr): Pass the current gsi to
283 2014-05-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
285 * config/arm/arm.c (arm_option_override): Use the SCHED_PRESSURE_MODEL
286 enum name for PARAM_SCHED_PRESSURE_ALGORITHM.
288 2014-05-15 Jakub Jelinek <jakub@redhat.com>
290 PR tree-optimization/61158
291 * fold-const.c (fold_binary_loc): If X is zero-extended and
292 shiftc >= prec, make sure zerobits is all ones instead of
293 invoking undefined behavior.
295 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
297 * regcprop.h: New file.
298 * regcprop.c (skip_debug_insn_p): New decl.
299 (replace_oldest_value_reg): Check skip_debug_insn_p.
300 (copyprop_hardreg_forward_bb_without_debug_insn): New function.
301 * shrink-wrap.c: Include regcprop.h.
302 (prepare_shrink_wrap): Call
303 copyprop_hardreg_forward_bb_without_debug_insn.
305 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
307 * shrink-wrap.h: Update comment.
308 * shrink-wrap.c: Update comment.
309 (next_block_for_reg): Rename to live_edge_for_reg.
310 (live_edge_for_reg): Allow live_edge->dest has two predecessors.
311 (move_insn_for_shrink_wrap): Split live_edge.
312 (prepre_shrink_wrap): One more parameter for move_insn_for_shrink_wrap.
314 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
316 * config/sparc/sparc-protos.h (sparc_absnegfloat_split_legitimate):
318 * config/sparc/sparc.c (sparc_absnegfloat_split_legitimate): Likewise.
319 * config/sparc/sparc.md (fptype_ut699): New attribute.
320 (in_branch_delay): Return false if -mfix-ut699 is specified and
321 fptype_ut699 is set to single.
322 (truncdfsf2): Add fptype_ut699 attribute.
323 (fix_truncdfsi2): Likewise.
324 (floatsisf2): Change fptype attribute.
325 (fix_truncsfsi2): Likewise.
326 (negtf2_notv9): Delete.
327 (negtf2_v9): Likewise.
328 (negtf2_hq): New instruction.
329 (negtf2): New instruction and splitter.
330 (negdf2_notv9): Rewrite.
331 (abstf2_notv9): Delete.
332 (abstf2_hq_v9): Likewise.
333 (abstf2_v9): Likewise.
334 (abstf2_hq): New instruction.
335 (abstf2): New instruction and splitter.
336 (absdf2_notv9): Rewrite.
338 2014-05-14 Cary Coutant <ccoutant@google.com>
341 * opts.c (common_handle_option): Don't special-case "-g".
342 (set_debug_level): Default to at least level 2 with "-g".
344 2014-05-14 DJ Delorie <dj@redhat.com>
346 * config/msp430/msp430.c (msp430_builtin): Add
347 MSP430_BUILTIN_DELAY_CYCLES.
348 (msp430_init_builtins): Register void __delay_cycles(long long).
349 (msp430_builtin_decl): Add it.
350 (cg_magic_constant): New.
351 (msp430_expand_delay_cycles): New.
352 (msp430_expand_builtin): Call it.
353 (msp430_print_operand_raw): Change integer printing from "int" to
355 * config/msp430/msp430.md (define_constants): Add delay_cycles tags.
356 (delay_cycles_start): New.
357 (delay_cycles_end): New.
358 (delay_cycles_32): New.
359 (delay_cycles_32x): New.
360 (delay_cycles_16): New.
361 (delay_cycles_16x): New.
362 (delay_cycles_2): New.
363 (delay_cycles_1): New.
364 * doc/extend.texi: Document __delay_cycles().
366 2014-05-14 Sandra Loosemore <sandra@codesourcery.com>
368 * config/nios2/nios2.md (nios2_cbranch): Fix paste-o in
369 length attribute computation.
371 2014-05-14 Richard Sandiford <rdsandiford@googlemail.com>
374 * print-rtl.c (print_rtx): Suppress uids if flag_dump_unnumbered.
376 2014-05-14 Richard Sandiford <r.sandiford@uk.ibm.com>
379 * config/sparc/sparc.md: Fix types of low and high in DI constant
380 splitter. Use gen_int_mode in some other splitters.
382 2014-05-14 Martin Jambor <mjambor@suse.cz>
385 * ipa-prop.c (ipa_modify_formal_parameters): Reset DECL_LANG_SPECIFIC.
387 2014-05-14 James Norris <jnorris@codesourcery.com>
389 * omp-low.c (expand_parallel_call): Remove shadow variable.
390 (expand_omp_taskreg): Likewise.
392 2014-05-14 Ilya Tocar <ilya.tocar@intel.com>
394 * common/config/i386/i386-common.c
395 (OPTION_MASK_ISA_CLFLUSHOPT_SET): Define.
396 (OPTION_MASK_ISA_XSAVES_SET): Ditto.
397 (OPTION_MASK_ISA_XSAVEC_SET): Ditto.
398 (OPTION_MASK_ISA_CLFLUSHOPT_UNSET): Ditto.
399 (OPTION_MASK_ISA_XSAVES_UNSET): Ditto.
400 (OPTION_MASK_ISA_XSAVEC_UNSET): Ditto.
401 (ix86_handle_option): Handle OPT_mxsavec, OPT_mxsaves, OPT_mclflushopt.
402 * config.gcc (i[34567]86-*-*): Add clflushoptintrin.h,
403 xsavecintrin.h, xsavesintrin.h.
405 * config/i386/clflushoptintrin.h: New.
406 * config/i386/xsavecintrin.h: Ditto.
407 * config/i386/xsavesintrin.h: Ditto.
408 * config/i386/cpuid.h (bit_CLFLUSHOPT): Define.
411 * config/i386/driver-i386.c (host_detect_local_cpu): Handle
412 -mclflushopt, -mxsavec, -mxsaves, -mno-xsaves, -mno-xsavec,
414 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
415 OPTION_MASK_ISA_CLFLUSHOPT, OPTION_MASK_ISA_XSAVEC,
416 OPTION_MASK_ISA_XSAVES.
417 * config/i386/i386.c (ix86_target_string): Handle -mclflushopt,
419 (PTA_CLFLUSHOPT) Define.
422 (ix86_option_override_internal): Handle new options.
423 (ix86_valid_target_attribute_inner_p): Ditto.
424 (ix86_builtins): Add IX86_BUILTIN_XSAVEC, IX86_BUILTIN_XSAVEC64,
425 IX86_BUILTIN_XSAVES, IX86_BUILTIN_XRSTORS, IX86_BUILTIN_XSAVES64,
426 IX86_BUILTIN_XRSTORS64, IX86_BUILTIN_CLFLUSHOPT.
427 (bdesc_special_args): Add __builtin_ia32_xsaves,
428 __builtin_ia32_xrstors, __builtin_ia32_xsavec, __builtin_ia32_xsaves64,
429 __builtin_ia32_xrstors64, __builtin_ia32_xsavec64.
430 (ix86_init_mmx_sse_builtins): Add __builtin_ia32_clflushopt.
431 (ix86_expand_builtin): Handle new builtins.
432 * config/i386/i386.h (TARGET_CLFLUSHOPT) Define.
433 (TARGET_CLFLUSHOPT_P): Ditto.
434 (TARGET_XSAVEC): Ditto.
435 (TARGET_XSAVEC_P): Ditto.
436 (TARGET_XSAVES): Ditto.
437 (TARGET_XSAVES_P): Ditto.
438 * config/i386/i386.md (ANY_XSAVE): Add UNSPECV_XSAVEC, UNSPECV_XSAVES.
439 (ANY_XSAVE64)" Add UNSPECV_XSAVEC64, UNSPECV_XSAVES64.
440 (attr xsave): Add xsavec, xsavec64, xsaves, xsaves64.
442 (ANY_XRSTOR64): Ditto.
444 (xrstor): Change into <xrstor>.
445 (xrstor_rex64): Change into <xrstor>_rex64.
446 (xrstor64): Change into <xrstor>64
448 * config/i386/i386.opt (mclflushopt): New.
451 * config/i386/x86intrin.h: Add clflushoptintrin.h, xsavesintrin.h,
453 * doc/invoke.texi: Document new options.
455 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
457 PR rtl-optimization/60866
458 * sel-sched-ir (sel_init_new_insn): New parameter old_seqno.
459 Default it to -1. Pass it down to init_simplejump_data.
460 (init_simplejump_data): New parameter old_seqno. Pass it down
461 to get_seqno_for_a_jump.
462 (get_seqno_for_a_jump): New parameter old_seqno. Use it for
463 initializing new jump seqno as a last resort. Add comment.
464 (sel_redirect_edge_and_branch): Save old seqno of the conditional
465 jump and pass it down to sel_init_new_insn.
466 (sel_redirect_edge_and_branch_force): Likewise.
468 2014-05-14 Georg-Johann Lay <avr@gjlay.de>
470 * config/avr/avr.h (REG_CLASS_CONTENTS): Use unsigned suffix for
471 shifted values to avoid build warning.
473 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
475 * cfgcleanup.c (try_forward_edges): Use location_t for locations.
476 * cfgrtl.c (rtl_merge_blocks): Fix comment.
477 (cfg_layout_merge_blocks): Likewise.
478 * except.c (emit_to_new_bb_before): Remove prev_bb local variable.
480 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
482 PR rtl-optimization/60901
483 * config/i386/i386.c (ix86_dependencies_evaluation_hook): Check that
484 bb predecessor belongs to the same scheduling region. Adjust comment.
486 2014-05-13 Peter Bergner <bergner@vnet.ibm.com>
488 * doc/sourcebuild.texi: (dfp_hw): Document.
489 (p8vector_hw): Likewise.
490 (powerpc_eabi_ok): Likewise.
491 (powerpc_elfv2): Likewise.
492 (powerpc_htm_ok): Likewise.
493 (ppc_recip_hw): Likewise.
496 2014-05-13 Cary Coutant <ccoutant@google.com>
498 * opts.c (finish_options): Use -ggnu-pubnames with -gsplit-dwarf.
500 2014-05-13 David Malcolm <dmalcolm@redhat.com>
502 * gengtype-parse.c (require3): Eliminate in favor of...
504 (require_template_declaration): Update to support optional single *
507 * gengtype.c (get_ultimate_base_class): Add a non-const overload.
508 (create_user_defined_type): Handle a single level of explicit
509 pointerness within template arguments.
510 (struct write_types_data): Add field "kind".
511 (filter_type_name): Handle "*" character.
512 (write_user_func_for_structure_ptr): Require a write_types_data
513 rather than just a prefix string, so that we can look up the kind
514 of the wtd and use it as an index into wrote_user_func_for_ptr,
515 ensuring that such functions are written at most once. Support
516 subclasses by invoking the marking function of the ultimate base class.
517 (write_user_func_for_structure_body): Require a write_types_data
518 rather than just a prefix string, so that we can pass this to
519 write_user_func_for_structure_ptr.
520 (write_func_for_structure): Likewise.
521 (ggc_wtd): Add initializer of new "kind" field.
524 * gengtype.h (enum write_types_kinds): New.
525 (struct type): Add field wrote_user_func_for_ptr to the "s"
528 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
530 * fold-const.c (optimize_bit_field_compare): Use wi:: operations
531 instead of const_binop.
532 (fold_binary_loc): Likewise.
534 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
536 * tree-dfa.h (get_addr_base_and_unit_offset_1): Update array index
537 calculation to match get_ref_base_and_extent.
539 2014-05-13 Catherine Moore <clm@codesourcery.com>
540 Sandra Loosemore <sandra@codesourcery.com>
542 * configure.ac: Fix assembly for explicit JALR relocation check.
543 * configure: Regenerate.
545 2014-05-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
547 * config/arm/arm.c (neon_itype): Remove NEON_RESULTPAIR.
548 (arm_init_neon_builtins): Remove handling of NEON_RESULTPAIR.
549 Remove associated type declarations and initialisations.
550 (arm_expand_neon_builtin): Likewise.
551 (neon_emit_pair_result_insn): Delete.
552 * config/arm/arm_neon_builtins (vtrn, vzip, vuzp): Delete.
553 * config/arm/neon.md (neon_vtrn<mode>): Delete.
554 (neon_vzip<mode>): Likewise.
555 (neon_vuzp<mode>): Likewise.
557 2014-05-13 Richard Biener <rguenther@suse.de>
560 * tree-inline.c (remap_gimple_stmt): Clear tail call flag,
561 it needs revisiting whether the call still may be tail-called.
563 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
565 * rtl.def (SYMBOL_REF): Remove middle "0" field.
566 * rtl.h (block_symbol): Reduce number of fields to 2.
567 (rtx_def): Add u2.symbol_ref_flags.
568 (SYMBOL_REF_FLAGS): Use it.
569 (SYMBOL_REF_DATA, SET_SYMBOL_REF_DECL, SYMBOL_REF_DECL)
570 (SET_SYMBOL_REF_CONSTANT, SYMBOL_REF_CONSTANT): Lower index.
571 * gengtype.c (adjust_field_rtx_def): Remove SYMBOL_REF_FLAGS handling.
572 Lower index of SYMBOL_REF_DATA.
573 * print-rtl.c (print_rtx): Lower index for SYMBOL_REF_DATA.
574 Print SYMBOL_REF_FLAGS at the same time.
575 * genattrtab.c (attr_rtx_1): Only initialize 1 "0" SYMBOL_REF field.
577 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
579 * rtl.def (VAR_LOCATION): Remove "i" field.
580 * rtl.h (rtx_def): Add u2.var_location_status.
581 (PAT_VAR_LOCATION_STATUS): Use it.
582 (gen_rtx_VAR_LOCATION): Declare.
583 * gengenrtl.c (excluded_rtx): Add VAR_LOCATION.
584 * emit-rtl.c (gen_rtx_VAR_LOCATION): New function.
585 * var-tracking.c (emit_note_insn_var_location): Remove casts.
587 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
589 * rtl.def (scratch): Fix outdated comment and remove "0" field.
590 * gengtype.c (adjust_field_rtx_def): Update accordingly.
592 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
594 * rtl.def (DEBUG_INSN, INSN, JUMP_INSN, CALL_INSN, JUMP_TABLE_DATA)
595 (BARRIER, CODE_LABEL, NOTE): Remove first "i" field.
596 * rtl.h (rtx_def): Add insn_uid to u2 field.
597 (RTX_FLAG_CHECK8): Delete in favor of...
598 (RTL_INSN_CHAIN_FLAG_CHECK): ...this new macro.
599 (INSN_DELETED_P): Update accordingly.
600 (INSN_UID): Use u2.insn_uid.
601 (INSN_CHAIN_CODE_P): Define.
602 (PREV_INSN, NEXT_INSN, BLOCK_FOR_INSN, PATTERN, INSN_LOCATION)
603 (INSN_CODE, REG_NOTES, CALL_INSN_FUNCTION_USAGE, CODE_LABEL_NUMBER)
604 (NOTE_DATA, NOTE_DELETED_LABEL_NAME, NOTE_BLOCK, NOTE_EH_HANDLER)
605 (NOTE_BASIC_BLOCK, NOTE_VAR_LOCATION, NOTE_CFI, NOTE_LABEL_NUMBER)
606 (NOTE_KIND, LABEL_NAME, LABEL_NUSES, JUMP_LABEL, LABEL_REFS): Lower
608 * print-rtl.c (print_rtx): Print INSN_UIDs before the main loop.
609 Update indices for insn-chain rtxes.
610 * gengtype.c (gen_rtx_next): Adjust test for insn-chain rtxes.
611 (adjust_field_rtx_def): Lower '0' indices for all insn-chain rtxes.
612 * emit-rtl.c (gen_label_rtx): Update gen_rtx_LABEL call.
613 * caller-save.c (init_caller_save): Update gen_rtx_INSN calls.
614 * combine.c (try_combine): Likewise.
615 * ira.c (setup_prohibited_mode_move_regs): Likewise.
617 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
619 * rtl.def (REG): Remove middle field.
620 * rtl.h (rtx_def): Add orignal_regno to u2.
621 (ORIGINAL_REGNO): Use it instead of field 1.
622 (REG_ATTRS): Lower field index accordingly.
623 * gengtype.c (adjust_field_rtx_def): Remove handling of
624 ORIGINAL_REGNO. Move REG_ATTRS index down.
625 * print-rtl.c (print_rtx): Move ORIGINAL_REGNO handling to the
626 code that prints the REGNO.
628 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
630 * print-rtl.c (print_rtx): Guard whole '0' block with ifndef
633 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
635 * rtl.h (rtx_def): Mark u2 as GTY ((skip)).
637 2014-05-13 Bin Cheng <bin.cheng@arm.com>
639 * tree-ssa-loop-ivopts.c (contain_complex_addr_expr): New.
640 (alloc_iv): Lower base expressions containing ADDR_EXPR.
642 2014-05-13 Ian Bolton <ian.bolton@arm.com>
644 * config/aarch64/aarch64-protos.h
645 (aarch64_hard_regno_caller_save_mode): New prototype.
646 * config/aarch64/aarch64.c (aarch64_hard_regno_caller_save_mode):
648 * config/aarch64/aarch64.h (HARD_REGNO_CALLER_SAVE_MODE): New macro.
650 2014-05-13 Christian Bruel <christian.bruel@st.com>
652 * target.def (mode_switching): New hook vector.
653 (mode_emit, mode_needed, mode_after, mode_entry): New hooks.
654 (mode_exit, modepriority_to_mode): Likewise.
655 * mode-switching.c (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Hookify.
656 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
657 * target.h: Include tm.h and hard-reg-set.h.
658 * doc/tm.texi.in (EMIT_MODE_SET, MODE_NEEDED, MODE_AFTER, MODE_ENTRY)
659 (MODE_EXIT, MODE_PRIORITY_TO_MODE): Delete and hookify.
660 * doc/tm.texi Regenerate.
661 * config/sh/sh.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
662 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
663 * config/sh/sh.c (sh_emit_mode_set, sh_mode_priority): Hookify.
664 (sh_mode_needed, sh_mode_after, sh_mode_entry, sh_mode_exit): Likewise.
665 * config/i386/i386.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
666 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
667 * config/i386/i386-protos.h (ix86_mode_needed, ix86_mode_after)
668 (ix86_mode_entrym, ix86_emit_mode_set): Remove external declaration.
669 * config/i386/i386.c (ix86_mode_needed, ix86_mode_after,
670 (ix86_mode_exit, ix86_mode_entry, ix86_mode_priority)
671 (ix86_emit_mode_set): Hookify.
672 * config/epiphany/epiphany.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY):
674 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
675 * config/epiphany/epiphany-protos.h (epiphany_mode_needed)
676 (emit_set_fp_mode, epiphany_mode_entry_exit, epiphany_mode_after)
677 (epiphany_mode_priority_to_mode): Remove declaration.
678 * config/epiphany/epiphany.c (emit_set_fp_mode): Hookify.
679 (epiphany_mode_needed, epiphany_mode_priority_to_mode): Likewise.
680 (epiphany_mode_entry, epiphany_mode_exit, epiphany_mode_after):
682 (epiphany_mode_priority_to_mode): Change priority type. Hookify.
683 (epiphany_mode_needed, epiphany_mode_entry_exit): Hookify.
684 (epiphany_mode_after, epiphany_mode_entry, emit_set_fp_mode): Hookify.
686 2014-05-13 Jakub Jelinek <jakub@redhat.com>
689 * config/i386/i386.c (ix86_expand_set_or_movmem): If count_exp
690 is const0_rtx, return immediately. Don't test count == 0 when
693 2014-05-13 Zhenqiang Chen <zhenqiang.chen@linaro.org>
695 * Makefile.in: add shrink-wrap.o.
696 * config/i386/i386.c: include "shrink-wrap.h"
697 * function.c: Likewise.
698 (requires_stack_frame_p, next_block_for_reg,
699 move_insn_for_shrink_wrap, prepare_shrink_wrap,
700 dup_block_and_redirect): Move to shrink-wrap.c
701 (thread_prologue_and_epilogue_insns): Extract three code segments
702 as functions in shrink-wrap.c
703 * function.h: Move #ifdef HAVE_simple_return ... #endif block to
705 * shrink-wrap.c: New file.
706 * shrink-wrap.h: New file.
708 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
710 * doc/extend.texi: Reflect current numbers of pragmas. Remove
711 reference to Solaris.
713 2014-05-12 Mike Stump <mikestump@comcast.net>
716 * genattrtab.c (filename): Add.
717 (convert_set_attr_alternative): Improve error message.
718 (check_defs): Restore read_md_filename for error messages.
719 (gen_insn): Save filename.
721 2014-05-12 Dimitris Papavasiliou <dpapavas@gmail.com>
723 * doc/invoke.texi: Document new switches -Wno-shadow-ivar,
724 -fno-local-ivars and -fivar-visibility.
725 * c-family/c.opt: Make -Wshadow also implicitly enable
728 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
730 * doc/tm.texi: Remove reference to deleted macro.
731 * doc/tm.texi.in: Likewise.
733 2014-05-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
736 * config/avr/avr.c (avr_out_store_psi): Use correct constant
739 2014-05-12 Georg-Johann Lay <avr@gjlay.de>
741 * config/arm/arm.h (License): Add GCC Runtime Library Exception.
742 * config/arm/aout.h (License): Same.
743 * config/arm/bpabi.h (License): Same.
744 * config/arm/elf.h (License): Same.
745 * config/arm/linux-elf.h (License): Same.
746 * config/arm/linux-gas.h (License): Same.
747 * config/arm/netbsd-elf.h (License): Same.
748 * config/arm/uclinux-eabi.h (License): Same.
749 * config/arm/uclinux-elf.h (License): Same.
750 * config/arm/vxworks.h (License): Same.
752 2014-05-11 Jakub Jelinek <jakub@redhat.com>
754 * tree.h (OMP_CLAUSE_LINEAR_STMT): Define.
755 * tree.c (omp_clause_num_ops): Increase OMP_CLAUSE_LINEAR
756 number of operands to 3.
757 (walk_tree_1): Walk all operands of OMP_CLAUSE_LINEAR.
758 * tree-nested.c (convert_nonlocal_omp_clauses,
759 convert_local_omp_clauses): Handle OMP_CLAUSE_DEPEND.
760 * gimplify.c (gimplify_scan_omp_clauses): Handle
761 OMP_CLAUSE_LINEAR_STMT.
762 * omp-low.c (lower_rec_input_clauses): Fix typo.
763 (maybe_add_implicit_barrier_cancel, lower_omp_1): Add
764 cast between Fortran boolean_type_node and C _Bool if
767 2014-05-11 Richard Sandiford <rdsandiford@googlemail.com>
769 PR tree-optimization/61136
770 * wide-int.h (multiple_of_p): Define a version that doesn't return
772 * fold-const.c (extract_muldiv_1): Use wi::multiple_of_p instead of an
773 integer_zerop/const_binop pair.
774 (multiple_of_p): Likewise, converting both operands to widest_int
777 2014-05-09 Teresa Johnson <tejohnson@google.com>
779 * cgraphunit.c (analyze_functions): Use correct dump file.
781 2014-05-09 Florian Weimer <fweimer@redhat.com>
783 * cfgexpand.c (stack_protect_decl_p): New function, extracted from
785 (stack_protect_return_slot_p): New function.
786 (expand_used_vars): Call stack_protect_decl_p and
787 stack_protect_return_slot_p for -fstack-protector-strong.
789 2014-05-09 Kenneth Zadeck <zadeck@naturalbridge.com>
792 * fold-const.c (fold_binary_loc): Changed width of mask.
794 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
796 * config/avr/avr-fixed.md (round<mode>3): Use -1U instead of -1 in
797 unsigned int initializers for regno_in, regno_out.
799 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
802 * config/avr/avr.md (cc): Add new attribute set_vzn.
803 (addqi3, addqq3, adduqq3, subqi3, subqq3, subuqq3, negqi2) [cc]:
804 Set cc insn attribute to set_vzn instead of set_zn for alternatives
805 with INC, DEC or NEG.
806 * config/avr/avr.c (avr_notice_update_cc): Handle SET_VZN.
807 (avr_out_plus_1): ADIW sets cc0 to CC_SET_CZN.
808 INC, DEC and ADD+ADC set cc0 to CC_CLOBBER.
810 2014-05-09 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
813 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
815 * wide-int.cc (UTItype): Define.
816 (UDWtype): Define for appropriate W_TYPE_SIZE.
818 2014-05-09 Richard Biener <rguenther@suse.de>
820 * Makefile.in (GTFILES): Remove tree-ssa-propagate.c.
821 * tree-ssa-propagate.c: Do not include gt-tree-ssa-propagate.h.
822 (interesting_ssa_edges, varying_ssa_edges): Move out of GC space.
823 (add_ssa_edge, process_ssa_edge_worklist, ssa_prop_init,
824 ssa_propagate): Adjust.
826 2014-05-08 Jeff Law <law@redhat.com>
828 PR tree-optimization/61009
829 * tree-ssa-threadedge.c (thread_through_normal_block): Return a
830 tri-state rather than a boolean. When a block is too big to
831 thread through, inform caller via negative return value.
832 (thread_across_edge): If a block was too big for normal threading,
833 then it's too big for a joiner too, so remove temporary equivalences
834 and return immediately.
836 2014-05-08 Manuel López-Ibáñez <manu@gcc.gnu.org>
837 Matthias Klose <doko@ubuntu.com>
840 * optc-gen.awk: Fix option handling for -Wunused-parameter.
842 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
845 * config/i386/i386.c (PTA_HASWELL): Remove PTA_RTM.
847 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
850 * config/alpha/alpha.c: Include gimple-iterator.h.
851 (alpha_gimple_fold_builtin): New function. Move
852 ALPHA_BUILTIN_UMULH folding from ...
853 (alpha_fold_builtin): ... here.
854 (TARGET_GIMPLE_FOLD_BUILTIN): New define.
856 2014-05-08 Wei Mi <wmi@google.com>
859 * config/i386/i386.c (ix86_compute_frame_layout): Update
860 preferred_stack_boundary for call, expanded from tls descriptor.
861 * config/i386/i386.md (*tls_global_dynamic_32_gnu): Update RTX
862 to depend on SP register.
863 (*tls_local_dynamic_base_32_gnu): Ditto.
864 (*tls_local_dynamic_32_once): Ditto.
865 (tls_global_dynamic_64_<mode>): Set
866 ix86_tls_descriptor_calls_expanded_in_cfun.
867 (tls_local_dynamic_base_64_<mode>): Ditto.
868 (tls_global_dynamic_32): Set
869 ix86_tls_descriptor_calls_expanded_in_cfun. Update RTX
870 to depend on SP register.
871 (tls_local_dynamic_base_32): Ditto.
873 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
875 * config/arm/arm_neon.h: Update comment.
876 * config/arm/neon-docgen.ml: Delete.
877 * config/arm/neon-gen.ml: Delete.
878 * doc/arm-neon-intrinsics.texi: Update comment.
880 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
882 * config/arm/arm_neon_builtins.def (vadd, vsub): Only define the v2sf
884 (vand, vorr, veor, vorn, vbic): Remove.
885 * config/arm/neon.md (neon_vadd, neon_vsub, neon_vadd_unspec): Adjust
887 (neon_vsub_unspec): Likewise.
888 (neon_vorr, neon_vand, neon_vbic, neon_veor, neon_vorn): Remove.
890 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
892 * config/arm/arm_neon.h (vadd_s8): GNU C implementation
893 (vadd_s16): Likewise.
894 (vadd_s32): Likewise.
895 (vadd_f32): Likewise.
897 (vadd_u16): Likewise.
898 (vadd_u32): Likewise.
899 (vadd_s64): Likewise.
900 (vadd_u64): Likewise.
901 (vaddq_s8): Likewise.
902 (vaddq_s16): Likewise.
903 (vaddq_s32): Likewise.
904 (vaddq_s64): Likewise.
905 (vaddq_f32): Likewise.
906 (vaddq_u8): Likewise.
907 (vaddq_u16): Likewise.
908 (vaddq_u32): Likewise.
909 (vaddq_u64): Likewise.
911 (vmul_s16): Likewise.
912 (vmul_s32): Likewise.
913 (vmul_f32): Likewise.
915 (vmul_u16): Likewise.
916 (vmul_u32): Likewise.
918 (vmulq_s8): Likewise.
919 (vmulq_s16): Likewise.
920 (vmulq_s32): Likewise.
921 (vmulq_f32): Likewise.
922 (vmulq_u8): Likewise.
923 (vmulq_u16): Likewise.
924 (vmulq_u32): Likewise.
926 (vsub_s16): Likewise.
927 (vsub_s32): Likewise.
928 (vsub_f32): Likewise.
930 (vsub_u16): Likewise.
931 (vsub_u32): Likewise.
932 (vsub_s64): Likewise.
933 (vsub_u64): Likewise.
934 (vsubq_s8): Likewise.
935 (vsubq_s16): Likewise.
936 (vsubq_s32): Likewise.
937 (vsubq_s64): Likewise.
938 (vsubq_f32): Likewise.
939 (vsubq_u8): Likewise.
940 (vsubq_u16): Likewise.
941 (vsubq_u32): Likewise.
942 (vsubq_u64): Likewise.
944 (vand_s16): Likewise.
945 (vand_s32): Likewise.
947 (vand_u16): Likewise.
948 (vand_u32): Likewise.
949 (vand_s64): Likewise.
950 (vand_u64): Likewise.
951 (vandq_s8): Likewise.
952 (vandq_s16): Likewise.
953 (vandq_s32): Likewise.
954 (vandq_s64): Likewise.
955 (vandq_u8): Likewise.
956 (vandq_u16): Likewise.
957 (vandq_u32): Likewise.
958 (vandq_u64): Likewise.
960 (vorr_s16): Likewise.
961 (vorr_s32): Likewise.
963 (vorr_u16): Likewise.
964 (vorr_u32): Likewise.
965 (vorr_s64): Likewise.
966 (vorr_u64): Likewise.
967 (vorrq_s8): Likewise.
968 (vorrq_s16): Likewise.
969 (vorrq_s32): Likewise.
970 (vorrq_s64): Likewise.
971 (vorrq_u8): Likewise.
972 (vorrq_u16): Likewise.
973 (vorrq_u32): Likewise.
974 (vorrq_u64): Likewise.
976 (veor_s16): Likewise.
977 (veor_s32): Likewise.
979 (veor_u16): Likewise.
980 (veor_u32): Likewise.
981 (veor_s64): Likewise.
982 (veor_u64): Likewise.
983 (veorq_s8): Likewise.
984 (veorq_s16): Likewise.
985 (veorq_s32): Likewise.
986 (veorq_s64): Likewise.
987 (veorq_u8): Likewise.
988 (veorq_u16): Likewise.
989 (veorq_u32): Likewise.
990 (veorq_u64): Likewise.
992 (vbic_s16): Likewise.
993 (vbic_s32): Likewise.
995 (vbic_u16): Likewise.
996 (vbic_u32): Likewise.
997 (vbic_s64): Likewise.
998 (vbic_u64): Likewise.
999 (vbicq_s8): Likewise.
1000 (vbicq_s16): Likewise.
1001 (vbicq_s32): Likewise.
1002 (vbicq_s64): Likewise.
1003 (vbicq_u8): Likewise.
1004 (vbicq_u16): Likewise.
1005 (vbicq_u32): Likewise.
1006 (vbicq_u64): Likewise.
1007 (vorn_s8): Likewise.
1008 (vorn_s16): Likewise.
1009 (vorn_s32): Likewise.
1010 (vorn_u8): Likewise.
1011 (vorn_u16): Likewise.
1012 (vorn_u32): Likewise.
1013 (vorn_s64): Likewise.
1014 (vorn_u64): Likewise.
1015 (vornq_s8): Likewise.
1016 (vornq_s16): Likewise.
1017 (vornq_s32): Likewise.
1018 (vornq_s64): Likewise.
1019 (vornq_u8): Likewise.
1020 (vornq_u16): Likewise.
1021 (vornq_u32): Likewise.
1022 (vornq_u64): Likewise.
1024 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1026 * wide-int.cc (UTItype): Define.
1027 (UDWtype): Define for appropriate W_TYPE_SIZE.
1029 2014-05-08 Marc Glisse <marc.glisse@inria.fr>
1031 PR tree-optimization/59100
1032 * tree-ssa-phiopt.c: Include tree-inline.h.
1033 (neutral_element_p, absorbing_element_p): New functions.
1034 (value_replacement): Handle conditional binary operations with a
1035 neutral or absorbing element.
1037 2014-05-08 Richard Biener <rguenther@suse.de>
1039 * tree-ssa-sccvn.c (vn_get_expr_for): Valueize operands before
1040 folding the expression.
1041 (valueize_expr): Remove.
1042 (visit_reference_op_load): Do not valueize the result of
1044 (simplify_binary_expression): Likewise.
1045 (simplify_unary_expression): Likewise.
1047 2014-05-08 Richard Biener <rguenther@suse.de>
1049 * gimplify.c (gimplify_call_expr): Use saved fnptrtype for
1050 looking at TYPE_ARG_TYPES.
1052 2014-05-08 Richard Biener <rguenther@suse.de>
1054 * gimple-fold.c (gimple_fold_stmt_to_constant_1): Remove
1055 pointer propagation special-case.
1057 2014-05-08 Bin Cheng <bin.cheng@arm.com>
1059 * tree-affine.c (tree_to_aff_combination): Handle MEM_REF for
1060 core part of address expressions.
1062 2014-05-08 Alan Modra <amodra@gmail.com>
1065 * config/rs6000/rs6000.c (expand_block_move): Allow 64-bit
1066 loads and stores when -mno-strict-align at any alignment.
1067 (expand_block_clear): Similarly. Also correct calculation of
1070 2014-05-07 Thomas Preud'homme <thomas.preudhomme@arm.com>
1073 * tree-complex.c (expand_complex_move): Keep line info when expanding
1075 * tree-ssa-uninit.c (warn_uninit): New argument. Ignore assignment
1076 of complex expression. Use new argument to display correct location
1077 for values coming from phi statement.
1078 (warn_uninitialized_vars): Adapt to new signature of warn_uninit.
1079 (warn_uninitialized_phi): Pass location of phi argument to
1081 * tree-ssa.c (ssa_undefined_value_p): For SSA_NAME initialized by a
1082 COMPLEX_EXPR, recurse on each part of the COMPLEX_EXPR.
1084 2014-05-07 Segher Boessenkool <segher@kernel.crashing.org>
1086 * config/rs6000/predicates.md (indexed_address_mem): New.
1087 * config/rs6000/rs6000.md (type): Remove load_ext, load_ext_u,
1088 load_ext_ux, load_ux, load_u, store_ux, store_u, fpload_ux, fpload_u,
1089 fpstore_ux, fpstore_u.
1090 (sign_extend, indexed, update): New.
1091 (cell_micro): Adjust.
1092 (*zero_extend<mode>di2_internal1, *zero_extendsidi2_lfiwzx,
1093 *extendsidi2_lfiwax, *extendsidi2_nocell, *extendsfdf2_fpr,
1094 *movsi_internal1, *movsi_internal1_single, *movhi_internal,
1095 *movqi_internal, *movcc_internal1, mov<mode>_hardfloat,
1096 *mov<mode>_softfloat, *mov<mode>_hardfloat32, *mov<mode>_hardfloat64,
1097 *mov<mode>_softfloat64, *movdi_internal32, *movdi_internal64,
1098 *mov<mode>_string, *ldmsi8, *ldmsi7, *ldmsi6, *ldmsi5, *ldmsi4,
1099 *ldmsi3, *stmsi8, *stmsi7, *stmsi6, *stmsi5, *stmsi4, *stmsi3,
1100 *movdi_update1, movdi_<mode>_update, movdi_<mode>_update_stack,
1101 *movsi_update1, *movsi_update2, movsi_update, movsi_update_stack,
1102 *movhi_update1, *movhi_update2, *movhi_update3, *movhi_update4,
1103 *movqi_update1, *movqi_update2, *movqi_update3, *movsf_update1,
1104 *movsf_update2, *movsf_update3, *movsf_update4, *movdf_update1,
1105 *movdf_update2, load_toc_aix_si, load_toc_aix_di, probe_stack_<mode>,
1106 *stmw, *lmw, as well as 10 anonymous patterns): Adjust.
1108 * config/rs6000/dfp.md (movsd_store, movsd_load): Adjust.
1109 * config/rs6000/vsx.md (*vsx_movti_32bit, *vsx_extract_<mode>_load,
1110 *vsx_extract_<mode>_store): Adjust.
1111 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_microcoded_insn,
1112 is_cracked_insn, insn_must_be_first_in_group,
1113 insn_must_be_last_in_group): Adjust.
1115 * config/rs6000/40x.md (ppc403-load, ppc403-store, ppc405-float):
1117 * config/rs6000/440.md (ppc440-load, ppc440-store, ppc440-fpload,
1118 ppc440-fpstore): Adjust.
1119 * config/rs6000/476.md (ppc476-load, ppc476-store, ppc476-fpload,
1120 ppc476-fpstore): Adjust.
1121 * config/rs6000/601.md (ppc601-load, ppc601-store, ppc601-fpload,
1122 ppc601-fpstore): Adjust.
1123 * config/rs6000/603.md (ppc603-load, ppc603-store, ppc603-fpload):
1125 * config/rs6000/6xx.md (ppc604-load, ppc604-store, ppc604-fpload):
1127 * config/rs6000/7450.md (ppc7450-load, ppc7450-store, ppc7450-fpload,
1128 ppc7450-fpstore): Adjust.
1129 * config/rs6000/7xx.md (ppc750-load, ppc750-store): Adjust.
1130 * config/rs6000/8540.md (ppc8540_load, ppc8540_store): Adjust.
1131 * config/rs6000/a2.md (ppca2-load, ppca2-fp-load, ppca2-fp-store):
1133 * config/rs6000/cell.md (cell-load, cell-load-ux, cell-load-ext,
1134 cell-fpload, cell-fpload-update, cell-store, cell-store-update,
1135 cell-fpstore, cell-fpstore-update): Adjust.
1136 * config/rs6000/e300c2c3.md (ppce300c3_load, ppce300c3_fpload,
1137 ppce300c3_store, ppce300c3_fpstore): Adjust.
1138 * config/rs6000/e500mc.md (e500mc_load, e500mc_fpload, e500mc_store,
1139 e500mc_fpstore): Adjust.
1140 * config/rs6000/e500mc64.md (e500mc64_load, e500mc64_fpload,
1141 e500mc64_store, e500mc64_fpstore): Adjust.
1142 * config/rs6000/e5500.md (e5500_load, e5500_fpload, e5500_store,
1143 e5500_fpstore): Adjust.
1144 * config/rs6000/e6500.md (e6500_load, e6500_fpload, e6500_store,
1145 e6500_fpstore): Adjust.
1146 * config/rs6000/mpc.md (mpccore-load, mpccore-store, mpccore-fpload):
1148 * config/rs6000/power4.md (power4-load, power4-load-ext,
1149 power4-load-ext-update, power4-load-ext-update-indexed,
1150 power4-load-update-indexed, power4-load-update, power4-fpload,
1151 power4-fpload-update, power4-store, power4-store-update,
1152 power4-store-update-indexed, power4-fpstore, power4-fpstore-update):
1154 * config/rs6000/power5.md (power5-load, power5-load-ext,
1155 power5-load-ext-update, power5-load-ext-update-indexed,
1156 power5-load-update-indexed, power5-load-update, power5-fpload,
1157 power5-fpload-update, power5-store, power5-store-update,
1158 power5-store-update-indexed, power5-fpstore, power5-fpstore-update):
1160 * config/rs6000/power6.md (power6-load, power6-load-ext,
1161 power6-load-update, power6-load-update-indexed,
1162 power6-load-ext-update, power6-load-ext-update-indexed, power6-fpload,
1163 power6-fpload-update, power6-store, power6-store-update,
1164 power6-store-update-indexed, power6-fpstore, power6-fpstore-update):
1166 * config/rs6000/power7.md (power7-load, power7-load-ext,
1167 power7-load-update, power7-load-update-indexed,
1168 power7-load-ext-update, power7-load-ext-update-indexed, power7-fpload,
1169 power7-fpload-update, power7-store, power7-store-update,
1170 power7-store-update-indexed, power7-fpstore, power7-fpstore-update):
1172 * config/rs6000/power8.md (power8-load, power8-load-update,
1173 power8-load-ext, power8-load-ext-update, power8-fpload,
1174 power8-fpload-update, power8-store, power8-store-update-indexed,
1175 power8-fpstore, power8-fpstore-update): Adjust.
1176 * config/rs6000/rs64.md (rs64a-load, rs64a-store, rs64a-fpload):
1178 * config/rs6000/titan.md (titan_lsu_load, titan_lsu_fpload,
1179 titan_lsu_store, titan_lsu_fpstore): Adjust.
1180 * config/rs6000/xfpu.md (fp-load, fp-store): Adjust.
1182 2014-05-07 Oleg Endo <olegendo@gcc.gnu.org>
1185 * config/sh/sh-mem.cc (sh_expand_strlen): Use loop when emitting
1186 unrolled byte insns. Emit address increments after move insns.
1188 2014-05-07 David Malcolm <dmalcolm@redhat.com>
1190 * gimple.h (gimple_builtin_call_types_compatible_p): Accept a
1191 const_gimple, rather than a gimple.
1192 (gimple_call_builtin_p): Likewise, for the three variants.
1194 * gimple.c (gimple_builtin_call_types_compatible_p): Likewise.
1195 (gimple_call_builtin_p): Likewise, for the three variants.
1197 2014-05-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
1199 PR tree-optimization/61095
1200 * tree-ssanames.c (get_nonzero_bits): Fix type extension in wi::shwi.
1202 2014-05-07 Richard Biener <rguenther@suse.de>
1204 PR tree-optimization/61034
1205 * tree-ssa-alias.c (call_may_clobber_ref_p_1): Export.
1206 (maybe_skip_until): Use translate to take into account
1207 lattices when trying to do disambiguations.
1208 (get_continuation_for_phi_1): Likewise.
1209 (get_continuation_for_phi): Adjust for added translate arguments.
1210 (walk_non_aliased_vuses): Likewise.
1211 * tree-ssa-alias.h (get_continuation_for_phi): Adjust prototype.
1212 (walk_non_aliased_vuses): Likewise.
1213 (call_may_clobber_ref_p_1): Declare.
1214 * tree-ssa-sccvn.c (vn_reference_lookup_3): Also disambiguate against
1215 calls. Stop early if we are only supposed to disambiguate.
1216 * tree-ssa-pre.c (translate_vuse_through_block): Adjust.
1218 2014-05-07 Joern Rennecke <joern.rennecke@embecosm.com>
1220 * config/epiphany/epiphany.c (epiphany_handle_interrupt_attribute):
1221 Emit an error when the function has arguments.
1223 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
1225 * cfgloop.h (unswitch_loops): Remove.
1226 * doc/passes.texi: Remove references to loop-unswitch.c
1227 * timevar.def (TV_LOOP_UNSWITCH): Remove.
1229 2014-05-07 Evgeny Stupachenko <evstupac@gmail.com>
1231 * tree-vect-data-refs.c (vect_grouped_load_supported): New
1232 check for loads group of length 3.
1233 (vect_permute_load_chain): New permutations for loads group of
1235 * tree-vect-stmts.c (vect_model_load_cost): Change cost
1236 of vec_perm_shuffle for the new permutations.
1238 2014-05-07 Alan Lawrence <alan.lawrence@arm.com>
1240 * config/aarch64/arm_neon.h (vtrn1_f32, vtrn1_p8, vtrn1_p16, vtrn1_s8,
1241 vtrn1_s16, vtrn1_s32, vtrn1_u8, vtrn1_u16, vtrn1_u32, vtrn1q_f32,
1242 vtrn1q_f64, vtrn1q_p8, vtrn1q_p16, vtrn1q_s8, vtrn1q_s16, vtrn1q_s32,
1243 vtrn1q_s64, vtrn1q_u8, vtrn1q_u16, vtrn1q_u32, vtrn1q_u64, vtrn2_f32,
1244 vtrn2_p8, vtrn2_p16, vtrn2_s8, vtrn2_s16, vtrn2_s32, vtrn2_u8,
1245 vtrn2_u16, vtrn2_u32, vtrn2q_f32, vtrn2q_f64, vtrn2q_p8, vtrn2q_p16,
1246 vtrn2q_s8, vtrn2q_s16, vtrn2q_s32, vtrn2q_s64, vtrn2q_u8, vtrn2q_u16,
1247 vtrn2q_u32, vtrn2q_u64): Replace temporary asm with __builtin_shuffle.
1249 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
1251 * loop-unswitch.c: Delete.
1253 2014-05-07 Richard Biener <rguenther@suse.de>
1255 * config.gcc: Always set need_64bit_hwint to yes.
1257 2014-05-07 Chung-Ju Wu <jasonwucj@gmail.com>
1259 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Have it in favor
1260 of using optimize_size.
1262 2014-05-06 Mike Stump <mikestump@comcast.net>
1264 * wide-int.h (wi::int_traits <HOST_WIDE_INT>): Always define.
1266 2014-05-06 Joseph Myers <joseph@codesourcery.com>
1268 * config/i386/sse.md (*mov<mode>_internal)
1269 (*<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>)
1270 (*<sse2_avx_avx512f>_loaddqu<mode><mask_name>)
1271 (<sse>_andnot<mode>3, <code><mode>3, *andnot<mode>3)
1272 (*<code><mode>3, *andnot<mode>3<mask_name>)
1273 (<mask_codefor><code><mode>3<mask_name>): Only consider
1274 TARGET_SSE_PACKED_SINGLE_INSN_OPTIMAL for modes of size 16.
1276 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
1279 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
1281 * lra-constraints.c (valid_address_p): Move earlier in file.
1282 Add a constraint argument to the address_info version.
1283 (satisfies_memory_constraint_p): New function.
1284 (satisfies_address_constraint_p): Likewise.
1285 (process_alt_operands, curr_insn_transform): Use them.
1286 (process_address): Pass the constraint to valid_address_p when
1287 checking address operands.
1289 2014-05-06 Richard Sandiford <r.sandiford@uk.ibm.com>
1291 * lto-cgraph.c (compute_ltrans_boundary): Make node variables local
1292 to their respective blocks. Fix inadvertent use of "node".
1294 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
1296 * emit-rtl.c (init_derived_machine_modes): New functionm, split
1298 (init_emit_once): ...here.
1299 * rtl.h (init_derived_machine_modes): Declare.
1300 * toplev.c (do_compile): Call it even if no_backend.
1302 2014-05-06 Kenneth Zadeck <zadeck@naturalbridge.com>
1303 Mike Stump <mikestump@comcast.net>
1304 Richard Sandiford <rdsandiford@googlemail.com>
1305 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1307 * alias.c (ao_ref_from_mem): Use wide-int interfaces.
1308 (rtx_equal_for_memref_p): Update comment.
1309 (adjust_offset_for_component_ref): Use wide-int interfaces.
1310 * builtins.c (get_object_alignment_2): Likewise.
1311 (c_readstr): Likewise.
1312 (target_char_cast): Add comment.
1313 (determine_block_size): Use wide-int interfaces.
1314 (expand_builtin_signbit): Likewise.
1315 (fold_builtin_int_roundingfn): Likewise.
1316 (fold_builtin_bitop): Likewise.
1317 (fold_builtin_bswap): Likewise.
1318 (fold_builtin_logarithm): Use signop.
1319 (fold_builtin_pow): Likewise.
1320 (fold_builtin_memory_op): Use wide-int interfaces.
1321 (fold_builtin_object_size): Likewise.
1322 * cfgloop.c (alloc_loop): Initialize nb_iterations_upper_bound and
1323 nb_iterations_estimate.
1324 (record_niter_bound): Use wide-int interfaces.
1325 (get_estimated_loop_iterations_int): Likewise.
1326 (get_estimated_loop_iterations): Likewise.
1327 (get_max_loop_iterations): Likewise.
1328 * cfgloop.h: Include wide-int.h.
1329 (struct nb_iter_bound): Change bound to widest_int.
1330 (struct loop): Change nb_iterations_upper_bound and
1331 nb_iterations_estimate to widest_int.
1332 (record_niter_bound): Switch to use widest_int.
1333 (get_estimated_loop_iterations): Likewise.
1334 (get_max_loop_iterations): Likewise.
1335 (gcov_type_to_double_int): Rename to gcov_type_to_wide_int and
1336 update for wide-int.
1337 * cgraph.c (cgraph_add_thunk): Use wide-int interfaces.
1338 * combine.c (try_combine): Likewise.
1339 (subst): Use CONST_SCALAR_INT_P rather than CONST_INT_P.
1340 * config/aarch64/aarch64.c (aapcs_vfp_sub_candidate): Use wide-int
1342 (aarch64_float_const_representable_p): Likewise.
1343 * config/arc/arc.c: Include wide-int.h.
1344 (arc_can_use_doloop_p): Use wide-int interfaces.
1345 * config/arm/arm.c (aapcs_vfp_sub_candidate): Likewise.
1346 (vfp3_const_double_index): Likewise.
1347 * config/avr/avr.c (avr_out_round): Likewise.
1348 (avr_fold_builtin): Likewise.
1349 * config/bfin/bfin.c (bfin_local_alignment): Likewise.
1350 (bfin_can_use_doloop_p): Likewise.
1351 * config/darwin.c (darwin_mergeable_constant_section): Likewise.
1352 (machopic_select_rtx_section): Update to handle CONST_WIDE_INT.
1353 * config/i386/i386.c: Include wide-int.h.
1354 (ix86_data_alignment): Use wide-int interfaces.
1355 (ix86_local_alignment): Likewise.
1356 (ix86_emit_swsqrtsf): Update real_from_integer.
1357 * config/msp430/msp430.c (msp430_attr): Use wide-int interfaces.
1358 * config/nds32/nds32.c (nds32_insert_attributes): Likewise.
1359 * config/rs6000/predicates.md (any_operand): Add const_wide_int.
1360 (zero_constant): Likewise.
1361 (input_operand): Likewise.
1362 (splat_input_operand): Likewise.
1363 (non_logical_cint_operand): Change const_double to const_wide_int.
1364 * config/rs6000/rs6000.c (num_insns_constant): Handle CONST_WIDE_INT.
1365 (easy_altivec_constant): Remove comment.
1366 (paired_expand_vector_init): Use CONSTANT_P.
1367 (rs6000_legitimize_address): Handle CONST_WIDE_INT.
1368 (rs6000_emit_move): Update checks.
1369 (rs6000_aggregate_candidate): Use wide-int interfaces.
1370 (rs6000_expand_ternop_builtin): Likewise.
1371 (rs6000_output_move_128bit): Handle CONST_WIDE_INT.
1372 (rs6000_assemble_integer): Likewise.
1373 (rs6000_hash_constant): Likewise.
1374 (output_toc): Likewise.
1375 (rs6000_rtx_costs): Likewise.
1376 (rs6000_emit_swrsqrt); Update call to real_from_integer.
1377 * config/rs6000/rs6000-c.c: Include wide-int.h.
1378 (altivec_resolve_overloaded_builtin): Use wide-int interfaces.
1379 * config/rs6000/rs6000.h (TARGET_SUPPORTS_WIDE_INT): New.
1380 * config/rs6000/rs6000.md: Use const_scalar_int_operand.
1381 Handle CONST_WIDE_INT.
1382 * config/sol2-c.c (solaris_pragma_align): Change low to unsigned HWI.
1383 Use tree_fits_uhwi_p.
1384 * config/sparc/sparc.c: Include wide-int.h.
1385 (sparc_fold_builtin): Use wide-int interfaces.
1386 * config/vax/vax.c: Include wide-int.h.
1387 (vax_float_literal): Use real_from_integer.
1388 * coretypes.h (struct hwivec_def): New.
1390 (const_hwivec): New.
1391 * cse.c (hash_rtx_cb): Handle CONST_WIDE_INT.
1392 (equiv_constant): Handle CONST_WIDE_INT.
1393 * cselib.c (rtx_equal_for_cselib_1): Use CASE_CONST_UNIQUE.
1394 (cselib_hash_rtx): Handle CONST_WIDE_INT.
1395 * dbxout.c (stabstr_U): Use wide-int interfaces.
1396 (dbxout_type): Update to use cst_fits_shwi_p.
1397 * defaults.h (LOG2_BITS_PER_UNIT): Define.
1398 (TARGET_SUPPORTS_WIDE_INT): Add default.
1399 * dfp.c: Include wide-int.h.
1400 (decimal_real_to_integer2): Use wide-int interfaces and rename to
1401 decimal_real_to_integer.
1402 * dfp.h (decimal_real_to_integer2): Return a wide_int and rename to
1403 decimal_real_to_integer.
1404 * doc/generic.texi (Constant expressions): Update for wide_int.
1405 * doc/rtl.texi (const_double): Likewise.
1406 (const_wide_int, CONST_WIDE_INT, CONST_WIDE_INT_VEC): New.
1407 (CONST_WIDE_INT_NUNITS, CONST_WIDE_INT_ELT): New.
1408 * doc/tm.texi.in (REAL_VALUE_TO_INT): Remove.
1409 (REAL_VALUE_FROM_INT): Remove.
1410 (TARGET_SUPPORTS_WIDE_INT): New.
1411 * doc/tm.texi: Regenerate.
1412 * dojump.c (prefer_and_bit_test): Use wide-int interfaces.
1413 * double-int.h: Include wide-int.h.
1414 (struct wi::int_traits): New.
1415 * dwarf2out.c (get_full_len): New.
1416 (dw_val_equal_p): Add case dw_val_class_wide_int.
1417 (size_of_loc_descr): Likewise.
1418 (output_loc_operands): Likewise.
1419 (insert_double): Remove.
1420 (insert_wide_int): New.
1422 (print_die): Add case dw_val_class_wide_int.
1423 (attr_checksum): Likewise.
1424 (attr_checksum_ordered): Likewise.
1425 (same_dw_val_p): Likewise.
1426 (size_of_die): Likewise.
1427 (value_format): Likewise.
1428 (output_die): Likewise.
1429 (double_int_type_size_in_bits): Rename to offset_int_type_size_in_bits.
1431 (clz_loc_descriptor): Use wide-int interfaces.
1432 (mem_loc_descriptor): Likewise. Handle CONST_WIDE_INT.
1433 (loc_descriptor): Use wide-int interfaces. Handle CONST_WIDE_INT.
1434 (round_up_to_align): Use wide-int interfaces.
1435 (field_byte_offset): Likewise.
1436 (insert_double): Rename to insert_wide_int. Use wide-int interfaces.
1437 (add_const_value_attribute): Handle CONST_WIDE_INT. Update
1438 CONST_DOUBLE handling. Use wide-int interfaces.
1439 (add_bound_info): Use tree_fits_uhwi_p. Use wide-int interfaces.
1440 (gen_enumeration_type_die): Use add_AT_wide.
1441 (hash_loc_operands): Add case dw_val_class_wide_int.
1442 (compare_loc_operands): Likewise.
1443 * dwarf2out.h: Include wide-int.h.
1444 (wide_int_ptr): New.
1445 (enum dw_val_class): Add dw_val_class_wide_int.
1446 (struct dw_val_struct): Add val_wide.
1447 * emit-rtl.c (const_wide_int_htab): New.
1448 (const_wide_int_htab_hash): New.
1449 (const_wide_int_htab_eq): New.
1450 (lookup_const_wide_int): New.
1451 (const_double_htab_hash): Use wide-int interfaces.
1452 (const_double_htab_eq): Likewise.
1453 (rtx_to_double_int): Conditionally compile for wide-int.
1454 (immed_double_int_const): Rename to immed_wide_int_const and
1455 update for wide-int.
1456 (immed_double_const): Conditionally compile for wide-int.
1457 (init_emit_once): Use wide-int interfaces.
1458 * explow.c (plus_constant): Likewise.
1459 * expmed.c (mask_rtx): Move further up file. Use wide-int interfaces.
1460 (lshift_value): Use wide-int interfaces.
1461 (expand_mult): Likewise.
1462 (choose_multiplier): Likewise.
1463 (expand_smod_pow2): Likewise.
1464 (make_tree): Likewise.
1465 * expr.c (convert_modes): Consolidate handling of constants.
1466 Use wide-int interfaces.
1467 (emit_group_load_1): Add note.
1468 (store_expr): Update comment.
1469 (get_inner_reference): Use wide-int interfaces.
1470 (expand_constructor): Update comment.
1471 (expand_expr_real_2): Use wide-int interfaces.
1472 (expand_expr_real_1): Likewise.
1473 (reduce_to_bit_field_precision): Likewise.
1474 (const_vector_from_tree): Likewise.
1475 * final.c: Include wide-int-print.h.
1476 (output_addr_const): Handle CONST_WIDE_INT. Use CONST_DOUBLE_AS_INT_P.
1477 * fixed-value.c: Include wide-int.h.
1478 (fixed_from_string): Use wide-int interfaces.
1479 (fixed_to_decimal): Likewise.
1480 (fixed_convert_from_real): Likewise.
1481 (real_convert_from_fixed): Likewise.
1482 * fold-const.h (mem_ref_offset): Return an offset_int.
1483 (div_if_zero_remainder): Remove code parameter.
1484 * fold-const.c (div_if_zero_remainder): Remove code parameter.
1485 Use wide-int interfaces.
1486 (may_negate_without_overflow_p): Use wide-int interfaces.
1487 (negate_expr_p): Likewise.
1488 (fold_negate_expr): Likewise.
1489 (int_const_binop_1): Likewise.
1490 (const_binop): Likewise.
1491 (fold_convert_const_int_from_int): Likewise.
1492 (fold_convert_const_int_from_real): Likewise.
1493 (fold_convert_const_int_from_fixed): Likewise.
1494 (fold_convert_const_fixed_from_int): Likewise.
1495 (all_ones_mask_p): Take an unsigned size. Use wide-int interfaces.
1496 (sign_bit_p): Use wide-int interfaces.
1497 (make_range_step): Likewise.
1498 (build_range_check): Likewise. Pass an integer of the correct type
1499 instead of using integer_one_node.
1500 (range_predecessor): Pass an integer of the correct type instead
1501 of using integer_one_node.
1502 (range_successor): Likewise.
1503 (merge_ranges): Likewise.
1504 (unextend): Use wide-int interfaces.
1505 (extract_muldiv_1): Likewise.
1506 (fold_div_compare): Likewise.
1507 (fold_single_bit_test): Likewise.
1508 (fold_sign_changed_comparison): Likewise.
1509 (try_move_mult_to_index): Update calls to div_if_zero_remainder.
1510 (fold_plusminus_mult_expr): Use wide-int interfaces.
1511 (native_encode_int): Likewise.
1512 (native_interpret_int): Likewise.
1513 (fold_unary_loc): Likewise.
1514 (pointer_may_wrap_p): Likewise.
1515 (size_low_cst): Likewise.
1516 (mask_with_tz): Likewise.
1517 (fold_binary_loc): Likewise.
1518 (fold_ternary_loc): Likewise.
1519 (multiple_of_p): Likewise.
1520 (tree_call_nonnegative_warnv_p): Update calls to
1521 tree_int_cst_min_precision and real_from_integer.
1522 (fold_negate_const): Use wide-int interfaces.
1523 (fold_abs_const): Likewise.
1524 (fold_relational_const): Use tree_int_cst_lt.
1525 (round_up_loc): Use wide-int interfaces.
1526 * genemit.c (gen_exp): Add CONST_WIDE_INT case.
1527 * gengenrtl.c (excluded_rtx): Add CONST_WIDE_INT case.
1528 * gengtype.c: Remove include of double-int.h.
1529 (do_typedef): Use wide-int interfaces.
1530 (open_base_files): Add wide-int.h.
1531 (main): Add offset_int and widest_int typedefs.
1532 * gengtype-lex.l: Handle "^".
1533 (CXX_KEYWORD): Add "static".
1534 * gengtype-parse.c (require3): New.
1535 (require_template_declaration): Handle constant template arguments
1536 and nested templates.
1537 * gengtype-state.c: Don't include "double-int.h".
1538 * genpreds.c (write_one_predicate_function): Update comment.
1539 (write_tm_constrs_h): Add check for hval and lval use in
1541 * genrecog.c (validate_pattern): Add CONST_WIDE_INT case.
1542 (add_to_sequence): Likewise.
1543 * gensupport.c (struct std_pred_table): Add const_scalar_int_operand
1544 and const_double_operand.
1545 * gimple.c (preprocess_case_label_vec_for_gimple): Use wide-int
1547 * gimple-fold.c (get_base_constructor): Likewise.
1548 (fold_array_ctor_reference): Likewise.
1549 (fold_nonarray_ctor_reference): Likewise.
1550 (fold_const_aggregate_ref_1): Likewise.
1551 (gimple_val_nonnegative_real_p): Likewise.
1552 (gimple_fold_indirect_ref): Likewise.
1553 * gimple-pretty-print.c (dump_ssaname_info): Likewise.
1554 * gimple-ssa-strength-reduction.c: Include wide-int-print.h.
1555 (struct slsr_cand_d): Change index to be widest_int.
1556 (struct incr_info_d): Change incr to be widest_int.
1557 (alloc_cand_and_find_basis): Use wide-int interfaces.
1558 (slsr_process_phi): Likewise.
1559 (backtrace_base_for_ref): Likewise. Return a widest_int.
1560 (restructure_reference): Take a widest_int instead of a double_int.
1561 (slsr_process_ref): Use wide-int interfaces.
1562 (create_mul_ssa_cand): Likewise.
1563 (create_mul_imm_cand): Likewise.
1564 (create_add_ssa_cand): Likewise.
1565 (create_add_imm_cand): Take a widest_int instead of a double_int.
1566 (slsr_process_add): Use wide-int interfaces.
1567 (slsr_process_cast): Likewise.
1568 (slsr_process_copy): Likewise.
1569 (dump_candidate): Likewise.
1570 (dump_incr_vec): Likewise.
1571 (replace_ref): Likewise.
1572 (cand_increment): Likewise. Return a widest_int.
1573 (cand_abs_increment): Likewise.
1574 (replace_mult_candidate): Take a widest_int instead of a double_int.
1575 (replace_unconditional_candidate): Use wide-int interfaces.
1576 (incr_vec_index): Take a widest_int instead of a double_int.
1577 (create_add_on_incoming_edge): Likewise.
1578 (create_phi_basis): Use wide-int interfaces.
1579 (replace_conditional_candidate): Likewise.
1580 (record_increment): Take a widest_int instead of a double_int.
1581 (record_phi_increments): Use wide-int interfaces.
1582 (phi_incr_cost): Take a widest_int instead of a double_int.
1583 (lowest_cost_path): Likewise.
1584 (total_savings): Likewise.
1585 (analyze_increments): Use wide-int interfaces.
1586 (ncd_with_phi): Take a widest_int instead of a double_int.
1587 (ncd_of_cand_and_phis): Likewise.
1588 (nearest_common_dominator_for_cands): Likewise.
1589 (insert_initializers): Use wide-int interfaces.
1590 (all_phi_incrs_profitable): Likewise.
1591 (replace_one_candidate): Likewise.
1592 (replace_profitable_candidates): Likewise.
1593 * godump.c: Include wide-int-print.h.
1594 (go_output_typedef): Use wide-int interfaces.
1595 * graphite-clast-to-gimple.c (gmp_cst_to_tree): Likewise.
1596 * graphite-sese-to-poly.c (tree_int_to_gmp): Likewise.
1597 (build_loop_iteration_domains): Likewise.
1598 * hooks.h: Include wide-int.h rather than double-int.h.
1599 (hook_bool_dint_dint_uint_bool_true): Delete.
1600 (hook_bool_wint_wint_uint_bool_true): Declare.
1601 * hooks.c (hook_bool_dint_dint_uint_bool_true): Removed.
1602 (hook_bool_wint_wint_uint_bool_true): New.
1603 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): Use wide-int
1605 (ubsan_expand_si_overflow_mul_check): Likewise.
1606 * ipa-devirt.c (get_polymorphic_call_info): Likewise.
1607 * ipa-prop.c (compute_complex_assign_jump_func): Likewise.
1608 (get_ancestor_addr_info): Likewise.
1609 (ipa_modify_call_arguments): Likewise.
1610 * loop-doloop.c (doloop_modify): Likewise.
1611 (doloop_optimize): Likewise.
1612 * loop-iv.c (iv_number_of_iterations): Likewise.
1613 * loop-unroll.c (decide_unroll_constant_iterations): Likewise.
1614 (unroll_loop_constant_iterations): Likewise.
1615 (decide_unroll_runtime_iterations): Likewise.
1616 (unroll_loop_runtime_iterations): Likewise.
1617 (decide_peel_simple): Likewise.
1618 (decide_unroll_stupid): Likewise.
1619 * lto-streamer-in.c (streamer_read_wi): Add.
1620 (input_cfg): Use wide-int interfaces.
1621 (lto_input_tree_1): Likewise.
1622 * lto-streamer-out.c (streamer_write_wi): Add.
1623 (hash_tree): Use wide-int interfaces.
1624 (output_cfg): Likewise.
1625 * Makefile.in (OBJS): Add wide-int.o and wide-int-print.o.
1626 (GTFILES): Add wide-int.h and signop.h.
1627 (TAGS): Look for .cc files too.
1628 * omp-low.c (scan_omp_1_op): Use wide-int interfaces.
1629 * optabs.c (expand_subword_shift): Likewise.
1630 (expand_doubleword_shift): Likewise.
1631 (expand_absneg_bit): Likewise.
1632 (expand_copysign_absneg): Likewise.
1633 (expand_copysign_bit): Likewise.
1634 * postreload.c (reload_cse_simplify_set): Likewise.
1635 * predict.c (predict_iv_comparison): Likewise.
1636 * pretty-print.h: Include wide-int-print.h.
1638 * print-rtl.c (print_rtx): Add CONST_WIDE_INT case.
1639 * print-tree.c: Include wide-int-print.h.
1640 (print_node_brief): Use wide-int interfaces.
1641 (print_node): Likewise.
1642 * read-rtl.c (validate_const_wide_int): New.
1643 (read_rtx_code): Add CONST_WIDE_INT case.
1644 * real.c: Include wide-int.h.
1645 (real_to_integer2): Delete.
1646 (real_to_integer): New function, returning a wide_int.
1647 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
1648 (ten_to_ptwo): Update call to real_from_integer.
1649 (real_digit): Likewise.
1650 * real.h: Include signop.h, wide-int.h and insn-modes.h.
1651 (real_to_integer2, REAL_VALUE_FROM_INT, REAL_VALUE_FROM_UNSIGNED_INT)
1652 (REAL_VALUE_TO_INT): Delete.
1653 (real_to_integer): Declare a wide-int form.
1654 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
1655 * recog.c (const_int_operand): Improve comment.
1656 (const_scalar_int_operand): New.
1657 (const_double_operand): Add a separate definition for CONST_WIDE_INT.
1658 * rtlanal.c (commutative_operand_precedence): Handle CONST_WIDE_INT.
1659 (split_double): Likewise.
1660 * rtl.c (DEF_RTL_EXPR): Handle CONST_WIDE_INT.
1661 (rtx_size): Likewise.
1662 (rtx_alloc_stat_v): New.
1663 (rtx_alloc_stat): Now calls rtx_alloc_stat_v.
1664 (cwi_output_hex): New.
1665 (iterative_hash_rtx): Handle CONST_WIDE_INT.
1666 (cwi_check_failed_bounds): New.
1667 * rtl.def (CONST_WIDE_INT): New.
1668 * rtl.h: Include <utility> and wide-int.h.
1669 (struct hwivec_def): New.
1670 (CWI_GET_NUM_ELEM): New.
1671 (CWI_PUT_NUM_ELEM): New.
1672 (struct rtx_def): Add num_elem and hwiv.
1673 (CASE_CONST_SCALAR_INT): Modify for TARGET_SUPPORTS_WIDE_INT.
1674 (CASE_CONST_UNIQUE): Likewise.
1675 (CASE_CONST_ANY): Likewise.
1676 (CONST_SCALAR_INT_P): Likewise.
1677 (CONST_WIDE_INT_P): New.
1679 (HWIVEC_CHECK): New.
1680 (cwi_check_failed_bounds): New.
1682 (HWIVEC_CHECK): New.
1683 (CONST_WIDE_INT_VEC) New.
1684 (CONST_WIDE_INT_NUNITS) New.
1685 (CONST_WIDE_INT_ELT) New.
1686 (rtx_mode_t): New type.
1687 (wi::int_traits <rtx_mode_t>): New.
1689 (wi::min_value): New.
1690 (wi::max_value): New.
1692 (const_wide_int_alloc): New.
1693 (immed_wide_int_const): New.
1694 * sched-vis.c (print_value): Handle CONST_WIDE_INT.
1695 * sel-sched-ir.c (lhs_and_rhs_separable_p): Update comment.
1696 * signop.h: New file.
1697 * simplify-rtx.c (mode_signbit_p): Handle CONST_WIDE_INT.
1698 (simplify_const_unary_operation): Use wide-int interfaces.
1699 (simplify_binary_operation_1): Likewise.
1700 (simplify_const_binary_operation): Likewise.
1701 (simplify_const_relational_operation): Likewise.
1702 (simplify_immed_subreg): Likewise.
1703 * stmt.c (expand_case): Likewise.
1704 * stor-layout.h (set_min_and_max_values_for_integral_type): Take a
1705 signop rather than a bool.
1706 * stor-layout.c (layout_type): Use wide-int interfaces.
1707 (initialize_sizetypes): Update calls to
1708 set_min_and_max_values_for_integral_type.
1709 (set_min_and_max_values_for_integral_type): Take a signop rather
1710 than a bool. Use wide-int interfaces.
1711 (fixup_signed_type): Update accordingly. Remove
1712 HOST_BITS_PER_DOUBLE_INT limit.
1713 (fixup_unsigned_type): Likewise.
1714 * system.h (STATIC_CONSTANT_P): New.
1715 (STATIC_ASSERT): New.
1716 * target.def (can_use_doloop_p): Take widest_ints rather than
1718 * target.h: Include wide-int.h rather than double-int.h.
1719 * targhooks.h (can_use_doloop_if_innermost): Take widest_ints rather
1721 * targhooks.c (default_cxx_get_cookie_size): Use tree_int_cst_lt
1722 rather than INT_CST_LT_UNSIGNED.
1723 (can_use_doloop_if_innermost): Take widest_ints rather than
1725 * tree-affine.c: Include wide-int-print.h.
1726 (double_int_ext_for_comb): Delete.
1727 (wide_int_ext_for_comb): New.
1728 (aff_combination_zero): Use wide-int interfaces.
1729 (aff_combination_const): Take a widest_int instead of a double_int.
1730 (aff_combination_elt): Use wide-int interfaces.
1731 (aff_combination_scale): Take a widest_int instead of a double_int.
1732 (aff_combination_add_elt): Likewise.
1733 (aff_combination_add_cst): Likewise.
1734 (aff_combination_add): Use wide-int interfaces.
1735 (aff_combination_convert): Likewise.
1736 (tree_to_aff_combination): Likewise.
1737 (add_elt_to_tree): Take a widest_int instead of a double_int.
1738 (aff_combination_to_tree): Use wide-int interfaces.
1739 (aff_combination_remove_elt): Likewise.
1740 (aff_combination_add_product): Take a widest_int instead of
1742 (aff_combination_mult): Use wide-int interfaces.
1743 (aff_combination_expand): Likewise.
1744 (double_int_constant_multiple_p): Delete.
1745 (wide_int_constant_multiple_p): New.
1746 (aff_combination_constant_multiple_p): Take a widest_int pointer
1747 instead of a double_int pointer.
1748 (print_aff): Use wide-int interfaces.
1749 (get_inner_reference_aff): Take a widest_int pointer
1750 instead of a double_int pointer.
1751 (aff_comb_cannot_overlap_p): Take widest_ints instead of double_ints.
1752 * tree-affine.h: Include wide-int.h.
1753 (struct aff_comb_elt): Change type of coef to widest_int.
1754 (struct affine_tree_combination): Change type of offset to widest_int.
1755 (double_int_ext_for_comb): Delete.
1756 (wide_int_ext_for_comb): New.
1757 (aff_combination_const): Use widest_int instead of double_int.
1758 (aff_combination_scale): Likewise.
1759 (aff_combination_add_elt): Likewise.
1760 (aff_combination_constant_multiple_p): Likewise.
1761 (get_inner_reference_aff): Likewise.
1762 (aff_comb_cannot_overlap_p): Likewise.
1763 (aff_combination_zero_p): Use wide-int interfaces.
1764 * tree.c: Include tree.h.
1765 (init_ttree): Use make_int_cst.
1766 (tree_code_size): Removed code for INTEGER_CST case.
1767 (tree_size): Add INTEGER_CST case.
1768 (make_node_stat): Update comment.
1769 (get_int_cst_ext_nunits, build_new_int_cst, build_int_cstu): New.
1770 (build_int_cst_type): Use wide-int interfaces.
1771 (double_int_to_tree): Likewise.
1772 (double_int_fits_to_tree_p): Delete.
1773 (force_fit_type_double): Delete.
1774 (force_fit_type): New.
1775 (int_cst_hash_hash): Use wide-int interfaces.
1776 (int_cst_hash_eq): Likewise.
1777 (build_int_cst_wide): Delete.
1778 (wide_int_to_tree): New.
1779 (cache_integer_cst): Use wide-int interfaces.
1780 (build_low_bits_mask): Likewise.
1781 (cst_and_fits_in_hwi): Likewise.
1782 (real_value_from_int_cst): Likewise.
1783 (make_int_cst_stat): New.
1784 (integer_zerop): Use wide_int interfaces.
1785 (integer_onep): Likewise.
1786 (integer_all_onesp): Likewise.
1787 (integer_pow2p): Likewise.
1788 (integer_nonzerop): Likewise.
1789 (tree_log2): Likewise.
1790 (tree_floor_log2): Likewise.
1791 (tree_ctz): Likewise.
1792 (int_size_in_bytes): Likewise.
1793 (mem_ref_offset): Return an offset_int rather than a double_int.
1794 (build_type_attribute_qual_variant): Use wide_int interfaces.
1795 (type_hash_eq): Likewise
1796 (tree_int_cst_equal): Likewise.
1797 (tree_int_cst_lt): Delete.
1798 (tree_int_cst_compare): Likewise.
1799 (tree_fits_shwi_p): Use wide_int interfaces.
1800 (tree_fits_uhwi_p): Likewise.
1801 (tree_int_cst_sign_bit): Likewise.
1802 (tree_int_cst_sgn): Likewise.
1803 (tree_int_cst_min_precision): Take a signop rather than a bool.
1804 (simple_cst_equal): Use wide_int interfaces.
1805 (compare_tree_int): Likewise.
1806 (iterative_hash_expr): Likewise.
1807 (int_fits_type_p): Likewise. Use tree_int_cst_lt rather than
1809 (get_type_static_bounds): Use wide_int interfaces.
1810 (tree_int_cst_elt_check_failed): New.
1811 (build_common_tree_nodes): Reordered to set prec before filling in
1813 (int_cst_value): Check cst_and_fits_in_hwi.
1814 (widest_int_cst_value): Use wide_int interfaces.
1815 (upper_bound_in_type): Likewise.
1816 (lower_bound_in_type): Likewise.
1817 (num_ending_zeros): Likewise.
1818 (drop_tree_overflow): Likewise.
1819 * tree-call-cdce.c (check_pow): Update call to real_from_integer.
1820 (gen_conditions_for_pow_cst_base): Likewise.
1821 * tree-cfg.c: Include wide-int.h and wide-int-print.h.
1822 (group_case_labels_stmt): Use wide-int interfaces.
1823 (verify_gimple_assign_binary): Likewise.
1824 (print_loop): Likewise.
1825 * tree-chrec.c (tree_fold_binomial): Likewise.
1826 * tree-core.h (struct tree_base): Add int_length.
1827 (struct tree_int_cst): Change rep of value.
1828 * tree-data-ref.c (dr_analyze_innermost): Use wide-int interfaces.
1829 (dr_may_alias_p): Likewise.
1830 (max_stmt_executions_tree): Likewise.
1831 * tree.def (INTEGER_CST): Update comment.
1832 * tree-dfa.c (get_ref_base_and_extent): Use wide-int interfaces.
1833 * tree-dfa.h (get_addr_base_and_unit_offset_1): Likewise.
1834 * tree-dump.c: Include wide-int.h and wide-int-print.h.
1835 (dequeue_and_dump): Use wide-int interfaces.
1836 * tree.h: Include wide-int.h.
1837 (NULL_TREE): Moved to earlier loc in file.
1838 (TREE_INT_CST_ELT_CHECK): New.
1839 (tree_int_cst_elt_check_failed): New.
1841 (TREE_INT_CST): Delete.
1842 (TREE_INT_CST_LOW): Use wide-int interfaces.
1843 (TREE_INT_CST_HIGH): Delete.
1844 (TREE_INT_CST_NUNITS): New.
1845 (TREE_INT_CST_EXT_NUNITS): Likewise.
1846 (TREE_INT_CST_OFFSET_NUNITS): Likewise.
1847 (TREE_INT_CST_ELT): Likewise.
1848 (INT_CST_LT): Delete.
1849 (tree_int_cst_elt_check): New (two forms).
1850 (type_code_size): Update comment.
1851 (make_int_cst_stat, make_int_cst): New.
1852 (tree_to_double_int): Delete.
1853 (double_int_fits_to_tree_p): Delete.
1854 (force_fit_type_double): Delete.
1855 (build_int_cstu): Replace with out-of-line function.
1856 (build_int_cst_wide): Delete.
1857 (tree_int_cst_lt): Define inline.
1858 (tree_int_cst_le): New.
1859 (tree_int_cst_compare): Define inline.
1860 (tree_int_cst_min_precision): Take a signop rather than a bool.
1861 (wi::int_traits <const_tree>): New.
1862 (wi::int_traits <tree>): New.
1863 (wi::extended_tree): New.
1864 (wi::int_traits <wi::extended_tree>): New.
1865 (wi::to_widest): New.
1866 (wi::to_offset): New.
1867 (wi::fits_to_tree_p): New.
1868 (wi::min_value): New.
1869 (wi::max_value): New.
1870 * tree-inline.c (remap_gimple_op_r): Use wide-int interfaces.
1871 (copy_tree_body_r): Likewise.
1872 * tree-object-size.c (compute_object_offset): Likewise.
1873 (addr_object_size): Likewise.
1874 * tree-predcom.c: Include wide-int-print.h.
1875 (struct dref_d): Change type of offset to widest_int.
1876 (dump_dref): Call wide-int printer.
1877 (aff_combination_dr_offset): Use wide-int interfaces.
1878 (determine_offset): Take a widest_int pointer rather than a
1880 (split_data_refs_to_components): Use wide-int interfaces.
1881 (suitable_component_p): Likewise.
1882 (order_drefs): Likewise.
1883 (add_ref_to_chain): Likewise.
1884 (valid_initializer_p): Likewise.
1885 (determine_roots_comp): Likewise.
1886 * tree-pretty-print.c: Include wide-int-print.h.
1887 (dump_generic_node): Use wide-int interfaces.
1888 * tree-sra.c (sra_ipa_modify_expr): Likewise.
1889 * tree-ssa-address.c (addr_for_mem_ref): Likewise.
1890 (move_fixed_address_to_symbol): Likewise.
1891 (move_hint_to_base): Likewise.
1892 (move_pointer_to_base): Likewise.
1893 (move_variant_to_index): Likewise.
1894 (most_expensive_mult_to_index): Likewise.
1895 (addr_to_parts): Likewise.
1896 (copy_ref_info): Likewise.
1897 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
1898 (indirect_refs_may_alias_p): Likewise.
1899 (stmt_kills_ref_p_1): Likewise.
1900 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
1901 * tree-ssa-ccp.c: Update comment at top of file. Include
1903 (struct prop_value_d): Change type of mask to widest_int.
1904 (extend_mask): New function.
1905 (dump_lattice_value): Use wide-int interfaces.
1906 (get_default_value): Likewise.
1907 (set_constant_value): Likewise.
1908 (set_value_varying): Likewise.
1909 (valid_lattice_transition): Likewise.
1910 (set_lattice_value): Likewise.
1911 (value_to_double_int): Delete.
1912 (value_to_wide_int): New.
1913 (get_value_from_alignment): Use wide-int interfaces.
1914 (get_value_for_expr): Likewise.
1915 (do_dbg_cnt): Likewise.
1916 (ccp_finalize): Likewise.
1917 (ccp_lattice_meet): Likewise.
1918 (bit_value_unop_1): Use widest_ints rather than double_ints.
1919 (bit_value_binop_1): Likewise.
1920 (bit_value_unop): Use wide-int interfaces.
1921 (bit_value_binop): Likewise.
1922 (bit_value_assume_aligned): Likewise.
1923 (evaluate_stmt): Likewise.
1924 (ccp_fold_stmt): Likewise.
1925 (visit_cond_stmt): Likewise.
1926 (ccp_visit_stmt): Likewise.
1927 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Likewise.
1928 (constant_pointer_difference): Likewise.
1929 (associate_pointerplus): Likewise.
1930 (combine_conversions): Likewise.
1931 * tree-ssa-loop.h: Include wide-int.h.
1932 (struct tree_niter_desc): Change type of max to widest_int.
1933 * tree-ssa-loop-im.c (mem_refs_may_alias_p): Use wide-int interfaces.
1934 * tree-ssa-loop-ivcanon.c (remove_exits_and_undefined_stmts): Likewise.
1935 (remove_redundant_iv_tests): Likewise.
1936 (canonicalize_loop_induction_variables): Likewise.
1937 * tree-ssa-loop-ivopts.c (alloc_iv): Likewise.
1938 (constant_multiple_of): Take a widest_int pointer instead of
1939 a double_int pointer.
1940 (get_computation_aff): Use wide-int interfaces.
1941 (ptr_difference_cost): Likewise.
1942 (difference_cost): Likewise.
1943 (get_loop_invariant_expr_id): Likewise.
1944 (get_computation_cost_at): Likewise.
1945 (iv_elimination_compare_lt): Likewise.
1946 (may_eliminate_iv): Likewise.
1947 * tree-ssa-loop-niter.h (estimated_loop_iterations): Use widest_int
1948 instead of double_int.
1949 (max_loop_iterations): Likewise.
1950 (max_stmt_executions): Likewise.
1951 (estimated_stmt_executions): Likewise.
1952 * tree-ssa-loop-niter.c: Include wide-int-print.h.
1953 (split_to_var_and_offset): Use wide-int interfaces.
1954 (determine_value_range): Likewise.
1955 (bound_difference_of_offsetted_base): Likewise.
1956 (bounds_add): Take a widest_int instead of a double_int.
1957 (number_of_iterations_ne_max): Use wide-int interfaces.
1958 (number_of_iterations_ne): Likewise.
1959 (number_of_iterations_lt_to_ne): Likewise.
1960 (assert_loop_rolls_lt): Likewise.
1961 (number_of_iterations_lt): Likewise.
1962 (number_of_iterations_le): Likewise.
1963 (number_of_iterations_cond): Likewise.
1964 (number_of_iterations_exit): Likewise.
1965 (finite_loop_p): Likewise.
1966 (derive_constant_upper_bound_assign): Likewise.
1967 (derive_constant_upper_bound): Return a widest_int.
1968 (derive_constant_upper_bound_ops): Likewise.
1969 (do_warn_aggressive_loop_optimizations): Use wide-int interfaces.
1970 (record_estimate): Take a widest_int rather than a double_int.
1971 (record_nonwrapping_iv): Use wide-int interfaces.
1972 (double_int_cmp): Delete.
1973 (wide_int_cmp): New.
1974 (bound_index): Take a widest_int rather than a double_int.
1975 (discover_iteration_bound_by_body_walk): Use wide-int interfaces.
1976 (maybe_lower_iteration_bound): Likewise.
1977 (estimate_numbers_of_iterations_loop): Likewise.
1978 (estimated_loop_iterations): Take a widest_int pointer than than
1979 a double_int pointer.
1980 (estimated_loop_iterations_int): Use wide-int interfaces.
1981 (max_loop_iterations): Take a widest_int pointer than than
1982 a double_int pointer.
1983 (max_loop_iterations_int): Use wide-int interfaces.
1984 (max_stmt_executions): Take a widest_int pointer than than
1985 a double_int pointer.
1986 (estimated_stmt_executions): Likewise.
1987 (n_of_executions_at_most): Use wide-int interfaces.
1988 (scev_probably_wraps_p): Likewise.
1989 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Update calls
1991 * tree-scalar-evolution.c (simplify_peeled_chrec): Use wide-int
1993 * tree-ssanames.c (set_range_info): Use wide_int_refs rather than
1994 double_ints. Adjust for trailing_wide_ints <3> representation.
1995 (set_nonzero_bits): Likewise.
1996 (get_range_info): Return wide_ints rather than double_ints.
1997 Adjust for trailing_wide_ints <3> representation.
1998 (get_nonzero_bits): Likewise.
1999 (duplicate_ssa_name_range_info): Adjust for trailing_wide_ints <3>
2001 * tree-ssanames.h (struct range_info_def): Replace min, max and
2002 nonzero_bits with a trailing_wide_ints <3>.
2003 (set_range_info): Use wide_int_refs rather than double_ints.
2004 (set_nonzero_bits): Likewise.
2005 (get_range_info): Return wide_ints rather than double_ints.
2006 (get_nonzero_bits): Likewise.
2007 * tree-ssa-phiopt.c (jump_function_from_stmt): Use wide-int interfaces.
2008 * tree-ssa-pre.c (phi_translate_1): Likewise.
2009 * tree-ssa-reassoc.c (decrement_power): Use calls to real_from_integer.
2010 (acceptable_pow_call): Likewise.
2011 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Use wide-int
2013 (vn_reference_fold_indirect): Likewise.
2014 (vn_reference_maybe_forwprop_address): Likewise.
2015 (valueize_refs_1): Likewise.
2016 * tree-ssa-structalias.c (get_constraint_for_ptr_offset): Likewise.
2017 * tree-ssa-uninit.c (is_value_included_in): Use wide-int interfaces,
2018 tree_int_cst_lt and tree_int_cst_le.
2019 * tree-streamer-in.c (unpack_ts_base_value_fields): Use wide-int
2021 (streamer_alloc_tree): Likewise.
2022 * tree-streamer-out.c (pack_ts_int_cst_value_fields): Likewise.
2023 (streamer_write_tree_header): Likewise.
2024 (streamer_write_integer_cst): Likewise.
2025 * tree-switch-conversion.c (emit_case_bit_tests): Likewise.
2026 (build_constructors): Likewise.
2027 (array_value_type): Likewise.
2028 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list): Likewise.
2029 (vect_check_gather): Likewise.
2030 * tree-vect-generic.c (build_replicated_const): Likewise.
2031 (expand_vector_divmod): Likewise.
2032 * tree-vect-loop.c (vect_transform_loop): Likewise.
2033 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
2034 (vect_do_peeling_for_alignment): Likewise.
2035 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
2036 * tree-vrp.c: Include wide-int.h.
2037 (operand_less_p): Use wide-int interfaces and tree_int_cst_lt.
2038 (extract_range_from_assert): Use wide-int interfaces.
2039 (vrp_int_const_binop): Likewise.
2040 (zero_nonzero_bits_from_vr): Take wide_int pointers rather than
2041 double_int pointers.
2042 (ranges_from_anti_range): Use wide-int interfaces.
2043 (quad_int_cmp): Delete.
2044 (quad_int_pair_sort): Likewise.
2045 (extract_range_from_binary_expr_1): Use wide-int interfaces.
2046 (extract_range_from_unary_expr_1): Likewise.
2047 (adjust_range_with_scev): Likewise.
2048 (masked_increment): Take and return wide_ints rather than double_ints.
2049 (register_edge_assert_for_2): Use wide-int interfaces.
2050 (check_array_ref): Likewise.
2051 (search_for_addr_array): Likewise.
2052 (maybe_set_nonzero_bits): Likewise.
2053 (union_ranges): Pass an integer of the correct type instead of
2054 using integer_one_node.
2055 (intersect_ranges): Likewise.
2056 (simplify_truth_ops_using_ranges): Likewise.
2057 (simplify_bit_ops_using_ranges): Use wide-int interfaces.
2058 (range_fits_type_p): Likewise.
2059 (simplify_cond_using_ranges): Likewise. Take a signop rather than
2061 (simplify_conversion_using_ranges): Use wide-int interfaces.
2062 (simplify_float_conversion_using_ranges): Likewise.
2063 (vrp_finalize): Likewise.
2064 * value-prof.c (gimple_divmod_fixed_value_transform): Likewise.
2065 (gimple_stringops_transform): Likewise.
2066 * varasm.c (decode_addr_const): Likewise.
2067 (const_hash_1): Likewise.
2068 (const_rtx_hash_1): Likewise
2069 (output_constant): Likewise.
2070 (array_size_for_constructor): Likewise.
2071 (output_constructor_regular_field): Likewise.
2072 (output_constructor_bitfield): Likewise.
2073 * var-tracking.c (loc_cmp): Handle CONST_WIDE_INT.
2074 * mkconfig.sh: Include machmode.h to pick up BITS_PER_UNIT for
2076 * gencheck.c: Define BITS_PER_UNIT.
2079 * wide-int-print.cc: New.
2080 * wide-int-print.h: New.
2082 2014-05-06 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2084 * config/avr/avr.c (avr_can_eliminate): Mark unused argument.
2086 2014-05-06 Richard Biener <rguenther@suse.de>
2088 * tree-pass.h (TODO_verify_ssa, TODO_verify_flow,
2089 TODO_verify_stmts, TODO_verify_rtl_sharing): Remove.
2090 (TODO_verify_all): Adjust.
2091 * asan.c: Remove references to TODO_verify_ssa, TODO_verify_flow,
2092 TODO_verify_stmts and TODO_verify_rtl_sharing.
2093 * bb-reorder.c: Likewise.
2094 * cfgexpand.c: Likewise.
2095 * cprop.c: Likewise.
2097 * function.c: Likewise.
2098 * fwprop.c: Likewise.
2100 * gimple-ssa-isolate-paths.c: Likewise.
2101 * gimple-ssa-strength-reduction.c: Likewise.
2102 * ipa-split.c: Likewise.
2103 * loop-init.c: Likewise.
2104 * loop-unroll.c: Likewise.
2105 * lower-subreg.c: Likewise.
2106 * modulo-sched.c: Likewise.
2107 * postreload-gcse.c: Likewise.
2108 * predict.c: Likewise.
2109 * recog.c: Likewise.
2110 * sched-rgn.c: Likewise.
2111 * store-motion.c: Likewise.
2112 * tracer.c: Likewise.
2113 * trans-mem.c: Likewise.
2114 * tree-call-cdce.c: Likewise.
2115 * tree-cfg.c: Likewise.
2116 * tree-cfgcleanup.c: Likewise.
2117 * tree-complex.c: Likewise.
2118 * tree-eh.c: Likewise.
2119 * tree-emutls.c: Likewise.
2120 * tree-if-conv.c: Likewise.
2121 * tree-into-ssa.c: Likewise.
2122 * tree-loop-distribution.c: Likewise.
2123 * tree-object-size.c: Likewise.
2124 * tree-parloops.c: Likewise.
2125 * tree-pass.h: Likewise.
2126 * tree-sra.c: Likewise.
2127 * tree-ssa-ccp.c: Likewise.
2128 * tree-ssa-copy.c: Likewise.
2129 * tree-ssa-copyrename.c: Likewise.
2130 * tree-ssa-dce.c: Likewise.
2131 * tree-ssa-dom.c: Likewise.
2132 * tree-ssa-dse.c: Likewise.
2133 * tree-ssa-forwprop.c: Likewise.
2134 * tree-ssa-ifcombine.c: Likewise.
2135 * tree-ssa-loop-ch.c: Likewise.
2136 * tree-ssa-loop-ivcanon.c: Likewise.
2137 * tree-ssa-loop.c: Likewise.
2138 * tree-ssa-math-opts.c: Likewise.
2139 * tree-ssa-phiopt.c: Likewise.
2140 * tree-ssa-phiprop.c: Likewise.
2141 * tree-ssa-pre.c: Likewise.
2142 * tree-ssa-reassoc.c: Likewise.
2143 * tree-ssa-sink.c: Likewise.
2144 * tree-ssa-strlen.c: Likewise.
2145 * tree-ssa-tail-merge.c: Likewise.
2146 * tree-ssa-uncprop.c: Likewise.
2147 * tree-switch-conversion.c: Likewise.
2148 * tree-tailcall.c: Likewise.
2149 * tree-vect-generic.c: Likewise.
2150 * tree-vectorizer.c: Likewise.
2151 * tree-vrp.c: Likewise.
2153 * var-tracking.c: Likewise.
2154 * bt-load.c: Likewise.
2155 * cfgcleanup.c: Likewise.
2156 * combine-stack-adj.c: Likewise.
2157 * combine.c: Likewise.
2158 * compare-elim.c: Likewise.
2159 * config/epiphany/resolve-sw-modes.c: Likewise.
2160 * config/i386/i386.c: Likewise.
2161 * config/mips/mips.c: Likewise.
2162 * config/s390/s390.c: Likewise.
2163 * config/sh/sh_treg_combine.cc: Likewise.
2164 * config/sparc/sparc.c: Likewise.
2167 * final.c: Likewise.
2168 * ifcvt.c: Likewise.
2169 * mode-switching.c: Likewise.
2170 * passes.c: Likewise.
2171 * postreload.c: Likewise.
2173 * reg-stack.c: Likewise.
2174 * regcprop.c: Likewise.
2175 * regrename.c: Likewise.
2178 2014-05-06 Richard Biener <rguenther@suse.de>
2181 * bitmap.c (debug_bitmap): Dump to stderr, not stdout.
2182 * tree-ssa-structalias.c (dump_solution_for_var): Likewise.
2184 2014-05-05 Jan Hubicka <hubicka@ucw.cz>
2187 * ipa-devirt.c (get_class_context): Allow POD to change to non-POD.
2189 2014-05-05 Radovan Obradovic <robradovic@mips.com>
2190 Tom de Vries <tom@codesourcery.com>
2192 * target.def (call_fusage_contains_non_callee_clobbers): New
2194 * doc/tm.texi.in (@node Stack and Calling): Add Miscellaneous Register
2196 (@node Miscellaneous Register Hooks): New node.
2197 (@hook TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): New hook.
2198 * doc/tm.texi: Regenerate.
2200 2014-05-05 Marek Polacek <polacek@redhat.com>
2203 * opts.c (common_handle_option): Call error_at instead of warning_at.
2205 2014-05-05 Richard Biener <rguenther@suse.de>
2207 * passes.c (execute_function_todo): Don't reset TODO_verify_ssa
2208 from last_verified if update_ssa ran. Move TODO_verify_rtl_sharing
2209 under the TODO_verify_il umbrella.
2211 2014-05-05 Richard Biener <rguenther@suse.de>
2213 * passes.c (execute_function_todo): Move TODO_verify_flow under
2214 the TODO_verify_ul umbrella.
2216 2014-05-05 Richard Biener <rguenther@suse.de>
2219 * fold-const.c (fold_binary_loc): Consistently avoid canonicalizing
2220 X & CST away from a CST that is the mask of a mode.
2222 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2224 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
2225 int argument to enum machine_mode.
2226 (picochip_class_max_nregs): Ditto.
2227 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
2228 (picochip_class_max_nregs): Ditto.
2230 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2232 * target.def: Add new target hook.
2233 * doc/tm.texi: Regenerate.
2234 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
2235 * targhooks.c (default_keep_leaf_when_profiled): New function.
2237 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
2238 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
2240 2014-05-05 Bin Cheng <bin.cheng@arm.com>
2242 PR tree-optimization/60363
2243 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
2244 (copy_phi_args): New parameters. Call get_value_locus_in_path.
2245 (update_destination_phis): New parameter.
2246 (create_edge_and_update_destination_phis): Ditto.
2247 (ssa_fix_duplicate_block_edges): Pass new arguments.
2248 (thread_single_edge): Ditto.
2250 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
2252 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
2253 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
2254 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
2255 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
2256 Use RS6000_BTM_HARD_FLOAT.
2257 (BU_MISC_2): Likewise.
2258 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
2259 RS6000_BTM_HARD_FLOAT.
2260 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
2262 (rs6000_invalid_builtin): Add hard floating builtin support.
2263 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
2264 hard float builtins.
2265 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
2267 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
2269 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
2270 Add missing function* argument.
2272 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
2274 * lra-constraints.c (valid_address_p): Move earlier in file.
2275 Add a constraint argument to the address_info version.
2276 (satisfies_memory_constraint_p): New function.
2277 (satisfies_address_constraint_p): Likewise.
2278 (process_alt_operands, curr_insn_transform): Use them.
2279 (process_address): Pass the constraint to valid_address_p when
2280 checking address operands.
2282 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
2284 * config/mips/mips.c (mips_isa_rev): New variable.
2285 (mips_set_architecture): Set it.
2286 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
2288 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
2289 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
2290 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
2291 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
2292 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
2293 conditions in terms of mips_isa_rev.
2294 (mips_isa_rev): Declare.
2296 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
2298 * config/sh/sh-mem.cc: Use tabs instead of spaces.
2299 (prob_unlikely, prob_likely): Make variables const.
2301 2014-05-03 Denis Chertykov <chertykov@gmail.com>
2303 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
2305 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
2307 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
2309 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
2311 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
2312 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
2313 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
2315 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
2316 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
2318 Replace usage of ROUND_REG with sh_round_reg.
2319 Use CEIL instead of ROUND_ADVANCE.
2321 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
2324 * config/sh/sh.c: Include stdlib headers before everything else.
2326 2014-05-02 Jakub Jelinek <jakub@redhat.com>
2328 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
2329 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
2330 (gimplify_adjust_omp_clauses): Simd region is never
2331 directly nested in combined parallel. Instead, for linear
2332 with copyin/copyout, if in combined for simd loop, make decl
2333 firstprivate/lastprivate on OMP_FOR.
2334 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
2335 expand_omp_for_static_chunk): When setting endvar, also set
2336 fd->loop.v to the same value.
2338 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
2340 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
2342 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
2344 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
2347 2014-05-02 Marek Polacek <polacek@redhat.com>
2349 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
2351 2014-05-02 Kito Cheng <kito@0xlab.org>
2353 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
2354 to a C expression marco.
2355 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
2356 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
2357 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
2358 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
2359 HONOR_REG_ALLOC_ORDER.
2360 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
2362 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2364 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
2366 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2368 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
2370 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
2372 * tree-if-conv.c (is_cond_scalar_reduction): New function.
2373 (convert_scalar_cond_reduction): Likewise.
2374 (predicate_scalar_phi): Add recognition and transformation
2375 of simple conditioanl reduction to be vectorizable.
2377 2014-05-01 Marek Polacek <polacek@redhat.com>
2380 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
2382 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
2384 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
2385 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
2386 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
2387 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
2388 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
2389 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
2390 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
2391 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
2393 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
2395 * config/arc/arc.opt (mlra): Move comment above option name
2396 to avoid mis-parsing as language options.
2398 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2400 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
2401 * config/sol2.h: ... here.
2402 * config/sol2-10.h: Remove.
2404 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
2405 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
2406 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
2407 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
2408 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
2409 * config/sol2.h: ... here.
2410 (SECTION_NAME_FORMAT): Don't redefine.
2411 (STARTFILE_ARCH32_SPEC): Rename to ...
2412 (STARTFILE_ARCH_SPEC): ... this.
2413 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
2414 * config/sparc/sol2.h: ... here.
2415 (SECTION_NAME_FORMAT): Don't undef.
2416 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
2417 (SUBTARGET_EXTRA_SPECS): Remove.
2418 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
2420 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
2421 (MD_STARTFILE_PREFIX): Remove.
2422 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
2423 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
2424 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
2425 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
2426 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
2427 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
2428 * config/i386/sol2.h: ... here.
2429 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
2430 * config/i386/sol2-bi.h: Remove.
2431 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
2432 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
2434 * config/i386/t-sol2-64: Rename to ...
2435 * config/i386/t-sol2: ... this.
2436 * config/sparc/t-sol2-64: Rename to ...
2437 * config/sparc/t-sol2: ... this.
2439 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
2440 sol2_tm_file_head, sol2_tm_file_tail.
2441 Include ${cpu_type}/sol2.h before sol2.h.
2443 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
2444 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
2445 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
2446 Reflect i386/t-sol2-64 renaming.
2447 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
2448 Reflect sparc/t-sol2-64 renaming.
2450 2014-04-30 Richard Biener <rguenther@suse.de>
2452 * passes.c (execute_function_todo): Move TODO_verify_stmts
2453 and TODO_verify_ssa under the TODO_verify_il umbrella.
2454 * tree-ssa.h (verify_ssa): Adjust prototype.
2455 * tree-ssa.c (verify_ssa): Add parameter to tell whether
2456 we should verify SSA operands.
2457 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
2458 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
2459 whether we should verify whether not throwing stmts have EH info.
2460 * graphite-scop-detection.c (create_sese_edges): Adjust.
2461 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
2462 * tree-eh.c (lower_try_finally_switch): Do not add the
2463 default case label twice.
2465 2014-04-30 Marek Polacek <polacek@redhat.com>
2467 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
2468 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
2469 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
2470 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
2472 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
2474 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
2475 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
2476 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
2477 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
2478 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
2479 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
2480 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
2481 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
2483 2014-04-29 David Malcolm <dmalcolm@redhat.com>
2485 * tree-cfg.c (dump_function_to_file): Dump the return type of
2486 functions, in a line to itself before the function body, mimicking
2487 the layout of a C function.
2489 2014-04-29 Jakub Jelinek <jakub@redhat.com>
2491 PR tree-optimization/60971
2492 * tree-tailcall.c (process_assignment): Reject conversions which
2495 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
2497 * calls.c (initialize_argument_information): Always treat
2498 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
2499 (expand_call): Likewise.
2500 (emit_library_call_calue_1): Likewise.
2501 * expr.c (PUSH_ARGS_REVERSED): Do not define.
2502 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
2505 2014-04-29 Nick Clifton <nickc@redhat.com>
2507 * config/msp430/msp430.md (umulsidi): Fix typo.
2508 (mulhisi3): Enable even inside interrupt handlers.
2509 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
2510 bigger return address pushed in large mode.
2512 2014-04-29 Nick Clifton <nickc@redhat.com>
2514 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
2515 (arc_init_reg_tables): Use a machine_mode enum to iterate over
2517 * config/m32r/m32r.c (init_reg_tables): Likewise.
2518 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
2519 enum to hold the modes.
2521 2014-04-29 Richard Biener <rguenther@suse.de>
2523 * dominance.c (free_dominance_info): Add overload with
2525 (dom_info_state): Likewise.
2526 (dom_info_available_p): Likewise.
2527 * basic-block.h (free_dominance_info, dom_info_state,
2528 dom_info_available_p): Declare overloads.
2529 * passes.c (execute_function_todo): Verify that verifiers
2530 don't change dominator info state. Drop dominator info
2531 for IPA pass invocations.
2532 * cgraph.c (release_function_body): Restore asserts that
2533 dominator information is released.
2535 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
2537 * doc/invoke.texi: Fix typo.
2538 * tree-vrp.c: Fix typos.
2539 * gimple.c (infer_nonnull_range): Reorder operands of an && condition.
2541 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2543 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
2545 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
2547 * config/aarch64/aarch64-builtins.c
2548 (aarch64_types_storestruct_lane_qualifiers): New.
2549 (TYPES_STORESTRUCT_LANE): Likewise.
2550 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
2551 (st3_lane): Likewise.
2552 (st4_lane): Likewise.
2553 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
2554 (vec_store_lanesci_lane<mode>): Likewise.
2555 (vec_store_lanesxi_lane<mode>): Likewise.
2556 (aarch64_st2_lane<VQ:mode>): Likewise.
2557 (aarch64_st3_lane<VQ:mode>): Likewise.
2558 (aarch64_st4_lane<VQ:mode>): Likewise.
2559 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
2560 * config/aarch64/arm_neon.h
2561 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
2562 use new macro arguments.
2563 (__ST3_LANE_FUNC): Likewise.
2564 (__ST4_LANE_FUNC): Likewise.
2565 * config/aarch64/iterators.md (V_TWO_ELEM): New.
2566 (V_THREE_ELEM): Likewise.
2567 (V_FOUR_ELEM): Likewise.
2569 2014-04-28 David Malcolm <dmalcolm@redhat.com>
2571 * doc/gimple.texi: Replace the description of the now-defunct
2572 union gimple_statement_d with a diagram showing the
2573 gimple_statement_base class hierarchy and its relationships to
2574 the GSS_ and GIMPLE_ enums.
2576 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
2578 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
2579 * config/aarch64/aarch64.c
2580 (aarch64_cannot_change_mode_class): Weaken conditions.
2581 (aarch64_modes_tieable_p): New.
2582 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
2584 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
2586 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2587 (loadsync_<mode>): Change mode.
2588 (load_quadpti, store_quadpti): New.
2589 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2590 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2592 2014-04-28 Martin Jambor <mjambor@suse.cz>
2594 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
2595 same alias type as the original statement.
2596 (subreplacement_assignment_data): New type.
2597 (handle_unscalarized_data_in_subtree): New type of parameter,
2598 generate new memory accesses with same alias type as the original
2600 (load_assign_lhs_subreplacements): Likewise.
2601 (sra_modify_constructor_assign): Generate new memory accesses with
2602 same alias type as the original statement.
2604 2014-04-28 Richard Biener <rguenther@suse.de>
2606 * tree-pass.h (TODO_verify_il): Define.
2607 (TODO_verify_all): Complete properly.
2608 * passes.c (execute_function_todo): Move existing loop-closed
2609 SSA verification under TODO_verify_il.
2610 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
2611 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
2612 Fix tree sharing issue.
2614 2014-04-28 Richard Biener <rguenther@suse.de>
2617 * builtins.def (DEF_C11_BUILTIN): Add.
2618 (BUILT_IN_ALIGNED_ALLOC): Likewise.
2619 * coretypes.h (enum function_class): Add function_c11_misc.
2620 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
2621 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
2622 (call_may_clobber_ref_p_1): Likewise.
2623 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
2624 (mark_all_reaching_defs_necessary_1): Likewise.
2625 (propagate_necessity): Likewise.
2626 (eliminate_unnecessary_stmts): Likewise.
2627 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
2629 2014-04-28 Richard Biener <rguenther@suse.de>
2631 * tree-vrp.c (vrp_var_may_overflow): Remove.
2632 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
2633 with overflow immediately bump to one before that value and
2634 let iteration figure out overflow status.
2636 2014-04-28 Richard Biener <rguenther@suse.de>
2638 * configure.ac: Do valgrind header checks unconditionally.
2639 Add --enable-valgrind-annotations.
2640 * system.h: Guard valgrind header inclusion with
2641 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
2642 * alloc-pool.c (pool_alloc, pool_free): Use
2643 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
2644 to guard possibly dead code.
2645 * config.in: Regenerated.
2646 * configure: Likewise.
2648 2014-04-28 Jeff Law <law@redhat.com>
2650 PR tree-optimization/60902
2651 * tree-ssa-threadedge.c
2652 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
2653 over real defs when invalidating outputs from statements that do not
2654 produce useful outputs for threading.
2656 2014-04-28 Richard Biener <rguenther@suse.de>
2658 PR tree-optimization/60979
2659 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
2660 SCOPs that end in a block with a successor with abnormal
2663 2014-04-28 Richard Biener <rguenther@suse.de>
2665 * tree-pass.h (execute_pass_list): Adjust prototype.
2666 * passes.c (pass_manager::execute_early_local_passes): Adjust.
2667 (do_per_function): Change callback signature, push all actual
2668 work to the callbals.
2669 (do_per_function_toporder): Likewise.
2670 (execute_function_dump): Adjust.
2671 (execute_function_todo): Likewise.
2672 (clear_last_verified): Likewise.
2673 (verify_curr_properties): Likewise.
2674 (update_properties_after_pass): Likewise.
2675 (execute_pass_list_1): Split out from ...
2676 (execute_pass_list): ... here. Adjust.
2677 (execute_ipa_pass_list): Likewise.
2678 * cgraphunit.c (cgraph_add_new_function): Adjust.
2679 (analyze_function): Likewise.
2680 (expand_function): Likewise.
2681 * cgraph.c (release_function_body): Free dominance info
2682 here instead of asserting it was magically freed elsewhere.
2684 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
2686 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
2687 * configure: Regenerate.
2688 * config/sparc/sparc.opt (muser-mode): New option.
2689 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
2691 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
2692 * doc/invoke.texi (SPARC options): Document -muser-mode.
2694 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
2696 * cselib.c (find_slot_memmode): Delete.
2697 (cselib_hasher): Change compare_type to a struct.
2698 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
2700 (preserve_constants_and_equivs): Adjust for new compare_type.
2701 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
2702 (wrap_constant): Delete.
2703 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
2705 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
2707 * doc/install.texi (Building with profile feedback): Remove
2710 2014-04-26 Tom de Vries <tom@codesourcery.com>
2712 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
2715 2014-04-25 Cary Coutant <ccoutant@google.com>
2718 * dwarf2out.c (should_move_die_to_comdat): A type definition
2719 can contain a subprogram definition, but don't move it to a
2721 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
2722 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
2724 (clone_tree_hash): Rename to...
2725 (clone_tree_partial): ...this; change callers. Copy
2726 DW_TAG_subprogram DIEs as declarations.
2727 (copy_decls_walk): Don't copy children of a declaration into a
2730 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
2733 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
2736 2014-04-25 Jiong Wang <jiong.wang@arm.com>
2738 * config/arm/predicates.md (call_insn_operand): Add long_call check.
2739 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
2741 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
2744 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2746 * config/arm/arm.c (arm_cortex_a8_tune): Initialise T16-related fields.
2748 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2750 PR tree-optimization/60930
2751 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
2752 creating a multiply candidate by folding two constant
2753 multiplicands when the result overflows.
2755 2014-04-25 Jakub Jelinek <jakub@redhat.com>
2757 PR tree-optimization/60960
2758 * tree-vect-generic.c (expand_vector_operation): Only call
2759 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
2761 2014-04-25 Tom de Vries <tom@codesourcery.com>
2763 * expr.c (clobber_reg_mode): New function.
2764 * expr.h (clobber_reg): New function.
2766 2014-04-25 Tom de Vries <tom@codesourcery.com>
2768 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
2771 2014-04-25 Radovan Obradovic <robradovic@mips.com>
2772 Tom de Vries <tom@codesourcery.com>
2774 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
2776 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
2777 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
2778 new argument to find_all_hard_reg_sets call.
2780 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2782 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
2783 Use HOST_WIDE_INT_C for mask literal.
2784 (aarch_rev16_shleft_mask_imm_p): Likewise.
2786 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
2789 * config/sparc/sparc.md (ashlsi3_extend): Delete.
2791 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
2793 PR preprocessor/56540
2794 * config/i386/i386-c.c (ix86_target_macros): Define
2795 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
2797 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2799 * configure.ac (tga_func): Remove.
2800 (LIB_TLS_SPEC): Remove.
2801 * configure: Regenerate.
2802 * config.in: Regenerate.
2803 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
2805 2014-04-25 Richard Biener <rguenther@suse.de>
2808 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
2809 call stmt use/clobber sets during stmt walk instead of
2810 walking the possibly incomplete set of caller edges.
2812 2014-04-25 Richard Biener <rguenther@suse.de>
2815 * passes.c (apply_ipa_transforms): Inline into only caller ...
2816 (execute_one_pass): ... here. Properly bring in function
2817 bodies for nodes we want to apply IPA transforms to.
2819 2014-04-24 Cong Hou <congh@google.com>
2821 PR tree-optimization/60896
2822 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
2823 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
2824 (vect_mark_pattern_stmts): Set the def type of all statements in
2825 PATTERN_DEF_SEQ as vect_internal_def.
2827 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
2829 * doc/extend.texi (PowerPC Built-in Functions): Document new
2830 powerpc extended divide, bcd, pack/unpack 128-bit, builtin functions.
2831 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
2833 * config/rs6000/predicates.md (const_0_to_3_operand): New
2834 predicate to match 0..3 integer constants.
2836 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
2837 to support adding miscellaneous builtin functions.
2838 (BU_DFP_MISC_2): Likewise.
2839 (BU_P7_MISC_1): Likewise.
2840 (BU_P7_MISC_2): Likewise.
2841 (BU_P8V_MISC_3): Likewise.
2842 (BU_MISC_1): Likewise.
2843 (BU_MISC_2): Likewise.
2844 (DIVWE): Add extended divide builtin functions.
2847 (DIVWEUO): Likewise.
2851 (DIVDEUO): Likewise.
2852 (DXEX): Add decimal floating-point builtin functions.
2855 (DDEDPDQ): Likewise.
2857 (DENBCDQ): Likewise.
2864 (CDTBCD): Add new BCD builtin functions.
2868 (BCDADD_LT): Likewise.
2869 (BCDADD_EQ): Likewise.
2870 (BCDADD_GT): Likewise.
2871 (BCDADD_OV): Likewise.
2873 (BCDSUB_LT): Likewise.
2874 (BCDSUB_EQ): Likewise.
2875 (BCDSUB_GT): Likewise.
2876 (BCDSUB_OV): Likewise.
2877 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
2878 (UNPACK_TD): Likewise.
2879 (PACK_TF): Likewise.
2880 (UNPACK_TF): Likewise.
2881 (UNPACK_TF_0): Likewise.
2882 (UNPACK_TF_1): Likewise.
2883 (PACK_V1TI): Likewise.
2884 (UNPACK_V1TI): Likewise.
2886 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
2887 support for decimal floating point builtin functions.
2888 (rs6000_expand_ternop_builtin): Add checks for the new builtin
2889 functions that take constant arguments.
2890 (rs6000_invalid_builtin): Add decimal floating point builtin support.
2891 (rs6000_init_builtins): Setup long double, _Decimal64, and
2892 _Decimal128 types for new builtin functions.
2893 (builtin_function_type): Set the unsigned flags appropriately for
2894 the new builtin functions.
2895 (rs6000_opt_masks): Add support for decimal floating point builtin
2898 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
2899 floating point builtin functions.
2900 (RS6000_BTM_COMMON): Likewise.
2901 (RS6000_BTI_long_double): Likewise.
2902 (RS6000_BTI_dfloat64): Likewise.
2903 (RS6000_BTI_dfloat128): Likewise.
2904 (long_double_type_internal_node): Likewise.
2905 (dfloat64_type_internal_node): Likewise.
2906 (dfloat128_type_internal_node): Likewise.
2908 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
2909 2.07 bcd arithmetic instructions.
2910 (UNSPEC_BCDSUB): Likewise.
2911 (UNSPEC_BCD_OVERFLOW): Likewise.
2912 (UNSPEC_BCD_ADD_SUB): Likewise.
2913 (bcd_add_sub): Likewise.
2914 (BCD_TEST): Likewise.
2915 (bcd<bcd_add_sub>): Likewise.
2916 (bcd<bcd_add_sub>_test): Likewise.
2917 (bcd<bcd_add_sub>_test2): Likewise.
2918 (bcd<bcd_add_sub>_<code>): Likewise.
2919 (peephole2 for combined bcd ops): Likewise.
2921 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
2922 decimal floating point builtin functions.
2923 (UNSPEC_DENBCD): Likewise.
2924 (UNSPEC_DXEX): Likewise.
2925 (UNSPEC_DIEX): Likewise.
2926 (UNSPEC_DSCLI): Likewise.
2927 (UNSPEC_DSCRI): Likewise.
2928 (D64_D128): Likewise.
2929 (dfp_suffix): Likewise.
2930 (dfp_ddedpd_<mode>): Likewise.
2931 (dfp_denbcd_<mode>): Likewise.
2932 (dfp_dxex_<mode>): Likewise.
2933 (dfp_diex_<mode>): Likewise.
2934 (dfp_dscli_<mode>): Likewise.
2935 (dfp_dscri_<mode>): Likewise.
2937 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
2939 (UNSPEC_CDTBCD): Likewise.
2940 (UNSPEC_CBCDTD): Likewise.
2941 (UNSPEC_DIVE): Add support for new extended divide builtin functions.
2942 (UNSPEC_DIVEO): Likewise.
2943 (UNSPEC_DIVEU): Likewise.
2944 (UNSPEC_DIVEUO): Likewise.
2945 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
2946 pack/unpack 128-bit types.
2947 (UNSPEC_PACK_128BIT): Likewise.
2948 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
2949 (udiv<mode>3): Use idiv_ldiv mode attribute.
2950 (div<mode>3): Likewise.
2951 (addg6s): Add new BCD builtin functions.
2954 (UNSPEC_DIV_EXTEND): Add support for new extended divide instructions.
2955 (div_extend): Likewise.
2956 (div<div_extend>_<mode>"): Likewise.
2957 (FP128_64): Add support for new builtin functions to pack/unpack
2959 (unpack<mode>): Likewise.
2960 (unpacktf_0): Likewise.
2961 (unpacktf_1): Likewise.
2962 (unpack<mode>_dm): Likewise.
2963 (unpack<mode>_nodm): Likewise.
2964 (pack<mode>): Likewise.
2965 (unpackv1ti): Likewise.
2966 (packv1ti): Likewise.
2968 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
2970 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
2973 2014-04-24 Jakub Jelinek <jakub@redhat.com>
2975 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
2976 * gimplify.c (omp_is_private): Change last argument's type to int.
2977 Only diagnose lastprivate if the simd argument is 1, only diagnose
2978 linear if the simd argument is 2.
2979 (gimplify_omp_for): Adjust omp_is_private callers. When adding
2980 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
2981 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
2982 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
2983 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
2984 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
2985 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
2986 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
2987 * tree-nested.c (convert_nonlocal_omp_clauses,
2988 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
2990 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
2993 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
2996 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
2998 * flag-types.h (enum ivar_visibility): Add.
3000 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
3002 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
3003 function * argument.
3005 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
3007 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
3009 2014-04-24 Radovan Obradovic <robradovic@mips.com>
3010 Tom de Vries <tom@codesourcery.com>
3012 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
3013 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
3015 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
3016 * emit-rtl.c (try_split): Same.
3018 2014-04-24 Radovan Obradovic <robradovic@mips.com>
3019 Tom de Vries <tom@codesourcery.com>
3021 * common.opt (fuse-caller-save): New option.
3023 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
3025 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of
3026 elements for big-endian.
3028 2014-04-24 Richard Biener <rguenther@suse.de>
3030 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
3031 during TER and instead use the sepops interface for expanding
3032 non-GIMPLE_SINGLE_RHS.
3034 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3036 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
3037 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
3039 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3041 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
3042 assembler 64-bit option.
3043 * configure: Regenerate.
3045 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3047 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
3048 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
3049 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
3050 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
3051 (TARGET_CRYPTO): Take TARGET_SIMD into account.
3053 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3055 * config/aarch64/aarch64-builtins.c
3056 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
3057 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
3058 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
3059 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
3061 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
3062 (Vrevsuff): New mode attribute.
3064 2014-04-24 Terry Guo <terry.guo@arm.com>
3066 * config/arm/arm.h (machine_function): Define variable
3067 after_arm_reorg here.
3068 * config/arm/arm.c (after_arm_reorg): Remove the definition.
3069 (arm_split_constant): Update the way to access variable
3072 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
3074 2014-04-23 Tom de Vries <tom@codesourcery.com>
3076 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
3078 2014-04-23 David Malcolm <dmalcolm@redhat.com>
3080 * is-a.h: Update comments to reflect the following changes to the
3081 "pointerness" of the API, making the template parameter match the
3082 return type, allowing use of is-a.h with typedefs of pointers.
3083 (is_a_helper::cast): Return a T rather then a pointer to a T, so
3084 that the return type matches the parameter to the is_a_helper.
3086 (dyn_cast): Likewise.
3088 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
3089 pointer from the is-a.h API.
3091 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
3092 (is_a_helper <cgraph_node *>::test): ...this, matching change to
3094 (is_a_helper <varpool_node>::test): Likewise, convert to...
3095 (is_a_helper <varpool_node *>::test): ...this.
3097 (varpool_first_variable): Update for removal of implicit pointer
3098 from the is-a.h API.
3099 (varpool_next_variable): Likewise.
3100 (varpool_first_static_initializer): Likewise.
3101 (varpool_next_static_initializer): Likewise.
3102 (varpool_first_defined_variable): Likewise.
3103 (varpool_next_defined_variable): Likewise.
3104 (cgraph_first_defined_function): Likewise.
3105 (cgraph_next_defined_function): Likewise.
3106 (cgraph_first_function): Likewise.
3107 (cgraph_next_function): Likewise.
3108 (cgraph_first_function_with_gimple_body): Likewise.
3109 (cgraph_next_function_with_gimple_body): Likewise.
3110 (cgraph_alias_target): Likewise.
3111 (varpool_alias_target): Likewise.
3112 (cgraph_function_or_thunk_node): Likewise.
3113 (varpool_variable_node): Likewise.
3114 (symtab_real_symbol_p): Likewise.
3115 * cgraphunit.c (referred_to_p): Likewise.
3116 (analyze_functions): Likewise.
3117 (handle_alias_pairs): Likewise.
3118 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
3119 * gimple-ssa.h (gimple_vuse_op): Likewise.
3120 (gimple_vdef_op): Likewise.
3121 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
3122 * gimple.c (gimple_build_asm_1): Likewise.
3123 (gimple_build_try): Likewise.
3124 (gimple_build_resx): Likewise.
3125 (gimple_build_eh_dispatch): Likewise.
3126 (gimple_build_omp_for): Likewise.
3127 (gimple_omp_for_set_clauses): Likewise.
3129 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
3130 (is_a_helper <gimple_statement_asm *>::test): ...this.
3131 (is_a_helper <gimple_statement_bind>::test): Convert to...
3132 (is_a_helper <gimple_statement_bind *>::test): ...this.
3133 (is_a_helper <gimple_statement_call>::test): Convert to...
3134 (is_a_helper <gimple_statement_call *>::test): ...this.
3135 (is_a_helper <gimple_statement_catch>::test): Convert to...
3136 (is_a_helper <gimple_statement_catch *>::test): ...this.
3137 (is_a_helper <gimple_statement_resx>::test): Convert to...
3138 (is_a_helper <gimple_statement_resx *>::test): ...this.
3139 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
3140 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
3141 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
3142 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
3143 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
3144 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
3145 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
3146 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
3147 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
3148 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
3149 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
3150 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
3151 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
3152 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
3153 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
3154 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
3155 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
3156 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
3157 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
3158 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
3159 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
3160 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
3161 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
3162 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
3163 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
3164 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
3165 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
3166 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
3167 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
3168 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
3169 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
3170 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
3171 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
3172 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
3173 (is_a_helper <gimple_statement_phi>::test): Convert to...
3174 (is_a_helper <gimple_statement_phi *>::test): ...this.
3175 (is_a_helper <gimple_statement_transaction>::test): Convert to...
3176 (is_a_helper <gimple_statement_transaction *>::test): ...this.
3177 (is_a_helper <gimple_statement_try>::test): Convert to...
3178 (is_a_helper <gimple_statement_try *>::test): ...this.
3179 (is_a_helper <gimple_statement_wce>::test): Convert to...
3180 (is_a_helper <gimple_statement_wce *>::test): ...this.
3181 (is_a_helper <const gimple_statement_asm>::test): Convert to...
3182 (is_a_helper <const gimple_statement_asm *>::test): ...this.
3183 (is_a_helper <const gimple_statement_bind>::test): Convert to...
3184 (is_a_helper <const gimple_statement_bind *>::test): ...this.
3185 (is_a_helper <const gimple_statement_call>::test): Convert to...
3186 (is_a_helper <const gimple_statement_call *>::test): ...this.
3187 (is_a_helper <const gimple_statement_catch>::test): Convert to...
3188 (is_a_helper <const gimple_statement_catch *>::test): ...this.
3189 (is_a_helper <const gimple_statement_resx>::test): Convert to...
3190 (is_a_helper <const gimple_statement_resx *>::test): ...this.
3191 (is_a_helper <const gimple_statement_eh_dispatch>::test): Convert to...
3192 (is_a_helper <const gimple_statement_eh_dispatch *>::test): ...this.
3193 (is_a_helper <const gimple_statement_eh_filter>::test): Convert to...
3194 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
3195 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
3197 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
3199 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
3201 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
3203 (is_a_helper <const gimple_statement_omp_return>::test): Convert to...
3204 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
3205 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
3207 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
3208 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
3210 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
3211 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
3212 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
3213 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert to...
3214 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
3215 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
3217 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
3218 (is_a_helper <const gimple_statement_omp_target>::test): Convert to...
3219 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
3220 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
3222 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
3223 (is_a_helper <const gimple_statement_omp_single>::test): Convert to...
3224 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
3225 (is_a_helper <const gimple_statement_omp_teams>::test): Convert to...
3226 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
3227 (is_a_helper <const gimple_statement_omp_task>::test): Convert to...
3228 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
3229 (is_a_helper <const gimple_statement_phi>::test): Convert to...
3230 (is_a_helper <const gimple_statement_phi *>::test): ...this.
3231 (is_a_helper <const gimple_statement_transaction>::test): Convert to...
3232 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
3233 (is_a_helper <const gimple_statement_with_ops>::test): Convert to...
3234 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
3235 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
3236 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
3237 (is_a_helper <const gimple_statement_with_memory_ops>::test): Convert
3239 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
3241 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert to...
3242 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
3244 (gimple_use_ops): Update for removal of implicit pointer from the
3246 (gimple_set_use_ops): Likewise.
3247 (gimple_vuse): Likewise.
3248 (gimple_vdef): Likewise.
3249 (gimple_vuse_ptr): Likewise.
3250 (gimple_vdef_ptr): Likewise.
3251 (gimple_set_vuse): Likewise.
3252 (gimple_set_vdef): Likewise.
3253 (gimple_omp_return_set_lhs): Likewise.
3254 (gimple_omp_return_lhs): Likewise.
3255 (gimple_omp_return_lhs_ptr): Likewise.
3256 (gimple_call_fntype): Likewise.
3257 (gimple_call_set_fntype): Likewise.
3258 (gimple_call_set_internal_fn): Likewise.
3259 (gimple_call_use_set): Likewise.
3260 (gimple_call_clobber_set): Likewise.
3261 (gimple_bind_vars): Likewise.
3262 (gimple_bind_set_vars): Likewise.
3263 (gimple_bind_body_ptr): Likewise.
3264 (gimple_bind_set_body): Likewise.
3265 (gimple_bind_add_stmt): Likewise.
3266 (gimple_bind_block): Likewise.
3267 (gimple_bind_set_block): Likewise.
3268 (gimple_asm_ninputs): Likewise.
3269 (gimple_asm_noutputs): Likewise.
3270 (gimple_asm_nclobbers): Likewise.
3271 (gimple_asm_nlabels): Likewise.
3272 (gimple_asm_input_op): Likewise.
3273 (gimple_asm_input_op_ptr): Likewise.
3274 (gimple_asm_output_op): Likewise.
3275 (gimple_asm_output_op_ptr): Likewise.
3276 (gimple_asm_set_output_op): Likewise.
3277 (gimple_asm_clobber_op): Likewise.
3278 (gimple_asm_set_clobber_op): Likewise.
3279 (gimple_asm_label_op): Likewise.
3280 (gimple_asm_set_label_op): Likewise.
3281 (gimple_asm_string): Likewise.
3282 (gimple_catch_types): Likewise.
3283 (gimple_catch_types_ptr): Likewise.
3284 (gimple_catch_handler_ptr): Likewise.
3285 (gimple_catch_set_types): Likewise.
3286 (gimple_catch_set_handler): Likewise.
3287 (gimple_eh_filter_types): Likewise.
3288 (gimple_eh_filter_types_ptr): Likewise.
3289 (gimple_eh_filter_failure_ptr): Likewise.
3290 (gimple_eh_filter_set_types): Likewise.
3291 (gimple_eh_filter_set_failure): Likewise.
3292 (gimple_eh_must_not_throw_fndecl): Likewise.
3293 (gimple_eh_must_not_throw_set_fndecl): Likewise.
3294 (gimple_eh_else_n_body_ptr): Likewise.
3295 (gimple_eh_else_e_body_ptr): Likewise.
3296 (gimple_eh_else_set_n_body): Likewise.
3297 (gimple_eh_else_set_e_body): Likewise.
3298 (gimple_try_eval_ptr): Likewise.
3299 (gimple_try_cleanup_ptr): Likewise.
3300 (gimple_try_set_eval): Likewise.
3301 (gimple_try_set_cleanup): Likewise.
3302 (gimple_wce_cleanup_ptr): Likewise.
3303 (gimple_wce_set_cleanup): Likewise.
3304 (gimple_phi_capacity): Likewise.
3305 (gimple_phi_num_args): Likewise.
3306 (gimple_phi_result): Likewise.
3307 (gimple_phi_result_ptr): Likewise.
3308 (gimple_phi_set_result): Likewise.
3309 (gimple_phi_arg): Likewise.
3310 (gimple_phi_set_arg): Likewise.
3311 (gimple_resx_region): Likewise.
3312 (gimple_resx_set_region): Likewise.
3313 (gimple_eh_dispatch_region): Likewise.
3314 (gimple_eh_dispatch_set_region): Likewise.
3315 (gimple_omp_critical_name): Likewise.
3316 (gimple_omp_critical_name_ptr): Likewise.
3317 (gimple_omp_critical_set_name): Likewise.
3318 (gimple_omp_for_clauses): Likewise.
3319 (gimple_omp_for_clauses_ptr): Likewise.
3320 (gimple_omp_for_set_clauses): Likewise.
3321 (gimple_omp_for_collapse): Likewise.
3322 (gimple_omp_for_index): Likewise.
3323 (gimple_omp_for_index_ptr): Likewise.
3324 (gimple_omp_for_set_index): Likewise.
3325 (gimple_omp_for_initial): Likewise.
3326 (gimple_omp_for_initial_ptr): Likewise.
3327 (gimple_omp_for_set_initial): Likewise.
3328 (gimple_omp_for_final): Likewise.
3329 (gimple_omp_for_final_ptr): Likewise.
3330 (gimple_omp_for_set_final): Likewise.
3331 (gimple_omp_for_incr): Likewise.
3332 (gimple_omp_for_incr_ptr): Likewise.
3333 (gimple_omp_for_set_incr): Likewise.
3334 (gimple_omp_for_pre_body_ptr): Likewise.
3335 (gimple_omp_for_set_pre_body): Likewise.
3336 (gimple_omp_parallel_clauses): Likewise.
3337 (gimple_omp_parallel_clauses_ptr): Likewise.
3338 (gimple_omp_parallel_set_clauses): Likewise.
3339 (gimple_omp_parallel_child_fn): Likewise.
3340 (gimple_omp_parallel_child_fn_ptr): Likewise.
3341 (gimple_omp_parallel_set_child_fn): Likewise.
3342 (gimple_omp_parallel_data_arg): Likewise.
3343 (gimple_omp_parallel_data_arg_ptr): Likewise.
3344 (gimple_omp_parallel_set_data_arg): Likewise.
3345 (gimple_omp_task_clauses): Likewise.
3346 (gimple_omp_task_clauses_ptr): Likewise.
3347 (gimple_omp_task_set_clauses): Likewise.
3348 (gimple_omp_task_child_fn): Likewise.
3349 (gimple_omp_task_child_fn_ptr): Likewise.
3350 (gimple_omp_task_set_child_fn): Likewise.
3351 (gimple_omp_task_data_arg): Likewise.
3352 (gimple_omp_task_data_arg_ptr): Likewise.
3353 (gimple_omp_task_set_data_arg): Likewise.
3354 (gimple_omp_taskreg_clauses): Likewise.
3355 (gimple_omp_taskreg_clauses_ptr): Likewise.
3356 (gimple_omp_taskreg_set_clauses): Likewise.
3357 (gimple_omp_taskreg_child_fn): Likewise.
3358 (gimple_omp_taskreg_child_fn_ptr): Likewise.
3359 (gimple_omp_taskreg_set_child_fn): Likewise.
3360 (gimple_omp_taskreg_data_arg): Likewise.
3361 (gimple_omp_taskreg_data_arg_ptr): Likewise.
3362 (gimple_omp_taskreg_set_data_arg): Likewise.
3363 (gimple_omp_task_copy_fn): Likewise.
3364 (gimple_omp_task_copy_fn_ptr): Likewise.
3365 (gimple_omp_task_set_copy_fn): Likewise.
3366 (gimple_omp_task_arg_size): Likewise.
3367 (gimple_omp_task_arg_size_ptr): Likewise.
3368 (gimple_omp_task_set_arg_size): Likewise.
3369 (gimple_omp_task_arg_align): Likewise.
3370 (gimple_omp_task_arg_align_ptr): Likewise.
3371 (gimple_omp_task_set_arg_align): Likewise.
3372 (gimple_omp_single_clauses): Likewise.
3373 (gimple_omp_single_clauses_ptr): Likewise.
3374 (gimple_omp_single_set_clauses): Likewise.
3375 (gimple_omp_target_clauses): Likewise.
3376 (gimple_omp_target_clauses_ptr): Likewise.
3377 (gimple_omp_target_set_clauses): Likewise.
3378 (gimple_omp_target_child_fn): Likewise.
3379 (gimple_omp_target_child_fn_ptr): Likewise.
3380 (gimple_omp_target_set_child_fn): Likewise.
3381 (gimple_omp_target_data_arg): Likewise.
3382 (gimple_omp_target_data_arg_ptr): Likewise.
3383 (gimple_omp_target_set_data_arg): Likewise.
3384 (gimple_omp_teams_clauses): Likewise.
3385 (gimple_omp_teams_clauses_ptr): Likewise.
3386 (gimple_omp_teams_set_clauses): Likewise.
3387 (gimple_omp_sections_clauses): Likewise.
3388 (gimple_omp_sections_clauses_ptr): Likewise.
3389 (gimple_omp_sections_set_clauses): Likewise.
3390 (gimple_omp_sections_control): Likewise.
3391 (gimple_omp_sections_control_ptr): Likewise.
3392 (gimple_omp_sections_set_control): Likewise.
3393 (gimple_omp_for_set_cond): Likewise.
3394 (gimple_omp_for_cond): Likewise.
3395 (gimple_omp_atomic_store_set_val): Likewise.
3396 (gimple_omp_atomic_store_val): Likewise.
3397 (gimple_omp_atomic_store_val_ptr): Likewise.
3398 (gimple_omp_atomic_load_set_lhs): Likewise.
3399 (gimple_omp_atomic_load_lhs): Likewise.
3400 (gimple_omp_atomic_load_lhs_ptr): Likewise.
3401 (gimple_omp_atomic_load_set_rhs): Likewise.
3402 (gimple_omp_atomic_load_rhs): Likewise.
3403 (gimple_omp_atomic_load_rhs_ptr): Likewise.
3404 (gimple_omp_continue_control_def): Likewise.
3405 (gimple_omp_continue_control_def_ptr): Likewise.
3406 (gimple_omp_continue_set_control_def): Likewise.
3407 (gimple_omp_continue_control_use): Likewise.
3408 (gimple_omp_continue_control_use_ptr): Likewise.
3409 (gimple_omp_continue_set_control_use): Likewise.
3410 (gimple_transaction_body_ptr): Likewise.
3411 (gimple_transaction_label): Likewise.
3412 (gimple_transaction_label_ptr): Likewise.
3413 (gimple_transaction_set_body): Likewise.
3414 (gimple_transaction_set_label): Likewise.
3416 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
3417 * ipa-inline-analysis.c (inline_write_summary): Likewise.
3418 * ipa-ref.c (ipa_record_reference): Likewise.
3419 * ipa-reference.c (analyze_function): Likewise.
3420 (ipa_reference_write_optimization_summary): Likewise.
3421 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
3422 (address_taken_from_non_vtable_p): Likewise.
3423 (comdat_can_be_unshared_p_1): Likewise.
3424 * lto-cgraph.c (lto_output_ref): Likewise.
3425 (add_references): Likewise.
3426 (compute_ltrans_boundary): Likewise.
3427 (output_symtab): Likewise.
3428 (input_ref): Likewise.
3429 (input_cgraph_1): Likewise.
3430 (output_cgraph_opt_summary): Likewise.
3431 * lto-streamer-out.c (lto_output): Likewise.
3432 (output_symbol_p): Likewise.
3433 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
3434 (lsei_start_function_in_partition): Likewise.
3435 (lsei_next_variable_in_partition): Likewise.
3436 (lsei_start_variable_in_partition): Likewise.
3437 * symtab.c (insert_to_assembler_name_hash): Likewise.
3438 (unlink_from_assembler_name_hash): Likewise.
3439 (symtab_unregister_node): Likewise.
3440 (symtab_remove_node): Likewise.
3441 (dump_symtab_node): Likewise.
3442 (verify_symtab_base): Likewise.
3443 (verify_symtab_node): Likewise.
3444 (symtab_make_decl_local): Likewise.
3445 (symtab_alias_ultimate_target): Likewise.
3446 (symtab_resolve_alias): Likewise.
3447 (symtab_get_symbol_partitioning_class): Likewise.
3448 * tree-phinodes.c (allocate_phi_node): Likewise.
3449 (reserve_phi_args_for_new_edge): Likewise.
3450 (remove_phi_args): Likewise.
3451 * varpool.c (varpool_node_for_asm): Likewise.
3452 (varpool_remove_unreferenced_decls): Likewise.
3454 2014-04-23 Jeff Law <law@redhat.com>
3456 PR tree-optimization/60902
3457 * tree-ssa-threadedge.c
3458 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
3459 invalidate outputs from statements that do not produce useful
3460 outputs for threading.
3462 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
3464 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
3465 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
3466 machine descriptions for Stack Smashing Protector.
3468 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
3470 * aarch64.md (<optab>_rol<mode>3): New pattern.
3471 (<optab>_rolsi3_uxtw): Likewise.
3472 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
3474 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
3476 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
3477 (arm_cortex_a12_tune): Likewise.
3479 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3481 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
3483 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3485 * config/arm/arm.md (arm_rev16si2): New pattern.
3486 (arm_rev16si2_alt): Likewise.
3487 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
3489 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3491 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
3492 (rev16<mode>2_alt): Likewise.
3493 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
3494 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
3495 (aarch_rev16_shleft_mask_imm_p): Likewise.
3496 (aarch_rev16_p_1): Likewise.
3497 (aarch_rev16_p): Likewise.
3498 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
3499 (aarch_rev16_shright_mask_imm_p): Likewise.
3500 (aarch_rev16_shleft_mask_imm_p): Likewise.
3502 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3504 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
3505 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
3507 (cortex_a53_extra_costs): Likewise.
3508 (cortex_a57_extra_costs): Likewise.
3509 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
3510 (cortexa7_extra_costs): Likewise.
3511 (cortexa8_extra_costs): Likewise.
3512 (cortexa12_extra_costs): Likewise.
3513 (cortexa15_extra_costs): Likewise.
3514 (v7m_extra_costs): Likewise.
3515 (arm_new_rtx_costs): Handle BSWAP.
3517 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3519 * config/arm/arm.c (cortexa8_extra_costs): New table.
3520 (arm_cortex_a8_tune): New tuning struct.
3521 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
3523 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3525 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
3527 2014-04-23 Richard Biener <rguenther@suse.de>
3529 * Makefile.in (OBJS): Remove loop-unswitch.o.
3530 * tree-pass.h (make_pass_rtl_unswitch): Remove.
3531 * passes.def (pass_rtl_unswitch): Likewise.
3532 * loop-init.c (gate_rtl_unswitch): Likewise.
3533 (rtl_unswitch): Likewise.
3534 (pass_data_rtl_unswitch): Likewise.
3535 (pass_rtl_unswitch): Likewise.
3536 (make_pass_rtl_unswitch): Likewise.
3537 * rtl.h (reversed_condition): Likewise.
3538 (compare_and_jump_seq): Likewise.
3539 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
3541 * loop-unroll.c (compare_and_jump_seq): Likewise.
3543 2014-04-23 Richard Biener <rguenther@suse.de>
3545 PR tree-optimization/60903
3546 * tree-ssa-loop-im.c (analyze_memory_references): Remove
3547 commented code block.
3548 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
3549 loop flags to newly created BBs and edges.
3551 2014-04-23 Nick Clifton <nickc@redhat.com>
3553 * config/msp430/msp430.c (msp430_handle_option): Move function
3555 (msp430_option_override): Simplify mcu and mcpu option handling.
3556 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
3557 support for -mhwmult command line option.
3558 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
3559 -mhwmult command line option.
3560 (msp430_hwmult_enabled): Delete.
3561 (msp43o_output_labelref): Add support for -mhwmult command line option.
3562 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
3563 (umulsidi3): Likewise.
3564 * config/msp430/msp430.opt (mmcu): Add Report attribute.
3565 (mcpu, mlarge, msmall): Likewise.
3566 (mhwmult): New option.
3567 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
3569 (msp430_is_f5_mcu): Remove prototype.
3570 (msp430_use_f5_series_hwmult): Add prototype.
3571 * config/msp430/msp430-opts.h: New file.
3572 * common/config/msp430: New directory.
3573 * common/config/msp430/msp430-common.c: New file.
3574 * config.gcc (msp430): Remove target_has_targetm_common.
3575 * doc/invoke.texi: Document -mhwmult command line option.
3577 2014-04-23 Nick Clifton <nickc@redhat.com>
3579 * config/i386/cygwin.h (ENDFILE_SPEC): Include
3580 default-manifest.o if it can be found in the search path.
3581 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
3583 2014-04-23 Terry Guo <terry.guo@arm.com>
3585 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
3587 2014-04-23 Richard Biener <rguenther@suse.de>
3590 * tree-inline.c (declare_return_variable): Use mark_addressable.
3592 2014-04-23 Richard Biener <rguenther@suse.de>
3595 * loop-init.c (loop_optimizer_init): Make sure to apply
3596 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
3598 2014-04-22 Jakub Jelinek <jakub@redhat.com>
3601 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
3603 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
3604 if flag_sanitize_undefined_trap_on_error.
3605 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
3606 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
3607 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
3608 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
3609 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
3610 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
3611 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
3612 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
3613 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
3614 * ubsan.c (ubsan_instrument_unreachable): Return
3615 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
3616 (ubsan_expand_null_ifn): Emit __builtin_trap ()
3617 if flag_sanitize_undefined_trap_on_error and
3618 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
3619 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
3620 instrument_bool_enum_load): Emit __builtin_trap () if
3621 flag_sanitize_undefined_trap_on_error and
3622 __builtin_handle_*_abort () if !flag_sanitize_recover.
3623 * doc/invoke.texi (-fsanitize-recover,
3624 -fsanitize-undefined-trap-on-error): Document.
3626 2014-04-22 Christian Bruel <christian.bruel@st.com>
3628 * config/sh/sh.md (mov<mode>): Replace movQIHI.
3629 Force immediates to SImode.
3631 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
3633 * config/nios2/nios2.md (UNSPEC_ROUND): New.
3635 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
3636 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
3637 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
3638 (nios2_fpu_insn): Add entry for round.
3639 (N2FPU_NO_ERRNO_P): Define.
3640 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
3642 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
3644 2014-04-22 Richard Henderson <rth@redhat.com>
3646 * config/aarch64/aarch64 (addti3, subti3): New expanders.
3647 (add<GPI>3_compare0): Remove leading * from name.
3648 (add<GPI>3_carryin): Likewise.
3649 (sub<GPI>3_compare0): Likewise.
3650 (sub<GPI>3_carryin): Likewise.
3651 (<su_optab>mulditi3): New expander.
3652 (multi3): New expander.
3653 (madd<GPI>): Remove leading * from name.
3655 2014-04-22 Martin Jambor <mjambor@suse.cz>
3657 * cgraphclones.c (cgraph_function_versioning): Copy
3658 ipa_transforms_to_apply instead of asserting it is empty.
3660 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
3663 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
3664 on count_exp to get mode.
3666 2014-04-22 Andrew Pinski <apinski@cavium.com>
3668 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
3669 Handle TLS for ILP32.
3670 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
3671 (tlsie_small_<mode>): this and handle PTR.
3672 (tlsie_small_sidi): New pattern.
3673 (tlsle_small): Change to an expand to handle ILP32.
3674 (tlsle_small_<mode>): New pattern.
3675 (tlsdesc_small): Rename to ...
3676 (tlsdesc_small_<mode>): this and handle PTR.
3678 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3680 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
3682 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3684 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
3685 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
3686 (aarch64_types_signed_poly_qualifiers): Likewise.
3687 (aarch64_types_unsigned_signed_qualifiers): Likewise.
3688 (aarch64_types_poly_signed_qualifiers): Likewise.
3689 (TYPES_REINTERP_SS): Type macro added.
3690 (TYPES_REINTERP_SU): Likewise.
3691 (TYPES_REINTERP_SP): Likewise.
3692 (TYPES_REINTERP_US): Likewise.
3693 (TYPES_REINTERP_PS): Likewise.
3694 (aarch64_fold_builtin): New expression folding added.
3695 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
3696 Declarations removed.
3697 (REINTERP_SS): Declarations added.
3698 (REINTERP_US): Likewise.
3699 (REINTERP_PS): Likewise.
3700 (REINTERP_SU): Likewise.
3701 (REINTERP_SP): Likewise.
3702 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
3703 (vreinterpretq_p8_f64): Likewise.
3704 (vreinterpret_p16_f64): Likewise.
3705 (vreinterpretq_p16_f64): Likewise.
3706 (vreinterpret_f32_f64): Likewise.
3707 (vreinterpretq_f32_f64): Likewise.
3708 (vreinterpret_f64_f32): Likewise.
3709 (vreinterpret_f64_p8): Likewise.
3710 (vreinterpret_f64_p16): Likewise.
3711 (vreinterpret_f64_s8): Likewise.
3712 (vreinterpret_f64_s16): Likewise.
3713 (vreinterpret_f64_s32): Likewise.
3714 (vreinterpret_f64_s64): Likewise.
3715 (vreinterpret_f64_u8): Likewise.
3716 (vreinterpret_f64_u16): Likewise.
3717 (vreinterpret_f64_u32): Likewise.
3718 (vreinterpret_f64_u64): Likewise.
3719 (vreinterpretq_f64_f32): Likewise.
3720 (vreinterpretq_f64_p8): Likewise.
3721 (vreinterpretq_f64_p16): Likewise.
3722 (vreinterpretq_f64_s8): Likewise.
3723 (vreinterpretq_f64_s16): Likewise.
3724 (vreinterpretq_f64_s32): Likewise.
3725 (vreinterpretq_f64_s64): Likewise.
3726 (vreinterpretq_f64_u8): Likewise.
3727 (vreinterpretq_f64_u16): Likewise.
3728 (vreinterpretq_f64_u32): Likewise.
3729 (vreinterpretq_f64_u64): Likewise.
3730 (vreinterpret_s64_f64): Likewise.
3731 (vreinterpretq_s64_f64): Likewise.
3732 (vreinterpret_u64_f64): Likewise.
3733 (vreinterpretq_u64_f64): Likewise.
3734 (vreinterpret_s8_f64): Likewise.
3735 (vreinterpretq_s8_f64): Likewise.
3736 (vreinterpret_s16_f64): Likewise.
3737 (vreinterpretq_s16_f64): Likewise.
3738 (vreinterpret_s32_f64): Likewise.
3739 (vreinterpretq_s32_f64): Likewise.
3740 (vreinterpret_u8_f64): Likewise.
3741 (vreinterpretq_u8_f64): Likewise.
3742 (vreinterpret_u16_f64): Likewise.
3743 (vreinterpretq_u16_f64): Likewise.
3744 (vreinterpret_u32_f64): Likewise.
3745 (vreinterpretq_u32_f64): Likewise.
3747 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3749 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
3750 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
3751 (vreinterpret_p8_s8): Likewise.
3752 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
3753 (vreinterpret_p8_s16): Likewise.
3754 (vreinterpret_p8_s32): Likewise.
3755 (vreinterpret_p8_s64): Likewise.
3756 (vreinterpret_p8_f32): Likewise.
3757 (vreinterpret_p8_u8): Likewise.
3758 (vreinterpret_p8_u16): Likewise.
3759 (vreinterpret_p8_u32): Likewise.
3760 (vreinterpret_p8_u64): Likewise.
3761 (vreinterpret_p8_p16): Likewise.
3762 (vreinterpretq_p8_s8): Likewise.
3763 (vreinterpretq_p8_s16): Likewise.
3764 (vreinterpretq_p8_s32): Likewise.
3765 (vreinterpretq_p8_s64): Likewise.
3766 (vreinterpretq_p8_f32): Likewise.
3767 (vreinterpretq_p8_u8): Likewise.
3768 (vreinterpretq_p8_u16): Likewise.
3769 (vreinterpretq_p8_u32): Likewise.
3770 (vreinterpretq_p8_u64): Likewise.
3771 (vreinterpretq_p8_p16): Likewise.
3772 (vreinterpret_p16_s8): Likewise.
3773 (vreinterpret_p16_s16): Likewise.
3774 (vreinterpret_p16_s32): Likewise.
3775 (vreinterpret_p16_s64): Likewise.
3776 (vreinterpret_p16_f32): Likewise.
3777 (vreinterpret_p16_u8): Likewise.
3778 (vreinterpret_p16_u16): Likewise.
3779 (vreinterpret_p16_u32): Likewise.
3780 (vreinterpret_p16_u64): Likewise.
3781 (vreinterpret_p16_p8): Likewise.
3782 (vreinterpretq_p16_s8): Likewise.
3783 (vreinterpretq_p16_s16): Likewise.
3784 (vreinterpretq_p16_s32): Likewise.
3785 (vreinterpretq_p16_s64): Likewise.
3786 (vreinterpretq_p16_f32): Likewise.
3787 (vreinterpretq_p16_u8): Likewise.
3788 (vreinterpretq_p16_u16): Likewise.
3789 (vreinterpretq_p16_u32): Likewise.
3790 (vreinterpretq_p16_u64): Likewise.
3791 (vreinterpretq_p16_p8): Likewise.
3792 (vreinterpret_f32_s8): Likewise.
3793 (vreinterpret_f32_s16): Likewise.
3794 (vreinterpret_f32_s32): Likewise.
3795 (vreinterpret_f32_s64): Likewise.
3796 (vreinterpret_f32_u8): Likewise.
3797 (vreinterpret_f32_u16): Likewise.
3798 (vreinterpret_f32_u32): Likewise.
3799 (vreinterpret_f32_u64): Likewise.
3800 (vreinterpret_f32_p8): Likewise.
3801 (vreinterpret_f32_p16): Likewise.
3802 (vreinterpretq_f32_s8): Likewise.
3803 (vreinterpretq_f32_s16): Likewise.
3804 (vreinterpretq_f32_s32): Likewise.
3805 (vreinterpretq_f32_s64): Likewise.
3806 (vreinterpretq_f32_u8): Likewise.
3807 (vreinterpretq_f32_u16): Likewise.
3808 (vreinterpretq_f32_u32): Likewise.
3809 (vreinterpretq_f32_u64): Likewise.
3810 (vreinterpretq_f32_p8): Likewise.
3811 (vreinterpretq_f32_p16): Likewise.
3812 (vreinterpret_s64_s8): Likewise.
3813 (vreinterpret_s64_s16): Likewise.
3814 (vreinterpret_s64_s32): Likewise.
3815 (vreinterpret_s64_f32): Likewise.
3816 (vreinterpret_s64_u8): Likewise.
3817 (vreinterpret_s64_u16): Likewise.
3818 (vreinterpret_s64_u32): Likewise.
3819 (vreinterpret_s64_u64): Likewise.
3820 (vreinterpret_s64_p8): Likewise.
3821 (vreinterpret_s64_p16): Likewise.
3822 (vreinterpretq_s64_s8): Likewise.
3823 (vreinterpretq_s64_s16): Likewise.
3824 (vreinterpretq_s64_s32): Likewise.
3825 (vreinterpretq_s64_f32): Likewise.
3826 (vreinterpretq_s64_u8): Likewise.
3827 (vreinterpretq_s64_u16): Likewise.
3828 (vreinterpretq_s64_u32): Likewise.
3829 (vreinterpretq_s64_u64): Likewise.
3830 (vreinterpretq_s64_p8): Likewise.
3831 (vreinterpretq_s64_p16): Likewise.
3832 (vreinterpret_u64_s8): Likewise.
3833 (vreinterpret_u64_s16): Likewise.
3834 (vreinterpret_u64_s32): Likewise.
3835 (vreinterpret_u64_s64): Likewise.
3836 (vreinterpret_u64_f32): Likewise.
3837 (vreinterpret_u64_u8): Likewise.
3838 (vreinterpret_u64_u16): Likewise.
3839 (vreinterpret_u64_u32): Likewise.
3840 (vreinterpret_u64_p8): Likewise.
3841 (vreinterpret_u64_p16): Likewise.
3842 (vreinterpretq_u64_s8): Likewise.
3843 (vreinterpretq_u64_s16): Likewise.
3844 (vreinterpretq_u64_s32): Likewise.
3845 (vreinterpretq_u64_s64): Likewise.
3846 (vreinterpretq_u64_f32): Likewise.
3847 (vreinterpretq_u64_u8): Likewise.
3848 (vreinterpretq_u64_u16): Likewise.
3849 (vreinterpretq_u64_u32): Likewise.
3850 (vreinterpretq_u64_p8): Likewise.
3851 (vreinterpretq_u64_p16): Likewise.
3852 (vreinterpret_s8_s16): Likewise.
3853 (vreinterpret_s8_s32): Likewise.
3854 (vreinterpret_s8_s64): Likewise.
3855 (vreinterpret_s8_f32): Likewise.
3856 (vreinterpret_s8_u8): Likewise.
3857 (vreinterpret_s8_u16): Likewise.
3858 (vreinterpret_s8_u32): Likewise.
3859 (vreinterpret_s8_u64): Likewise.
3860 (vreinterpret_s8_p8): Likewise.
3861 (vreinterpret_s8_p16): Likewise.
3862 (vreinterpretq_s8_s16): Likewise.
3863 (vreinterpretq_s8_s32): Likewise.
3864 (vreinterpretq_s8_s64): Likewise.
3865 (vreinterpretq_s8_f32): Likewise.
3866 (vreinterpretq_s8_u8): Likewise.
3867 (vreinterpretq_s8_u16): Likewise.
3868 (vreinterpretq_s8_u32): Likewise.
3869 (vreinterpretq_s8_u64): Likewise.
3870 (vreinterpretq_s8_p8): Likewise.
3871 (vreinterpretq_s8_p16): Likewise.
3872 (vreinterpret_s16_s8): Likewise.
3873 (vreinterpret_s16_s32): Likewise.
3874 (vreinterpret_s16_s64): Likewise.
3875 (vreinterpret_s16_f32): Likewise.
3876 (vreinterpret_s16_u8): Likewise.
3877 (vreinterpret_s16_u16): Likewise.
3878 (vreinterpret_s16_u32): Likewise.
3879 (vreinterpret_s16_u64): Likewise.
3880 (vreinterpret_s16_p8): Likewise.
3881 (vreinterpret_s16_p16): Likewise.
3882 (vreinterpretq_s16_s8): Likewise.
3883 (vreinterpretq_s16_s32): Likewise.
3884 (vreinterpretq_s16_s64): Likewise.
3885 (vreinterpretq_s16_f32): Likewise.
3886 (vreinterpretq_s16_u8): Likewise.
3887 (vreinterpretq_s16_u16): Likewise.
3888 (vreinterpretq_s16_u32): Likewise.
3889 (vreinterpretq_s16_u64): Likewise.
3890 (vreinterpretq_s16_p8): Likewise.
3891 (vreinterpretq_s16_p16): Likewise.
3892 (vreinterpret_s32_s8): Likewise.
3893 (vreinterpret_s32_s16): Likewise.
3894 (vreinterpret_s32_s64): Likewise.
3895 (vreinterpret_s32_f32): Likewise.
3896 (vreinterpret_s32_u8): Likewise.
3897 (vreinterpret_s32_u16): Likewise.
3898 (vreinterpret_s32_u32): Likewise.
3899 (vreinterpret_s32_u64): Likewise.
3900 (vreinterpret_s32_p8): Likewise.
3901 (vreinterpret_s32_p16): Likewise.
3902 (vreinterpretq_s32_s8): Likewise.
3903 (vreinterpretq_s32_s16): Likewise.
3904 (vreinterpretq_s32_s64): Likewise.
3905 (vreinterpretq_s32_f32): Likewise.
3906 (vreinterpretq_s32_u8): Likewise.
3907 (vreinterpretq_s32_u16): Likewise.
3908 (vreinterpretq_s32_u32): Likewise.
3909 (vreinterpretq_s32_u64): Likewise.
3910 (vreinterpretq_s32_p8): Likewise.
3911 (vreinterpretq_s32_p16): Likewise.
3912 (vreinterpret_u8_s8): Likewise.
3913 (vreinterpret_u8_s16): Likewise.
3914 (vreinterpret_u8_s32): Likewise.
3915 (vreinterpret_u8_s64): Likewise.
3916 (vreinterpret_u8_f32): Likewise.
3917 (vreinterpret_u8_u16): Likewise.
3918 (vreinterpret_u8_u32): Likewise.
3919 (vreinterpret_u8_u64): Likewise.
3920 (vreinterpret_u8_p8): Likewise.
3921 (vreinterpret_u8_p16): Likewise.
3922 (vreinterpretq_u8_s8): Likewise.
3923 (vreinterpretq_u8_s16): Likewise.
3924 (vreinterpretq_u8_s32): Likewise.
3925 (vreinterpretq_u8_s64): Likewise.
3926 (vreinterpretq_u8_f32): Likewise.
3927 (vreinterpretq_u8_u16): Likewise.
3928 (vreinterpretq_u8_u32): Likewise.
3929 (vreinterpretq_u8_u64): Likewise.
3930 (vreinterpretq_u8_p8): Likewise.
3931 (vreinterpretq_u8_p16): Likewise.
3932 (vreinterpret_u16_s8): Likewise.
3933 (vreinterpret_u16_s16): Likewise.
3934 (vreinterpret_u16_s32): Likewise.
3935 (vreinterpret_u16_s64): Likewise.
3936 (vreinterpret_u16_f32): Likewise.
3937 (vreinterpret_u16_u8): Likewise.
3938 (vreinterpret_u16_u32): Likewise.
3939 (vreinterpret_u16_u64): Likewise.
3940 (vreinterpret_u16_p8): Likewise.
3941 (vreinterpret_u16_p16): Likewise.
3942 (vreinterpretq_u16_s8): Likewise.
3943 (vreinterpretq_u16_s16): Likewise.
3944 (vreinterpretq_u16_s32): Likewise.
3945 (vreinterpretq_u16_s64): Likewise.
3946 (vreinterpretq_u16_f32): Likewise.
3947 (vreinterpretq_u16_u8): Likewise.
3948 (vreinterpretq_u16_u32): Likewise.
3949 (vreinterpretq_u16_u64): Likewise.
3950 (vreinterpretq_u16_p8): Likewise.
3951 (vreinterpretq_u16_p16): Likewise.
3952 (vreinterpret_u32_s8): Likewise.
3953 (vreinterpret_u32_s16): Likewise.
3954 (vreinterpret_u32_s32): Likewise.
3955 (vreinterpret_u32_s64): Likewise.
3956 (vreinterpret_u32_f32): Likewise.
3957 (vreinterpret_u32_u8): Likewise.
3958 (vreinterpret_u32_u16): Likewise.
3959 (vreinterpret_u32_u64): Likewise.
3960 (vreinterpret_u32_p8): Likewise.
3961 (vreinterpret_u32_p16): Likewise.
3962 (vreinterpretq_u32_s8): Likewise.
3963 (vreinterpretq_u32_s16): Likewise.
3964 (vreinterpretq_u32_s32): Likewise.
3965 (vreinterpretq_u32_s64): Likewise.
3966 (vreinterpretq_u32_f32): Likewise.
3967 (vreinterpretq_u32_u8): Likewise.
3968 (vreinterpretq_u32_u16): Likewise.
3969 (vreinterpretq_u32_u64): Likewise.
3970 (vreinterpretq_u32_p8): Likewise.
3971 (vreinterpretq_u32_p16): Likewise.
3973 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3975 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
3977 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator extended.
3979 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
3980 (vqnegd_s64): Likewise.
3981 (vqabs_s64): Likewise.
3982 (vqabsd_s64): Likewise.
3984 2014-04-22 Richard Henderson <rth@redhat.com>
3986 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
3987 computation to the top of the loop.
3989 2014-04-22 Renlin <renlin.li@arm.com>
3990 Jiong Wang <jiong.wang@arm.com>
3992 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
3993 * config/aarch64/aarch64.c (aarch64_layout_frame)
3994 (aarch64_initial_elimination_offset): Likewise.
3996 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
3998 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
4001 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
4003 * machmode.h (bitwise_mode_for_mode): Declare.
4004 * stor-layout.h (bitwise_type_for_mode): Likewise.
4005 * stor-layout.c (bitwise_mode_for_mode): New function.
4006 (bitwise_type_for_mode): Likewise.
4007 * builtins.c (fold_builtin_memory_op): Use it instead of
4008 int_mode_for_mode and build_nonstandard_integer_type.
4010 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4012 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
4013 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
4014 (*-*-solaris2*): Simplify.
4015 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
4016 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
4017 *-*-solaris2.9* handling.
4019 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
4021 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
4022 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
4024 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
4025 * configure: Regenerate.
4027 * config/i386/sol2-9.h: Remove.
4029 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
4030 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
4031 Remove Solaris 9 references.
4033 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
4035 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
4036 (floatuns<GPI:mode><GPF:mode>2): Remove.
4037 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
4038 and floatuns conversions.
4039 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
4040 and floatuns conversions.
4041 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
4042 (w1,w2): New mode attributes for inequal width conversions.
4044 2014-04-22 Renlin Li <Renlin.Li@arm.com>
4046 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
4047 the output asm format.
4049 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
4051 * config/aarch64/aarch64-simd.md
4052 (aarch64_cm<optab>di): Always split.
4053 (*aarch64_cm<optab>di): New.
4054 (aarch64_cmtstdi): Always split.
4055 (*aarch64_cmtstdi): New.
4057 2014-04-22 Jakub Jelinek <jakub@redhat.com>
4059 PR tree-optimization/60823
4060 * omp-low.c (ipa_simd_modify_function_body): Go through
4061 all SSA_NAMEs and for those refering to vector arguments
4062 which are going to be replaced adjust SSA_NAME_VAR and,
4063 if it is a default definition, change it into a non-default
4064 definition assigned at the beginning of function from new_decl.
4065 (ipa_simd_modify_stmt_ops): Rewritten.
4066 * tree-dfa.c (set_ssa_default_def): When removing default def,
4067 check for NULL loc instead of NULL *loc.
4069 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4071 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
4072 restrictions on core registers for DImode values in Thumb2.
4074 2014-04-22 Ian Bolton <ian.bolton@arm.com>
4076 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
4077 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
4079 2014-04-22 Ian Bolton <ian.bolton@arm.com>
4081 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
4082 (*iordi_notzesidi_di): Likewise.
4083 (*iordi_notsesidi_di): Likewise.
4085 2014-04-22 Ian Bolton <ian.bolton@arm.com>
4087 * config/arm/arm-protos.h (tune_params): New struct members.
4088 * config/arm/arm.c: Initialise tune_params per processor.
4089 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
4090 for speed, based on new tune_params.
4092 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
4094 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro added.
4095 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added macro.
4096 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment corrected.
4097 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
4098 * config/aarch64/arm_neon.h (vrnd_f64): Added.
4099 (vrnda_f64): Likewise.
4100 (vrndi_f64): Likewise.
4101 (vrndm_f64): Likewise.
4102 (vrndn_f64): Likewise.
4103 (vrndp_f64): Likewise.
4104 (vrndx_f64): Likewise.
4106 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
4108 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
4109 GET_MODE_SIZE argument is enum machine_mode.
4111 2014-04-22 Jakub Jelinek <jakub@redhat.com>
4114 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
4115 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
4117 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
4120 * asan.c (asan_emit_stack_protection): Force the base to align to
4121 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
4122 appropriate bits if STRICT_ALIGNMENT.
4123 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
4125 (expand_used_vars): Leave a space in the stack frame for alignment
4126 if STRICT_ALIGNMENT.
4128 2014-04-21 David Malcolm <dmalcolm@redhat.com>
4130 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
4132 (gimple_store_p): Likewise.
4133 (gimple_assign_load_p): Likewise.
4134 (gimple_assign_cast_p): Likewise.
4135 (gimple_clobber_p): Likewise.
4137 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
4138 rather than a gimple.
4139 (gimple_assign_cast_p): Likewise.
4141 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
4144 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
4145 If mode is DDmode and TARGET_E500_DOUBLE allow move.
4147 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
4148 more debug information for E500 if -mdebug=reg.
4150 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
4153 * config/i386/i386.c (ix86_expand_builtin)
4154 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
4155 register for target RTX.
4156 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
4158 2014-04-18 Cong Hou <congh@google.com>
4160 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
4161 the widen-mult pattern by handling two operands with different sizes,
4162 and operands whose size is smaller than half of the result type.
4164 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
4166 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
4167 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
4168 (do_estimate_edge_time): Compute it.
4169 * ipa-inline.c (want_inline_small_function_p): Bypass
4170 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
4172 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
4174 * ipa-inline.c (spec_rem): New static variable.
4175 (dump_overall_stats): New function.
4176 (dump_inline_stats): New function.
4178 2014-04-18 Richard Henderson <rth@redhat.com>
4180 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
4181 to GET_MODE_SIZE, not a reg_class_t.
4183 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4185 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
4186 (vsx_xxmrglw_<mode>): Likewise.
4188 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
4191 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
4192 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
4193 (rs6000_init_hard_regno_mode_ok): Likewise.
4195 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
4197 * ipa-inline.c (inline_small_functions): Account only non-cold
4199 * doc/invoke.texi (inline-unit-growth): Update documentation.
4201 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
4203 * config/rs6000/rs6000.md (addti3, subti3): New.
4205 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
4208 * config/i386/i386.c (ix86_expand_clear): Remove outdated
4209 comment. Check optimize_insn_for_size_p instead of
4210 optimize_insn_for_speed_p.
4212 2014-04-17 Martin Jambor <mjambor@suse.cz>
4214 * gimple-iterator.c (gsi_start_edge): New function.
4215 * gimple-iterator.h (gsi_start_edge): Declare.
4216 * tree-sra.c (single_non_eh_succ): New function.
4217 (disqualify_ops_if_throwing_stmt): Renamed to
4218 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
4219 having one non-EH successor BB.
4220 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
4221 generate loads into replacements.
4222 (sra_modify_assign): Likewise and and also use the simple path for
4224 (sra_modify_function_body): Commit statements on edges.
4226 2014-04-17 Richard Biener <rguenther@suse.de>
4229 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
4230 comparison results and add clarifying comment.
4232 2014-04-17 Jakub Jelinek <jakub@redhat.com>
4234 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
4235 (blank_mode): Initialize it.
4236 (emit_mode_size_inline, emit_mode_nunits_inline,
4237 emit_mode_inner_inline): New functions.
4238 (emit_insn_modes_h): Call them and surround their output with
4239 #if GCC_VERSION >= 4001 ... #endif.
4240 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
4241 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
4242 mode_* arrays if the argument is __builtin_constant_p.
4243 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
4244 is enum machine_mode.
4246 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
4248 * passes.c (opt_pass::execute): Adjust.
4249 (pass_manager::execute_pass_mode_switching): Likewise.
4250 (early_local_passes::execute): Likewise.
4251 (execute_one_pass): Pass cfun to the pass's execute method.
4252 * tree-pass.h (opt_pass::execute): Add function * argument.
4253 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
4254 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
4255 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
4256 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
4257 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
4258 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
4259 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
4260 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
4261 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
4262 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
4263 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
4264 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
4265 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
4266 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
4267 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
4268 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
4269 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
4270 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
4271 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
4272 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
4273 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
4274 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
4275 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
4276 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
4277 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
4278 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
4279 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
4280 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
4281 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
4284 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
4286 * passes.c (opt_pass::gate): Take function * argument.
4287 (gate_all_early_local_passes): Merge into
4288 (early_local_passes::gate): this.
4289 (gate_all_early_optimizations): Merge into
4290 (all_early_optimizations::gate): this.
4291 (gate_all_optimizations): Mege into
4292 (all_optimizations::gate): this.
4293 (gate_all_optimizations_g): Merge into
4294 (all_optimizations_g::gate): this.
4295 (gate_rest_of_compilation): Mege into
4296 (rest_of_compilation::gate): this.
4297 (gate_postreload): Merge into
4298 (postreload::gate): this.
4299 (dump_one_pass): Pass cfun to the pass's gate method.
4300 (execute_ipa_summary_passes): Likewise.
4301 (execute_one_pass): Likewise.
4302 (ipa_write_summaries_2): Likewise.
4303 (ipa_write_optimization_summaries_1): Likewise.
4304 (ipa_read_summaries_1): Likewise.
4305 (ipa_read_optimization_summaries_1): Likewise.
4306 (execute_ipa_stmt_fixups): Likewise.
4307 * tree-pass.h (opt_pass::gate): Add function * argument.
4308 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
4309 combine-stack-adj.c, combine.c, compare-elim.c,
4310 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
4311 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
4312 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
4313 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
4314 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
4315 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
4316 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
4317 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
4318 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
4319 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
4320 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
4321 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
4322 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
4323 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
4324 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
4325 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
4326 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
4327 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
4328 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
4329 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
4330 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
4331 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
4332 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
4333 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
4334 var-tracking.c, vtable-verify.c, web.c: Adjust.
4336 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
4338 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
4339 * configure: Regenerate.
4341 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
4343 * passes.c (dump_one_pass): don't check pass->has_gate.
4344 (execute_ipa_summary_passes): Likewise.
4345 (execute_one_pass): Likewise.
4346 (ipa_write_summaries_2): Likewise.
4347 (ipa_write_optimization_summaries_1): Likewise.
4348 (ipa_read_optimization_summaries_1): Likewise.
4349 (execute_ipa_stmt_fixups): Likewise.
4350 * tree-pass.h (pass_data::has_gate): Remove.
4351 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
4352 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
4353 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
4354 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
4355 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
4356 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
4357 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
4358 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
4359 gimple-low.c, gimple-ssa-isolate-paths.c,
4360 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
4361 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
4362 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
4363 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
4364 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
4365 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
4366 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
4367 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
4368 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
4369 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
4370 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
4371 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
4372 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
4373 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
4374 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
4375 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
4376 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
4377 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
4378 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
4379 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
4380 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
4381 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
4384 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
4386 * pass_manager.h (pass_manager::register_dump_files_1): Remove
4388 * passes.c (pass_manager::register_dump_files_1): Merge into
4389 (pass_manager::register_dump_files): this, and remove its handling of
4390 properties since the pass always has the properties anyway.
4391 (pass_manager::pass_manager): Adjust.
4393 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
4395 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
4396 * passes.c (pass_manager::register_dump_files_1): Remove dead code
4397 dealing with properties.
4398 (pass_manager::register_dump_files): Adjust.
4400 2014-03-20 Mark Wielaard <mjw@redhat.com>
4402 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
4403 then represent the bound as normal constant value.
4405 2014-04-17 Jakub Jelinek <jakub@redhat.com>
4408 Forward port from 4.8 branch
4409 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
4411 * config/i386/bmiintrin.h (_blsi_u32): New.
4415 (_blsmsk_u32): Ditto.
4416 (_blsmsk_u64): Ditto.
4417 (_tzcnt_u32): Ditto.
4418 (_tzcnt_u64): Ditto.
4420 2014-04-17 Kito Cheng <kito@0xlab.org>
4422 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
4424 2014-04-17 Richard Biener <rguenther@suse.de>
4427 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
4428 boolean results for comparisons.
4430 2014-04-17 Richard Biener <rguenther@suse.de>
4432 PR tree-optimization/60836
4433 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
4434 initial PHI args to be gimple values.
4436 2014-04-17 Richard Biener <rguenther@suse.de>
4438 PR tree-optimization/60841
4439 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
4440 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
4441 of stmts to SLP build.
4442 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
4443 (vect_analyze_slp): Likewise.
4444 (vect_analyze_slp_instance): Likewise.
4445 (vect_build_slp_tree): Limit overall SLP tree growth.
4446 * tree-vectorizer.h (vect_analyze_data_refs,
4447 vect_analyze_slp): Adjust prototypes.
4449 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
4451 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
4454 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
4456 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
4457 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
4458 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
4459 for TARGET_SLOW_PSHUFB
4461 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
4463 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
4464 * config/i386/i386.c (intel_cost): Ditto.
4466 2014-04-17 Joey Ye <joey.ye@arm.com>
4468 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
4470 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
4472 * opts.c (common_handle_option): Disable -fipa-reference coorectly
4475 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
4477 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
4478 (type_all_derivations_known_p): New predicate.
4479 (type_all_ctors_visible_p): New predicate.
4480 (type_possibly_instantiated_p): New predicate.
4481 (get_odr_type): Compute all_derivations_known.
4482 (dump_odr_type): Dump the flag.
4483 (maybe_record_type): Cleanup.
4484 (record_target_from_binfo): Add bases_to_consider array;
4485 record bases for types w/o instances and skip CXX destructor.
4486 (possible_polymorphic_call_targets_1): Add bases_to_consider
4487 and consider_construction parameters; check if type may have instance.
4488 (get_polymorphic_call_info): Set maybe_in_construction to true
4489 when we know nothing.
4490 (record_targets_from_bases): Skip CXX destructors; they are
4491 never called for types in construction.
4492 (possible_polymorphic_call_targets): Do not record target when
4493 type may not have instance.
4495 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
4498 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
4499 external aliases alive, too.
4501 2014-04-16 Andrew Pinski <apinski@cavium.com>
4503 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
4506 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
4508 * final.c (compute_alignments): Do not apply loop alignment to a block
4509 falling through to the exit.
4511 2014-04-16 Catherine Moore <clm@codesourcery.com>
4513 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
4514 Adjust constraints for microMIPS store patterns.
4516 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
4518 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
4520 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
4522 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
4523 (append_use): Run at -O0.
4524 (append_vdef): Likewise.
4525 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
4526 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
4528 2014-04-16 Jakub Jelinek <jakub@redhat.com>
4530 PR tree-optimization/60844
4531 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
4532 (propagate_op_to_single_use, remove_visited_stmt_chain,
4533 linearize_expr, repropagate_negates, reassociate_bb): Use it
4534 instead of gsi_remove.
4536 2014-04-16 Martin Jambor <mjambor@suse.cz>
4538 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
4539 ipa_transforms_to_apply.
4540 (cgraph_function_versioning): Assert that old_node has empty
4541 ipa_transforms_to_apply.
4542 * trans-mem.c (ipa_tm_create_version): Likewise.
4543 * tree-inline.c (tree_function_versioning): Do not duplicate
4544 ipa_transforms_to_apply.
4546 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4549 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
4551 Pass necessary as flags on 64-bit Solaris/x86.
4552 Use lowercase relocs for x86_64-*-*.
4553 * configure: Regenerate.
4555 2014-04-15 Jan Hubicka <jh@suse.cz>
4557 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
4558 (maybe_record_node, likely_target_p): Use it.
4560 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4563 Revert following patch
4565 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
4568 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
4569 software floating point or no floating point registers, do not
4570 allow any type in the FPRs. Eliminate a test for SPE SIMD types
4571 in GPRs that occurs after we tested for GPRs that would never be
4574 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
4575 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
4576 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
4577 specifically allow DDmode, since that does not use the SPE SIMD
4580 2014-03-21 Mark Wielaard <mjw@redhat.com>
4582 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
4583 as unsigned or int depending on type and value used.
4585 2014-04-15 Richard Biener <rguenther@suse.de>
4587 PR rtl-optimization/56965
4588 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
4589 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
4591 * alias.c (true_dependence_1): Do not call
4592 nonoverlapping_component_refs_p.
4593 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
4594 nonoverlapping_component_refs_p.
4595 (indirect_refs_may_alias_p): Likewise.
4597 2014-04-15 Teresa Johnson <tejohnson@google.com>
4599 * cfg.c (dump_bb_info): Fix flags check.
4600 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
4602 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4604 PR rtl-optimization/60663
4605 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
4608 2014-04-15 Richard Biener <rguenther@suse.de>
4610 * lto-streamer.h (LTO_major_version): Bump to 4.
4612 2014-04-15 Richard Biener <rguenther@suse.de>
4614 * common.opt (lto_partition_model): New enum.
4615 (flto-partition=): Merge separate options with a single with argument,
4616 add -flto-partition=one support.
4617 * flag-types.h (enum lto_partition_model): Declare.
4618 * opts.c (finish_options): Remove duplicate -flto-partition=
4620 * lto-wrapper.c (run_gcc): Adjust.
4622 2014-04-15 Richard Biener <rguenther@suse.de>
4624 * alias.c (ncr_compar): New function.
4625 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
4627 2014-04-15 Richard Biener <rguenther@suse.de>
4629 * alias.c (record_component_aliases): Do not walk BINFOs.
4631 2014-04-15 Richard Biener <rguenther@suse.de>
4633 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
4634 Add struct function argument and adjust.
4635 (find_func_aliases_for_call): Likewise.
4636 (find_func_aliases): Likewise.
4637 (find_func_clobbers): Likewise.
4638 (intra_create_variable_infos): Likewise.
4639 (compute_points_to_sets): Likewise.
4640 (ipa_pta_execute): Adjust. Do not push/pop cfun.
4642 2014-04-15 Richard Biener <rguenther@suse.de>
4644 * tree.c (iterative_hash_expr): Use enum tree_code_class
4645 to store TREE_CODE_CLASS.
4646 (tree_block): Likewise.
4647 (tree_set_block): Likewise.
4648 * tree.h (fold_build_pointer_plus_loc): Use
4649 convert_to_ptrofftype_loc.
4651 2014-04-15 Jakub Jelinek <jakub@redhat.com>
4654 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
4657 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
4659 * cfgloop.h (struct loop): Move force_vectorize down.
4660 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
4661 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
4662 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
4663 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
4664 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
4665 * tree-core.h (enum annot_expr_kind): Add new kind values.
4666 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
4667 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
4669 * tree.def (ANNOTATE_EXPR): Tweak comment.
4671 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
4673 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
4676 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
4678 * tree.h (TYPE_IDENTIFIER): Declare.
4679 * tree.c (subrange_type_for_debug_p): Use it.
4680 * godump.c (go_format_type): Likewise.
4681 * dwarf2out.c (is_cxx_auto, modified_type_die,
4682 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
4683 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
4685 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
4688 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
4690 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
4692 * config/i386/i386.c (examine_argument): Return bool. Return true if
4693 parameter should be passed in memory.
4694 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
4695 (construct_container): Update calls to examine_argument.
4696 (function_arg_advance_64): Ditto.
4697 (return_in_memory_32): Merge with ix86_return_in_memory.
4698 (return_in_memory_64): Ditto.
4699 (return_in_memory_ms_64): Ditto.
4701 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
4703 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
4704 * coverage.c (coverage_compute_profile_id): Handle externally visible
4707 2014-04-14 Martin Jambor <mjambor@suse.cz>
4709 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
4710 DECL_DISREGARD_INLINE_LIMITS functions.
4712 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
4715 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
4717 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
4720 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
4721 optimize_insn_for_speed_p instead of
4722 optimize_function_for_speed_p.
4724 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
4726 * doc/invoke.texi (free): Document AArch64.
4728 2014-04-14 Richard Biener <rguenther@suse.de>
4730 PR tree-optimization/60042
4731 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
4732 (insert_into_preds_of_block): Do not prevent PHI insertion
4733 for REFERENCE exprs here ...
4734 (eliminate_dom_walker::before_dom_children): ... but prevent
4735 their use here under similar conditions when applied to the
4736 IL after PRE optimizations.
4738 2014-04-14 Richard Biener <rguenther@suse.de>
4740 * passes.def: Move early points-to after early SRA.
4742 2014-04-14 Richard Biener <rguenther@suse.de>
4744 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
4745 check for which sign-changes we allow when forwarding
4746 a converted value into a switch.
4748 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
4750 * stor-layout.c (place_field): Finalize non-constant offset for the
4753 2014-04-14 Richard Biener <rguenther@suse.de>
4755 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
4757 (expand_switch_using_bit_tests_p): Likewise.
4758 (process_switch): Compute and pass on speed_p based on the
4760 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
4761 optimize_bb_for_speed_p.
4763 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
4765 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
4766 * function.h (struct function): Rename has_force_vect_loops into
4767 has_force_vectorize_loops.
4768 * lto-streamer-in.c (input_cfg): Adjust for renaming.
4769 (input_struct_function_base): Likewise.
4770 * lto-streamer-out.c (output_cfg): Likewise.
4771 (output_struct_function_base): Likewise.
4772 * omp-low.c (expand_omp_simd): Likewise.
4773 * tree-cfg.c (move_sese_region_to_fn): Likewise.
4774 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
4775 (version_loop_for_if_conversion): Likewise.
4776 (tree_if_conversion): Likewise.
4777 (main_tree_if_conversion): Likewise.
4778 (gate_tree_if_conversion): Likewise.
4779 * tree-inline.c (copy_loops): Likewise.
4780 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
4781 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
4782 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
4783 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
4784 * tree-vectorizer.c (vectorize_loops): Likewise.
4785 * tree-vectorizer.h (unlimited_cost_model): Likewise.
4787 2014-04-14 Richard Biener <rguenther@suse.de>
4790 * lto-streamer-out.c (wrap_refs): New function.
4791 (lto_output): Wrap symbol references in global initializes in
4792 type-preserving MEM_REFs.
4794 2014-04-14 Christian Bruel <christian.bruel@st.com>
4796 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
4798 2014-04-14 Christian Bruel <christian.bruel@st.com>
4800 * config/sh/sh.md (setmemqi): New expand pattern.
4801 * config/sh/sh.h (CLEAR_RATIO): Define.
4802 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
4803 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
4805 2014-04-14 Richard Biener <rguenther@suse.de>
4808 * fold-const.c (negate_expr_p): Don't negate directional rounding
4810 (fold_negate_expr): Likewise.
4812 2014-04-14 Richard Biener <rguenther@suse.de>
4814 PR tree-optimization/59817
4815 PR tree-optimization/60453
4816 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
4817 recursion to catch all CHRECs in the scalar evolution and restrict
4818 the predicate for the remains appropriately.
4820 2014-04-12 Catherine Moore <clm@codesourcery.com>
4822 * config/mips/constraints.md: Add new register constraint "kb".
4823 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
4824 (*movhi_internal): Likewise.
4825 (*movqi_internal): Likewise.
4826 * config/mips/mips.h (M16_STORE_REGS): New register class.
4827 (REG_CLASS_NAMES): Add M16_STORE_REGS.
4828 (REG_CLASS_CONTENTS): Likewise.
4829 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
4831 2014-04-11 Tobias Burnus <burnus@net-b.de>
4834 * doc/invoke.texi (-Wformat-signedness): Document it.
4835 (Wformat=2): Mention that this enables -Wformat-signedness.
4837 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
4839 * common/config/epiphany/epiphany-common.c
4840 (epiphany_option_optimization_table): Enable section anchors by
4841 default at -O1 or higher.
4842 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
4843 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
4844 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
4845 carries no extra cost.
4846 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
4847 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
4848 * config/epiphany/predicates.md (memclob_operand): New predicate.
4849 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
4850 Use memclob_operand predicate and X constraint for operand 3.
4852 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
4854 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
4855 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
4858 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
4860 PR rtl-optimization/60651
4861 * mode-switching.c (optimize_mode_switching): Make sure to emit
4862 sets of a lower numbered entity before sets of a higher numbered
4863 entity to a mode of the same or lower priority.
4864 When creating a seginfo for a basic block that starts with a code
4865 label, move the insertion point past the code label.
4866 (new_seginfo): Document and enforce requirement that
4867 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
4868 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
4869 * doc/tm.texi: Regenerate.
4871 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
4874 * config/arc/arc.c (arc_save_restore): Fix assert typo.
4876 2013-04-11 Jakub Jelinek <jakub@redhat.com>
4878 * BASE-VER: Set to 4.10.0.
4880 2014-04-11 Tobias Burnus <burnus@net-b.de>
4883 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
4884 * doc/gcc.texi (Service): Update description in the @menu
4885 * doc/invoke.texi (Option Summary): Remove misplaced and
4888 2014-04-11 Steve Ellcey <sellcey@mips.com>
4889 Jakub Jelinek <jakub@redhat.com>
4892 * expr.c (convert_move): Use emit_store_flag_force instead of
4893 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
4896 2014-04-11 Richard Biener <rguenther@suse.de>
4899 * varasm.c (assemble_alias): Avoid endless error reporting
4900 recursion by setting TREE_ASM_WRITTEN.
4902 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4904 * config/s390/s390.md: Add a splitter for NOT rtx.
4906 2014-04-11 Jakub Jelinek <jakub@redhat.com>
4908 PR rtl-optimization/60663
4909 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
4911 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
4912 Jakub Jelinek <jakub@redhat.com>
4915 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
4916 flag from decl_node to node.
4918 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4921 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
4922 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
4923 ameliorating the cases where it can be.
4925 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
4928 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
4930 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
4931 (loadsync_<mode>): Change mode.
4932 (load_quadpti, store_quadpti): New.
4933 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
4934 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
4935 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
4937 2014-04-09 Cong Hou <congh@google.com>
4940 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
4943 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4945 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
4946 instead of vnor to exploit possible fusion opportunity in the
4948 (altivec_expand_vec_perm_const_le): Likewise.
4950 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
4952 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
4953 (loadsync_<mode>): Change mode.
4954 (load_quadpti, store_quadpti): New.
4955 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
4956 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
4958 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
4961 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
4962 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
4963 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
4965 2014-04-08 Richard Biener <rguenther@suse.de>
4968 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
4969 a 64bit widest int print double-int similar to on HWI64 hosts.
4971 2014-04-08 Richard Biener <rguenther@suse.de>
4973 PR tree-optimization/60785
4974 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
4975 default defs properly.
4977 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
4979 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
4980 (Weffc++): Likewise.
4982 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
4984 * ipa-devirt.c (maybe_record_node): When node is not recorded,
4985 set completep to false rather than true.
4987 2014-04-07 Douglas B Rupp <rupp@adacore.com>
4990 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
4991 ARM_TARGET2_DWARF_FORMAT.
4993 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
4996 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
4997 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
5000 2014-04-07 Richard Biener <rguenther@suse.de>
5002 PR tree-optimization/60766
5003 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
5004 (may_eliminate_iv): Convert cand_value_at result to desired type.
5006 2014-04-07 Jason Merrill <jason@redhat.com>
5009 * common.opt (-fno-gnu-unique): Add.
5010 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
5012 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5014 * haifa-sched.c: Fix outdated function reference and minor
5015 grammar errors in introductory comment.
5017 2014-04-07 Richard Biener <rguenther@suse.de>
5020 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
5022 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
5024 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
5027 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
5028 size accounting for thunks.
5029 (pa_asm_output_mi_thunk): Use final_start_function() and
5030 final_end_function() to output function start and end directives.
5032 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
5034 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
5035 device specific ISA/ feature information. Remove short_sp and
5036 errata_skip ds. Add avr_device_specific_features enum to have device
5038 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
5039 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
5040 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
5041 updated device specific info.
5042 * config/avr/avr-mcus.def: Merge device specific details to
5043 dev_attribute field.
5044 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
5046 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
5047 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
5048 assembler if RMW isa supported by current device.
5049 * config/avr/genmultilib.awk: Update as device info structure changed.
5050 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
5052 2014-04-04 Cong Hou <congh@google.com>
5054 PR tree-optimization/60656
5055 * tree-vect-stmts.c (supportable_widening_operation):
5056 Fix a bug that elements in a vector with vect_used_by_reduction
5057 property are incorrectly reordered when the operation on it is not
5058 consistant with the one in reduction operation.
5060 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
5062 PR rtl-optimization/60155
5063 * gcse.c (record_set_data): New function.
5064 (single_set_gcse): New function.
5065 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
5066 (hoist_code): Likewise.
5067 (get_pressure_class_and_nregs): Likewise.
5069 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
5071 * explow.c (probe_stack_range): Emit a final optimization blockage.
5073 2014-04-04 Anthony Green <green@moxielogic.com>
5075 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
5078 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
5081 * lto-cgraph.c (input_overwrite_node): Check that partitioning
5082 flags are set only during streaming.
5083 * ipa.c (process_references, walk_polymorphic_call_targets,
5084 symtab_remove_unreachable_nodes): Drop bodies of always inline
5085 after early inlining.
5086 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
5088 2014-04-04 Jakub Jelinek <jakub@redhat.com>
5089 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5092 * dwarf2out.c (const_ok_for_output_1): Reject expressions
5095 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5098 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
5100 (cortex_a53_fdivd): Likewise.
5102 2014-04-04 Martin Jambor <mjambor@suse.cz>
5105 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
5107 * cgraph.c (clone_of_p): Also return true if thunks match.
5108 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
5109 cgraph_function_or_thunk_node and an obsolete comment.
5110 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
5112 (build_function_decl_skip_args): Likewise.
5113 (set_new_clone_decl_and_node_flags): New function.
5114 (duplicate_thunk_for_node): Likewise.
5115 (redirect_edge_duplicating_thunks): Likewise.
5116 (cgraph_clone_node): New parameter args_to_skip, pass it to
5117 redirect_edge_duplicating_thunks which is called instead of
5118 cgraph_redirect_edge_callee.
5119 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
5120 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
5122 2014-04-04 Jeff Law <law@redhat.com>
5125 * config/arm/predicates.md (const_int_I_operand): New predicate.
5126 (const_int_M_operand): Similarly.
5127 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
5129 (insv_t2, extv_reg, extzv_t2): Likewise.
5130 (load_multiple_with_writeback): Similarly for const_int_I_operand.
5131 (pop_multiple_with_writeback_and_return): Likewise.
5132 (vfp_pop_multiple_with_writeback): Likewise
5134 2014-04-04 Richard Biener <rguenther@suse.de>
5137 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
5138 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
5140 * gimplify.h (gimple_add_tmp_var_fn): Declare.
5141 * gimplify.c (gimple_add_tmp_var_fn): New function.
5142 * gimple-expr.h (create_tmp_reg_fn): Declare.
5143 * gimple-expr.c (create_tmp_reg_fn): New function.
5144 * gimple-low.c (record_vars_into): Don't change cfun.
5145 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
5146 code generation without cfun.
5148 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
5151 * Makefile.in (install-driver): Fix shell scripting.
5153 2014-04-03 Cong Hou <congh@google.com>
5155 PR tree-optimization/60505
5156 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
5157 threshold of number of iterations below which no vectorization
5159 * tree-vect-loop.c (new_loop_vec_info):
5160 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
5161 * tree-vect-loop.c (vect_analyze_loop_operations):
5162 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
5163 * tree-vect-loop.c (vect_transform_loop):
5164 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
5165 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
5166 of iterations of the loop and see if we should build the epilogue.
5168 2014-04-03 Richard Biener <rguenther@suse.de>
5170 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
5171 (streamer_tree_cache_create): Adjust.
5172 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
5173 to allow optional nodes array.
5174 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
5175 (streamer_tree_cache_append): Likewise.
5176 (streamer_tree_cache_create): Create nodes array optionally
5177 as specified by parameter.
5178 * lto-streamer-out.c (create_output_block): Avoid maintaining
5179 the node array in the writer cache.
5180 (DFS_write_tree): Remove assertion.
5181 (produce_asm_for_decls): Free the out decl state hash table early.
5182 * lto-streamer-in.c (lto_data_in_create): Adjust for
5183 streamer_tree_cache_create prototype change.
5185 2014-04-03 Richard Biener <rguenther@suse.de>
5187 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
5188 set TREE_CHAIN to NULL_TREE.
5190 2014-04-03 Richard Biener <rguenther@suse.de>
5192 PR tree-optimization/60740
5193 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
5194 over all GIMPLE_COND operands.
5196 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
5198 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
5199 (Weffc++): Remove Scott's numbering, merge lists and reference
5202 2014-04-03 Nick Clifton <nickc@redhat.com>
5204 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
5207 2014-04-03 Martin Jambor <mjambor@suse.cz>
5209 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
5210 mention gcc_unreachable before failing.
5211 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
5214 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
5217 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
5218 inconsistent code and instead mark the context inconsistent.
5219 (possible_polymorphic_call_targets): For inconsistent contexts
5220 return empty complete list.
5222 2014-04-02 Anthony Green <green@moxielogic.com>
5224 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
5225 (extendqisi2, extendhisi2): Define.
5226 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
5227 (WCHAR_TYPE): Change to unsigned int.
5229 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5231 PR tree-optimization/60733
5232 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
5233 insertion point for PHI candidates to be the end of the feeding
5234 block for the PHI argument.
5236 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
5238 PR rtl-optimization/60650
5239 * lra-constraints.c (process_alt_operands): Decrease reject for
5240 earlyclobber matching.
5242 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5244 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
5246 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5248 * config/spu/spu.c (pad_bb): Do not crash when the last
5249 insn is CODE_FOR_blockage.
5251 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5253 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
5254 lies outside the target mode.
5256 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
5259 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
5260 software floating point or no floating point registers, do not
5261 allow any type in the FPRs. Eliminate a test for SPE SIMD types
5262 in GPRs that occurs after we tested for GPRs that would never be
5265 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
5266 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
5267 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
5268 specifically allow DDmode, since that does not use the SPE SIMD
5271 2014-04-02 Richard Biener <rguenther@suse.de>
5274 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
5275 MODE_INTs. Properly use negv_optab.
5276 (expand_abs): Likewise.
5278 2014-04-02 Richard Biener <rguenther@suse.de>
5281 * Makefile.in (install-driver): Guard extra installs with special
5284 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
5286 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5289 2014-04-01 Richard Henderson <rth@redhat.com>
5292 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
5293 alternative enabled before register allocation.
5295 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
5297 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
5298 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
5300 (nios2_large_got_address): Remove unneeded 'sym' parameter.
5301 (nios2_got_address): Update nios2_large_got_address call site.
5302 (nios2_delegitimize_address): New function.
5303 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
5304 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
5305 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
5307 2014-04-01 Martin Husemann <martin@duskware.de>
5309 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
5312 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
5314 PR rtl-optimization/60604
5315 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
5316 check from register_operand.
5317 (register_operand): Redefine in terms of general_operand.
5318 (nonmemory_operand): Use register_operand for the non-constant cases.
5320 2014-04-01 Richard Biener <rguenther@suse.de>
5322 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
5324 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
5326 * doc/invoke.texi (mapp-regs): Clarify.
5328 2014-03-31 Ulrich Drepper <drepper@gmail.com>
5330 * config/i386/avx512fintrin.h (__v32hi): Define type.
5331 (__v64qi): Likewise.
5332 (_mm512_set1_epi8): Define.
5333 (_mm512_set1_epi16): Define.
5334 (_mm512_set4_epi32): Define.
5335 (_mm512_set4_epi64): Define.
5336 (_mm512_set4_pd): Define.
5337 (_mm512_set4_ps): Define.
5338 (_mm512_setr4_epi64): Define.
5339 (_mm512_setr4_epi32): Define.
5340 (_mm512_setr4_pd): Define.
5341 (_mm512_setr4_ps): Define.
5342 (_mm512_setzero_epi32): Define.
5344 2014-03-31 Martin Jambor <mjambor@suse.cz>
5347 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
5348 callsite_arguments_match_p. Updated all callers. Also check types of
5349 corresponding formal parameters and actual arguments.
5350 (not_all_callers_have_enough_arguments_p) Renamed to
5351 some_callers_have_mismatched_arguments_p.
5353 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
5355 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
5357 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
5360 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
5363 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
5365 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
5366 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
5368 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
5369 Use FMAMODE_NOVF512 mode iterator.
5370 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
5371 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
5372 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
5373 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
5375 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
5376 Use VF_128_256 mode iterator.
5377 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
5380 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
5382 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
5383 static chain if needed.
5385 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
5388 * lra-constraints.c (index_part_to_reg): New.
5389 (process_address): Use it.
5391 2014-03-27 Jeff Law <law@redhat.com>
5392 Jakub Jelinek <jakub@redhat.com>
5395 * expr.c (do_tablejump): Use simplify_gen_binary rather than
5396 gen_rtx_{PLUS,MULT} to build up the address expression.
5398 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
5399 creating non-canonical RTL.
5401 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
5404 * ipa-inline.c (want_inline_small_function_p): Short circuit large
5405 functions; reorganize to make cheap checks first.
5406 (inline_small_functions): Do not estimate growth when dumping;
5408 * ipa-inline.h (inline_summary): Add min_size.
5409 (growth_likely_positive): New function.
5410 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
5411 (set_cond_stmt_execution_predicate): Cleanup.
5412 (estimate_edge_size_and_time): Compute min_size.
5413 (estimate_calls_size_and_time): Likewise.
5414 (estimate_node_size_and_time): Likewise.
5415 (inline_update_overall_summary): Update min_size.
5416 (do_estimate_edge_time): Likewise.
5417 (do_estimate_edge_size): Update.
5418 (do_estimate_edge_hints): Update.
5419 (growth_likely_positive): New function.
5421 2014-03-28 Jakub Jelinek <jakub@redhat.com>
5424 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
5425 also if addr has VOIDmode.
5427 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5429 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
5430 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
5432 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
5433 instructions as well as AdvancedSIMD loads.
5435 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5437 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
5438 Use crypto_aese type.
5439 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
5440 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
5441 crypto_aese, crypto_aesmc. Move to types.md.
5442 * config/arm/types.md (crypto_aes): Split into crypto_aese,
5444 * config/arm/iterators.md (crypto_type): Likewise.
5446 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
5448 * cgraph.c: Include expr.h and tree-dfa.h.
5449 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
5452 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
5455 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
5456 regs from checking multi-reg pseudos.
5458 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5460 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
5462 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5464 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
5465 if it would clobber the stack pointer, even temporarily.
5467 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
5469 * mode-switching.c: Make small adjustments to the top comment.
5471 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
5473 * config/rs6000/constraints.md (wD constraint): New constraint to
5474 match the constant integer to get the top DImode/DFmode out of a
5475 vector in a VSX register.
5477 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
5478 match the constant integer to get the top DImode/DFmode out of a
5479 vector in a VSX register.
5481 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
5484 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
5487 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
5488 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
5490 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
5491 Optimize vec_extract of 64-bit values, where the value being
5492 extracted is in the top word, where we can use scalar
5493 instructions. Add direct move and store support. Combine the big
5494 endian/little endian vector select load support into a single insn.
5495 (vsx_extract_<mode>_internal1): Likewise.
5496 (vsx_extract_<mode>_internal2): Likewise.
5497 (vsx_extract_<mode>_load): Likewise.
5498 (vsx_extract_<mode>_store): Likewise.
5499 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
5500 combined into vsx_extract_<mode>_load.
5501 (vsx_extract_<mode>_one_le): Likewise.
5503 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
5504 define the top 64-bit vector element.
5506 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
5509 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5510 Document vec_vbpermq builtin.
5513 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
5514 enable use of xxsldwi and xxpermdi builtin functions.
5515 (vec_xxpermdi): Likewise.
5517 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5518 Document use of vec_xxsldwi and vec_xxpermdi builtins.
5520 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
5522 PR rtl-optimization/60650
5523 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
5525 (find_spills_for): New.
5526 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
5527 Spill all pseudos on the second iteration.
5529 2014-03-27 Marek Polacek <polacek@redhat.com>
5532 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
5535 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5537 * config/s390/s390.c (s390_can_use_return_insn): Check for
5538 call-saved FPRs on 31 bit.
5540 2014-03-27 Jakub Jelinek <jakub@redhat.com>
5543 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
5544 if they need regimplification, just drop them instead of
5545 calling gimple_regimplify_operands on them.
5547 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
5550 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
5551 (aarch64_frame_pointer_required): Adjust logic.
5552 (aarch64_can_eliminate): Adjust logic.
5553 (aarch64_override_options_after_change): Adjust logic.
5555 2014-03-27 Dehao Chen <dehao@google.com>
5557 * ipa-inline.c (early_inliner): Update node's inline info.
5559 2014-03-26 Dehao Chen <dehao@google.com>
5561 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
5562 compiler inserted conditional jumps for NAN float check.
5564 2014-03-26 Jakub Jelinek <jakub@redhat.com>
5566 * ubsan.h (ubsan_create_data): Change second argument's type
5567 to const location_t *.
5568 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
5570 (ubsan_create_data): Change second argument to const location_t *PLOC.
5571 Create Loc field whenever PLOC is non-NULL.
5572 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
5573 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
5577 * real.c (real_to_integer2): Change type of low to UHWI.
5579 2014-03-26 Tobias Burnus <burnus@net-b.de>
5581 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
5582 (CILK_SELF_SPECS): New define.
5583 (driver_self_specs): Use it.
5585 2014-03-26 Richard Biener <rguenther@suse.de>
5587 * tree-pretty-print.c (percent_K_format): Implement special
5588 case for LTO and its stripped down BLOCK tree.
5590 2014-03-26 Jakub Jelinek <jakub@redhat.com>
5593 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
5595 * tree-vrp.c (simplify_internal_call_using_ranges): If only
5596 one range is range_int_cst_p, but not both, at least optimize
5597 addition/subtraction of 0 and multiplication by 0 or 1.
5598 * gimple-fold.c (gimple_fold_call): Fold
5599 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
5600 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
5601 INTEGER_CSTs, try to fold at least x * 0 and y - y.
5603 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
5605 PR rtl-optimization/60452
5606 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
5607 <case REG>: Return 1 for invalid offsets from the frame pointer.
5609 2014-03-26 Marek Polacek <polacek@redhat.com>
5612 * doc/extend.texi (C Extensions): Mention variable-length arrays in
5615 2014-03-26 Marek Polacek <polacek@redhat.com>
5618 * doc/extend.texi (Designated Inits): Describe what happens to omitted
5621 2014-03-26 Marek Polacek <polacek@redhat.com>
5624 * ira-color.c (update_conflict_hard_regno_costs): Perform the
5625 multiplication in unsigned type.
5627 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
5629 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
5631 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
5633 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
5635 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
5638 * cif-code.def (UNREACHABLE) New code.
5639 * ipa-inline.c (inline_small_functions): Skip edges to
5640 __builtlin_unreachable.
5641 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
5642 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
5643 predicate to __bulitin_unreachable.
5644 (set_cond_stmt_execution_predicate): Fix issue when
5645 invert_tree_comparison returns ERROR_MARK.
5646 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
5647 propagate to inline clones.
5648 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
5650 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
5651 * cgraphclones.c (cgraph_clone_node): If call destination is already
5652 ureachable, do not redirect it back.
5653 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
5656 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
5658 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
5659 Do not modify inline clones.
5661 2014-03-25 Jakub Jelinek <jakub@redhat.com>
5663 * config/i386/i386.md (general_sext_operand): New mode attr.
5664 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
5665 don't generate (sign_extend (const_int)).
5666 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
5667 operands[2]. Use We constraint instead of <i> and
5668 <general_sext_operand> predicate instead of <general_operand>.
5669 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
5670 * config/i386/constraints.md (We): New constraint.
5671 * config/i386/predicates.md (x86_64_sext_operand,
5672 sext_operand): New predicates.
5674 2014-03-25 Martin Jambor <mjambor@suse.cz>
5677 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
5678 inconsistent devirtualizations to __builtin_unreachable.
5680 2014-03-25 Marek Polacek <polacek@redhat.com>
5683 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
5685 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
5687 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
5688 order of elements for big-endian.
5690 2014-03-25 Richard Biener <rguenther@suse.de>
5693 * gimplify-me.c (gimple_regimplify_operands): Update the
5696 2014-03-25 Martin Jambor <mjambor@suse.cz>
5699 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
5700 (lto_output_varpool_node): Likewise.
5701 (input_overwrite_node): Likewise.
5702 (input_varpool_node): Likewise.
5704 2014-03-25 Richard Biener <rguenther@suse.de>
5706 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
5707 (run_gcc): Likewise.
5709 2014-03-25 Jakub Jelinek <jakub@redhat.com>
5711 * combine.c (simplify_compare_const): Add MODE argument.
5712 Handle mode_width 0 as very large mode_width.
5713 (try_combine, simplify_comparison): Adjust callers.
5715 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
5716 type to avoid signed integer overflow.
5717 * explow.c (plus_constant): Likewise.
5719 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
5721 * doc/generic.texi: Correct typos.
5723 2014-03-24 Tobias Burnus <burnus@net-b.de>
5725 * doc/invoke.texi (-flto): Expand section about
5726 using static libraries with LTO.
5728 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5730 PR rtl-optimization/60501
5731 * optabs.def (addptr3_optab): New optab.
5732 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
5733 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
5734 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
5736 * lra.c (emit_add3_insn): Use the addptr pattern if available.
5738 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
5740 2014-03-24 Ulrich Drepper <drepper@gmail.com>
5742 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
5745 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
5746 (_mm256_undefined_ps): Define.
5747 (_mm256_undefined_pd): Define.
5748 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
5749 (_mm_undefined_pd): Define.
5750 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
5751 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
5752 (_mm512_undefined_ps): Define.
5753 (_mm512_undefined_pd): Define.
5754 Use _mm*_undefined_*.
5755 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
5757 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
5759 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
5760 (lshr_simd): DI mode added.
5761 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
5762 (aarch64_ushr_simddi): Likewise.
5763 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
5764 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
5765 (vshrd_n_u64): Likewise.
5767 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5769 * Makefile.in (s-macro_list): Depend on cc1.
5771 2014-03-23 Teresa Johnson <tejohnson@google.com>
5773 * ipa-utils.c (ipa_print_order): Use specified dump file.
5775 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
5777 PR rtl-optimization/60601
5778 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
5780 * gcc.c (eval_spec_function): Initialize save_growing_value.
5782 2014-03-22 Jakub Jelinek <jakub@redhat.com>
5785 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
5786 code == MINUS_EXPR, never swap op0 with op1.
5788 * toplev.c (init_local_tick): Avoid signed integer multiplication
5790 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
5791 shift by first operand's bitsize.
5793 2014-03-21 Jakub Jelinek <jakub@redhat.com>
5796 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
5798 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
5799 TARGET_ISA_64BIT_P(x).
5801 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5803 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
5804 pattern for vector nor instead of subtract from splat(-1).
5805 (altivec_expand_vec_perm_const_le): Likewise.
5807 2014-03-21 Richard Henderson <rth@twiddle.net>
5810 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
5811 related insns after epilogue_completed.
5813 2014-03-21 Martin Jambor <mjambor@suse.cz>
5816 * cgraph.h (symtab_node): New flag body_removed.
5817 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
5818 when removing bodies.
5819 * symtab.c (dump_symtab_base): Dump body_removed flag.
5820 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
5821 had their bodies removed.
5823 2014-03-21 Martin Jambor <mjambor@suse.cz>
5826 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
5829 2014-03-21 Richard Biener <rguenther@suse.de>
5831 PR tree-optimization/60577
5832 * tree-core.h (struct tree_base): Document nothrow_flag use
5833 in VAR_DECL_NONALIASED.
5834 * tree.h (VAR_DECL_NONALIASED): New.
5835 (may_be_aliased): Adjust.
5836 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
5838 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
5840 * expr.c (expand_expr_real_1): Remove outdated comment.
5842 2014-03-20 Jakub Jelinek <jakub@redhat.com>
5845 * ira.c (adjust_cleared_regs): Call copy_rtx on
5846 *reg_equiv[REGNO (loc)].src_p before passing it to
5847 simplify_replace_fn_rtx.
5850 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
5851 into CONST, put pic register as first operand of PLUS. Use
5852 gen_const_mem for both 32-bit and 64-bit PIC got loads.
5854 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5856 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
5858 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
5860 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
5861 around for store forwarding issue in the FPU on the UT699.
5862 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
5863 loads and operations if -mfix-ut699 is specified.
5864 (divtf3_hq): Tweak attribute.
5865 (sqrttf2_hq): Likewise.
5867 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
5869 * calls.c (store_one_arg): Remove incorrect const qualification on the
5870 type of the temporary.
5871 * cfgexpand.c (expand_return): Likewise.
5872 * expr.c (expand_constructor): Likewise.
5873 (expand_expr_real_1): Likewise.
5875 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
5877 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
5880 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
5883 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
5885 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
5887 * config/arm/aarch-common-protos.h
5888 (alu_cost_table): Fix spelling of "extend".
5889 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
5891 2014-03-19 Richard Biener <rguenther@suse.de>
5894 * tree-core.h (tree_type_common): Re-order pointer members
5895 to reduce recursion depth during GC walks.
5897 2014-03-19 Marek Polacek <polacek@redhat.com>
5900 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
5901 before accessing it.
5903 2014-03-19 Richard Biener <rguenther@suse.de>
5906 * lto-streamer-in.c (input_function): In WPA stage do not drop
5909 2014-03-19 Jakub Jelinek <jakub@redhat.com>
5911 PR tree-optimization/60559
5912 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
5913 with build_zero_cst assignment.
5915 2014-03-18 Kai Tietz <ktietz@redhat.com>
5917 PR rtl-optimization/56356
5918 * sdbout.c (sdbout_parms): Verify that parms'
5919 incoming argument is valid.
5920 (sdbout_reg_parms): Likewise.
5922 2014-03-18 Richard Henderson <rth@redhat.com>
5925 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
5926 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
5927 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
5929 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
5931 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
5932 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
5933 Italicize plugin event names in description. Explain that
5934 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
5935 Remind that no GCC functions should be called after PLUGIN_FINISH.
5936 Explain what pragmas with expansion are.
5938 2014-03-18 Martin Liska <mliska@suse.cz>
5940 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
5941 gimple call statement is update.
5942 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
5943 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
5945 2014-03-18 Jakub Jelinek <jakub@redhat.com>
5948 * ubsan.c (ubsan_instrument_unreachable): Call
5949 initialize_sanitizer_builtins.
5950 (ubsan_pass): Likewise.
5953 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
5954 varpool_finalize_decl instead of rest_of_decl_compilation.
5956 2014-03-18 Richard Biener <rguenther@suse.de>
5958 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
5959 by using bitmap_and_compl instead of bitmap_and_compl_into.
5960 (df_rd_transfer_function): Likewise.
5962 2014-03-18 Richard Biener <rguenther@suse.de>
5964 * doc/lto.texi (fresolution): Fix typo.
5966 2014-03-18 Richard Biener <rguenther@suse.de>
5968 * doc/invoke.texi (flto): Update for changes in 4.9.
5970 2014-03-18 Richard Biener <rguenther@suse.de>
5972 * doc/loop.texi: Remove section on the removed lambda framework.
5973 Update loop docs with recent changes in preserving loop structure.
5975 2014-03-18 Richard Biener <rguenther@suse.de>
5977 * doc/lto.texi (-fresolution): Document.
5979 2014-03-18 Richard Biener <rguenther@suse.de>
5981 * doc/contrib.texi: Adjust my name.
5983 2014-03-18 Jakub Jelinek <jakub@redhat.com>
5986 * internal-fn.c: Include diagnostic-core.h.
5987 (expand_BUILTIN_EXPECT): New function.
5988 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
5989 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
5990 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
5991 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
5993 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
5994 Revert 3 argument __builtin_expect code.
5995 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
5996 * gimple-fold.c (gimple_fold_call): Likewise.
5997 * tree.h (fold_builtin_expect): New prototype.
5998 * builtins.c (build_builtin_expect_predicate): Add predictor
5999 argument, if non-NULL, create 3 argument __builtin_expect.
6000 (fold_builtin_expect): No longer static. Add ARG2 argument,
6001 pass it through to build_builtin_expect_predicate.
6002 (fold_builtin_2): Adjust caller.
6003 (fold_builtin_3): Handle BUILT_IN_EXPECT.
6004 * internal-fn.def (BUILTIN_EXPECT): New.
6006 2014-03-18 Tobias Burnus <burnus@net-b.de>
6009 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
6010 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
6011 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
6013 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
6016 * predict.c (combine_predictions_for_bb): Fix up formatting.
6017 (expr_expected_value_1, expr_expected_value): Add predictor argument,
6018 fill what it points to if non-NULL.
6019 (tree_predict_by_opcode): Adjust caller, use the predictor.
6020 * predict.def (PRED_COMPARE_AND_SWAP): Add.
6022 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
6024 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
6025 proper constant for the store mode.
6027 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
6029 * symtab.c (change_decl_assembler_name): Fix transparent alias
6032 2014-03-16 Renlin Li <Renlin.Li@arm.com>
6034 * config/aarch64/aarch64.c: Correct the comments about the
6035 aarch64 stack layout.
6037 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
6039 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
6040 check for GF_OMP_FOR_KIND_FOR.
6042 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
6044 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
6045 ymm and zmm register names.
6047 2014-03-17 Jakub Jelinek <jakub@redhat.com>
6050 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
6051 note creation for the 2010-08-31 changes.
6053 2014-03-17 Marek Polacek <polacek@redhat.com>
6056 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
6057 as -fno-tree-loop-vectorize.
6058 (expand_omp_simd): Likewise.
6060 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
6062 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
6063 (eligible_for_call_delay): New prototype.
6064 * config/sparc/sparc.c (tls_call_delay): Rename into...
6065 (eligible_for_call_delay): ...this. Return false if the instruction
6066 cannot be put in the delay slot of a branch.
6067 (eligible_for_restore_insn): Simplify.
6068 (eligible_for_return_delay): Return false if the instruction cannot be
6069 put in the delay slot of a branch and simplify.
6070 (eligible_for_sibcall_delay): Return false if the instruction cannot be
6071 put in the delay slot of a branch.
6072 * config/sparc/sparc.md (fix_ut699): New attribute.
6073 (tls_call_delay): Delete.
6074 (in_call_delay): Reimplement.
6075 (eligible_for_sibcall_delay): Rename into...
6076 (in_sibcall_delay): ...this.
6077 (eligible_for_return_delay): Rename into...
6078 (in_return_delay): ...this.
6079 (in_branch_delay): Reimplement.
6080 (in_uncond_branch_delay): Delete.
6081 (in_annul_branch_delay): Delete.
6083 2014-03-14 Richard Henderson <rth@redhat.com>
6086 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
6087 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
6088 (*floathi<X87MODEF>2_i387_with_temp): Remove.
6089 (floathi splitters): Remove.
6090 (float<SWI48x>xf2): New pattern.
6091 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
6092 code that tried to handle DImode for 32-bit, but which was excluded
6093 by the pattern's condition. Drop allocation of stack temporary.
6094 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
6095 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
6096 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
6097 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
6098 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
6099 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
6100 (*float<SWI48><MODEF>2_sse_interunit): Remove.
6101 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
6102 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
6103 (*float<SWI48x><X87MODEF>2_i387): Remove.
6104 (all float _with_temp splitters): Remove.
6105 (*float<SWI48x><MODEF>2_i387): New pattern.
6106 (*float<SWI48><MODEF>2_sse): New pattern.
6107 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
6108 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
6110 2014-03-14 Jakub Jelinek <jakub@redhat.com>
6111 Marek Polacek <polacek@redhat.com>
6114 * common.opt (dump_base_name_prefixed): New Variable.
6115 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
6116 if x_dump_base_name_prefixed is already set, set it at the end.
6118 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
6120 PR rtl-optimization/60508
6121 * lra-constraints.c (get_reload_reg): Add new parameter
6123 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
6124 Pass the new parameter values.
6126 2014-03-14 Richard Biener <rguenther@suse.de>
6128 * common.opt: Revert unintented changes from r205065.
6131 2014-03-14 Richard Biener <rguenther@suse.de>
6134 * cfghooks.c (split_block): Properly adjust all loops the
6135 block was a latch of.
6137 2014-03-14 Martin Jambor <mjambor@suse.cz>
6140 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
6143 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
6146 * config/avr/avr.c (avr_set_current_function): Pass function name
6147 through default_strip_name_encoding before sanity checking instead
6148 of skipping the first char of the assembler name.
6150 2014-03-13 Richard Henderson <rth@redhat.com>
6153 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
6154 (ix86_force_to_memory, ix86_free_from_memory): Remove.
6155 * config/i386/i386-protos.h: Likewise.
6156 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
6157 in the expander instead of a splitter.
6158 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
6159 any possibility of requiring a memory.
6160 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
6161 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
6162 (fp branch splitters): Update for ix86_split_fp_branch.
6163 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
6164 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
6165 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
6166 (*fop_<MODEF>_2_i387): Remove f/r alternative.
6167 (*fop_<MODEF>_3_i387): Likewise.
6168 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
6169 (splitters for the fop_* register patterns): Remove.
6170 (fscalexf4_i387): Rename from *fscalexf4_i387.
6171 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
6173 2014-03-13 Jakub Jelinek <jakub@redhat.com>
6175 PR tree-optimization/59779
6176 * tree-dfa.c (get_ref_base_and_extent): Use double_int
6177 type for bitsize and maxsize instead of HOST_WIDE_INT.
6179 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
6181 PR rtl-optimization/57320
6182 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
6183 the CFG after thread_prologue_and_epilogue_insns.
6185 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
6187 PR rtl-optimization/57189
6188 * lra-constraints.c (process_alt_operands): Disfavor spilling
6191 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
6193 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
6195 2014-03-13 Jakub Jelinek <jakub@redhat.com>
6197 PR tree-optimization/59025
6199 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
6200 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
6202 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
6205 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
6206 calls of avr_out_plus_1.
6208 2014-03-13 Bin Cheng <bin.cheng@arm.com>
6210 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
6211 BB's single pred and update the father loop's latch info later.
6213 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
6215 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
6219 (VEC_base): Likewise.
6220 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
6221 registers, we need to swap double words in little endian mode.
6223 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
6224 to be a container mode for 128-bit integer operations added in ISA
6225 2.07. Unlike TImode and PTImode, the preferred register set is
6226 the Altivec/VMX registers for the 128-bit operations.
6228 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
6230 (rs6000_split_128bit_ok_p): Likewise.
6232 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
6233 macros for creating ISA 2.07 normal and overloaded builtin
6234 functions with 3 arguments.
6235 (BU_P8V_OVERLOAD_3): Likewise.
6236 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
6237 for use as overloaded functions.
6238 (VPERM_1TI_UNS): Likewise.
6239 (VSEL_1TI): Likewise.
6240 (VSEL_1TI_UNS): Likewise.
6241 (ST_INTERNAL_1ti): Likewise.
6242 (LD_INTERNAL_1ti): Likewise.
6243 (XXSEL_1TI): Likewise.
6244 (XXSEL_1TI_UNS): Likewise.
6245 (VPERM_1TI): Likewise.
6246 (VPERM_1TI_UNS): Likewise.
6247 (XXPERMDI_1TI): Likewise.
6248 (SET_1TI): Likewise.
6249 (LXVD2X_V1TI): Likewise.
6250 (STXVD2X_V1TI): Likewise.
6251 (VEC_INIT_V1TI): Likewise.
6252 (VEC_SET_V1TI): Likewise.
6253 (VEC_EXT_V1TI): Likewise.
6254 (EQV_V1TI): Likewise.
6255 (NAND_V1TI): Likewise.
6256 (ORC_V1TI): Likewise.
6257 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
6258 added in ISA 2.07. Add both normal 'altivec' builtins, and the
6260 (VADDUQM): Likewise.
6261 (VSUBCUQ): Likewise.
6262 (VADDEUQM): Likewise.
6263 (VADDECUQ): Likewise.
6264 (VSUBEUQM): Likewise.
6265 (VSUBECUQ): Likewise.
6267 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
6268 __int128_t and __uint128_t types.
6269 (__uint128_type): Likewise.
6270 (altivec_categorize_keyword): Add support for vector __int128_t,
6271 vector __uint128_t, vector __int128, and vector unsigned __int128
6272 as a container type for TImode operations that need to be done in
6273 VSX/Altivec registers.
6274 (rs6000_macro_to_expand): Likewise.
6275 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
6276 to support 128-bit integer instructions vaddcuq, vadduqm,
6277 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
6278 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
6280 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
6281 for V1TImode, and set up preferences to use VSX/Altivec registers.
6282 Setup VSX reload handlers.
6283 (rs6000_debug_reg_global): Likewise.
6284 (rs6000_init_hard_regno_mode_ok): Likewise.
6285 (rs6000_preferred_simd_mode): Likewise.
6286 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
6287 (easy_altivec_constant): Likewise.
6288 (output_vec_const_move): Likewise.
6289 (rs6000_expand_vector_set): Convert V1TImode set and extract to
6291 (rs6000_expand_vector_extract): Likewise.
6292 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
6294 (rs6000_const_vec): Add support for V1TImode.
6295 (rs6000_emit_le_vsx_load): Swap double words when loading or
6296 storing TImode/V1TImode.
6297 (rs6000_emit_le_vsx_store): Likewise.
6298 (rs6000_emit_le_vsx_move): Likewise.
6299 (rs6000_emit_move): Add support for V1TImode.
6300 (altivec_expand_ld_builtin): Likewise.
6301 (altivec_expand_st_builtin): Likewise.
6302 (altivec_expand_vec_init_builtin): Likewise.
6303 (altivec_expand_builtin): Likewise.
6304 (rs6000_init_builtins): Add support for V1TImode type. Add
6305 support for ISA 2.07 128-bit integer builtins. Define type names
6306 for the VSX/Altivec vector types.
6307 (altivec_init_builtins): Add support for overloaded vector
6308 functions with V1TImode type.
6309 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
6310 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
6312 (rs6000_split_128bit_ok_p): Likewise.
6313 (rs6000_handle_altivec_attribute): Create V1TImode from vector
6314 __int128_t and vector __uint128_t.
6316 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
6317 and mode attributes.
6324 (VS_scalar): Likewise.
6325 (VS_double): Likewise.
6326 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
6328 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
6329 we support the ISA 2.07 128-bit integer arithmetic instructions.
6330 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
6331 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
6332 and TImode types for use with the builtin functions.
6333 (V1TI_type_node): Likewise.
6334 (unsigned_V1TI_type_node): Likewise.
6335 (intTI_type_internal_node): Likewise.
6336 (uintTI_type_internal_node): Likewise.
6338 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
6339 128-bit builtin functions.
6340 (UNSPEC_VADDEUQM): Likewise.
6341 (UNSPEC_VADDECUQ): Likewise.
6342 (UNSPEC_VSUBCUQ): Likewise.
6343 (UNSPEC_VSUBEUQM): Likewise.
6344 (UNSPEC_VSUBECUQ): Likewise.
6345 (VM): Add V1TImode to vector mode iterators.
6347 (VI_unit): Likewise.
6348 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
6349 (altivec_vaddcuq): Likewise.
6350 (altivec_vsubuqm): Likewise.
6351 (altivec_vsubcuq): Likewise.
6352 (altivec_vaddeuqm): Likewise.
6353 (altivec_vaddecuq): Likewise.
6354 (altivec_vsubeuqm): Likewise.
6355 (altivec_vsubecuq): Likewise.
6357 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
6359 (BOOL_128): Likewise.
6360 (BOOL_REGS_OUTPUT): Likewise.
6361 (BOOL_REGS_OP1): Likewise.
6362 (BOOL_REGS_OP2): Likewise.
6363 (BOOL_REGS_UNARY): Likewise.
6364 (BOOL_REGS_AND_CR0): Likewise.
6366 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
6367 128-bit integer builtin support.
6368 (vec_vadduqm): Likewise.
6369 (vec_vaddecuq): Likewise.
6370 (vec_vaddeuqm): Likewise.
6371 (vec_vsubecuq): Likewise.
6372 (vec_vsubeuqm): Likewise.
6373 (vec_vsubcuq): Likewise.
6374 (vec_vsubuqm): Likewise.
6376 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
6377 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
6378 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
6379 128-bit integer add/subtract to ISA 2.07.
6381 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
6383 * config/arc/arc.c (arc_predicate_delay_insns):
6384 Fix third argument passed to conditionalize_nonjump.
6386 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
6388 * config/aarch64/aarch64-builtins.c
6389 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
6390 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
6391 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
6392 instead of __builtin_lfloor.
6393 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
6395 2014-03-12 Jakub Jelinek <jakub@redhat.com>
6397 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
6398 (tree_ssa_ifcombine_bb_1): New function.
6399 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
6400 is an empty forwarder block to then_bb or vice versa and then_bb
6401 and else_bb are effectively swapped.
6403 2014-03-12 Christian Bruel <christian.bruel@st.com>
6406 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
6407 REG_CFA_DEF_CFA note.
6408 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
6409 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
6411 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
6413 PR tree-optimization/60454
6414 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
6416 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6418 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
6419 Do not define target_cpu_default2 to generic.
6420 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
6421 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
6422 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
6424 2014-03-12 Jakub Jelinek <jakub@redhat.com>
6425 Marc Glisse <marc.glisse@inria.fr>
6427 PR tree-optimization/60502
6428 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
6429 instead of build_low_bits_mask.
6431 2014-03-12 Jakub Jelinek <jakub@redhat.com>
6434 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
6435 if there are multiple uses, but op doesn't live on E edge.
6436 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
6437 clobber stmts before __builtin_unreachable.
6439 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
6441 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
6442 hard_frame_pointer_rtx.
6443 * cse.c (cse_insn): Remove volatile check.
6444 * cselib.c (cselib_process_insn): Likewise.
6445 * dse.c (scan_insn): Likewise.
6447 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
6449 * config/arc/arc.c (conditionalize_nonjump): New function,
6451 (arc_ifcvt): ... this.
6452 (arc_predicate_delay_insns): Use it.
6454 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
6456 * config/arc/predicates.md (extend_operand): During/after reload,
6457 allow const_int_operand.
6458 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
6459 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
6460 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
6462 (umulsi3_highpart_i): Likewise.
6464 2014-03-11 Richard Biener <rguenther@suse.de>
6466 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
6467 Add asserts to guard possible wrong-code bugs.
6469 2014-03-11 Richard Biener <rguenther@suse.de>
6471 PR tree-optimization/60429
6472 PR tree-optimization/60485
6473 * tree-ssa-structalias.c (set_union_with_increment): Properly
6474 take into account all fields that overlap the shifted vars.
6475 (do_sd_constraint): Likewise.
6476 (do_ds_constraint): Likewise.
6477 (get_constraint_for_ptr_offset): Likewise.
6479 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
6481 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
6482 (nios2_compute_frame_layout):
6483 Add calculation of cfun->machine->fp_save_offset.
6484 (nios2_expand_prologue): Correct setting of frame pointer register
6486 (nios2_expand_epilogue): Update recovery of stack pointer from
6487 frame pointer accordingly.
6488 (nios2_initial_elimination_offset): Update calculation of offset
6489 for eliminating to HARD_FRAME_POINTER_REGNUM.
6491 2014-03-10 Jakub Jelinek <jakub@redhat.com>
6494 * ipa.c (symtab_remove_unreachable_nodes): Don't call
6495 cgraph_get_create_node on VAR_DECLs.
6497 2014-03-10 Richard Biener <rguenther@suse.de>
6500 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
6502 2014-03-08 Douglas B Rupp <rupp@gnat.com>
6504 * config/vms/vms.opt (vms_float_format): New variable.
6506 2014-03-08 Tobias Burnus <burnus@net-b.de>
6508 * doc/invoke.texi (-fcilkplus): Update implementation status.
6510 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
6511 Richard Biener <rguenther@suse.de>
6513 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
6514 consistently accross all TUs.
6515 (run_gcc): Enable -fshort-double automatically at link at link-time
6516 and disallow override.
6518 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
6521 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
6522 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
6523 if they can't be used.
6525 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6527 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
6528 for Solaris 11/x86 ld.
6529 * configure: Regenerate.
6531 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6533 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
6534 (LIB_TLS_SPEC): Save as ld_tls_libs.
6535 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
6536 (HAVE_AS_IX86_TLSLDM): New test.
6537 * configure, config.in: Regenerate.
6538 * config/i386/i386.c (legitimize_tls_address): Fall back to
6539 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
6540 cannot support TLS_MODEL_LOCAL_DYNAMIC.
6541 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
6542 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
6544 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
6546 * common.opt (fira-loop-pressure): Mark as optimization.
6548 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
6550 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
6551 an OpenMP mappable type.
6553 2014-03-06 Matthias Klose <doko@ubuntu.com>
6555 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
6556 MULTILIB_OSDIRNAMES is not defined.
6558 2014-03-06 Jakub Jelinek <jakub@redhat.com>
6559 Meador Inge <meadori@codesourcery.com>
6562 * config/arm/arm.c (arm_tls_symbol_p): Remove.
6563 (arm_legitimize_address): Call legitimize_tls_address for any
6564 arm_tls_referenced_p expression, handle constant addend. Call it
6565 before testing for !TARGET_ARM.
6566 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
6568 2014-03-06 Richard Biener <rguenther@suse.de>
6574 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
6576 * tree-streamer.c (record_common_node): Assert we don't record
6577 nodes with type double.
6578 (preload_common_node): Skip type double, complex double and double
6579 pointer since it is now frontend dependent due to fshort-double option.
6581 2014-03-06 Richard Biener <rguenther@suse.de>
6583 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
6584 or -fno-lto is specified and the linker has full plugin support.
6585 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
6586 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
6587 * lto-wrapper.c (merge_and_complain): Merge compile-time
6588 optimization levels.
6589 (run_gcc): And pass it through to the link options.
6591 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
6595 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
6597 * cselib.c (remove_useless_values): Skip to avoid quadratic
6598 behavior if the condition moved from...
6599 (cselib_process_insn): ... here holds.
6601 2014-03-05 Jakub Jelinek <jakub@redhat.com>
6604 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
6605 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
6608 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
6609 (TM_H): Add x86-tune.def.
6611 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6613 * config/aarch64/aarch64.c (generic_tunings):
6614 Use cortexa57_extra_costs.
6616 2014-03-05 Jakub Jelinek <jakub@redhat.com>
6619 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
6620 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
6621 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
6624 2014-03-04 Heiher <r@hev.cc>
6626 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
6627 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
6629 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
6631 * config/i386/predicates.md (const2356_operand): Change to ...
6632 (const2367_operand): ... this.
6633 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
6635 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6636 (*avx512pf_scatterpf<mode>sf): Ditto.
6637 (avx512pf_scatterpf<mode>df): Ditto.
6638 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6639 (*avx512pf_scatterpf<mode>df): Ditto.
6640 * config/i386/i386.c (ix86_expand_builtin): Update
6641 incorrect hint operand error message.
6643 2014-03-04 Richard Biener <rguenther@suse.de>
6645 * lto-section-in.c (lto_get_section_data): Fix const cast.
6647 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
6649 * tree-streamer.c (record_common_node): Assert we don't record
6650 nodes with type double.
6651 (preload_common_node): Skip type double, complex double and double
6652 pointer since it is now frontend dependent due to fshort-double option.
6654 2014-03-04 Richard Biener <rguenther@suse.de>
6657 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
6658 (lto_input_toplevel_asms): Likewise.
6659 * lto-section-in.c (lto_get_section_data): Instead do it here
6662 2014-03-04 Richard Biener <rguenther@suse.de>
6664 PR tree-optimization/60382
6665 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
6666 dead PHIs a reduction.
6668 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
6670 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
6672 (_mm_prefetch): Move out of GCC target("sse") pragma.
6673 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
6674 GCC target("prfchw") pragma.
6675 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
6677 * config/i386/i386.c (ix86_option_override_internal): Enable
6678 -mprfchw with -mprefetchwt1.
6680 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
6682 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
6685 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
6687 * opts.h (CL_PCH_IGNORE): Define.
6688 * targhooks.c (option_affects_pch_p):
6689 Return false for options that have CL_PCH_IGNORE set.
6690 * opt-functions.awk: Process PchIgnore.
6691 * doc/options.texi: Document PchIgnore.
6693 * config/arc/arc.opt (misize): Add PchIgnore property.
6695 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6697 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
6698 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
6699 constraint on constants to permit them being loaded into
6700 GENERAL_REGS or BASE_REGS.
6702 2014-03-03 Nick Clifton <nickc@redhat.com>
6704 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
6705 anti-cacnonical alternatives.
6706 (negandhi3_real): New pattern.
6707 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
6709 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
6711 * config/avr/avr-mcus.def: Remove atxmega16x1.
6712 * config/avr/avr-tables.opt: Regenerate.
6713 * config/avr/t-multilib: Regenerate.
6714 * doc/avr-mmcu.texi: Regenerate.
6716 2014-03-03 Tobias Grosser <tobias@grosser.es>
6717 Mircea Namolaru <mircea.namolaru@inria.fr>
6719 PR tree-optimization/58028
6720 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
6723 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6725 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
6726 not handled by recognizers.
6728 2014-03-03 Jakub Jelinek <jakub@redhat.com>
6731 * function.c (expand_function_end): Don't emit
6732 clobber_return_register sequence if clobber_after is a BARRIER.
6733 * cfgexpand.c (construct_exit_block): Append instructions before
6734 return_label to prev_bb.
6736 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6738 * config/rs6000/constraints.md: Document reserved use of "wc".
6740 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
6743 * ipa.c (function_and_variable_visibility): When dissolving comdat
6744 group, also set all symbols to local.
6746 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
6751 2013-12-14 Jan Hubicka <jh@suse.cz>
6753 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
6755 2014-03-02 Jon Beniston <jon@beniston.com>
6761 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
6762 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
6763 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
6764 (simple_return, *simple_return): New patterns
6765 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
6766 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
6768 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
6770 * dwarf2out.c (gen_subprogram_die): Tidy.
6772 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
6775 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
6776 (*mov_t_msb_neg_negc): ... this new insn.
6778 2014-02-28 Jason Merrill <jason@redhat.com>
6781 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
6784 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
6787 * dwarf2out.c (decltype_auto_die): New static.
6788 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
6789 (gen_type_die_with_usage): Handle 'decltype(auto)'.
6790 (is_cxx_auto): Likewise.
6792 2014-02-28 Ian Bolton <ian.bolton@arm.com>
6794 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
6795 we are not using general regs only.
6797 2014-02-28 Richard Biener <rguenther@suse.de>
6800 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
6801 previous fix and only allow to remove trivial pre-headers
6802 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
6803 (remove_forwarder_block): Properly update the latch of a loop.
6805 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
6808 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
6809 (cselib_preserved_hash_table): New.
6810 (preserve_constants_and_equivs): Move preserved vals to it.
6811 (cselib_find_slot): Look it up first.
6812 (cselib_init): Initialize it.
6813 (cselib_finish): Release it.
6814 (dump_cselib_table): Dump it.
6816 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
6819 * cselib.c (remove_useless_values): Skip to avoid quadratic
6820 behavior if the condition moved from...
6821 (cselib_process_insn): ... here holds.
6823 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
6826 * var-tracking.c (vt_initialize): Apply the same condition to
6827 preserve the CFA base value.
6829 2014-02-28 Joey Ye <joey.ye@arm.com>
6832 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
6833 if reload in progress or completed.
6835 2014-02-28 Tobias Burnus <burnus@net-b.de>
6838 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
6841 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
6843 * doc/tm.texi.in (Condition Code Status): Update documention for
6844 relative locations of cc0-setter and cc0-user.
6846 2014-02-27 Jeff Law <law@redhat.com>
6848 PR rtl-optimization/52714
6849 * combine.c (try_combine): When splitting an unrecognized PARALLEL
6850 into two independent simple sets, if I3 is a jump, ensure the
6851 pattern we place into I3 is a (set (pc) ...).
6853 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
6854 Jeff Law <law@redhat.com>
6856 PR rtl-optimization/49847
6857 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
6858 are in different blocks.
6859 * doc/tm.texi (Condition Code Status): Update documention for
6860 relative locations of cc0-setter and cc0-user.
6862 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
6865 * lra.c (lra_emit_add): Check SUBREG too.
6867 2014-02-27 Andreas Schwab <schwab@suse.de>
6869 * config/m68k/m68k.c (m68k_option_override): Disable
6870 -flive-range-shrinkage for classic m68k.
6871 (m68k_override_options_after_change): Likewise.
6873 2014-02-27 Marek Polacek <polacek@redhat.com>
6876 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
6877 -Wmaybe-uninitialized.
6879 2014-02-27 Alan Modra <amodra@gmail.com>
6882 * reload1.c (emit_input_reload_insns): When reload_override_in,
6883 set old to rl->in_reg when rl->in_reg is a subreg.
6885 2014-02-26 Richard Biener <rguenther@suse.de>
6888 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
6890 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
6892 * common/config/i386/predicates.md (const1256_operand): Remove.
6893 (const2356_operand): New.
6894 (const_1_to_2_operand): Remove.
6895 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
6896 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6897 (*avx512pf_gatherpf<mode>sf): Ditto.
6898 (avx512pf_gatherpf<mode>df): Ditto.
6899 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6900 (*avx512pf_gatherpf<mode>df): Ditto.
6901 (avx512pf_scatterpf<mode>sf): Ditto.
6902 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6903 (*avx512pf_scatterpf<mode>sf): Ditto.
6904 (avx512pf_scatterpf<mode>df): Ditto.
6905 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6906 (*avx512pf_scatterpf<mode>df): Ditto.
6907 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
6909 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
6911 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
6912 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
6913 (_mm512_mask_testn_epi64_mask): Move to ...
6914 * config/i386/avx512cdintrin.h: Here.
6915 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
6916 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
6917 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
6918 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
6919 TARGET_AVX512F from TARGET_AVX512CD.
6921 2014-02-26 Richard Biener <rguenther@suse.de>
6924 * ipa.c (walk_polymorphic_call_targets): Properly guard
6925 call to inline_update_overall_summary.
6927 2014-02-26 Bin Cheng <bin.cheng@arm.com>
6930 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
6931 and latches only if requested. Fix latch if it is removed.
6932 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
6933 LOOPS_HAVE_PREHEADERS.
6935 2014-02-25 Andrew Pinski <apinski@cavium.com>
6937 * builtins.c (expand_builtin_thread_pointer): Create a new target
6938 when the target is NULL.
6940 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
6942 PR rtl-optimization/60317
6943 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
6944 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
6945 * lra-assigns.c: Include params.h.
6946 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
6947 other reload pseudos considerations.
6949 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6951 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
6952 to use canonical form for nor<mode>3.
6954 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6957 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
6960 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
6962 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
6963 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
6964 (ix86_handle_option): Handle OPT_mprefetchwt1.
6965 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
6966 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
6968 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
6969 OPTION_MASK_ISA_PREFETCHWT1.
6970 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
6971 (PTA_PREFETCHWT1): New.
6972 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
6973 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
6974 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
6975 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
6976 (*prefetch_avx512pf_<mode>_: Change into ...
6977 (*prefetch_prefetchwt1_<mode>: This.
6978 * config/i386/i386.opt (mprefetchwt1): New.
6979 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
6980 (_mm_prefetch): Handle intent to write.
6981 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
6983 2014-02-25 Richard Biener <rguenther@suse.de>
6986 * emit-rtl.c (mem_attrs_htab): Remove.
6987 (mem_attrs_htab_hash): Likewise.
6988 (mem_attrs_htab_eq): Likewise.
6989 (set_mem_attrs): Always allocate new mem-attrs when something changed.
6990 (init_emit_once): Do not allocate mem_attrs_htab.
6992 2014-02-25 Richard Biener <rguenther@suse.de>
6995 * lto-opts.c (lto_write_options): Output non-explicit conservative
6996 -fwrapv, -fno-trapv and -fno-strict-overflow.
6997 * lto-wrapper.c (merge_and_complain): Handle merging those options.
6998 (run_gcc): And pass them through.
7000 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
7002 * sel-sched.c (calculate_new_fences): New parameter ptime.
7003 Calculate it as a maximum over all fence cycles.
7004 (sel_sched_region_2): Adjust the call to calculate_new_fences.
7005 Print the final schedule timing when sched_verbose.
7007 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
7009 PR rtl-optimization/60292
7010 * sel-sched.c (fill_vec_av_set): Do not reset target availability
7011 bit fot the fence instruction.
7013 2014-02-24 Alangi Derick <alangiderick@gmail.com>
7015 * calls.h: Fix typo in comment.
7017 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
7019 * config/pa/pa.c (pa_output_move_double): Don't valididate when
7020 adjusting offsetable addresses.
7022 2014-02-24 Guozhi Wei <carrot@google.com>
7024 * sparseset.h (sparseset_pop): Fix the wrong index.
7026 2014-02-24 Walter Lee <walt@tilera.com>
7028 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
7029 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
7031 * common/config/tilegx/tilegx-common.c
7032 (TARGET_DEFAULT_TARGET_FLAGS): Define.
7033 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
7035 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
7036 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
7037 (tilegx_gimplify_va_arg_expr): Handle big endian.
7038 (tilegx_expand_unaligned_load): Ditto.
7039 (tilegx_expand_unaligned_store): Ditto.
7040 (TARGET_RETURN_IN_MSB): New.
7041 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
7042 (TARGET_ENDIAN_DEFAULT): New.
7043 (TARGET_BIG_ENDIAN): Handle big endian.
7044 (BYTES_BIG_ENDIAN): Ditto.
7045 (WORDS_BIG_ENDIAN): Ditto.
7046 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
7049 * config/tilegx/tilegx.md (extv): Handle big endian.
7051 (insn_st<n>): Ditto.
7052 (insn_st<n>_add<bitsuffix>): Ditto.
7053 (insn_stnt<n>): Ditto.
7054 (insn_stnt<n>_add<bitsuffix>):Ditto.
7055 (vec_interleave_highv8qi): Handle big endian.
7056 (vec_interleave_highv8qi_be): New.
7057 (vec_interleave_highv8qi_le): New.
7058 (insn_v1int_h): Handle big endian.
7059 (vec_interleave_lowv8qi): Handle big endian.
7060 (vec_interleave_lowv8qi_be): New.
7061 (vec_interleave_lowv8qi_le): New.
7062 (insn_v1int_l): Handle big endian.
7063 (vec_interleave_highv4hi): Handle big endian.
7064 (vec_interleave_highv4hi_be): New.
7065 (vec_interleave_highv4hi_le): New.
7066 (insn_v2int_h): Handle big endian.
7067 (vec_interleave_lowv4hi): Handle big endian.
7068 (vec_interleave_lowv4hi_be): New.
7069 (vec_interleave_lowv4hi_le): New.
7070 (insn_v2int_l): Handle big endian.
7071 (vec_interleave_highv2si): Handle big endian.
7072 (vec_interleave_highv2si_be): New.
7073 (vec_interleave_highv2si_le): New.
7074 (insn_v4int_h): Handle big endian.
7075 (vec_interleave_lowv2si): Handle big endian.
7076 (vec_interleave_lowv2si_be): New.
7077 (vec_interleave_lowv2si_le): New.
7078 (insn_v4int_l): Handle big endian.
7079 * config/tilegx/tilegx.opt (mbig-endian): New option.
7080 (mlittle-endian): New option.
7081 * doc/install.texi: Document tilegxbe-linux.
7082 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
7084 2014-02-24 Martin Jambor <mjambor@suse.cz>
7087 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
7088 there are no parameter descriptors.
7090 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
7092 PR rtl-optimization/60268
7093 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
7094 initialization to ...
7095 (sched_rgn_init): ... here.
7096 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
7098 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
7100 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
7103 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
7105 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
7108 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
7110 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
7111 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
7113 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
7115 * config/microblaze/predicates.md: Add cmp_op predicate.
7116 * config/microblaze/microblaze.md: Add branch_compare instruction
7117 which uses cmp_op predicate and emits cmp insn before branch.
7118 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
7119 to microblaze_expand_conditional_branch and consolidate logic.
7120 (microblaze_expand_conditional_branch): emit branch_compare
7121 insn instead of handling cmp op separate from branch insn.
7123 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7125 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
7128 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7130 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
7131 define_insn with define_expand and new define_insn
7132 *altivec_lve<VI_char>x_internal.
7133 (altivec_stve<VI_char>x): Replace define_insn with define_expand
7134 and new define_insn *altivec_stve<VI_char>x_internal.
7135 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
7137 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
7139 (altivec_expand_stvex_be): New function.
7141 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
7143 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
7144 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
7145 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
7146 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
7148 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
7151 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
7152 instead of emit_move_insn.
7154 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7156 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
7158 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
7159 gen_altivec_vsumsws.
7161 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7163 * config/rs6000/altivec.md (altivec_lvxl): Rename as
7164 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
7165 (altivec_lvxl_<mode>): New define_expand incorporating
7166 -maltivec=be semantics where needed.
7167 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
7168 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
7169 semantics where needed.
7170 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
7171 (altivec_stvx_<mode>): New define_expand incorporating
7172 -maltivec=be semantics where needed.
7173 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
7174 VM2 iterator instead of V4SI.
7175 (altivec_stvxl_<mode>): New define_expand incorporating
7176 -maltivec=be semantics where needed.
7177 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
7178 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
7179 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
7180 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
7181 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
7182 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
7183 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
7184 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
7185 ALTIVEC_BUILTIN_STVXL.
7186 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
7187 (altivec_expand_stvx_be): Likewise.
7188 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
7189 (altivec_expand_lvx_be): Likewise.
7190 (altivec_expand_stvx_be): Likewise.
7191 (altivec_expand_builtin): Add cases for
7192 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
7193 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
7194 (altivec_init_builtins): Add definitions for
7195 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
7196 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
7198 2014-02-21 Catherine Moore <clm@codesourcery.com>
7200 * doc/invoke.texi (mvirt, mno-virt): Document.
7201 * config/mips/mips.opt (mvirt): New option.
7202 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
7204 2014-02-21 Richard Biener <rguenther@suse.de>
7206 PR tree-optimization/60276
7207 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
7208 (STMT_VINFO_MIN_NEG_DIST): New macro.
7209 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
7210 STMT_VINFO_MIN_NEG_DIST.
7211 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
7212 made for negative dependence distances still hold.
7214 2014-02-21 Richard Biener <rguenther@suse.de>
7217 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
7218 DECL_INITIAL for globals not in the current function context.
7220 2014-02-21 Jakub Jelinek <jakub@redhat.com>
7222 PR tree-optimization/56490
7223 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
7224 * tree-ssa-uninit.c: Include params.h.
7225 (compute_control_dep_chain): Add num_calls argument, return false
7226 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
7227 num_calls to recursive call.
7228 (find_predicates): Change dep_chain into normal array,
7229 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
7230 variable and adjust compute_control_dep_chain caller.
7231 (find_def_preds): Likewise.
7233 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
7235 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
7236 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
7238 2014-02-21 Nick Clifton <nickc@redhat.com>
7240 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
7241 (pushhi1): Likewise.
7242 (popqi1): Add mode to pre_dec.
7245 2014-02-21 Jakub Jelinek <jakub@redhat.com>
7247 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
7248 mode for mask of V8SFmode permutation.
7250 2014-02-20 Richard Henderson <rth@redhat.com>
7253 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
7254 a new pseudo for OLDVAL.
7256 2014-02-20 Jakub Jelinek <jakub@redhat.com>
7259 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
7260 gen_reg_rtx if d->testing_p.
7261 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
7262 if d->testing_p and we will certainly return true.
7263 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
7266 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
7268 * emit-rtl.c (gen_reg_rtx): Assert that
7269 crtl->emit.regno_pointer_align_length is non-zero.
7271 2014-02-20 Richard Henderson <rth@redhat.com>
7274 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
7275 on failure the store back into EXPECT.
7277 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
7278 Sandra Loosemore <sandra@codesourcery.com>
7280 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
7281 * config/nios2/nios2.c (nios2_function_profiler): Add
7282 -fPIC (flag_pic == 2) support.
7283 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
7284 (nios2_large_offset_p): New function.
7285 (nios2_unspec_reloc_p): Move up position, update to use
7286 nios2_large_offset_p.
7287 (nios2_unspec_address): Remove function.
7288 (nios2_unspec_offset): New function.
7289 (nios2_large_got_address): New function.
7290 (nios2_got_address): Add large offset support.
7291 (nios2_legitimize_tls_address): Update usage of removed and new
7293 (nios2_symbol_binds_local_p): New function.
7294 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
7295 (nios2_legitimize_address): Update to use nios2_large_offset_p.
7296 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
7297 (nios2_print_operand): Merge H/L processing, add hiadj/lo
7298 processing for (const (unspec ...)).
7299 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
7301 2014-02-20 Richard Biener <rguenther@suse.de>
7303 * tree-cfg.c (replace_uses_by): Mark altered BBs before
7304 doing the substitution.
7305 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
7307 2014-02-20 Martin Jambor <mjambor@suse.cz>
7310 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
7311 info when checking whether lattices are bottom.
7313 2014-02-20 Richard Biener <rguenther@suse.de>
7316 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
7319 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
7322 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
7323 parameter specifying the scaling.
7324 (inline_call): Update.
7325 (want_inline_recursively): Guard division by zero.
7326 (recursive_inlining): Update.
7327 * ipa-inline.h (clone_inlined_nodes): Update.
7329 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
7332 * config/i386/i386.c (classify_argument): Pass structures of size
7333 64 bytes or less in register.
7335 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
7336 Kirill Yukhin <kirill.yukhin@intel.com>
7338 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
7339 (_mm_rcp28_round_ss): Ditto.
7340 (_mm_rsqrt28_round_sd): Ditto.
7341 (_mm_rsqrt28_round_ss): Ditto.
7342 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
7343 (_mm_rcp14_round_ss): Ditto.
7344 (_mm_rsqrt14_round_sd): Ditto.
7345 (_mm_rsqrt14_round_ss): Ditto.
7346 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
7347 the first input operand, get rid of match_dup.
7348 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
7350 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
7352 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
7353 operand as the first input operand, set type attribute.
7354 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
7356 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
7357 operand as the first input operand, set type attribute.
7359 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7361 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
7364 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
7367 * config/i386/i386.c (construct_container): Remove TFmode check
7368 for X86_64_INTEGER_CLASS.
7370 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
7373 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
7374 only when -Wpsabi is enabled.
7376 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
7379 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
7380 passing arrays in registers are the same as for structs, so remove the
7381 special case for them.
7383 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
7385 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
7386 destination type, extract only the valid bits if the source type is not
7387 integral and has a different mode.
7389 2014-02-19 Richard Biener <rguenther@suse.de>
7392 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
7395 2014-02-19 Richard Biener <rguenther@suse.de>
7398 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
7399 (ipa_modify_call_arguments): Emit an argument load explicitely and
7400 preserve virtual SSA form there and for the replacement call.
7401 Do not update SSA form nor free dominance info.
7403 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
7405 * ipa.c (function_and_variable_visibility): Also clear WEAK
7406 flag when disolving COMDAT_GROUP.
7408 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
7410 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
7411 * ipa-prop.c (ipa_set_jf_known_type): Return early when
7413 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
7414 do more sanity checks.
7415 (detect_type_change): Return true when giving up early.
7416 (compute_complex_assign_jump_func): Fix type parameter of
7417 ipa_set_ancestor_jf.
7418 (compute_complex_ancestor_jump_func): Likewise.
7419 (update_jump_functions_after_inlining): Fix updating of
7421 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
7423 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
7425 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
7426 inline clones when edge disappears.
7428 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
7431 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
7432 Split 64-bit moves into 2 patterns. Do not allow the use of
7433 direct move for TDmode in little endian, since the decimal value
7434 has little endian bytes within a word, but the 64-bit pieces are
7435 ordered in a big endian fashion, and normal subreg's of TDmode are
7437 (mov<mode>_64bit_dm): Likewise.
7438 (movtd_64bit_nodm): Likewise.
7440 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
7442 PR tree-optimization/60174
7443 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
7444 statement of an SSA_NAME that occurs in an abnormal PHI node.
7446 2014-02-18 Jakub Jelinek <jakub@redhat.com>
7449 * final.c (SEEN_BB): Remove.
7450 (SEEN_NOTE, SEEN_EMITTED): Renumber.
7451 (final_scan_insn): Don't force_source_line on second
7452 NOTE_INSN_BASIC_BLOCK.
7454 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
7457 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
7458 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
7459 (type_natural_mode): Warn ABI change when %zmm register is not
7460 available for AVX512F vector value passing.
7462 2014-02-18 Kai Tietz <ktietz@redhat.com>
7465 * config/i386/i386.c (ix86_expand_prologue): Use value in
7466 rax register as displacement when restoring %r10 or %rax.
7467 Fix wrong offset when restoring both registers.
7469 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
7471 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
7472 assertion with conditional return.
7474 2014-02-18 Jakub Jelinek <jakub@redhat.com>
7475 Uros Bizjak <ubizjak@gmail.com>
7478 * config/i386/driver-i386.c (host_detect_local_cpu): If
7479 YMM state is not saved by the OS, also clear has_f16c. Move
7480 CPUID 0x80000001 handling before YMM state saving checking.
7482 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
7484 PR rtl-optimization/58960
7485 * haifa-sched.c (alloc_global_sched_pressure_data): New,
7486 factored out from ...
7487 (sched_init): ... here.
7488 (free_global_sched_pressure_data): New, factored out from ...
7489 (sched_finish): ... here.
7490 * sched-int.h (free_global_sched_pressure_data): Declare.
7491 * sched-rgn.c (nr_regions_initial): New static global.
7492 (haifa_find_rgns): Initialize it.
7493 (schedule_region): Disable sched-pressure for the newly
7496 2014-02-17 Richard Biener <rguenther@suse.de>
7498 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
7499 release SSA defs of pattern stmts.
7501 2014-02-17 Richard Biener <rguenther@suse.de>
7503 * tree-inline.c (expand_call_inline): Release the virtual
7504 operand defined by the call we are about to inline.
7506 2014-02-17 Richard Biener <rguenther@suse.de>
7508 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
7510 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
7511 Ilya Tocar <ilya.tocar@intel.com>
7513 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
7514 arguments order in builtin.
7515 (_mm512_permutexvar_epi64): Ditto.
7516 (_mm512_mask_permutexvar_epi64): Ditto
7517 (_mm512_maskz_permutexvar_epi32): Ditto
7518 (_mm512_permutexvar_epi32): Ditto
7519 (_mm512_mask_permutexvar_epi32): Ditto
7521 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7523 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
7524 (p8_vmrgow): Likewise.
7526 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7528 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
7531 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
7534 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
7535 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
7536 into 64-bit and 32-bit moves. On 64-bit moves, add support for
7537 using direct move instructions on ISA 2.07. Also adjust
7538 instruction length for 64-bit.
7539 (mov<mode>_64bit, TFmode/TDmode): Likewise.
7540 (mov<mode>_32bit, TFmode/TDmode): Likewise.
7542 2014-02-15 Alan Modra <amodra@gmail.com>
7546 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
7547 find_replacement on parts of insn rtl that might be reloaded.
7549 2014-02-15 Richard Biener <rguenther@suse.de>
7551 PR tree-optimization/60183
7552 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
7553 (tree_ssa_phiprop): Calculate and free post-dominators.
7555 2014-02-14 Jeff Law <law@redhat.com>
7557 PR rtl-optimization/60131
7558 * ree.c (get_extended_src_reg): New function.
7559 (combine_reaching_defs): Use it rather than assuming location of REG.
7560 (find_and_remove_re): Verify first operand of extension is
7561 a REG before adding the insns to the copy list.
7563 2014-02-14 Roland McGrath <mcgrathr@google.com>
7565 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
7566 * configure: Regenerated.
7567 * config.in: Regenerated.
7568 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
7569 instead of ASM_SHORT.
7571 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
7572 Richard Earnshaw <rearnsha@arm.com>
7574 PR rtl-optimization/59535
7575 * lra-constraints.c (process_alt_operands): Encourage alternative
7576 when unassigned pseudo class is superset of the alternative class.
7577 (inherit_reload_reg): Don't inherit when optimizing for code size.
7578 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
7579 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
7580 modes not less than 4 for Thumb1.
7582 2014-02-14 Kyle McMartin <kyle@redhat.com>
7585 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
7587 2014-02-14 Richard Biener <rguenther@suse.de>
7589 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
7590 (get_frame_arg): Drop the assert with langhook types_compatible_p.
7591 Do not strip INDIRECT_REFs.
7593 2014-02-14 Richard Biener <rguenther@suse.de>
7596 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
7597 DECL_FUNCTION_SPECIFIC_TARGET.
7598 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
7599 * tree-streamer-out.c (pack_ts_target_option): Remove.
7600 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
7601 (write_ts_function_decl_tree_pointers): Do not stream
7602 DECL_FUNCTION_SPECIFIC_TARGET.
7603 * tree-streamer-in.c (unpack_ts_target_option): Remove.
7604 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
7605 (lto_input_ts_function_decl_tree_pointers): Do not stream
7606 DECL_FUNCTION_SPECIFIC_TARGET.
7608 2014-02-14 Jakub Jelinek <jakub@redhat.com>
7610 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
7611 (get_initial_def_for_induction, vectorizable_induction): Ignore
7612 debug stmts when looking for exit_phi.
7613 (vectorizable_live_operation): Fix up condition.
7615 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
7617 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
7618 nreverse() because it changes the content of original tree list.
7620 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
7622 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
7623 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
7625 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
7627 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
7628 GNU coding standards.
7630 2014-02-13 Jakub Jelinek <jakub@redhat.com>
7633 * dwarf2out.c (gen_subprogram_die): Don't call
7634 add_calling_convention_attribute if subr_die is old_die.
7636 2014-02-13 Sharad Singhai <singhai@google.com>
7638 * doc/optinfo.texi: Fix order of nodes.
7640 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
7642 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
7643 operands[2], not operands[3].
7645 2014-02-13 Richard Biener <rguenther@suse.de>
7648 * doc/install.texi (ISL): Update recommended version to 0.12.2,
7649 mention the possibility of an in-tree build.
7650 (CLooG): Update recommended version to 0.18.1, mention the
7651 possibility of an in-tree build and clarify that the ISL
7652 bundled with CLooG does not work.
7654 2014-02-13 Jakub Jelinek <jakub@redhat.com>
7657 * expr.c (compress_float_constant): If x is a hard register,
7658 extend into a pseudo and then move to x.
7660 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
7662 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
7663 caused by bad second argument to warning_at() with -mhotpatch and
7664 nested functions (e.g. with gfortran).
7666 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
7668 * opts.c (option_name): Remove "enabled by default" rider.
7670 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
7672 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
7674 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
7675 Uros Bizjak <ubizjak@gmail.com>
7678 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
7679 * configure: Regenerated.
7681 2014-02-12 Richard Biener <rguenther@suse.de>
7683 * vec.c (vec_prefix::calculate_allocation): Move as
7684 inline variant to vec.h.
7685 (vec_prefix::calculate_allocation_1): New out-of-line version.
7686 * vec.h (vec_prefix::calculate_allocation_1): Declare.
7687 (vec_prefix::m_has_auto_buf): Rename to ...
7688 (vec_prefix::m_using_auto_storage): ... this.
7689 (vec_prefix::calculate_allocation): Inline the easy cases
7690 and dispatch to calculate_allocation_1 which doesn't need the
7692 (va_heap::reserve): Use gcc_checking_assert.
7693 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
7694 m_using_auto_storage.
7695 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
7697 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
7698 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
7699 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
7701 2014-02-12 Richard Biener <rguenther@suse.de>
7703 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
7704 when we found a dependence.
7706 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
7708 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
7710 (maybe_fold_stmt): ... into this new function.
7711 * omp-low.c (lower_omp): Update comment.
7713 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
7716 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
7719 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
7721 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
7722 identifiers in comments.
7723 (cortexa53_extra_costs): Likewise.
7724 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
7725 (cortexa7_extra_costs): Likewise.
7726 (cortexa12_extra_costs): Likewise.
7727 (cortexa15_extra_costs): Likewise.
7728 (v7m_extra_costs): Likewise.
7730 2014-02-12 Richard Biener <rguenther@suse.de>
7733 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
7734 of posix_memalign being successful.
7735 (lower_stmt): Restrict lowering of posix_memalign to when
7736 -ftree-bit-ccp is enabled.
7738 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
7740 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
7742 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
7744 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
7746 PR rtl-optimization/60116
7747 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
7748 other_insn once the combination has been validated.
7750 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
7753 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
7755 * ipa-devirt.c: Include demangle.h
7756 (odr_violation_reported): New static variable.
7757 (add_type_duplicate): Update odr_violations.
7758 (maybe_record_node): Add completep parameter; update it.
7759 (record_target_from_binfo): Add COMPLETEP parameter;
7760 update it as needed.
7761 (possible_polymorphic_call_targets_1): Likewise.
7762 (struct polymorphic_call_target_d): Add nonconstruction_targets;
7763 rename FINAL to COMPLETE.
7764 (record_targets_from_bases): Sanity check we found the binfo;
7765 fix COMPLETEP updating.
7766 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
7767 parameter, fix computing of COMPLETEP.
7768 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
7769 at LTO time do demangling.
7770 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
7771 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
7773 (gimple_get_virt_method_for_binfo): Likewise.
7774 * gimple-fold.h (gimple_get_virt_method_for_binfo,
7775 gimple_get_virt_method_for_vtable): Update prototypes.
7777 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
7780 * genautomata.c (add_presence_absence): Fix typo with
7781 {final_}presence_list.
7783 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
7786 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
7787 for VSX/Altivec vectors that land in GPR registers.
7789 2014-02-11 Richard Henderson <rth@redhat.com>
7790 Jakub Jelinek <jakub@redhat.com>
7793 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
7794 around drhs if type conversion to lacc->type is not useless.
7796 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7798 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
7800 (cortex-a57.cortex-a53): Likewise.
7801 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
7803 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7805 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
7808 2014-02-11 Renlin Li <Renlin.Li@arm.com>
7810 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
7811 add_options_for_arm_vfp3.
7813 2014-02-11 Jeff Law <law@redhat.com>
7816 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
7817 object with an undesirable mode.
7819 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7822 * config/i386/sol2-9.h: New file.
7823 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
7824 *-*-solaris2.9*): Use it.
7826 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
7828 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
7829 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
7831 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
7833 * config/microblaze/microblaze.c: Extend mcpu version format
7835 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
7837 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
7839 2014-02-10 Richard Henderson <rth@redhat.com>
7842 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
7843 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
7844 ms-abi vs -mno-accumulate-outgoing-args.
7845 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
7846 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
7849 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
7852 * cfgexpand.c (expand_asm_operands): Attach source location to
7853 ASM_INPUT rtx objects.
7854 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
7856 2014-02-10 Nick Clifton <nickc@redhat.com>
7858 * config/mn10300/mn10300.c (popcount): New function.
7859 (mn10300_expand_prologue): Include saved registers in stack usage
7862 2014-02-10 Jeff Law <law@redhat.com>
7865 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
7866 when changing the SET_DEST of a prior insn to avoid an input reload.
7868 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
7870 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
7871 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
7872 -mcall-openbsd, or -mcall-linux.
7873 (CC1_ENDIAN_BIG_SPEC): Remove.
7874 (CC1_ENDIAN_LITTLE_SPEC): Remove.
7875 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
7876 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
7877 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
7878 and %cc1_endian_default.
7879 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
7881 2014-02-10 Richard Biener <rguenther@suse.de>
7883 PR tree-optimization/60115
7884 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
7885 MEM_REF handling. Properly verify that the accesses are not
7886 out of the objects bound.
7888 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7890 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
7893 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
7895 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
7896 proper constants and fix formatting.
7897 (possible_polymorphic_call_targets): Fix formatting.
7899 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
7900 Ilya Tocar <ilya.tocar@intel.com>
7902 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
7903 (_mm512_loadu_epi32): Renamed into...
7904 (_mm512_loadu_si512): This.
7905 (_mm512_storeu_epi32): Renamed into...
7906 (_mm512_storeu_si512): This.
7907 (_mm512_maskz_ceil_ps): Removed.
7908 (_mm512_maskz_ceil_pd): Ditto.
7909 (_mm512_maskz_floor_ps): Ditto.
7910 (_mm512_maskz_floor_pd): Ditto.
7911 (_mm512_floor_round_ps): Ditto.
7912 (_mm512_floor_round_pd): Ditto.
7913 (_mm512_ceil_round_ps): Ditto.
7914 (_mm512_ceil_round_pd): Ditto.
7915 (_mm512_mask_floor_round_ps): Ditto.
7916 (_mm512_mask_floor_round_pd): Ditto.
7917 (_mm512_mask_ceil_round_ps): Ditto.
7918 (_mm512_mask_ceil_round_pd): Ditto.
7919 (_mm512_maskz_floor_round_ps): Ditto.
7920 (_mm512_maskz_floor_round_pd): Ditto.
7921 (_mm512_maskz_ceil_round_ps): Ditto.
7922 (_mm512_maskz_ceil_round_pd): Ditto.
7923 (_mm512_expand_pd): Ditto.
7924 (_mm512_expand_ps): Ditto.
7925 * config/i386/i386.c (ix86_builtins): Remove
7926 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
7927 (bdesc_args): Ditto.
7928 * config/i386/predicates.md (const1256_operand): New.
7929 (const_1_to_2_operand): Ditto.
7930 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
7931 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
7932 (*avx512pf_gatherpf<mode>sf): Ditto.
7933 (avx512pf_gatherpf<mode>df): Ditto.
7934 (*avx512pf_gatherpf<mode>df_mask): Ditto.
7935 (*avx512pf_gatherpf<mode>df): Ditto.
7936 (avx512pf_scatterpf<mode>sf): Ditto.
7937 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
7938 (*avx512pf_scatterpf<mode>sf): Ditto.
7939 (avx512pf_scatterpf<mode>df): Ditto.
7940 (*avx512pf_scatterpf<mode>df_mask): Ditto.
7941 (*avx512pf_scatterpf<mode>df): Ditto.
7942 (avx512f_expand<mode>): Removed.
7943 (<shift_insn><mode>3<mask_name>): Change predicate type.
7945 2014-02-08 Jakub Jelinek <jakub@redhat.com>
7947 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
7948 not at the end of datarefs vector use ordered_remove to avoid
7949 reordering datarefs vector.
7952 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
7953 mark local addressable non-static vars as GOVD_PRIVATE
7954 instead of GOVD_LOCAL.
7955 * omp-low.c (lower_omp_for): Move gimple_bind_vars
7956 and BLOCK_VARS of gimple_bind_block to new_stmt rather
7960 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
7961 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
7962 assume_aligned or alloc_align attributes.
7963 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
7964 arguments. Handle also assume_aligned and alloc_align attributes.
7965 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
7966 calls to functions with assume_aligned or alloc_align attributes.
7967 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
7969 2014-02-08 Terry Guo <terry.guo@arm.com>
7971 * doc/invoke.texi: Document ARM -march=armv7e-m.
7973 2014-02-08 Jakub Jelinek <jakub@redhat.com>
7975 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
7976 flag on __cilkrts_rethrow builtin.
7979 * ipa-cp.c (determine_versionability): Fail at -O0
7980 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
7981 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
7984 2014-02-04 Jakub Jelinek <jakub@redhat.com>
7987 * tree-inline.c (copy_forbidden): Fail for
7988 __attribute__((optimize (0))) functions.
7990 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
7992 * varpool.c: Include pointer-set.h.
7993 (varpool_remove_unreferenced_decls): Variables in other partitions
7994 will not be output; be however careful to not lose information
7997 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
7999 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
8000 lookup in the vtable constructor.
8002 2014-02-07 Jeff Law <law@redhat.com>
8005 * config/m68k/m68k.md (ashldi_extsi): Turn into a
8006 define_insn_and_split.
8008 * ipa-inline.c (inline_small_functions): Fix typos.
8010 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
8012 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
8013 (s390_can_use_return_insn): Declare.
8014 * config/s390/s390.h (EPILOGUE_USES): Define.
8015 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
8017 (s390_chunkify_start): Handle return JUMP_LABELs.
8018 (s390_early_mach): Emit a main_pool instruction on the entry edge.
8019 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
8020 (s390_can_use_return_insn): New functions.
8021 (s390_fix_long_loop_prediction): Handle conditional returns.
8022 (TARGET_SET_UP_BY_PROLOGUE): Define.
8023 * config/s390/s390.md (ANY_RETURN): New code iterator.
8024 (*creturn, *csimple_return, return, simple_return): New patterns.
8026 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
8028 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
8029 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
8030 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
8031 REG_CFA_RESTORE list when deciding not to restore a register.
8033 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
8035 * config/s390/s390.c: Include tree-pass.h and context.h.
8036 (s390_early_mach): New function, split out from...
8037 (s390_emit_prologue): ...here.
8038 (pass_data_s390_early_mach): New pass structure.
8039 (pass_s390_early_mach): New class.
8040 (s390_option_override): Create and register early_mach pass.
8041 Move to end of file.
8043 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
8045 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
8046 to match for the exit block.
8048 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8050 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
8051 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
8052 Reject misaligned operands.
8054 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8056 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
8058 2014-02-07 Richard Biener <rguenther@suse.de>
8061 * gimple-low.c (lower_builtin_posix_memalign): New function.
8062 (lower_stmt): Call it to lower posix_memalign in a way
8063 to make alignment info accessible.
8065 2014-02-07 Jakub Jelinek <jakub@redhat.com>
8068 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
8069 __builtin_setjmp_receiver.
8071 2014-02-07 Richard Biener <rguenther@suse.de>
8074 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
8075 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
8076 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
8077 Handle BUILT_IN_POSIX_MEMALIGN.
8078 (find_func_clobbers): Likewise.
8079 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
8080 (call_may_clobber_ref_p_1): Likewise.
8082 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
8085 * ipa-devirt.c (record_target_from_binfo): Remove overactive
8088 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
8091 * lto-cgraph.c (lto_output_node): Use
8092 symtab_get_symbol_partitioning_class.
8093 (lto_output_varpool_node): likewise.
8094 (symtab_get_symbol_partitioning_class): Move here from
8096 * cgraph.h (symbol_partitioning_class): Likewise.
8097 (symtab_get_symbol_partitioning_class): Declare.
8099 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
8101 * ggc.h (ggc_internal_cleared_alloc): New macro.
8102 * vec.h (vec_safe_copy): Handle memory stats.
8103 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
8104 * target-globals.c (save_target_globals): Likewise.
8106 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
8109 * expr.c (emit_move_resolve_push): Export; be bit more selective
8110 on when to clear alias set.
8111 * expr.h (emit_move_resolve_push): Declare.
8112 * function.h (struct function): Add tail_call_marked.
8113 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
8114 * config/i386/i386-protos.h (ix86_expand_push): Remove.
8115 * config/i386/i386.md (TImode move expander): De not call
8117 (FP push expanders): Preserve memory attributes.
8118 * config/i386/sse.md (push<mode>1): Remove.
8119 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
8120 (ix86_expand_push): Remove.
8121 * config/i386/mmx.md (push<mode>1): Remove.
8123 2014-02-06 Jakub Jelinek <jakub@redhat.com>
8125 PR rtl-optimization/60030
8126 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
8127 lopart with paradoxical subreg before shifting it up by hprec.
8129 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8131 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
8132 Remove extra newline at end of file.
8133 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
8134 (arm_issue_rate): Handle cortexa57.
8135 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
8136 (cortex-a57.cortex-a53): Likewise.
8138 2014-02-06 Jakub Jelinek <jakub@redhat.com>
8141 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
8142 don't record in REG_FRAME_RELATED_EXPR registers not set in that
8144 (arm_expand_prologue): Adjust all callers.
8145 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
8146 info, registers also at the lowest numbered registers side. Use
8147 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
8151 * var-tracking.c (adjust_mems): Before adding a SET to
8152 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
8154 2014-02-06 Alan Modra <amodra@gmail.com>
8157 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
8158 change SDmode to DDmode when lra_in_progress.
8160 2014-02-06 Jakub Jelinek <jakub@redhat.com>
8163 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
8164 free_data_ref on the dr first, and before goto again also set dr
8165 to the next dr. For simd_lane_access, free old datarefs[i] before
8166 overwriting it. For get_vectype_for_scalar_type failure, don't
8167 free_data_ref if simd_lane_access.
8169 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
8172 * tree.h (opts_for_fn): New inline function.
8173 (opt_for_fn): Define.
8174 * config/i386/i386.c (ix86_function_regparm): Use
8175 opt_for_fn (decl, optimize) instead of optimize.
8177 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
8179 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
8180 for SYMBOL_REF in large memory model.
8182 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8184 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
8186 (cortex-a57): Likewise.
8187 (cortex-a57.cortex-a53): Likewise.
8189 2014-02-06 Yury Gribov <y.gribov@samsung.com>
8190 Kugan Vivekanandarajah <kuganv@linaro.org>
8192 * config/arm/arm.c (arm_vector_alignment_reachable): Check
8194 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
8196 2014-02-06 Richard Biener <rguenther@suse.de>
8198 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
8199 set_loop_copy and initialize_original_copy_tables.
8201 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
8203 * config/aarch64/aarch64-simd.md
8204 (aarch64_ashr_simddi): Change QI to SI.
8206 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
8207 Jakub Jelinek <jakub@redhat.com>
8210 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
8213 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8215 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
8216 CODE_FOR_altivec_vpku[hw]um to
8217 CODE_FOR_altivec_vpku[hw]um_direct.
8218 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
8219 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
8220 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
8221 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
8223 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8225 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
8226 generation for -maltivec=be.
8227 (altivec_vsumsws): Simplify redundant test.
8229 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8231 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
8232 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
8233 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
8234 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
8235 gen_altivec_vpkuwum.
8236 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
8238 (altivec_vpks<VI_char>ss): Likewise.
8239 (altivec_vpks<VI_char>us): Likewise.
8240 (altivec_vpku<VI_char>us): Likewise.
8241 (altivec_vpku<VI_char>um): Likewise.
8242 (altivec_vpku<VI_char>um_direct): New (copy of
8243 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
8245 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
8246 target is little endian and -maltivec=be is not specified.
8247 (*altivec_vupkhs<VU_char>_direct): New (copy of
8248 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
8249 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
8250 target is little endian and -maltivec=be is not specified.
8251 (*altivec_vupkls<VU_char>_direct): New (copy of
8252 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
8253 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
8254 little endian and -maltivec=be is not specified.
8255 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
8256 little endian and -maltivec=be is not specified.
8258 2014-02-05 Richard Henderson <rth@redhat.com>
8261 * combine-stack-adj.c: Revert r206943.
8262 * sched-int.h (struct deps_desc): Add last_args_size.
8263 * sched-deps.c (init_deps): Initialize it.
8264 (sched_analyze_insn): Add OUTPUT dependencies between insns that
8265 contain REG_ARGS_SIZE notes.
8267 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
8269 * lto-cgraph.c (asm_nodes_output): Make global.
8270 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
8271 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
8272 (driver_handle_option): Handle OPT_fwpa.
8274 2014-02-05 Jakub Jelinek <jakub@redhat.com>
8277 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
8278 a comment typo and formatting issue. If odr_hash hasn't been
8279 created, return vNULL and set *completep to false.
8282 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
8283 bb with no successors.
8285 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
8288 * doc/invoke.texi (-march): Clarify documentation for ARM.
8292 2014-02-05 Richard Biener <rguenther@suse.de>
8294 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
8295 when not vectorizing because of too many alias checks.
8296 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
8297 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
8299 2014-02-05 Nick Clifton <nickc@redhat.com>
8301 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
8302 accept extended registers in any mode when compiling for the MN10300.
8304 2014-02-05 Yury Gribov <y.gribov@samsung.com>
8306 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
8307 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
8308 sanitization attributes.
8309 (can_inline_edge_p): Likewise.
8310 (sanitize_attrs_match_for_inline_p): New function.
8312 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
8314 * ipa-prop.c (detect_type_change): Shor circuit testing of
8315 type changes on THIS pointer.
8317 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
8320 * config/pa/pa.c (legitimize_tls_address): Return original address
8321 if not passed a SYMBOL_REF rtx.
8322 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
8324 (pa_emit_move_sequence): Simplify TLS source operands.
8325 (pa_legitimate_constant_p): Reject all TLS constants.
8326 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
8327 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
8329 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
8331 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
8332 groups when we know they are controlled by LTO.
8333 * varasm.c (default_binds_local_p_1): If object is in other partition,
8334 it will be resolved locally.
8336 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
8338 * config/host-linux.c (linux_gt_pch_use_address): Don't
8339 use SSIZE_MAX because it is not always defined.
8341 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
8344 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
8345 threshold for pseudo splitting.
8346 (update_ebb_live_info): Process call argument hard registers and
8347 hard registers from insn definition too.
8348 (max_small_class_regs_num): New constant.
8349 (inherit_in_ebb): Update live hard regs through EBBs. Update
8350 reloads_num only for small register classes. Don't split for
8353 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
8356 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
8359 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
8361 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
8364 2014-02-04 Marek Polacek <polacek@redhat.com>
8366 * gdbinit.in (pel): Define.
8368 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
8370 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
8373 2014-02-04 Richard Biener <rguenther@suse.de>
8376 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
8377 in function context local.
8378 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
8379 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
8380 similar to LTO_imported_decl_ref.
8382 2014-02-04 Jakub Jelinek <jakub@redhat.com>
8384 PR tree-optimization/60002
8385 * cgraphclones.c (build_function_decl_skip_args): Clear
8388 PR tree-optimization/60023
8389 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
8390 false to gsi_replace.
8391 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
8392 has been in some EH region and vec_stmt could throw, add
8393 vec_stmt into the same EH region.
8394 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
8395 has no lhs, ignore it.
8396 * internal-fn.c (expand_MASK_LOAD): Likewise.
8399 * tree-inline.c (copy_forbidden): Fail for
8400 __attribute__((optimize (0))) functions.
8403 * omp-low.c (simd_clone_struct_copy): If from->inbranch
8404 is set, copy one less argument.
8405 (expand_simd_clones): Don't subtract clone_info->inbranch
8406 from simd_clone_struct_alloc argument.
8408 PR rtl-optimization/57915
8409 * recog.c (simplify_while_replacing): If all unary/binary/relational
8410 operation arguments are constant, attempt to simplify those.
8413 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
8414 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
8416 2014-02-04 Richard Biener <rguenther@suse.de>
8418 PR tree-optimization/60012
8419 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
8420 TBAA disambiguation to all DDRs.
8422 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8425 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
8426 (LINK_SPEC): Use it for -shared, -shared-libgcc.
8428 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
8431 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
8433 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
8435 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
8436 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
8438 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
8441 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
8442 to figure out targets of polymorphic calls with known decl.
8443 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
8444 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
8445 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
8446 (get_polymorphic_call_info): ... here.
8447 (get_polymorphic_call_info_from_invariant): New function.
8449 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
8451 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
8452 lookup via vtable pointer; check for type consistency
8453 and turn inconsitent facts into UNREACHABLE.
8454 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
8455 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
8456 type inconsistent querries; return UNREACHABLE instead.
8458 2014-02-03 Richard Henderson <rth@twiddle.net>
8461 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
8462 already processed this node.
8463 (normalize_one_pred_1): Pass along mark_set.
8464 (normalize_one_pred): Create and destroy a pointer_set_t.
8465 (normalize_one_pred_chain): Likewise.
8467 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
8469 PR gcov-profile/58602
8470 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
8472 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
8475 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
8476 -fno-devirtualize; try to devirtualize by the knowledge of
8477 virtual table pointer given by aggregate propagation.
8478 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
8479 (ipa_print_node_jump_functions): Dump also offset that
8480 is relevant for polymorphic calls.
8481 (determine_known_aggregate_parts): Add arg_type parameter; use it
8482 instead of determining the type from pointer type.
8483 (ipa_compute_jump_functions_for_edge): Update call of
8484 determine_known_aggregate_parts.
8485 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
8486 (gimple_get_virt_method_for_binfo): ... here; simplify using
8487 vtable_pointer_value_to_vtable.
8488 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
8489 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
8490 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
8491 (vtable_pointer_value_to_vtable): Break out from ...; handle also
8493 (vtable_pointer_value_to_binfo): ... here.
8494 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
8496 2014-02-03 Teresa Johnson <tejohnson@google.com>
8498 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
8499 redef of outer loop index variable.
8501 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
8505 * doc/extend.texi (Function Attributes): Typo.
8507 2014-02-03 Cong Hou <congh@google.com>
8509 PR tree-optimization/60000
8510 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
8511 if the vectorized statement is a store. A store statement can only
8512 appear at the end of pattern statements.
8514 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
8516 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
8517 (ix86_option_override_internal): Default long double to 64-bit for
8518 32-bit Bionic and to 128-bit for 64-bit Bionic.
8520 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
8521 TARGET_LONG_DOUBLE_128 is true.
8522 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
8524 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
8525 (mlong-double-64): Negate -mlong-double-128.
8526 (mlong-double-128): New option.
8528 * config/i386/i386-c.c (ix86_target_macros): Define
8529 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
8531 * doc/invoke.texi: Document -mlong-double-128.
8533 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
8535 PR rtl-optimization/60024
8536 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
8538 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
8540 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
8542 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
8544 PR rtl-optimization/57662
8545 * sel-sched.c (code_motion_path_driver): Do not mark already not
8546 existing blocks in the visiting bitmap.
8548 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
8550 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
8551 on the insn being emitted.
8553 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
8554 Will Deacon <will.deacon@arm.com>
8556 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
8558 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8560 * config/arm/arm-tables.opt: Regenerate.
8562 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8564 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
8565 for vector types other than V16QImode.
8566 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
8567 define_expand, and call altivec_expand_vec_perm_le when producing
8568 code with little endian element order.
8569 (*altivec_vperm_<mode>_internal): New insn having previous
8570 behavior of altivec_vperm_<mode>.
8571 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
8572 altivec_expand_vec_perm_le when producing code with little endian
8574 (*altivec_vperm_<mode>_uns_internal): New insn having previous
8575 behavior of altivec_vperm_<mode>_uns.
8577 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8579 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
8580 (altivec_vsumsws): Add handling for -maltivec=be with a little
8582 (altivec_vsumsws_direct): New.
8583 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
8584 gen_altivec_vsumsws.
8586 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
8588 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
8589 vtable_pointer_value_to_binfo): New functions.
8590 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
8591 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
8593 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
8595 * config/nios2/nios2.md (load_got_register): Initialize GOT
8596 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
8597 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
8599 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
8601 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
8602 preserverd by passthrough, do not propagate the type.
8604 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
8606 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
8607 (mips_atomic_assign_expand_fenv): New function.
8608 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
8610 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
8612 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
8613 (__builtin_mips_set_fcsr): Likewise.
8614 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
8615 MIPS_USI_FTYPE_VOID.
8616 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
8617 (mips16_expand_set_fcsr): Likewise.
8618 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
8619 (mips16_set_fcsr_stub): Likewise.
8620 (mips16_get_fcsr_one_only_stub): New class.
8621 (mips16_set_fcsr_one_only_stub): Likewise.
8622 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
8623 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
8624 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
8625 (hard_float): New availability predicate.
8626 (mips_builtins): Add get_fcsr and set_fcsr.
8627 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
8628 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
8629 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
8630 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
8631 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
8634 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
8636 * config/mips/mips.c (mips_one_only_stub): New class.
8637 (mips_need_mips16_rdhwr_p): Replace with...
8638 (mips16_rdhwr_stub): ...this new variable.
8639 (mips16_stub_call_address): New function.
8640 (mips16_rdhwr_one_only_stub): New class.
8641 (mips_expand_thread_pointer): Use mips16_stub_call_address.
8642 (mips_output_mips16_rdhwr): Delete.
8643 (mips_finish_stub): New function.
8644 (mips_code_end): Use it to handle rdhwr stubs.
8646 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
8649 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
8650 when calculating size of integer atomic types.
8652 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
8654 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
8656 2014-02-01 Jakub Jelinek <jakub@redhat.com>
8658 PR tree-optimization/60003
8659 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
8660 * profile.c (branch_prob): Use gimple_call_builtin_p
8661 to check for BUILT_IN_SETJMP_RECEIVER.
8662 * tree-inline.c (copy_bb): Call notice_special_calls.
8664 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
8667 * lra-constraints.c (process_alt_operands): Update reload_sum only
8670 2014-01-31 Richard Henderson <rth@redhat.com>
8673 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
8674 until after else_eh is processed.
8676 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
8678 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
8679 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
8680 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
8681 in smmintrin.h, remove them.
8682 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
8683 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
8684 * config/i386/i386.md (ROUND_SAE): Fix value.
8685 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
8686 (const48_operand): New.
8687 * config/i386/subst.md (round), (round_expand): Use
8688 const_4_or_8_to_11_operand.
8689 (round_saeonly), (round_saeonly_expand): Use const48_operand.
8691 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
8693 * config/i386/constraints.md (Yk): Swap meaning with k.
8694 * config/i386/i386.md (movhi_internal): Change Yk to k.
8695 (movqi_internal): Ditto.
8696 (*k<logic><mode>): Ditto.
8699 (kandn<mode>): Ditto.
8700 (*<code>hi_1): Ditto.
8701 (*<code>qi_1): Ditto.
8702 (kxnor<mode>): Ditto.
8703 (kortestzhi): Ditto.
8704 (kortestchi): Ditto.
8706 (*one_cmplhi2_1): Ditto.
8707 (*one_cmplqi2_1): Ditto.
8708 * config/i386/sse.md (): Change k to Yk.
8709 (avx512f_load<mode>_mask): Ditto.
8710 (avx512f_blendm<mode>): Ditto.
8711 (avx512f_store<mode>_mask): Ditto.
8712 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
8713 (avx512f_storedqu<mode>_mask): Ditto.
8714 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
8716 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
8717 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
8718 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
8719 (avx512f_maskcmp<mode>3): Ditto.
8720 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
8721 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
8722 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
8723 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
8724 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
8725 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
8726 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
8727 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
8728 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
8729 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
8730 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
8731 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
8732 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
8733 (vec_extract_lo_<mode>_maskm): Ditto.
8734 (vec_extract_hi_<mode>_maskm): Ditto.
8735 (avx512f_vternlog<mode>_mask): Ditto.
8736 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
8737 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
8738 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
8739 (avx512f_<code>v8div16qi2_mask): Ditto.
8740 (avx512f_<code>v8div16qi2_mask_store): Ditto.
8741 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
8742 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
8743 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
8744 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
8745 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
8746 (*avx512pf_gatherpf<mode>df_mask): Ditto.
8747 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
8748 (*avx512pf_scatterpf<mode>df_mask): Ditto.
8749 (avx512cd_maskb_vec_dupv8di): Ditto.
8750 (avx512cd_maskw_vec_dupv16si): Ditto.
8751 (avx512f_vpermi2var<mode>3_maskz): Ditto.
8752 (avx512f_vpermi2var<mode>3_mask): Ditto.
8753 (avx512f_vpermi2var<mode>3_mask): Ditto.
8754 (avx512f_vpermt2var<mode>3_maskz): Ditto.
8755 (*avx512f_gathersi<mode>): Ditto.
8756 (*avx512f_gathersi<mode>_2): Ditto.
8757 (*avx512f_gatherdi<mode>): Ditto.
8758 (*avx512f_gatherdi<mode>_2): Ditto.
8759 (*avx512f_scattersi<mode>): Ditto.
8760 (*avx512f_scatterdi<mode>): Ditto.
8761 (avx512f_compress<mode>_mask): Ditto.
8762 (avx512f_compressstore<mode>_mask): Ditto.
8763 (avx512f_expand<mode>_mask): Ditto.
8764 * config/i386/subst.md (mask): Change k to Yk.
8765 (mask_scalar_merge): Ditto.
8768 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
8770 * doc/extend.texi (Vector Extensions): Document ?: in C++.
8772 2014-01-31 Richard Biener <rguenther@suse.de>
8775 * builtins.c (fold_builtin_memory_op): Make sure to not
8776 use a floating-point mode or a boolean or enumeral type for
8779 2014-01-30 DJ Delorie <dj@redhat.com>
8781 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
8782 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
8783 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
8784 whenever main() has an epilogue.
8786 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8788 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
8789 unused variable "field".
8790 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
8791 (vsx_mergeh_<mode>): Likewise.
8792 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
8793 (altivec_vmrghh): Likewise.
8794 (altivec_vmrghw): Likewise.
8795 (altivec_vmrglb): Likewise.
8796 (altivec_vmrglh): Likewise.
8797 (altivec_vmrglw): Likewise.
8798 (altivec_vspltb): Add missing uses.
8799 (altivec_vsplth): Likewise.
8800 (altivec_vspltw): Likewise.
8801 (altivec_vspltsf): Likewise.
8803 2014-01-30 Jakub Jelinek <jakub@redhat.com>
8806 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
8807 frame related instructions.
8809 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
8811 PR rtl-optimization/59959
8812 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
8813 any reload of register whose subreg is invalid.
8815 2014-01-30 Jakub Jelinek <jakub@redhat.com>
8817 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
8818 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
8819 Add missing return type - void.
8821 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8823 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
8824 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
8825 remove element index adjustment for endian (now handled in vsx.md
8827 (altivec_expand_vec_perm_const): Use
8828 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
8829 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
8830 (vsx_xxspltw_<mode>): Adjust element index for little endian.
8831 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
8832 define_expand and a new define_insn *altivec_vspltb_internal;
8833 adjust for -maltivec=be on a little endian target.
8834 (altivec_vspltb_direct): New.
8835 (altivec_vsplth): Divide into a define_expand and a new
8836 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
8837 little endian target.
8838 (altivec_vsplth_direct): New.
8839 (altivec_vspltw): Divide into a define_expand and a new
8840 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
8841 little endian target.
8842 (altivec_vspltw_direct): New.
8843 (altivec_vspltsf): Divide into a define_expand and a new
8844 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
8845 a little endian target.
8847 2014-01-30 Richard Biener <rguenther@suse.de>
8849 PR tree-optimization/59993
8850 * tree-ssa-forwprop.c (associate_pointerplus): Check we
8851 can propagate form the earlier stmt and avoid the transform
8852 when the intermediate result is needed.
8854 2014-01-30 Alangi Derick <alangiderick@gmail.com>
8856 * README.Portability: Fix typo.
8858 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
8860 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
8861 comparison_operator with ordered_comparison_operator.
8863 2014-01-30 Nick Clifton <nickc@redhat.com>
8865 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
8866 Rename to mn10300_store_multiple_regs.
8867 * config/mn10300/mn10300.c: Likewise.
8868 * config/mn10300/mn10300.md (store_movm): Fix typo: call
8869 store_multiple_regs.
8870 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
8871 Call mn10300_store_multiple_regs.
8873 2014-01-30 Nick Clifton <nickc@redhat.com>
8874 DJ Delorie <dj@redhat.com>
8876 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
8877 %fp 2 to keep registers after it properly word-aligned.
8878 (rl78_alloc_physical_registers_umul): Handle the case where both
8879 input operands are the same.
8881 2014-01-30 Richard Biener <rguenther@suse.de>
8883 PR tree-optimization/59903
8884 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
8887 2014-01-30 Jason Merrill <jason@redhat.com>
8890 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
8893 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
8895 2014-01-30 Richard Biener <rguenther@suse.de>
8897 PR tree-optimization/59951
8898 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
8900 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
8903 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
8904 SFmode to DFmode case.
8906 2014-01-29 DJ Delorie <dj@redhat.com>
8908 * config/msp430/msp430.opt (-minrt): New.
8909 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
8911 (ENDFILE_SPEC): Likewise.
8913 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
8915 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
8916 (estimate_function_body_sizes): Use it.
8918 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
8921 * dwarf2out.c (is_cxx_auto): New.
8922 (is_base_type): Use it.
8923 (gen_type_die_with_usage): Likewise.
8925 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8927 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
8928 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
8929 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
8930 -maltivec=be with LE targets.
8931 (vsx_mergeh_<mode>): Likewise.
8932 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
8933 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
8934 (altivec_vmrghb): Replace with define_expand and new
8935 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
8936 (altivec_vmrghb_direct): New define_insn.
8937 (altivec_vmrghh): Replace with define_expand and new
8938 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
8939 (altivec_vmrghh_direct): New define_insn.
8940 (altivec_vmrghw): Replace with define_expand and new
8941 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
8942 (altivec_vmrghw_direct): New define_insn.
8943 (*altivec_vmrghsf): Adjust for endianness.
8944 (altivec_vmrglb): Replace with define_expand and new
8945 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
8946 (altivec_vmrglb_direct): New define_insn.
8947 (altivec_vmrglh): Replace with define_expand and new
8948 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
8949 (altivec_vmrglh_direct): New define_insn.
8950 (altivec_vmrglw): Replace with define_expand and new
8951 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
8952 (altivec_vmrglw_direct): New define_insn.
8953 (*altivec_vmrglsf): Adjust for endianness.
8954 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
8955 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
8956 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
8957 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
8958 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
8959 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
8960 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
8961 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
8963 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
8965 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
8966 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
8969 2014-01-29 Richard Biener <rguenther@suse.de>
8971 PR tree-optimization/58742
8972 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
8973 associate_pointerplus_align.
8974 (associate_pointerplus_diff): New function.
8975 (associate_pointerplus): Likewise. Call associate_pointerplus_align
8976 and associate_pointerplus_diff.
8978 2014-01-29 Richard Biener <rguenther@suse.de>
8980 * lto-streamer.h (LTO_major_version): Bump to 3.
8981 (LTO_minor_version): Reset to 0.
8983 2014-01-29 Renlin Li <Renlin.Li@arm.com>
8985 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
8986 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
8987 (arm_file_start): Generate correct asm header for armv7ve.
8988 * config/arm/bpabi.h: Add multilib support for armv7ve.
8989 * config/arm/driver-arm.c: Change the architectures of cortex-a7
8990 and cortex-a15 to armv7ve.
8991 * config/arm/t-aprofile: Add multilib support for armv7ve.
8992 * doc/invoke.texi: Document -march=armv7ve.
8994 2014-01-29 Richard Biener <rguenther@suse.de>
8996 PR tree-optimization/58742
8997 * tree-ssa-forwprop.c (associate_plusminus): Return true
8998 if we changed sth, defer EH cleanup to ...
8999 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
9000 (simplify_mult): New function.
9002 2014-01-29 Jakub Jelinek <jakub@redhat.com>
9005 PR tree-optimization/59920
9006 * tree.c (build_common_builtin_nodes): Remove
9007 __builtin_setjmp_dispatcher initialization.
9008 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
9009 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
9010 instead of gsi_after_labels + manually skipping debug stmts.
9011 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
9012 ignore bbs with IFN_ABNORMAL_DISPATCHER.
9013 * tree-inline.c (copy_edges_for_bb): Remove
9014 can_make_abnormal_goto argument, instead add abnormal_goto_dest
9015 argument. Ignore computed_goto_p stmts. Don't call
9016 make_abnormal_goto_edges. If a call might need abnormal edges
9017 for non-local gotos, see if it already has an edge to
9018 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
9019 with true argument, don't do anything then, otherwise add
9020 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
9021 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
9023 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
9024 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
9025 (lower_stmt): Don't set data->calls_builtin_setjmp.
9026 (lower_builtin_setjmp): Adjust comment.
9027 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
9028 * tree-cfg.c (found_computed_goto): Remove.
9029 (factor_computed_gotos): Remove.
9030 (make_goto_expr_edges): Return bool, true for computed gotos.
9031 Don't call make_abnormal_goto_edges.
9032 (build_gimple_cfg): Don't set found_computed_goto, don't call
9033 factor_computed_gotos.
9034 (computed_goto_p): No longer static.
9035 (make_blocks): Don't set found_computed_goto.
9036 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
9037 (make_edges): If make_goto_expr_edges returns true, push bb
9038 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
9039 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
9040 vector. Record mapping between bbs and OpenMP regions if there
9041 are any, adjust make_gimple_omp_edges caller. Call
9042 handle_abnormal_edges.
9043 (make_abnormal_goto_edges): Remove.
9044 * tree-cfg.h (make_abnormal_goto_edges): Remove.
9045 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
9046 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
9047 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
9048 * internal-fn.def (ABNORMAL_DISPATCHER): New.
9049 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
9050 filling *region also set *region_idx to (*region)->entry->index.
9053 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
9054 For REGs set ORIGINAL_REGNO.
9056 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
9058 * doc/md.texi: Mention that a target shouldn't implement
9059 vec_widen_(s|u)mul_even/odd pair if it is less efficient
9062 2014-01-29 Jakub Jelinek <jakub@redhat.com>
9064 PR tree-optimization/59594
9065 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
9066 a copy of the datarefs vector rather than the vector itself.
9068 2014-01-28 Jason Merrill <jason@redhat.com>
9071 * dwarf2out.c (auto_die): New static.
9072 (gen_type_die_with_usage): Handle C++1y 'auto'.
9073 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
9076 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
9079 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
9080 (SPEC_X32): Likewise.
9081 (SPEC_64): Likewise.
9082 * config/i386/i386.c (ix86_option_override_internal): Turn off
9083 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
9085 (x86_file_start): Output .code16gcc for TARGET_16BIT.
9086 * config/i386/i386.h (TARGET_16BIT): New macro.
9087 (TARGET_16BIT_P): Likewise.
9088 * config/i386/i386.opt: Add m16.
9089 * doc/invoke.texi: Document -m16.
9091 2014-01-28 Jakub Jelinek <jakub@redhat.com>
9093 PR preprocessor/59935
9094 * input.c (location_get_source_line): Bail out on when line number
9095 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
9097 2014-01-28 Richard Biener <rguenther@suse.de>
9099 PR tree-optimization/58742
9100 * tree-ssa-forwprop.c (associate_plusminus): Handle
9101 pointer subtraction of the form (T)(P + A) - (T)P.
9103 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9105 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
9108 2014-01-28 Richard Biener <rguenther@suse.de>
9111 2014-01-28 Richard Biener <rguenther@suse.de>
9113 PR rtl-optimization/45364
9114 PR rtl-optimization/59890
9115 * var-tracking.c (local_get_addr_clear_given_value): Handle
9116 already cleared slot.
9117 (val_reset): Handle not allocated local_get_addr_cache.
9118 (vt_find_locations): Use post-order on the inverted CFG.
9120 2014-01-28 Richard Biener <rguenther@suse.de>
9122 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
9124 2014-01-28 Richard Biener <rguenther@suse.de>
9126 PR rtl-optimization/45364
9127 PR rtl-optimization/59890
9128 * var-tracking.c (local_get_addr_clear_given_value): Handle
9129 already cleared slot.
9130 (val_reset): Handle not allocated local_get_addr_cache.
9131 (vt_find_locations): Use post-order on the inverted CFG.
9133 2014-01-28 Alan Modra <amodra@gmail.com>
9135 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
9136 * configure.ac <recursive call for build != host>: Define
9137 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
9138 and LD_FOR_BUILD too.
9139 * configure: Regenerate.
9141 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
9143 * config/i386/i386.c (get_builtin_code_for_version): Separate
9144 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
9145 Broadwell from Haswell.
9147 2014-01-27 Steve Ellcey <sellcey@mips.com>
9149 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
9150 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
9151 * config/mips/mips.c (mips_option_override): Change setting
9153 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
9154 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
9155 Change from Mask to Var.
9157 2014-01-27 Jeff Law <law@redhat.com>
9159 * ipa-inline.c (inline_small_functions): Fix typo.
9161 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
9163 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
9164 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
9165 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
9166 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
9167 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
9168 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
9169 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
9170 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
9171 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
9172 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
9173 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
9174 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
9175 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
9176 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
9177 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
9178 (_mm512_storeu_epi64): Ditto.
9179 (_mm512_cmpge_epi32_mask): Ditto.
9180 (_mm512_cmpge_epu32_mask): Ditto.
9181 (_mm512_cmpge_epi64_mask): Ditto.
9182 (_mm512_cmpge_epu64_mask): Ditto.
9183 (_mm512_cmple_epi32_mask): Ditto.
9184 (_mm512_cmple_epu32_mask): Ditto.
9185 (_mm512_cmple_epi64_mask): Ditto.
9186 (_mm512_cmple_epu64_mask): Ditto.
9187 (_mm512_cmplt_epi32_mask): Ditto.
9188 (_mm512_cmplt_epu32_mask): Ditto.
9189 (_mm512_cmplt_epi64_mask): Ditto.
9190 (_mm512_cmplt_epu64_mask): Ditto.
9191 (_mm512_cmpneq_epi32_mask): Ditto.
9192 (_mm512_cmpneq_epu32_mask): Ditto.
9193 (_mm512_cmpneq_epi64_mask): Ditto.
9194 (_mm512_cmpneq_epu64_mask): Ditto.
9195 (_mm512_expand_pd): Ditto.
9196 (_mm512_expand_ps): Ditto.
9197 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
9198 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
9199 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
9200 * config/i386/i386.c (ix86_builtins): Add
9201 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
9202 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
9203 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
9204 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
9205 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
9206 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
9207 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
9208 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
9209 IX86_BUILTIN_PMOVUSQW512_MEM.
9210 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
9211 __builtin_ia32_pmovsqd512mem_mask,
9212 __builtin_ia32_pmovqd512mem_mask,
9213 __builtin_ia32_pmovusqw512mem_mask,
9214 __builtin_ia32_pmovsqw512mem_mask,
9215 __builtin_ia32_pmovqw512mem_mask,
9216 __builtin_ia32_pmovusdw512mem_mask,
9217 __builtin_ia32_pmovsdw512mem_mask,
9218 __builtin_ia32_pmovdw512mem_mask,
9219 __builtin_ia32_pmovqb512mem_mask,
9220 __builtin_ia32_pmovusqb512mem_mask,
9221 __builtin_ia32_pmovsqb512mem_mask,
9222 __builtin_ia32_pmovusdb512mem_mask,
9223 __builtin_ia32_pmovsdb512mem_mask,
9224 __builtin_ia32_pmovdb512mem_mask.
9225 (bdesc_args): Add __builtin_ia32_expanddf512,
9226 __builtin_ia32_expandsf512.
9227 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
9228 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
9229 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
9230 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
9231 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
9232 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
9233 (avx512f_<code>v8div16qi2_mask_store): This.
9234 (avx512f_expand<mode>): New.
9236 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
9238 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
9240 (_mm512_mask_prefetch_i64gather_pd): Ditto.
9241 (_mm512_prefetch_i32scatter_pd): Ditto.
9242 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
9243 (_mm512_prefetch_i64scatter_pd): Ditto.
9244 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
9245 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
9246 (_mm512_mask_prefetch_i64gather_ps): Ditto.
9247 (_mm512_prefetch_i32scatter_ps): Ditto.
9248 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
9249 (_mm512_prefetch_i64scatter_ps): Ditto.
9250 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
9251 * config/i386/i386-builtin-types.def: Define
9252 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
9253 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
9254 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
9255 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
9256 IX86_BUILTIN_SCATTERPFQPD.
9257 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
9258 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
9259 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
9260 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
9261 __builtin_ia32_scatterpfqps.
9262 (ix86_expand_builtin): Expand new built-ins.
9263 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
9264 fix memory access data type.
9265 (*avx512pf_gatherpf<mode>_mask): Ditto.
9266 (*avx512pf_gatherpf<mode>): Ditto.
9267 (avx512pf_scatterpf<mode>): Ditto.
9268 (*avx512pf_scatterpf<mode>_mask): Ditto.
9269 (*avx512pf_scatterpf<mode>): Ditto.
9270 (GATHER_SCATTER_SF_MEM_MODE): New.
9271 (avx512pf_gatherpf<mode>df): Ditto.
9272 (*avx512pf_gatherpf<mode>df_mask): Ditto.
9273 (*avx512pf_scatterpf<mode>df): Ditto.
9275 2014-01-27 Jakub Jelinek <jakub@redhat.com>
9278 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
9279 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
9282 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
9284 * common/config/arm/arm-common.c
9285 (arm_rewrite_mcpu): Handle multiple names.
9287 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
9289 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
9291 * gimple-builder.h (create_gimple_tmp): Delete.
9293 2014-01-27 Christian Bruel <christian.bruel@st.com>
9295 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
9298 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
9300 * config/pa/pa.md (call): Generate indirect long calls to non-local
9301 functions when outputing 32-bit code.
9302 (call_value): Likewise except for special call to buggy powf function.
9304 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
9305 portable runtime and PIC indirect calls.
9306 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
9307 and PIC call sequences. Use ldo instead of blr to set return register
9308 in PIC call sequence.
9310 2014-01-25 Walter Lee <walt@tilera.com>
9312 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
9313 avoid clobbering a live register.
9315 2014-01-25 Walter Lee <walt@tilera.com>
9317 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
9318 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
9319 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
9320 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
9322 2014-01-25 Walter Lee <walt@tilera.com>
9324 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
9325 arguments on even registers.
9326 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
9328 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
9329 (BIGGEST_ALIGNMENT): Ditto.
9330 (BIGGEST_FIELD_ALIGNMENT): Ditto.
9332 2014-01-25 Walter Lee <walt@tilera.com>
9334 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
9335 insns before bundling.
9336 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
9338 2014-01-25 Walter Lee <walt@tilera.com>
9340 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
9341 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
9342 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
9344 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
9346 * config/mips/constraints.md (kl): Delete.
9347 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
9348 define expands, using...
9349 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
9350 instructions for MIPS16.
9351 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
9352 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
9354 2014-01-25 Walter Lee <walt@tilera.com>
9356 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
9360 2014-01-25 Walter Lee <walt@tilera.com>
9362 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
9363 (TARGET_EXPAND_TO_RTL_HOOK): Define.
9365 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
9367 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
9370 2014-01-25 Jakub Jelinek <jakub@redhat.com>
9372 * print-rtl.c (in_call_function_usage): New var.
9373 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
9374 EXPR_LIST mode as mode and not as reg note name.
9377 * cfgloopmanip.c (copy_loop_info): If
9378 loop->warned_aggressive_loop_optimizations, make sure
9379 the flag is set in target loop too.
9381 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
9383 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
9385 * builtins.def: Likewise.
9386 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
9387 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
9388 * ira.c (ira_setup_eliminable_regset): Likewise.
9389 * omp-low.c (gate_expand_omp): Likewise.
9390 (execute_lower_omp): Likewise.
9391 (diagnose_sb_0): Likewise.
9392 (gate_diagnose_omp_blocks): Likewise.
9393 (simd_clone_clauses_extract): Likewise.
9396 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9398 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
9399 correction for little endian...
9400 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
9403 2014-01-24 Jeff Law <law@redhat.com>
9405 PR tree-optimization/59919
9406 * tree-vrp.c (find_assert_locations_1): Do not register asserts
9407 for non-returning calls.
9409 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
9411 * common/config/aarch64/aarch64-common.c
9412 (aarch64_rewrite_mcpu): Handle multiple names.
9413 * config/aarch64/aarch64.h
9414 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
9416 2014-01-24 Dodji Seketeli <dodji@redhat.com>
9418 * input.c (add_file_to_cache_tab): Handle the case where fopen
9421 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
9424 * config/i386/i386.md (pushsf splitter): Get stack adjustment
9425 from push operand if code of push isn't PRE_DEC.
9427 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
9430 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
9431 -mquad-memory-atomic. Update -mquad-memory documentation to say
9432 it is only used for non-atomic loads/stores.
9434 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
9435 -mquad-memory or -mquad-memory-atomic switches.
9437 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
9438 -mquad-memory-atomic to ISA 2.07 support.
9440 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
9441 to separate support of normal quad word memory operations (ldq, stq)
9442 from the atomic quad word memory operations.
9444 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
9445 support to separate non-atomic quad word operations from atomic
9446 quad word operations. Disable non-atomic quad word operations in
9447 little endian mode so that we don't have to swap words after the
9448 load and before the store.
9449 (quad_load_store_p): Add comment about atomic quad word support.
9450 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
9451 options printed with -mdebug=reg.
9453 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
9454 -mquad-memory-atomic as the test for whether we have quad word
9455 atomic instructions.
9456 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
9457 or -mp8-vector are used, allow byte/half-word atomic operations.
9459 * config/rs6000/sync.md (load_lockedti): Insure that the address
9460 is a proper indexed or indirect address for the lqarx instruction.
9461 On little endian systems, swap the hi/lo registers after the lqarx
9463 (load_lockedpti): Use indexed_or_indirect_operand predicate to
9464 insure the address is valid for the lqarx instruction.
9465 (store_conditionalti): Insure that the address is a proper indexed
9466 or indirect address for the stqcrx. instruction. On little endian
9467 systems, swap the hi/lo registers before doing the stqcrx.
9469 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
9470 insure the address is valid for the stqcrx. instruction.
9472 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
9473 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
9474 type of quad memory support is available.
9476 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
9479 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
9480 there is a danger of looping.
9482 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
9484 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
9485 force flag_ira_loop_pressure if set via command line.
9487 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
9489 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
9490 (ashr_simd): New builtin handling DI mode.
9491 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
9492 (aarch64_sshr_simddi): New match pattern.
9493 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
9494 (vshrd_n_s64): Likewise.
9495 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
9497 2014-01-23 Nick Clifton <nickc@redhat.com>
9499 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
9500 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
9501 favour of mcu specific scripts.
9502 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
9505 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
9506 Alex Velenko <Alex.Velenko@arm.com>
9508 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
9509 (vaddv_s16): Likewise.
9510 (vaddv_s32): Likewise.
9511 (vaddv_u8): Likewise.
9512 (vaddv_u16): Likewise.
9513 (vaddv_u32): Likewise.
9514 (vaddvq_s8): Likewise.
9515 (vaddvq_s16): Likewise.
9516 (vaddvq_s32): Likewise.
9517 (vaddvq_s64): Likewise.
9518 (vaddvq_u8): Likewise.
9519 (vaddvq_u16): Likewise.
9520 (vaddvq_u32): Likewise.
9521 (vaddvq_u64): Likewise.
9522 (vaddv_f32): Likewise.
9523 (vaddvq_f32): Likewise.
9524 (vaddvq_f64): Likewise.
9525 (vmaxv_f32): Likewise.
9526 (vmaxv_s8): Likewise.
9527 (vmaxv_s16): Likewise.
9528 (vmaxv_s32): Likewise.
9529 (vmaxv_u8): Likewise.
9530 (vmaxv_u16): Likewise.
9531 (vmaxv_u32): Likewise.
9532 (vmaxvq_f32): Likewise.
9533 (vmaxvq_f64): Likewise.
9534 (vmaxvq_s8): Likewise.
9535 (vmaxvq_s16): Likewise.
9536 (vmaxvq_s32): Likewise.
9537 (vmaxvq_u8): Likewise.
9538 (vmaxvq_u16): Likewise.
9539 (vmaxvq_u32): Likewise.
9540 (vmaxnmv_f32): Likewise.
9541 (vmaxnmvq_f32): Likewise.
9542 (vmaxnmvq_f64): Likewise.
9543 (vminv_f32): Likewise.
9544 (vminv_s8): Likewise.
9545 (vminv_s16): Likewise.
9546 (vminv_s32): Likewise.
9547 (vminv_u8): Likewise.
9548 (vminv_u16): Likewise.
9549 (vminv_u32): Likewise.
9550 (vminvq_f32): Likewise.
9551 (vminvq_f64): Likewise.
9552 (vminvq_s8): Likewise.
9553 (vminvq_s16): Likewise.
9554 (vminvq_s32): Likewise.
9555 (vminvq_u8): Likewise.
9556 (vminvq_u16): Likewise.
9557 (vminvq_u32): Likewise.
9558 (vminnmv_f32): Likewise.
9559 (vminnmvq_f32): Likewise.
9560 (vminnmvq_f64): Likewise.
9562 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
9564 * config/aarch64/aarch64-simd.md
9565 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
9566 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
9567 (*aarch64_mul3_elt<mode>): Likewise.
9568 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
9569 (*aarch64_mul3_elt_to_64v2df): Likewise.
9570 (*aarch64_mla_elt<mode>): Likewise.
9571 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
9572 (*aarch64_mls_elt<mode>): Likewise.
9573 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
9574 (*aarch64_fma4_elt<mode>): Likewise.
9575 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
9576 (*aarch64_fma4_elt_to_64v2df): Likewise.
9577 (*aarch64_fnma4_elt<mode>): Likewise.
9578 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
9579 (*aarch64_fnma4_elt_to_64v2df): Likewise.
9580 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
9581 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
9582 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
9583 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
9584 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
9585 (aarch64_sqdmull_lane<mode>_internal): Likewise.
9586 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
9588 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
9590 * config/aarch64/aarch64-simd.md
9591 (aarch64_be_checked_get_lane<mode>): New define_expand.
9592 * config/aarch64/aarch64-simd-builtins.def
9593 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
9594 New builtin definition.
9595 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
9596 Use new safe be builtin.
9598 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
9600 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
9602 (aarch64_be_st1<mode>): Likewise.
9603 (aarch_ld1<VALL:mode>): Define_expand modified.
9604 (aarch_st1<VALL:mode>): Likewise.
9605 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
9606 (UNSPEC_ST1): Likewise.
9608 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
9610 * config/microblaze/microblaze.md: Add trap insn and attribute
9612 2014-01-23 Dodji Seketeli <dodji@redhat.com>
9614 PR preprocessor/58580
9615 * input.h (location_get_source_line): Take an additional line_size
9617 (void diagnostics_file_cache_fini): Declare new function.
9618 * input.c (struct fcache): New type.
9619 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
9620 New static constants.
9621 (diagnostic_file_cache_init, total_lines_num)
9622 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
9623 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
9624 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
9625 (get_next_line, read_next_line, goto_next_line, read_line_num):
9626 New static function definitions.
9627 (diagnostic_file_cache_fini): New function.
9628 (location_get_source_line): Take an additional output line_len
9629 parameter. Re-write using lookup_or_add_file_to_cache_tab and
9631 * diagnostic.c (diagnostic_finish): Call
9632 diagnostic_file_cache_fini.
9633 (adjust_line): Take an additional input parameter for the length
9634 of the line, rather than calculating it with strlen.
9635 (diagnostic_show_locus): Adjust the use of
9636 location_get_source_line and adjust_line with respect to their new
9637 signature. While displaying a line now, do not stop at the first
9638 null byte. Rather, display the zero byte as a space and keep
9639 going until we reach the size of the line.
9640 * Makefile.in: Add vec.o to OBJS-libcommon
9642 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
9643 Ilya Tocar <ilya.tocar@intel.com>
9645 * config/i386/avx512fintrin.h (_mm512_kmov): New.
9646 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
9647 (__builtin_ia32_kmov16): Ditto.
9648 * config/i386/i386.md (UNSPEC_KMOV): New.
9651 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
9653 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
9654 (_mm512_storeu_si512): Ditto.
9656 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
9659 * rtl.h (get_referenced_operands): Declare.
9660 * recog.c (get_referenced_operands): New function.
9661 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
9662 operands have been referenced when recording LO_SUM references.
9664 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
9666 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
9668 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
9670 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
9671 Enable for generic and recent AMD targets.
9673 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
9675 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
9676 ARG_SIZE note when adjustment was eliminated.
9678 2014-01-22 Jeff Law <law@redhat.com>
9680 PR tree-optimization/59597
9681 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
9682 in file. Accept new argument REGISTERING and use it to modify
9683 dump output appropriately.
9684 (register_jump_thread): Corresponding changes.
9685 (mark_threaded_blocks): Reinstate code to cancel unprofitable
9686 thread paths involving joiner blocks. Add code to dump cancelled
9687 jump threading paths.
9689 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
9691 PR rtl-optimization/59477
9692 * lra-constraints.c (inherit_in_ebb): Process call for living hard
9693 regs. Update reloads_num and potential_reload_hard_regs for all insns.
9695 2014-01-22 Tom Tromey <tromey@redhat.com>
9697 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
9699 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
9701 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
9703 PR rtl-optimization/59896
9704 * lra-constraints.c (process_alt_operands): Check unused note for
9705 matched operands of insn with no output reloads.
9707 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
9709 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
9710 (mips_move_from_gpr_cost): Likewise.
9712 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
9714 PR rtl-optimization/59858
9715 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
9716 ira_class_hard_regs_num.
9717 (process_alt_operands): Increase reject for dying matched operand.
9719 2014-01-21 Jakub Jelinek <jakub@redhat.com>
9722 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
9723 smaller than size, perform several stores or loads and stores
9724 at dst + count - size to store or copy all of size bytes, rather
9725 than just last modesize bytes.
9727 2014-01-20 DJ Delorie <dj@redhat.com>
9729 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
9730 that CLOBBERs are REGs before propogating their values.
9732 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
9735 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
9736 (cgraph_inline_failed_type): New function.
9737 * cgraph.h (DEFCIFCODE): Add type.
9738 (cgraph_inline_failed_type_t): New enum.
9739 (cgraph_inline_failed_type): New prototype.
9740 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
9741 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
9742 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
9743 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
9744 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
9745 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
9746 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
9747 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
9748 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
9749 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
9750 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
9751 OPTIMIZATION_MISMATCH.
9752 * tree-inline.c (expand_call_inline): Emit errors during
9753 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
9755 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
9758 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
9759 mode attribute in insn output.
9761 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
9763 * output.h (output_constant): Delete.
9764 * varasm.c (output_constant): Make private.
9766 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
9768 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
9770 2014-01-20 Jakub Jelinek <jakub@redhat.com>
9773 * tree.h (fold_builtin_strcat): New prototype.
9774 * builtins.c (fold_builtin_strcat): No longer static. Add len
9775 argument, if non-NULL, don't call c_strlen. Optimize
9776 directly into __builtin_memcpy instead of __builtin_strcpy.
9777 (fold_builtin_2): Adjust fold_builtin_strcat caller.
9778 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
9780 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
9782 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
9783 for SImode_address_operand operands, having only a REG argument.
9785 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
9787 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
9788 loader name using mbig-endian.
9789 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
9791 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
9793 * doc/invoke.texi (-march): Clarify documentation for AArch64.
9797 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
9799 * config/aarch64/aarch64-protos.h
9800 (aarch64_cannot_change_mode_class_ptr): Declare.
9801 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
9802 aarch64_cannot_change_mode_class_ptr): New.
9803 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
9804 backend hook aarch64_cannot_change_mode_class.
9806 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
9808 * common/config/aarch64/aarch64-common.c
9809 (aarch64_handle_option): Don't handle any option order logic here.
9810 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
9811 selected_cpu, warn on architecture version mismatch.
9812 (aarch64_override_options): Fix parsing order for option strings.
9814 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9815 Iain Sandoe <iain@codesourcery.com>
9818 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
9819 warning. Amend comment to reflect current functionality.
9821 2014-01-20 Richard Biener <rguenther@suse.de>
9824 * builtins.c (fold_builtin_strcat): Remove case better handled
9825 by tree-ssa-strlen.c.
9827 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
9829 * config/aarch64/aarch64.opt
9830 (mcpu, march, mtune): Make case-insensitive.
9832 2014-01-20 Jakub Jelinek <jakub@redhat.com>
9835 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
9836 if operands[1] is a REG or ZERO_EXTEND of a REG.
9838 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
9840 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
9842 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
9844 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
9845 long non-pic millicode calls.
9847 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9849 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
9851 2014-01-19 Kito Cheng <kito@0xlab.org>
9853 * builtins.c (expand_movstr): Check movstr expand done or fail.
9855 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
9856 H.J. Lu <hongjiu.lu@intel.com>
9859 * config/i386/i386.md (*lea<mode>): Zero-extend return register
9860 to DImode for zero-extended addresses.
9862 2014-01-19 Jakub Jelinek <jakub@redhat.com>
9864 PR rtl-optimization/57763
9865 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
9866 on the new indirect jump_insn and increment LABEL_NUSES (label).
9868 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
9872 * config.gcc (x86_archs): New variable.
9873 (x86_64_archs): Likewise.
9874 (x86_cpus): Likewise.
9875 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
9876 --with-arch/--with-cpu= options.
9877 Support --with-arch=/--with-cpu={nehalem,westmere,
9878 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
9880 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
9882 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
9883 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
9885 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
9887 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
9889 2014-01-18 Jakub Jelinek <jakub@redhat.com>
9892 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
9893 clear cpp_get_options (parse_in)->warn_unused_macros for
9894 ix86_target_macros_internal with cpp_define.
9896 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
9898 * jump.c (delete_related_insns): Keep (use (insn))s.
9899 * reorg.c (redundant_insn): Check for barriers too.
9901 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
9903 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
9905 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
9907 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
9908 call to $$dyncall when TARGET_LONG_CALLS is true.
9910 2014-01-17 Jeff Law <law@redhat.com>
9912 * ree.c (combine_set_extension): Temporarily disable test for
9913 changing number of hard registers.
9915 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
9918 * ipa-inline-analysis.c (inline_free_summary):
9919 Do not free summary of aliases.
9921 2014-01-17 Jakub Jelinek <jakub@redhat.com>
9924 * gimplify.c (gimplify_expr): Use create_tmp_var
9925 instead of create_tmp_var_raw. If cond doesn't have
9926 integral type, don't add the IFN_ANNOTATE builtin at all.
9928 2014-01-17 Martin Jambor <mjambor@suse.cz>
9931 * ipa-cp.c (prev_edge_clone): New variable.
9932 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
9933 Also resize prev_edge_clone vector.
9934 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
9935 (ipcp_edge_removal_hook): New function.
9936 (ipcp_driver): Register ipcp_edge_removal_hook.
9938 2014-01-17 Andrew Pinski <apinski@cavium.com>
9939 Steve Ellcey <sellcey@mips.com>
9942 * config/mips/mips.c (mips_print_operand): Check operand mode instead
9945 2014-01-17 Jeff Law <law@redhat.com>
9948 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
9949 so that pass_ccp runs first.
9951 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
9953 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
9954 (ix86_adjust_cost): Use !TARGET_XXX.
9955 (do_reorder_for_imul): Likewise.
9956 (swap_top_of_ready_list): Likewise.
9957 (ix86_sched_reorder): Likewise.
9959 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
9961 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
9962 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
9963 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
9964 (intel_memset): New. Duplicate slm_memset.
9965 (intel_cost): New. Duplicate slm_cost.
9966 (m_INTEL): New macro.
9967 (processor_target_table): Add "intel".
9968 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
9969 with PROCESSOR_INTEL for "intel".
9970 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
9971 PROCESSOR_SILVERMONT.
9972 (ix86_issue_rate): Likewise.
9973 (ix86_adjust_cost): Likewise.
9974 (ia32_multipass_dfa_lookahead): Likewise.
9975 (swap_top_of_ready_list): Likewise.
9976 (ix86_sched_reorder): Likewise.
9977 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
9978 instead of TARGET_OPT_AGU.
9979 * config/i386/i386.h (TARGET_INTEL): New.
9980 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
9981 (processor_type): Add PROCESSOR_INTEL.
9982 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
9983 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
9985 2014-01-17 Marek Polacek <polacek@redhat.com>
9988 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
9991 2014-01-17 Richard Biener <rguenther@suse.de>
9993 PR tree-optimization/46590
9994 * opts.c (default_options_table): Add entries for
9995 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
9996 all enabled at -O1 but not for -Og.
9997 * common.opt (fbranch-count-reg): Remove Init(1).
9998 (fmove-loop-invariants): Likewise.
9999 (ftree-pta): Likewise.
10001 2014-01-17 Jakub Jelinek <jakub@redhat.com>
10003 * config/i386/i386.c (ix86_data_alignment): For compatibility with
10004 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
10005 decls to at least the GCC 4.8 used alignments.
10008 * tree-nested.c (convert_nonlocal_reference_stmt,
10009 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
10010 of GIMPLE_BIND stmts, adjust associated decls.
10012 2014-01-17 Richard Biener <rguenther@suse.de>
10014 PR tree-optimization/46590
10015 * vec.h (vec<>::bseach): New member function implementing
10016 binary search according to C89 bsearch.
10017 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
10018 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
10019 bitmap pointer again. Make accesses_in_loop a flat array.
10020 (mem_ref_obstack): New global.
10021 (outermost_indep_loop): Adjust for mem_ref->stored changes.
10022 (mark_ref_stored): Likewise.
10023 (ref_indep_loop_p_2): Likewise.
10024 (set_ref_stored_in_loop): New helper function.
10025 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
10026 (memref_free): Adjust.
10027 (record_mem_ref_loc): Simplify.
10028 (gather_mem_refs_stmt): Adjust.
10029 (sort_locs_in_loop_postorder_cmp): New function.
10030 (analyze_memory_references): Sort accesses_in_loop after
10031 loop postorder number.
10032 (find_ref_loc_in_loop_cmp): New function.
10033 (for_all_locs_in_loop): Find relevant cluster of locs in
10034 accesses_in_loop and iterate without recursion.
10035 (execute_sm): Avoid uninit warning.
10036 (struct ref_always_accessed): Simplify.
10037 (ref_always_accessed::operator ()): Likewise.
10038 (ref_always_accessed_p): Likewise.
10039 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
10040 loop postorder numbers here.
10041 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
10044 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
10047 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
10048 on decls for which assemble_alias has been called.
10050 2014-01-17 Nick Clifton <nickc@redhat.com>
10052 * config/msp430/msp430.opt: (mcpu): New option.
10053 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
10054 (msp430_option_override): Parse target_cpu. If the MCU name
10055 matches a generic string, clear target_mcu.
10056 (msp430_attr): Allow numeric interrupt values up to 63.
10057 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
10058 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
10060 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
10062 * config/msp430/msp430.md (popm): Use %J rather than %I.
10063 (addsi3): Use msp430_nonimmediate_operand for operand 2.
10064 (addhi_cy_i): Use immediate_operand for operand 2.
10065 * doc/invoke.texi: Document -mcpu option.
10067 2014-01-17 Richard Biener <rguenther@suse.de>
10069 PR rtl-optimization/38518
10070 * df.h (df_analyze_loop): Declare.
10071 * df-core.c: Include cfgloop.h.
10072 (df_analyze_1): Split out main part of df_analyze.
10073 (df_analyze): Adjust.
10074 (loop_inverted_post_order_compute): New function.
10075 (loop_post_order_compute): Likewise.
10076 (df_analyze_loop): New function avoiding whole-function
10077 postorder computes.
10078 * loop-invariant.c (find_defs): Use df_analyze_loop.
10079 (find_invariants): Adjust.
10080 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
10082 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
10084 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
10085 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
10087 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
10089 * ipa-ref.c (ipa_remove_stmt_references): Fix references
10090 traversal when removing references.
10092 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
10095 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
10097 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
10099 PR middle-end/56791
10100 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
10101 pushing a reload for an autoinc when we had previously reloaded an
10102 inner part of the address.
10104 2014-01-16 Jakub Jelinek <jakub@redhat.com>
10106 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
10108 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
10109 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
10110 when not giving up or versioning for alias only because of
10112 (vect_analyze_data_ref_dependences): Set to true.
10113 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
10115 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
10116 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
10119 PR middle-end/58344
10120 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
10123 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
10124 operand 0 predicate for gathers, use a new pseudo as subtarget.
10126 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
10128 PR middle-end/59609
10129 * lra-constraints.c (process_alt_operands): Add printing debug info.
10130 Check absence of input/output reloads for matched operands too.
10132 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
10134 PR rtl-optimization/59835
10135 * ira.c (ira_init_register_move_cost): Increase cost for
10138 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
10140 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
10142 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
10145 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
10146 non-register objects. Use gen_(high/low)part more consistently.
10149 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
10152 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
10153 endian support, remove tests for WORDS_BIG_ENDIAN.
10154 (p8_mfvsrd_3_<mode>): Likewise.
10155 (reload_gpr_from_vsx<mode>): Likewise.
10156 (reload_gpr_from_vsxsf): Likewise.
10157 (p8_mfvsrd_4_disf): Likewise.
10159 2014-01-16 Richard Biener <rguenther@suse.de>
10161 PR rtl-optimization/46590
10162 * lcm.c (compute_antinout_edge): Use postorder iteration.
10163 (compute_laterin): Use inverted postorder iteration.
10165 2014-01-16 Nick Clifton <nickc@redhat.com>
10167 PR middle-end/28865
10168 * varasm.c (output_constant): Return the number of bytes actually
10170 (output_constructor_array_range): Update the field size with the
10171 number of bytes emitted by output_constant.
10172 (output_constructor_regular_field): Likewise. Also do not
10173 complain if the total number of bytes emitted is now greater
10174 than the expected fieldpos.
10175 * output.h (output_constant): Update prototype and descriptive comment.
10177 2014-01-16 Marek Polacek <polacek@redhat.com>
10179 PR middle-end/59827
10180 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
10181 it is error_mark_node.
10183 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
10185 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
10186 VALID_AVX256_REG_OR_OI_MODE.
10188 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
10190 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
10191 current procedure should be profiled.
10193 2014-01-15 Andrew Pinski <apinski@cavium.com>
10195 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
10196 of moving from/to the STACK_REG register class.
10198 2014-01-15 Richard Henderson <rth@redhat.com>
10201 * reginfo.c (global_regs_decl): Globalize.
10202 * rtl.h (global_regs_decl): Declare.
10203 * ira.c (do_reload): Diagnose frame_pointer_needed and it
10204 reserved via global_regs.
10206 2014-01-15 Teresa Johnson <tejohnson@google.com>
10208 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
10210 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
10212 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
10213 and vmulosh rather than call gen_vec_widen_smult_*.
10214 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
10215 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
10216 (vec_widen_smult_even_v16qi): Likewise.
10217 (vec_widen_umult_even_v8hi): Likewise.
10218 (vec_widen_smult_even_v8hi): Likewise.
10219 (vec_widen_umult_odd_v16qi): Likewise.
10220 (vec_widen_smult_odd_v16qi): Likewise.
10221 (vec_widen_umult_odd_v8hi): Likewise.
10222 (vec_widen_smult_odd_v8hi): Likewise.
10223 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
10224 vmuloub rather than call gen_vec_widen_umult_*.
10225 (vec_widen_umult_lo_v16qi): Likewise.
10226 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
10227 vmulosb rather than call gen_vec_widen_smult_*.
10228 (vec_widen_smult_lo_v16qi): Likewise.
10229 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
10230 rather than call gen_vec_widen_umult_*.
10231 (vec_widen_umult_lo_v8hi): Likewise.
10232 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
10233 rather than call gen_vec_widen_smult_*.
10234 (vec_widen_smult_lo_v8hi): Likewise.
10236 2014-01-15 Jeff Law <law@redhat.com>
10238 PR tree-optimization/59747
10239 * ree.c (find_and_remove_re): Properly handle case where a second
10240 eliminated extension requires widening a copy created for elimination
10241 of a prior extension.
10242 (combine_set_extension): Ensure that the number of hard regs needed
10243 for a destination register does not change when we widen it.
10245 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
10247 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
10248 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
10249 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
10250 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
10251 (avr-*-rtems*): Likewise.
10252 (bfin*-rtems*): Likewise.
10253 (moxie-*-rtems*): Likewise.
10254 (h8300-*-rtems*): Likewise.
10255 (i[34567]86-*-rtems*): Likewise.
10256 (lm32-*-rtems*): Likewise.
10257 (m32r-*-rtems*): Likewise.
10258 (m68k-*-rtems*): Likewise.
10259 (microblaze*-*-rtems*): Likewise.
10260 (mips*-*-rtems*): Likewise.
10261 (powerpc-*-rtems*): Likewise.
10262 (sh-*-rtems*): Likewise.
10263 (sparc-*-rtems*): Likewise.
10264 (sparc64-*-rtems*): Likewise.
10265 (v850-*-rtems*): Likewise.
10266 (m32c-*-rtems*): Likewise.
10268 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
10270 PR rtl-optimization/59511
10271 * ira.c (ira_init_register_move_cost): Use memory costs for some
10272 cases of register move cost calculations.
10273 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
10274 instead of BB frequency.
10275 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
10276 * lra-assigns.c (find_hard_regno_for): Ditto.
10278 2014-01-15 Richard Biener <rguenther@suse.de>
10280 PR tree-optimization/59822
10281 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
10282 (vectorizable_load): Use it to hoist defs of uses of invariant
10283 loads out of the loop.
10285 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
10286 Kugan Vivekanandarajah <kuganv@linaro.org>
10289 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
10292 2014-01-15 Richard Biener <rguenther@suse.de>
10294 PR rtl-optimization/59802
10295 * lcm.c (compute_available): Use inverted postorder to seed
10296 the initial worklist.
10298 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10301 * config/s390/s390.c (s390_preferred_reload_class): Don't return
10302 ADDR_REGS for invalid symrefs in non-PIC code.
10304 2014-01-15 Jakub Jelinek <jakub@redhat.com>
10307 * builtins.c (determine_block_size): Initialize *probable_max_size
10308 even if len_rtx is CONST_INT.
10310 2014-01-14 Andrew Pinski <apinski@cavium.com>
10312 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
10313 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
10314 (cortexa53_tunings): Likewise.
10315 (aarch64_sched_issue_rate): New function.
10316 (TARGET_SCHED_ISSUE_RATE): Define.
10318 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
10320 * ira-costs.c (find_costs_and_classes): Add missed
10321 ira_init_register_move_cost_if_necessary.
10323 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
10326 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
10328 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
10331 * config/i386/i386.c (type_natural_mode): Add a bool parameter
10332 to indicate if type is used for function return value. Warn ABI
10333 change if the vector mode isn't available for function return value.
10334 (ix86_function_arg_advance): Pass false to type_natural_mode.
10335 (ix86_function_arg): Likewise.
10336 (ix86_gimplify_va_arg): Likewise.
10337 (function_arg_32): Don't warn ABI change.
10338 (ix86_function_value): Pass true to type_natural_mode.
10339 (ix86_return_in_memory): Likewise.
10340 (ix86_struct_value_rtx): Removed.
10341 (TARGET_STRUCT_VALUE_RTX): Likewise.
10343 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
10345 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
10346 converting a conditional jump into a conditional return.
10348 2014-01-14 Richard Biener <rguenther@suse.de>
10350 PR tree-optimization/58921
10351 PR tree-optimization/59006
10352 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
10353 hoisting invariant stmts.
10354 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
10355 invariant loads on the preheader edge if possible.
10357 2014-01-14 Joey Ye <joey.ye@arm.com>
10359 * doc/plugin.texi (Building GCC plugins): Update to C++.
10361 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
10363 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
10364 (_mm_rcp28_round_ss): Ditto.
10365 (_mm_rsqrt28_round_sd): Ditto.
10366 (_mm_rsqrt28_round_ss): Ditto.
10367 (_mm_rcp28_sd): Ditto.
10368 (_mm_rcp28_ss): Ditto.
10369 (_mm_rsqrt28_sd): Ditto.
10370 (_mm_rsqrt28_ss): Ditto.
10371 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
10372 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
10373 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
10374 (IX86_BUILTIN_RCP28SD): Ditto.
10375 (IX86_BUILTIN_RCP28SS): Ditto.
10376 (IX86_BUILTIN_RSQRT28SD): Ditto.
10377 (IX86_BUILTIN_RSQRT28SS): Ditto.
10378 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
10379 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
10380 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
10381 (ix86_expand_special_args_builtin): Expand new FTYPE.
10382 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
10383 (srcp14<mode>): Make insn unary.
10384 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
10385 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
10386 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
10387 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
10388 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
10389 Fix rounding: make it SAE only.
10390 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
10392 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
10394 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
10395 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
10396 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
10397 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
10398 (round_saeonly_mask_scalar_operand4): Ditto.
10399 (round_saeonly_mask_scalar_op3): Ditto.
10400 (round_saeonly_mask_scalar_op4): Ditto.
10402 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10404 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
10405 Implement -maltivec=be for vec_insert and vec_extract.
10407 2014-01-10 DJ Delorie <dj@redhat.com>
10409 * config/msp430/msp430.md (call_internal): Don't allow memory
10410 references with SP as the base register.
10411 (call_value_internal): Likewise.
10412 * config/msp430/constraints.md (Yc): New. For memory references
10413 that don't use SP as a base register.
10415 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
10416 "an integer without a # prefix"
10417 * config/msp430/msp430.md (epilogue_helper): Use it.
10419 2014-01-13 Jakub Jelinek <jakub@redhat.com>
10422 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
10423 AVX512F gather builtins.
10424 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
10425 on gather decls with INTEGER_TYPE masktype.
10426 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
10427 directly into the builtin rather than hoisting it before loop.
10429 PR tree-optimization/59387
10430 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
10431 (scev_const_prop): If folded_casts and type has undefined overflow,
10432 use force_gimple_operand instead of force_gimple_operand_gsi and
10433 for each added stmt if it is assign with
10434 arith_code_with_undefined_signed_overflow, call
10435 rewrite_to_defined_overflow.
10436 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
10437 gimple-fold.h instead.
10438 (arith_code_with_undefined_signed_overflow,
10439 rewrite_to_defined_overflow): Moved to ...
10440 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
10441 rewrite_to_defined_overflow): ... here. No longer static.
10442 Include gimplify-me.h.
10443 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
10444 rewrite_to_defined_overflow): New prototypes.
10446 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10448 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
10450 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
10452 * builtins.c (get_object_alignment_2): Minor tweak.
10453 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
10455 2014-01-13 Christian Bruel <christian.bruel@st.com>
10457 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
10458 optimized non constant lengths.
10460 2014-01-13 Jakub Jelinek <jakub@redhat.com>
10463 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
10464 load as __atomic_load_N if possible.
10466 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
10468 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
10470 (rs6000_expand_builtin): Adjust call.
10472 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
10475 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
10476 * config/rs6000/rs6000.c: Include target-globals.h.
10477 (rs6000_set_current_function): Instead of doing target_reinit
10478 unconditionally, use save_target_globals_default_opts and
10479 restore_target_globals.
10481 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
10483 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
10484 (rs6000_expand_builtin): Handle mffs and mtfsf.
10485 (rs6000_init_builtins): Define mffs and mtfsf.
10486 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
10487 (rs6000_mffs): New pattern.
10488 (rs6000_mtfsf): New pattern.
10490 2014-01-11 Bin Cheng <bin.cheng@arm.com>
10492 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
10493 Start narrowing with START. Apply candidate-use pair
10494 and check overall cost in narrowing.
10495 (iv_ca_prune): Pass new argument.
10497 2014-01-10 Jeff Law <law@redhat.com>
10499 PR middle-end/59743
10500 * ree.c (combine_reaching_defs): Ensure the defining statement
10501 occurs before the extension when optimizing extensions with
10502 different source and destination hard registers.
10504 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
10507 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
10508 vtables into the type inheritance graph.
10510 2014-01-10 Jakub Jelinek <jakub@redhat.com>
10512 PR rtl-optimization/59754
10513 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
10514 modes in the REGNO != REGNO case.
10516 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10518 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
10520 2014-01-10 Jakub Jelinek <jakub@redhat.com>
10522 PR tree-optimization/59745
10523 * tree-predcom.c (tree_predictive_commoning_loop): Call
10524 free_affine_expand_cache if giving up because components is NULL.
10526 * target-globals.c (save_target_globals): Allocate < 4KB structs using
10527 GC in payload of target_globals struct instead of allocating them on
10528 the heap and the larger structs separately using GC.
10529 * target-globals.h (struct target_globals): Make regs, hard_regs,
10530 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
10531 of GTY((skip)) and change type to void *.
10532 (reset_target_globals): Cast loads from those fields to corresponding
10535 2014-01-10 Steve Ellcey <sellcey@mips.com>
10538 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
10539 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
10540 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
10542 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
10545 * aarch64-modes.def (CC_Zmode): New flags mode.
10546 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
10547 represents an equality.
10548 (aarch64_get_condition_code): Handle CC_Zmode.
10549 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
10551 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10553 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
10554 extraction in good case.
10556 2014-01-10 Richard Biener <rguenther@suse.de>
10558 PR tree-optimization/59374
10559 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
10560 checking after SLP discovery. Mark stmts not participating
10561 in any SLP instance properly.
10563 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10565 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
10566 when handling a SET rtx.
10568 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10570 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
10571 (cortex-a57): Likewise.
10572 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
10574 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10576 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
10577 non-iwmmxt builtins.
10579 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
10583 * ipa-devirt.c record_target_from_binfo): Take as argument
10584 stack of binfos and lookup matching one for virtual inheritance.
10585 (possible_polymorphic_call_targets_1): Update.
10587 2014-01-10 Huacai Chen <chenhc@lemote.com>
10589 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
10590 kernel strings for Loongson-2E/2F/3A.
10592 2014-01-10 Jakub Jelinek <jakub@redhat.com>
10594 PR middle-end/59670
10595 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
10596 is_gimple_call before calling gimple_call_internal_p.
10598 2014-01-09 Steve Ellcey <sellcey@mips.com>
10600 * Makefile.in (TREE_FLOW_H): Remove.
10601 (TREE_SSA_H): Add file names from tree-flow.h.
10602 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
10603 * tree.h: Remove tree-flow.h reference.
10604 * hash-table.h: Remove tree-flow.h reference.
10605 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
10606 reference with tree-ssa-loop.h.
10608 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10610 * doc/invoke.texi: Add -maltivec={be,le} options, and document
10611 default element-order behavior for -maltivec.
10612 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
10613 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
10614 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
10615 when targeting big endian, at least for now.
10616 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
10618 2014-01-09 Jakub Jelinek <jakub@redhat.com>
10620 PR middle-end/47735
10621 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
10622 var satisfies use_register_for_decl, just take into account type
10623 alignment, rather than decl alignment.
10625 PR tree-optimization/59622
10626 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
10627 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
10628 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
10629 Don't devirtualize for inplace at all. For targets.length () == 1,
10630 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
10632 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
10634 * config/i386/i386.md (cpu): Remove the unused btver1.
10636 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
10638 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
10640 2014-01-09 Jakub Jelinek <jakub@redhat.com>
10643 * tree-core.h (struct target_globals): New forward declaration.
10644 (struct tree_target_option): Add globals field.
10645 * tree.h (TREE_TARGET_GLOBALS): Define.
10646 (prepare_target_option_nodes_for_pch): New prototype.
10647 * target-globals.h (struct target_globals): Define even if
10648 !SWITCHABLE_TARGET.
10649 * tree.c (prepare_target_option_node_for_pch,
10650 prepare_target_option_nodes_for_pch): New functions.
10651 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
10652 * config/i386/i386.c: Include target-globals.h.
10653 (ix86_set_current_function): Instead of doing target_reinit
10654 unconditionally, use save_target_globals_default_opts and
10655 restore_target_globals.
10657 2014-01-09 Richard Biener <rguenther@suse.de>
10659 PR tree-optimization/59715
10660 * tree-cfg.h (split_critical_edges): Declare.
10661 * tree-cfg.c (split_critical_edges): Export.
10662 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
10664 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
10666 * cfgexpand.c (expand_stack_vars): Optionally disable
10667 asan stack protection.
10668 (expand_used_vars): Likewise.
10669 (partition_stack_vars): Likewise.
10670 * asan.c (asan_emit_stack_protection): Optionally disable
10671 after return stack usage.
10672 (instrument_derefs): Optionally disable memory access instrumentation.
10673 (instrument_builtin_call): Likewise.
10674 (instrument_strlen_call): Likewise.
10675 (asan_protect_global): Optionally disable global variables protection.
10676 * doc/invoke.texi: Added doc for new options.
10677 * params.def: Added new options.
10678 * params.h: Likewise.
10680 2014-01-09 Jakub Jelinek <jakub@redhat.com>
10682 PR rtl-optimization/59724
10683 * ifcvt.c (cond_exec_process_if_block): Don't call
10684 flow_find_head_matching_sequence with 0 longest_match.
10685 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
10686 non-active insns if !stop_after.
10687 (try_head_merge_bb): Revert 2014-01-07 changes.
10689 2014-01-08 Jeff Law <law@redhat.com>
10691 * ree.c (get_sub_rtx): New function, extracted from...
10692 (merge_def_and_ext): Here.
10693 (combine_reaching_defs): Use get_sub_rtx.
10695 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
10697 * cgraph.h (varpool_variable_node): Do not choke on null node.
10699 2014-01-08 Catherine Moore <clm@codesourcery.com>
10701 * config/mips/mips.md (simple_return): Attempt to use JRC
10703 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
10705 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
10707 PR rtl-optimization/59137
10708 * reorg.c (steal_delay_list_from_target): Call update_block for
10710 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
10712 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10714 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
10715 two duplicate entries.
10717 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
10720 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
10722 * config/mips/mips.c (mips_truncated_op_cost): New function.
10723 (mips_rtx_costs): Adjust test for BADDU.
10724 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
10726 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
10728 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
10729 (*baddu_si): ...this new pattern.
10731 2014-01-08 Jakub Jelinek <jakub@redhat.com>
10734 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
10736 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
10738 PR middle-end/57748
10739 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
10741 (expand_expr, expand_normal): Adjust.
10742 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
10743 inner_reference_p. Use inner_reference_p to expand inner references.
10744 (store_expr): Adjust.
10745 * cfgexpand.c (expand_call_stmt): Adjust.
10747 2014-01-08 Rong Xu <xur@google.com>
10749 * gcov-io.c (gcov_var): Move from gcov-io.h.
10750 (gcov_position): Ditto.
10751 (gcov_is_error): Ditto.
10752 (gcov_rewrite): Ditto.
10753 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
10754 only part to libgcc/libgcov.h.
10756 2014-01-08 Marek Polacek <polacek@redhat.com>
10758 PR middle-end/59669
10759 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
10761 2014-01-08 Marek Polacek <polacek@redhat.com>
10764 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
10766 2014-01-08 Jakub Jelinek <jakub@redhat.com>
10768 PR rtl-optimization/59649
10769 * stor-layout.c (get_mode_bounds): For BImode return
10770 0 and STORE_FLAG_VALUE.
10772 2014-01-08 Richard Biener <rguenther@suse.de>
10774 PR middle-end/59630
10775 * gimple.h (is_gimple_builtin_call): Remove.
10776 (gimple_builtin_call_types_compatible_p): New.
10777 (gimple_call_builtin_p): New overload.
10778 * gimple.c (is_gimple_builtin_call): Remove.
10779 (validate_call): Rename to ...
10780 (gimple_builtin_call_types_compatible_p): ... this and export. Also
10781 check return types.
10782 (validate_type): New static function.
10783 (gimple_call_builtin_p): New overload and adjust.
10784 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
10785 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
10786 (gimple_fold_stmt_to_constant_1): Likewise.
10787 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
10789 2014-01-08 Richard Biener <rguenther@suse.de>
10791 PR middle-end/59471
10792 * gimplify.c (gimplify_expr): Gimplify register-register type
10793 VIEW_CONVERT_EXPRs to separate stmts.
10795 2014-01-07 Jeff Law <law@redhat.com>
10797 PR middle-end/53623
10798 * ree.c (combine_set_extension): Handle case where source
10799 and destination registers in an extension insn are different.
10800 (combine_reaching_defs): Allow source and destination registers
10801 in extension to be different under limited circumstances.
10802 (add_removable_extension): Remove restriction that the
10803 source and destination registers in the extension are the same.
10804 (find_and_remove_re): Emit a copy from the extension's
10805 destination to its source after the defining insn if
10806 the source and destination registers are different.
10808 PR middle-end/59285
10809 * ifcvt.c (merge_if_block): If we are merging a block with more than
10810 one successor with a block with no successors, remove any BARRIER
10811 after the second block.
10813 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
10815 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
10817 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
10820 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
10821 for 14-bit register offsets when INT14_OK_STRICT is false.
10823 2014-01-07 Roland Stigge <stigge@antcom.de>
10824 Michael Meissner <meissner@linux.vnet.ibm.com>
10827 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
10828 Only check TFmode for SPE constants. Don't check TImode or TDmode.
10830 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
10832 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
10835 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
10837 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
10838 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
10839 rtx is const0_rtx or not.
10841 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
10844 * target-globals.c (save_target_globals): Remove this_fn_optab
10846 * toplev.c: Include optabs.h.
10847 (target_reinit): Temporarily restore the global options if another
10848 set of options are in force.
10850 2014-01-07 Jakub Jelinek <jakub@redhat.com>
10852 PR rtl-optimization/58668
10853 * cfgcleanup.c (flow_find_cross_jump): Don't count
10854 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
10855 to determine what is counted.
10856 (flow_find_head_matching_sequence): Use active_insn_p to determine
10858 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
10860 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
10861 determine what is counted.
10863 PR tree-optimization/59643
10864 * tree-predcom.c (split_data_refs_to_components): If one dr is
10865 read and one write, determine_offset fails and the write isn't
10866 in the bad component, just put the read into the bad component.
10868 2014-01-07 Mike Stump <mikestump@comcast.net>
10869 Jakub Jelinek <jakub@redhat.com>
10872 * tree-core.h (struct tree_optimization_option): Change optabs
10873 type from unsigned char * to void *.
10874 * optabs.c (init_tree_optimization_optabs): Adjust
10875 TREE_OPTIMIZATION_OPTABS initialization.
10877 2014-01-06 Jakub Jelinek <jakub@redhat.com>
10880 * config/i386/i386.h (struct machine_function): Add
10881 no_drap_save_restore field.
10882 * config/i386/i386.c (ix86_save_reg): Use
10883 !cfun->machine->no_drap_save_restore instead of
10884 crtl->stack_realign_needed.
10885 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
10886 this function clears frame_pointer_needed. Set
10887 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
10888 and DRAP reg is needed.
10890 2014-01-06 Marek Polacek <polacek@redhat.com>
10893 * doc/implement-c.texi: Mention that other integer types are
10894 permitted as bit-field types in strictly conforming mode.
10896 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
10898 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
10899 is newly allocated.
10901 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
10903 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
10905 2014-01-06 Martin Jambor <mjambor@suse.cz>
10908 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
10910 * ipa-prop.c (ipa_print_node_params): Fix indentation.
10912 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
10916 * var-tracking.c (add_stores): Preserve the value of the source even if
10917 we don't record the store.
10919 2014-01-06 Terry Guo <terry.guo@arm.com>
10921 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
10923 2014-01-05 Iain Sandoe <iain@codesourcery.com>
10926 * config/darwin.c (darwin_function_section): Adjust return values to
10927 correspond to optimisation changes made in r206070.
10929 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
10931 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
10932 from prefetch_block tune setting.
10933 (nocona_cost): Correct size of prefetch block to 64.
10935 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
10937 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
10938 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
10939 used to save the static chain register in the computation of the offset
10940 from which the FP registers need to be restored.
10942 2014-01-04 Jakub Jelinek <jakub@redhat.com>
10944 PR tree-optimization/59519
10945 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
10946 ICE if get_current_def (current_new_name) is already non-NULL, as long
10947 as it is a phi result of some other phi in *new_exit_bb that has
10950 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
10951 or vmovdqu* for misaligned_operand.
10952 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
10953 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
10954 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
10955 aligned_mem for AVX512F masked aligned load and store builtins and for
10956 non-temporal moves.
10958 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
10960 PR tree-optimization/59651
10961 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
10962 Address range for negative step should be added by TYPE_SIZE_UNIT.
10964 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
10966 * config/m68k/m68k.c (handle_move_double): Handle pushes with
10967 overlapping registers also for registers other than the stack pointer.
10969 2014-01-03 Marek Polacek <polacek@redhat.com>
10972 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
10975 2014-01-03 Jakub Jelinek <jakub@redhat.com>
10978 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
10981 * config/i386/i386.md (MODE_SIZE): New mode attribute.
10982 (push splitter): Use <P:MODE_SIZE> instead of
10983 GET_MODE_SIZE (<P:MODE>mode).
10984 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
10985 (mov -1, reg peephole2): Likewise.
10986 * config/i386/sse.md (*mov<mode>_internal,
10987 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
10988 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
10989 *<code><mode>3, *andnot<mode>3<mask_name>,
10990 <mask_codefor><code><mode>3<mask_name>): Likewise.
10991 * config/i386/subst.md (mask_mode512bit_condition,
10992 sd_mask_mode512bit_condition): Likewise.
10994 2014-01-02 Xinliang David Li <davidxl@google.com>
10996 PR tree-optimization/59303
10997 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
10998 (dump_predicates): Better output format.
10999 (pred_equal_p): New function.
11000 (is_neq_relop_p): Ditto.
11001 (is_neq_zero_form_p): Ditto.
11002 (pred_expr_equal_p): Ditto.
11003 (pred_neg_p): Ditto.
11004 (simplify_pred): Ditto.
11005 (simplify_preds_2): Ditto.
11006 (simplify_preds_3): Ditto.
11007 (simplify_preds_4): Ditto.
11008 (simplify_preds): Ditto.
11009 (push_pred): Ditto.
11010 (push_to_worklist): Ditto.
11011 (get_pred_info_from_cmp): Ditto.
11012 (is_degenerated_phi): Ditto.
11013 (normalize_one_pred_1): Ditto.
11014 (normalize_one_pred): Ditto.
11015 (normalize_one_pred_chain): Ditto.
11016 (normalize_preds): Ditto.
11017 (normalize_cond_1): Remove function.
11018 (normalize_cond): Ditto.
11019 (is_gcond_subset_of): Ditto.
11020 (is_subset_of_any): Ditto.
11021 (is_or_set_subset_of): Ditto.
11022 (is_and_set_subset_of): Ditto.
11023 (is_norm_cond_subset_of): Ditto.
11024 (pred_chain_length_cmp): Ditto.
11025 (convert_control_dep_chain_into_preds): Type change.
11026 (find_predicates): Ditto.
11027 (find_def_preds): Ditto.
11028 (destroy_predicates_vecs): Ditto.
11029 (find_matching_predicates_in_rest_chains): Ditto.
11030 (use_pred_not_overlap_with_undef_path_pred): Ditto.
11031 (is_pred_expr_subset): Ditto.
11032 (is_pred_chain_subset_of): Ditto.
11033 (is_included_in): Ditto.
11034 (is_superset_of): Ditto.
11036 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
11038 Update copyright years.
11040 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
11042 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
11043 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
11044 config/arc/arc.md, config/arc/arc.opt,
11045 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
11046 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
11047 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
11048 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
11049 config/linux-protos.h, config/linux.c, config/winnt-c.c,
11050 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
11051 vtable-verify.c, vtable-verify.h: Use the standard form for the
11054 2014-01-02 Tobias Burnus <burnus@net-b.de>
11056 * gcc.c (process_command): Update copyright notice dates.
11057 * gcov-dump.c: Ditto.
11059 * doc/cpp.texi: Bump @copying's copyright year.
11060 * doc/cppinternals.texi: Ditto.
11061 * doc/gcc.texi: Ditto.
11062 * doc/gccint.texi: Ditto.
11063 * doc/gcov.texi: Ditto.
11064 * doc/install.texi: Ditto.
11065 * doc/invoke.texi: Ditto.
11067 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
11069 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
11071 2014-01-01 Jakub Jelinek <jakub@redhat.com>
11073 * config/i386/sse.md (*mov<mode>_internal): Guard
11074 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
11076 PR rtl-optimization/59647
11077 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
11078 new_rtx into UNSIGNED_FLOAT rtxes.
11080 Copyright (C) 2014 Free Software Foundation, Inc.
11082 Copying and distribution of this file, with or without modification,
11083 are permitted in any medium without royalty provided the copyright
11084 notice and this notice are preserved.